Control and Safety Analysis of Intensified Chemical Processes [1 ed.] 9783527352623, 9783527843633

Citation preview

Control and Safety Analysis of Intensified Chemical Processes

Control and Safety Analysis of Intensified Chemical Processes Edited by Dipesh Shikchand Patle and Gade Pandu Rangaiah

Editors

Motilal Nehru National Institute of Technology MNNIT Allahabad Campus Allahabad, Prayagraj 211004, India

All books published by WILEY-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.

Dr. Gade Pandu Rangaiah

Library of Congress Card No.: applied for

Dr. Dipesh Shikchand Patle

National University of Singapore Engineering Drive 4 Singapore SN, 117576 Cover Image: Courtesy of Dr. Dipesh

Shikchand Patle, arhendrix/Shutterstock

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at . © 2024 WILEY-VCH GmbH, Boschstraße 12, 69469 Weinheim, Germany All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Print ISBN: 978-3-527-35262-3 ePDF ISBN: 978-3-527-84363-3 ePub ISBN: 978-3-527-84364-0 oBook ISBN: 978-3-527-84365-7 Typesetting

Straive, Chennai, India

To my father (Shikchand Patle) and mother (Sushila Patle), who made me what I am today and To my dearest vivacious wife “Dipa” and lovely daughter “Meha,” for their love, selfless devotion, support, and encouragement. Dipesh Shikchand Patle

To my mother (Gade Sakuntala Devi) and father (Gade Gopala Swami), who made me what I am today and To my dearest wife (Krishna Kumari) for her selfless devotion to supporting and strengthening our family, my academic career, and my contributions. Gade Pandu Rangaiah

vii

Contents Preface xv

Part I 1 1.1 1.2 1.3 1.4 1.5

2

2.1 2.2 2.3 2.4 2.5 2.6 2.6.1 2.6.2 2.6.3 2.6.4 2.6.5 2.7 2.8 2.9

Overview and Background 1

Introduction 3 Dipesh Shikchand Patle and Gade Pandu Rangaiah Process Intensification 3 Need for Control and Safety Analysis of Intensified Chemical Processes 5 Studies on Control and Safety Analysis of Intensified Chemical Processes 7 Scope and Organization of the Book 9 Conclusions 12 References 13 Applications and Potential of Process Intensification in Chemical Process Industries 15 Chirla C.S. Reddy Introduction 15 Benefits of Process Intensification Techniques 16 Static Mixers 17 Process Intensification for Separation Vessels 18 Process Intensification for Distillation 21 Process Intensification for Heating 24 Steam Injection Heater 24 Steam/Electric Heaters as a Replacement for Fired Heaters 25 Process Intensification for Flue Gas Heat Recovery 26 Process Heat Exchangers 26 Sonic Horn 27 Steam Compression 27 Process Intensification for Carbon Capture 30 Process Intensification for Vacuum Systems 31

viii

Contents

2.10 2.11 2.12 2.12.1 2.12.2 2.13 2.14 2.15

Process Intensification for Water Deaeration 33 Process Intensification for Development of Inherently Safer Design (ISD) 33 Process Intensification for Reducing Pressure Relief and Handling Requirements 35 Non-safety Instrumented Solutions for Pressure Relief Systems 37 Safety Instrumented System (SIS) Solutions for Reducing Pressure Relief Requirements 39 Process Intensification for Wastewater Recovery 41 Challenges of Process Intensification Techniques 43 Conclusions 44 References 45

Part II

3

3.1 3.2 3.2.1 3.2.2 3.2.3 3.2.4 3.3 3.3.1 3.3.2 3.3.3 3.3.4 3.3.5 3.3.6 3.4 3.4.1 3.4.2 3.4.2.1 3.4.2.2 3.5 3.6 3.6.1 3.6.2

Procedures and Software for Simulation, Control and Safety Analysis 47

Simulation and Optimization of Intensified Chemical Processes 49 Zemin Feng and Gade Pandu Rangaiah Introduction 49 Simulation of Chemical Processes 50 Usefulness of Process Simulation 50 Commercial Process Simulators 52 Free Process Simulators 53 Computational Methods for Process Simulation 53 Procedure for Simulation of (Intensified) Chemical Processes 56 Problem Analysis 56 Basic Process Flow Design 57 Process Intensification and Integration 57 Model Construction 57 Simulation and Convergence 59 Results Analysis 59 Optimization of (Intensified) Chemical Processes 59 Mathematical Optimization Methods 59 Optimization of Chemical Processes with a Process Simulator 62 Optimization Using MATLAB 62 Optimization Using Python 63 Challenges in the Simulation/Optimization of Intensified Chemical Processes 65 Case Study 66 Problem Analysis 66 Process Flow Design 67

Contents

3.6.3 3.6.4 3.6.4.1 3.6.4.2 3.6.4.3 3.6.5 3.7

Model Construction 69 Simulation and Convergence 70 Process Simulation 70 Economic Evaluation Criterion 71 Process Optimization 73 Results and Analysis 75 Conclusions 78 References 79

4

Dynamic Simulation and Control of Intensified Chemical Processes 83 Zemin Feng and Gade Pandu Rangaiah Introduction 83 Dynamic Simulation of Chemical Processes 84 Understanding Dynamic Simulation 84 Applications of Dynamic Simulation 87 Dynamic Simulation Software 88 Dynamic Simulation and Control Procedure 91 Dynamic Simulation and Control of Intensified Chemical Processes 98 Challenges Due to Process Intensification 100 Process Control 100 Controlled, Manipulated, and Disturbance Variables 101 Typical Control Loop 101 Control Degrees of Freedom 101 Case Study 102 Steady-state Simulation and Optimization 103 Preparation/Initialization for Dynamic Simulation 103 Control Structure Design 107 Composition Control Scheme 108 Temperature Control Scheme 110 Tuning of Controller Parameters 112 Analysis of Dynamic Simulation Results 112 Conclusions 120 References 121

4.1 4.2 4.2.1 4.2.2 4.2.3 4.3 4.4 4.4.1 4.5 4.5.1 4.5.2 4.5.3 4.6 4.6.1 4.6.2 4.6.3 4.6.3.1 4.6.3.2 4.6.4 4.6.5 4.7

5 5.1 5.2 5.2.1 5.2.1.1 5.2.1.2 5.2.1.3

Safety Analysis of Intensified Chemical Processes 125 Masrina Mohd Nadzir, Zainal Ahmad, and Syamsul Rizal Abd Shukor Introduction 125 Safety Analysis in Chemical Process Industry 126 Safety Analysis Tools 128 Hazard Identification 128 Risk Assessment 130 Inherently Safer Design (ISD) 131

ix

x

Contents

5.2.1.4 5.2.1.5 5.2.1.6 5.2.1.7 5.3 5.3.1 5.3.2 5.3.2.1 5.3.2.2 5.3.3 5.4 5.5 5.5.1 5.5.2 5.5.3 5.5.4 5.5.5 5.6 5.7

Safety Instrumented Systems 132 Human Factors and Safety Culture 132 Regulatory Framework and Compliance 134 Monitoring and Continuous Improvement 135 Process Intensification and Safety Analysis 136 Impacts of Process Intensification on Safety 136 Safety Analysis in Intensified Process Design 137 Hazard Identification Techniques for Process Intensification Technologies 138 Risk Assessment for Process Intensification Technologies 140 Inherently Safer Design Principles Intensified Processes 141 Safety Management Systems for Intensified Processes 144 Safety Training and Competency for Intensified Processes 146 Importance of Safety Training and Competency 146 Developing Safety Training and Competency Programs 147 Utilizing a Blended Learning Approach 148 Assessing Training Effectiveness and Continual Improvement 148 Benefits of Effective Safety Training and Competency Management 148 Case Studies of Safety Analysis in Intensified Processes 149 Conclusions 151 References 151

Part III Control and Safety Analysis of Intensified Chemical Processes 155 6

6.1 6.2 6.3 6.4 6.5 6.6 6.6.1 6.6.2 6.6.3 6.6.4 6.7 6.8

Control of Hybrid Reactive–Extractive Distillation Systems for Ternary Azeotropic Mixtures 157 Zong Yang Kong and Hao-Yeh Lee Introduction 157 Steady-state Design of the RED 160 Dynamic Simulation Setup 161 Inventory Control Setup 162 Sensitivity Analysis 163 Quality Control Structures 165 Control Structure 1 (CS 1) – Simple Temperature Control 165 Control Structure 2 (CS 2) – Triple Point Temperature Control 168 Control Structure 3 (CS 3) – Triple Point Temperature Control Using SVD Analysis 170 Feedforward Control Structure 3 (FF-CS 3) 172 Control Performance Evaluation 177 Conclusions 178 Acknowledgements 179 Acronyms 179 Nomenclature 180 References 180

Contents

7

7.1 7.2 7.3 7.4 7.4.1 7.4.2 7.4.3 7.4.4 7.5

8

8.1 8.2 8.2.1 8.2.1.1 8.2.2 8.3 8.3.1 8.4 8.4.1 8.4.2 8.4.3 8.5 8.5.1 8.5.1.1 8.5.2 8.5.3 8.5.4 8.5.4.1 8.5.5 8.6

Process Design and Control of Reactive Distillation in Recycle Systems 183 Mihai Daniel Moraru, Costin Sorin Bildea, and Anton Alexandru Kiss Introduction 183 Design of Reactive Distillation Processes 184 Control of Reactive Distillation Processes 188 Case Study: RD Coupled with a Distillation–Reactor System and Recycle 192 Basis of Design and Basic Data 192 Process Design 198 Process Control 201 Discussion 204 Conclusions 204 References 205

Dynamics and Control of Middle-vessel Batch Distillation with Vapor Recompression 209 Radhika Gandu, Akash Burolia, Dipesh Shikchand Patle, and Gara Uday Bhaskar Babu Introduction 209 Conventional Middle-vessel Batch Distillation 211 A Systematic Simulation Approach of CMVBD 212 Model Equations 213 Constant Composition Control 216 Single-stage Vapor Recompression in Middle-vessel Batch Distillation 216 A Systematic Simulation Approach of SiVRMVBD 216 Performance Specifications 218 Energy Savings 218 Total Annual Cost 218 Greenhouse Gas Emissions 219 Results and Discussion 219 Conventional Middle-vessel Batch Distillation Column 219 Dynamic Composition Profiles 219 Single-stage Vapor Recompression in Middle-vessel Batch Distillation 222 Energetic, Economic, and Environmental Performance: CMVBD vs. SiVRMVBD 225 Constant Composition Control 226 SiVRMVBD-GSPI 229 Energetic, Economic, and Environmental Performance: CMVBD vs. Controlled CMVBD and SiVRMVBD 232 Conclusions 234 References 234

xi

xii

Contents

9

9.1 9.2 9.3 9.3.1 9.3.2 9.3.3 9.4 9.5 9.5.1 9.5.2 9.5.3 9.5.4 9.6 9.7 9.8

10

10.1 10.2 10.3 10.3.1 10.3.2 10.3.3 10.4 10.4.1 10.4.2 10.4.3 10.4.4 10.5 10.5.1 10.5.2 10.5.3 10.5.4

Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices 237 Savyasachi Shrikhande, Gunawant K. Deshpande, Gade Pandu Rangaiah, and Dipesh Shikchand Patle Introduction 237 Safety Indices for Process Safety Assessment 239 Description of Distillation Systems 241 Conventional Sequence of Columns 241 Dividing-Wall Column 241 Dividing-Wall Column with Mechanical Vapor Recompression 243 Selection of Safety Indices 244 Results and Discussion 245 Conventional Sequence of Columns 245 Dividing-Wall Column 251 Dividing-Wall Column with Mechanical Vapor Recompression 253 Comparative Analysis 255 Survey of Engineers and Discussion of their Responses 257 Improved PRI 262 Conclusions 263 Acknowledgments 263 References 264

Dynamic Safety Analysis of Intensified Extractive Distillation Processes with Independent Protection Layers 269 Chengtian Cui and Meng Qi Introduction 269 Preliminary 271 Process Studied 272 Process Intensification Measures 272 Steady-state Process Design 273 Process Intensification Analysis 274 Dynamics and Control 276 Control Basis 276 BPCS #1 279 BPCS #2 279 BPCS #3 282 Safety Analysis 284 Process #1 Safety Analysis 285 Process #2 Safety Analysis 286 Process #3 Safety Analysis 288 Dynamic Safety Analysis of Process #3 with IPLs 289

Contents

10.6

Conclusions 292 Acknowledgments 293 References 293

11

Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts 295 Juan G. Segovia-Hernández, César Ramírez-Márquez, Gabriel Contreras-Zarazúa, Eduardo Sánchez-Ramírez, and Juan J. Quiroz-Ramírez Introduction 295 Methodology 302 Control Behavior Analysis 306 Singular Value Decomposition 306 Methyl Ethyl Ketone 307 Methyl Ethyl Ketone Production Through a Conventional Process 308 MEK Production from Non-renewable Sources 308 Purification of MEK Through Process-Intensified Schemes 308 Intensification of Alcohol-to-Jet Fuel Process 313 Process Modeling and Optimization 314 Results 316 New Processes for Furfural and Co-products 318 Results 321 Lactic Acid 324 Lactic Acid Production by Reactive Distillation 325 Design and Synthesis of Intensified Processes 326 Optimization 326 Results and Discussion 327 Future and Perspectives 329 Conclusions 329 Acknowledgments 330 References 330

11.1 11.2 11.2.1 11.2.1.1 11.3 11.3.1 11.3.1.1 11.3.2 11.4 11.4.1 11.4.2 11.5 11.5.1 11.6 11.6.1 11.6.2 11.6.3 11.6.4 11.7 11.8

12

12.1 12.2 12.2.1 12.2.2 12.3 12.3.1 12.3.1.1 12.3.1.2

Analysis of Safety and Economic Objectives for Intensified Algal Biodiesel Process 335 Gunavant Deshpande, Ashish N. Sawarkar, and Dipesh Shikchand Patle Introduction 335 Process Development 337 Process Development of Alternative 1 337 Process Development of Alternative 2 340 Multi-Objective Optimization 342 Objective Functions 344 Break-Even Cost 344 Individual Risk (IR) 345

xiii

xiv

Contents

12.3.2 12.4 12.4.1 12.4.2 12.5 12.6

Simple Additive Weighting (SAW) Method 347 Results and Discussion 347 Minimization of BEC and IR for Alternative 1 348 Minimization of BEC and IR for Alternative 2 350 Comparative Analysis 352 Conclusions 353 References 354 Index 359

xv

Preface Process intensification refers to the use of novel equipment and/or methods to substantially improve the performance of a process or unit. It results in a reduced inventory of hazardous chemicals/materials, leading to less risk. Also, it can result in a significantly smaller footprint of the process and reduced consumption of utilities while improving productivity. Process intensification has significant potential for both new and existing plants. Many chemical plants, designed decades ago, are in operation across the world, and there is a strong need to improve their performance. They might have been designed optimally for economic, technological, and societal constraints; however, owing to technological advances, stricter environmental regulations, and demanding market scenarios, existing chemical processes need to be retrofitted/revamped, perhaps utilizing process intensification, to improve their performance. An intensified process may have a higher degree of difficulty pertaining to its controllability and safety. For instance, intensified processes may require higher energy inputs and/or higher operating temperatures or may involve several phenomena simultaneously, warranting a more complex control system. In addition, process intensification may lead to reduced control degrees of freedom, which can make the job of control engineers more challenging. Control and safety studies on intensified chemical processes generally require computational tools and process simulators, as performing them on a live plant is risky. Nonetheless, appropriate models have to be developed. Hence, it is essential to study the intricacies imparted in a chemical process as a result of intensification strategies so as to devise a suitable control structure, leading to safer operations. Fortunately, better technical alternatives (such as more efficient process equipment) as well as effective simulation/optimization techniques are available today that can be employed to achieve the goals of the study. A literature review suggests that there has been a steady increase in research on the control and/or safety of intensified chemical processes since the year 2000; in particular, the increase has been significant in recent years. However, there is no book that specifically focuses on control and safety analysis of intensified chemical processes, though there are books with one or two chapters on these aspects. This book in your hands addresses this gap. Its chapters are contributed by active and leading researchers in the field from India, Malaysia, Mexico, Netherlands, People’s Republic of China, Romania, Singapore, Taiwan, United Kingdom, and

xvi

Preface

United States of America. All chapters, the first submission as well as the revision, were peer-reviewed anonymously by at least two experts. For the benefit of readers with diverse backgrounds, this book is organized into three parts. Chapters 1 and 2 in Part I provide an overview of process intensification, its applications and developments, the need for control and safety analysis, and the principles, potential, and challenges of intensified chemical processes. Part II (Chapters 3, 4, and 5) focuses on an overview of simulation and optimization methods, common programs for simulation and optimization, interfacing of simulators and optimizers, an overview of dynamic simulation and control, programs for dynamic simulation and control, tuning of controllers, criteria for control assessment, an overview of safety analysis, and methodologies and tools for safety analysis. In Part III, Chapters 6–12 present studies on control and safety analysis of important chemical processes that are intensified using hybrid reactive-extractive distillation, reactive distillation, middle-vessel batch distillation with vapor recompression, intensified extractive distillation column, dividing-wall column, and dividing-wall column with mechanical vapor recompression, among others. This book provides chemical engineering researchers, practitioners, and students with process intensification techniques and their effects on process controllability and safety. Its contents can be readily adopted as part of special courses on process systems engineering, process design and integration, plantwide control of chemical processes, and safety and hazards analysis of chemical processes, aimed at undergraduate/postgraduate students. Manufacturers are likely to continue to intensify the existing processes further in view of the ever-increasing competition and strict regulations. This book is expected to provide them with a ready understanding of the effects of intensification techniques in terms of control and safety, both of which are of paramount importance. In essence, this book will be a useful resource for researchers working in process systems engineering, for students studying related courses, and for practitioners interested in process intensification, control, and safety. It is also expected that this book will contribute to further developments and improvements in the existing state of the art. At this stage, we are glad that this book has materialized in the year 2024 within two years since its conceptualization. We express our deep sense of gratitude to all the contributors and reviewers for their efforts and cooperation in bringing the book to its fruitful conclusion. We thank Wiley for agreeing to publish this book and helping to disseminate the knowledge. We are grateful to Motilal Nehru National Institute of Technology Allahabad (India) and the National University of Singapore (Singapore) for providing us with the required facilities/resources for research and the freedom to publish this book. Finally, we acknowledge the direct and indirect support of our family members, without which this book would not have materialized so smoothly. September 2023

Dipesh Shikchand Patle Motilal Nehru National Institute of Technology Allahabad, Prayagraj, India Gade Pandu Rangaiah National University of Singapore, Singapore

1

Part I Overview and Background

3

1 Introduction Dipesh Shikchand Patle 1 and Gade Pandu Rangaiah 2,3 1 Department Chemical Engineering, Motilal Nehru National Institute of Technology, Allahabad, 211004, Uttar Pradesh, India 2 Department of Chemical & Biomolecular Engineering, National University of Singapore, Singapore, 117585, Singapore 3 School of Chemical Engineering, Vellore Institute of Technology, Vellore, 632014, Tamil Nadu, India

1.1 Process Intensification In recent years, process intensification has been one of the most active areas of research in chemical engineering. It refers to the use of novel equipment and/or methods to drastically improve the performance of a process or unit. Process intensification could result in a substantial reduction in equipment size, a reduction in energy consumption, an increase in product yield, a reduction in waste, or anything that ultimately leads to economical and sustainable technologies (Stankiewicz and Moulijn, 2000). It is dissimilar to merely scaling up the plant to make it more economical. One of the most important features of process intensification is that the changes it brings are drastic in nature and are revolutionary rather than evolutionary (Stankiewicz and Drinkenburg, 2004). The advent of process intensification in the chemical engineering field was marked in 1983, when the very first paper was published on the application of centrifugal fields (i.e. “HiGee,” also known as rotating packed bed) in distillation processes (Ramshaw, 1983). In the 1980s, process intensification was mainly referred to as a substantial reduction in the size of the equipment (Stankiewicz and Drinkenburg, 2004). The first definition of process intensification by Ramshaw is “devising an exceedingly compact plant which reduces both the ‘main plant item’ and the installations costs” (Ramshaw, 1983). The definition of process intensification has changed, and now it is not limited only to the substantial reduction in size. Currently, process intensification refers to novel equipment and techniques that lead to inexpensive and sustainable processes. At present, conventional methods are becoming unsustainable because of the stricter environmental policies and agreements across the world. One of the definitions of process intensification that is deemed suitable in the current era is “Any chemical engineering development that leads to substantially smaller, cleaner, safer Control and Safety Analysis of Intensified Chemical Processes, First Edition. Edited by Dipesh Shikchand Patle and Gade Pandu Rangaiah. © 2024 WILEY-VCH GmbH. Published 2024 by WILEY-VCH GmbH.

4

1 Introduction

and more energy efficient technology or that combine[s] multiple operations into fewer devices (or a single apparatus)” (Baldea, 2015). Gerven and Stankiewicz (2009) reported four main goals of process intensification: 1. 2. 3. 4.

Maximize the effectiveness of intermolecular and intramolecular events. Deliver the same physical and chemical environment to the molecules. Optimize the driving forces of the process and increase the specific surface area. Maximize the synergy of partial processes.

The four approaches to achieve the above goals are reported as follows (Gerven and Stankiewicz, 2009): 1. The first approach is structure (spatial domain); it provides different explicit spatial structuring into different equipment or reactant assemblies, which can achieve the above goals. For example, static mixers with a specific structure, heterogeneous catalysts with a specific structure/shape, etc. to achieve goals 1, 2, and 3. 2. The second approach is energy (thermodynamic domain), which provides efficient energy distribution and utilization of different forms of energy available to attain the above goals. For example, ultrasound/microwave-assisted reactors where all molecules experience the same physical and/or chemical environment to achieve goals 1, 2, and 3). 3. The third approach is synergy (i.e. functional domain), in which various phenomena function synergically to achieve goals 1, 3, and 4. An example of this is reactive distillation (RD). 4. The fourth approach is time (i.e. temporal domain), which provides control over time scales. For example, an oscillatory baffle flow reactor facilitates controlled energy input. Stankiewicz and Drinkenburg (2004) presented a toolbox to achieve process intensification. It includes equipment and methods for process intensification. The former dealt with the change in hardware to achieve intensification, whereas the latter dealt with the change in the way a process is carried out (software) to achieve intensification. Currently, climate change is one of the major concerns, and chemical/process industries are considerably responsible for it. These industries continue to contribute to the disruption of the environment, biosphere, biochemical flow, and land systems (Steffen et al., 2015). These have prompted the formation of several regulations, international/national treaties, procedures, and consumer preferences that seek sustainable alternatives. The reduction of greenhouse gasses’ emissions, addressal of resource paucity, and revisiting material management are among many efforts that may enhance sustainability. These sustainable changes can be achieved through integrated innovation, which is interrelated with process intensification (López-Guajardo et al., 2022).

1.2 Need for Control and Safety Analysis of Intensified Chemical Processes

Process intensification is well recognized in current trends because of its synergy with Industry 4.0 and circular chemistry (López-Guajardo et al., 2022). The objective of circular chemistry is to (re)investigate the better use of resources by decreasing the consumption of the resources, improving efficiency, increasing the life of products, and maintaining them in the production cycle (Keijer et al., 2019). Industry 4.0, conceptualized in the year 2011, revolutionizes the way different process systems function within an integrated framework. This is related to the application of new and advanced technological structures, for example, the Internet of Things (IoT), big data and analytics, artificial intelligence (AI), cloud computing, and cybersecurity (Sharma et al., 2021). The technological advancements have modernized conventional industrial operations, procedures, and instruments (Canas et al., 2021), leading to improvement in the overall process of circularity (circular economy) by means of adopting sustainable practices. Thus, a clear overlap is present between the benefits of Industry 4.0 and the main objectives of process intensification. To achieve sustainable development, López-Guajardo et al. (2022) suggested a process intensification 4.0 strategy, which is the confluence of Industry 4.0, circular chemistry, and process intensification.

1.2 Need for Control and Safety Analysis of Intensified Chemical Processes Despite the numerous benefits of process intensification, the controllability and safety of the intensified processes must be evaluated, as intensification may introduce certain specific hazards into the process. In the drive towards newer and/or better processes, industries should ensure that new hazards are not created on account of process intensification. Potential problems due to process intensification include the following (Etchells, 2005): ●

●

●

Some process intensification technologies require excessive electricity inputs (e.g. for microwaves, excessive voltages, or electromagnetic radiation) or must be operated at higher temperatures and/or pressures. The high-energy sources may introduce new hazards that have to be considered when applied to hazardous substances, e.g. whether it is safe or not to use microwaves on thermally unstable substances or mixtures. Control and safety of intensified systems can be challenging. Although process intensification may reduce hazards (e.g. by substantially reducing the material inventory), control degrees of freedom may be reduced due to intensification. This reduction may affect the controllability of the process. Due to better mixing, intensified reactors have the potential to dramatically increase reaction rates. In comparison to conventional reactors, this might result in a substantially higher rate of energy release in some cases, and it might also, in some situations, affect the reaction chemistry. If the reaction has not been

5

6

1 Introduction

●

●

●

●

properly evaluated before, it could have serious safety implications (e.g. if the increased reaction produces a gas rather than a liquid). While there are tested and proven methods for determining the anticipated reaction thermochemistry in conventional reactors, these methods are less reliable for some recent types of intensified equipment (Etchells, 2005). Rotating equipment, which may be introduced on account of process intensification, may not be appropriate for materials that are friction sensitive, or those that could ignite or explode under friction. Hybrid or integrated intensified systems will invariably have many smaller parts with more connections. This will increase the chances of seal leakages due to the greater number of joints. Heat integration with mechanical vapor recompression may result in potential fire and explosion hazards if the gas compressor seal fails. The short residence time in intensified processes requires control instruments that respond rapidly and detect deviations (Klais et al., 2009).

The abovementioned factors require a careful analysis of the controllability and safety of the intensified process, without which the intensified process, though efficient, may become infeasible to operate. At various stages of design and operation, the risk of accidents has to be investigated. Chemical process safety can be broadly classified into four tiers of protection, as presented in Figure 1.1 (Ebrahimi

Risk due to potential incidents

Inherent safety layer of protection

Remaining risk

Passive safety layer of protection

Remaining risk

Active safety layer of protection Remaining risk Procedural safety layer of protection (Operational stage) Actual risk

Figure 1.1

Layers of protection for a chemical process and their effect on reducing the risk.

1.3 Studies on Control and Safety Analysis of Intensified Chemical Processes

et al., 2012). Even if the inherent safety layer is frequently the most effective one, it is only logical to have additional layers of protection. It is rarely probable that all these levels will fail simultaneously.

1.3 Studies on Control and Safety Analysis of Intensified Chemical Processes In this section, we analyze the studies on the control and safety of intensified chemical processes in the last two decades. First, the Scopus database was searched on 7 September 2023, for articles containing “intensification or intensified” in their title, abstract, and/or keywords; this search resulted in a total of 142,492 articles in all subject areas, without any limitations. Out of these entries, 8470 articles ( journal papers, books, and book chapters, while excluding conference papers, review articles, notes, editorials, etc.) in all languages are in the chemical engineering field from the year 2000 to 2022. These include some articles unrelated to process intensification; on the other hand, they do not include some articles related to process integration, such as those on improving the energy efficiency of distillation systems. Overall, 8470 is a reasonable estimate of articles related to process intensification. Figure 1.2, based on these articles, clearly indicates that there has been an increasing interest and research in process intensification in chemical engineering. As evident from the sharp rise in the number of published articles from the year 2016 to 2022, process intensification has received significant interest from researchers in recent years. Subsequently, to understand the research interest and contributions on control and/or safety of intensified chemical processes, the Scopus database was searched for articles containing “intensification OR intensified” AND “control OR safety” in their title, abstract, and/or keywords; the rest of the criteria are as mentioned in the above paragraph. This resulted in 1148 articles, which are assumed to cover the control and/or safety of intensified chemical processes. The search results (Figure 1.3)

Number of articles

1200 1000 800 600 400 200 0 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2022 Year

Figure 1.2 Number of articles ( journal research articles, books, and book chapters) with intensification OR intensified in the article title, abstract, and/or keywords, published in all languages in the chemical engineering field, in each year from the year 2000 to 2022.

7

1 Introduction

Number of Articles

160 140 120 100 80 60 40 20 0 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2022 Year

Figure 1.3 Number of articles ( journal papers, books, and book chapters) annually published from the year 2000 to 2022 on control and/or safety of intensified chemical processes.

confirm that there has been a gradual and steady increase in the research on the control and/or safety of intensified chemical processes since the year 2000; particularly, the increase has been significant in recent years except for a dip in the year 2020. The 1148 articles on control and/or safety of intensified chemical processes were published in more than 150 journals or book chapters from the year 2000 to 2022. Of these articles, 92.9% are research articles, 6.1% are book chapters, and 1% are books. The top 10 journals that published a higher number of these articles are presented in Figure 1.4. Among them, the maximum number of articles has appeared in the Chemical Engineering and Processing: Process Intensification journal, which is

AlChE Journal Separation and Purification Technology Chemical Engineering Transactions Computers and Chemical Engineering

Journals

8

Chemical Engineering Research and Design Computer Aided Chemical Engineering Chemical Engineering Science Chemical Engineering Journal Industrial and Engineering Chemistry Research Chemical Engineering and Processing Process Intensification 0

20

40

60

80

100

120

Number of articles

Figure 1.4 The journals that published the highest number of articles on control and/or safety of intensified chemical processes from the year 2000 to 2022.

Countries

1.4 Scope and Organization of the Book

Russian Federation Netherlands Italy India Mexico France United Kingdom Germany United States China 0

30

60

90

120 150 180 210 240 270 Number of articles

Authors

Figure 1.5 The countries that contributed the highest number of articles on control and/or safety of intensified chemical processes from the year 2000 to 2022. Baldea, M. Rong, B.G. Hessel, V. Errico, M. Drioli, E. Pistikopoulos, E.N. Hernández, S. Contreras-Zarazúa, G. Kiss, A.A. Sánchez-Ramírez, E. Ramírez-Márquez, C. Segovia-Hernández, J.G. 0

10

20 30 Number of articles

40

50

Figure 1.6 Researchers, who published 10 or more articles on control and/or safety of intensified chemical processes from the year 2000 to 2022.

not surprising. As shown in Figure 1.5, the largest number of articles on control/ safety analysis of intensified processes have come from China, followed by the United States, Germany, the United Kingdom, and others. Finally, Figure 1.6 lists the 12 active researchers, each of whom has contributed 10 or more articles; in particular, Prof. J.G. Segovia-Hernández is the most active researcher with 45 articles. The editors are pleased that five of these active researchers have contributed chapters to this book.

1.4 Scope and Organization of the Book For better understanding and for the convenience of novice readers, researchers, as well as practitioners, this book is divided into three parts. Part I (i.e. this chapter

9

10

1 Introduction

and Chapter 2) provides an overview of process intensification, its applications and developments, need for control and safety analysis, and industrial applications of process intensification. Chapter 2 will be of interest to all who are interested in process intensification. Chapters 3–5 in Part II focus on providing the basics and overview of steady-state simulation, optimization methods, programs and simulators, dynamic simulation and control, safety analysis, and popular methodologies and tools for safety analysis. These chapters will be of interest to young researchers, who may be new to control and safety field. Lastly, Chapters 6–12 in Part III deal with studies on control and safety analysis of many intensified chemical processes, including hybrid reactive-extractive distillation, RD, middle vessel batch distillation with vapor recompression, intensified extractive distillation column, dividing wall column (DWC), and DWC with mechanical vapor recompression. These advanced chapters in Part III will be of interest to researchers and practitioners working in the field. The scope of each chapter is outlined as follows. Chapter 2, prepared by a practitioner with extensive industrial experience and keen interest in the latest developments, presents an overview of many process intensification techniques employed in industrial practice and discusses their relative merits and associated practical challenges. This chapter provides a brief and handy reference relating to process intensification for practicing engineers, researchers, and students. Two experienced academicians in process simulation, optimization, and control prepared both Chapters 3 and 4. The general approach to process simulation, common commercial process simulators, and free process simulation software are outlined in Chapter 3. This is very useful for novice readers to refresh the basics and find a suitable process simulation platform for their applications. Then, the main steps in the simulation of intensified chemical processes are presented. Next, popular approaches for single- and multi-objective optimization of chemical processes and the linking of common optimizers with a commercial process simulator are described. Also, challenges in the simulation and optimization of intensified chemical processes are outlined. Finally, process simulation and optimization of intensified chemical processes are illustrated, taking DWC as an example. Chapter 4 on dynamic simulation is important to understand and evaluate the transient behavior of chemical processes as well as the performance of the associated control system. Dynamic simulation is more complex and computationally challenging than steady-state simulation. Chapter 4 first introduces the principles and applications of dynamic simulation of chemical processes as well as the common software for dynamic simulation. Then, a detailed procedure involving five steps for dynamic simulation and control of intensified chemical processes is presented. Next, variables related to process control, typical control loop, and control degrees of freedom are outlined. Subsequently, the need for and difficulties in the dynamic simulation and control of intensified chemical processes are summarized. Finally, a case study on dynamic simulation and control of an intensified chemical process, namely, extractive DWC, is presented and discussed. Chapter 5, prepared by three researchers with experience in chemical process safety and control, addresses the tools and procedures for detecting the positive and

1.4 Scope and Organization of the Book

negative effects of process intensification on process safety. It examines the potential impacts of process intensification on process safety, considering positive effects such as reduced inventories of hazardous materials and lower energy consumption as well as negative effects such as increased process complexity and new safety concerns associated with novel technologies. Also, Chapter 5 highlights the critical role of safety considerations in the design, operation, and management of process intensification technologies in industries. Then, several case studies illustrate the application of safety analysis for implementing process intensification technologies, demonstrating how hazard identification, risk assessment, and inherently safer design principles can contribute to the safe operation of intensified processes. Chapter 6, prepared by two active researchers in process control, outlines the progressive steps required to develop a robust control structure for the reactive-extractive distillation system for the recovery of isopropanol and diisopropyl ether from wastewater. In this chapter, several full control schemes (from the simplest temperature control to the complicated feedforward control) are presented in an evolutionary manner so that readers can grasp the advantages and drawbacks of each scheme. This chapter then proposes a suitable control scheme with triple-point temperature control that can regulate the purity of the products closer to the respective setpoint, despite several disturbances. Three leading researchers in the design and control of distillation processes contributed to Chapter 7, which focuses on the design and control of RD columns with an application for the production of 4-hydroxybutyl acrylate. This work frames the RD columns within a process where recycle streams (to these columns) are present, as is typical in industrial practice. The control is focused on developing a plantwide strategy to achieve the material inventory (in other words, balancing the reaction stoichiometry), the desired production rate, and product purity. Several process disturbances (e.g. flowrate and composition changes) are implemented to test the proposed control structure of the plant. Chapter 8 is prepared by four researchers with an interest in process simulation and control. It presents a study on improving the performance of middle-vessel batch distillation without vapor bypass but using a vapor recompression column for separating ternary (methanol/ethanol/1-propanol) zeotropic mixtures. It shows that single-stage vapor-recompressed middle-vessel batch distillation with a gain-scheduled proportional-integral controller is the best configuration in terms of energetic, economic, environmental, productivity, purity, and practical feasibility compared to conventional middle-vessel batch distillation. Chapter 9 is the contribution of four researchers active in process systems engineering. It analyzes three process alternatives, namely, a conventional sequence of columns (CSCs), DWC, and DWC with multistage vapor recompression (MVR), for different safety indices, namely, individual risk, damage index, and process route index. Later, the safety prospects of the three processes are discussed based on industry experts’ recommendations obtained through a survey. Similarity and dissimilarity between the safety prospects depicted by the considered indices and those suggested by the experts are discussed. Finally, a modification to the process route

11

12

1 Introduction

index is proposed to make it more comprehensive for analyzing the safety of processes and apply it to CSC, DWC, and DWC-MVR systems. Chapter 10, contributed by two researchers active in process safety, presents a dynamic safety analysis of intensified extractive distillation processes for the separation of an azeotropic mixture of acetonitrile and water, focusing on column overpressure as the primary safety issue. For this analysis, multiple layers of protection, including basic process controls, alarms, safety instrumented systems, and pressure relief systems, are simulated to mitigate potential hazards arising from various hazardous scenarios. The effectiveness of these protection layers is then assessed using Aspen Plus for steady-state process design, Aspen Dynamics for dynamic simulation and control, and scenario-based dynamic safety analysis. Chapter 11, prepared by five established researchers in process modeling, optimization, and control, describes the operational and safety aspects of bioprocesses, taking the separation and purification of methyl ethyl ketone, furfural, and lactic acid as an example. Various intensified separation and purification processes, such as RD and split-wall columns, are used. In this chapter, intensification techniques, proposed designs for each separation system, and the results are presented and discussed in terms of their operability and safety. The study reported in Chapter 11 shows that controllability and security are essential for the new intensified technologies and the enormous potential of developing intensified processes to make bioprocesses safer and more operable. Finally, Chapter 12 by three researchers active in process development, safety, and control analyzes economic and safety objectives for ultrasound assisted and ionic liquid catalyzed in situ biodiesel production from wet microalgae (referred to as alternative 1), and the same process (i.e. alternative 1) is intensified by DWC and MVR (referred to as alternative 2). Both processes are modeled in Aspen Plus V10 and optimized using an MS Excel-based program for the elitist non-dominated sorting genetic algorithm. This study takes individual risk as the safety objective and break-even cost as the economic objective. One solution from the Pareto-optimal front of each alternative is chosen using the simple additive weighing method. Then, the chosen optimal solutions for the two alternatives are compared.

1.5 Conclusions In essence, this book is intended to provide the chemical engineering academia, students, and practitioners with process intensification techniques/technologies and their influence on process controllability and safety. It is organized in such a way that readers can study one or more chapters of their interest independent of the other chapters. The contents of this book can be readily adopted as part of special/elective courses on process systems engineering, process design and integration, plantwide control of chemical processes, and safety and hazards analysis for chemical processes for under-/post-graduate students. Manufacturers are likely to continue to intensify the existing processes even further owing to the rising competition and regulations. This book will provide them with a comprehensive understanding of

References

the effects of process intensification on control and safety, both of which are indispensable. Finally, it is hoped that this book will be valuable to researchers and practitioners interested in process intensification and will also help in further research, developments, and industrial applications of process intensification.

References Baldea, M. (2015). From process integration to process intensification. Computers & Chemical Engineering 81: 104–114. Canas, H., Mula, J., Díaz-Madronero, M., and Campuzano-Bolarín, F. (2021). Implementing Industry 4.0 principles. Computers & Industrial Engineering 158: 107379. Ebrahimi, F., Virkki-Hatakka, T., and Turunen, I. (2012). Safety analysis of intensified processes. Chemical Engineering and Processing: Process Intensification 52: 28–33. Etchells (2005). Process intensification: safety pros and cons. Process Safety and Environmental Protection 83: 85–89. Gerven, T.V. and Stankiewicz, A. (2009). Structure, energy, synergy, time—the fundamentals of process intensification. Industrial and Engineering Chemistry Research 48 (5): 2465–2474. Keijer, T., Bakker, V., and Slootweg, J.C. (2019). Circular chemistry to enable a circular economy. Nature Chemistry 11: 190–195. Klais, O., Westphal, F., Benaissa, W. et al. (2009). Guidance on safety/health for process intensification including MS design. Part III: risk analysis. Chemical Engineering Technology 32 (11): 1831–1844. López-Guajardo, E.A., Delgado-Licona, F., Álvarez, A.J. et al. (2022). Process intensification 4.0: a new approach for attaining new, sustainable and circular processes enabled by machine learning. Chemical Engineering and Processing: Process Intensification 180: 108671. Ramshaw, C. (1983). ‘Higee’ distillation – an example of process intensification. The Chemical Engineer 389 (2): 13–14. Sharma, A.K., Bhandari, R., Pinca-Bretotean, C. et al. (2021). A study of trends and industrial prospects of Industry 4.0. Materials Today Proceedings 47: 2364–2369. Stankiewicz, A. and Drinkenburg, A.A.H. (2004). Process intensification: history, philosophy, principles. In: Re-Engineering the Chemical Processing Plant: Process Intensification, 1e (ed. A. Stankiewicz and J.A. Mouljin), 14–40. New York, NY, USA: Marcel Dekker. Stankiewicz, A.I. and Moulijn, J.A. (2000). Process intensification: transforming chemical engineering. Chemical Engineering Progress 96 (1): 22–34. Steffen, W., Richardson, K., Rockstrom, J. et al. (2015). Planetary boundaries: guiding human development on a changing planet. Science 347: 1259855.

13

15

2 Applications and Potential of Process Intensification in Chemical Process Industries Chirla C.S. Reddy Singapore Refining Company Private Limited, Singapore, 628260, Singapore

2.1 Introduction Process intensification is essential for the sustainable development of chemical processing plants. Process intensification techniques can be used during the entire life cycle of the plant (from design to construction and operations). However, the opportunity to effectively use process intensification is highest in the design phase of the process plant. This chapter describes many proven process intensification techniques used in process industries. Section 2.2 outlines the benefits of process intensification techniques. Section 2.3 presents static mixtures for mixing and chemical reactions. Process intensification techniques for process separation vessels, gas–liquid and liquid–liquid separators are discussed in Section 2.4. Next, heat-integrated, heat pump-assisted, dividing-wall, and reactive distillation (RD) columns are discussed in Section 2.5. Process intensification techniques for heating, flue gas (FG) waste heat recovery, and process heat exchange are presented in Section 2.6. Section 2.7 presents steam ejectors for steam compression. Rotating packed beds for carbon capture is discussed in Section 2.8. Process intensification techniques applicable to vacuum systems are discussed in Section 2.9. Next, Section 2.10 discusses compact degassing technology for water deaeration. Section 2.11 describes process intensification applications for the development of inherently safer design (ISD) processes. Safety-instrumented system solutions for reducing pressure relief and safe disposal requirements are presented in Section 2.12. Section 2.13 covers process intensification techniques applicable to wastewater (WW) recovery. The challenges of process intensification techniques are discussed in Section 2.14. Finally, this chapter ends with conclusions in Section 2.15. Learning outcomes of this chapter on process intensification techniques are: ● ●

●

Outline the key attributes of process intensification techniques. Discuss the commonly used process intensification techniques/technologies in process industries, along with their advantages and disadvantages. Elaborate on the role of process intensification techniques for developing ISD solutions.

Control and Safety Analysis of Intensified Chemical Processes, First Edition. Edited by Dipesh Shikchand Patle and Gade Pandu Rangaiah. © 2024 WILEY-VCH GmbH. Published 2024 by WILEY-VCH GmbH.

16

2 Applications and Potential of Process Intensification in Chemical Process Industries ● ●

Describe control systems for process intensification applications. Explain the importance of process intensification techniques/technologies for the sustainability of process industries.

2.2 Benefits of Process Intensification Techniques Process intensification techniques help process plants to: ●

●

●

● ● ● ● ●

occupy less footprint (typically 10-fold volume reduction, less piping, lower equipment count, etc.), consume less energy resources (e.g. electricity, fossil oil, gas, etc., due to enhanced heat, mass transfer, mixing, and energy efficiency), make processes simpler (lower number of process steps that can dramatically reduce net energy consumption, increase product quantity and/or quality, etc.), require low feed rates (with the ability to use reactants at higher concentrations), lower process interruptions and downtime, lower operating and maintenance costs, produce less waste, and/or be safer and more sustainable.

Process-intensified equipment may use rotation (in a cyclone or on a rotor), vibration (e.g. high-frequency ultrasound used for fouling prevention in HEs), and/or mixing (e.g. static mixing, covered in Section 2.3) for enhancing mass and/or heat transfer. Many benefits of process intensification are summarized in Figure 2.1. Process intensification also enables higher process flexibility, increased inherent safety, and distributed manufacturing capability. Process safety benefits from process intensification are: ●

●

●

●

● ●

Reduction of equipment and/or plant size results in a smaller volume of toxic and flammable inventories, thereby reducing the consequences of toxic release, fire, and explosions. Mitigating the risk of thermal runaway in highly exothermic chemical reactions by carrying out the process in smaller reactors. Due to enhanced mass and heat transfer rates in compact reactors, solvent, energy, and cooling requirements are reduced. Reduces the number of process operations, which leads to fewer transfer operations and less pipework, preventing the source of hazardous leakages. Lower pressure relief and effluent handling system requirements. Smaller equipment responds more quickly to safety interlock actions, such as shutting off reactant feeds, emptying a reactor to a reaction quench tank, depressurizing the reactor, and increasing cooling flow.

Process intensification techniques can be segregated into three distinct classifications involving the development of innovative equipment, methods, and plant designs. Process intensification covers both reactive and nonreactive equipment. An overview of process intensification equipment, methods, and techniques is shown

2.3 Static Mixers

Benefits of process intensification

Business: • Lower capital, operating and maintenance costs • Innovation • Faster growth from research to commercialization

Figure 2.1

Process: • Higher selectivity, yields and/ or purity • Lower byproducts generation • Higher reaction rates • Better quality products • Increased process flexibility & inventory reduction • Smaller plot space requirements • Easier start-up and shutdown operations • Easier revamps and retrofits • Continuous processes

Sustainability: • Greener processes • Improved process safety • Lower resource requirements • Lower wastage • Lower energy consumption and hence lower emissions

Benefits of process intensification for business, process, and sustainability.

in Figure 2.2. Wang et al. (2017) have presented a comprehensive review of process intensification techniques used in solid handling applications.

2.3 Static Mixers A static mixer (SM) is a motionless mixer inserted into a pipeline or a closed housing for manipulating fluid streams to divide, recombine, accelerate/decelerate, spread, swirl, or form layers as they flow through it. Fluid components flowing through SM are brought into very intimate contact and achieve very good mixing. SMs are widely used in process plants mainly for liquid–liquid, gas–liquid, solid–liquid, and solid–solid mixings, such as for mixing oils, resins, dilution of process streams, chemical injection, gas-liquid dispersions, pH control, and pipeline reactions. Generally, SM avoids the need for expensive mixing tanks, which are traditionally used for the mixing of fluids. The uses of SM for mixing applications are summarized in Figure 2.3. SMs rely on external pumps and/or compressor/blower to move fluid components across the mixer elements. Hence, fluid pressure drop is the key basis for selecting the appropriate SM design. The shape, length, and number of mixer elements are chosen to achieve the desired mixing, without exceeding the allowable pressure drop. The advantages and disadvantages of SMs are compiled in Table 2.1. Generally, injected fluid flow is adjusted continuously in proportion to the main fluid flow using a ratio control. A typical control arrangement for an SM is illustrated in Figure 2.4. Liquid jet ejectors (LJEs) can be used for mixing contents inside pipelines and storage tanks1 , agitating, and supplying oxygen to biological reactors in WW plants2 . 1 https://www.koerting.de/en/tank-mixing-systems.html. 2 https://www.koerting.de/en/water-treatment.html.

17

18

2 Applications and Potential of Process Intensification in Chemical Process Industries

Reactors

Static mixer reactor Spinning disk reactor Micro reactors Heat exchanger reactors Rotating packed bed reactor

Equipment Equipment for nonreactive operations

Multifunctional reactors

Process intensification

Methods

Static mixer Compact heat exchanger Active and passive techniques for heat exchanger Centrifuges Rotating packed beds Reactive distillation Reactive extraction Heat integrated reactor Membrane reactors Fuel cell

Separations

Externally heat integrated distillation Dividing wall distillation Heat pump assisted distillation Membrane distillation Absorptive distillation Membrane absorption Demisters/Vane separators/Coalescers

Alternative energy sources

Centrifugal field Solar energy Ultrasound Electric field

Other methods

Plant design

Figure 2.2

Re-engineering of existing plants

Modifications to apply inherently safer methods

Greenfield plants

Modular and inherently safer designs

Overview of process intensification equipment, methods, and plant design.

SMs not only provide good mixing but also provide good heat transfer (with jacketed design) and uniform residence time. Hence, they are also used for reaction processes. For example, Sulzer’s static-mixer reactor (Figure 2.5) has mixing elements made of heat-transfer tubes. It can be successfully used in processes that require simultaneous mixing and heat transfer, such as in nitration or neutralization reactions.

2.4 Process Intensification for Separation Vessels Gas/vapor–liquid (G/V–L), liquid–liquid (L–L), and liquid–liquid–gas/vapor (L–L–G/V) separations are usually carried out in horizontal or vertical separation vessels, using the difference in fluid densities and providing sufficient residence space. This traditional design leads to bigger separators and hence results in large inventories inside the separators. Separator size can be reduced significantly (up to 50% or more) by using internal separation devices such as feed Schoepentoeter, liquid–liquid coalescers, vane separators, and/or demister pads. Feed Schoepentoeter is mainly installed at the feed nozzle, inside the G/V–L separator. It separates G/V–L mixture effectivity with a liquid separation efficiency of ∼95% from a gas stream. It reduces the liquid load to vane separators and/or demister pads

2.4 Process Intensification for Separation Vessels

Static mixers

Mixing of miscible liquids

Mixing of immiscible fluids

Liquidliquid

Figure 2.3

Mixing and reaction, with heat transfer

Mixing of solids

Liquidgas

Liquidsolid

Uses of SMs.

RY

FC-2

Secondary fluid

Mixing ratio

FI-2

FCV-2 FI-1

Mixed fluid

Main fluid

Legend: FI = Flow indicator/transmitter FC = Flow controller RY = Ratio relay

Figure 2.4

Static mixer

Typical arrangement and control system for a SM, installed in a pipeline.

and enables efficient liquid droplet separation from the gas/vapor stream. Vane separators and/or wire-type demister pads are installed on the upper portion of G/V–L separator in the gas region. They can remove liquid particles with sizes > 10 microns, from the gas/vapor stream, with a separation efficiency of 99–99.5%. A typical gravity-type vertical G/V–L separator is shown in Figure 2.6a, and its intensification with feed Schoepentoeter, vane, and demister pad is shown in Figure 2.6b. For most applications, a demister pad alone is sufficient. If assured liquid droplet separation (>99.9% separation of liquid particles with sizes >8 microns) is desired in applications such as liquid droplet knockout drums at the suction of reciprocating compressors, a combination of feed Schoepentoeter, wire demister, and/or vane separator is used. Good level control in G/V–L separator is essential for achieving good separation of liquid droplets from gas/vapor stream.

19

20

2 Applications and Potential of Process Intensification in Chemical Process Industries

Table 2.1

Advantages and disadvantages of SMs.

Advantages ● ●

●

●

●

●

●

●

●

Produce nearly homogenous mixing. Compact size (compared to stirred tank with agitator) and can be built into an existing pipeline. Requires lower capital cost compared to stirred tank mixer. No moving parts (e.g. shaft, bearings, seals, or drive motor) and hence incur negligible maintenance. if properly applied, consumes less energy compared to rotating agitators. Silent operations do not require electrical power supply and are thus suitable for working in potentially explosive atmospheres Supports optional heating of the mixture using jacket design and steam/hot oils. Heating can also be provided without jacket, using external electrical coils, installed on the static mixer’s surface. Available in all commercial materials of construction. Custom-designed by vendors as per the process mixing needs.

Disadvantages ●

●

● ●

●

Short residence time. Not suitable for longer residence times (∼30 min to a day or more). Require precise dosing of components, using good control, with constant concentrations (for both main and injected fluids) as per mixing ratio. Not well suitable for batch operations. Not well suited for blending very high-viscosity fluids. May be clogged by solids.

Figure 2.5 Sulzer’s static-mixer reactor. Source: Sulzer Ltd / //https://www.sulzer.com/ en/shared/products/polymer-reaction-technology / last accessed September 26, 2023.

2.5 Process Intensification for Distillation Vapor

Vapor Demister pad

Gas/vapor and liquid Feed

Gas/vapor and liquid feed

Gravity type G/V-L separator

Vane separator

Intensified G/V-L separator

Schoepentoeter LI

LC

LI

Liquid LC

LCV Liquid

LCV

Figure 2.6

(a) Gravity-type G/V–L separator and (b) Intensified G/V–L separator.

The gravity-type L–L–G separator is generally big as it requires a large residence time for separating dispersed phase liquid particles with sizes 1, the system is attenuated and can be controlled. An attenuation ratio in the range of 4–10 is preferred for a system.

97

98

4 Dynamic Simulation and Control of Intensified Chemical Processes

Figure 4.11 Transient response for a step change in the set point of a controller.

y(t) B2

B1

ysp

y(∞)

y(tp) tp

ts

The overshoot (or maximum deviation), 𝜎, of a transient response is defined as: 𝜎=

y(tp ) − y(∞) y(∞)

× 100%

(4.7)

Here, y(tp ) is the value of the response at the first peak, and y(∞) is the value when the system reaches a new steady state. The smaller the overshoot is, the better the control performance. The Steady-state deviation (offset) measures the performance of the controller to trace its set point under various disturbances. It is defined as the error between ysp and y(∞). The integral of error criteria quantify the performance of the control system over time. The popular criteria are the integral of absolute error (IAE), integral of squared error (ISE), and integral of product of time and absolute error (ITAE); they are defined as follows: T

IAE =

∫0

|e(t)|dt

(4.8)

e2 (t)dt

(4.9)

T

ISE =

∫0

T

ITAE =

∫0

t|e(t)|dt

(4.10)

Here, e(t) is the error between the controller set point and process variable at any time. Minimizing ISE tends to eliminate large errors quickly, i.e. faster dynamic response with smaller oscillations. IAE tends to produce slower response than ISE but with less sustained oscillations. ITAE adds time weight on error, which produces a system with a shorter setting time than IAE and ISE.

4.4 Dynamic Simulation and Control of Intensified Chemical Processes An intensified chemical process (ICP) often involves appropriate integration of multiple unit operations such as pervaporation, distillation, heat pump, reactor, chromatography, filtration, and centrifugation. Designing and optimizing the integration of two or more of these unit operations while maintaining efficient and seamless operations can be complex (D’Souza et al. 2013). Typical example of process intensification is the reactive distillation (RD or catalytic distillation)

4.4 Dynamic Simulation and Control of Intensified Chemical Processes

process for producing methyl acetate via esterification of acetic acid and methanol in Eastman Chemical (Stankiewicz and Moulijn, 2002). The old production process consisted of 28 equipment, including reactor, distillation columns, decanter, and heat exchangers. However, the intensified process consists of only one RD column, including its reboiler and condenser, which significantly reduces the capital/operating costs and (size of) equipment. Process intensification aims to increase the driving forces for heat/mass transfer and/or reaction in chemical processes. Although the intensified process increases the effectiveness of equipment, its integrated structure and integration of phenomena may lead to complex dynamics and difficulties in its operation and control. Hence, dynamic simulation of ICPs is essential for the following issues: Process Robustness: Ensuring the robustness of ICPs is crucial for consistent and reliable operation. Factors such as fouling, variability in feedstock quality, and process disturbances can affect the performance and efficiency of the process (Becker et al., 2023). To develop strategies to mitigate the effects of these factors, dynamic simulation and control are essential to investigate the operability of ICPs in various situations. Process Safety: Safety considerations are important in ICPs, where higher pressures, temperatures, and/or flow rates may be involved. Adequate measures should be implemented to prevent accidents, ensure equipment integrity, and protect personnel. Risk assessments, safety protocols, and equipment design that comply with safety standards are essential to address these challenges. Further, process conditions related to safety (such as overpressure and thermal runaway of reactor) often vary with time and should be assessed via dynamic simulation. Control and Automation: ICPs such as RD and DWC often require advanced control and automation systems for efficient and reliable operation. Implementing robust control strategies that can handle the complexity of ICPs, monitor critical process parameters, and adjust operation in real time is essential for achieving high process efficiency. Equipment Design: ICPs may involve specialized equipment and technologies, such as DWC, RD, and pervaporation. Designing these systems to be cost effective for different process conditions can be challenging. Further, ensuring that this equipment can be easily integrated into existing facilities or scaled up for commercial production is crucial (Ruiz-Ruiz et al., 2013). Finally, dynamic simulation and control are essential for evaluating the operation flexibility of ICPs. Validation and Regulatory Compliance: ICPs also must meet regulatory requirements and undergo validation to ensure product quality, safety, and consistency. These require dynamic simulation to validate the operability of ICPs for their industrial acceptance. Startup and Shutdown: The operation complexity is an important limitation in the industrial applications of ICPs. Dynamic simulation can approximate the dynamic behavior of these processes and thus provide guidance for their startup and shutdown.

99

100

4 Dynamic Simulation and Control of Intensified Chemical Processes

4.4.1

Challenges Due to Process Intensification

Compared to conventional processes, ICPs often have larger driving forces for heat/mass transfer and/or reaction, resulting in smaller size of equipment; they may also integrate two or more unit operations. These lead to complex behavior, which makes the dynamic simulation and control of ICPs more difficult than that of conventional processes. These difficulties are outlined below. ●

●

●

Model Validation: Accuracy of basic data (e.g. heat/mass transfer coefficients, reaction rates, and phase equilibrium data) has significant impact on the dynamic simulation of ICPs due to the effects of error accumulation in time scale and interactions among process variables. An appropriate thermodynamic model (fluid package) such as PR (Peng–Robinson), Wilson, or NRTL (nonrandom two-liquid) model, along with its parameter values for accurately predicting properties of pure components and their mixtures, is essential for realistic simulation. Hence, availability and reliability of physicochemical properties of pure components and binary interaction parameters of the selected thermodynamic model (parameters of the model used for estimating heat/mass transfer coefficients and reaction kinetics) must be checked and validated using suitable experimental or literature data. Numerical Solution and Convergence: Dynamic model of a chemical process consists of a large set of DAEs and ODEs/PDEs, which are solved using a numerical method over a specified period of time to find the dynamic behavior of the state variables. In particular, dynamic model of an ICP often involves even more number of equations and state variables as well as more and/or complex interactions among process variables; all these can make the solution of this dynamic model challenging with convergence difficulties and/or require more computing time compared to solving the dynamic model of a conventional process. Complex Process Control: ICPs often involve increased interactions of process variables and/or unit operations, which can make their dynamic behavior complex and their control challenging compared to conventional chemical processes. For example, interactions between the two sides of the dividing wall in a DWC make the control of vapor split ratio more difficult, but this control is important for efficient operation of DWC (Halvorsen and Skogestad, 2004). Further, compared to conventional processes, smaller volume of equipment and faster energy/mass transfer of ICPs require higher reliability and faster response of control systems for handling process disturbances. Therefore, MPC is promising for the operation of ICPs; e.g. MPC and temperature difference control were investigated for extractive DWC by Feng et al. (2018), and nonlinear MPC of DWC was studied by Qian et al. (2023).

4.5 Process Control This section outlines controlled, manipulated, and disturbance variables, feedback control loops, and CDOF for the benefit of novice readers.

4.5 Process Control

4.5.1

Controlled, Manipulated, and Disturbance Variables

Process variables for designing a PWC scheme often include flow rate, temperature, and pressure of streams entering/leaving the units/equipment, liquid levels in tanks, valve openings, and pressure, temperature, and composition in the units/equipment itself. They can be classified as process input, output, and disturbance variables. Process-controlled variables (i.e. output variables) give information about the state of the process. They are often state variables such as temperature, pressure, level, and composition. For example, CA and T in Eqs. (4.1) and (4.2) are the controlled variables for the CSTR system. Process-manipulated variables are one type of input variables that can be manipulated by the process operator or control system to maintain the controlled variables at the desired set points for optimal operation of the process. They are often valve openings and shaft work for compressors. For example, CA0 , F, and QR in Eqs. (4.1) and (4.2) are the manipulated variables for the CSTR system. Process disturbance variables are also input variables. They affect the dynamic response of controlled variables and may or may not be measurable. Moreover, they cannot be adjusted by the control system. For example, T 0 in Eqs. (4.1) and (4.2) is a measurable disturbance variable.

4.5.2

Typical Control Loop

Figure 4.12 shows a typical control loop for controlling a process, which can be a unit/equipment such as a distillation column and a heat exchanger. The controller seeks to maintain the measured process variable, ym , at the specified set point, r. Whenever the controller error, e(t) is not equal to zero, the controller will change its output signal to the valve to adjust the valve opening (i.e. manipulated variable); then, the process variable, y changes as per the manipulated variable and the disturbance variable. The measured process variable is then sent to the controller to calculate the new controller error. This entire procedure is repeated throughout the process operation. A PID controller or advanced multivariable controller, e.g. linear or nonlinear MPC, can be used for process control. The model in MPC includes interaction among variables and, therefore, predicts more accurately the dynamic response of the process variable over a specified time range. Thus, MPC often shows better performance than PID controller.

4.5.3

Control Degrees of Freedom

CDOF refers to the number of variables that can be manipulated independently in the given process; it is very important for the design of a PWC system. If a model is built for dynamic simulation and control system design, CDOF also refers to the degrees of freedom of the mathematical model, i.e. the difference between the total number of unknown variables and the number of independent chemical/physical equations, which is given by: CDOF = number of unknown variables − number of independent equations (4.11)

101

102

4 Dynamic Simulation and Control of Intensified Chemical Processes

Controller error e(t) = r – ym Setpoint r + –

e

Manipulated variable

Controller output signal

Controller Gc(s)

Measured process variable signal ym

uc

Valve Gr(s)

u

Disturbance variable

Process variable

Process Gp(s)

y

d

Measurement sensor/transmitter Gm(s)

Figure 4.12 Typical control loop block diagram; G(s) is the Laplace transfer of the relevant differential equation (i.e. of controller, valve, process, and measurement sensor/transmitter as indicated by its subscript).

Values of CDOF variables must be specified before the model equations can be solved. See Murthy Konda et al. (2006) for further details. In general, there are three cases, depending on the value of CDOF. (i) CDOF = zero: the model equations can be solved in principle without requiring the values of any variables. (ii) CDOF > zero: there are more unknown variables than independent equations relating them. In this situation, the values of DOF variables must be specified before the remaining variables can be determined. For example, if F and T 0 in Eqs. (4.3) and (4.4) are fixed, DOF of CSTR system is 2, i.e. CA0 and QR ; thus, CDOF of CSTR system is 2, and CA0 and QR can be adjusted to satisfy the desired reaction conversion (CA ) and reaction temperature (T). (iii) CDOF < zero: there are more independent equations than unknown variables. In this situation, the model is overspecified and cannot be solved.

4.6 Case Study Extractive dividing-wall column (EDWC) is a thermally coupled extractive distillation, which is an example of intensification, for separating mixtures with azeotropes and/or low relative volatility; it has been used in industrial practice (Yildirim et al., 2011). The extractive section and entrainer recovery section in EDWC share one reboiler, reducing the number and size of equipment as well as capital cost compared to conventional extractive distillation. In this section, a case study on EDWC for separating toluene and 2-methoxy ethanol (MEA) mixture using dimethyl sulfoxide (DMSO) as the entrainer is used to illustrate the dynamic simulation and control procedures described in the earlier sections of this chapter. This EDWC was first proposed by Li et al. (2017) and then used to investigate its controllability (Feng et al., 2018, 2019a). In the following sub-sections, composition and temperature control schemes were designed and compared for handling flow rate and composition

4.6 Case Study

disturbances in the feed stream to EDWC. Since PI control is widely used for the operation of distillation processes, it is chosen for investigating the control of EDWC; see Feng et al. (2018, 2019a) for advanced control (i.e. MPC) of EDWC.

4.6.1

Steady-state Simulation and Optimization

The steady-state simulation and optimization of EDWC process for separating equimolar mixture of toluene and MEA using DMSO entrainer were studied by Li et al. (2017). The product purity of both toluene and MEA is 99.5 wt%, whereas the purity of recycled DMSO is 99.82 wt%. As shown in Figure 4.13a, EDWC consists of three sections: extractive section (C1), rectifying section (C2), and striping section (C3); hence, EDWC is simulated via three columns (Figure 4.13b), which is its thermodynamic equivalent model in the absence of heat transfer across the wall. Extractive and rectifying sections of EDWC have 40 and 12 ideal stages, including their respective condenser, while striping section has 10 ideal stages, including its reboiler. The fresh feed enters the 28th stage of C1 at a flow rate of 100 kmol/h. The recycled entrainer, after cooling to 101.25 ∘ C by cooling water and mixed with a small amount of makeup, is fed on the 6th stage of C1. The reflux ratios of C1 and C2 are 1.2966 and 0.3347, respectively. The molar flow rate ratio of entrainer-to-feed (S/F) is 0.7, whereas the vapor split ratio (RV , which refers to the flow rate fraction going to the bottom of C1 from the top of C3) is 0.65. The pressure at the top of C1 is set as 1.01 bar, while that at the top of C2 is 1.21 bar, for balancing the pressure at the bottom of C1 and C2 sections. The pressure drop on each stage in the three sections is set at 0.0068 atm.

4.6.2

Preparation/Initialization for Dynamic Simulation

Before dynamic simulation of a chemical process, additional information about the physical size of equipment, such as tanks, reflux drums, flash vessels, and the diameter and number of trays of columns, must be specified. In addition, the required pumps and valves must be added before converting the steady-state simulation to a dynamic one. Figure 4.14 displays the steady-state model of EDWC in Aspen Plus, after adding the required pumps and valves. The compressor C101 in Figure 4.14 is used to increase the pressure of the stream V1 from the top of C3, overcoming the pressure drop in valves V103 and V104. A large pressure increase in compressor will lead to a large increase in the temperature of stream V1; this will change the process configuration. Hence, the discharge pressure of compressor C101 is specified as 1.35 bar, which provides a 6.8 kPa pressure drop in each of the valves, V103 and V104. The pressure drop in the remaining valves is assumed as 2 bar for effectively handling the expected process disturbances since, if valve pressure drop is too small, the change in flow rate is fairly small when the valve opening changes from 50% to 100%. Pumps P101–P105 are added to increase the pressure of the streams entering valves for satisfying their respective pressure drops. In addition, the pressure of each stream before entering the column must be equal to the corresponding feed stage pressure in the column; this can be checked

103

104

4 Dynamic Simulation and Control of Intensified Chemical Processes Condenser1

QC1

D1, Toluene

Makeup DMSO

RR1 S/F

C1

Condenser2

QC2

FEED C2

RR2

D2, MEA

Rv

C3

Cooler E1

Reboiler QR

Entrainer

(a)

DMSO

Condenser1 Duty: –1053.43 kW

Makeup: 0.05 kmol/h DMSO: 100 wt%

FEED: 105°C, 3.2 bar 100.00 kmol/h MEA: 50 mol% Toluene: 50 mol%

E1 Duty: –323.68 kW 101.25°C

2

–715.98 kW

1.21 bar 129.91°C

D1:

6 C1 28

50.10 kmol/h MEA: 0.41 wt% Toluene: 99.50 wt% DMSO: 0.09 wt%

C2

40

RR1: 1.2966

12

2

Rv = 0.65 L1: 146.21°C, 1.28 bar 235.93 kmol/h MEA: 65.49 wt% Toluene: 0.30 wt% DMSO: 34.21 wt% B: 69.94 kmol/h 2-MEA: 0.18 wt% DMSO: 99.82 wt%

(b)

Condenser2 Duty:

1.01 bar 110.44°C

D2: 49.96 kmol/h MEA: 99.50 wt% Toluene: 0.48 wt% DMSO: 0.02 wt%

RR2: 0.3346

41 C3 49

V1: 150.74°C, 1.29 bar 178.51 kmol/h MEA: 91.17 wt% Toluene: 0.40 wt% DMSO: 8.43 wt%

L2: 150.7°C, 1.28 bar 12.53 kmol/h MEA: 58.26 wt% Toluene: 0.09 wt% DMSO: 41.65 wt%

Reboiler Duty: 2048.53 kW 202.05°C 1.35 bar

Figure 4.13 (a) EDWC structure along with some potential manipulated variables (shown in blue color) and (b) three-column configuration, which is the thermodynamic equivalent model, of EDWC with stream data for the separation of toluene and MEA using DMSO as the entrainer.

4.6 Case Study

Figure 4.14 Steady-state process model of three columns for EDWC, constructed in Aspen Plus, after adding the required pumps and valves.

by Pressure Checker in the Dynamic Mode in Aspen Plus, as described in Section 4.3. For distillation control, reflux drum and sump volumes should be large enough to provide five minutes of holdup when half full (Luyben and Chien, 2009). Herein, reflux drums and sumps are sized for 20 minutes of total holdup when they are 100% full to avoid any effect of level on the control of EDWC. Figure 4.15 depicts the hydraulic results for the reflux drum and column base of C1 for EDWC in Aspen Plus. For the RadFrac model of distillation in Aspen Plus, the first stage often represents the reflux drum/condenser; thus, the volumetric flow rate of the liquid to the reflux drum is 13.4127 m3 /h. Then, the reflux drum volume is calculated by: ) ( ( )2 m3 D ÷ 60 × 20 min = 4.47 m3 H = 13.4127 (4.12) V =π 2 h Here, D is the reflux drum diameter and H is the reflux drum height, which is assumed as two times D. Thus, D and H of the reflux drum are 1.42 and 2.84 m, respectively. Since there is no reboiler for C1, the liquid from the last (40th ) stage is the volumetric flow rate of the liquid to the column base, i.e. 20.9566 m3 /h, which results in D and H of C1 column base as 1.64 and 3.28 m, respectively; this is assuming H is two times D. Figure 4.16 depicts the hydraulic results of C3 of EDWC in Aspen Plus. Since the last (10th ) stage of C3 is the reboiler of EDWC, volumetric flow rate of the liquid to C3 column base is that from the 9th stage of C3, i.e. 20.2942 m3 /h. Table 4.3 summarizes the calculated D and H of all reflux drums and column bases of EDWC sections. Simple tray hydraulics is used for dynamic simulation of columns. As shown in Figure 4.7, D and weir height of columns for each section are also given in Table 4.3.

105

106

4 Dynamic Simulation and Control of Intensified Chemical Processes

Liquid from reflux drum

Liquid to column base Figure 4.15 Aspen Plus.

Hydraulics results for the reflux drum and column base of C1 of EDWC in

Liquid to column base Figure 4.16

Hydraulics results for C3 column base of EDWC in Aspen Plus.

4.6 Case Study

Table 4.3 Diameter and height of the reflux drum and column base, as well as diameter and weir height of column for EDWC sections. EDWC sections Item

C1

C2

C3

Diameter of the reflux drum, m

1.42

1.09

—

Height of the reflux drum, m

2.84

2.18

—

Diameter of the column base, m

1.64

0.62

1.63

Height of the column base, m

3.28

1.24

3.26

Diameter of the column section, m

1.1

0.8

1.2

Weir height in the column section, m

0.05

0.05

0.05

Once the required dynamic parameters for dynamic mode are specified in Aspen Plus, the pressure checker in Figure 4.6 is used to check whether the pressure of each stream equals to the pressure of the corresponding equipment or column feed stage. If the flowsheet is configurated to be fully pressure driven, then click “Pressure Driven” button to convert the steady-state simulation in Aspen Plus to dynamic simulation in Aspen Dynamics. Note that the dynamic model in Aspen Dynamics must be successfully initialized by clicking Initialization button in the run mode window before conducting dynamic simulation.

4.6.3

Control Structure Design

EDWC introduces a new CDOF, namely, vapor split ratio (RV ) compared to conventional extractive distillation. RV has significant effect on the operability and operating cost of EDWC. Also, it depends significantly on the hydraulic conditions inside the column, e.g. dividing-wall installation, liquid resistance, and pressure drop on both sides of the dividing wall. Therefore, it is hard to directly manipulate RV in practice. A few studies (Dwivedi et al., 2012; Harvianto, 2019; Li et al., 2020) experimentally investigated the feasibility of adjusting RV by changing the vapor flow area on both sides of the dividing wall, but this is difficult in industrial practice. Until now, there have been no reports on industrial implementation of manipulating RV for DWC operation. This restricts the industrial applications of DWC. Previous studies indicate that the control structure with manipulating RV can achieve better control for rejecting feed composition disturbances than that without manipulating RV for the operation of EDWC (Tututi-Avila et al., 2014). Hence, RV is chosen as the manipulated variable to obtain good control of EDWC. Unlike only one condenser in DWC for separating ternary mixtures, EDWC has two condensers; it is divided by a dividing wall into two sections (C1 and C2 in Figure 4.13a), each with its own condenser, which provides an alternative to indirectly manipulating RV by adjusting the operating pressure at the top of C1 or C2. See Luyben (2018) and Feng et al. (2019a) for details on this.

107

108

4 Dynamic Simulation and Control of Intensified Chemical Processes

Another critical process variable is S/F, which has a significant impact on the product purities and energy consumption of EDWC. Large S/F increases both the relative volatility of light key to heavy key and the reboiler duty of EDWC for recovering the entrainer, DMSO. However, the control scheme with fixed S/F often exhibits more robust performance than that with variable S/F for EDWC operation (Tututi-Avila et al., 2014). Hence, S/F is fixed at a constant value in the control schemes presented in this section. Both composition control (CC) and temperature control (TC) schemes are designed and compared for EDWC. 4.6.3.1 Composition Control Scheme

The inventory control loops for CC scheme were designed as follows (Figure 4.17). (1) The feed flow rate is flow-controlled (FC1). (2) The flow rate of recycled entrainer, DMSO from the bottom of C3 is controlled by a flow controller (FC3), whose set point is continuously adjusted by the value of S/F (mass flow rate ratio of 0.6283 = molar flow rate ratio of 0.7) along with the variation of feed flow rate. This is a typical cascade control scheme. (3) The operating pressure of C1 and C2 is controlled by manipulating their respective condenser duties (PC1 and PC2). (4) The operating pressure of C3 is controlled by manipulating shaft work of compressor, C101 (PC3). (5) The vapor flow rate to the bottom of C1 is controlled by a flow controller (FC2), whose set point is adjusted by the value of RV (RV = 0.65 at normal operating state). (6) The reflux flow rates of C1 and C2 are ratioed to the feed flow rate with mass ratios of 0.7108 (R1 /F) and 0.1513 (R2 /F), respectively, instead of using reflux ratios. This is because R1 /F and R2 /F can respond quickly to the variation of feed flow rate, and it is a feedforward control structure. (7) The reflux drum levels of C1 and C2 are controlled by manipulating their respective distillate flow rates (LC1 and LC2). (8) The sump levels of C1 and C2 are controlled by manipulating their respective bottom liquid flow rates (LC3 and LC4). (9) The sump level of C3 is controlled by make-up DMSO flow rate with reverse action (LC5). (10) The temperature of recycled entrainer, before mixing with make-up DMSO, is controlled by manipulating the duty of cooler E1 (TC5). For the operation of EDWC, the product purity of toluene (C1 distillate), MEA (C2 distillate), and recycled entrainer DMSO (C3 bottom product) should be controlled at the desired values. In general, it is more effective to control impurity levels than to control purity levels for the operation of distillation (Ling and Luyben, 2010). This is because a change in impurity from 1 to 1.5 mol% is more sensitive to the variation of manipulated variable than a change in purity from 99 to 89.5 mol%. Hence, impurity of MEA in toluene (D1 in Figure 4.13b) can be used to control toluene product purity by manipulating reflux ratio of C1. From Figure 4.13b, toluene and DMSO in MEA product (D2) are 0.44 and 0.05 wt%, respectively.

4.6 Case Study

Figure 4.17

CC scheme for the operation of EDWC constructed in Aspen Dynamics.

Toluene impurity in MEA products comes from the bottom liquid of C1 (L1). Hence, it will not change much if toluene impurity in L1 is controlled by manipulating RV at the desired value. Steady-state and dynamic simulations of EDWC indicate that DMSO in MEA products becomes the main impurity that varies when the feed composition changes. Figure 4.13b shows that DMSO mole fraction in MEA product is 0.02 wt%, which is too small and cannot be used as the controlled variable to manipulate R2 /F. Therefore, composition-sensitive analysis based on steady-state simulation is used to select the suitable DMSO concentration on the stage of C2 to manipulate R2 /F. Figure 4.18 shows the liquid phase DMSO mole fraction profile in C2 with reflux ratio of C2 (RR2 ) when toluene purity from top of C1, DMSO purity from bottom of C3, and impurity of toluene from bottom of C1 are fixed; these data were obtained by conducting steady-state simulation of EDWC. Herein, toluene purity from C2 distillate is fixed by adjusting the reflux ratio of C1, while DMSO purity from bottom stream of C3 is fixed by adjusting the boil up ratio of C3. The impurity of toluene from the bottom stream of C1 (L1) is fixed by adjusting RV . Figure 4.18 shows that the change of DMSO mole fraction on the 6th stage of C2 with variation of RR2 is larger than that on other stages of C2; in other words, DMSO mole fraction on the 6th stage is more sensitive to RR2 , and so it can be used for manipulating RR2 . The quality control loops for CC scheme are as follows (Figure 4.17). (1) Manipulated variable, R1 /F is used to control MEA impurity in C1 distillate for controlling toluene product purity (CC1). (2) Manipulated variable, R2 /F is used to control mole fraction of DMSO on 6th stage of C2 for controlling MEA product purity (CC2).

109

110

4 Dynamic Simulation and Control of Intensified Chemical Processes

Figure 4.18 Sensitive analysis of liquid phase DMSO mole fraction on stages in C2 with reflux ratio, when toluene purity from top of C1, DMSO purity from bottom of C3, and impurity of toluene from bottom of C1 are fixed.

(3) Manipulated variable, RV is used to control impurity of toluene in C1 bottom stream (CC3). (4) Manipulated variable, QR /F (ratio of reboiler duty of EDWC to feed flow rate), is used to control impurity of MEA in the bottom stream of C3 for controlling recycled entrainer, DMSO purity (CC4). Herein, the utilization of feedforward structures R1 /F, R2 /F, and QR /F can reduce or eliminate the time delay of dynamic response to feed flow rate and feed composition disturbances. All the composition control loops have a five minutes dead time to account for the time delay of composition measurement. 4.6.3.2 Temperature Control Scheme

The inventory control loops for TC scheme are the same as those for CC scheme. For quality control loops, open-loop sensitive tests employing ±0.1% step changes in manipulated variables are used to select the sensitive tray temperature for TC scheme; Figure 4.19 shows results of these tests for step changes in RR1 , RR2 , RV , and QR . Temperatures on the 13th stage in C1 (T 13 , Figure 4.19a) and the 9th stage in C2 (T 9 , Figure 4.19b) have the largest changes and can be used to manipulate RR1 and RR2 (i.e. R1 /F and R2 /F for feedforward control structure), respectively. Temperature on the 34th stage (T 34 ) has the largest change for ±0.1% changes in RV (Figure 4.19c), and so it can be used to manipulate RV . Figure 4.19d shows that the temperature on the 43rd stage (T 43 ) has the largest change, and so it can be selected as the sensitive tray temperature for pairing with the manipulated variable, QR , to control recycled entrainer/DMSO purity. Figure 4.20 shows the overall TC scheme for the operation of EDWC. All quality control loops in this scheme are as follows: (1) T 13 is used to manipulate R1 /F for controlling toluene product purity (TC1). (2) T 9 is used to manipulate R2 /F for controlling MEA product purity (TC2).

(c)

(b)

Stage

ΔT (°C)

Stage

ΔT (°C)

(a)

ΔT (°C)

ΔT (°C)

4.6 Case Study

Stage

(d)

Stage

Figure 4.19 Open-loop test results: temperature change profiles in C1 (plots (a) and (c)), C2 (plot (b)) and C3 (plot (d)).

Figure 4.20

TC scheme for the operation of EDWC constructed in Aspen Dynamics.

111

112

4 Dynamic Simulation and Control of Intensified Chemical Processes

(3) T 34 is used to manipulate RV for controlling the impurity of toluene in L1 (TC3). (4) T 43 is used to manipulate QR /F for controlling recycled entrainer purity (TC4). All the above temperature control loops have one minute dead time to account for the time delay of temperature measurement.

4.6.4

Tuning of Controller Parameters

PI controller is used in the designed control schemes for EDWC operation. The parameters of flow, pressure, and level controllers are from Table 4.2; among them, proportional gain of level controllers is set at 5. The tuning procedure described in Section 4.3 along with Tyreus–Luyben method is used to find the parameters of temperature and composition controllers, which are given in Table 4.4.

4.6.5

Analysis of Dynamic Simulation Results

In this section, the dynamic performance of designed control (CC and TC) schemes is tested by giving disturbances of ±20% changes in the feed flowrate (F ± 20%) and ±10% changes in toluene mole fraction in the feed stream (zT ± 10%); each disturbance was introduced at the time of 0.5 hours. IAE given in Eq. (4.8) is used as the quantitative criterion to evaluate performance of CC and TC schemes. Here, the total simulation time of each test was 22 hours due to the slow dynamic response of CC scheme. Figure 4.21 shows the closed-loop dynamic responses of purities of toluene, MEA, and recycled DMSO, in CC and TC schemes, when subjected to ±20% changes in the feed flow rate. In CC scheme, product purities return to their respective set points, but after relatively larger transient deviations (blue and red solid curves) than those in TC scheme (black dash and purple dash dot curves). In Figure 4.21, value shown for each curve in each plot is the steady-state deviation at t = 22 hours. There are Table 4.4

Tuning parameters of controllers. Manipulated variable

Proportional Gain, K p

Integral time, 𝝉 I (min)

Wd1,MEA

R1 /F

0.21

216.48

xC2-6,DMSO

R2 /F

0.14

52.80

Scheme

Control loop

Controlled variable

CC scheme

CC1 CC2

TC scheme

CC3

wb1,toluene

RV

0.30

106.92

CC4

wb3,MEA

QR /F

0.98

75.24

TC5

TC

QCooler

0.36

4.62

TC1

T 13

R1 /F

1.10

30.36

TC2

T9

R2 /F

3.91

17.16

TC3

T 34

RV

3.86

18.48

TC4

T 43

QR /F

1.11

13.2

TC5

TC

QCooler

0.36

4.62

4.6 Case Study

4.43E–04

–1.60E–05

–4.56E–04 1.31E–05

(a)

–9.74E–04

7.13E–06

(b)

Time (h)

–1.20E–05

MEA (wt%)

Toluene (wt%)

7.39E–04

Time (h)

5.72E–06

–3.16E–05

Toluene in L1 (wt%)

DMSO (wt%)

1.95E–05

Time (h)

7.33E–06

–4.64E–04 –4.78E–06

3.26E–06

(c)

6.12E–04

(d)

Time (h)

Figure 4.21 Dynamic responses of toluene (plot (a)), MEA (plot (b)), and DMSO (plot (c)) purities, and toluene impurity in L1 (plot (d)) in CC and TC schemes for ±20% changes in the feed flow rate; given value for each curve is the corresponding steady-state deviation at 22 hours.

larger steady-state deviations for the feed flow rate changes in TC scheme compared to CC scheme, especially for toluene and MEA purities. This is because extractive (C1) and rectifying (C2) sections are typically multi-component separation columns and have strong interaction among process variables that lead to significant effect of composition profile in the column on the operating temperature of the selected sensitive tray. The steady-state deviation of DMSO purity is very small in TC scheme since C3 section approximates a binary separation column, i.e. the main components in C3 are only MEA and DMSO. Figure 4.21d shows that toluene impurity in L1 has larger steady-state deviations in TC scheme than those in CC scheme, which directly leads to large steady-state deviations in toluene and MEA purity in TC scheme. This is because toluene impurity in L1 is the main reason for the change in toluene impurity in MEA product and MEA impurity in toluene product. These steady-state deviations are relatively small; if required, they can be reduced by temperature difference control schemes; for details, see Feng et al. (2018). In addition, responses in CC scheme are slower than those in TC scheme; for example, toluene purity returns to a new steady-state value at 22 hours for CC scheme and 6 hours for TC scheme. This is a typical characteristic of composition controller, mainly due to time delay of composition measurement and inherent

113

4 Dynamic Simulation and Control of Intensified Chemical Processes

2.96E–05

1.59E–04

8.59E–04

–1.30E–04

MEA (wt%)

–1.04E–03

–1.15E–03

–4.10E–05

(a)

(b)

Time (h)

3.00E–05

3.15E–05

–3.59E–05 –3.11E–06

Time (h)

–3.55E–06 Toluene in L1 (wt%)

Toluene (wt%)

9.88E–04

DMSO (wt%)

114

7.48E–04

9.17E–06

(c)

Time (h)

(d)

–5.59E–04

Time (h)

Figure 4.22 Dynamic responses of toluene (plot (a)), MEA (plot (b)) and DMSO (plot (c)) purities, and toluene impurity in L1 (plot (d)) in CC and TC schemes for ±10% changes in toluene mole fraction in the feed stream; the given value for each curve is the corresponding steady-state deviation at 22 hours.

composition dynamic features (e.g. as given in Table 4.4, K p of composition controllers is smaller than that of temperature controllers while 𝜏 I of composition controllers is larger than that of temperature controllers); for this and other reasons (e.g. costs), TC scheme is often preferred in industrial practice. Figure 4.22 compares the closed-loop responses of purities of toluene, MEA, and recycled DMSO in CC and TC schemes, when subjected to ±10% changes in toluene mole fraction in the feed stream. Like those for ±20% feed flow rate disturbances, responses of toluene and MEA purities for feed composition disturbances display larger transient deviations in CC scheme than those in TC scheme. For example, when toluene mole fraction in the feed stream decreases by 10%, toluene purity at the top of C1 first decreases to 98.76 wt% (−0.74 wt%), then slowly increases to 99.79 wt% (+0.29 wt%), and then returns to the target purity. The large steady-state deviation in the response of toluene impurity in L1 in TC scheme leads to relatively large steady-state deviations in the responses of toluene and MEA purity, compared to those in CC scheme. On the other hand, CC scheme shows a slower and more oscillatory response than TC scheme for feed composition disturbances as well. Figure 4.23 shows the dynamic responses of the controlled variables in CC scheme, for feed flow rate and composition disturbances. Herein, xMEA in D1 (plots

xMEA in D1 (wt%)

xDMSO on 6th stage of C2 (wt%)

4.6 Case Study

(e)

Time (h)

(d)

Time (h)

xMEA in B1 (°C)

Time (h)

xMEA in B1 (°C)

(c)

(b) xDMSO on 6th stage of C2 (wt%)

Time (h)

xMEA in D1 (wt%)

(a)

Time (h)

(f)

Time (h)

Figure 4.23 Dynamic responses of controlled variables in CC scheme; xMEA in D1 (plots (a) and (c)) and B1 (plots (e) and (f)) are the impurity of MEA in C1 distillate and C3 bottom streams, respectively; xDMSO on the 6th stage of C2 is the mole fraction of DMSO on the 6th stage of C2 (plots (b) and (d)).

(a) and (c) in Figure 4.23) and B1 (plots (e) and (f)) are the impurities of MEA in C1 distillate and C3 bottom stream (i.e. recycled entrainer) and are the controlled variables of CC1 and CC3, respectively. xDMSO on the 6th stage of C2 (plots (b) and (d) in Figure 4.23) is the mole fraction of DMSO on 6th stage of C2, and it is the controlled variable of CC2. The dynamic responses of the controlled variable, i.e. toluene impurity in L1, for controller CC3, are already shown in Figures 4.21 and 4.22. All the controlled variables return to their respective set points, although their responses are slow, for both feed flow rate and composition disturbances, especially for feed composition changes. Figure 4.24 shows the dynamic responses of sensitive tray temperatures of temperature controllers in TC scheme for both feed flow rate and composition disturbances. All these controlled temperatures return to their set points in about four hours after a disturbance in the feed (compared to around 20 hours for controlled compositions to return to their set points in CC scheme in Figure 4.23). This further confirms that temperature controllers often respond faster to (measured or unmeasured) process disturbances than composition controllers; therefore, for this and lower costs, TC scheme is preferred over CC scheme in industrial practice.

115

Time (h)

(e)

(g)

Time (h)

(h)

Time (h)

(c)

Time (h)

(f)

Time (h)

Time (h)

T43 (°C)

T9 (°C)

T34 (°C)

(b)

T43 (°C)

(d))

T34 (°C)

T9 (°C)

T13 (°C)

Time (h)

T13 (°C)

(a)

Time (h)

Figure 4.24 Dynamic responses of controlled variables (sensitive temperatures) in TC scheme, to feed flow rate disturbances (plots (a–c) and (g)) and feed composition disturbances (plots (d–f) and (h)).

R2IF

R1IF

R2IF

4.6 Case Study

Time (h)

(c)

(b)

Time (h)

QR (GJ/h)

Time (h)

RV

(a)

Time (h)

(d)

Time (h)

Figure 4.25 Dynamic responses of four manipulated variables of quality controllers in TC and CC schemes: disturbances in the feed flow rate.

Figure 4.25 shows the closed-loop responses of four manipulated variables of quality controllers in both CC and TC schemes for ±20% changes in feed flow rate. Final values of these manipulated variables at time of 22 hours are very close, which indicates the feasibility and effectiveness of TC scheme for replacing CC scheme. The large transient deviations in the responses of R1 /F and RV in TC scheme result in faster dynamic responses in TC scheme than those in CC scheme, for rejecting feed flow rate disturbances. In addition, RV is approximately constant after oscillation. For ±10% changes in toluene mole fraction in the feed stream, closed-loop responses of manipulated variables in CC and TC schemes are presented in Figure 4.26. All these manipulated variables reach steady state by 22 hours; their steady-state values are different due to component/mass conservation of the system. In addition, TC scheme responds faster than CC scheme, to changes in toluene content in the feed stream. Note that the dynamic responses of the manipulated variables of level, pressure, and flow controllers are not shown in Figures 4.25 and 4.26 and are not discussed in this section since they only relate to material balances of the system. IAE values at T = 22 hours for the transient responses of purities of two products and recycled DMSO under different disturbances are given in Figure 4.27. For toluene and MEA purities in TC scheme under feed flow rate disturbances, they are much smaller than those in CC scheme. When toluene mole fraction in the feed stream increases by 10%, IAE for toluene purity in TC scheme is much smaller than that in CC scheme, whereas IAE for MEA purity in TC scheme is larger than that in

117

R2IF

R1IF

4 Dynamic Simulation and Control of Intensified Chemical Processes

(a)

(b)

Time (h)

RV

QR (GJ/h)

Time (h)

(c)

(d)

Time (h)

Time (h)

Figure 4.26 Dynamic responses of manipulated variables in TC and CC schemes: disturbances in the feed composition. 0.04

0.04

0.03

0.03

0.02 0.01 0.00

(a)

IAE for MEA

IAE for Toluene

Disturbance

0.02 0.01 0.00

F + 20% F − 20% T + 10% T − 10%

F + 20% F − 20% T + 10% T − 10%

(b)

Disturbance

0.005

IAE for DMSO

118

0.004 0.003 0.002 0.001 0.000

(c)

F + 20% F − 20% T + 10% T − 10%

Disturbance

Figure 4.27 IAE values of responses of purities under feed flow rate and feed composition disturbances in TC and CC schemes: (plot (a)) toluene product, (plot (b)) MEA product, and (plot (c)) recycled DMSO.

4.6 Case Study

CC scheme. This is due to large steady-state deviations in the response of MEA purity in TC scheme. However, when toluene mole fraction in the feed stream decreases by 10%, IAE for toluene purity in TC scheme is much larger than that in CC scheme, whereas IAE for MEA purity in TC and CC schemes is very close. This observation is different from that for the increase in toluene mole fraction in the feed stream, and it may be due to complex interactions in intensified distillation (i.e. EDWC). All the IAE values for the dynamic responses of purities of two products and recycled DMSO, shown in Figure 4.27, are presented in Tables 4.5 and 4.6. The steady-state errors (SSEs) values for the dynamic responses of purities of two products and recycled DMSO shown in Figures 4.21 and 4.22, are also presented in Tables 4.5 and 4.6. In summary, both TC and CC schemes can effectively handle ±20% step changes in feed flow rate and ±10% step changes in feed composition, which indicates that EDWC can be efficiently operated under such process disturbances. From process safety point of view, these dynamic simulation results can help the users to conduct model-based hazard and operability analysis (HAZOP) (Berdour, 2018). For example, Patle et al. (2021) proposed a model-based HAZOP that involves Table 4.5 IAE and final steady-state error for the dynamic responses of purities of two products and recycled DMSO for CC scheme. Disturbance Purity of

Item

F + 20%

F − 20%

zT + 10%

zT − 10%

Toluene

SSE

−1.60E-05

1.31E-05

−1.30E-04

1.59E-04

MEA DMSO

IAE

0.004433

0.006458

0.014662

0.032849

SSE

−1.20E-05

7.13E-06

−4.10E-05

2.96E-05

IAE

0.003031

0.006114

0.023958

0.022941

SSE

5.72E-06

3.26E-06

2.85E-06

−3.11E-06

IAE

9.84E-04

0.001817

9.39E-04

0.002158

Table 4.6 IAE values and final steady-state error for the dynamic responses of purities of two products and recycled DMSO for TC scheme. Disturbance Purity of

Item

F + 20%

Toluene

SSE

4.43E-04

MEA DMSO

F − 20%

zT + 10%

zT − 10%

−4.56E-04

−1.04E-03

9.88E-04

IAE

0.009029

0.009286

0.022741

0.021171

SSE

−9.74E-04

7.39E-04

8.59E-04

−1.15E-03

IAE

0.020703

SSE

−3.16E-05

IAE

0.001723

0.0154 1.95E-05 0.00134

0.017621

0.023467

−3.59E-05

3.15E-05

5.28E-04

4.56E-04

119

120

4 Dynamic Simulation and Control of Intensified Chemical Processes

dynamic simulations; they illustrated the proposed method for safety analysis of a formic acid process involving a reactive DWC. In practice, process safety is guaranteed by both the process control system (PCS) and the process safety interlock system (PSIS). Herein, PCS aims to monitor and control the process (controlled) variables within predefined limits for safe and efficient operation, e.g. sensitive tray temperatures in Figure 4.24 are controlled at the desired setpoint by adjusting their corresponding manipulated variables. PSIS is a protective response initiated on the detection of a process hazard, e.g. a valve fails to open or close. PCS and PSIS share the measurement devices and final control elements (e.g. valves); however, they have relatively independent software systems. PSIS consists of one or more logic conditions that detect out-of-limit process conditions and respond by driving the finial control elements to a safe state through a series of logic operations, e.g. pump P105 in Figures 4.17 and 4.20 will be shut down when the sump level of C3 decreases to its lower safety limit.

4.7 Conclusions Dynamic simulation is essential to understand the dynamic characteristics, operability, and controllability of processes, and for process design, process revamping, safety analysis, process control, designing (semi-)batch processes, designing process start-up and shut-down procedures, and operator training. This chapter provides a comprehensive procedure for dynamic simulation and control of (intensified) chemical processes. The procedure has 5 main steps: (i) steady-state process simulation and optimization; (ii) initialization for dynamic simulation, including the size of equipment and addition of required pumps and valves, as well as specification of required process parameters such as pressure drop for fluid flow; (iii) inventory and quality control loop design; (iv) tuning of temperature and composition controllers; and (v) results analysis and comparison for selecting and optimizing the process control scheme. Challenges in dynamic simulation and control of ICPs are outlined. Finally, a case study for the dynamic simulation and control of EDWC process for separating toluene and MEA mixture using DMSO as the entrainer is presented in detail to illustrate the dynamic simulation and control procedure as well as the challenges described earlier. It includes detailed analysis of control performance, such as oscillations and steady-state deviations in responses of product purities and control and controlled variables, as well as IAE comparison.

Acronyms ACM ATV CC CDOF CSTR

aspen custom modeler auto-tune variation composition control control degrees of freedom continuous stirred tank reactor

References

DAE DMSO DWC EDWC HAZOP IAE ICP ISE ITAE MEA MPC NRTL ODE PCS PDE PID PR PSIS PWC RD SSE TC

differential-algebraic equation dimethyl sulfoxide dividing-wall column extractive dividing-wall column hazard and operability analysis integral of absolute error intensified chemical process(es) integral of squared error integral of product of time and absolute error 2-methoxy ethanol model predictive control nonrandom two-liquid ordinary differential equation process control system partial differential equation proportional-integral-derivative Peng-Robinson process safety interlock system plantwide control reactive distillation steady-state error temperature control

References Becker, L., Sturm, J., Eiden, F., and Holtmann, D. (2023). Analyzing and understanding the robustness of bioprocesses. Trends in Biotechnology 41 (8): 1013–1026. https://doi .org/10.1016/j.tibtech.2023.03.002. Berdour, F., Villemur, C., Olivier-Maget, N., and Gabas, N. (2018). Dynamic simulation for risk analysis: application to an exothermic reaction. Process Safety and Environmental Protection 113: 149–163. D’Souza, R.N., Azevedo, A.M., Aires-Barros, M.R. et al. (2013). Emerging technologies for the integration and intensification of downstream bioprocesses. Pharmaceutical Bioprocessing 1 (5): 423–440. Dwivedi, D., Strandberg, J.P., Halvorsen, I.J. et al. (2012). Active vapor split control for dividing-wall columns. Industrial & Engineering Chemistry Research 51 (46): 15176–15183. Egger, T. and Fieg, G. (2017). Enzymatic catalyzed reactive dividing wall column: experiments and model validation. AICHE Journal 63 (6): 2198–2211. El Wajeh, M., Mhamdi, A., and Mitsos, A. (2023). Dynamic modeling and plantwide control of a production process for biodiesel and glycerol. Industrial & Engineering Chemistry Research 62 (27): 10559–10576. Feng, Z., Shen, W., Rangaiah, G.P., and Dong, L. (2018). Proportional-integral control and model predictive control of extractive dividing-wall column based on

121

122

4 Dynamic Simulation and Control of Intensified Chemical Processes

temperature differences. Industrial & Engineering Chemistry Research 57 (31): 10572–10590. Feng, Z., Shen, W., Rangaiah, G.P., and Dong, L. (2019a). Closed-loop identification and model predictive control of extractive dividing-wall column. Chemical Engineering and Processing: Process Intensification 142: 107552. Feng, Z., Shen, W., Rangaiah, G.P., and Dong, L. (2019b). Process development, assessment, and control of reactive dividing-wall column with vapor recompression for producing n-propyl acetate. Industrial & Engineering Chemistry Research 58 (1): 276–295. Gani, R., Soerensen, E.L., and Perregaard, J. (1992). Design and analysis of chemical process through DYNSIM. Industrial & Engineering Chemistry Research 31 (1): 244–254. Harvianto, G., Kim, K.H., Kang, K.J., and Lee, M. (2019). Optimal operation of a dividing wall column using an enhanced active vapor distributor. Chemical Engineering Research and Design 144: 512–519. Halvorsen, I.J. and Skogestad, S. (2004). Shortcut analysis of optimal operation of Petlyuk distillation. Industrial & Engineering Chemistry Research 43 (14): 3994–3999. Kallakuri, R. and Bahuguna, P.C. (2021). Role of operator training simulators in hydrocarbon industry – a review. International Journal of Simulation and Modelling 20 (4): 649–660. Li, C., Li, J., Li, D. et al. (2020). Experimental study and CFD numerical simulation of an innovative vapor splitter in dividing wall column. AICHE Journal 66 (8): 16266. Li, L.M., Guo, L.J., Tu, Y.Q. et al. (2017). Comparison of different extractive distillation processes for 2-methoxyethanol/toluene separation: design and control. Computers & Chemical Engineering 99: 117–134. Ling, H. and Luyben, W.L. (2010). Temperature control of the BTX divided-wall column. Industrial & Engineering Chemistry Research 49 (1): 189–203. Luyben, W.L. (2013). Distillation Design and Control Using Aspen Simulation. New York: Wiley. Luyben, W.L. and Chien, I.-L. (2009). Design and Control of Distillation Systems for Separating Azeotropes. Hoboken, NJ: Wiley. Luyben, W.L. (2018). Vapor split manipulation in extractive divided-wall distillation columns. Chemical Engineering and Processing Process Intensification 126: 132–140. Murthy Konda, N.V.S.N., Rangaiah, G.P., and Krishnaswamy, P.R. (2006). A simple and effective procedure for control degrees of freedom. Chemical Engineering Science 61: 1184–1194. Panunzi, A., Moroni, M., Mazzelli, A., and Bravi, M. (2022). Industrial case-study-based computational fluid dynamic (CFD) modeling of stirred and aerated bioreactors. Industrial & Engineering Chemistry Research 7 (29): 25152–25163. Patle, D.S., Agrawal, V., Sharma, S., and Rangaiah, G.P. (2021). Plantwide control and process safety of formic acid process having a reactive dividing-wall column and three material recycles. Computers and Chemical Engineering 147: 107248. Patle, D.S., Ahmad, Z., and Rangaiah, G.P. (2014). Operator training simulators in the chemical industry: review, issues, and future directions. Reviews in Chemical Engineering 32 (2): 199–216.

References

Peng, J., Edgar, T.F., and Eldridge, R.B. (2003). Dynamic rate-based and equilibrium models for a packed reactive distillation column. Chemical Engineering Science 58 (12): 2671–2680. Premkumar, R. and Rangaiah, G.P. (2009). Retrofitting conventional column systems to dividing-wall columns. Chemical Engineering Research and Design 87 (1): 47–60. Qian, X., Lin, K.H., Jia, S. et al. (2023). Nonlinear model predictive control for dividing wall columns. AICHE Journal 69 (2): e18062. Rafiei, M. and Ricardez-Sandoval, L.A. (2019). New frontiers, challenges, and opportunities in integration of design and control for enterprise-wide sustainability. Computers and Chemical Engineering 132: 106610. Ruiz-Ruiz, F., Benavides, J., and Rito-Palomares, M. (2013). Scaling-up of a B-phycoerythrin production and purification bioprocess involving aqueous two-phase systems: practical experiences. Process Biochemistry 48 (4): 738–745. Shitahun, A., Ruge, V., Gebremedhin, M. et al. (2013). Model-based dynamic optimization with OpenModelica and CasADi. IFAC Proceedings Volumes 46 (21): 446–451. Sigue, S., Abderafi, S., Vaudreuil, S., and Bounahmidi, T. (2023). Design and steady-state simulation of a CSP-ORC power plant using an open-source co-simulation framework combining SAM and DWSIM. Thermal Science and Engineering Process 37: 101580. Stankiewicz, A. and Moulijn, J.A. (2002). Process intensification. Industrial and Engineering Chemistry Research 41: 1920–1924. Tututi-Avila, S., Jiménez-Gutiérrezb, A., and Hahn, J. (2014). Control analysis of an extractive dividing-wall column used for ethanol dehydration. Chemical Engineering and Processing: Process Integration 82: 88–100. Wolff, E.A. and Skogestad, S. (1995). Operation of integrated three-product (Petlyuk) distillation columns. Industrial & Engineering Chemistry Research 34: 2094–2103. Yildirim, O., Kiss, A.A., and Kenig, E.Y. (2011). Dividing wall columns in chemical process industry: a review on current activities. Separation and Purification Technology 80: 403–417.

123

125

5 Safety Analysis of Intensified Chemical Processes Masrina Mohd Nadzir, Zainal Ahmad, and Syamsul Rizal Abd Shukor School of Chemical Engineering, Engineering Campus, Universiti Sains Malaysia, 14300 Nibong Tebal, Penang, Malaysia

5.1 Introduction Chemical process industry (CPI) has always been subjected to tight safety regulations by local authorities and meeting international safety standards. Global competition and meeting market’s ever-growing needs represent a multitude of challenges to which the CPI must respond quickly. Process intensification is a promising technology that is touted to be the future of CPIs that aim to improve efficiency and sustainability by increasing their output per unit of time, reducing energy consumption, and minimizing waste (Curcio, 2013). Although originally defined as a strategy to downsize chemical plants to meet specific production targets (Ramshaw, 1999), process intensification has evolved to encompass a holistic approach to process design that aims to enhance efficiency, reduce environmental footprint, and improve process safety (Reay et al., 2008), including businesses, among other benefits. These benefits are reflected in the fact that the use of microreactors and other process intensification devices can help reduce the size of chemical plants, resulting in lower material and energy consumption. The downsizing of facilities also leads to a reduction in waste generation and emissions, which ultimately helps reduce the environmental footprint of the process. Process intensification technologies can also improve process safety by minimizing the handling of hazardous materials and the potential for breakthrough reactions (Stankiewicz and Moulijn, 2000). For example, continuous-flow processing, which minimizes the inventory of reactive chemicals, can reduce the risk of accidents. Intensified processes are generally considered safer owing to their smaller footprint and reduced inventory of hazardous materials or energy. Nevertheless, it is important to recognize that in many intensified processes, multiple phenomena occur simultaneously in a confined space and at a high rate, complicating process development and control, which leads to hazardous conditions (Stankiewicz and Moulijn, 2000). The implementation of process intensification technology often leads to novel technologies with limited experience, and thus it may not be entirely Control and Safety Analysis of Intensified Chemical Processes, First Edition. Edited by Dipesh Shikchand Patle and Gade Pandu Rangaiah. © 2024 WILEY-VCH GmbH. Published 2024 by WILEY-VCH GmbH.

126

5 Safety Analysis of Intensified Chemical Processes

safe and could lead to unforeseen accidents (Curcio, 2013). Moreover, optimizing a process involves trade-offs between technological, economic, safety, and environmental standards. Hence, reaching an agreement on several safety precautions and investment expenses for process intensification technology can be particularly challenging. It is crucial to realize that energy-saving measures or their recycling may result in unacceptable risks leading to catastrophic industrial accidents. Luyben and Hendershot (2004) highlight the potential conflicts, particularly between the desire to limit the potential scale of accidents (limitation) and the need to manage the rapid changes that can occur in smaller, more intensified systems (intensification). Intensification in chemical processes, which involves using smaller liquid holdups, can enhance safety and environmental conditions by reducing potential leak volumes, but it also increases the system’s sensitivity to disturbances, leading to rapid, potentially unsafe changes before corrective actions can be implemented. This creates a conflict with the limitation strategy, which seeks to mitigate equipment failure effects, as smaller equipment can react more swiftly to changes, causing larger shifts in operating conditions and process output. Therefore, the need to understand and implement proper safety measures is paramount. It is essential to understand proper safety analysis for an intensified process, as the need for detailed safety analysis is recommended at the early design stage for any given process intensification technology (Ebrahimi et al., 2012). The realization of inherently safe process design is a fundamental approach for solving safety problems in chemical processes, particularly for process intensification technology, which is critical for preventing serious safety accidents. As rightly pointed out by Trevor Kletz, “what you don’t have can’t leak,” and he emphasized the importance of inherently safe design concepts in mitigating safety risks (Kletz, 1978). This chapter presents a comprehensive safety analysis, particularly for intensified processes, that aims to increase the efficiency and sustainability of chemical processes. Key challenges and opportunities associated with safety in process intensification are discussed with the inclusion of relevant developments in safety assessment, hazard identification, risk analysis, and mitigation strategies. Additionally, it highlights the notable absence of quantitative analysis on the dynamics’ impact on inherently safer design (ISD) and demonstrates this through four illustrative examples on the dynamic disadvantages of small-holdup designs. These examples include a two-column distillation system, two alternative nitration CSTR reactor systems, a CSTR nitration reactor subject to instrumentation failure, and a distillation column with conventional tray liquid holdups compared to one with increased holdups. This chapter provides practical guidance for engineers and researchers involved in the design, operation, and maintenance of process intensification systems, as well as for regulators and policymakers who need to ensure the safe deployment of process intensification technologies.

5.2 Safety Analysis in Chemical Process Industry The primary purpose of process safety analysis (PSA) is to identify and assess potential risks and hazards associated with a process. In particular, PSA aims to predict

5.2 Safety Analysis in Chemical Process Industry

and prevent possible catastrophic failures that could lead to accidents, injuries, and environmental damage. PSA is also used to evaluate the safety of process modifications or expansions and to develop plans for responding to emergency situations. The objectives of PSA can be broadly categorized into three main areas: risk identification, risk assessment, and risk management. Through PSA, sources of potential hazards within a given process can be identified and classified according to their level of risk. The rapid advancement of chemical technologies in recent decades has necessitated the critical importance of risk assessment in managing potential hazards (Khan and Amyotte, 2003). These hazards, if left unaddressed, possess the capacity to cause catastrophic consequences for personnel, the environment, and industrial assets. In this context, effective decision-making regarding safety-related issues in chemical process facilities relies heavily on accurate and relevant risk information (Yuan et al., 2023). Access to risk information is key to the successful management of industrial facilities. Gaining insight into potential risks allows for better risk management decisions, whereas a lack of information can result in suboptimal decision-making. Risk assessment is particularly crucial in the design and operation of chemical processes where high temperatures, pressures, and/or reactant concentrations can pose significant hazards (Kletz, 1980). Historically, the CPI has relied on accumulated industry experience for risk assessment in its facilities and operations, resulting in a commendable safety record compared with other sectors (Carter and Hirst, 2000). However, as the development and implementation of novel process technologies increases, the applicability of historical experience in ensuring safety may diminish (Mannan et al., 2005), with the CPI’s dependence on such experience potentially proving inadequate for addressing the risks associated with these emerging technologies. This can undermine the effectiveness of conventional risk assessment approaches for emerging technologies, necessitating the development of techniques to predict and assess risks in the absence of relevant historical data. Putting things into context, potentially unseen hazardous process intensification technologies lack the prior experience necessary to assess the safety aspects of advanced technology in their designs. The absence of relevant historical data has led researchers, scientists, and industrial practitioners to develop techniques for predicting and assessing risks. Ensuring the safe and reliable operation of chemical processes requires thorough safety analysis, which involves assessing and evaluating potential risks and hazards to identify and mitigate them (Crowl and Louvar, 2019). This involves assessing and evaluating the potential risks and hazards associated with a chemical process to identify and mitigate these risks. Various safety analysis techniques, such as hazard and operability (HAZOP) studies , failure modes and effects analysis (FMEA), and layer of protection analysis (LOPA), have proven helpful in mitigating risks (Khan and Abbasi, 1998). Comprehensive risk assessment encompasses several steps, from hazard identification to risk assessment to risk control (Aven, 2016). Compliance with relevant standards and guidelines, such as those of the American Institute of Chemical Engineers (AIChE) or the International Organization for Standardization (ISO), further strengthens the risk management process (Bridges et al., 2011).

127

128

5 Safety Analysis of Intensified Chemical Processes

Additionally, it is essential to consider other factors that may impact safety, including the design and layout of the facility, the accessibility of emergency shutdown systems and procedures, and the training and qualifications of operators. Implementing reliable safety management systems (SMSs) that can recognize and eliminate possible risks before they materialize is also crucial for ensuring the continued safe operation of chemical processes (Khan and Amyotte, 2003).

5.2.1

Safety Analysis Tools

PSA is the systematic evaluation of potential risks and hazards that may arise during the operation of a process, with the aim of identifying, assessing, and controlling risks associated with hazardous chemicals, equipment, and procedures in the workplace (Kletz, 1999). Conducting a comprehensive safety analysis of a chemical process is of utmost importance, particularly when incorporating process intensification technologies, as these may introduce potential hazards resulting from elevated temperatures, pressures, and/or reactant concentrations (Stankiewicz and Moulijn, 2000). Primarily, this analysis serves to safeguard personnel’s well-being by identifying potential hazards and instituting appropriate preventive measures. Furthermore, it contributes to environmental protection by detecting and mitigating the risks associated with the unintended release of harmful chemicals or the occurrence of fires and explosions (Mannan, 2012). A thorough safety evaluation also enhances the reliability and efficiency of a process by identifying and addressing potential hazards, thereby mitigating the likelihood of unscheduled shutdowns or disruptions (Crowl and Louvar, 2019). Compliance with regulatory requirements is imperative to circumvent punitive consequences such as fines, penalties, and legal action. Safety analysis is an integral component of risk assessment in the design and operation of chemical processes to ensure their safe and reliable operation. By employing methodologies such as HAZOP analysis, FMEA, fault tree analysis (FTA), event tree analysis (ETA), and probabilistic risk assessment techniques and implementing robust SMSs, potential hazards can be effectively identified and mitigated, ensuring the safety of workers and the environment (Khan and Abbasi, 1998; Aven, 2016). It is imperative that organizations prioritize risk assessment and safety analysis in the development and operation of chemical technologies to minimize the potential for catastrophic consequences (Tweeddale, 2003). 5.2.1.1 Hazard Identification

Hazard identification is a crucial initial step in ensuring safety within the CPIs. It is a process of identifying and recognizing potential dangers or risks in a given situation. Effective hazard identification plays a critical role in the development of appropriate control measures to mitigate or eliminate risks that may cause injury, property damage, environmental degradation, or undesirable outcomes. It provides a foundation for risk assessment and management, enabling the maintenance of safety, environmental protection, and overall well-being in various contexts. Therefore, prioritizing hazard identification is paramount, as it contributes to the enhancement of safety, environmental protection, and overall well-being.

5.2 Safety Analysis in Chemical Process Industry

This process entails recognizing potential sources of harm and their associated consequences. Various systematic techniques can be employed to identify hazards, among the widely used methodologies, including: ●

●

●

●

●

Process Hazard Analysis (PHA): A comprehensive assessment of the potential hazards associated with a process includes the identification of possible causes and consequences of accidents as well as recommendations for risk mitigation. Process hazard analysis (PHA) is a comprehensive assessment technique employed to identify potential hazards associated with chemical processes (Khan and Amyotte, 2003). This method examines the possible causes and consequences of accidents and provides recommendations for risk mitigation (Center for Chemical Process Safety, 2008). The PHA process typically encompasses a multidisciplinary team of experts, which ensures that various perspectives are considered in hazard identification and risk evaluation (Lees, 2012). Hazard and Operability Study (HAZOP): A structured and systematic examination of a process that employs a set of guidewords to evaluate potential deviations from the intended design and their potential consequences for the facility as a whole. The strength of this methodology lies in its ability to adhere to a plant’s process flow diagrams (PFDs) and piping and instrumentation diagrams (P&IDs) and divide the design into manageable sections with defined boundaries called nodes, ensuring analysis of each piece of equipment in the process. The analysis is carried out by a small, carefully chosen interdisciplinary team that has adequate expertise and knowledge of the facility and is able to answer the majority of the field’s queries (Dunjó et al., 2010). Failure Modes and Effects Analysis (FMEA): A bottom-up approach that systematically analyzes individual components or process steps to identify failure modes and assess their effects on the overall system. The process involves evaluating as many components, assemblies, and subsystems as possible to discern the probability, severity, and detectability of each failure mode to prioritize risks and identify the most critical areas that require modification to mitigate the risk of failure and enhance reliability (Pasman, 2015). FMEA is a data-driven and team-oriented approach that aims to determine the relative impact of various failure modes on productivity goals, thereby facilitating the prioritization of the most critical failure modes for further analysis and improvement. What-If Analysis: A qualitative risk assessment tool that explores potential hazards by posing hypothetical questions about process deviations and their potential consequences. This method involves brainstorming sessions with experienced multidisciplinary teams to analyze various scenarios, including deviations from standard operating conditions, equipment malfunctions, and human errors (Center for Chemical Process Safety, 2008). By asking “what-if” questions, formulated based on experience and applied to existing drawings and process descriptions, the team identifies possible risks, evaluates their likelihood and severity, and recommends mitigation measures. Checklist Analysis: A systematic qualitative approach to hazard identification, which utilizes predefined lists of questions, prompts, or items related to potential hazards and risks associated with a process or operation. Checklist analysis

129

130

5 Safety Analysis of Intensified Chemical Processes

aimed at ensuring compliance with standard practices and identifying potential hazards. Its primary objective is to identify hazards and assess their consequences by comparing the current state of the system against established criteria or standards (Khan and Abbasi, 1998). The process typically involves a multidisciplinary team of experts who systematically evaluate each item in the checklist to determine the presence of potential hazards and the need for further analysis or action (Ericson, 2016). This technique can be applied at any stage of a process’s lifetime and serves as an accessible method for familiarizing inexperienced personnel with a process and providing a common basis for management review of analysts’ assessments. A typical checklist analysis entails using a list of specific items to identify known hazards, design deficiencies, and potential incident scenarios associated with common process equipment and operations. For optimal effectiveness, checklists should be tailored to an individual company, plant, or product (Manuele, 2008). The simplicity and versatility of checklist analysis make it an essential tool for hazard identification, ensuring equipment conforms with accepted standards, and identifying areas requiring further evaluation. In some cases, analysts may combine a general checklist with another hazard evaluation method to uncover common hazards that the checklist alone might miss. 5.2.1.2 Risk Assessment

Risk, within the context of process safety in the chemical industry, can be succinctly defined as the mathematical product of the frequency or likelihood of a hazardous event’s occurrence and the severity of the potential consequences resulting from such an event. This definition encapsulates two fundamental dimensions of risk: probability and impact. Frequency or likelihood refers to the statistical probability of a specific hazardous event occurring within a given time frame. This is often determined through historical data, predictive models, and expert judgment. It is a measure of the system’s vulnerability to potential threats, taking into consideration the effectiveness of existing safety measures and controls. Whereas, severity of consequence quantifies the potential impact of a hazardous event should it occur. This could encompass a range of outcomes, from minor operational disruptions to catastrophic incidents leading to significant human, environmental, and financial losses. The severity is typically assessed based on worst-case scenarios, thereby ensuring that risk management strategies are robust and capable of mitigating the most severe potential outcomes. By conceptualizing risk as the product of these two dimensions, we can create a comprehensive and quantifiable measure of potential harm. This approach allows for a more nuanced understanding of risk, enabling the prioritization of risk management efforts and the allocation of resources in a manner that effectively reduces the likelihood of hazardous events and mitigates their potential impacts. This risk-based approach is fundamental to enhancing process safety in the CPI and is a cornerstone of proactive SMSs. Risk assessment is an inherent part of a broader risk management strategy that involves systematic evaluation of the likelihood and consequences of identified hazards. In the context of process intensification, a risk assessment framework is a

5.2 Safety Analysis in Chemical Process Industry

Table 5.1

Examples of risk assessment method widely used in industry.

Method

Description

Layer of Protection Analysis (LOPA)

A semi-quantitative method that evaluates the adequacy of existing safety measures by comparing the frequency of initiating events with the required risk reduction.

Fault Tree Analysis (FTA)

A top-down, deductive approach that quantifies the probability of a specific undesired event occurring by identifying combinations of initiating events and failure modes.

Event Tree Analysis (ETA)

An inductive approach that estimates the likelihood and consequences of various potential outcomes following an initiating event.

Risk Matrix

An essential tool used in risk assessment methodologies, including but not limited to HAZOP and FMEA. It is a two-dimensional grid that categorizes risks based on their severity and likelihood of occurrence. It serves as a visual representation of the risks involved in a particular situation, providing a systematic and intuitive way to prioritize hazards based on their potential impact and the likelihood of occurrence.

structured approach designed to identify, analyze, and manage potential hazards associated with the application of these novel processes. This process helps prioritize risks and develop appropriate risk management strategies, ultimately enhancing the safety and reliability of intensified processes. Risk assessment can involve both qualitative and quantitative methods. Qualitative risk analysis is quick but subjective, whereas quantitative risk analysis is optional, more detailed, objective, and involves contingency reserves and go/no-go decisions. Quantitative analysis requires more time and is more complex, with quantitative data being difficult to collect and quality data being prohibitively expensive (Volkan, 2021). Both or either quantitative and qualitative risk assessment methods can be employed, depending on how well the risk is known, and if it can be evaluated and prioritized in a timely manner, it may be possible to reduce the possible negative effects or increase the possible positive effects and take advantage of the opportunities (Bansal, 2019). Among the widely used risk assessments are given in Table 5.1: 5.2.1.3 Inherently Safer Design (ISD)

ISD aims to eliminate or minimize hazards at the source, rather than relying on engineered safety systems or administrative controls. The concept of inherent safety introduced by Kletz (1978, 1996) has been widely recognized by industry players, where this approach emphasizes the importance of addressing safety concerns during the conceptual and preliminary design phases, which can significantly reduce the likelihood of accidents, environmental impacts, and related costs. The four main strategies of ISD include (Gupta and Edwards, 2002): ●

Minimization: Reducing the quantity of hazardous materials in the process by using smaller quantities, less hazardous substances, or less hazardous operating conditions. This reduces the potential impact of a release or accident.

131

132

5 Safety Analysis of Intensified Chemical Processes ●

●

●

Substitution: Replace hazardous materials or processes with less hazardous alternatives. For example, use a less toxic or flammable solvent or a less reactive chemical in a reaction. Attenuation: Controlling process conditions to reduce the likelihood of hazardous events, such as explosions or runaway reactions. This can involve reducing temperatures, pressures, or reaction rates and using inherently safer equipment designs. Simplification: Designing processes and facilities to be less complex makes them easier to understand, operate, and maintain. This can reduce the potential for human error, equipment failures, or process upsets.

ISD is an iterative process that relies on informed decision-making. This approach necessitates taking into account the entire life cycle, the full spectrum of hazards and risks, and the possibility of risk transfer from one impacted population to another, with the inclusion of various technical and economic feasibility options. 5.2.1.4 Safety Instrumented Systems

Safety Instrumented Systems (SISs) are engineered solutions consisting of hardware and software controls designed to provide layers of protection in the event of equipment failure or process deviations (Lees, 2012). These systems are critical for maintaining process safety and protecting people, the environment, and assets. SISs are typically employed in CPIs where there is a potential for hazardous events and where the failure of equipment or process control systems could lead to severe consequences. SIS is designed following IEC 61511, IEC 61508, and ISA S84.01 standard guidelines. In general, SIS should function as designed for safety by isolating the process plant/system/unit during an emergency and must be independent of all other control systems for the same equipment or process. Key concepts related to SIS include: ●

●

●

Safety Instrumented Functions (SIFs): SIFs are the individual safety-related tasks performed by a SIS. These functions are intended to bring a process to a safe state in the event of a specific hazardous condition or process deviation. For example, a safety valve that releases pressure from a vessel when it exceeds a predefined threshold is an example of a SIF. Safety Integrity Level (SIL): A measure of the reliability of a SIF ranges from SIL 1 (lowest reliability) to SIL 4 (highest reliability), with higher SILs indicating a greater degree of risk reduction. The required SIL for a specific SIF is determined based on the assessed risk associated with the relevant hazardous event. Performance Requirements: The specification of necessary reliability, response time, and other performance criteria for each SIFs ensures that the SIF can effectively mitigate the associated risk. The specified performance requirements should be based on a thorough risk assessment, considering the potential consequences of a hazardous event and the likelihood of its occurrence.

5.2.1.5 Human Factors and Safety Culture

Human factors and safety culture play a critical role in ensuring process safety in the CPI. These factors can significantly impact safety performance by influencing the attitudes, behaviors, and decision-making of individuals at all levels within an

5.2 Safety Analysis in Chemical Process Industry

organization. A positive safety culture fosters an environment in which safety is a shared value and priority and where individuals feel responsible for their own safety and the safety of others. The successful adoption and operation of intensified processes require a strong safety culture and a thorough understanding of the potential human factors that may impact process safety. Key elements of promoting a positive safety culture within the context of process intensification technology include: ●

●

●

●

●

●

Understanding Novel Technologies: Process intensification often involves the use of new and innovative technologies that can introduce new hazards and challenges. Ensuring that employees have a comprehensive understanding of the intensified processes and associated risks is crucial to maintaining safe operations. This understanding can be achieved through targeted training programs and ongoing education on the specific technologies employed. Encouraging Open Communication and Reporting of Safety Concerns: A strong safety culture is characterized by open communication and willingness to report safety concerns, creating an environment in which individuals feel comfortable discussing safety issues and sharing lessons learned without fear of blame or retribution. Encouraging open communication and reporting of safety concerns, incidents, and near misses is vital for maintaining a positive safety culture. This openness is particularly important when dealing with process intensification technologies, as it can help identify potential issues with novel technologies and processes early on, allowing for timely mitigation of risks. Implementing a Just Culture: A just culture is one in which individuals are not unfairly punished for honest mistakes but are held accountable for willful violations of safety rules and procedures or for failing to learn from previous incidents, can help build trust and promote a positive safety culture. Fostering a just culture within an organization can promote a sense of responsibility and trust among employees, leading to improved safety performance when working with intensified processes. A just culture recognizes that human error is inevitable and that learning from mistakes is essential for improving safety performance. Proactive Management of Human Factors: In the implementation of process intensification technologies, it is essential to proactively address potential human factors that could impact process safety. This may include ergonomic design considerations, workload assessment, and human error analysis. By addressing these factors during the design and implementation stages of intensified processes, organizations can minimize the potential for accidents related to human error. Maintaining Competency and Training: Ensuring that employees have the necessary skills and knowledge to safely operate intensified processes is crucial. Comprehensive training programs should be developed and implemented, with a focus on the unique aspects and potential hazards associated with process intensification technology. Regular refresher training and ongoing competency assessments can help maintain a high level of skill and awareness among employees and foster a strong understanding of process safety principles. Continuous Improvement and Learning From Incidents: A strong safety culture is characterized by a commitment to continuous improvement and learning from incidents, both within the organization and from external sources. This involves

133

134

5 Safety Analysis of Intensified Chemical Processes

regularly reviewing and refining SMSs, processes, and practices, as well as seeking opportunities to learn from both internal and external sources. In the context of process intensification technologies, organizations should actively seek lessons learned from other companies and industries that have implemented similar technologies. This can help to identify potential risks and best practices for managing the unique challenges associated with process intensification. 5.2.1.6 Regulatory Framework and Compliance

Compliance with the regulatory framework governing process safety in the CPI is crucial for ensuring safe operations, particularly with process intensification technology. Key regulations and standards include the following: ●

●

●

Process Safety Management (PSM) Standard: PSM is a systematic approach for managing hazards associated with the process industries. It is a comprehensive regulation established by the U.S. Occupational Safety and Health Administration (OSHA) designed to identify, evaluate, and control process hazards in order to prevent catastrophic accidents (Mannan et al., 2005). PSM involves the use of various tools and techniques, including hazard analysis, risk assessment, and process design, to identify potential hazards and assess their associated risks. This also involves the development of procedures and systems to prevent and mitigate the effects of potential accidents. When implementing process intensification technology, organizations must ensure that their intensified processes are in compliance with PSM requirements, including employee training programs. Adequate resources, training, and oversight must be provided to ensure that workers are competent and knowledgeable about process safety hazards and controls. Seveso Directive: A European Union directive that sets safety requirements for establishments storing or handling large quantities of hazardous substances to prevent major accidents and limit their consequences (Mannan et al., 2005). Process intensification often leads to reduced inventories of hazardous materials, but organizations must still ensure compliance with the Seveso Directive when implementing intensified processes. Other National and International Standards: Organizations implementing process intensification technology should also be aware of other relevant national and international standards that may apply to their specific industries or regions. These may include guidelines from organizations such as the Center for Chemical Process Safety (CCPS), the AIChE, the ISO, and others.

Understanding and adhering to these regulations and standards is vital for maintaining safe operations and avoiding potential legal consequences and financial penalties associated with noncompliance. In the context of process intensification technology, adherence to regulatory frameworks and standards necessitates a multifaceted approach, such as: ● ●

comprehending the pertinent regulations to ensure compliance; conducting rigorous hazard identification and risk assessment for innovative technologies and process designs;

5.2 Safety Analysis in Chemical Process Industry ●

●

●

formulating and executing relevant policies, procedures, and training programs to guarantee safe operations; undertaking regular internal and external audits and inspections to validate compliance and identify areas requiring corrective actions; and meticulously managing modifications to processes, equipment, or operating procedures to maintain regulatory conformity, which encompasses hazard assessments for proposed alterations and updates to associated policies, procedures, and training.

5.2.1.7 Monitoring and Continuous Improvement

Monitoring and continuous improvement are essential components for maintaining safety in the CPI, particularly when implementing process intensification technology. The adoption of intensified processes often introduces new risks and challenges, making it crucial for organizations to continuously monitor their operations and update their safety practices to stay ahead of potential hazards. Key aspects of this process include: ●

●

●

●

Safety Performance Indicators: Establishing relevant safety performance indicators for process intensification safety allows organizations to track and assess their performance over time. These indicators can include metrics related to incident rates, near-miss reporting, equipment reliability, and employee training, among others. Regularly reviewing and analyzing performance indicator data can help organizations identify trends and areas of concern that warrant further investigation or intervention. Regular Audits and Inspections: Conducting systematic and periodic evaluations of SMSs, processes, and facilities is essential for monitoring and continuous improvement. These assessments can help identify areas of noncompliance with safety regulations, potential hazards, and opportunities for improvement in the design, operation, and maintenance of intensified processes. Management of Change (MOC): Effective MOC processes are crucial for maintaining safety during process intensification. Organizations must ensure that any changes to equipment, processes, or procedures are thoroughly evaluated for potential hazards and that necessary updates to safety practices are implemented in a timely manner. Continuous Improvement Culture: Fostering a culture of continuous improvement, in which employees are encouraged to identify opportunities for enhancing safety and are empowered to make changes, is essential for maintaining safety in process intensification. This culture can be supported through regular communication, recognition of safety achievements, and engagement in safety-related decision-making.

Regular monitoring of safety performance and the implementation of continuous improvement initiatives are critical components of maintaining safety in process intensification. By establishing and tracking relevant safety performance indicators, organizations can proactively address potential safety concerns and strive for excellence in PSM.

135

136

5 Safety Analysis of Intensified Chemical Processes

5.3 Process Intensification and Safety Analysis Process intensification and safety analysis are intricately correlated. Process intensification technology encompasses a wide range of innovations, including the use of microstructured equipment, and the integration of multiple unit operations into a single process step results in enhanced process efficiency, productivity, and sustainability while reducing the environmental impact (Keil, 2018). These advantages make process intensification technology an attractive option for process design and optimization. More importantly, it significantly improves the safety aspects by reducing the number of hazardous operations, minimizing the usage of harmful substances, reducing waste generations, and improving sustainability while simultaneously reducing the environmental footprint of industrial processes. Process intensification technology often involves the use of novel and complex equipment, which may pose unique safety risks if not adequately assessed and controlled. Designing and implementing process intensification strategies must take safety considerations into account, as these new technologies may require modifications in traditional safety protocols and may require additional safety barriers to minimize the possibility of incidents or accidents (Mannan et al., 2016). The integration of process intensification with safety analysis can lead to safer, more efficient, and more sustainable processes. Safety analysis tools can be used to identify and assess potential risks associated with the novel process intensification equipment, which can then be used to develop appropriate control strategies, such as process modifications, equipment redesigns, or the introduction of additional safety measures. This can significantly reduce the likelihood of incidents and accidents, making the process safer for workers and the surrounding environment. Some specific safety issues that must be considered in process intensification include (but not limited to): ●

●

●

●

Chemical Reactions: Intensified processes may involve higher temperatures, pressures, or concentrations, which can increase the risk of chemical reactions, including runaway reactions that can lead to explosions or fires. Equipment Design: The use of novel equipment and technologies may require new safety measures to prevent equipment failure, leaks, or other hazards. Human Factors: The introduction of new processes or equipment may require additional training for operators or maintenance staff to ensure safe operation. Hazard Identification and Risk Assessment: Safety analysis must be performed to identify potential hazards and risks associated with the new process or equipment and to assess their likelihood and severity.

5.3.1

Impacts of Process Intensification on Safety

The use of novel and complex equipment in process intensification technology can pose unique safety risks that must be carefully assessed and controlled to ensure safe operations. The design and operation of process intensification equipment often differ significantly from traditional chemical processes and, as a result, require a specialized safety approach. A thorough understanding of the specific hazards

5.3 Process Intensification and Safety Analysis

associated with these new technologies is essential to ensure their safe integration into the overall process (Klais et al., 2010). One of the primary safety concerns associated with process intensification equipment is the potential for unexpected reactions or unintended consequences owing to highly dynamic processes such as rapid heat and mass transfer rates or fast reactions. The use of microreactors, microchannels, and microfluidic systems in process intensification can lead to rapid and intensive mixing of reagents, which may increase the risk of reaction runaway and the release of hazardous chemicals (Lutze et al., 2010). These dynamic processes may result in transient behavior, which can be difficult to predict and control, thereby introducing new risks and challenges. Consequently, safety assessments must consider the potential hazards of the chemicals and processes involved, as well as the possibility of accidental release. Process intensification technology often involves the use of unconventional operating conditions, such as high temperatures, pressures, and concentrations (Van Gerven and Stankiewicz, 2009), which pose unique safety challenges. High-pressure systems are more susceptible to mechanical failure, which can result in equipment damage or hazardous releases. Pressure vessels, piping, and valves associated with process intensification equipment must be designed, installed, and maintained to ensure safe operation and minimize the risk of failure. In addition, process intensification technology often requires the use of nonconventional materials, such as ceramics, composites, and metals with unique properties. These materials may present distinct safety hazards, such as brittleness, corrosion, or toxicity, which must be carefully considered during the safety assessment process. These novel conditions necessitate the development of new materials, equipment designs, and safety measures to ensure adequate containment and control (Lutze et al., 2010). To address these safety concerns, safety assessments must be conducted throughout the entire life cycle of process intensification equipment, from design to decommissioning. Adequate safety measures, such as protective equipment, interlocks, and emergency shutdown systems, must be implemented to minimize the risk of incidents or accidents. Safety assessments must be tailored to the specific requirements of process intensification equipment, considering the potential hazards of the chemicals and processes involved, the use of nonconventional materials, and the unique safety challenges posed by high-pressure and high-temperature systems. The impact of process intensification on process safety is multifaceted, with both positive and negative aspects. By incorporating safety considerations into the design and implementation of process intensification technologies, chemical and processing industries can harness the benefits of process intensification while ensuring the safety of their operations. It is essential to develop robust safety methodologies and adopt ISD principles, such as passive safety measures and fail-safe designs, to minimize potential hazards and ensure the safe operation of intensified processes (Etchells, 2005).

5.3.2

Safety Analysis in Intensified Process Design

The technological advancement of intensified processes offers numerous advantages, as discussed earlier, but these benefits can come at the cost of increased safety

137

138

5 Safety Analysis of Intensified Chemical Processes

concerns owing to the inherent nature of intensified processes. Intensified processes often involve novel technology with limited experience, increasing the likelihood of unpredicted accidents. Thus, it is crucial to incorporate safety analysis during the design stage (Ebrahimi et al., 2012) to identify and mitigate potential hazards. Incorporating safety analysis during the design stage of intensified processes is essential for several reasons: ● ●

●

●

Unique hazards that may not be present in traditional processes. Unconventional operating conditions, such as higher temperatures, pressures, or concentrations, can exacerbate safety concerns. Smaller equipment and compact designs that increase the complexity of equipment in intensified processes may introduce new safety challenges related to handling, maintenance, and potential failure points. Interconnected processes often involve the integration of multiple process steps, which can lead to unexpected interactions and safety concerns that may not be apparent when analyzing individual units.

Many common safety assessment techniques rely in part on empirical data, making them less suitable for evaluating innovative technologies. Conventional safety assessment methods often overlook the safety benefits associated with process intensification (Ebrahimi et al., 2008). For instance, the intrinsic safety index developed by Rahman et al. (2005) and several other methods inadequately consider factors such as smaller size and efficient temperature control that are typical of intensified processes. Nevertheless, it is evident that these factors significantly affect the intrinsic safety of any process. For the aforementioned reasons, safety results derived from existing methods may prove to be deceptive and unreliable. Consequently, there is a need for the development of novel methods to more accurately investigate the safety of innovative and intensified processes. However, existing, widely used safety analyses are often adopted and tailored to address the inherent properties and characteristics of process intensification technologies. Thus, it is important to have a thorough understanding of the equipment and its environment. The extent of the required risk assessments depends on the complexity of the process and the extent and nature of the risks involved. In the development of new processes, it is crucial to integrate hazard and risk analysis at every stage of the process design. This iterative and stage-by-stage approach aids in identifying and mitigating risks promptly, circumventing the high costs of preventive measures that may arise in the latter stages of process design. It is pivotal to adopt comprehensive risk identification practices, utilizing a range of risk analysis tools that are best suited to each specific phase of project development. This ensures a robust and thorough evaluation of potential risks throughout the entire project lifecycle, contributing to the overall safety of the intensified chemical processes. 5.3.2.1 Hazard Identification Techniques for Process Intensification Technologies

The safe design and operation of process intensification technologies require the identification of hazards associated with the novel aspects of these processes.

5.3 Process Intensification and Safety Analysis

Various techniques are available for identifying hazardous situations, which all require a rigorous, thorough, and systematic approach by a team of multidisciplinary experts. To ensure success, it is crucial to first identify and analyze possible scenarios that could lead to accidents of varying degrees of severity. Without a structured identification system, hazards can be overlooked, leading to incomplete risk evaluations and potential losses. The widely used traditional hazard identification techniques, as shown in Section 5.2.1, are still applicable with some fine tuning for evaluating and performing a thorough risk analysis and management. Some common techniques used for intensified processes include the following: ●

●

●

Checklist: This methodology is the simplest and most commonly used method to identify hazards based on a set of questions with answers such as yes or no, and the affirmative or negative responses determine the hazards present. The quality of the checklist depends on the experience of the person who prepares and applies it, but literature provides numerous examples that can be adapted to specific uses. In process intensification technologies, it can be useful for collecting information to employ more complex techniques. Checklists are applicable to management systems and projects at all stages, with appropriate checklists for each stage. The primary function of a checklist is to serve as a comprehensive and systematic tool for conducting a final review, ensuring that no important tasks or details are overlooked. HAZOP: The traditional HAZOP approach is adapted to consider the unique aspects of process intensification technologies, such as novel equipment, materials, or operating conditions. The unique characteristics of process intensification technology need to be identified as new types of equipment uncommon to traditional types of processes that may have different operating principles, modes of failure, and safety features that need to be considered. There is also the possibility that new materials that have not been tested in traditional processes could affect process safety. Consequently, modification of the HAZOP study methodology is required to address these unique characteristics. Therefore, it may be necessary to form a specialized team that includes experts in relevant areas such as materials science, equipment design, and process safety. Uncommon scenarios related to the new use of materials, novel equipment, or their interactions may need to be developed to identify potential hazards and operability problems. It may also be beneficial to extend the HAZOP’s focus on the operational phase of a process to consider the entire life cycle of the process intensification technology. This could involve identifying potential hazards and operability problems that could occur at different stages of the technology’s life cycle, from initial design and manufacture through disposal or recycling. This broader perspective could help to ensure that the technology is safe and efficient not just during operation but throughout its entire life cycle. However, this would likely require significant modifications to the traditional HAZOP methodology and the formation of a specialized team with expertise in life cycle analysis. FMEA: The unique characteristics of process intensification technologies require modifying the traditional FMEA approach to focus on the identification of failure

139

140

5 Safety Analysis of Intensified Chemical Processes

●

modes specific to intensified processes, taking into account the potential interactions between integrated process steps. These may include issues related to system integration, such as the compatibility of different unit operations and the impact of process conditions on each unit operation. It is essential to consider the interactions between these unit operations when assessing risks, as changes in process conditions in one unit operation may impact the performance of another unit operation downstream. It is also crucial to examine the manner in which the incorporation of numerous unit operations might influence the overarching process dynamics. Other potential failure modes may include issues related to material compatibility, such as corrosion or fouling of membranes or catalysts. FTA: FTA is a structured approach for identifying and analyzing system potential failures to evaluate and improve the reliability of complex systems, highlighting the complex interconnections and dependencies within these processes. FTA is the construction of fault trees that represent the logical relationships between the potential causes and consequences of failures. The inherent nature of process intensification technologies, which often involve multiple unit operations, high temperatures and pressures, and the use of new materials and equipment, can increase process complexity and introduce new failure modes that may not be captured by traditional FTA. Adapting the traditional FTA approach to process intensification technology requires several modifications. Identifying potential failure modes requires a thorough understanding of the unique processes and characteristics of process intensification technologies. Traditional FTA relies on historical data and expert knowledge to identify potential failures; however, this may not be sufficient for process intensification technologies due to the lack of historical data and technological novelty. Therefore, it may be necessary to use different and uncommon methods, such as simulation and modeling, to identify potential failure modes. Concomitantly, it is important to consider the dynamic nature of process intensification technology processes, as they are often highly dynamic with rapid changes in the process and operating conditions. Therefore, it is essential to consider the time-dependent behavior of the system when analyzing potential failure modes.

Engaging all stakeholders, including process engineers, materials scientists, equipment designers, and safety experts, in the hazard identification process is vital to ensure a comprehensive understanding of potential failure modes and to collaboratively develop effective mitigation strategies tailored to the unique complexities of process intensification technologies. 5.3.2.2 Risk Assessment for Process Intensification Technologies

Risk assessment for process intensification technologies involves evaluating the likelihood and severity of a wide range of potential hazards specific to intensified processes. Traditional risk assessment methods are typically based on linear analysis of individual hazards, which may not be adequate for process intensification processes. Quantitative and qualitative risk assessments can be tailored to estimate the risk associated with process intensification technologies by considering the probability

5.3 Process Intensification and Safety Analysis

of hazardous events, consequences of these events, and exposure of people and the environment to the unique aspects of these processes.

5.3.3

Inherently Safer Design Principles Intensified Processes

Chemical processes have always been subjected to safety analysis, which requires comprehensive approaches toward safe processes. Over the years, several methods and indices have been developed to perform different types of analyses in areas such as reliability, safety, human factors, and quality control. The American Petroleum Institute (API) has created RP 14C, a safety analysis method based on a number of conventional hazards analysis methodologies, including FMEA and HAZOPS (API 14C, 2017). The premises of RP 14C listed below provide a good platform for designing an inherently safe process intensification systems. ●

● ●

● ●

Process components function in the same manner, regardless of specific facility design. Each process component is analyzed for “worst case” input and output conditions. If fully protected when analyzed alone, the analysis will be valid for that component in any configuration. If every component is protected, the system will be protected. When components are assembled into a system, some devices can be eliminated.

The ISD concept aims to create processes that are intrinsically secure from the outset, thereby minimizing accident risks and enhancing safety through hazard elimination or reduction (Hendershot, 2012a). An ISD evaluation may not always result in a clear choice, as one option may be inherently safer with respect to one or more specific hazards, but it may introduce new hazards or increase the magnitude of other hazards. It is all about which of the alternatives identified and evaluated is the best option (Hendershot, 2012b). The principles of process intensification technology are well-suited to the concept, which process intensification fundamentals build on the premise of eliminating or minimizing the unique hazards associated with the chemical processes involved. The following four primary principles can be adapted to address the specific challenges of intensified processes: ●

Minimization: Process intensification technology often leads to a smaller size of process equipment and less quantity of hazardous materials in the system, thus decreasing the potential for accidents and the severity of their consequences. For instance, one of the most illustrative examples of process intensification technology that incorporates the principle of minimization is microreactor technology. Microreactors are devices with channels or cavities the size of a few hundred micrometers that allow reactions to take place. By reducing the size of the reactor, the surface-to-volume ratio is greatly increased, leading to intensified heat and mass transfer, resulting in improved safety and process efficiency. This improved heat transfer allows reactions that are exothermic to be carried out more safely, as the heat can be removed more efficiently, thereby reducing the risk of thermal runaway reactions. The minute size of the microreactors means

141

142

5 Safety Analysis of Intensified Chemical Processes

●

●

that the quantity of hazardous substances inside the reactor at any given time is minimized, which reduces the potential consequences of release. High-gravity (HiGee) technologies are another example of process intensification technologies that can be used to minimize the amount of hazardous substances in a process. Centrifugal forces are used to intensify the mass transfer processes, enabling the equipment to be significantly downsized. By using HiGee technologies, it is possible to reduce the size of equipment, such as distillation columns, which can reduce the inventory of hazardous substances, consequently reducing the potential for hazards and improving process efficiency. However, this reduced-size equipment also poses some challenges that need to be addressed before it can be widely adopted in the chemical industry. Some of these challenges include fabrication and characterization of microstructures, design and optimization of flow patterns and reaction conditions, development and testing of novel catalysts and materials, prevention and removal of fouling and clogging, detection and monitoring of process variables, safety and reliability assessment, and economic and environmental evaluation. Additionally, while the application of process intensification technologies can greatly aid in minimizing the amount of hazardous substances, it is crucial to maintain a holistic view of the process and to consider potential trade-offs. For instance, while the use of microreactors can reduce the inventory of hazardous substances, it may lead to an increase in the number of potential leakage points owing to the increased number of units. Careful risk analysis is required when implementing such technologies. Therefore, further research and development are needed to overcome these barriers and to demonstrate the feasibility and benefits of these technologies for various chemical processes. Substitution: Substitution refers to the replacement of hazardous materials, processes, or methods with less-hazardous alternatives that can perform the same or better functions. The concept of “safer is better” underlines this principle and serves as a proactive approach in mitigating potential harm. Process intensification offers several opportunities for substitution, such as utilizing alternative solvents or catalysts that are less flammable, explosive, corrosive, or toxic than the conventional solvents and leveraging green chemistry principles. This can lead to a reduced risk of fire, explosion, and toxic exposure; lower the risk of accidents and injuries; and reduce the environmental impact and regulatory burden of chemical processes. The use of green solvents presents a quintessential example of process intensification that embraces this principle. Traditional solvents used in various chemical processes often pose significant environmental and health risks. Philosophically, process intensification encourages the use of green solvents such as supercritical fluids, ionic liquids, and water under specific conditions. Supercritical carbon dioxide (Jaouhari et al., 2020), for instance, is used as an environmentally benign solvent in various processes, such as extraction and chromatography. Deep eutectic solvents (DES), having low vapor pressure, low volatility, nonflammability, biodegradability, and low toxicity, among others, can be used to replace organic

5.3 Process Intensification and Safety Analysis

●

●

●

●

solvents and ionic liquids, thus reducing emissions and fire hazards (Ullah et al., 2023). Moderation: Moderation is a proactive strategy that involves reducing the impact of hazardous events by diluting or attenuating the hazard at the source rather than relying on protective measures such as alarms or relief valves. It aims to reduce the potential for accidents by moderating and controlling the process conditions, such as temperature, pressure, and concentration; the use of diluents or retardants; and the control of release rates. The focus is on limiting the energy potential of a process and moderating the conditions under which it operates. This will further contribute to enhanced safety. This is best illustrated by heat exchanger reactors, which integrate the reaction and heat exchange into a single unit, thereby improving the control of exothermic reactions (Anxionnaz et al., 2008). By efficiently removing the heat of reaction, the process conditions can be moderated and the risk of thermal runaway reactions can be reduced. Microreactor technology also fits well with the moderation principle. The small size and high surface-to-volume ratio allow for fast heat and mass transfer, which enables high conversion rates and selectivity. This means that fewer reactants and intermediates are needed and stored in the system, reducing the risk of fire, explosion, or toxic release in case of a leak or rupture. The high heat transfer coefficient of microreactors enables efficient cooling or heating of the reaction mixture, preventing hot spots or runaway reactions that translate into improved thermal management (Wang et al., 2020). This is particularly important for highly exothermic or endothermic reactions, which can pose significant challenges for conventional reactors (Renken and Kiwi-Minsker, 2010). The short residence time and low hold-up volume of microreactors allow for rapid response to changes in process parameters, such as temperature, pressure, flow rate, or concentration. This enables precise control over the reaction conditions and product quality, avoiding undesired side reactions or by-products that can compromise safety. Simplification: Simplification focuses on reducing the complexity of process operations and equipment, which can minimize potential failure points and human error and thus improve the efficiency and sustainability of chemical processes. By simplifying process operations and equipment, we can reduce the number of steps, components, and interfaces that can cause failures or errors (Smith, 2019). This consolidation of several functions reduces the overall footprint of the process, resulting in lower capital and operating costs as well as the environmental impact of the process, thereby reducing the potential for errors and enhancing overall safety. Process intensification inherently promotes simplification through the integration of multiple unit operations into a single unit. Process intensification also allows for a modular and compact design, which can reduce the footprint, material and energy consumption, and installation time of the process (Kim et al., 2017). The continuous process scheme advocated in process intensification technologies can improve efficiency by reducing the need for batch processing. For instance, a continuous fermentation process can produce more product in a shorter time than

143

144

5 Safety Analysis of Intensified Chemical Processes

a batch fermentation process. In addition, implementation of specific automation mechanisms for process intensification technologies can improve efficiency and consistency by reducing the need for manual intervention. Simplification can enhance the performance, reliability, and safety of chemical processes, as well as facilitate their scale-up and adaptation to different conditions and applications. Process simplification can increase sustainability by reducing the use of resources and energy.

5.4 Safety Management Systems for Intensified Processes SMSs are integral to the safe operation of intensified processes in various industrial settings. These structured and comprehensive frameworks play a crucial role in process intensification by systematically identifying, assessing, and managing the safety risks associated with intensified processes. Structured approaches comprise a comprehensive set of policies, procedures, and practices designed to manage safety risks and mitigate hazards. Intensified processes, which embody the implementation of innovative strategies to enhance process efficiency and effectiveness, are subject to a plethora of safety hazards. The inherent challenges posed by intensified processes, such as increased potential for accidents, runaway reactions, fires, explosions, and toxic releases, necessitate a proactive and systematic approach to safety in order to effectively identify and control the hazards and risks associated with these processes. The complexity and heightened interconnectivity of these processes underscore the necessity for a robust and systematic approach to safety management. Given the multifaceted nature of these risks, it is paramount that safety professionals, process designers, and management recognize the value of SMS and continue to develop and implement these systems to improve safety in industrial processes. Implementing an SMS tailored to intensified processes involves several key components, including hazard identification and assessment, risk management strategies, safety performance monitoring, and continuous improvement initiatives. These components work in unison to ensure that safety aspects are considered throughout the design, development, and operation of the intensified processes. By integrating these components, an SMS can effectively address the unique challenges associated with process intensification, thereby mitigating the potential for accidents and ensuring the overall safety and integrity of these complex systems. An SMS tailored to intensified processes should consist of the following components: ●

Hazard Identification and Assessment: The hazard identification and assessment process is the cornerstone of any SMS, particularly crucial in intensified processes due to the increased complexity and interdependence of process components. This involves identifying the possible sources of hazards in the process design, operation, and maintenance, such as chemical reactivity, thermal instability, pressure build-up, material compatibility, human error, and external factors.

5.4 Safety Management Systems for Intensified Processes

●

●

●

Hazards should be assessed in terms of likelihood and severity using appropriate methods such as HAZOP analysis, FTA, and ETA, as discussed in Section 5.3.2. Risk Management Strategies: This involves selecting and implementing the most effective and feasible measures to prevent or mitigate the hazards and risks identified in the previous step. Risk management in intensified processes often necessitates a multi-faceted approach where the measures should follow the hierarchy of controls, which prioritizes elimination, substitution, engineering controls, administrative controls, personal protective equipment (PPE), and failsafe mechanisms. These measures should also be compatible with the objectives and principles of process intensification, such as minimizing waste and energy consumption, maximizing efficiency and productivity, and enhancing flexibility and reliability. Safety Performance Monitoring: This involves measuring and evaluating the effectiveness of SMS and risk management strategies through regular tracking of safety metrics and key performance indicators (KPIs) to achieve desired safety outcomes. Surveillance of key safety indicators, such as process temperature, pressure, flow rate, accident frequency, incidence of near-miss events or accidents, severity rates, safety audits, inspection results, and compliance with regulations and standards, are essential components of this system. Monitoring should also involve feedback from various stakeholders, such as operators, managers, regulators, and customers. This data should be utilized to adjust safety management strategies, improve hazard identification and assessment processes, and identify opportunities for continuous improvement. Continuous Improvement Initiatives: A SMS is not a static entity but rather a dynamic system that must continuously adapt and improve to meet evolving safety challenges. Continuous improvement initiatives may include regular safety audits, safety culture assessments, and the implementation of lessons learned from safety incidents. These initiatives involve identifying and implementing opportunities to improve SMS and risk management strategies based on the results of safety performance monitoring. The improvement should follow a cyclic process of plan-do-check-act (PDCA), which involves planning improvement actions, executing them, checking their effectiveness, and acting on the results. The improvement should also involve learning from best practices and lessons learned from other organizations or industries that use intensified processes. This continuous improvement should involve all levels of the organization and provide a feedback loop that integrates the results of performance monitoring and other safety initiatives. In intensified processes, continuous improvement can involve the integration of new technologies and methodologies to enhance safety performance.

Safe operation of intensified processes necessitates the implementation of a robust SMS. The complex and interconnected nature of these intensified processes necessitates a comprehensive and systematic approach to safety management by integrating these key components into a unified SMS that can effectively manage the unique safety challenges posed by process intensification, enhance operational efficiency,

145

146

5 Safety Analysis of Intensified Chemical Processes

protect workers, minimize environmental impacts, and maintain public confidence in industry practices.

5.5 Safety Training and Competency for Intensified Processes Safety training and competency development are critical components of managing the risks associated with intensified processes in the CPI, especially when dealing with process intensification technologies, as these technologies often involve novel methods, materials, and equipment that can introduce new risks and hazards in CPIs. Ensuring that personnel have adequate knowledge, skills, and abilities to safely operate and maintain intensified processes can help minimize the risk of accidents related to human error, contribute to the overall process safety performance, and contribute to a strong safety culture. In this section, we discuss the importance of developing and maintaining appropriate safety training and competency programs for personnel working with intensified processes and how effective training and competency management can support safe operation.

5.5.1

Importance of Safety Training and Competency

As process intensification introduces novel technologies, materials, and operational strategies, it is essential that personnel receive appropriate training to effectively manage associated safety challenges (Sitter et al., 2019; Dias and Ierapetritou, 2019). Comprehensive safety training and competency development programs should cover: ●

●

●

Understanding the Underlying Principles of Intensified Processes and Their Potential Hazards: Novel equipment and innovative approaches of process intensification technologies may introduce unfamiliar hazards and potential risks that require specialized knowledge and skills to manage. Safety training and competency development should address the unique characteristics of process intensification technologies and their associated hazards, enabling personnel to identify, assess, and mitigate risks effectively. Adapting to Changing Operational Strategies and Maintenance of Intensified Process Equipment: Ensuring that personnel are competent in managing new strategies that often require changes in operational strategies, such as increased automation or more complex control systems, is crucial for maintaining safe operations and minimizing the potential for human error. Safety training should provide personnel with a comprehensive understanding of the operational strategies associated with process intensification technologies and equip them with the skills necessary to effectively respond to process deviations, equipment malfunctions, and other challenges. Compliance with Regulatory Requirements: As process intensification technologies continue to evolve, regulatory requirements and industry standards may

5.5 Safety Training and Competency for Intensified Processes

●

change to address new safety concerns. Ensuring that personnel have up-to-date knowledge of applicable regulations and best practices is essential for maintaining compliance and avoiding potential penalties and liabilities. Safety training and competency development should include regular updates on relevant regulatory requirements and industry standards, enabling personnel to adapt to changing expectations and to maintain safe operations. These include adaptation of hazard identification, risk assessment, and ISD principles in the context of intensified processes. Emergency response and incident management procedures specific to intensified processes should be part of the curriculum for a comprehensive training and competency development program. Building a Strong Safety Culture: A robust safety culture is fundamental to the success of any organization, particularly in the CPI, where the potential consequences of accidents can be severe. Investing in safety training and competency development for process intensification technologies demonstrates an organization’s commitment to safety, reinforces the importance of continuous learning, and encourages personnel to actively engage in safety improvement efforts. By fostering a strong safety culture, organizations can enhance overall process safety performance, reduce the likelihood of accidents, and promote a safe and productive work environment.

5.5.2

Developing Safety Training and Competency Programs

The effectiveness of safety training and competency programs related to process intensification technologies requires organizations to adopt a systematic and comprehensive approach that addresses the unique characteristics and challenges associated with these innovative technologies. The key aspects to consider when developing safety training and competency programs for process intensification technologies to ensure the effectiveness of the programs are as follows: ●

●

Conducting Training Needs Assessments: Regular training needs assessments need to be conducted to identify the specific knowledge, skills, and abilities related to intensified processes for personnel to work safely with process intensification technologies. This involves evaluating the competencies of individual employees, identifying gaps in their knowledge and skills, and determining the most appropriate training intervention to address these gaps. Training needs assessments should be conducted periodically to account for changes in process intensification technologies, industry best practices, and regulatory requirements. Developing Customized Training Curricula: Effective safety training programs should be tailored to the specific requirements of process intensification technologies and the organization’s unique operational context. Training curricula should cover topics such as the underlying principles of process intensification and those discussed in Sections 5.3.2, 5.5.3, and 5.4, including relief systems, SIS, fire safety systems, and emergency response procedures specific to process intensification technologies. More importantly, training materials should be developed in collaboration with subject-matter experts to ensure that the content is accurate, up-to-date, and relevant to the target audience. Evaluating and

147

148

5 Safety Analysis of Intensified Chemical Processes

updating training materials and competency requirements should be conducted regularly to reflect advancements in process intensification technology, industry best practices, and lessons learned from incidents and near-misses.

5.5.3

Utilizing a Blended Learning Approach

In view of current educational technology advancements, to maximize the effectiveness of safety training and competency development programs, organizations should adopt a blended learning approach that combines various instructional methods and learning formats. This may include classroom-based instruction, hands-on training, computer-based simulations, and on-the-job coaching and mentoring. By providing a diverse range of learning opportunities and flexibilities, organizations can cater to different learning styles and preferences, ensuring that personnel develop a comprehensive understanding of process intensification technologies and their associated safety challenges.

5.5.4

Assessing Training Effectiveness and Continual Improvement

Regular assessments of training effectiveness are crucial to ensuring that safety training and competency programs remain relevant and impactful. This can involve evaluating participant feedback, conducting knowledge and skill assessments, and tracking KPIs related to process safety performance. Based on the outcomes of these assessments, organizations can identify areas for improvement, update training materials and methods, and implement necessary changes to enhance the overall effectiveness of their safety training and competency programs.

5.5.5 Benefits of Effective Safety Training and Competency Management Investing in comprehensive safety training and competency development programs for process intensification technologies can yield significant benefits for organizations operating in the CPI. These benefits extend beyond simply ensuring regulatory compliance and can positively impact the overall safety performance, productivity, and reputation of the organization. Among the benefits of such investments are: ●

●

Enhanced Process Safety Performance: One of the primary benefits of effective safety training and competency management is the potential to improve overall process safety performance. By ensuring that personnel have the requisite knowledge, skills, and abilities to safely operate and maintain intensified processes, we can reduce the likelihood of accidents, incidents, and near misses related to human error in the operation and maintenance of intensified processes. In turn, this can contribute to a reduction in the frequency and severity of process safety events, leading to a safer work environment and potentially lower insurance premiums. Increased Operational Efficiency: Well-trained and competent personnel are better equipped to manage the unique challenges associated with process

5.6 Case Studies of Safety Analysis in Intensified Processes

●

intensification technologies such as advanced control systems, increased automation, and novel materials and equipment. Investing in safety training and competency development can ensure that the organization’s workforce is capable of operating and maintaining intensified processes efficiently, resulting in reduced downtime, fewer equipment malfunctions, optimized resource utilization, and reduced regulatory noncompliance risks. Stronger Safety Culture: Effective safety training and competency management can contribute to the development of a robust safety culture within an organization. Employees who are consistently provided with the resources and opportunities to enhance their safety knowledge and skills are more likely to feel valued and engaged in safety improvement efforts. This, in turn, enhances personnel confidence and fosters a sense of shared responsibility and accountability for process safety, leading to better decision-making, more proactive hazard identification and risk mitigation, and ultimately, improved safety performance in managing the unique safety challenges associated with intensified processes. This would develop a strong safety culture that values continuous learning and improvement.

Effective and comprehensive safety training and competency management are essential components for ensuring the safe operation of intensified processes in the CPI. Developing and maintaining robust training and competency programs would lead to minimization of the risk of accidents related to human error and contribute to overall process safety performance resulting in positive economic valuation to the organization.

5.6 Case Studies of Safety Analysis in Intensified Processes In recent years, process intensification has gained significant attention in the CPI. Since the process intensification concept was popularized by Colin Ramshaw in the 1970s, various novel, multifunctional, and integrated equipment have been developed that bring forth design concepts that further improve CPIs, particularly with smaller footprints and economical processes that are safer and more sustainable. Although process intensification offers numerous benefits, it also poses significant safety challenges that must be addressed to ensure safe operation of intensified processes. In this section, a few examples of process intensification applications in the CPI are presented, and a safety analysis is discussed to ensure the safe operation of intensified processes. Microstructured reactors and heat exchangers are common examples of process intensification technology. An integrated microreactor and heat exchanger provide significant energy savings by recovering heat from the reaction process and using it to preheat reactants or drive other processes. The combined functions of a reactor and a heat exchanger, which can lead to a more streamlined process, increase productivity by reducing the amount of time and equipment required to complete a chemical reaction. This would improve the environmental impact and enhance

149

150

5 Safety Analysis of Intensified Chemical Processes

safety performance. One such example is illustrated by Benaïssa et al. (2008), where a prototype of a “heat-exchanger/reactor” provided by Alfa Laval Vicarb, called the open plate reactor (OPR), was used for the esterification of propionic anhydride by 2-butanol. The authors conducted a safety assessment using risk assessment methods, HAZOP study, and simulations to determine the optimal operating conditions for safe control. Bibliographic sources and calorimetric tools were used to assess the risk of providing information about chemical hazards and synthesis, whereas the HAZOP study was applied to the intensified process to identify potential hazards and provide a number of runaway scenarios. The feasibility of the reaction in the “heat-exchanger/reactor” was assessed via a simulation model, which was then used to estimate the temperature and concentration profiles during synthesis. Finally, the behavior of the process was simulated following probable malfunctions to calculate the adiabatic temperature rise and maximum temperature of the synthesis reaction. They concluded that the thermal inertia of the “heat-exchanger/reactor” allows it to be intrinsically safer. This shows that the proposed safety methodology for the operation of a continuous intensified reactor can be used as a guide for the design of new intensified reactors and to ensure safe operation of the process. The method needs to be validated with experimental studies and can be used as a guide for the design of new intensified reactors. A generalized modular representation framework was proposed by Tian et al. (2018) to discover intensification opportunities from process fundamentals and incorporate flexibility analysis, inherent safety analysis, and controller design. The framework comprises three interactive toolboxes: process intensification/synthesis toolbox, process simulation/optimization toolbox, and process operability/control/ safety toolbox. The proposed framework was applied to a heat exchanger network synthesis problem for thermal intensification, demonstrating its capability to enhance safety and operability at the early design stage. This framework was also applied to hybrid separation/reaction systems to exploit the synergetic potential of reactive distillation. Another example of process intensification is reactive distillation, which combines separation and reaction steps in a single unit. Reactive distillation has been used to produce various chemicals, including esters, ethers, and alcohols. The safety analysis of reactive distillation involves identifying potential hazards, such as runaway reactions, thermal decomposition, and overpressure. Qi et al. (2019) investigated the concept of ISD through process intensification and proposed a novel safety assessment methodology for evaluating the safety of process intensification-based ISD options. The proposed methodology was applied to a case study of the C3-alkyne hydrogenation distillation process. Four intensification alternatives were analyzed: reactive distillation, thermally coupled distillation of a side-rectifying section, thermally coupled distillation of a side-stripping section, and fully thermally coupled distillation. The authors highlighted that applying process intensification strategies in the design stage may result in adverse safety effects, and eliminating units may cause risk transfer and risk accumulation, which could lead to serious consequences when accidents occur. The study showed that process

References

intensification-based ISD options can significantly reduce the inherent hazards associated with the process. Therefore, the proposed methodology can help identify and mitigate these risks and assist in achieving ISD through process intensification.

5.7 Conclusions This chapter highlights the critical role of safety considerations in the design, operation, and management of process intensification technologies in the CPI. As process intensification technologies continue to gain traction due to their potential for enhanced efficiency, reduced environmental impact, and improved economic performance, it is essential that safety concerns are adequately addressed to ensure the long-term viability and acceptance of these innovative approaches. Various safety challenges associated with process intensification technologies have been discussed, including increased process complexity, novel materials and equipment, and the potential for unforeseen hazards. The importance of integrating safety analysis during the design stage of intensified processes was emphasized, utilizing methods such as hazard identification techniques, risk assessment, and ISD principles. In addition, the role of SMSs in maintaining the safe operation of intensified processes was explored, highlighting the need for effective safety training and competency management as well as continuous monitoring and improvement efforts. Furthermore, the potential impacts of process intensification technologies on process safety were examined, considering both the positive effects, such as reduced inventories of hazardous materials and lower energy consumption, and potential negative effects, such as increased process complexity and new safety concerns associated with novel technologies. The application of safety analysis in real-world process intensification technology implementations was illustrated through several case studies, demonstrating how the safe operation of intensified processes can be contributed by hazard identification, risk assessment, and ISD principles. In summary, the successful integration of safety considerations into the design, operation, and management of process intensification technologies is crucial for sustainable development of the CPI. By adopting a proactive and systematic approach to safety, organizations can harness the full potential of process intensification technologies while minimizing the risk of accidents and incidents. This, in turn, can contribute to improved safety performance, increased operational efficiency, and enhanced organizational reputation, positioning the industry for continued growth and innovation in the future.

References American Petroleum Institute (2017). API RP 14C: Analysis, Design, Installation, and Testing of Safety Systems for Offshore Production Facilities. 8th Edition (February 2017), Washington, DC.

151

152

5 Safety Analysis of Intensified Chemical Processes

Anxionnaz, Z., Cabassud, M., Gourdon, C., and Tochon, P. (2008). Heat exchanger/ reactors (HEX reactors): concepts, technologies: state-of-the-art. Chemical Engineering and Processing: Process Intensification 47 (12): 2029–2050. Aven, T. (2016). Risk assessment and risk management: review of recent advances on their foundation. European Journal of Operational Research 253 (1): 1–13. Bansal, S. (2019). Differentiating quantitative risk and qualitative risk analysis. iZenBridge. https://www.izenbridge.com/blog/differentiating-quantitative-riskanalysis-and-qualitative-risk-analysis/ (accessed 17 January 2023). Benaïssa, W., Gabas, N., Cabassud, M. et al. (2008). Evaluation of an intensified continuous heat-exchanger reactor for inherently safer characteristics. Journal of Loss Prevention in the Process Industries 21 (5): 528–536. Bridges, W., Creighton, J., and Tugwell, J. (2011). Process safety management standard: inspection criteria and enforcement policy. Professional Safety 56 (11): 36–43. Carter, D. and Hirst, I. (2000). ‘Worst case’ methodology for the initial assessment of societal risk from proposed major accident installations. Journal of Hazardous Materials 71 (1-3): 117–128. Center for Chemical Process Safety (2008). Guidelines for Hazard Evaluation Procedures. New York, USA: AIChE. Crowl, D.A. and Louvar, J.F. (2019). Chemical Process Safety: Fundamentals with Applications, 4e. Pearson. Curcio, S. (2013). Process intensification in the chemical industry: a review. In: Sustainable Development in Chemical Engineering: Innovative Technologies (ed. A. Basile, V. Piemonte, and M. de Falco), 95–118. Wiley. Reay, D., Ramshaw, C., and Harvey, A. (ed.) (2008). Chapter 2 – Process intensification – an Overview. In: Isotopes in Organic Chemistry, Process Intensification, 21–45. Butterworth-Heinemann. Dias, L.S. and Ierapetritou, M.G. (2019, 2019). Optimal operation and control of intensified processes – challenges and opportunities. Current Opinion in Chemical Engineering 25: 82–86. Dunjó, J., Fthenakis, V., Vílchez, J.A., and Arnaldos, J. (2010). Hazard and operability (HAZOP) analysis. A literature review. Journal of Hazardous Materials 173 (1–3): 19–32. https://doi.org/10.1016/j.jhazmat.2009.08.076. Ebrahimi, F., Kolehmainen, E., and Turunen, I. (2008). Safety advantages of on-site microprocesses. Organic Process Research and Development 13: 965–969. Ebrahimi, F., Virkki-Hatakka, T., and Turunen, I. (2012). Safety analysis of intensified processes. Chemical Engineering and Processing: Process Intensification 52: 28–33. Ericson, C.A. (2016). Hazard Analysis Techniques for System Safety, 2e. Wiley. Etchells, J.C. (2005). Process intensification: safety pros and cons. Process Safety and Environmental Protection 83 (2): 85–89. Gupta, J.P. and Edwards, D.W. (2002). Inherently safer design—present and future. Process Safety and Environmental Protection 80 (3): 115–125. Hendershot, D.C. (2012a). Inherently safer design: the fundamentals. Chemical Engineering Progress 108 (1): 40–42. Hendershot, D.C. (2012b). New focus on inherently safer design. Journal of Chemical Health and Safety 19 (6): 43.

References

Jaouhari, T., Zhang, F., Tassaing, T. et al. (2020). Process intensification for the synthesis of ultra-small organic nanoparticles with supercritical CO2 in a microfluidic system. Chemical Engineering Journal 397: 125333. Keil, F. (2018). Process intensification. Reviews in Chemical Engineering 34 (2): 135–200. Khan, F.I. and Abbasi, S.A. (1998). Techniques and methodologies for risk analysis in chemical process industries. Journal of Loss Prevention in the Process Industries 11 (4): 261–277. Khan, F.I. and Amyotte, P.R. (2003). How to make inherent safety practice a reality. Canadian Journal of Chemical Engineering 81: 2–16. Kim, Y.H., Park, L.K., Yiacoumi, S., and Tsouris, C. (2017). Modular chemical process intensification: a review. Annual Review of Chemical and Biomolecular Engineering 8: 359–380. Klais, O., Westphal, F., Benaissa, W. et al. (2010). Guidance on safety/health for process intensification including MS design. Part III: Risk analysis. Chemical Engineering & Technology 33 (3): 444–454. Kletz, T.A. (1978). What you don’t have can’t leak. Chemistry and Industry, (May 6), 287–292. Kletz, T.A. (1980). Benefits and risks: their assessment in relation to human needs. Endeavour 4 (2): 46–51. Kletz, T.A. (1996). Inherently safer design: the growth of an idea. Process Safety Progress 15 (1): 5–8. Kletz, T.A. (1999). What went wrong. Case Histories of Process Plant Disasters. Houston: Gulf Publishing. Lees, F.P. (2012). Lees’ Loss Prevention in the Process Industries: Hazard Identification, Assessment, and Control. Butterworth-Heinemann. Lutze, P., Gani, R., and Woodley, J.M. (2010). Process intensification: a perspective on process synthesis. Chemical Engineering and Processing: Process Intensification 49 (6): 547–558. Luyben, W.L. and Hendershot, D.C. (2004). Dynamic disadvantages of intensification in inherently safer process design. Industrial & Engineering Chemistry Research 43 (2): 384–396. Mannan, M.S., West, H.H., Krishna, K. et al. (2005). The legacy of Bhopal: the impact over the last 20 years and future direction. Journal of Loss Prevention in the Process Industries 18 (4–6): 218–224. Mannan, M.S. (ed.) (2012). Lees’ Loss Prevention in the Process Industries Hazard Identification, Assessment and Control, 4e. Butterworth-Heinemann. Mannan, M.S., Reyes-Valdes, O., Jain, P. et al. (2016). The evolution of process safety: current status and future direction. Annual Review of Chemical and Biomolecular Engineering 7: 135–162. Manuele, F.A. (2008). Advanced Safety Management Focusing on Z10 and Serious Injury Prevention. Wiley. Tweeddale, M. (ed.) (2003). Managing Risk and Reliability of Process Plants. Gulf Professional Publishing.

153

154

5 Safety Analysis of Intensified Chemical Processes

Pasman, H.J. (2015). A system perspective for assessing and avoiding low-probability, high consequence events. In: Risk Analysis and Control for Industrial Processes – Gas, Oil and Chemicals. Butterworth-Heinemann. Qi, M., Zhu, J., Xu, J. et al. (2019). Investigation of inherently safer design through process intensification: novel safety assessment methodology and case study in C3 -alkyne hydrogenation distillation process. Industrial & Engineering Chemistry Research 58 (12): 4866–4880. Rahman, M., Heikkilä, A.M., and Hurme, M. (2005). Comparison of inherent safety indices in process concept evaluation. Journal of Loss Prevention in the Process Industries 18: 327–334. Ramshaw, C. (1999). Process intensification and green chemistry. Green Chemistry 1 (1): G15–G17. Renken, A. and Kiwi-Minsker, L. (2010). Microstructured catalytic reactors. Advances in Catalysis 53: 47–122. Sitter, S., Chen, Q., and Grossmann, I.E. (2019). An overview of process intensification methods. Current Opinion in Chemical Engineering 25: 87–94. Smith, S.F. (2019). Process intensification: getting more from less. https://assets .thermofisher.com/TFS-Assets/BPD/Reference-Materials/process-intensificationgetting-more-less-article.pdf (accessed 17 January 2023). Stankiewicz, A. and Moulijn, J.A. (2000). Process intensification: transforming chemical engineering. Chemical Engineering Progress 96 (1): 22–34. Tian, Y., Sam Mannan, M., and Pistikopoulos, E.N. (2018). Towards a systematic framework for the synthesis of operable process intensification systems. Computer Aided Chemical Engineering 44: 2383–2388. Ullah, A., Zhang, Y., Liu, C. et al. (2023). Process intensification strategies for green solvent mediated biomass pretreatment. Bioresource Technology 369: 128394. Van Gerven, T. and Stankiewicz, A.I. (2009). Structure, energy, synergy, time – the fundamentals of process intensification. Industrial Engineering & Chemical Research 48: 2465–2474. Volkan, E. (2021). Risk assessment and analysis methods: qualitative and quantitative. ISACA Journal 2: 1–6. Wang, J., Ni, L., Cui, J. et al. (2020). Thermal hazard analysis of styrene polymerization in microreactor of varying diameter. Processes 8 (12): 1650. Yuan, S., Reniers, G., Yang, M., and Bai, Y. (2023). Cost-effective maintenance of safety and security barriers in the chemical process industries via genetic algorithm. Process Safety and Environmental Protection 170: 356–371.

155

Part III Control and Safety Analysis of Intensified Chemical Processes

157

6 Control of Hybrid Reactive–Extractive Distillation Systems for Ternary Azeotropic Mixtures Zong Yang Kong 1 and Hao-Yeh Lee 2 1 Department of Engineering, School of Engineering and Technology, Sunway University, Bandar Sunway 47500, Selangor, Malaysia 2 Department of Chemical Engineering, National Taiwan University of Science and Technology, Taipei City 106335, Taiwan

6.1 Introduction The application of hybrid reactive-extractive distillation (RED) for ternary azeotropic separation has become increasingly popular, as indicated by the growing number of studies in the last two years (i.e. 2020–2022). The concept of RED is analogous to that of reactive distillation (RD), which combines reaction and separation processes in a single unit operation. However, a key difference is that RED uses a solvent to enhance the separation of azeotropic mixtures by altering the relative volatility of the components, which is not typically used in traditional RD systems. While both RED and RD involve reaction and separation, RED is specifically designed for azeotropic separation, which requires the use of extractive distillation. The solvent used in the RED can either be injected externally or obtained within the RED system, depending on the type of configuration, which makes the process flexible. In comparison to the traditional RD or the conventional extractive distillation (CED) that are widely used for azeotropic separation, the RED has the advantage of using less energy because it relies on the heat produced by the reaction for the subsequent separation process. To our knowledge, the RED was first introduced by Shen and coworkers in 2020 for recovering tetrahydrofuran (THF) and ethanol from waste effluent (Su et al., 2020). Note however that their RED system was based on a triple column configuration where the RD and ED take place consecutively in different columns (Figure 6.1a). In their proposed flowsheet, the first column is a reactive distillation column (RDC) where the hydration reaction occurred between ethylene oxide (EO) and water to form ethylene glycol (EG), according to Eq. (6.1), which comes out from the bottom of the RDC. The xi in Eq. (6.1) refers to the mole fraction of the individual component. The balanced mixture in the distillate commonly contains azeotrope, and it is sent to the 2nd column for the purpose of extractive distillation, together with an externally injected solvent. Here, it is worth re-emphasizing that the solvent can Control and Safety Analysis of Intensified Chemical Processes, First Edition. Edited by Dipesh Shikchand Patle and Gade Pandu Rangaiah. © 2024 WILEY-VCH GmbH. Published 2024 by WILEY-VCH GmbH.

158

6 Control of Hybrid Reactive–Extractive Distillation Systems for Ternary Azeotropic Mixtures Solvent (DMSO/EG) Cooler

RD

SRC

EC

D3 (EtOH)

D2 (THF)

Feed (THF/EtOH/water) D1 (THF + EtOH) Reactant (EO)

B2 (EtOH + DMSO/EG) B3 (DMSO/EG) Purge

B1 (EG)

(a)

Solvent (EG) Cooler

REDC

SRC

D1 (EA)

D2 (EtOH)

Feed (EA/EtOH/water) Reactant (EO)

B1 (EtOH + EG)

(b)

Figure 6.1

B2 (EG)

Purge

The process configuration for (a) Triple column RED and (b) double column RED.

either be obtained within the RED system (e.g. EG produced from the bottom of RDC) or it can be an entirely new solvent (e.g. dimethyl sulfoxide [DMSO]) that comes from external injection. In the 2nd column (i.e. extractive distillation column [EDC]), high purity of THF is obtained from the distillate, while the balance mixture containing ethanol and solvent (i.e. either EG or DMSO) is channeled to the solvent recovery column (SRC) for subsequent separation and solvent recovery. The high-purity ethanol is obtained from the distillate of the SRC, while the recovered solvent obtained from the bottom of the SRC is cooled prior to returning to the EDC. Altogether, their triple column RED provides a lower total annual cost (TAC) and CO2 emission by about 63% and 84% with respect to the conventional pressure-swing distillation (i.e. their base case). Instead of using a triple column RED configuration, another possible alternative is to combine the RD and ED as a single column, which is the main interest of this work (Figure 6.1b). Such a hybrid configuration requires only two distillation columns in comparison to the triple column

6.1 Introduction

RED (Figure 6.1a). Such a configuration first appeared in the work of Wang et al. (2021), who studied the application of RED for the recovery of ethyl acetate (EA) and ethanol from waste effluent. In such a configuration (Figure 6.1b), the 1st column is known as the reactive-extractive distillation column (REDC) where the EG produced from the RD is directly used as solvent for the subsequent azeotropic separation within the same column. The 2nd column is an SRC, analogous to those in the triple column RED (Figure 6.1a). In the double column RED studied by Wang et al. (2021), high purity of EA and ethanol are obtained from the distillate of the REDC and SRC columns, respectively. Such a configuration was reported to provide 35% and 20% lower TAC and CO2 emissions in comparison to the triple column RED, which has 56% and 100% lower TAC and CO2 emissions than the CED, respectively. EO + Water → EG

(6.1)

( ) ) −9547 mol 9 e = 3.15 × 10 xwater xEO T s cm3 To our knowledge, there are about 14 studies that have worked on the RED system for the past two years (i.e. 2020–2022). Some of these studies focused on the triple column RED (Su et al., 2020; Wang et al., 2021), while the majority of the studies worked on the double column RED (Yan et al., 2022; Zhang et al., 2021a) for azeotropic separation. There are also studies that applied various process intensification techniques to the RED, such as dividing walls (Liu et al., 2022; Yang et al., 2022) and side-stream (Yang et al., 2023), to enhance the separation process performance. Two recent review papers are available that critically summarize the recent studies on RED, in chronological order, which can be referred to by interested readers (Kong et al., 2022a,b). While there are already more than 10 studies that have worked on the steady-state design of the RED and its corresponding intensified processes, only two studies have analyzed the control performance of the RED (Liu et al., 2023; Wu and Chien, 2022). Developing an effective control scheme for the RED is a challenging task. This is because an excessive amount of EO entering the RED system will result in unreacted EO, and when released into the atmosphere, the flammable and hazardous substance may lead to undesirable catastrophic incidents. On the other hand, a deficit of EO entering the process will lead to water impurity in the product streams. An online composition monitoring system on the fresh feed stream can be implemented to overcome the aforementioned challenge but such a system often comes with a very high cost and is difficult to maintain. Thus, it is necessary to find alternative control schemes that can withstand the variation in feed composition without the need for expensive online composition analyzers. In addition, the simultaneous separation and reaction of the various components in a single column has resulted in the product composition being very sensitive to the variations in operating variables. These challenges suggest the need for proper planning to develop an effective control strategy to tighten the purity of the products in the RED. In this chapter, we presented a systematic step (i.e. evolutionary approach) for setting up an effective control scheme for a RED system. To date, there are various book chapters that outline the detailed methodology for setting up different control (

r

159

160

6 Control of Hybrid Reactive–Extractive Distillation Systems for Ternary Azeotropic Mixtures

schemes for RD and ED. To our knowledge, there is no book chapter that provides systematic steps for setting up the control strategy for a RED system. This is likely because the RED is a very new system that has appeared only in the last two years. This highlights the contribution of this chapter. Here, the recovery of isopropanol (IPA) and diisopropyl ether (DIPE) from waste effluent using RED was used as a case study, reproduced from our previous work (Kong et al., 2022c).

6.2 Steady-state Design of the RED Figure 6.2 shows the process flowsheet for the recovery of IPA and DIPE from waste effluent using RED, reproduced from Kong et al. (2022c). Here, the fresh feed stream containing 10 000 kg/h of DIPE, IPA, and water at a ratio of 0.3/0.1/0.6 is fed to a pre-concentration column (PC) so that the excessive water content is removed prior to the RED. This prevents the composition of the mixture from exceeding its azeotropic composition. A portion of water is discharged through the bottom of the PC, while the remaining mixture of about 4250 kg/h (i.e. 59.83 kmol/h) that contains azeotropes is diverted to the RED through the distillate of the PC, together with the EO (reactant). In the REDC, the water reacts with EO to form EG, according to Eq. (6.1), and the EG is consequently employed as a solvent for the separation between DIPE and IPA in the same column. The high purity of DIPE is obtained via the distillate in the REDC, while the balance mixture containing IPA and EG is sent to the recovery column (i.e. SRC), where IPA of high purity is obtained via the EG 313.15 K 4095 kg/h 0.9998 EG 0.0002 Water

Purge 856.52 kg/h 0.9998 EG 0.0002 Water

QC = –0.524 MW

Cooler D1 335.36 K; 1 atm 4249.29 kg/h QC1 = –1.036 MW 0.7060 DIPE 0.2353 IPA 0.0587 Water Trace EG Trace EO PC

NF1 = 11 F 298 K 10000 kg/h 0.3 DIPE 0.1 IPA 0.6 Water

NFS = 2 NF2 = 4

RR1 = 0.739 NT1 = 17 D1 = 0.789 m

376.7 K QR1 = 1.503 MW

341.58 K; 1 atm QC2 = –0.242 MW

EO 609.6 kg/h 1 EO

B1 5750.71 kg/h 1 Water

NFR = 13

REDC

D2 2999.7 kg/h 0.9997 DIPE Trace IPA Trace Water 0.0002 EG Trace EO

RR2 = 0.201 NT2 = 31 D2 = 0.82 m

393.69 K QR2 = 0.087 MW

B2 5954.36 kg/h 0.0002 DIPE 0.1679 IPA 0.0002 Water 0.8314 EG 0.0002 EO

NF3 = 13

354.80 K; 1 atm QC3 = –0.227 MW

SRC

D3 1002.85 kg/h 0.0011 DIPE 0.9971 IPA 0.0004 Water Trace EG 0.0013 EO

RR3 = 0.2077 NT3 = 24 D3 = 0.435 m

475.2 K QR3 = 0.555 MW

B3 4951.52 kg/h 0.9998 EG 0.0002 Water

Figure 6.2 The process configuration for the recovery of IPA and DIPE from waste effluent using RED.

6.3 Dynamic Simulation Setup

distillate. The recovered EG is cooled and recycled back to the REDC. The purity of DIPE and IPA were kept identical to those of Kong et al. (2022c) at 99.97 and 99.7 wt%, respectively. In this chapter, only the RED section (i.e. second column and third column in Figure 6.2) is discussed (highlighted in blue in Figure 6.2), since the goal of this chapter is to work on the control study (i.e. dynamic simulation) of RED.

6.3 Dynamic Simulation Setup The steady-state flowsheet of the RED in Figure 6.2 can be converted from Aspen Plus to Aspen Plus Dynamics for the purpose of dynamic simulation. This can be done either via pressure-driven or flow-driven. Flow-driven simulations in Aspen Plus Dynamics are designed to capture the dynamic behavior of flowrate and their influence on process variables, making them particularly well-suited for systems that experience variations in flowrate, such as pipeline networks. Conversely, pressure-driven simulations focus on the effects of pressure differentials on process variables, making them highly applicable to processes where pressure changes play a significant role, such as distillation columns. In this chapter, the pressure-driven method is employed because it is more reflective of reality than flow-driven. Pumps, compressors, and control valves are crucial components in a dynamic system that need to be included in the simulation to ensure accuracy and reliability. Pressure drops through distillation columns also need to be considered to capture the true behavior of the process. By using a pressure-driven simulation approach, Aspen Plus Dynamics can provide a more realistic representation of the dynamic behavior of a process (Luyben, 2013). Prior to the conversion, both distillation towers need to be sized reasonably. The diameter of both towers and their pressure drop are calculated automatically using the column internal feature made available in Aspen Plus, with a tray spacing and weir height of 0.6096 and 0.1016 m, respectively. The size (i.e. volume) of the column bottom and reflux drum were determined based on a holdup time of 10 minutes when the liquid was at 50% level. After both columns have been sized accordingly, another crucial step is to ensure that there is a realistic pressure drop across all the streams so that the component can flow from higher pressure to lower pressure, which is generally the case in actual practice. Incorporating sufficient pressure drop across a control valve is another important step for dynamic controllability. The detailed guide to these fundamental setups is made available in Luyben’s book (Luyben, 2013). Note that during the conversion from Aspen Plus to Aspen Plus Dynamics, there may be a small drop in the purity of products due to the pressure drop in the column for the sake of dynamic simulation, which was not explicitly accounted for during the steady-state design. We consider such a marginal drop to be reasonable since the purpose of this chapter is to explore the effectiveness of the developed control schemes in returning the purity of products purities back to their nominal values during external disturbances.

161

162

6 Control of Hybrid Reactive–Extractive Distillation Systems for Ternary Azeotropic Mixtures

6.4 Inventory Control Setup After converting the flowsheet from Aspen Plus to Aspen Plus Dynamics, the subsequent step is to implement inventory control that ensures material balance during external disturbances. It is worth noting that Aspen Plus offers a “pressure checker” button that allows us to confirm if the flowsheet is pressure-driven. If the flowsheet is indeed pressure-driven, the simulation file is automatically converted to the Aspen Plus Dynamics environment, which is visible on the computer. In the event of an unsuccessful conversion, a pop-up window will display an error message indicating the specific issues. These issues often involve inadequate pressure drop within the distillation column or an inconsistency between the pressure of the feed inlet stream and the pressure of the inlet stage in the column. Figure 6.3 shows the inventory control loops in the RED. Analogous to the typical inventory control structure found in RD or ED system, the level of the reflux drum in both columns is controlled using the distillate flowrate. The level of the reboiler sump in both columns, on the other hand, is controlled using the bottom flowrate. The pressure of the condenser in both columns is maintained by using its corresponding duty. The fresh feed and solvent streams are equipped with flow controllers together with a ratio to maintain the solvent to feed ratio, analogous to the typical ED or RD system. In addition to the flow ratio between the feed and solvent streams, one additional flow ratio is required in the RED system, i.e. the ratio between the fresh feed and the reactant EO (see the blue line in Figure 6.3). As for the controller setting, a proportional (P) controller with K c = 2 and 𝜏 i = 9999 min was used for the level of both reflux drums and reboiler sumps. A proportional and integral (PI) controller with K c = 20 and 𝜏 i = 12 min was used for the pressure of the condenser. The settings of the flow controllers are set at K c = 0.5 and 𝜏 i = 0.3 min. These controller settings are consistent with those recommended by Luyben (2013).

FC 3

Purge

PC 21

PC 11

R1 LC 21

LC 11

Solvent

2

SRC

REDC

4

Feed FC 1

13 13

Reactant LC 22

FC 2 R2

Figure 6.3

31

LC 12

The inventory control for RED.

24

6.5 Sensitivity Analysis

6.5 Sensitivity Analysis Once the inventory control has been established, it is crucial to install the quality control loop next to ensure that the product qualities (i.e. purities) can be held closely near their nominal values during operational variations. Generally, two different types of controllers are possible, i.e. the composition controller or the temperature controller (TC). In this chapter, we focused mainly on the TC because the cost and maintenance associated with the composition controller are usually expensive, which makes them less favorable in the industry. Nonetheless, there is a brief discussion that covers the temporary usage of pseudo-composition controllers for the purpose of closed-loop sensitivity analysis in Section 6.6.4 (i.e. used for identifying new setpoints for the TC). Prior to setting up the TC, it is crucial to first find out the location of the temperature-sensitive tray in both columns, and this can be done via various techniques such as using the gradient of the temperature profile (i.e. slope criterion) (Li et al., 2013), sensitivity analysis (Arifin and Chien, 2008), or using singular value decomposition (SVD) analysis (Zhang et al., 2021b). Each of these techniques possessed its own advantages and drawbacks (i.e. characteristics). In this chapter, we started the sensitive tray identification using the open-loop sensitivity analysis, while the SVD analysis is discussed in Section 6.6.3 when it is being used as an alternative method (i.e. complement method) to identify an alternative temperature-sensitive tray. The slope criterion method is relatively straightforward and, hence, is not demonstrated in this chapter. Interested readers can refer to Luyben’s book (Luyben, 2013) for more information about the slope criterion method. The open-loop sensitivity analysis was conducted by first installing the TCs, whose control variable (CV) is the tray(s) temperature, while the manipulating variable (MV) is either the reboiler duty (Qr), reflux ratio (RR), or EO to feed ratio that are not being used for inventory control. Here, one can choose to preliminary determine the temperature-sensitive tray by looking at the column temperature 120

Figure 6.4 Tray temperature profile for REDC.

110

T for REDC (°C)

100 90 80 70 60 50 40 30

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Trays

163

6 Control of Hybrid Reactive–Extractive Distillation Systems for Ternary Azeotropic Mixtures 0.16

0.02

0.12 0.01

ΔT (K)

ΔT (K)

0.08 0.04

0.00

0.00 –0.01 –0.04 –0.08

+0.01% EO –0.01% EO

+0.01% Qr –0.01% Qr 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Trays (a)

–0.02

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Trays (b)

0.16

0.015

0.12 0.010 0.08 ΔT (K)

ΔT (K)

0.005

0.000

0.04 0.00

–0.04 –0.005 +0.01% RR –0.01% RR –0.010

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Trays (c)

–0.08 –0.12

+0.01% Qr –0.01% Qr 0

2

(d)

4

6

8 10 12 14 16 18 20 22 24 26 Trays

0.010

0.005

ΔT (K)

164

0.000

–0.005 +0.01% RR –0.01% RR –0.010

0

2

(e)

4

6

8 10 12 14 16 18 20 22 24 26 Trays

Figure 6.5 Sensitivity analysis for REDC using (a) Qr, (b) EO to feed ratio, (c) RR, and SRC using (d) Qr and (e) RR.

profile. The tray that displays the highest temperature gradient can be selected as the preliminary temperature-sensitive tray. Figure 6.4 elucidates the preliminary temperature-sensitive tray selection for the REDC column. Once the TCs have been installed, one can change the TC from “automatic control” to “manual control” so that it is possible to input a small change (e.g. ±0.01%) to the free MVs (e.g. Qr, RR, and EO to feed ratio). The temperature after such a marginal change is recorded, and the difference between the new temperature and

6.6 Quality Control Structures

steady-state temperature is calculated as ΔT. Once the ΔT for both the positive and negative small change has been determined, one can plot the ΔT vs. tray graph, and the tray(s) that displayed the largest ΔT is chosen as the temperaturesensitive tray(s). Figure 6.5 shows the results of the sensitivity analysis for both REDC and SRC. For the REDC, three different MVs are available, i.e. the Qr, RR, and EO to feed ratio, while there are two MVs available for the SRC, i.e. the Qr and RR. From Figure 6.5, it appears that the REDC contains three different temperature-sensitive zones, i.e. zone 1 (2nd tray to 3rd tray), zone 2 (11th tray to 19th tray), and zone 3 (26th tray to 30th tray). For the SRC, two different zones were observed, i.e. zone 1 (11th tray to 13th tray) and zone 2 (21st tray to 25th tray). Once the temperature-sensitive tray(s) are determined, a close-loop relay feedback test is used to determine the PI tuning constants using the IMC-PI tuning rule. A detailed guide on how to conduct the closed-loop relay feedback test is extensively outlined in Luyben’s book (Luyben, 2013). Finally, it is necessary to also install a TC for controlling the solvent temperature using the cooler duty.

6.6 Quality Control Structures 6.6.1

Control Structure 1 (CS 1) – Simple Temperature Control

Once the temperature-sensitive trays have been determined, we develop next the first control structure. To start with, we would like to use only the simple TCs that are commonly employed in the ED and RD systems to control the product qualities in the RED system. Here, only the Qr and RR in both columns are used as the MVs, while the EO to feed ratio is kept at a fixed value. The purpose of this control structure is to explore the performance of the TC when the EO to feed ratio is not used as a MV. In control structure 1 (CS 1), we selected stage 3 and stage 29 in the REDC as the CV, which are manipulated using RR and Qr in the same column, respectively. This is since the RR is nearer to the top of the column (i.e. near stage 3) while the Qr is closer to the bottom of the column (i.e. near stage 29). For the same reason, stages 12 and 23 of the SRC were selected as the CV, controlled by using RR and Qr, respectively. Figure 6.6 shows the flowsheet of CS 1, with the tuning parameters of the TC given in Table 6.1. The control performance of CS 1 is tested using a ±10% throughput change and a ±5% feed composition change, the results of which are depicted in Figure 6.7. The disturbances are injected at t = 1.5 hours. From Figure 6.7a,b, it was observed that CS 1 enables both products’ purities to return back to their nominal point in less than two hours during the +10% throughput disturbance. During the −10% throughput disturbance, there is a marginal deviation in the DIPE purity, while the IPA purity returns to its nominal value in less than two hours. Note, however, that the new steady-state value of the DIPE purity is higher than the product specification, and thus, we consider such a deviation to be reasonable. Upon inspecting the temperature-sensitive tray (Figure 6.7c,e,g,i), it was observed that all the temperatures returned to their nominal values quickly, except for the temperature of 3rd tray

165

166

6 Control of Hybrid Reactive–Extractive Distillation Systems for Ternary Azeotropic Mixtures

TC 5

FC 3

Purge

PC 11

PC 21

R1 LC 11

Solvent Feed

2

3

4 FC 1

12

Table 6.1

TC 4

13 29

TC 2

FC 2

Figure 6.6

13

TC 1

Reactant

R2

LC 21 SRC

REDC

23

TC 3 LC 22

LC 12

31

24

Process flowsheet for CS 1. The tuning parameters for all TCs for CS 1. 𝝉i

Controller

Kc

TC 1

3.498 92

5.28

TC 2

1.821 488

11.88

TC 3

1.958 724

6.6

TC 4

4.089 741

11.88

TC 5

7.803 71

13.2

during the −10% throughput disturbance. This was due to the saturation of the reflux flowrate reaching 0 kg/h (Figure 6.7d), which prevented the temperature of 3rd tray from returning to its nominal value. Such observations aligned with those reported by Wu and Chien (2022), who worked on the control of RED. To overcome this issue, some other alternative control structure needs to be explored. In terms of composition disturbances, it was observed from Figure 6.7b that the IPA purity dropped significantly during the −5% composition disturbance. During composition disturbances, the temperature of all temperature-sensitive trays returns to their nominal values swiftly, suggesting that the temperature of the columns is well maintained. This means that there is no direct relation between the significant drop in the purity of products during the feed composition disturbance and the column temperature profile. Therefore, we attributed the significant deviation in purity of products to the fact that CS 1 used a fixed EO to feed flowrate, and any variation in the feed composition disturbance that cannot be detected instantaneously will result in incorrect amount of EO entering the system because the EO to feed ratio is not adjusted automatically.

6.6 Quality Control Structures 1.00 Purity of IPA (mol%)

Purity of DIPE (mol%)

0.9998 0.9996 0.9994

+ 10% throughput – 10% throughput + 5% feed comp – 5% feed comp

0.9992 0.9990 0

2

4

(a)

6 Time (h)

8

46 44 42 4

8

10

T 29 in REDC (°C)

108 106 104 102 100 0

2

4

(e)

6 Time (h)

8

10

800 600 400 200 0 –200 0

10

2

4

6 Time (h)

8

10

2

4

6 Time (h)

8

10

2

4

6 Time (h)

8

10

2

4

6 Time (h)

8

10

0.68 0.66 0.64 0.62 0.60 0.58

Reflux mass flowrate SRC (kg/h)

T 12 in SRC (°C)

8

0

240

97 96

0

2

4

(g)

6 Time (h)

8

Reboiler duty SRC (GJ/h)

148 146 144

(i)

Figure 6.7

2

4

6 Time (h)

220 210 200 190

(h)

150

0

230

180 0

10

152 T 23 in SRC (°C)

6 Time (h)

(f)

98

142

4

1400 1200 1000

99

95

2

(d) Reboiler duty REDC (GJ/h)

6 Time (h)

0.96 0.95 0

Reflux mass flowrate REDC (kg/h)

T 3 in REDC (°C)

48

2

0.97

(b)

50

(c)

0.98

10

52

40 0

0.99

8

10

2.2 2.1 2.0 1.9 1.8 1.7 1.6 0

(j)

Control performance of CS 1.

Note that excess amount of EO entering the system will result in unconsumed (i.e. unreacted) EO, while a deficit of EO will lead to water impurities. To overcome this bottleneck, Wu and Chien (2022) suggested the use the EO to feed ratio as one of the MVs to control one of the temperature-sensitive trays so that any changes to the tray temperature resulting from external variations will lead to the automatic

167

168

6 Control of Hybrid Reactive–Extractive Distillation Systems for Ternary Azeotropic Mixtures

adjustment of the EO to feed ratio. We explore next such control structure. Here, note that another alternative approach is to install an online composition controller on the fresh feed stream so that the EO to feed ratio can also be adjusted automatically. Such an online composition controller nonetheless comes with a heavy price and maintenance cost.

6.6.2

Control Structure 2 (CS 2) – Triple Point Temperature Control

From the sensitivity analysis in Figure 6.5, we found that the REDC contains three different temperature-sensitive zones, while the SRC contains two different zones. Since there are three MVs for the REDC (i.e. Qr, RR, and EO to feed ratio) and two MVs for the SRC (i.e. Qr and RR), it is possible to pair one MV with one CV. In the REDC, we decided to pair 3rd tray with RR, 14th tray with EO to feed ratio, and 29th tray with Qr, owing to the nature of their location, which is close to the corresponding MV. The pairing for SRC remains the same as CS 1. We call such a control structure CS 2, the flowsheet of which is given in Figure 6.8. The tuning parameters for all TCs are given in Table 6.2. The control performance of CS 2 is tested using the same throughput and feed composition disturbances, injected at t = 1.5 hours. The results are given in Figure 6.9. From Figure 6.9a, it was observed that the DIPE purity returned to a value close to its nominal value quickly during the +10% throughput disturbance, with a very small (i.e. acceptable) deviation. A similar result was observed for the IPA purity during the +10% throughput disturbance (Figure 6.9d), while the IPA purity experienced a significant drop during the −10% throughput disturbance. This is likely caused by the temperature of 3rd tray in the REDC that failed to return to its nominal value. This abnormality was caused by the same reason as CS 1, i.e. TC 6

FC 3

Purge

PC 21

PC 11

R1 LC 21

LC 11

2

Solvent

4

Feed

SRC

REDC

3 12

13

Reactant

14

29

TC 2

TC 5

23

TC 4 LC 22

FC 2 31

R2 TC 3

Figure 6.8

13

TC 1

FC 1

Process flowsheet for CS 2.

LC 12

24

6.6 Quality Control Structures

Table 6.2

The tuning parameters for all TCs for CS 2. 𝝉i

TC 1

3.132 762

6.6

TC 2

1.464 717

10.56

TC 3

0.309 906

11.88

TC 4

2.058 045

6.6

TC 5

3.186 95

11.88

TC 6

0.092 748

13.2

52

0.9996

50

T 3 in REDC (°C)

0.9998

0.9994 0.9992

+ 10% throughput – 10% throughput + 5% feed comp – 5% feed comp

0.9990 0.9988 0

2

(a)

4 6 Time (h)

8

44

0

2

(b)

4 6 Time (h)

8

10

Feed to reactant ratio

1000

0.97 0.96

800 600 400 200 0

0

2

(d)

4 6 Time (h)

8

10

–200 0

2

(e)

106 104 102 100 0

2

(g)

4 6 Time (h)

8

4 6 Time (h)

8

97 96

2

4 6 Time (h)

8

0.64 0.60 0.56 0.52 0

(j)

Figure 6.9

2

4 6 Time (h)

8

10

10

2

4 6 Time (h)

8

10

200 180

2

(k)

Control performance of CS 2.

4 6 Time (h)

2

4 6 Time (h)

8

10

148 146 144

(i)

220

160 0

8

150

142 0

10

Reboiler duty SRC (GJ/h)

Reflux mass flowrate SRC (kg/h)

0.68

4 6 Time (h)

(f)

240

0.72

2

0.21

152

(h)

0.76

10

0.22

0.20 0

10

98

95 0

10

8

0.23

T 23 in SRC (°C)

T 12 in SRC (°C)

108

4 6 Time (h)

0.24

99

110

2

0.25

1200

0.98

78 76 74 72 70 68 66 64 62 60 58 0

(c)

1400 Reflux mass flowrate REDC (kg/h)

Purity of IPA (mol%)

46

40

10

0.99

0.95

T 29 in REDC (°C)

48

42

1.00

Reboiler duty REDC (GJ/h)

T 14 in REDC (°C)

Kc

Purity of DIPE (mol%)

Controller

8

10

2.10 2.05 2.00 1.95 1.90 1.85 1.80 1.75 1.70 0

(l)

169

170

6 Control of Hybrid Reactive–Extractive Distillation Systems for Ternary Azeotropic Mixtures

the saturation of the RR during −10% throughput disturbance, and is consistent to those reported by existing studies (Wu and Chien, 2022). Other than the temperature of 3rd tray in the RED, the temperature of all other temperature-sensitive trays (Figure 6.9c,g,h,i) returns to their nominal values. Note that some of the temperature-sensitive trays took slightly longer duration to achieve steady-state in comparison to CS 1 (Figure 6.7), especially during the +5% feed composition disturbance. Upon inspecting the corresponding MVs in Figure 6.9f,j–l, it was observed that the RR also becomes saturated during the +5% feed composition disturbance, similar to those observed during the −10% throughput disturbance. Such observation was unexpected, and in our opinion, this could be attributed to the abnormal response in the feed-to-reactant ratio (Figure 6.9f). During the +5% feed composition disturbance, the feed to reactant ratio decreases to about 0.205, and such a value does not correspond to the asymmetric change (i.e. changes of same magnitude) as in during the −5% feed composition. We believe such an abnormality has disrupted the RR, which results in saturation. In our opinion, the abnormality was caused by the CV selected (i.e. temperature of stage 14) that is too close to the reactant feed stage (i.e. stage 13), and thus, any minor adjustment to the reactant flowrate will cause a high fluctuation in the temperature of stage 14, which makes it difficult to control at its nominal value. These abnormalities reflect the inferiority of CS 2. For the composition disturbances, CS 2 was found to provide similar performance as CS 1. In CS 2, the IPA purity drops more severely in comparison to CS 1. These observations suggest that using the triple-point TC that includes the EO to feed ratio as one of the MV does not contribute positively to maintaining the product’s qualities under feed composition variation. Such unfavorable findings can be attributed to the possibility that the temperature-sensitive trays were not determined precisely or that the control pairing between the MVs and CVs is not appropriate. To double-check the location of the temperature-sensitive trays and confirm the suitability of the control pairing, we perform next the SVD analysis.

6.6.3 Control Structure 3 (CS 3) – Triple Point Temperature Control Using SVD Analysis The equation involved for SVD analysis is given by Eq. (6.2) for a given column that contains n trays. It is a steady-state gain matrix, G, that contains the n rows and m columns, which represent the number of control and manipulated variables, respectively. ΔCv ΔMv ∑ G=U VT

G=

(6.2) (6.3)

In Eq. (6.2), the ΔCV and ΔM V denote the change in tray temperatures and step change in MVs, respectively. The gain matrix, G, can be decomposed via the SVD function in MATLAB, given by Eq. (6.3). The U and V in Eq. (6.3) are the orthonor∑ mal matrices, while is a diagonal matrix of the singular values. A comprehensive guideline for the usage of SVD function in MATLAB is available on the MathWorks

6.6 Quality Control Structures

Figure 6.10 REDC.

SVD analysis of

2.0

+0.01% Qr +0.01% RR +0.05% % EO

Singular vector (U)

1.5 1.0 0.5 0.0

–0.5 –1.0

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Trays

website (MathWorks, n.d.). Figure 6.10 shows the result of the SVD analysis for the REDC derived from the sensitivity analysis (i.e. steady-state gain) (Figure 6.5). From Figure 6.10, it was observed that 29th tray displayed the highest U vector value with Qr as MV, while 28th tray displayed the highest U vector value with EO to feed ratio as MV. Thus, we decided to control 28th tray and 29th tray using EO to feed ratio and Qr, respectively. For the RR as MV, it seems that there is no specific tray that displays the highest U vector value. Since we have paired 29th tray with Qr, we decided to pair 15th tray with RR because in 15th tray, the U vector value is the second highest when the RR is used as MV (i.e. lower than Qr but higher than EO to feed ratio). For the SRC, the control structure remains analogous to CS 1 and CS 2. Such a control structure is denoted as CS 3, the flowsheet of which is given in Figure 6.11. The tuning parameters for all TCs are given in Table 6.3.

TC 6

FC 3

Purge

PC 21

PC 11

R1 LC 11

Solvent Feed

2

LC 21 SRC

REDC

4 FC 1

13

Reactant FC 2

28

29

TC 5

TC 2

23

TC 4 LC 22

31

Figure 6.11

13

15

R2 TC 3

12

TC 1

LC 12

Process flowsheet for CS 3.

24

171

172

6 Control of Hybrid Reactive–Extractive Distillation Systems for Ternary Azeotropic Mixtures

Table 6.3

The tuning parameters for all TCs for CS 3.

Controller

Kc

𝝉i

TC 1

2.635 443

11.88

TC 2

1.841 113

TC 3

0.759 195

TC 4

1.691 34

6.6

TC 5

3.585 524

11.88

TC 6

0.092 748

13.2

7.92 15.839 999

The performance of CS 3 is tested using the same disturbances as in CS 1 and CS 2 for the sake of consistency, introduced at t = 1.5 hours. The performances are given in Figure 6.12. In comparison to CS 1 (Figure 6.7) and CS 2 (Figure 6.9), it appears that CS 3 enables the DIPE purity to hold tightly at its nominal value, with a small (i.e. reasonable) deviation, for both throughput and feed composition variation. In addition, CS 3 also provides minor improvement to IPA purity, where the final purity after 10 hours is closer to its nominal value relative to CS 1 and CS 2, for both throughput and feed composition. However, the problem of the reflux flowrate saturation to a 0 kg/h persists during the −10% throughput disturbance, identical to CS 1 and CS 2, which cause the CV (i.e. temperature of tray 15th ) to be unable to return to its nominal value. Wu and Chien (2022) attributed this saturation to the interaction among the quality control loops (e.g. Qr, RR, or EO to feed ratio). Nonetheless, the temperature of all other temperature-sensitive trays returned to their nominal values in a short duration. Here, it can be concluded that choosing a new temperature-sensitive location using the SVD analysis helps to enhance the control performance. To overcome the saturation of RR during −10% throughput disturbance and improve the control performance, Wu and Chien (2022) recommended conducting a closed-loop sensitivity analysis for the throughput disturbances in the steady-state process (i.e. Aspen Plus) to determine the ideal deviation in the temperature of the temperature-sensitive trays so that these deviations can be used as a new setpoint for the TCs. In practice, these deviations can be programmed into the distributed control system as new setpoints, or they can alternatively be adjusted by operators manually when there is a recognized throughput disturbance. Wu and Chien (2022) have successfully demonstrated such practice in their dynamic simulation, with satisfying purity of products when throughput disturbances are introduced. We will investigate next the application of such practices in our control scheme.

6.6.4

Feedforward Control Structure 3 (FF-CS 3)

The closed-loop sensitivity analysis for the ±10% throughput disturbances can be conducted in Aspen Plus by activating the “Design Spec” function to hold the purity of product at their desired specification while varying the MV (e.g. Qr,

85

102

0.9996

80

100

0.9994 0.9992

+ 10% throughput – 10% throughput + 5% feed comp – 5% feed comp

0.9990 0.9988 0

2

4

6

8

1.000

65 60 0

2

4

6

8

0.990 0.985 0.980 0.975 0

2

4

6

8

10

(d)

0

2

4

6

8

10

8

10

8

10

8

10

Time (h)

(c) 0.245

1000 800 600 400 200 0 –200 0

94

Time (h)

1200

2

(e)

Time (h)

96

92

Feed to reactant ratio

0.995

98

10

1400 Reflux mass flowrate REDC (kg/h)

Purity of IPA (mol%)

70

(b)

Time (h)

(a)

75

55

10

T 28 in REDC (°C)

0.9998 T 15 in REDC (°C)

Purity of DIPE (mol%)

6.6 Quality Control Structures

4

6

8

10

0.240 0.235 0.230 0.225 0.220 0

Time (h)

2

4

6

Time (h)

(f) 152

99

104 103 102 101 0

2

6

8

98 97 96 95 0

10

2

(h)

Time (h)

(g) 0.75

4

6

8

10

0.65 0.60 0.55 0

2

(j)

Figure 6.12

4 6 Time (h)

8

10

146 144 142

220 210 200 190 180 2

(k)

4 6 Time (h)

0

2

8

10

4

6

Time (h)

(i)

230

170 0

148

Reboiler duty SRC (GJ/h)

0.70

150

Time (h)

240 Reflux mass flowrate SRC (kg/h)

Reboiler duty REDC (GJ/h)

4

T 23 in SRC (°C)

T 12 in SRC (°C)

T 29 in REDC (°C)

105

2.2 2.1 2.0 1.9 1.8 1.7 1.6 0

(l)

2

4 6 Time (h)

Control performance of CS 3.

RR, or distillate/bottom rate). Once the “Design Spec” has been activated, the throughput disturbances can be incorporated, and the new column temperature profile can be recorded. Then, one can determine the ideal temperature deviation for the temperature-sensitive tray and use it as a new setpoint for the TC. Instead of using Aspen Plus, another alternative approach is to conduct the closed-loop sensitivity analysis in Aspen Plus Dynamics by installing a temporary composition controller to hold the purity of products at their desired setpoint. The latter

173

6 Control of Hybrid Reactive–Extractive Distillation Systems for Ternary Azeotropic Mixtures

TC 3

FC 3

Purge

PC 21

PC 11

R1 LC 11

Solvent Feed

2

LC 21 SRC

REDC

4 FC 1

12

CC 1

13

TC 2

13

Reactant CC 2

FC 2

LC 12

31

Figure 6.13

23

TC 1 LC 22

R2

24

Process flowsheet for the RED with temporary composition controllers.

approach (i.e. closed-loop sensitivity using Aspen Plus Dynamics) is demonstrated in this chapter. Figure 6.13 shows the control scheme with two composition controllers temporarily installed in the REDC to maintain the purity of products for DIPE. There is no composition controller required for the SRC because the TCs are adequate to maintain the IPA purity, provided that the DIPE impurity flowing from the bottom of REDC to SRC is properly controlled. The relay-feedback test is used to determine the appropriate ultimate gain and period for the composition controllers. Introducing the ±10% throughput disturbances into the control scheme with composition controllers (Figure 6.13) enables the ideal temperature deviation for the temperature-sensitive tray to be obtained, given in Figure 6.14. These ideal 4 Ideal deviation for REDC (%/%)

174

0% (Steady-state) +10% FF –10% FF

3 2 1 0 –1 –2 –3 –4 0

5

10

15 Trays

20

25

30

Figure 6.14 Ideal temperature deviation for REDC during −10% throughput change.

6.6 Quality Control Structures

temperature deviations can be introduced manually whenever there is a known throughput disturbance, and such changes can be regarded as a form of feedforward control scheme. The feedforward control scheme considers the effects of the external disturbances introduced to the system (i.e. typically via the fresh feed stream) and provides a corrective action before the disturbances have a chance to impact the Table 6.4 Numerical value of the ideal temperature deviation for sensitive trays in REDC during −10% throughput change. Temperature (o C) Steady state

−10% throughput

70.1

67.0 (−4%)

28

97.5

95.8 (−2%)

29

104.2

103.1 (−1%)

Tray

15

–10% throughput 59.8388 × 0.9 = 53.8549

(a)

(b)

Figure 6.15 The difference in control palette for (a) Steady-state and (b) −10% throughput disturbance for the FF-CS 3.

175

6 Control of Hybrid Reactive–Extractive Distillation Systems for Ternary Azeotropic Mixtures

85

102

0.9996

80

100

0.9994 0.9992

+ 10% throughput – 10% throughput + 5% feed comp – 5% feed comp

0.9990 0

2

4

6

8

10

0

2

4

6

8

0.984 0.980 0

2

4

6

8

10

600 400 200 0 0

2

(e)

4

6

8

103 102 101 100 99 0

2

4

6

8

10

96

0

2

(h)

0.75

4

6

8

Reflux mass flowrate SRC (kg/h)

0.65 0.60 0.55 0

2

(j)

Figure 6.16

4 6 Time (h)

8

10

210 200 190 0

2

4 6 Time (h)

2

8

10

Control performance of FF-CS 3 with SP change.

4

6

8

10

8

10

8

10

Time (h)

150 148 146 144 0

2

4

6

Time (h)

(i)

220

(k)

0

152

Time (h)

230

180

10

0.224

142

10

240

0.70

8

0.228

(f)

97

Time (h)

(g)

6

0.232

Time (h)

98

95

4

Time (h)

0.236

T of Stage 23 in SRC (°C)

T of Stage 12 in SRC (°C)

104

2

0.240

0.220

10

99

105

0

0.244

800

106

92

(c)

1000

Time (h)

(d)

94

Time (h)

1200

–200

96

90

Feed to reactant ratio

0.988

98

10

1400 Reflux mass flowrate REDC (kg/h)

Purity of IPA (mol%)

60

(b)

0.992

T 29 in REDC (°C)

65

55

0.996

98

70

Time (h)

(a)

0.976

75

Reboiler duty SRC (GJ/h)

0.9988

T 28 in REDC (°C)

0.9998 T 15 in REDC (°C)

Purity of DIPE (mol%)

system. It uses an external signal, such as the throughput or composition, to predict the disturbance before it affects the process and generates a corrective action. The feedforward control structure is important because it facilitates the system’s response to disturbances, leading to better product quality and energy efficiency. It additionally reduces the reliance on feedback control, which can be slow to respond to disturbances and may not be able to fully correct for the disturbance. For the sake of illustration, Table 6.4 provides the numerical value of the ideal temperature deviations for the sensitive trays in REDC during −10% throughput change, while the difference in the control palette for the fresh feed and TCs between the steady-state and −10% throughput disturbance is given by Figure 6.15.

Reboiler duty REDC (GJ/h)

176

2.1 2.0 1.9 1.8 1.7 1.6 0

(l)

2

4 6 Time (h)

6.7 Control Performance Evaluation

Here, the setpoints of the TCs in Figure 6.15 are adjusted (i.e. feedforward) based on the ideal deviation obtained from the closed-loop sensitivity analysis (Table 6.4) for a known throughput disturbance. The control performance is given by Figure 6.16. From Figure 6.16d, it was observed that manually changing the SP of the TCs provided a significant enhancement to the IPA product purity during the −10% throughput disturbance. Another important observation made in Figure 6.16 was that during the −10% throughput disturbance, all the TCs (i.e. 15th tray, 28th tray, and 29th tray) in the REDC reached a new steady-state value and this is expected, since the SP of these TCs has been adjusted manually following the ideal deviation from the close-loop sensitivity analysis (Table 6.4). In addition, no saturation was observed for the RR in the REDC (i.e. the final steady-state value of the RR in the REDC is about 4 kg/h instead of 0 kg/h). The performance of DIPE purity remains identical to CS 3 (Figure 6.12). Altogether, it can be concluded that manually changing the SP of the TCs based on the ideal temperature deviation derived from the closed-loop sensitivity analysis (i.e. feedforward control) enables a significant improvement in the control performance of RED. The controllability of FF-CS 3 is now capable of upholding the purity of products at a satisfactory value close to its SP at the end of dynamic simulation.

6.7 Control Performance Evaluation To further evaluate the control performance of the four control structures, the maximum transient deviation and oscillation amplitude are compared using the integral of absolute error (IAE), as described in Eq. (6.4). The IAE is chosen as it considers both transient deviation and oscillation amplitude, which are important indicators of dynamic performance. The IAE is calculated for each product (DIPE and IPA) over a period of 10 hours, including 1.5 hours before the introduction of disturbances. A lower IAE indicates a better dynamic performance. t=10

IAE =

∫0

|e(t)|dt

(6.4)

In Eq. (6.4), the term “e” denotes the deviation of the manipulated variable from its nominal value (e.g. product purity), while “t” represents the duration of dynamic simulation (e.g. 10 hours). To assess the ultimate performance of the four control structures, the sum of IAE values (SIAE) (Eq. (6.5)) for the purities of both products is calculated for the four disturbances tested (i.e. ±10% throughput and ±5% feed composition disturbances). SIAE =

n ∑ IAEi,j

(6.5)

j=1

The i and j in Eq. (6.5) represent the purity of products and the number of tested disturbances. Table 6.5 provides a summary of the SIAE values for the four control structures developed in this study. The SIAE values appear to be consistent with the control performance observed in previous sections. The SIAE value for CS 2

177

178

6 Control of Hybrid Reactive–Extractive Distillation Systems for Ternary Azeotropic Mixtures

Table 6.5 The SIAE comparison of the four control structures for RED. Control structures

SIAE

CS 1

0.2329

CS 2

0.3404

CS 3

0.1411

FF-CS 3

0.0976

is 46% higher than that of CS 1, which is due to the longer time required for certain temperature-sensitive trays to reach steady-state, especially during the +5% feed composition disturbance. This delay causes a longer transient time required for the purity of products to return to their nominal values. Additionally, during the composition disturbances, the IPA purity in CS 2 drops more severely than in CS 1, possibly due to imprecise determination of temperature-sensitive trays or inappropriate control pairing between the MVs and CVs. These anomalies indicate that CS 2 is inferior to CS 1, resulting in a considerably higher SIAE value. After employing the SVD analysis to select alternative temperature-sensitive trays and control pairing, CS 3 demonstrated a significant improvement of 39% and 59% in SIAE relative to CS 1 and CS 2, respectively. This improvement was attributed to the tighter control of DIPE purity at its nominal values, for both throughput and feed composition variations. Likewise, CS 3 provided minor improvements in IPA purity compared to CS 1 and CS 2, and all temperature-sensitive trays returned to their nominal values after a shorter duration. However, CS 3 still faced the issue of saturation of the RR during the negative throughput disturbance, and to overcome this issue, a feedforward control scheme was employed (FF-CS 3), which determined the ideal temperature deviations during the negative throughput disturbance through closed-loop sensitivity analysis in the steady-state process, and used these deviations as a new setpoint for the TCs. The FF-CS 3 structure provided significant enhancement to the IPA product purity during the negative throughput disturbance, with no observed saturation for the RR in the REDC. Therefore, the SIAE of FF-CS 3 is the lowest among the four control structures and is 58%, 71%, and 31% lower than CS 1, CS 2, and CS 3, respectively.

6.8 Conclusions Here we present a progressive step in setting up effective control schemes for upholding the purity of products of a RED system. Starting with setting up the inventory control structure, we also demonstrated the steps required to select the temperature-sensitive tray(s) through both sensitivity and SVD analyses, to setting up the quality control structures. In this chapter, several full control schemes were presented, in an evolutionary manner, so that our audience could grasp the advantages and drawbacks of each scheme, from using the simplest TC to the

Acronyms

complicated feedforward control scheme with setpoint change. The study showed that the use of triple point TC (CS 3) provided reasonable control performance, but there were still issues with saturation in the reflux flowrate and slight deviation in IPA purity during the −10% throughput disturbance. To address these issues, a feedforward control scheme was developed for CS 3 (FF-CS 3), where the SP of the TCs was manually adjusted based on ideal temperature deviations obtained from closed-loop sensitivity analysis. This approach significantly improved the control performance of RED. The newly developed FF-CS 3 can now maintain the purity of product at a satisfactory value close to its SP at the end of dynamic simulation.

Acknowledgements Z.Y. Kong gratefully acknowledges the support from Sunway University Malaysia. The authors also appreciate the technical support given by Ms. Y.Y. Chen from the National Taiwan University of Science and Technology.

Acronyms CV CED DIPE DMSO EA EDC EG EO IAE IPA MV P PC PI Qr RD RDC RED REDC RR SIAE SRC SVD TAC TC THF

control variable conventional extractive distillation diisopropyl ether dimethyl sulfoxide ethyl acetate extractive distillation column ethylene glycol ethylene oxide integral absolute error isopropanol manipulating variable proportional pre-concentration column proportional and integral reboiler duty reactive distillation reactive distillation column reactive-extractive distillation reactive-extractive distillation column reflux ratio sum of integral absolute error solvent recovery column singular value decomposition total annual cost temperature controller tetrahydrofuran

179

180

6 Control of Hybrid Reactive–Extractive Distillation Systems for Ternary Azeotropic Mixtures

Nomenclature xi Kc 𝜏 ∑i ΔCV ΔM V G m n U and V

mole fraction of the component gain (%%−1 ) integral time (min) diagonal matrix of the singular values change in tray temperatures step change in manipulating variable steady-gain matrix that contains n rows and m columns number of columns in SVD number of trays in SVD orthonormal matrices

References Arifin, S. and Chien, I.L. (2008). Design and control of an isopropyl alcohol dehydration process via extractive distillation using dimethyl sulfoxide as an entrainer. Industrial and Engineering Chemistry Research 47: 790–803. https://doi.org/10.1021/ie070996n. Kong, Z.Y., Sánchez-Ramírez, E., Yang, A. et al. (2022a). Process intensification from conventional to advanced distillations: past, present, and future. Chemical Engineering Research and Design 188: 378–392. https://doi.org/10.1016/j.cherd.2022 .09.056. Kong, Z.Y., Sunarso, J., and Yang, A. (2022b). Recent progress on hybrid reactive-extractive distillation for azeotropic separation: a short review. Frontiers in Chemical Engineering 4: 986411. Kong, Z.Y., Yang, A., Chua, J. et al. (2022c). Energy-efficient hybrid reactive-extractive distillation with a preconcentration column for recovering isopropyl alcohol and diisopropyl ether from wastewater: process design, optimization, and intensification. Industrial and Engineering Chemistry Research 61: 11156–11167. https://doi.org/10 .1021/acs.iecr.2c01768. Li, W., Shi, L., Yu, B. et al. (2013). New pressure-swing distillation for separating pressure-insensitive maximum boiling azeotrope via introducing a heavy entrainer: design and control. Industrial and Engineering Chemistry Research 52: 7836–7853. https://doi.org/10.1021/ie400274d. Liu, J., Wan, G., Dong, M. et al. (2023). Dynamic controllability strategy of reactive-extractive dividing wall column for the separation of water-containing ternary azeotropic mixture. Separation and Purification Technology 304: 122338. https://doi.org/10.1016/j.seppur.2022.122338. Liu, J., Yan, J., Liu, W. et al. (2022). Design and multi-objective optimization of reactive-extractive dividing wall column with organic Rankine cycles considering safety. Separation and Purification Technology 287: 120512. https://doi.org/10.1016/j .seppur.2022.120512. Luyben, W.L. (2013). Distillation Design and Control Using Aspen Simulation, 2e. Wiley.

References

MathWorks (n.d.). SVD. [WWW Document]. https://www.mathworks.com/help/ matlab/ref/double.svd.html (accessed 29 November 2023). Su, Y., Yang, A., Jin, S. et al. (2020). Investigation on ternary system tetrahydrofuran/ethanol/water with three azeotropes separation via the combination of reactive and extractive distillation. Journal of Cleaner Production 273: 123145. https://doi.org/10.1016/j.jclepro.2020.123145. Wang, C., Zhuang, Y., Liu, L. et al. (2021). Design and comparison of energy-saving double column and triple column reactive-extractive hybrid distillation processes for ternary multi-azeotrope dehydration. Separation and Purification Technology 259: 118211. https://doi.org/10.1016/j.seppur.2020.118211. Wu, T.-W. and Chien, I.-L. (2022). Novel control strategy of intensified hybrid reactive-extractive distillation process for the separation of water-containing ternary mixtures. Separation and Purification Technology 294: 121159. https://doi.org/10 .1016/j.seppur.2022.121159. Yan, J., Liu, J., Ren, J. et al. (2022). Design and multi-objective optimization of hybrid reactive-extractive distillation process for separating wastewater containing benzene and isopropanol. Separation and Purification Technology 290: 120915. https://doi.org/ 10.1016/j.seppur.2022.120915. Yang, A., Kong, Z.Y., and Sunarso, J. (2023). Design and optimisation of novel hybrid side-stream reactive-extractive distillation for recovery of isopropyl alcohol and ethyl acetate from wastewater. Chemical Engineering Journal 451: 138563. https://doi.org/ 10.1016/j.cej.2022.138563. Yang, A., Su, Y., Sun, S. et al. (2022). Towards sustainable separation of the ternary azeotropic mixture based on the intensified reactive-extractive distillation configurations and multi-objective particle swarm optimization. Journal of Cleaner Production 332: https://doi.org/10.1016/j.jclepro.2021.130116. Zhang, Y.R., Wu, T.W., and Chien, I.L. (2021a). Intensified hybrid reactive-extractive distillation process for the separation of water-containing ternary mixtures. Separation and Purification Technology 279: https://doi.org/10.1016/j.seppur.2021 .119712. Zhang, Q., Hou, W., Ma, Y. et al. (2021b). Dynamic control analysis of eco-efficient double side-stream ternary extractive distillation process. Computers and Chemical Engineering 147: 107232. https://doi.org/10.1016/j.compchemeng.2021.107232.

181

183

7 Process Design and Control of Reactive Distillation in Recycle Systems Mihai Daniel Moraru 1 , Costin Sorin Bildea 2 , and Anton Alexandru Kiss 3 1 Department of Technology, Engineering and Projects, Westlake Epoxy, Seattleweg 17, 3195 ND Pernis, The Netherlands 2 Department of Chemical and Biochemical Engineering, University Politehnica of Bucharest, Str. Gh. Polizu 1-7, 011061 Bucharest, Romania 3 Department of Chemical Engineering, Delft University of Technology, Van der Maasweg 9, 2629 HZ, Delft, The Netherlands

7.1

Introduction

Reactive distillation (RD) is an important process intensification technique that can be used as a promising alternative to classical reaction–separation sequences. The RD technology has already been applied productively in the chemical process industry, particularly to overcome the challenges of chemical-equilibrium-limited reactions. The main drivers for RD applications are economical (20–80% lower variable cost, CAPEX, and energy usage), environmental (low GHG emissions, avoidance or reduction of solvent use or salt waste), and social (enhanced safety due to smaller reactive holdup and lower run-away sensitivity). Kiss (2022) provides a general overview of RD processes, covering fundamentals, modeling, and design, as well as industrial applications. This work centers on the design and control of RD columns that are an integral part of a larger chemical process, where recycle streams (to these columns) are present, as typical to industrial practice. These recycle streams originate from the incomplete conversion of reactants or their conversion to undesired by-products. Therefore, the unreacted reactants need to be recovered and recycled, while the by-products need to be recovered and if possible reconverted into reactants followed by their recycle to the RD columns. These aspects have a major impact on achieving the component inventory in the process, as well as achieving the desired capacity and maintaining the required products purity. The previous work by Baki and Kaymak (2014) found that a limited number of studies on this topic is available in the open literature, such as the studies of Luyben et al. (2004) and Wang et al. (2011) on synthesis of butyl acetate and methanol by the trans-esterification of methyl acetate with butanol, Lin et al. (2008) on methyl acetate hydrolysis, Wang et al. (2010) on making dimethyl carbonate (DMC) and ethylene Control and Safety Analysis of Intensified Chemical Processes, First Edition. Edited by Dipesh Shikchand Patle and Gade Pandu Rangaiah. © 2024 WILEY-VCH GmbH. Published 2024 by WILEY-VCH GmbH.

184

7 Process Design and Control of Reactive Distillation in Recycle Systems

glycol by trans-esterification of ethylene carbonate (EC) and methanol, and Pathak et al. (2011) on cumene synthesis from benzene and propylene. More recent studies are those of Lee et al. (2019) on diphenyl carbonate and Patrascu et al. (2022) on DMC production by the indirect alcoholysis of urea. This work presents a case study of industrial importance, in which the RD column is an integral part of a larger chemical process, namely the production of 4-hydroxybutyl acrylate by esterification of acrylic acid with 1,4-butandiol. This process is used to illustrate key aspects of design and control of RD systems with recycles. The design is focused on developing the topology of the entire process and solving the mass and energy balance. The control is focused on developing a plantwide strategy to achieve the material inventory (or differently said, balancing the reactions’ stoichiometry), along with achieving the desired production rate and product purity. Several key perturbations and process changes (e.g. throughput and reactant composition) are implemented to test the control structure of the plant. All this work makes use of both steady-state and dynamic, rigorous process simulations.

7.2

Design of Reactive Distillation Processes

RD can be classified into homogeneous processes (either auto-catalyzed or homogeneously catalyzed) and heterogeneous processes (catalyzed by a solid catalyst), also known as catalytic distillation. Another classification commonly made is based on the phase characteristics of the component mixture, which can be homogeneous (i.e. a single liquid phase) or heterogeneous (i.e. two liquid phases; in particular, the mixture presents minimum boiling heterogeneous azeotrope that allows a relatively easy liquid–liquid split after condensing the distilled vapor). The following discussion refers to the former classification, as the equipment design and internals required are different for each class of RD. Equipment and internals. In the case of autocatalytic reactions, the rate can be affected only by pressure or temperature. In addition, homogeneous catalysis can influence the reaction rate by changing the catalyst concentration, so that the reaction rate can be properly adjusted to the needs of the RD equipment. Homogeneous catalysis is more flexible, but it needs difficult separation steps for catalyst recovery, or it leads to salt waste formation (due to the downstream neutralization of cheap acid / base catalysts). Heterogeneous catalysis is simpler in principle, but it needs more equipment volume and suffers from increased operating temperature and limited catalyst lifetime (Schoenmakers and Bessling, 2003). When using solid catalysts, a special construction is needed in practice – such as catalyst packed in ‘‘tea bags’’ on trays or sandwiched in structured packing (Sulzer Katapak) – in order to fix the solid catalyst particles in the reactive zone. Such constructions put a limit on the catalyst concentration that can be attained, as the reaction rate can be increased only up to the limit set by the attainable concentration range in an RD column. The choice of internals for RD distillation is much more limited as compared to classical distillation. Kiss (2022) provides an overview of the specific internals

7.2 Design of Reactive Distillation Processes

used in RD processes and their performance in terms of separation and reaction. Conventional internals can be used for non-catalyzed or homogeneously catalyzed RD processes, as no development of new internals is required. The homogeneous catalyst is usually fed together with the heavy liquid feed stream (or added to the reflux), and the liquid holdup of the column internals has to be maximized toward obtaining high or full conversion. The choice of internals depends on the reaction rate. Slow reactions require large liquid holdups and long residence times in an RD column. Tray columns are typically used for such cases, applying a bubbly flow regime to ensure a larger liquid holdup and a higher residence time. In the case of fast reactions, the separation efficiency is a more important selection criterion for internals than the liquid holdup. Random or structured packing can be used as internals to provide a higher specific surface area and great separation efficiency, as the chemical equilibrium can be reached within relatively short residence times (Schoenmakers and Bessling, 2003). Heterogeneously catalyzed RD processes require specific internals to immobilize the solid catalyst in the desired region. Moreover, it is important to ensure an adequate reaction rate by allowing sufficient contact between the liquid phase and the active sites of the catalyst. The most commonly used catalysts for RD processes are acidic ion-exchange (IEX) resins, such as Amberlyst or Lewatit. Various technologies for immobilizing solid catalysts have been developed and reported elsewhere (Kiss, 2022). Equipment design. RD sets clear specifications for the reaction conversion and the product compositions. Accordingly, the degrees of freedom (DoF) in an RD column have to be tuned accordingly to achieve these specifications while also optimizing an objective function (e.g. total annual cost). Several specifications are typically required (Luyben and Yu, 2008): ●

●

● ● ● ●

RD column pressure and the pressure drop along the stage or the full column. Note that setting the pressure fixes both the top and bottom column temperatures, which are constrained by the cold/hot utilities available on-site. If the bottom temperature is too high, one could use dilution with a reactant that is recovered afterward. Working at the highest acceptable temperature is better, as it leads to faster reaction rates. Total number of stages, the reactants feed locations, and the exit points of the product streams. This sets the RD configuration in terms of the rectification, reaction, and stripping sections. Top distillate (liquid, vapor, mixed) or bottom product, as ratios or absolute values. Type of the condenser (partial or total) and reboiler (kettle or thermosyphon). Reflux ratio or the vapor boilup ratio. Liquid holdup distribution on the reactive stages.

The usual design specifications in classical distillation are the concentrations of heavy key component (HKC) in the distillate and light key component (LKC) in the bottom product. The liquid holdup has no effect on the steady-state design of a simple distillation column, only on the dynamic behavior. The diameter of the column can be estimated using vapor-loading correlations, after determining the vapor rates

185

186

7 Process Design and Control of Reactive Distillation in Recycle Systems

(reflux ratio) needed to achieve a desired separation. Conversely, in the case of RD processes, the reaction rates strongly depend on the liquid holdup and the amount of catalyst on each stage, so these parameters are crucial. An iterative design procedure is required for RD columns, as the liquid holdup must be known prior to designing the column. A reasonable liquid holdup per tray is assumed initially, and the RD column is designed to reach the desired conversion and product purity. Afterwards, the column diameter is determined, along with the liquid height on the reactive trays that is required for the assumed holdup. Liquid heights of up to 10–15 cm are common, to avoid hydraulic pressure-drop limitations. If the determined liquid height is too high, then a smaller liquid holdup is assumed and all calculations are repeated (Luyben and Yu, 2008). For the hydraulic design of an RD column, a shortcut method can be used (Dimian et al., 2014): 1. 2. 3. 4.

5. 6.

7.

8. 9.

Estimate an average volumetric liquid flow rate for the operation. Assume a value for the liquid velocity at the ‘load point’ (ULP ): e.g. 10 m3 /m2 h. Assume a value for the number of theoretical stages per meter (NSTM), e.g. 2. Calculate the column diameter. Considering the specifications of the internals used, estimate the packing volume, the liquid, and the catalyst holdup per reactive stage. Introduce the above values in a process simulation, expressing the reaction rate in units that are compatible with the liquid holdup (mass, molar, or volumetric). Determine the total number of reactive stages required to achieve a target conversion. Check the temperatures, concentrations, and reaction rate profiles. Get from the simulation the liquid and gas flows, as well as fluid properties. Recalculate the load point velocity and liquid holdup by using specific correlations and diagrams. Check the hydraulic design by selecting packings with similar characteristics. Verify if the gas (vapor) load and the pressure drop are within the optimal region. Check all values and repeat the points 4 to 8 until acceptable values are achieved.

Other design and feasibility check methods for RD systems are available in literature. They can be classified into three groups: (1) graphical and topological methods, (2) optimization techniques, and (3) heuristic and evolutionary approaches. Kiss (2022) summarizes all these methods, including their principles, assumptions, and a brief pro/con analysis. RD process design. When the RD column is an integral part of a larger process involving recycles, the column design guidelines presented above are no different. What is different is that, in processes with recycles, the component mass balance around the column has to be solved first: in most cases, in a single-column RD process, the feed streams to and product streams from the column are well known, while in recycle systems, the recycle streams – which may implicitly be part of the feeds or fed individually to the RD column – are unknown. Roughly speaking, the process design (i.e. process topology) activity becomes important and, next to the equipment design activity, adds another layer of complexity to the overall design. However, one observation is that in processes with recycles, the process and equipment design activities are to some extent influencing each other due to their iterative nature. At first,

7.2 Design of Reactive Distillation Processes

the iterative calculations are simple, based on simplifying assumptions, and describe key characteristics of both design activities. Later, the calculations become rigorous, based on physicochemical principles, and describe a wider range of characteristics of these two design activities. This is in line with and closely follows the steps of the initial conceptual design methodology proposed by Douglas (1988) and of that revised by Dimian and Bildea (2008). Simply put, the philosophy – not necessarily a methodology – adopted here starts simple and adds complexity in a progressive manner. In practical steps, this is one way to put together a process and equipment design of such a process: (1) Initial-level calculations (spreadsheet or hand calculations are doable): • Understand the chemical components present in the process, the physical properties of pure components and mixtures, and in particular the azeotropy and phase equilibria of the system, as well as the stoichiometry of the main and secondary reactions. • Answer the question: Is it likely that RD can be employed in this process? – if the answer is no, then stop; else, proceed with investigating its technical feasibility (Shah et al., 2012). • Make an overall component mass balance, say at the input/output (black-box) level, based on the feeds and stoichiometries, in order to have a feel of the magnitude of component flows / concentrations. • Consider that inside the black box there is an RD column, and make a mass balance based on the most likely behavior of the column, in terms of overhead and bottom components, considering fresh feeds and stoichiometry. • Add additional black-box separation steps to recover reactants (for recycling) and eliminate products (from the process) that are present in the outlet streams of the column, but keep the splits at a high level. • Other additional black-box reactive or reactive-separation steps should also be implemented if deemed necessary. • Close the recycles, adjust the inputs, and re-iterate the calculations. (2) Intermediate-level calculations (the use of a process simulator is highly advised): • Use rigorous thermodynamic calculations for physical properties, especially when calculating the phase equilibria. • Implement chemical equilibrium-based reactions, preferable kinetic-based reactions, and phase equilibrium-based separations inside the RD column. Specify feasible recoveries and product purities. Determine and implement key column characteristics like the number of stages, amount of catalyst or hold-up volume for reaction, and stage pressure drop. Other calculations (such as rigorous hydraulics) can be postponed. Note that analogous to the classical Reaction–Separation–Recycle systems where the reactor is central and needs to be solved first, here, the design of the RD is in focus. This is because its product streams represent the starting point in developing the additional separation steps that are required. Of course, if multiple reactive(-separation) steps are foreseen/necessary, then one needs to decide which comes first.

187

188

7 Process Design and Control of Reactive Distillation in Recycle Systems

• Implement equilibrium-based calculations for other separations, reaction, or reactive-separation steps for reactants recycle and products elimination from the process. • Close the recycles, adjust inputs, and re-iterate the calculations accordingly. (3) Complex-level calculations (the use of a process simulator is unavoidable at this stage): • Add the required level of complexity deemed to be necessary for answering all questions when determining the techno-economic feasibility of the overall process. • Always re-iterate the calculations until convergence is achieved. Based on our experience, achieving the mass and energy balance of the overall plant using kinetic-based rate equations, thermodynamic-based phase equilibria separations, and closing the recycles is the most challenging task of the process design activity. The use of steady-state process simulators is unavoidable, while in some instances, the use of the dynamic process simulators is necessary for achieving the steady-state mass and energy balance. This has to do with the approach to solve and the numerical solvers used to find the solution of the system of equations. Luyben (2004) describes in detail how Aspen Dynamics is used to find a steady-state solution of a process, and several case studies are presented.

7.3

Control of Reactive Distillation Processes

RD is a great example of the subtle interaction between design and control. The steady-state and dynamic aspects of a process should be considered at all R&D stages and commercialization, from laboratory to pilot plant to full-scale production (Luyben and Yu, 2008; Sharma and Singh, 2010). Typically, the controllability of an RD column is improved by adding more reactive trays while taking into account the trade-off between steady-state design and dynamic controllability. The neat operation mode requires that the reactants are fed according to the stoichiometric ratio. For this reason, the control system must be able to detect any imbalance that leads to a gradual accumulation of a reactant, lower conversion, and product purity. Another option is to operate an RD column with an excess of reactant to make the control easier, but this requires the recovery and recycle of the (unreacted) excess reactant. Note that, even in the latter case, the stoichiometry still needs to be balanced out at the plant level. The control of an RD column is a rather challenging task in practice due to process nonlinearities and complex interactions between the VLE and chemical reactions. Various control methods can be used for RD processes, ranging from simple PID controllers (Moraru and Bildea, 2017; Moraru et al., 2022b) to advanced model predictive controllers (MPCs) – including dynamic matrix control (DMC), quadratic dynamic matrix control (QDMC), robust multivariable predictive control technology (RMPCT), generalized predictive control (GPC), and others (Sharma and Singh, 2010). Independent of the type of the control methodology adopted, the basic control structure has to achieve three main objectives: production rate, product purity, and

7.3 Control of Reactive Distillation Processes

reaction stoichiometry. There are only a few process control structures that are applied to the neat operation of both homo- and heterogeneous RD processes. These process control structures use inferential temperature control (or concentration analyzers) at some location in the RD column to balance the stoichiometry. Figure 7.1(a) illustrates three control structures, reported in the literature, which aim to fulfill the main control objectives, albeit in different ways. Notably, these process control structures apply to both homogeneous and heterogeneous RD processes. 1) Control structure S-1 (Roat et al., 1986) manipulates the reboiler duty (or the vapor boil up) to set the production rate. The flow rates of both fresh reactants are used (either directly or involving ratio control) to control two temperatures in the column. These two control loops work jointly to achieve the specified product purity and to feed the reactants in the correct stoichiometric ratio. Luyben and Yu (2008) applied this control structure to several processes for acetic acid esterification. 2) Control structure S-2 (Huang et al., 2004) fixes the flow rate of one reactant (throughput manipulator) and manipulates the flowrate of the other reactant (directly or involving ratio control), and the reboiler duty to control two temperatures in the column. This was applied to RD processes for the synthesis of butyl propionate and butyl acetate (Huang et al., 2004). Control structures based on the same idea were also applied to RD processes for the production of amyl acetate (Hung W.J. et al., 2006), triacetin (Hung et al., 2014), butyl and amyl acetates (Hung et al., 2006), n-propyl propionate (Xu et al., 2014), butyl levulinate (Chung et al., 2015), DMC (Wang et al., 2010), 1,3-dioxolane (Pan et al., 2020), various esters of acetic acid (Luyben and Yu, 2008), and diphenyl carbonate (Lee et al., 2019). 3) Control structure S-3 manipulates the flow rate of one reactant to set the production rate, while the ratio between the reboiler duty and limiting reactant flow rate is kept constant. The flowrate of the other reactant (or reactant ratio) is used to control one temperature in the column. Similar control structures were applied to RD processes for making butyl acrylate (Zeng et al., 2006) and butyl levulinate (Chung et al., 2015). Figure 7.1 (b) shows three novel control structures proposed recently by Moraru et al. (2022b), which can be applied to heterogeneous RD processes only (such as some esterifications). In all these new process control structures, the light reactant flowrate is used to set the production rate. The quality of the bottom product is controlled by the reboiler duty of the RD column. The feedback required to set the correct stoichiometric ratio between the fresh reactants is obtained from the measured reflux ratio (S-4) or reflux rate (S-5). In the case of control structure S-6, a part of the heavy reactant (the alcohol, immiscible with water) is fed to the decanter. The reflux ratio (or the reflux rate) is set, and the level of the organic phase is controlled by the alcohol flowrate. Plantwide control. Independent of the individual steps comprising a chemical process, the plantwide control (PWC) system must achieve its basic goals: setting the production rate, keeping the product purities at the required values, and

189

190

7 Process Design and Control of Reactive Distillation in Recycle Systems Condenser

S-1

Condenser

S-4

TC

TC

Decanter LC

PC

Decanter

Heavy reactant

LC

PC

LC

LC

Heavy reactant Water

Water

T sp

FC

TC

X

(L/D)sp

FC

FC

RC

X

RD

RD

Fsp

TC

FC

Fsp

T sp

FC

Light reactant LC

TC

Light reactant LC

Reboiler

T sp Reboiler

Ester

Condenser

S-2

Ester

Condenser

S-5

TC

TC

Decanter LC

PC

Decanter

Heavy reactant

LC

PC

LC

LC

Heavy reactant Water

Water

T sp

FC

X

Lsp

FC

TC

FC

X

RD

RD

Fsp

Fsp FC

FC

TC

Light reactant LC

T sp

TC

Light reactant LC

Reboiler

T sp Reboiler

Ester

Ester

Condenser

S-3

Condenser

S-6

Alcohol LC

TC

TC

Decanter PC

LC

Decanter PC

LC

Heavy reactant

LC

Heavy reactant Water

T

FC

X F

(L/D)sp RD

F

sp

FC

Light reactant

(a)

(F/F)sp X

RD

sp

FC

(Q/F)sp

X

FC TC

Water

FC

sp

X

TC

Light reactant

Reboiler LC

LC FC

Ester

T sp Reboiler

Ester

(b)

Figure 7.1 Control of reactive distillation processes involving a heterogeneous azeotrope. (a) Classic control structures from literature. (b) Newly proposed control structures

maintaining the component inventory (or in other words, balancing the reaction stoichiometry when chemical transformations are involved). Therefore, no matter how complex the individual step may or may not be, these objectives are valid also for processes with recycles in which reactive-distillation columns are present. Hence, the well-known principles and methodologies described in the literature can be used to

7.3 Control of Reactive Distillation Processes

develop the PWC system (e.g. Luyben et al., 1999; Skogestad and Postlethwaite, 2005; Dimian et al., 2014), including those pertaining to reactive-distillation columns and other unit operations. However, it is worth mentioning that for such processes, special attention has to be given to the control system of the RD column. This is because, due to its fundamental tasks, the column has to achieve simultaneously a certain conversion of reactant(s) and a certain level of separation. If the control system of the column fails to achieve these two goals, it will most likely lead to a failure of the PWC system in achieving the overall control objectives. We find that maintaining the component inventory is not an easy task, especially in processes with recycles and when chemical reactions are present. In most of the situations, the reason is the accumulation of one of the components: either the reaction section has insufficient capacity to achieve the conversion of reactants, or the separation section is not able to remove one of the components, while the recycle “helps” to build-up components within the process. The typical control design problem is to obtain a controller that follows certain specifications for a given system described by a certain model. This could be relatively simple for small systems such as standalone unit operations. Nonetheless, when addressing a full chemical plant including intensified units (such as RD columns), several key issues may arise. For example, the complete process model may be unavailable, the controlled or manipulated variables (MVs) may still have to be chosen, and the control specifications may be incomplete or unclear. Consequently, most traditional process control design techniques are not able to cope with large processes. From a mathematical viewpoint, PWC is actually a complex combinatorial problem involving a large number of decision variables. The PWC involves the following tasks (Correa de Godoy and Garcia, 2017): 1. 2. 3. 4. 5.

Decision on the control objectives; Choice of the MVs; Choice of the controlled variables (CVs); Selection of any extra measurements; Choice of the process control configuration (i.e. structure of the overall controller that interconnects all variables that are measured, manipulated, and controlled); 6. Choice of the type of controller. The PWC design implies the selection of controlled and manipulated variables, extra measurements, control configuration, and controller type. Typically, it includes all of the structural decisions of the control systems but not the actual design of the system. The major PWC design techniques proposed in the literature are addressed in a review paper (Correa de Godoy and Garcia, 2017). An integrated framework of simulation and heuristics is recommended to be used as a hybrid PWC technique, aided by computer simulations at each step (Konda et al., 2005). The steps of this integrated framework are: 1. Define the PWC objectives; 2. Determine the degrees of freedom usable for process control;

191

192

7 Process Design and Control of Reactive Distillation in Recycle Systems

3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

Identify and analyze the plantwide disturbances; Set the performance and controller tuning criteria; Select the production rate manipulator; Select the product quality manipulator; Select manipulators for more severe CVs; Select manipulators for less severe CVs; Design the control for the unit operations; Check the component material balances; Analyze and address the effects due to integration and recycles; Enhance the control system’s performance (if possible).

7.4 Case Study: RD Coupled with a Distillation–Reactor System and Recycle 7.4.1

Basis of Design and Basic Data

This industrially relevant case study refers to the synthesis of 4-hydroxybutyl acrylate (HBA). This reaction has been performed at lab scale using a solid-base catalyst, as reported by Yang et al. (2008a,b). However, the possibility of applying this reaction at a large scale using a new RD-based process was not yet investigated, although a preliminary conceptual design was reported in a conference paper (Moraru et al., 2020) and a detailed study of three RSR processes (Moraru et al., 2022a,b). In this respect, we present here the design and control of a process comprised of an RD column coupled with an additional separation step (say, for example, a distillation column) for the recovery of HBA product and a reactor for the hydrolysis of by-product 1,4-butanediol diacrylate (BDA) back to acrylic acid (AA) and HBA. The plant capacity considered is ∼23 kt/a of HBA at a purity of 99.4% HBA. Notably, HBA is used industrially to obtain homo-/co-polymers employed in many products, such as coatings, pressure-sensitive additives, or photosensitive resins. HBA is also used industrially in chemical syntheses, as it undergoes addition reactions with a large variety of inorganic and organic components (BASF brochures, 2016a,b). The usual product specifications for HBA are min. 97% purity, max. 0.3% acrylic acid, max. 0.5% diacrylate, and max. 0.1% water content (all by mass). The chemistry routes to make HBA at the industrial scale are: (1) direct esterification of AA with 1,4-butanediol (BD), leading to HBA and water as by-product, and 2) trans-esterification of methyl acrylate with BD, leading to HBA and methanol as by-product. Another undesired by-product that is formed in both routes is BDA, due to the reaction between the HBA product and the AA reactant present in the system (while water by-product is also obtained in this esterification reaction). The paper by Ostrowski et al. (2011) provides more information on the possible secondary reactions. The authors employ a computational quantum mechanical modeling method (density-functional theory) to calculate the activation energy of the

7.4 Case Study: RD Coupled with a Distillation–Reactor System and Recycle

side reactions in a number of esterification systems. These secondary reactions are the following (in order of increasing activation energy): (1) addition of acrylic acid to the double bond in acrylates, (2) additions of alcohols and water to acrylates, (3) dimerization of acrylic acid, and (4) addition of alcohol to the acid dimmer (leading to the same product as the addition of AA to an acrylate). It is worth noting that the lower the activation energy relative to one another, the easier the chemical reaction occurs, thus favoring the by-product formation in the lower activation energy reactions. Most data about these chemical routes and HBA manufacturing is available in the patent literature. A BASF patent (Dockner et al., 1995) describes the synthesis of 1,4-butanediol mono-acrylate via esterification of AA with BD in a process that leads to an aqueous solution of unreacted BD that is converted to tetrahydrofurane (THF) in the presence of a strong acid as catalyst, and eventually the THF is removed from the aqueous solution. Rohm and Haas Co. (Curtis, 2008) claim an enhanced process that yields high-purity hydroxyalkyl acrylates (HAA) prepared from acrylic acid and alkylene oxides. Mitsubishi Chemical (Tokuda et al., 2009) and Nippon Shokubai (Jinno et al., 2015) disclosed similar methods for the synthesis of HAA from acrylic acid and an alkylene oxide. Osaka Organic Chemical Industry Co. Ltd. (Sugiura et al., 2013) described a process for preparing HBA by the trans-esterification of an alkyl acrylate with BD in the presence of a dialkyltin oxide. A later patent from the same company (Tanaka et al., 2017) claims that HBA can be produced by the trans-esterification of MA with BD in the presence of a dialkyltin oxide catalyst (such as dioctyltin/dilauryltin/dibutyltin oxide), which can be recovered and reused. This reaction is carried out in a solvent (e.g. cyclohexane or methylcyclohexane) in the presence of an inhibitor (e.g. phenothiazine) that is needed to minimize polymerization. The same patent (Tanaka et al., 2017) also provides some details about the direct esterification reaction, particularly that by-product formation is significant. The process has a complex neutralization step due to the homogeneous acid used as a catalyst (sulfonic acid or para-toluene sulfonic acid), leading to waste salts generated in large amounts. Moreover, the unreacted AA remains in the HBA product, thus preventing its direct use as a raw material in some high-end applications. The paper by Yang et al. (2008a) reports that the direct esterification reaction can be carried out using a solid catalyst, namely a strongly acidic IEX resin such as Amberlyst 15. By using a heterogeneous catalyst, most of the problems related to using liquid catalysts can be avoided (including the recovery and re-use of the catalyst, difficult product recovery and purification, corrosion, and problems related to waste disposal). Nonetheless, the process remains challenging due to the side reaction that leads to the undesired BDA by-product. The chemistry of the HBA synthesis is rather complex, as it involves several reactions leading to many unwanted by-products. Yet, it is well known that using inhibitors can reduce the formation of by-products. Yang et al. (2008a) reported the HBA synthesis using two esterification reactions: the main reaction leading to the HBA product (Eq. (7.1)) and a side reaction forming the unwanted BDA by-product

193

194

7 Process Design and Control of Reactive Distillation in Recycle Systems

(Eq. (7.2)), along with water by-product that is formed in both esterification reactions, which are equilibrium-limited. The kinetic model proposed in that research study is described by Eqs. (7.3)–(7.8). O

O H2C

OH

HO

Acrylic acid AA

H 2C

1,4-Butanediol BD

O H2C

OH

4-Hydroxybutyl acrylate HBA

O OH

H2C

OH

O

H 2O

O OH O

O

H2C

CH2

O

H2O

O

Acrylic acid AA

(7.1)

Water

4-Hydroxybutyl acrylate HBA

1,4-Butanediol diacrylate BDA

(7.2)

Water

r1 = k1 (CAA CBD − (1∕Keq,1 )CHBA Cwater )

(7.3)

ln(Keq,1 ) = B1 ∕T + A1

(7.4)

k1 = k0,1 exp(−EA,1 ∕RT)

(7.5)

r2 = k2 (CAA CHBA − (1∕Keq,2 )CBDA Cwater )

(7.6)

ln(Keq,2 ) = B2 ∕T + A2

(7.7)

k2 = k0,2 exp(−EA,2 ∕RT)

(7.8)

In the Eqs. (7.3) to (7.5), the variable subscript 1 refers to the 1st chemical reaction, while in the Eqs. (7.6) to (7.8), the variable subscript 2 refers to the 2nd esterification reaction. Note that r [kmol/(kgcat s)] is the reaction rate, k (kmol/(kgcat s)/(kmol/m3 )2 ) is the forward reaction rate constant, Ci , are the liquid phase molar concentrations [kmol/m3 ] of component i (where i = AA, BD, HBA, BDA, water), K eq is the equilibrium constant based on molar concentration, A and B (K) are constants in the equilibrium constant equations, k0 (kmol/(kgcat s)/(kmol/m3 )2 ) is the pre-exponential factor, EA (kJ/kmol) is the activation energy, while R (=8.314 kJ/kmol/K) is the universal gas constant. Moraru et al. (2022a,b) used the experimental data reported in the paper of Yang et al. (2008a) to re-evaluate the values of all the kinetic parameters of this model. Table 7.1 gathers all these parameters, while Figure 7.2 provides a visual comparison between the reported experimental and calculated molar fractions Table 7.1

Parameters of the kinetic model based on the paper of Yang et al. (2008a).

Reaction

E A (kJ/kmol)

A

B(K)

k 0 (kmol/(kgcat s)/(kmol/m3 )2 )

1

58,300

−2.0212

1457.6

91.5

2

86,700

−0.4614

810.36

1,81,625

Notes: EA – as reported by Yang et al. (2008a), A and B – determined by regression using experimental data, k0 – determined by regression using experimental data. Source: Adapted from Yang et al. (2008a).

7.4 Case Study: RD Coupled with a Distillation–Reactor System and Recycle

0.7

AA

100 °C

BD

Composition (molfrac)

0.6

HBA

0.5

BDA

0.4

H2O AA

0.3

BD

0.2

HBA BDA

0.1

H2O

0.0 0

200

100

300

400

500

600

700

Time (min) 0.7

AA

110 °C

BD

Composition (molfrac)

0.6

HBA

0.5

BDA

0.4

H2O AA

0.3

BD

0.2

HBA BDA

0.1

H2O

0.0 0

100

200

300

400

500

Time (min) 0.7

AA

120 °C

BD

Composition (molfrac)

0.6

HBA

0.5

BDA

0.4

H 2O AA

0.3

BD HBA

0.2

BDA

0.1

H 2O

0.0 0

100

200 Time (min)

300

400

Figure 7.2 Comparison between experiment data (markers) from Yang et al. (2008a) and calculated (lines) mole fraction in time, at various temperatures; considering the same initial AA : BD mole ratio (1.85 : 1) and catalyst concentration (1.63% mass) for all experiments. Source: Adapted from Yang et al. (2008a).

195

196

7 Process Design and Control of Reactive Distillation in Recycle Systems

of the components. A good agreement between the experimental and calculated data is observed. During the first 200 minutes after the start, the experimental data compares well with the calculated data, but it begins to deviate toward the end, indicating that other side reactions may become more important at higher temperatures and longer residence times. It is worth noting that these calculations are made for a temperature outside the range in which the pre-exponential factors were regressed. The UNIQ-HOC property model was used in Aspen Plus to calculate the properties required for rigorous process simulations. This particular method makes use of the UNIQUAC activity coefficient model to describe the liquid phase behavior, and the Hayden-O’Connell (HOC) equation of state to describe the vapor phase and account for the dimerization of carboxylic acids in vapor phase. Note that the VLE and LLE use the same sets of binary interaction parameters (BIPs) of the activity coefficient model. The chemical system consists of five components: AA, BD, HBA, BDA, and water. However, the pure-component physical properties of only three of them (AA, BD, water) are present in the Aspen Plus database. The physical properties of the other two chemicals (HBA and BDA) were estimated based on their molecular structure using the Property Constant Estimation System (PCES) available in Aspen Plus (with the exception of HBA vapor pressure). Notably, only 2 out of 10 BIP sets (water/AA and water/BD pairs) for the UNIQUAC activity coefficient model are available in the Aspen Plus database. Also, the association parameter for water/AA used in the EOS is available in the Aspen Plus databank. All the remaining (missing) BIPs were estimated using the UNIFAC group contribution method. The databanks from which the parameters are retrieved are indicated according to the naming in Aspen Plus (i.e. APV100 and NISTV100). The vapor pressure of several components (AA, BD, water) is well described in the Aspen Plus database. For BDA, there are no experimental data reported, thus its vapor pressure was estimated using methods based on the molecular structure. In case of HBA, the Antoine coefficients used in the vapor pressure equation were derived from data available in two brochures from BASF (BASF, 2016a,b). More details on the physical properties are presented by Moraru et al. (2022b). Figure 7.3 shows the vapor pressure curves for all the components involved in the system. Water and AA are the lightest components, and BDA is the heaviest component; thus, an easy vapor–liquid-based separation should be possible. However, BD and HBA have very similar boiling points (suggesting a challenging fluid separation), and their vapor pressure curves cross each other, meaning that BD is lighter at higher pressures and HBA is lighter at lower pressures. In a process that employs a fixed-bed reactor followed by a classic separation equipment (such as distillation), this behavior indicates a difficult separation between BD and HBA. This is crucial in the process, as BD is a reactant that needs to be recovered and recycled back to the reaction section, while HBA is the main product that needs to be separated and highly purified.

7.4 Case Study: RD Coupled with a Distillation–Reactor System and Recycle

1

H2O

BD

P (bar)

AA BDA

0.1

0.01

0.001 75

HBA

100

125

150

175

200

225

250

T (°C)

Figure 7.3 Vapor pressure of all components. Vapor pressure for HBA (squares) are taken from BASF brochures (BASF, 2016a, 2016b). Table 7.2 Calculated singular points (SPs) including pure components and binary azeotropes – composition (mass-based) and boiling points at 0.1 bar. SP #

Type

Temp (C)

H2 O

AA

BD

HBA

BDA

0.1 bar 1

hom

100.0

1

—

—

—

—

2

hom

141.2

—

1

—

—

—

3

het-az

158.1

—

—

0.707

—

0.293

4

hom-az

159.9

—

—

0.516

0.484

—

5

hom

163.5

—

—

—

1

—

6

hom

163.8

—

—

1

—

—

7

hom

183.3

—

—

—

—

1

The azeotrope searches, based on the Distillation Synthesis tool available in Aspen Plus, reveal that there are two azeotropes (at 0.1 bar) present in the system. Table 7.2 lists the boiling points of these azeotropes as well as of the pure components. The close-boiling components BD and HBA form a minimum-boiling homogeneous azeotrope (singular point SP #4), while BD and BDA form a heterogeneous azeotrope (singular point SP #3). Figure 7.4 (constructed using the Distillation Synthesis tool in Aspen Plus) presents the ternary liquid–liquid diagram at 40 ∘ C for the ternary mixture water– BD–BDA, which exhibits two regions of liquid–liquid immiscibility gap. The other components (AA and HBA) do not present any immiscibility with the other components.

197

198

7 Process Design and Control of Reactive Distillation in Recycle Systems

BD (227.82 C) 0.9 0.8 0.7

223.67 C

0.6 0.5 0.4 0.3 0.2 0.1 BDA (264.84 C)

99.97 C

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

H2O (100.01 C)

Figure 7.4 Liquid–liquid equilibrium diagram (mass-based) for the ternary mixture water–BD–BDA at 40 ∘ C and 1.013 bar (including the azeotropes of water–BDA and BD–BDA).

7.4.2

Process Design

HBA could be conveniently manufactured in a process based on RD. The design of such a process considers a plant capacity of 22,900 t/a 4-HBA (at 99.4% mass purity). One of our recent studies (Moraru et al., 2022a) proved that a space-time yield of 0.91 kgproduct /kgcat /h in a catalytic fixed-bed reactor placed in a classical reactor-separation-recycle (RSR) process is enough to accomplish a similar production capacity. This is equivalent to a catalyst amount of 2500 kg; thus, the results described hereafter are based on considering this amount of catalyst present in the RD column. The detailed process flow diagram is shown in Figure 7.5, while the mass balance is summarized in Table 7.3. The RD column has a typical configuration, with the reactive zone (catalytic bed) placed in the middle of the column. The light reactant is added as vapor at the bottom of the reaction zone, while the heavy reactant is added as liquid at the top of the catalytic bed. As top distillated product of the RD column, high-purity water (>99.99% mass) is obtained. The heavier components (HBA and BDA) are removed as bottoms product, which is further split in the ideal separation block SEP (as the HBA/BDA split is very easy, a conventional distillation column can be used). In order to obtain on-spec HBA, it is imperative to avoid getting reactants in bottoms of the RD column; else these contaminating low-boiling components will end up in the lighter product (HBA) during the separation of HBA from BDA. The BDA stream is mixed with part of the H2 O stream (from the top of the RD column) and fresh BD reactant, and this mixed stream is fed to the reactor, where BDA is hydrolyzed back to AA and HBA. The outlet stream from the reactor is fed to the RD column, below the reactive zone,

7.4 Case Study: RD Coupled with a Distillation–Reactor System and Recycle 1383 kW 0.2 m3

LC

H2O

PC

BD

2

TSP FC

AA

TC

X

RD

F1SP

FC

TC

L/D = 2

7

Reactor

9

8

2B

TC

TSP

42

25,000 kg catalyst 238 tubes, 6 m length 0.17 m tube diameter

SEP

BD

FC

82 kW

3

Table 7.3

X

FC

TC

1842 kW

Figure 7.5

F7/F4 (= 0.09)

28

115 kW LC

X

2,500 kg cat 0.12–0.25 bar 49–196 °C 1.2 m diameter 2.83 m3 sump volume

30 31

1

10

6 FC

11 13

2A

5

F2B/F4 (= 0.262) 4

X

BDA HBA 11

Process flow diagram of the RD process for HBA production. Mass and energy balance of the RD process for HBA production.

Stream

1

2A

2B

3

4

T/C

35

35

35

195.6 190.6 49

49

49

87

106

49

168.1

P/bar

1.2

1.2

1.2

0.25

1.2

1.2

3

2.9

1.2

0.12

Flow/kg/h

1297 557 1081 6735

0.12

5

1.2

6

7

8

9

10

11

4125

1391

696

371

5577

5577

324

2611

Mass fractions —

—

—

—

—

—

—

—

—

—

—

—

1

1

0.067

0.055 1

H2 O

—

—

—

—

—

1

AA

1

—

—

—

—

46 ppm 46 ppm 46 ppm 3 ppm 0.048 46 ppm 1 ppb

BD

—

1

1

0.002 —

HBA

0.385

BDA

0.612 1

— 3 ppm

— 3 ppm

— 3 ppm

0.194 0.740

—

0.134 —

0.006

0.288

0.994

0.476 3 ppm

—

providing the required separation of the heavier components (HBA and BDA) from the lighter reactants (AA and BD) and water by-product. The equipment sizing is based on the rigorous process simulation results. The RD column includes 2500 kg of catalyst that ensures the same space-time-yield as the tubular reactor described in a previous research study (Moraru et al., 2022a,b). The RD column has 9 rectifying stages (excluding the condenser), and 11 stripping stages (excluding the reboiler). These stages are enough to obtain high-purity water as distillate product, and a bottoms stream that contains only HBA and BDA (free of alcohol and acid). Structured packing MellapakPlusTM 252Y (with an HETP of 0.4 m) is used for the rectifying and stripping sections. Considering a structured catalytic packing, the maximum amount of solid catalyst (mcat,max ) that can be hosted per tray is estimated as follows: 𝜋 mcat,max = HETP D2packing 𝜙cat 𝜌cat,bulk (7.9) 4 Dividing the total amount of catalyst to the number of 20 reactive trays, results in an actual amount of 125 kg of catalyst per tray. Sulzer KatapakTM -SP11 is employed for the reactive stages, having the solid catalyst immobilized in wire gauze layers which are combined with layers of MellapakPlus. The catalyst (with a bulk density of 𝜌cat,bulk = 740 kg/m3 ) occupies 𝜙cat ≈ 45% of the volume of this

199

200

7 Process Design and Control of Reactive Distillation in Recycle Systems

packing (Gotze et al., 2001). The separation efficiency is about 2 NTSM (m−1 ) which is equivalent to an HETP of 0.5 m (Gotze et al., 2001). Considering the density of the catalyst and a packing diameter (Dpacking ) equal to the column diameter (1.2 m) – which was determined using the pack sizing tool in Aspen Plus – this results in a maximum allowed amount of 188 kg of catalyst per tray. Therefore, the condition of not exceeding the maximum amount of catalyst per tray is satisfied (i.e. 125 kg catalyst < max. 188 kg/tray). Note that the approach of simulating the heterogeneous catalyst in an RD column (using the RADFRAC block in Aspen Plus) is well described in the book of Luyben and Yu (2008). The reactor is a fixed-bed tubular reactor operated adiabatically. The catalyst is inside the tubes. The amount of catalyst (25,000 kg) is determined for a BDA conversion of about 35%. Note that the BDA hydrolysis back to HBA and AA is equilibrium-limited, and only slightly exothermic. A brief sensitivity analysis shows that selecting a higher conversion does not justify the increase in the catalyst amount. By selecting a tube length of 6 m with a diameter of 0.17 m, the amount of catalyst required can be accommodated in a total of 238 tubes. The sump of the RD column and the reflux accumulator are sized considering a residence time of 10 minutes. The heat exchangers (AA evaporator, column reboiler and condenser, and reactor preheater) are designed using an overall heat transfer coefficient of 930 W/(m2 K). An instantaneous model is used in the dynamic simulations, as the heat exchanger hold-up is neglected. The separation between BDA and HBA is not implemented in the design. Although a distillation column would add some material holdup, its impact on the overall dynamics is evaluated to be small. The focus here was kept on the dynamics of the RD, capturing the influence of the recycle stream. The SEP block used in this simulation can be seen as the equivalent of a conventional column. Rough calculations show that this binary mixture can be split using a column with 30 theoretical stages. The key figures for the overall process are presented in the process flow diagram (see Figure 7.5). It is also worth mentioning some key information on how the final mass (and heat) balance was achieved. The most difficult task was to close the recycle, which was made in two iteration steps. In the first iteration, a preliminary mass balance was obtained in Aspen Plus, as mentioned, without being able to close the recycle. The preliminary sizing was also made, which meant finding the number of theoretical (catalytic, stripping, and rectification) stages as well as the column diameter, sump, and main hydraulic parameters; the reflux vessel was also sized (note that when exporting the steady-state simulation to dynamics, the holdup is necessary). Then, the simulation was exported to Aspen Dynamics. After implementing the main plantwide and some basic controls (please see the process control section), the recycle was step-wise added as feed to the RD column while the simulation was running. Once the simulation achieved steady state (i.e. the values of all variables remain do not change anymore), the steady-state mass (and heat) balance was considered to have been achieved. This ended the first iteration. In a second iteration, the preliminary mass (and heat) balance and the equipment sizing were updated in the Aspen Plus simulation. For some unknown reason, the recycle still did not close.

7.4 Case Study: RD Coupled with a Distillation–Reactor System and Recycle

Then, the simulation was exported in Aspen Dynamics, and the same procedure was followed to close the recycle. The mass (and heat) balance presented in Table 7.3 is the result of the second iteration step.

7.4.3

Process Control

The main objectives of any control structure are to achieve a desired production capacity, maintain product purity at given set points and keep component inventory, all of these despite any process perturbations. These are largely treated by well-known textbooks (e.g. Luyben et al., 1999; Skogestad and Postlethwaite, 2005; Dimian et al., 2014). Additional objectives can be defined, like the economic PWC, as largely described by Skogestad (2012). Here, we consider the first three objectives, namely the production capacity, product purity, and component inventory. All the dynamic simulations have been carried out using Aspen Dynamics. The control structure is presented in Figure 7.5, together with other process data. This control structure uses only PI controllers. At the plantwide level, the production capacity is set by the flow rate controller (marked in red) of the fresh acid. This is also known as the “throughput manipulator”. Note that there are also other options to set the production capacity. For example, one can choose the flow rate controller for the fresh alcohol that feeds the column. The product purity is infered by the temperature controller on stage 28, which manipulates the steam flow rate to the reboiler. By controlling the temperature on this stage at the required value, the bottom stream of the column is free of alcohol (the water and acid are lighter and are not expected to reach out in the bottoms). Thus, only the acrylate and the diacrylate are allowed to leave the column. This is important since the distillation column (modeled here as an ideal separation block) is left to perform the much simpler task of separating the acrylate as distillate at the required purity and the diacrilate as bottom products. If the temperature controller on stage 28 fails, alcohol and even acid will escape in the bottom stream of the RD column. Since these components are lighter than the acrylate, they will find their way out of the process with the distillate of the distillation column contaminating in this way the acrylate product. Therefore, although the temperature control on the distillation column is important (for not allowing the diacrylate to escape in the distillate), the temperature control of the RD column is the key control element to achieve required product purity. Note that the main idea, independent of the implemented control, is to prevent the reactants from escaping in the bottom stream of the RD column. In addition, a concentration controller in cascade with the temperature controller is also studied for two of the process changes. The results (i.e. with and without the concentration controller) are compared. Achieving component inventory is maybe the most difficult objective of the PWC structure. To a minimum, one needs to look at both the reaction stoichiometries, recycles, and at the inlet/outlet streams of the process when deciding on the control structure that might be capable of achieving the inventory. Regarding the process presented here, one key observation is that no reactants (i.e. acid and alcohol) leave

201

202

7 Process Design and Control of Reactive Distillation in Recycle Systems

the RD column. The other key observation is that only the acrylate and water leave the plant, while the diacrylate byproduct is recovered and converted to its reactants, which are recycled back to the RD column. Roughly speaking, the question that is asked is how to feed the alcohol reactant (remember that the acid reactant is fixed, setting the plant throughput) such that no accumulation or depletion of reactants will occur. The strategy proposed here is to infer the necessity of alcohol, or lack of it, by temperature measurement in the column, and adjusting the flow rate of fresh alcohol to the column accordingly. Note that the recycle stream contains all components, including alcohol; however, since the temperature measurement is a good indication of the alcohol inventory in the column from which neither alcohol nor acid escape, the control strategy should be able to properly adjust the fresh alcohol, or the ratio between the two fresh feeds as shown in Figure 7.5, to balance the stoichiometries. The simulation results (not presented here) show that setting the two fresh feeds in ratio responds faster at throughput changes, giving lower settling times. Using the temperature profile in the column, the slope criteria (i.e. ΔT = T n+1 −T n , where T is the temperature and n is the stage number) indicate that the temperature measurement and control should be made on stage 13. This is inside the top part of the reactive-distillation section, two stages below the stage where the fresh alcohol is fed to the column. The same procedure is applied to find a suitable stage for temperature control for product purity. Another important element of the proposed strategy for balancing the stoichiometry is in relation to the second feed of fresh alcohol, which is added to the hydrolysis reactor. Overall, it seems that the presence of alcohol increases the conversion of the diacrylate back to its reactants (i.e. acid and acrylate), which is a desired outcome. This is because the alcohol reacts with the resulting acid, forming even more acrylate. Thus, this second feed of fresh alcohol is added in ratio with the flow rate of the diacrylate resulted from the RD column. Another reason for feeding alcohol to the reactor is to have a homogeneous liquid phase, since water and the diacrylate present a large heterogeneous area (see Figure 7.4). The water stream to the hydrolysis reactor is also set in ratio with the diacrylate flow rate. For the rest of the variables, standard control is implemented. In the column, the top pressure is controlled by manipulating the cooling water flow rate. The level in the reflux vessel is controlled by manipulating the water flow removed from the process. The column is set to work at a fixed reflux ratio. The evaporation rate in the acid evaporator is controlled by manipulating the steam flow rate. The temperature at the reactor’s inlet is kept constant by manipulating the cooling water flow rate in the cooler. The reactor operation is adiabatic. The PI-controllers are tuned by choosing reasonable ranges for the process variable (PV) and the controller output (OP) and then setting the controller gain to 1 [%OP range]/[%PV range]. The integral time is set equal to the estimated time constant of the process. The column temperature controllers are tuned by assuming 1 minute measurement delay, using the ATV (Auto-Tune Variation) method to find the ultimate gain and the period of oscillations at the stability limit, and using the Tyreus–Luyben rules. The parameters of the concentration controller are determined using the open-loop method and the IMC (Internal Model Control) tuning

7.4 Case Study: RD Coupled with a Distillation–Reactor System and Recycle

(c)

Purity (massfrac) Time (h)

Purity (massfrac)

Purity (massfrac) Time (h)

Flow (kmol/h)

(b)

Time (h)

Flow (kmol/h)

(a)

Flow (kmol/h)

Flow (kmol/h)

Purity (massfrac)

rules. A 30 minutes sampling interval and a 30 minutes dead time are assumed when tuning this controller. The dynamic behavior of the plant is studied for two types of changes: (i) increase and decrease of the production capacity by approximately 10%; and (ii) contamination with 3% mass water of both fresh reactants. The changes are implemented one at a time, following the same procedure: After one hour of steady state, the new value of the MV is ramped during a two-hour interval, and the simulation continues until it reaches twenty process hours. The change of throughput is made by increasing/decreasing the fresh flow rate of the acid by 10% from its nominal value. The contamination of the fresh feeds is made by (simultaneously) increasing the water concentration from 0% to 3% mass and decreasing the acid / alcohol concentration from 100% to 97% mass. In the case of fresh alcohol, both fresh feeds are contaminated. All results are presented in Figure 7.6, each graph showing the acrylate and water molar flow rates on the Y1-axis, and their mass purity on the Y2-axis. The graphs on the top show the results for a 10% increase (a) and a 10% decrease (b) in the fresh acid flowrate. As expected, both the acrylate and the water flow rates increase/decrease accordingly by 10%. In both cases, the water purity remains constant while the acrylate purity modifies slightly. If necessary, the off-set of the acrylate purity can be eliminated by using a concentration controller, as shown by the results (interrupted line) in the case of increasing throughput. Figure 7.6c graph shows the results for fresh acid contamination with 3% mass water. Since the fresh acid is the throughput manipulator and is set to a fixed mass flow rate, the acrylate flow rate decreases. The water flow rate increases since more water is fed to the process. The water purity

(d)

Time (h)

Figure 7.6 Dynamic simulation results: acrylate and water flow rates and their purities for increase (a) and decrease (b) in capacity by approximately 10%, and the contamination of fresh acid (c) and fresh alcohol (d) with 5% mass water.

203

204

7 Process Design and Control of Reactive Distillation in Recycle Systems

remains unchanged, while the acrylate purity shows an offset. Figure 7.6d graph shows the results for fresh alcohol contamination with 3% mass water. As expected, the acrylate flow rate remains constant since the fresh acid is the throughput manipulator. The water purity remains unchanged, while the acrylate purity shows an offset, which can be easily eliminated by using a concentration controller (results indicated by the interrupted line).

7.4.4

Discussion

The manufacturing of 4-hydroxybutyl acrylate at industrial scale using catalytic distillation technology is promising and feasible. An annual production capacity of 22,900 t HBA can be achieved in an RD column using 10 m of KatapakTM -SP11 packing hosting a total of 2500 kg of solid catalyst. The additional rectification section (3.6 m) and a stripping section (4.4 m) – both of them using MellapakPlus structured packing – are sufficient to obtain high-purity water as top stream and a bottom stream (BDA and HBA) that is free of acid and alcohol. The BDA formed in the RD column is conveniently converted back to AA and HBA in a fixed-bed catalytic reactor with a total of 25,000 kg of catalyst Amberlyst 15. The process based on RD overcomes the difficult separation of BD/HBA/BDA, by completely reacting BD in the RD column, and thus achieving the necessary HBA specification in a subsequent distillation column required for the HBA/BDA split. The complete conversion of BD reactant in the RD column and the hydrolysis of BDA in the additional fixed-bed reactor are the key elements in characterizing the technical feasibility of this novel process. The proposed control structure is able to achieve the most important objectives (namely, production capacity, product purity, and component inventory) when an increase/decrease in process throughput is required (for production changes), or when the fresh reactants are contaminated with significant amounts of water. Due to material recycle, obtaining the mass and heat balance for this process is challenging. Both the Aspen Plus and Aspen Dynamics simulators were used to achieve steady-state mass and heat balance. In addition, a couple of iterations were necessary since the equipment design (mainly the column diameter and hydraulic parameters) had to be updated so that the system dynamics were properly captured in face of disturbances. Further iterations may refine the design, including the addition of a distillation column for HBA/BDA separation.

7.5

Conclusions

The design and control of RD columns in systems with recycles is more complex but closer to industrial reality. The design is focused on developing the topology of the entire process, which includes RD columns, and solving the mass and energy balance based on which key process performance indicators (e.g. reactant utilization, energy efficiency, material and energy intensity, and carbon dioxide emissions) can be analyzed. Using process simulators is unavoidable during process flowsheet

References

development due to complexities that both the process and equipment design pose and their integrated and iterative nature. Closing material recycles for large processes is particularly challenging to obtain steady-state mass (and energy) balances. The process control is focused on developing a plantwide strategy to achieve the material inventory (balancing the stoichiometry), along with achieving the desired production rate and product purity. Again, process simulators are key in testing and analyzing the feasibility of the proposed control structures. In the case of RD coupled with a distillation-reactor system and recycle (HBA process), the process design of the RD column as an integral part of a larger process is particularly challenging. This is because the material recycle is not easy to close, making it difficult to achieve the material balance of the final process topology. The hydrolysis reactor for BDA conversion back to AA and HBA is another key element of the process design in achieving a highly selective process. This is because the BDA formation in the column is significant, and therefore BDA must be converted to its reactants before recycling them to the RD column. The reactor is also an integral element for achieving the reaction stoichiometry, hence contributing to maintaining the component inventory. An undersized reactor would lead to insufficient hydrolysis of BDA, hence leading to its accumulation in the process. The rigorous dynamic simulations show that the proposed control structure achieves the main task of maintaining the component inventory as well as setting the desired production rate and keeping the product purity at the required values.

References Baki, R.Ö. and Kaymak, D.B. (2014). Plantwide control of a complex process involving a reactive distillation column. Chemical Engineering Communications 201 (4): 466–481. BASF (2016a). Petrochemicals specialty monomers, Technical information, TI/CP 1331e. BASF (2016b). Specialty monomers, Technical data, E-CPI/M 1610 booklet. Chung, Y.H., Peng, T.H., Lee, H.Y. et al. (2015). Design and control of reactive distillation system for esterification of levulinic acid and n-butanol. Industrial and Engineering Chemistry Research 54: 3341–3354. Correa de Godoy, R.J. and Garcia, C. (2017). Plantwide control: a review of design techniques, benchmarks, and challenges. Industrial & Engineering Chemistry Research 56: 7877–7887. Curtis, M.A. (2008). A process for the production of hydroxyalkyl (meth)acrylates, Patent No. EP1958930A2. Dimian, A.C. and Bildea, C.S. (2008). Chemical Process Design: Computer-aided Case Studies. Germany: Wiley-VCH Verlag GmbH & Co. KgaA. Dimian, A.C., Bildea, C.S., and Kiss, A.A. (2014). Integrated Design and Simulation of Chemical Processes. Amsterdam, The Netherlands: Elsevier. Dockner, T., Lermer, H., Bittins, K., Nestler, G., Brauch, G. (1995). Preparation of 1,4-butanediol mono(meth)acrylate by esterification of (meth)acrylic acid with

205

206

7 Process Design and Control of Reactive Distillation in Recycle Systems

1,4-butanediol, in which an aqueous solution of unconverted 1,4-butanediol is obtained, Patent No. US5637760A. Douglas, J.M. (1988). Conceptual Design of Chemical Processes. NY, USA: McGraw-Hill. Gotze, L., Bailer, O., Moritz, P., and von Scala, C. (2001). Reactive distillation with KATAPAK. Catalysis Today 69: 201–208. Huang, S.G., Kuo, C.L., Hung, S.B. et al. (2004). Temperature control of heterogeneous reactive distillation. AICHE Journal 50: 2203–2216. Hung, S.B., Lee, M.J., Tang, Y.T. et al. (2006). Control of different reactive distillation configurations. AICHE Journal 52: 1423–1440. Hung, S.K., Lee, C.S., Lee, H.Y. et al. (2014). Improved design and control of triacetin reactive distillation process for the utilization of glycerol. Industrial & Engineering Chemistry Research 53: 11989–12002. Jinno H., Ishida T., Takaki H., Nakamura M. (2015). Hydroxyalkyl (meth)acrylate and method for producing same, Patent No., US20150126767A1. Kiss, A.A. (2022). Reactive distillation. In: Process Intensification: By Reactive and Membrane-assisted Separations, 2e (ed. M. Skiborowski and A. Gorak). de Gruyter ISBN 978-3-11-072045-7. Konda, N.V.S.N.M., Rangaiah, G.P., and Krishnaswamy, P.R. (2005). Plantwide control of industrial processes: an integrated framework of simulation and heuristics. Industrial & Engineering Chemistry Research 44: 8300–8313. Lee, H.Y., Chen, C.Y., Chen, J.L. et al. (2019). Design and control of diphenyl carbonate reactive distillation process with thermally coupled and heat-integrated stages configuration. Computers & Chemical Engineering 121: 130–147. Lin, Y.D., Chen, J.H., Cheng, J.K. et al. (2008). Process alternatives for methyl acetate conversion using reactive distillation. 1. Hydrolysis. Chemical Engineering Science 63: 1668–1682. Luyben, W.L. (2004). Use of dynamic simulation to converge complex process flowsheets. Chemical Engineering Education 38: 142–149. Luyben, W.L., Pszalgowski, K.M., Schaefer, M.R., and Siddons, C. (2004). Design and control of conventional and reactive distillation processes for the production of butyl acetate. Industrial & Engineering Chemistry Research 43: 8014–8025. Luyben, W.L. and Yu, C.C. (2008). Reactive Distillation Design and Control. Hoboken, New-Jersey, USA: Wiley-AIChE. Luyben, W.L., Tyréus, B.D., and Luyben, M.L. (1999). Plantwide Process Control. New York, USA: McGraw-Hill. Moraru, M.D. and Bildea, C.S. (2017). Process for n-butyl acrylate production using reactive distillation: design, control and economic evaluation. Chemical Engineering Research and Design 125: 130–145. Moraru, M.D., Zaharia, E., and Bildea, C.S. (2020). Conceptual design of novel processes for 4-hydroxybutyl acrylate production. Computer Aided Chemical Engineering 48: 1267–1272. Moraru, M.D., Kiss, A.A., and Bildea, C.S. (2022a). Design and control of novel reaction–separation–recycle processes for the production of 4-hydroxybutyl acrylate. Chemical Engineering Research and Design 177: 801–814.

References

Moraru, M.D., Patrascu, I., Kiss, A.A., and Bildea, C.S. (2022b). A new class of control structures for heterogeneous reactive distillation processes. Chemical Engineering and Processing: Process Intensification 171: 108672. Ostrowski, S., Jamróz, M.E., and Dobrowolskia, J.C. (2011). Formation of heavy adducts in esterification of acrylic acid: a DFT study. Computational and Theoretical Chemistry 974: 100–108. Pan, Q., Gao, L., Li, J. et al. (2020). Process optimization and plantwide control for producing 1,3-dioxolane from aqueous formaldehyde solution and ethylene glycol. Separation and Purification Technology 236: 116235. Pathak, A.S., Agarwal, S., Gara, V., and Kaistha, N. (2011). Design and control of a vapor-phase conventional process and reactive distillation process for cumene production. Industrial & Engineering Chemistry Research 50: 3312–3326. Patrascu, I., Bildea, C.S., and Kiss, A.A. (2022). Novel eco-efficient reactive distillation process for dimethyl carbonate production by indirect alcoholysis of urea. Frontiers of Chemical Science and Engineering 16: 316–331. Roat, S., Downs, J., Vogel, E., and Doss, J. (1986). Integration of rigorous dynamic modeling and control system synthesis for distillation columns. In: Chemical Process Control – CPC III. Amsterdam: Elsevier. Schoenmakers, H.G. and Bessling, B. (2003). Reactive and catalytic distillation from an industrial perspective. Chemical Engineering and Processing 42: 145–155. Shah, M., Kiss, A.A., Zondervan, E., and de Haan, A.B. (2012). A systematic framework for the feasibility and technical evaluation of reactive distillation processes. Chemical Engineering and Processing: Process Intensification 60: 55–64. Sharma, N. and Singh, K. (2010). Control of reactive distillation column – a review. International Journal of Chemical Reactor Engineering 8: R5. Skogestad, S. and Postlethwaite, I. (2005). Multivariable Feedback Control: Analysis and Design, 2e. Chichester, UK: Wiley. Skogestad, S. (2012). Economic plantwide control. In: Plantwide Control: Recent Developments and Applications (ed. V. Kariwala and V.P. Rangaiah). Chichester, UK: Wiley. Sugiura T., Okamura K., Ito K., Kawakami N., Matsumoto S., Matsuno M. (2013). Process for preparing 4-hydroxybutyl acrylate, Patent No. WO 2013/157597, EP2860170A4. Tanaka Y., Okamura K., Ito K. (2017). Process for preparing 4-hydroxybutyl acrylate, Patent No. US009670129B2. Tokuda M., Kouno M., Morioka J. (2009). Method for producing hydroxyalkyl (meth)acrylate, Patent No. US8461374B2. Wang, S.J., Huang, H.P., and Yu, C.C. (2011). Design and control of a heat-integrated reactive distillation process to produce methanol and n-butyl acetate. Industrial & Engineering Chemistry Research 50: 1321–1329. Wang, S.J., Yu, C.C., and Huang, H.P. (2010). Plantwide design and control of DMC synthesis process via reactive distillation and thermally coupled extractive distillation. Computers and Chemical Engineering 34: 361–373.

207

208

7 Process Design and Control of Reactive Distillation in Recycle Systems

Xu, H., Ye, Q., Zhang, H. et al. (2014). Design and control of reactive distillation-recovery distillation flowsheet with a decanter for synthesis of N-propyl propionate. Chemical Engineering and Processing 85: 38–47. Yang, J.I., Cho, S.H., Kim, H.J. et al. (2008a). Production of 4-hydroxybutyl acrylate and its reaction kinetics over Amberlyst 15 catalyst. Canadian Journal of Chemical Engineering 85: 83–91. Yang, J.I., Cho, S.H., Park, J., and Lee, K.Y. (2008b). Esterification of acrylic acid with 1,4-butanediol in a batch distillation column reactor over Amberlyst 15 catalyst. Canadian Journal of Chemical Engineering 85: 883–888. Zeng, K.L., Kuo, C.L., and Chien, I.L. (2006). Design and control of butyl acrylate reactive distillation column system. Chemical Engineering Science 61: 4417–4431.

209

8 Dynamics and Control of Middle-vessel Batch Distillation with Vapor Recompression Radhika Gandu 1 , Akash Burolia 1 , Dipesh Shikchand Patle 2 , and Gara Uday Bhaskar Babu 1 1

Department of Chemical Engineering, National Institute of Technology, Warangal, 506004, Telangana, India Department of Chemical Engineering, Motilal Nehru National Institute of Technology Allahabad, Prayagraj, 211004, Uttar Pradesh, India 2

8.1 Introduction A batch distillation column is a widely used operation in small-scale industries, wherein multi-component azeotropic and zeotropic liquid mixtures are separated in a single batch. This operation is commonly used when the priority is to achieve very high product purity from valuable and expensive resources. Batch distillation’s flexibility makes it preferable to continuous distillation, as it can handle various feed compositions by changing operational parameters. Greenhouse gas (i.e. carbon dioxide (CO2 )) emissions from fossil fuel combustion are one of the numerous severe environmental problems of the current century. The batch distillation column separates multi-component liquid mixtures in a single batch operation. Also, multi-component batch distillation efficiently handles various feed fractions, difficult separations, easy separations, and various product specifications. This drives batch distillation suitable where the products’ demand and lifetime can vary immensely with time and be tentative, such as fine, pharmaceutical, and specialty chemicals. Although it has some advantages over continuous distillation, batch distillation has several inherent disadvantages: long batch time, increased energy usage, a high temperature in the reboiler, and a complex operation (Demicoli and Stichlmair, 2004). However, batch distillation has acquired revived interest because of the flexibility offered. To solve these issues, middle-vessel batch distillation (MVBD) column configurations have received intense attention in the industry and academia (Skouras and Skogestad, 2004; Gruetzmann et al., 2006). The MVBD strategy combines the batch rectifier and batch stripper in one column. For separating ternary mixtures, it is clear that the MVBD operation is much easier than the regular column operation (Warter et al., 2002) due to a more straightforward process to handle liquid fractions, i.e. distillate slop-cuts are not required compared to those in regular batch distillation operation. Hence, there is no production phase when separating ternary mixtures in an MVBD operation, lowering the batch Control and Safety Analysis of Intensified Chemical Processes, First Edition. Edited by Dipesh Shikchand Patle and Gade Pandu Rangaiah. © 2024 WILEY-VCH GmbH. Published 2024 by WILEY-VCH GmbH.

210

8 Dynamics and Control of Middle-vessel Batch Distillation with Vapor Recompression

operation period. But decrease in the middle component purity can be observed in the middle vessel. However, the product purity in a middle vessel of MVBD can be increased by withdrawing the top product. Thus, in this chapter, we considered the open-loop operation of MVBD, i.e. it involves product withdrawal, to show the benefits over closed-loop mode MVBD. Although there are extensive publications on regular batch distillation, only a few works have been published on the middle-vessel column for separating ternary mixtures. Therefore, the current chapter chose the MVBD operation in open-loop mode as a forthcoming candidate for additional improvement. Improving the performance of an MVBD, cost savings, and reduction of CO2 emissions is a significant challenge to satisfy the environmental marks as consented to in the Kyoto Protocol. Therefore, to further enhance the performance of open-loop MVBD, we explore the feasibility of heat integration in MVBD. Numerous heat integration schemes have recently been formulated to improve performance, especially continuous distillation. Energy may be integrated internally and externally for distillation operations. The internally heat-integrated distillation column (HIDiC) for continuous distillation (Naito et al., 2000; Gadalla et al., 2005; Suphanit, 2010; Shenvi et al., 2011) is the most familiar benchmark of internal heat integration. Conversely, the vapor recompression column (VRC) is established through external heat integration. Only a few papers in the literature deal with the external heat integration of batch distillation, i.e. VRC for separation of close-boiling point mixtures. The VRC is more feasible for separating close-boiling components in conventional and unconventional batch distillations due to the low-temperature difference of the column. As a result, it requires only low compression, which gives improved energy and TAC (total annualized cost) savings. The VRC approach can improve the performance of batch distillation for binary, multi-component mixtures, and reactive separations. A novel external energy-integrated VRC method was proposed for a conventional reactive batch column with close-boiling mixtures (Johri et al., 2011). Later, the VRC approach was explored for separating ternary closed-boiling components in unconventional batch columns, such as the middle-vessel batch column without vapor bypass and a side withdrawal (Babu et al., 2012a,b). Composition control is a difficult task in unsteady-state batch distillation. Many researchers are working on their concepts to control the composition with conventional methods and have developed control strategies for a quicker and smoother response. Reflux flow rate is generally chosen as a manipulated variable in batch distillation to maintain constant composition throughout the production operation. It must continuously increase to maintain constant distillate product purity in batch distillation (Monroy-Loperena and Alvarez-Ramirez, 2000). The performance of the column depends on the controller tuning and the type of feed mixture, i.e. azeotropic and zeotropic (Warter et al., 2004). The separation of multi-component zeotropic mixtures is easier in batch distillation. Still, it has two significant problems: high temperature difference and increased relative volatility,

8.2 Conventional Middle-vessel Batch Distillation

leading to high compressor power or compression ratio (CR). A close volatility system produces a sluggish response, making it easy to control at the desired purity, while a system with wide relative volatility gives a speedy response and a significant offset from the specified purity. Therefore, controlling systems with wide relative volatilities is challenging. Most works on VRC with constant composition control in the literature were focused on separating close-boiling mixtures (Babu and Jana, 2013; Vibhute and Jogwar, 2020; Parhi et al., 2020; Babu and Jana, 2014). Few papers have appeared to deal with VRC in conventional batch distillation to separate multicomponent zeotropic wide-boiling mixtures (Nair et al., 2017; Babu and Jana, 2013; Khan et al., 2012). Moreover, work has not been reported on open-loop MVBD for separating ternary zeotropic mixture, i.e. methanol/ethanol/1-propanol. This chapter aims to investigate the feasibility of VRC for separation of ternary zeotropic systems. Thus, we have chosen a ternary zeotropic mixture, i.e. methanol/ ethanol/1-propanol for the present study. Pure primary alcohols are very valuable as raw materials and solvents. Alcohol mixtures are produced as byproducts from the different synthesis processes. The pharmaceutical plant can produce several final products, which generate different waste solvent mixtures. The commercial importance of various alcohols differs significantly. The most important alcohols in the industry are methanol, ethanol, 1-propanol, 1-butanol, etc. This study considers a typical separation process wherein a mixture of three aliphatic alcohols, namely methanol, ethanol, and propanol, is required to be separated using batch distillation. This system does not exhibit azeotropic behavior or medium-difficulty separation. To the author’s knowledge, this is the first study on vapor recompression in an openloop MVBD and also constant composition control for separating a ternary zeotropic wide-boiling mixture.

8.2 Conventional Middle-vessel Batch Distillation Figure 8.1 depicts the schematic illustration of the conventional middle-vessel batch distillation (CMVBD) process without vapor bypass in closed-loop mode (thick line – startup phase; no product withdrawal) and open-loop mode (dotted lines – production phase; product withdrawn from top). It consists of three vessels: reboiler, middle vessel, and reflux drum. A middle vessel separates the CMVBD column into rectifying and stripping sections. This process simultaneously obtains light, intermediate, and heavy components from the reflux drum, middle vessel, and reboiler. The CMVBD operation consists of two operations: the startup phase and production phase. Startup phase: After the charge is loaded into the vessels, the column is operated at total reflux, i.e. no liquid is withdrawn. The light (Product 1) and heavy boilers (Product 3) will accumulate in the top and bottom vessels, respectively. During this phase, the concentration of the intermediate-boiling component in the middle vessel (Product 2) is increasing until it reaches its specification. Production phase: When the concentration of the intermediate-boiling component in the middle vessel reaches its specification, the lightest product can

211

212

8 Dynamics and Control of Middle-vessel Batch Distillation with Vapor Recompression

Condenser

VNT Reflux drum

NT R

D

MV+1 Product 1 VMV

LMV+1 Middle vessel (MV)

VMV–1 Vapor

1

Product 2

LMV Liquid

Tray

VB

L1

Product 3 Reboiler

Stream

Figure 8.1 Middle-vessel batch distillation without vapor bypass: ——— Startup phase (closed-loop); ……. . Production phase (open-loop).

be withdrawn from the top reflux drum with a constant distillate flow rate (D kmol/min). Both the middle component purity in the middle vessel and the heaviest component purity in the reboiler increase, while the lightest component purity in the reflux drum decreases during this phase.

8.2.1

A Systematic Simulation Approach of CMVBD

CMVBD column dynamics consists of combination of algebraic-differential equations based on certain assumptions. The model was devised based on the following assumptions: the liquid stream and vapor stream are thoroughly intermixed on individual trays; the molar vapor holdup is negligible compared to the molar liquid holdup; the delay time between trays is overlooked; it operates at

8.2 Conventional Middle-vessel Batch Distillation

atmospheric pressure; for liquid flow rate estimation, the nonlinear Francis–Weir formula is applicable; and vapor–liquid compositions are estimated by NRTL thermodynamic model. The model equations of CMVBD, i.e. the total continuity, component continuity, and energy balance equations, are summarized as follows. 8.2.1.1 Model Equations

Reboiler: Total continuity: Ṁ B = L1 − VB Component continuity: Ṁ B ẋ B, j = L1 x1, j − VB yB, j L Energy Equation: Ṁ B Ḣ B = QR + L1 H1 L − VB HB V Bottom Tray: Total continuity: Ṁ 1 = L2 + VB − L1 − V1 Component continuity: Ṁ 1 ẋ 1, j = L2 x2, j + VB yB, j − L1 x1, j − V1 y1, j L Energy Equation: Ṁ 1 Ḣ 1 = L2 H2 L + VB HB V − L1 H1 L − V1 H1 V Intermediate Trays: Total continuity: Ṁ N = LN+1 + VN−1 − LN − VN Component continuity: Ṁ N ẋ N, j = LN+1 xN+1, j + VN−1 yN−1, j − LN xN, j − VN yN, j L Energy Equation: Ṁ N Ḣ N = LN+1 HN+1 L + VN−1 HN−1 V − LN HN L − VN HN V Top Tray: Total continuity: Ṁ NT = R + VNT−1 − LNT − VNT Component continuity: Ṁ NT ẋ NT, j = R xD, j + VNT−1 yNT−1, j − LNT xNT, j − VNT yNT, j L Energy Equation: Ṁ NT Ḣ NT = L2 H2 L + VNT−1 HNT−1 V − LNT HNT L − VNT HNT V Condenser and Reflux drum: Total continuity: Ṁ D = VNT − R − D Component continuity: Ṁ D ẋ D, j = VNT yNT, j − (R + D)xD, j Middle vessel: Total continuity: Ṁ MV = LMV+1 − LMV − VMV + VMV−1 Component continuity: Ṁ MV ẋ MV, j = LMV+1 xMV+1, j − LMV xMV, j − VMV yMV, j + VMV−1 yMV−1, j L Energy Equation: Ṁ MV Ḣ MV = LMV+1 HMV+1 L − LMV HMV L − VMV HMV V + VMV−1 HMV−1 V ∑ Net feed rate: Ṁ MV = −(D) − (Ṁ N ) + Ṁ D + Ṁ B In the above modeling equations, L – liquid flow rate, V – vapor flow rate, M – molar holdup, R – reflux flow rate, D – distillate flow rate, QR – reboiler heat input, Ṁ – rate of change of holdup, ẋ – rate of change of liquid, Ḣ – rate of change of enthalpy, y – vapor phase composition, Subscripts B, 1, N, NT, MV, D, and j are reboiler, bottom tray 1, intermediate trays, top tray, middle vessel, reflux drum, and components, respectively. Superscripts L and V represent liquid and vapor enthalpies, respectively. Step 1: Define the column specifications (number of trays, column diameter, column length, weir height, weir length, tray efficiency), feed specifications (number of components, composition), initialize variables (hold up in the reboiler, holdup in

213

214

8 Dynamics and Control of Middle-vessel Batch Distillation with Vapor Recompression

the reflux drum, trays holdup, liquid phase initial compositions, trays initial liquid holdup, enthalpy constants, Antoine constants, reboiler heat input, distillate flow rate, pressure, etc.). Step 2: The feed mixture is introduced into the reboiler, middle vessel, trays, and reflux drum (about 97 % in reboiler and remaining in middle vessel, trays and reflux drum). Step 3: Constant heat input is provided to the reboiler. Step 4: Reboiler computations: Step 4.1. Calculations of bubble point: At bubble point temperature, summation of vapor phase molar fraction (SUMY) should be 1. The method used to find the bubble point temperature is as follows: Initial temperature (T) is assumed for i = 1 : 500 (iterations)

{ ( A( j)−

B(j)

)}

T+C(j) Calculate vapor pressure using Antoine equation Pv ( j) = 10 Calculate vapor phase compositions of components using Yeq ( j) = (Pv ( j) xB ( j) 𝛾(j))/P) ∑ Calculate the summation of vapor phase compositions, i.e. F(T) = Yeq ( j) If (F(T) − 1 < 0.000001) break T = T B (Record the T B i.e. the bubble point temperature) else (Solve iteratively using Newton–Raphson method) Derivative of F(T) is calculated by (Yeq ( j) * B( j) * 2.303)/(T + C( j)) * (T + C( j)) end New estimate of temperature is calculated by the following equation:

Tnew = T − (F(T)∕F ′ (T)) T = Tnew ′

(Function F(T) and F (T) are calculated based on new temperature, and this process is repeated till F(T) − 1 < 0.000001) end Finally, compute the actual vapor phase composition using tray efficiency. Here A, B, and C are Antione constants. Step 4.2. Enthalpy calculations: Calculate the liquid phase enthalpy (Hliq) and vapor phase enthalpy (Hvap) by the following equations: (

) ) BLV( j) ( 2 2 ∗ TB − TR Hvap( j) = (ALV( j) ∗ (TB − TR )) + 2 ) ( ) CLV( j) ( 3 ∗ TB − TR 3 + 3 ( )) ( ) ) ELV( j) ( 5 DLV( j) ( 4 ∗ TB − TR 4 + ∗ TB − TR 5 + 4 5

8.2 Conventional Middle-vessel Batch Distillation

Hvap =

∑

Hvap( j) ∗ yB ( j)

R ∗ TB ∗ TB ∗ B( j) ∗ 20.303 (C( j) + TB )2 ∑ Hliq = 𝜆(j) ∗ xB (j) 𝜆( j) =

Here, T R is reference temperature, T B is bubble point temperature, 𝜆 is latent heat, and ALV, BLV, CLV, DLV, ELV are enthalpy constants. Step 4.3. Calculations of average molecular weight (MW avg ) and density (DENSavg ). Initialize molecular weight (MW) and density (DENS) of components Calculate the average density using liquid phase composition from bubble point temperature calculation using the following equation: ∑ xB (j) ∗ MW(j) MWavg = Calculate the average molecular weight using liquid phase composition from bubble point temperature calculation using the following equation: ∑ xB (j) ∗ DENS(j) DENSavg = Step 4.4. Compute vapor flow rates from energy balance equations: The dynamics of internal energies are faster than composition. Therefore, setting zero the rate of change of enthalpy with time (left-hand side of energy balance equation) in energy balance equation transforms it into an algebraic equation to calculate vapor flow rate, i.e. V B = (QR + L1 H 1 L )/H B V Step 4.5. Compute the holdup (M B ) using total continuity equation (solve using numerical technique, i.e. Euler method). Step 4.6. Compute the liquid phase composition (xB ) by solving component equation (solve using numerical technique, i.e. Euler method). Step 5: The vapor departs through the bottom tray 1. Compute the liquid and vapor phase compositions on tray 1 using the equations stated in Step 4.1 to Step 4.6. In addition to bubble point, enthalpy, vapor flow rate, and liquid phase compositions, trays involve internal liquid flow rate calculations and are calculated using Francis–Weir formula given below. H1 ∗ MWavg LW1 = Atray ∗ DENSavg ) ( L1 = 1.84 ∗ WL ∗ DENSavg ∗ (LW1 − WH)1.5 L1 is liquid flow rate from tray 1, WH is weir height; H 1 is Tray 1 holdup; Atray is Tray area Step 6: The vapor departs to the intermediate trays and into the condenser from Tray 1. The vapor condenses in the total reflux condenser and fills the reflux drum. (Similarly to Tray 1, calculate the liquid phase composition, vapor phase composition, and internal liquid flow rates on intermediate trays, top tray, and reflux drum using the equations stated in Steps 4.1 to 4.6 and Step 5). Step 7: Total reflux operation starts, meaning the total condensed liquid is returned to the batch column’s top tray.

215

216

8 Dynamics and Control of Middle-vessel Batch Distillation with Vapor Recompression

Step 8: The liquid drops from tray to tray and reaches the reboiler. Step 9: The above steps repeat till the composition of the middle component (ethanol) in the middle vessel reaches maximum and the composition does not vary with time, then the batch operation is stated to have achieved a steady state. Step 10: Withdraw the lightest product (1-methanol) from the reflux drum with a fixed flow rate after the steady state. Once it reaches the specified purity, terminate the simulation. Then, collect the bottom product from the reboiler, middle product from middle vessel, and top product from the reflux drum.

8.2.2

Constant Composition Control

In the production phase operation of batch distillation, the lightest component is withdrawn with a fixed distillate flow rate from the reflux drum. As a result, the lightest composition, i.e. methanol composition, decreases with time. The control objective is to maintain the lightest component in the distillate, i.e. methanol, at a specified purity or setpoint. The set point is the steady-state value of the distillate’s lightest component at the startup phase’s end. In batch distillation, a reflux flow rate/reflux ratio/distillate flow rate is manipulated to maintain a constant distillate composition throughout the batch production operation. In this chapter, to maintain the constant distillate composition by manipulating reflux rate, two controllers, i.e. PI (proportional controller) and nonlinear GSPI (Gain Scheduling Proportional Integral Controller), have been designed (Adari and Jana, 2008; Babu et al., 2012a,b).

8.3 Single-stage Vapor Recompression in Middle-vessel Batch Distillation The schematic diagram of a vapor recompression with a single-stage compression in CMVBD is shown in Figure 8.2. The primary goal of single-stage vapor recompression in middle-vessel batch distillation (SiVRMVBD) is to utilize the heat from top vapor, resulting in reduced utility consumption, i.e. steam and cooling water. SiVRMVBD comprises an external single compressor as opposed to CMVBD. Therefore, SiVRMVBD simulation approach consists of additional steps compared to CMVBD, i.e. manipulation scenarios.

8.3.1

A Systematic Simulation Approach of SiVRMVBD

Step 1: The column specifications, feed specifications, tray efficiency, number of trays, liquid compositions, liquid holdup on all trays, etc., are defined. Step 2: The feed mixture is introduced into the reboiler, middle vessel, trays, and reflux drum. Step 3: Constant heat input is provided to the reboiler, and the vapor rises to the top tray.

8.3 Single-stage Vapor Recompression in Middle-vessel Batch Distillation

Condenser

VNT Reflux drum

NT R MV+1

D Product 1

VMV

LMV+1 Pin, Tin

Middle vessel (MV) VMV–1

Product 2

LMV

Vapor

Throttling valve

Compressor

Liquid Pout, Tout

1

Tray

VB

Product 3 L1 Reboiler

Stream

Figure 8.2 Single-stage vapor recompression in middle vessel batch distillation without vapor bypass.

Step 4: The vapor from the top tray (V NT ) is fed to the compressor. Here, the vapor is compressed to the required temperature (T out ). The latent heat of compressed vapor (𝜆), which is utilized as reboiler heat input, is thus condensed in the reboiler–condenser. Two manipulation criteria are employed to maintain: (i) temperature difference between compressor outlet required temperature (T out ) and reboiler temperature (T B ), i.e. ΔT T = T out − T B of at least 15∘ C for total condensation in reboiler. (ii) Same reboiler heat duty for CMVBD and SiVRMVBD: An overhead VNT manipulation was employed to maintain the same reboiler duty for fair comparison between CMVBD and SiVRMVBD.

217

218

8 Dynamics and Control of Middle-vessel Batch Distillation with Vapor Recompression

• If the energy released by the compressed vapor (QCV ) exceeds the required reboiler heat duty, the overhead vapor can be split into two parts: Part (V Comp ) of total vapor fed to the compressor to increase the temperature and rest (V Cond ) to the condenser. • If the latent heat (QCV ) from compressed vapor is less than the needed reboiler heat input, heat (QEA ) must be provided to the reboiler from external source. Step 5: The energy released by the compressed vapor (QCV ) is calculated from the following equation. QCV = VNT × 𝜆 Step 6: The compressed vapor is condensed in reboiler–condenser, which is at high pressure. The high-pressure liquid from reboiler–condenser is depressurized to the top tray pressure using a throttle valve and reaches reflux drum.

8.4 Performance Specifications 8.4.1

Energy Savings

As stated, batch distillation operations are highly energy-intensive processes. Therefore, energy consumption is one of the key performance indices to evaluate the thermal efficiency of VRCs in batch distillation. The energy consumption of SiVRMVBD (Qcons, SiVRMVBD ) is calculated by the following formula (Babu et al., 2012a,b). Qcons,SiVRMVBD = QR + 3Qcomp Here, QR is total reboiler heat duty and Qcomp is compressor duty. The energy consumption of CMVBD (Qconc, CMVBD ) is obtained from: Qconc,CMVBD = Batch time × QR Then, the corresponding energy savings are calculated from following equation: Qcons,CMVBD − Qcons,SiVRMVBD × 100 % Energy savings = Qcons,CMVBD

8.4.2

Total Annual Cost

The economic performance specification chosen in this chapter is total annual cost (TAC) with a payback period of 3 years. It is calculated using the equation given below: ( ) ( ) $ TCI TAC = TOC + year 3 TOC is total annual operating cost of respective columns, and TCI is total capital investment. The total capital investment consists of reboiler, total condenser, column shell, trays, and compressor. The steam, electricity, and cooling water cost values were adopted from Babu et al. (2012a,b). Marshell and Swift (M&S) index of 1704.9 was used in this chapter. The TCI and TOC were calculated using correlations from (Douglas 1988).

8.5 Results and Discussion

8.4.3

Greenhouse Gas Emissions

Batch distillation with a middle vessel is an energy-intensive operation and leads to significant greenhouse gas (CO2 ) emissions into the atmosphere. Therefore, reduction of CO2 emissions is must as per the Kyoto Protocol. A mathematical equation for calculating CO2 emissions from heat-integrated distillation columns was proposed by Gadalla et al. (2005). In this chapter, we used the same model. The SiVRMVBD scheme requires medium-pressure steam and electricity for the reboiler and compressor. The CMVBD needs medium-pressure steam to vaporize the reboiler liquid. This chapter used natural gas as a fuel to investigate CO2 emissions.

8.5 Results and Discussion In this study, ternary zeotropic wide-boiling mixture was investigated: methanol/ ethanol/1-propanol. The computer program was developed and simulated in MATLAB.

8.5.1

Conventional Middle-vessel Batch Distillation Column

CMVBD column process separates a moderately wide-boiling/zeotropic mixture of methanol, ethanol, and 1-propanol. Among these, methanol is the lightest component, ethanol is the intermediate component, and 1-propanol is the heaviest component. The liquid feed mixture at its bubble point is charged into the reboiler (97.5% total feed), condenser/reflux drum (0.65% of total feed), and on 49 trays (1.857% of total feed). The total number of trays, numbered from bottom to top, is 49, excluding the reboiler and condenser. A constant reboiler heat input of 4000 kJ/min is supplied throughout the operation. First, CMVBD operates in a startup phase mode and product is withdrawn from reflux drum in the production phase. The design and operating specifications are given in Table 8.1. 8.5.1.1 Dynamic Composition Profiles

The dynamic composition profiles of the startup phase and production phase of CMVBD are presented in this section. Figure 8.3 (Reflux drum), Figure 8.4 (Middle vessel), and Figure 8.5 (Reboiler) show the dynamic composition profiles at total reflux condition (no product is withdrawn). Figure 8.3 shows the composition profiles in the reflux drum at startup phase. The lightest component, methanol, attains the maximum purity of 0.999, while remaining components at their lowest purities. In the middle vessel, the middle component, ethanol, has a purity of 0.966 at the end of the startup phase, as shown in Figure 8.4. The heaviest component, 1-propanol, accumulates in the reboiler at the highest composition of 0.999, as shown in Figure 8.5. Steady-state results show the maximum achievable ethanol purity in the middle vessel, methanol purity in the reflux drum, and 1-propanol purity in the reboiler at 160.06 min. It is clear from the figures that the compositions remain unchanged at 160.06 min, indicating that the process reaches a steady state.

219

8 Dynamics and Control of Middle-vessel Batch Distillation with Vapor Recompression

Table 8.1

Operating conditions and column specifications.

System

Methanol/ethanol/1-propanol

Feed composition

0.15/0.7/0.15

Reboiler heat duty (kJ/min)

4000

Tray efficiency (%)

80

Number of trays

49

Total feed (kmol)

5.385

Initial holdup on trays (kmol)

0.098

Initial condenser holdup (kmol)

0.035

Initial middle vessel holdup (kmol)

0.25

Initial reboiler holdup (kmol)

5.25

1.0 Methanol 0.8 xD (mole fraction)

220

Ethanol 1-propanol

0.6

0.4

0.2

0.0 0

40

80

120

160

Time (min)

Figure 8.3

Distillate composition profile startup phase.

After a steady state, the top product, i.e. methanol from the reflux drum, is withdrawn at a fixed distillate flow rate of 0.01 kmol/min. The operation is terminated once it reaches the specified product purity (∼0.997). The production phase is 28.66 minutes. Therefore, the entire batch distillation process takes 188.72 minutes (startup phase time plus production phase time). It is observed during production phase that ethanol mole fraction in the middle vessel increases (Figure 8.6). The compositions at the end of steady-state and batch operation are shown in Table 8.2.

8.5 Results and Discussion

1.0

0.8 xMV (mole fraction)

Methanol Ethanol 0.6

1-propanol

0.4

0.2

0.0 0

40

80

120

160

Time (min)

Figure 8.4

Middle-vessel composition profile: startup phase.

1.0

xB (mole fraction)

0.8 Methanol Ethanol 1-propanol

0.6

0.4

0.2

0.0 0

40

80

120

160

Time (min)

Figure 8.5

Reboiler composition profile: startup phase.

It is evident from the results that the withdrawal of the top product increases the middle component’s purity when compared to CMVBD operating in closed-loop mode, i.e. total reflux condition. However, the top product composition decreases from 0.999 to 0.997.

221

8 Dynamics and Control of Middle-vessel Batch Distillation with Vapor Recompression

xD, Methanol (mol fraction)

0.9992 0.9988 0.9984 0.9980 0.9976

(a)

0.974 xMV, Ethanol (mol fraction)

222

0.972 0.970 0.968 0.966

(b) 160

165

170

175

180

185

190

Time (min)

Figure 8.6 (a) Lightest component (methanol) profile in reflux drum, (b) middle-boiling component (ethanol) profile in middle vessel: production phase.

Table 8.2

Compositions at startup phase and end of the batch operation. Reflux drum

Middle vessel

Reboiler

Component Startup phase At the end Startup phase At the end Startup phase At the end

Methanol Ethanol

0.999

0.997

1.00 × 10 −12

1-Propanol 4 × 10

−3

0.0023 2 × 10

−11

0.0299

0.025

1.2 × 10−11

4 × 10−13

0.966

0.973

0.0001

0.0001

0.0041

0.002

0.9999

0.9999

8.5.2 Single-stage Vapor Recompression in Middle-vessel Batch Distillation The primary goal of SiVRMVBD is to utilize the heat from top vapor, resulting in reduced utility consumption i.e. steam and cooling water. The SiVRMVBD is an unsteady state operation wherein the operational parameters, i.e. internal energy QCV and ΔT T , change with time. The objective of the chapter is to maintain the same dynamical performance (same reboiler duty of 4000 kJ/min) between CMVBD and SiVRMVBD for fair comparison and also ΔT T = T out − T B of at least 15∘ C for total condensation in reboiler–condenser. Iterative manipulation scenarios explained in Section 8.3 are used to attain the objectives. Now, to operate the SiVRMVBD at a fixed reboiler duty of 4000 kJ/min, the simulation approach, as explained earlier (Section 8.3), needs to be applied. We produced the manipulated variables profiles for the SiVRMVBD from the simulation approach.

8.5 Results and Discussion

4500 4000 3500 QR

Energy (kJ/min)

3000

QCV

2500

QEA

2000 1500 1000 500 0 0

50

100

150

200

Time (min)

Figure 8.7

Energy profile throughout the batch operation.

Figure 8.7 illustrates the adjustment of external energy from the source. It is observed that throughout the operation, the energy released by the compressed vapor is less than the required reboiler heat duty; hence external energy is necessary to make up 4000 kJ/min. However, in the production phase, the energy released by the compressed vapor is greater than the required reboiler heat duty; hence external energy is not required. Figure 8.8 demonstrates the vapor flow rate manipulation throughout the operation. It is clear that in steady state operation, the energy released by the compressed vapor is less than the required reboiler heat duty; hence vapor flow rate manipulation is unnecessary, i.e. condenser is not required in this period. However, in the production phase, the energy released by the compressed vapor is greater than the required reboiler heat duty; hence vapor flow rate manipulation is needed, i.e. condenser is required to condense some part of vapor. In this way, SiVRMVBD saves operating costs: energy from source (steam) and cooling water. Figure 8.9 shows the temperature difference of the column throughout the operation and it is observed that it increases continuously with time. It indicates that CR must increases continuously with time. Vapor recompression technique’s feasibility depends on the temperature difference of the column. At very high temperature difference in the column, compressor should operate at high CR and VRC may not give savings. Figure 8.10 shows CR manipulation to maintain at least 15 ∘ C. The temperature difference in the reboiler–condenser is maintained exactly at 15 ∘ C for complete condensation, as shown in Figure 8.11, by manipulating variable speed drive compressor.

223

8 Dynamics and Control of Middle-vessel Batch Distillation with Vapor Recompression

0.12

Vapor flow rate (Kmol/min)

0.10 0.08

VComp VCond

0.06

VNT

0.04 0.02 0.00 0

50

100

150

200

Time (min)

Figure 8.8

Vapor flow rate manipulation profile throughout the batch operation.

40 35 30 25 TB–TNT (0K)

224

20 15 10 5 0 –5

0

50

100

150

200

Time (min)

Figure 8.9

Temperature difference of the column throughout the batch operation.

8.5 Results and Discussion

2.2

CR

2.0

1.8

1.6

1.4 0

50

100

150

200

Time (min)

Figure 8.10

Compression ratio (CR) profile throughout the SiVRMVBD operation.

17

TOut–TB(0K)

16

15

14

13 0

50

100

150

200

Time (min)

Figure 8.11

ΔT T profile throughout the SiVRMVBD operation.

8.5.3 Energetic, Economic, and Environmental Performance: CMVBD vs. SiVRMVBD In this section, to show the benefits of the vapor recompression approach, the SiVRMVBD column is quantitatively compared with the CMVBD. It is clear from simulation results (Table 8.3) that the SiVRMVBD secures a 78.78% energy savings.

225

226

8 Dynamics and Control of Middle-vessel Batch Distillation with Vapor Recompression

Table 8.3

Energetic performance comparison. Energy consumption (kJ)

Column

CMVBD

SiVRMVBD

MVBD

7.54884 × 105

1.6013 × 105

Table 8.4

Economic comparative analysis.

Item

CMVBD

SiVRMVBD

Capital cost 2.2185 × 105

2.2185 × 105

Column tray

4

2.3070 × 10

2.3070 × 104

Reboiler

5.4386 × 104

5.4386 × 104

Column shell

3

Condenser

9.4298 × 10

424.834

Compressor

—

4.4142 × 104

3.0873 × 10

Total

5

3.4387 × 105

Operating cost 1.8023 × 104

Steam

3

411.6604

Cooling water

1.3947 × 10

11.8377

Electricity

—

3.1409 × 103

Total

1.9418 × 104

3.5644 × 103

TAC

1.2233 × 105

1.1819 × 105

%, TAC savings

3.384

Table 8.4 compares the economic performance between CMVBD and SiVRMVBD. The results clearly show that the 3.384% TAC savings (with a payback period of 3 years) is obtained with SiVRMVBD compared to that with CMVBD. Furthermore, the SiVRMVBD system reduces process costs by 81.6% compared to CMVBD. Finally, Table 8.5 compares the environmental performance of CMVBD and SiVRMVBD. It is evident from Table 8.5 that SiVRMVBD reduces CO2 emissions by 93.02% globally compared to CMVBD. Therefore, it is inferred that SiVRMVBD is the most promising technique pertaining to energy efficiency, cost-effectiveness, and minimal CO2 emissions, compared to CMVBD.

8.5.4

Constant Composition Control

As stated in section 8.2, withdrawing the lightest product from the reflux drum, increases the purity of the middle component in the middle vessel. However, the purity of top lightest product (i.e. methanol) decreases during production.

8.5 Results and Discussion

Table 8.5

Environmental comparative analysis. CMVBD

SiVRMVBD

Steam boiler

158.2473

4.2615

Gas turbine

—

37.8235

CO2 emissions (ton/yr)

Total local emissions (TLE)

158.2473

42.085

TLE savings (%)

—

73.405

Emissions saved at power station

—

26.7791

Total global emissions (TGE)

158.2473

11.0445

%, TGE savings

—

93.02

Therefore, it is required to switch on the control strategy after the startup phase to control the distillate composition. Here, the control objective is to recover the lightest component (i.e. methanol) at a constant purity by adjusting the reflux flow rate. In this chapter, the PI and GSPI controllers are designed for top product composition control by manipulating the reflux flow rate. The setpoint composition has been fixed at 0.999, i.e. the lightest component (methanol) composition value at the end of the steady state. It should be noted that both controllers are switched on only in the production phase. Figure 8.12 shows the profiles of controlled and manipulated variables for CMVBD with PI control, respectively. It is evident from this figure that the setpoint is maintained till 175.5 minutes. After that, the composition suddenly drops to a minimum value at the end of the operation. It is clear that the reflux flow rate increases (although it is not so apparent in the figure due to large scale on y-axis) with time till the end of the batch operation. Interestingly, the reflux flow rate suddenly increased rapidly at the end of batch operation to maintain the desired product composition. Figure 8.13 shows the response of CMVBD with the GSPI controller. The same trend has been observed, like the PI controller. However, compared to the PI controller, a smooth response is observed (although it is not so apparent in the figure due to large scale on y-axis), and it controls the composition till 173.24 minutes. The results are in agreement with the physical interpretation of batch operation. However, for quantitative calculations, simulations terminate for PI and GSPI control strategies at 172.94 minutes and 173.004 minutes, respectively, where the average purity of the product is 0.999. The performance improvements i.e., total distillate collected, purity and batch time achieved by the controlled MVBD over CMBD are summarized in Table 8.6. The PI and GSPI controllers provide more amounts of products with an average purity of 0.999 than that in CMVBD. Therefore, the GSPI controller offers more amount of products than PI-controlled CMVBD and CMVBD. Also, the middle vessel’s middle component purity is approximately the same. It is clear from the results that the total batch time for controlled CMVBD is less than CMVBD, which suggests that

227

xD (mol fraction)

8 Dynamics and Control of Middle-vessel Batch Distillation with Vapor Recompression

1.0 xD, Methanol Set point

0.9 0.8 0.7

Reflux flow rate (Kmol/min)

0.6

0.8 0.6 0.4 0.2 0.0 160

165

170

175

Time (min)

xD (mol fraction)

Figure 8.12 Controlled and manipulated variable profiles throughout the batch operation: PI controller.

1.0 0.9

xD, Methanol Set point

0.8 0.7 0.6

Reflux flow rate (Kmol/min)

228

1.0 0.8 0.6 0.4 0.2 0.0 160

162

164

166

168

170

172

174

Time (min)

Figure 8.13 Controlled and manipulated variable profiles throughout the batch operation: GSPI controller.

controlled CMVBD consumes less energy than CMVBD, consequently reducing CO2 emissions. The control performance specification i.e. Integral Square Error (ISE) is chosen to show the effectiveness of control strategies. The control tuning parameters and ISE of PI and GSPI controllers are shown in Table 8.7. The ISE value of CMVBD-GSPI is lesser as compared to CMVBD-PI, thus indicating that the GSPI controller has a smooth response.

8.5 Results and Discussion

Table 8.6

Performance specifications and comparative analysis.

Item

CMVBD

CMVBD-PI

CMVBD-GSPI

Total distillate (kmol)

0.2866

0.4246

0.6118

Startup phase (min)

160.066

160.066

160.066

Production phase (min)

28.655

12.874

12.942

Total batch time (min)

188.721

172.94

173.008

Composition at the end of batch operation (methanol/ethanol/1-propanol), mol fraction Reflux drum

0.997/0.02/0.0

0.999/0.001/0.0

0.999/0.001/0.0

Middle vessel

0.015/0.973/0.0118

0.0284/0.97/0.0046

0.0281/0.97/0.0049

Reboiler

0.00/0.0001/0.9999

0.00/0.0001/0.9999

0.00/0.0001/0.9999

Table 8.7 Controller parameters and performance specification. Tuning parameters Controller

K C /K C0

𝝉I

ISE

PI

1.51

0.1

5.2055 × 10−6

GSPI

2.52

0.1

1.2624 × 10−6

8.5.4.1 SiVRMVBD-GSPI

This subsection discusses the energy, vapor flow, and CR profiles of GSPI-controlled SiVRMVBD. The two primary goals of controlled single-stage vapor recompression in middle-vessel batch distillation (SiVRMVBD-GSPI) are: (i) to utilize the heat from top vapor, resulting in reduced utility consumption, i.e. steam and cooling water. (ii) To maximize the product, maintain a constant methanol composition at its set point. The SiVRMVBD-GSPI is an unsteady state operation wherein the operational parameter, i.e. internal energy, changes with time. The objective of the study is to maintain the same dynamical performance (same reboiler duty of 4000 kJ/min) between CMVBD and SiVRMVBD-GSPI for fair comparison and, temperature difference of at least 15 ∘ C is maintained for total condensation in reboiler–condenser. To attain the goals mentioned above, an iterative manipulation scenarios are proposed in Section 8.3. Now, SiVRMVBD-GSPI is operated at a fixed reboiler duty of 4000 kJ/min in the production phase only. It should be noted that the GSPI controller takes action when the production phase begins. Figure 8.14 illustrates the manipulation of external energy from the source. It is observed that throughout the steady state operation (profiles same as SiVRMVBD), the energy released by the compressed vapor is less than the required reboiler heat duty; hence, external energy from the source is necessary to make up 4000 kJ/min. However, in the production phase with GSPI control, the energy released by the

229

8 Dynamics and Control of Middle-vessel Batch Distillation with Vapor Recompression

12000 QR QEA QCV

Energy (kJ/min)

10000 8000 6000 4000 2000 0 0

50

100

150

200

Time (min)

Figure 8.14

Energy profile throughout the batch operation: SiVRMVBD-GSPI.

0.35 0.30 Vapor flow rate (Kmol/min)

230

VCond VComp VNT

0.25 0.20 0.15 0.10 0.05 0.00 –20

0

20

40

60

80

100 120 140 160 180 200

Time (min)

Figure 8.15 Vapor flow rate manipulation profile throughout the batch operation: SiVRMVBD-GSPI.

compressed vapor is greater and lower than the required reboiler heat duty; hence, external energy from the source is not required for some time and energy from the source is required for some operation time. Figure 8.15 demonstrates the vapor flow rate manipulation throughout the controlled operation. It is clear that in steady-state operation, the energy released by the compressed vapor is less than the required reboiler heat duty; hence, vapor flow rate manipulation is unnecessary, i.e. the condenser is not required in this period. However, in the production phase with GSPI control, the energy released by the compressed vapor is greater and lower than the required reboiler heat duty; hence vapor flow rate manipulation is needed, i.e. the

8.5 Results and Discussion

40 35

TB–TNT (0K)

30 25 20 15 10 5 0 –5 –20

Figure 8.16

0

20

40

60

80 100 120 Time (min)

140

160

180

200

Temperature difference of the column: SiVRMVBD-GSPI operation.

condenser is required to condense some part of the vapor. But in SiVRMVBD (without control), energy and vapor flow rate manipulations are straightforward, either lower or greater than required. In this way, SiVRMVBD-GSPI saves operating costs: energy from the source (steam) and cooling water. Figure 8.16 shows the temperature difference of the column throughout the operation, and it is observed that it increases continuously with time. It indicates that CR must increase continuously with time. Recall that vapor recompression technique’s feasibility depends on the temperature difference of the column. Figure 8.17 shows CR manipulation to maintain at least 15 ∘ C temperature difference. The temperature difference in the reboiler–condenser is maintained exactly

2.2

CR

2.0

1.8

1.6

1.4 –20

Figure 8.17

0

20

40

60

80 100 120 Time (min)

140

160

180

200

Compression ratio (CR) profile throughout the SiVRMVBD-GSPI operation.

231

8 Dynamics and Control of Middle-vessel Batch Distillation with Vapor Recompression

17

16 Tout–TB (0K)

232

15

14

13 –20

0

20

40

60

80

100

120

140

160

180

200

Time (min)

Figure 8.18

ΔT T profile throughout the SiVRMVBD-GSPI operation.

at 15 ∘ C for complete condensation, as shown in Figure 8.18, by manipulating variable-speed drive compressor.

8.5.5 Energetic, Economic, and Environmental Performance: CMVBD vs. Controlled CMVBD and SiVRMVBD This section compares the performance of controlled SiVRMVBD (SiVRMVBD-GSPI and SiVRMVBD-PI) and CMVBD (without control). As stated, batch distillation operations are highly energy-intensive processes. Therefore, energy consumption is a performance index to measure the thermal efficiency of vapor recompression in CMVBD. However, vapor recompression in CMVBD comprises additional equipment to raise the pressure, i.e. a compressor, as compared to the standard CMVBD; hence, it requires additional capital investment than CMVBD. Therefore, the economic feasibility study of SiVRMVBD is essential. Thus, the SiVRMVBD simulation approach consists of additional steps compared to CMVBD, i.e. manipulation scenarios. Table 8.8 shows the total energy consumption of respective conventional and controlled middle-vessel batch operations. It is clear from simulation results that the SiVRMVBD-PI secures a 79.57% energy savings compared to CMVBD. SiVRMVBDGSPI ensures 78.54% energy savings compared to CMVBD. It is observed that controlled SiVRMVBD shows improved energy savings when compared to controlled CMVBD and CMVBD. SiVRMVBD-PI and SiVRMVBD-GSPI secure almost the same energy savings (no significant change), but the advantage of SiVRMVBD-GSPI is that it provides more distillate product than SiVRMVBD-PI. Table 8.9 compares the economic performance between CMVBD and controlled SiVRMVBD. The results clearly show that 3.302% TAC savings (with

8.5 Results and Discussion

Table 8.8

Energetic performance comparison. Energy consumption (kJ)

Column

CMVBD

MVBD

7.54884 × 105

1.6013 × 105

MVBD-PI

5

6.91760 × 10

1.5418 × 105

MVBD-GSPI

6.92032 × 105

1.6194 × 105

Table 8.9

SiVRMVBD

Economic comparative analysis.

Item

CMVBD

SiVRMVBD-PI

SiVRMVBD-GSPI

Column shell

2.2185 × 105

2.2185 × 105

2.2185 × 105

Column tray

2.3070 × 104

2.3070 × 104

2.3070 × 104

4

4

5.4386 × 104

Capital cost

Reboiler

5.4386 × 10

5.4386 × 10

Condenser

9.4298 × 103

729.578

942.914

Compressor

—

4.3489 × 104

4.2955 × 104

Total

3.0873 × 105

3.4352 × 105

3.4320 × 105

1.8023 × 104

666.777

917.5255

3

Operating cost Steam Cooling water

1.3947 × 10

27.2007

40.361

Electricity

—

3.0844 × 103

3.0382 × 103

Total

1.9418 × 104

3.7783 × 103

3.9961 × 103

TAC

1.2233 × 105

1.1829 × 105

1.1840 × 105

3.302

3.212

%, TAC savings

a payback period of 3 years) can be obtained with SiVRMVBD-PI compared to that in CMVBD. SiVRMVBD-GSPI provides 3.212% TAC savings compared to that in CMVBD. It is clear from the results that vapor recompression reduces operating costs compared to CMVBD, consequently reducing CO2 emissions. Furthermore, it is evident from energy consumption that the controlled SiVRMVBD system is less efficient compared to CMVBD. The proposed controlled SiVRMVBD, reduces CO2 emissions by 94.02% globally. Therefore, SiVRMVBD-GSPI is the most promising controlled configuration with regard to energy, cost, CO2 emissions, amount of product, purity, and practical feasibility, compared to CMVBD.

233

234

8 Dynamics and Control of Middle-vessel Batch Distillation with Vapor Recompression

8.6 Conclusions Batch distillation is energy-intensive and contributes significantly to greenhouse gas emissions (e.g. carbon dioxide). VRC can reduce process costs and CO2 emissions in batch distillation. For the first time, the present research is focused on improving the performance of open-loop middle-vessel batch distillation for separating ternary zeotropic mixture, i.e. methanol/ethanol/1-propanol using VRC. Also, this study implemented PI and GSPI controllers to maintain the constant distillate composition. Furthermore, systematic simulation approaches have been developed for conventional CMVBD and SiVRMVBD. Firstly, a simulation approach has been implemented for closed-loop CMVBD and open-loop CMVBD. The results show that open-loop CMVBD provides increased middle component purity in the middle vessel compared to closed-loop CMVBD. However, the lightest component mole fraction in the distillate decreases in open-loop CMVBD. Then, VRC was implemented for open-loop CMVBD (SiVRMVBD), and results show improved energy and TAC savings of 78.78% and 3.21%, respectively, compared to CMVBD. Also, a 93% reduction in global CO2 emissions is achieved. Finally, PI and GSPI controllers have been implemented for CMVBD and SiVRMVBD to maintain constant distillation composition by manipulating the reflux flow rate. The results show that controlled SiVRMVBD provides improved performance. Overall, the results show that SiVRMVBD-GSPI is the best configuration in terms of energy, economics, environment, amount of product, purity, and practical feasibility compared to CMVBD.

References Adari, P.R. and Jana, A.K. (2008). Comparative control study of a high-purity ternary batch distillation. Chemical Product and Process Modeling 3 (1): https://doi.org/10 .2202/1934-2659.1174. Babu, G.U., Pal, E.K., and Jana, A.K. (2012a). An adaptive vapor recompression scheme for a ternary batch distillation with a side withdrawal. Industrial & Engineering Chemistry Research 51 (13): 4990–4997. Babu, G.U., Aditya, R., and Jana, A.K. (2012b). Economic feasibility of a novel energy efficient middle vessel batch distillation to reduce energy use. Energy 45 (1): 626–633. Babu, G.U. and Jana, A.K. (2014). Impact of vapor recompression in batch distillation on energy consumption, cost and CO2 emission: open-loop versus closed-loop operation. Applied Thermal Engineering 62 (2): 365–374. Babu, G.U. and Jana, A.K. (2013). Reducing total annualized cost and CO2 emissions in batch distillation: dynamics and control. AICHE Journal 59 (8): 2821–2832. Demicoli, D. and Stichlmair, J. (2004). Separation of ternary mixtures in a batch distillation column with side withdrawal. Computers and Chemical Engineering 28: 643–650. Douglas, J.M. (1988). Conceptual Design of Chemical Processes, 1e. New York: McGraw-Hill.

References

Gadalla, M.A., Olujic, Z., Jansens, P.J. et al. (2005). Reducing CO2 emissions and energy consumption of heat-integrated distillation systems. Environmental Science & Technology 39: 6860–6870. Gruetzmann, S., Fieg, G., and Kapala, T. (2006). Theoretical analysis and operating behaviour of a middle vessel batch distillation with cyclic operation. Chemical Engineering and Processing 45: 46–54. Johri, K., Babu, G.U., and Jana, A.K. (2011). Performance investigation of a variable speed vapor recompression reactive batch rectifier. AICHE Journal 57 (11): 3238–3242. Khan, M., Babu, G.U., and Jana, A.K. (2012). Improving energy efficiency and cost-effectiveness of batch distillation for separating wide boiling constituents. 1. Vapor recompression column. Industrial & Engineering Chemistry Research 51 (47): 15413–15422. Monroy-Loperena, R. and Alvarez-Ramirez, J. (2000). Output-feedback control of reactive batch distillation columns. Industrial and Engineering Chemistry Research 39: 378–386. Nair, R., Raykar, A., and Babu, G.U. (2017). Performance investigation of vapor recompressed batch distillation for separating ternary wide boiling constituents. Resource-Efficient Technologies 3 (4): 452–458. Naito, K., Nakaiwa, M., Huang, K. et al. (2000). Operation of a bench-scale ideal heat integrated distillation column (HIDiC): an experimental study. Computers and Chemical Engineering 24: 495–499. Parhi, S., Rangaiah, G.P., and Jana, A.K. (2020). Mixed-Integer dynamic optimization of conventional and vapor recompressed batch distillation for economic and environmental objectives. Chemical Engineering Research and Design 154: 70–85. Shenvi, A.A., Herron, D.M., and Agrawal, R. (2011). Energy efficiency limitations of the conventional heat integrated distillation column (HIDiC) configuration for binary distillation. Industrial and Engineering Chemistry Research 50: 119–130. Skouras, S. and Skogestad, S. (2004). Time (energy) requirements in closed batch distillation arrangements. Computers and Chemical Engineering 28: 829–837. Suphanit, B. (2010). Design of internally heat-integrated distillation column (HIDiC): uniform heat transfer area versus uniform heat distribution. Energy 35: 1505–1514. Vibhute, M.M. and Jogwar, S.S. (2020). Optimal operation and tracking control of vapor-recompressed batch distillation. AICHE Journal 66 (12): e17049. Warter, M., Demicoli, D., and Stichlmair, J. (2002). Batch distillation of zeotropic mixtures in a column with a middle vessel. Computer Aided Chemical Engineering 10: 385–390. Warter, M., Demicoli, D., and Stichlmair, J. (2004). Operation of a batch distillation column with a middle vessel: experimental results for the separation of zeotropic and azeotropic mixtures. Chemical Engineering and Processing: Process Intensification 43 (3): 263–272.

235

237

9 Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices Savyasachi Shrikhande 1 , Gunawant K. Deshpande 2 , Gade Pandu Rangaiah 3,4 , and Dipesh Shikchand Patle 2 1 Chemical Engineering and Analytical Science, University of Manchester, Manchester, M13 9PY, United Kingdom 2 Department of Chemical Engineering, Motilal Nehru National Institute of Technology Allahabad, Prayagraj, 211004, Uttar Pradesh, India 3 Department of Chemical and Biomolecular Engineering, National University of Singapore, Singapore, 117576, Singapore 4 School of Chemical Engineering, Vellore Institute of Technolog, Vellore, 632 014, Tamil Nadu, India

9.1 Introduction Equipment failures and human errors are considered as prime causes of accidents in the process industries (Ahmad et al., 2014). However, it is unrealistic to expect error-free performance from the operators throughout their work tenure. Therefore, it is of paramount importance that plants be designed so that they are user-friendly and can endure the deviations caused by the operator and/or equipment failures (Kletz and Amyotte, 2010). For inherently safer design of any process, several approaches can be followed (e.g. prevention of hazardous material leaks, reducing the use of hazardous materials, and avoiding harsh operating conditions). One of the most critical tasks in process safety is the identification and understanding of the process hazards, before moving on to hazard prevention and risk reduction (Kletz and Amyotte, 2010). Therefore, process safety evaluation during the initial stages of process development and design will aid in selecting a safer process route among viable alternatives. This can be performed through various indices that consider a variety of factors such as temperature, pressure, inventory, toxicity, explosiveness, and flammability. Among the common processes, distillation is a widely used unit operation, and it accounts for about 40% of the energy demand in the process industries (Dejanovic´ et al., 2010). Even a simple conventional arrangement can have an energy requirement of as high as 100 MW and a thermodynamic efficiency of ∼10% (Bruinsma and Spoelstra, 2010). Safety analysis of distillation is beneficial at the design stage (Argoti et al., 2019), and it helps in the prevention of disasters in the chemical industries (Gao et al., 2021). On the other hand, intensified distillation systems such as dividing-wall columns (DWCs) and columns with heat pumps (e.g. mechanical Control and Safety Analysis of Intensified Chemical Processes, First Edition. Edited by Dipesh Shikchand Patle and Gade Pandu Rangaiah. © 2024 WILEY-VCH GmbH. Published 2024 by WILEY-VCH GmbH.

238

9 Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices

vapor recompression, MVR) were studied for the separation of diverse mixtures in the academia; they are also employed in process industries. For example, about 15 years ago, Olujic et al. (2009) stated that more than 70 DWCs are in operation by BASF alone in various regions of the world. DWCs can significantly reduce capital cost, operating cost, and plot space requirements for both new systems and revamping existing systems (Premkumar and Rangaiah, 2009). However, intensified distillation systems may have a higher risk of accidents. In recent years, researchers around the globe have incorporated inherent safety as one of the performance criteria for the evaluation of chemical processes. For example, Vazquez-Castillo et al. (2019) reported a study on heat-integrated reactive distillation to produce ethyl levulinate; they included safety and environmental objectives in the design and optimization. Segovia-Hernandez et al. (2021) analyzed the safety and control of a process system of extraction and DWC for biobutanol production. Deshpande et al. (2022a) reported multi-objective optimization of ultrasound-assisted microalgal biodiesel process for safety, economic, and environmental objectives. Building on that, Deshpande et al. (2022b) optimized an intensified biodiesel process with MVR for multiple objectives. Sun et al. (2023) investigated economic, environmental, and safety aspects of a side-stream extractive distillation for separating azeotropic systems. Similarly, Zhai et al. (2023) analyzed economic, environmental, and safety aspects of extractive DWCs without/with feed pre-heating and heat pumps for separating a binary azeotrope. In the literature, there are many indices and/or methods for safety analysis, and they have been applied to different process systems to assess their inherent safety. However, there are not many studies on the application of safety indices to intensified distillation columns. Many studies on the intensification of distillation report the benefits of intensification (e.g. DWC or MVR) in terms of energy savings and cost; however, their ramifications pertaining to safety are not sufficiently quantified. This chapter addresses this gap in the literature by applying selected safety indices to three distillation systems for the separation of a liquid mixture of water, methanol, glycerol, hexane, and ionic liquid (IL) in a biodiesel process. The three systems are: conventional sequence of columns (CSC), DWC, and DWC with MVR (DWC–MVR). Moreover, inputs from experienced engineers on the safety of these systems are sought and discussed; they are also compared with the ranking of these systems using the safety indices. This practical analysis is rare in the open literature. Finally, a modified safety index and its application to these systems are presented. The rest of this chapter is organized as follows: Section 9.2 summarizes safety indices in the literature since 1999. Then, CSC, DWC, and DWC–MVR systems are described in Section 9.3. Next, the selection of safety indices for the present study is outlined in Section 9.4. Results from the application of the selected safety indices to the three distillation systems are presented and discussed in Section 9.5. A survey of experienced engineers on the safety of the three distillation systems and their responses are reported in Section 9.6. Based on these survey responses, a modified safety index and its application to these systems are described in Section 9.7. Finally, the main findings of this study are summarized in Section 9.8.

9.2 Safety Indices for Process Safety Assessment

9.2 Safety Indices for Process Safety Assessment Table 9.1 summarizes the safety indices proposed since 1999, which can be employed for assessing the inherent safety of any chemical process during its development and design phase. The first among these indices is the inherent safety index (ISI) proposed by Heikkilä (1999). It contains two parts – one pertaining to chemical safety and the other to process safety. The chemical safety index is further divided into two sub-indices for reaction hazards and hazardous substances. Likewise, the process safety contains two sub-indices for process conditions and process systems. Numerical scores are assigned to each of them based on various existing indices such

Table 9.1 Summary of safety indices that can be used during the development and design of chemical processes. Index

Key points

References

Inherent Safety Index (ISI)

Easy to execute; however, it includes subjective scaling.

Heikkilä (1999)

Safety, Health, and Environment (SHE)

Accounts for safety, health, and environmental aspects of a process.

Koller et al. (2000)

i-Safe

Accounts for a wide range of factors including unwanted or runaway reactions.

Palaniappan et al. (2002a, b)

Integrated Inherent Safety Index (I2SI)

Accounts for toxicity, fire and explosion, and environmental aspects of a process.

Khan and Amyotte (2004)

Process Route Index (PRI)

Calculates safety based on the influence of temperature and pressure on explosiveness.

Leong and Shariff (2009)

Toxic Release and Consequence Analysis Tool (TORCAT)

Index describing toxic releases of chemicals and their consequences.

Shariff and Zaini (2010)

Process Stream Index (PSI)

Focusses on the safety assessment of streams in a process.

Shariff et al. (2012)

Numerical Descriptive Inherent Safety Technique (NuDIST)

Numerical method-based index for inherent safety to ensure continuity in contrast to discrete indices.

Ahmad et al. (2014)

Individual Risk (IR)

Gives the probability of death of any person.

Medina-Herrera et al. (2014)

239

240

9 Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices

as Mond Index (Tyler, 1985) and Dow Fire and Explosion Index (AIChE and Dow Chemical Company, 1994) for factors such as toxic exposure and pressure parameters, respectively. The calculations for ISI are made based on the worst-case scenario. A safer process has a lower value of ISI. Unlike ISI, which focuses solely on process safety, the safety, health, and environment (SHE) index of Koller et al. (2000) incorporates various safety, health, and environmental aspects of each process route under investigation. The safety parameters include fire, acute toxicity, reaction, and decomposition (probability for undesired, runaway, and decomposition reactions). Health aspects include irritation and chronic toxicity. Five factors included under environmental aspect of SHE index are: water-mediated effects, aerial effects, degradation, solid waste, and accumulation. Each parameter value is calculated with its relevant fate index to find the effective dangerous property (EDP). Then, EDP values are converted onto an exponential scale and multiplied by the inventory (mass) of that chemical in the system to find the potential of danger (POD). Summation of POD for all components in the system and for all dangerous properties concludes the SHE calculation. A lower value of SHE indicates a safer process. Another method for the calculation of inherent safety is the i-safe method proposed by Palaniappan et al. (2002a, b). In this, various process routes are compared based on their overall safety index, which is a summation of overall chemical index and overall reaction index. There are four parameters (namely, flammability, toxicity, explosiveness, and reactivity rating) in the chemical index and another four parameters (namely, temperature, pressure, yield, and heat of reaction) in the reaction index. Khan and Amyotte (2004) proposed an integrated inherent safety index (I2SI) that considers a wide range of potential hazards, which aids in the investigation of different process routes. I2SI comprises three indices: damage index (DI), process and hazard control index (PHCI), and ISI. For each unit operation, DI, PHCI, and ISI are calculated, and subsequently, I2SI is calculated. A summation of I2SI values for all unit operations is the overall I2SI for the process. DI accounts for fire and explosion, acute and chronic toxicity, and environmental aspects, while PHCI accounts for process control systems and hazard control systems. A lower value of I2SI indicates a safer process. In addition to the above standalone indices, there are some indices that carry out inherent safety assessment using the data from process simulators such as Aspen Plus or HYSYS; they are: process route index (PRI), toxic release and consequence analysis tool (TORCAT), and process stream index (PSI), proposed by Leong and Shariff (2009), Shariff and Zaini (2010), and Shariff et al. (2012), respectively. PRI considers the effect of process conditions such as temperature and pressure on explosiveness, and it calculates inherent safety based on explosivity of the process alternative under investigation. On the other hand, TORCAT focuses on the toxicity parameter, whereas PSI focuses on the safety associated with streams rather than equipment in the process, to calculate inherent safety. Ahmad et al. (2014) proposed a numerical descriptive inherent safety technique (NuDIST), which incorporates the logistic function for hazard score assignment,

9.3 Description of Distillation Systems

thus providing a continuous scoring rubric. Like previous indices, NuDIST also has two sub-indices for chemical safety and process safety, which in turn have four sub-indices each: flammability, explosiveness, toxicity, and reactivity for chemical safety; and temperature, pressure, heat of reaction, and process inventory for process safety. Medina-Herrera et al. (2014) introduced a new methodology to calculate the probability of fault events that could cause damage to equipment or result in loss of life. This index utilizes quantitative risk analysis to categorize incidents into continuous and instantaneous events. Furthermore, HAZOP study is used to determine the possibility of instantaneous and continuous hazards in the process. The instantaneous release considers boiling liquid expanding vapor explosion (BLEVE), unconfined vapor cloud explosion (UVCE), flash fire, and toxic release, whereas jet fire, flash fire, and toxic release are categorized under continuous release events.

9.3 Description of Distillation Systems The three distillation systems used in this study are taken from Shrikhande et al. (2021), who reported a plantwide process to produce biodiesel from wet microalgal biomass. Any one of them can be employed in the separation section of this biodiesel process. As part of their study, Shrikhande et al. investigated DWC and DWC–MVR to improve CSC for the separation of a mixture of water, methanol, glycerol, hexane, and BBAIL (i.e. benzimidazolium based Brønsted acid ionic liquid) catalyst. The three distillation systems are described below.

9.3.1

Conventional Sequence of Columns

Figure 9.1 shows the schematic of CSC taken from Shrikhande et al. (2021); this CSC has two distillation columns, labeled as FRAC-3 and FRAC-4. A mixture of water, methanol, glycerol, hexane, and IL is sent to FRAC-3 operating at 1 atm. FRAC-3 has 7 stages, with feed entering on the 5th stage. A mixture of light components: methanol, hexane, and water, is obtained as the distillate (DIST-1 in Figure 9.1), and glycerol and IL are recovered from the bottoms (BOT-1 in Figure 9.1). DIST-1 is then fed on the 6th stage of FRAC-4 having 12 stages and operating at 1 atm. Distillate from FRAC-4 is 90 wt% methanol, 4 wt% water, and 6 wt% hexane, whereas 98 wt% water is obtained from the bottoms of FRAC-4. Note that pressure drops in the column, condenser, and reboiler are not considered in this study.

9.3.2

Dividing-Wall Column

Figure 9.2 shows the process flowsheet of DWC that can be chosen instead of FRAC-3 and FRAC-4 in CSC. This DWC has four sections: the top section (trays 2–5, marked in red) represents FRAC-4 section, the common section (left section: trays 1–4, marked in black) represents FRAC-3 section; right section (trays 6–11,

241

242

9 Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices DIST-1 F = 83865 kg/h T = 30 °C XWAT = 0.50 XMA = 0.46 XHEX = 0.04

Qc = 46 MW Qb = 47.2 MW 1 atm

1 atm

6

5 FEED F = 84339 kg/h T = 30 °C P = 1 atm XWAT = 0.50 XMA = 0.46 XGLY = 0.003 XHEX = 0.03 XIL = 0.007

Qc = 25.3 MW Qb = 28.2 MW

6

FRAC-3 H = 4.8 m c Dc = 3.81 m 7

FRAC-4

DIST-2 F = 42070 kg/h T = 63 °C XWAT = 0.04 XMA = 0.90 XHEX = 0.06

Hc = 9.6 m Dc = 3.19 m

11 12 BOT-2 F = 41796 kg/h T = 100 °C XWAT = 0.98 XMA = 0.02

BOT-1 F = 473 kg/h T = 137 °C XWAT = 0.13 XGLY = 0.54 XIL = 0.33

Figure 9.1 Two columns in CSC; here, F is the flow rate, T is the temperature, P is the pressure, X WAT , X MA , X GLY , X HEX , and X IL are mass fractions of water, methanol, glycerol, hexane, and ionic liquid, respectively, Qc and Qb are the duties of the condenser and reboiler, respectively, and Hc and Dc are the height and diameter of the column, respectively.

Qc = 53.5 MW Qb = 57.5 MW 1 atm

FEED F = 84339 kg/h T = 30 °C P = 1 atm XWAT = 0.50 XMA = 0.46 XGLY = 0.003 XHEX = 0.03 XIL = 0.007

DIST-2 F = 42070 kg/h T = 63 °C XWAT = 0.04 XMA = 0.90 XHEX = 0.06

2–5 DWC

1–4

6–11 11 5

6

Hc = 9.6 m Dc = 4.4 m

BOT-2 F = 41796 kg/h T = 100 °C XWAT = 0.98 XMA = 0.02

BOT-1 F = 473 kg/h T = 137 °C XWAT = 0.1 XGLY = 0.56 XIL = 0.34

Figure 9.2 Dividing-wall column system, see the caption of Figure 9.1 for the significance of symbols.

marked in red) represents FRAC-4 section, and the bottom section, which contains the 5th stage of FRAC-3 section, besides the single condenser and single reboiler. The feed is introduced into the DWC on the 3rd stage on the left section. The vapor generated from the reboiler and flowing upwards is split and sent to the 4th stage on the left section and the 11th stage on the right section. Similarly, the liquid flowing down from stage 5 in the top section is split into the 1st stage on the left section and the 6th stage on the right section. This design of DWC is adapted from Shrikhande et al. (2021). Water is removed from the 11th stage on the right section of the DWC, and methanol and a mixture of IL and glycerol are obtained in the distillate and bottoms, respectively.

9.3 Description of Distillation Systems FEED F = 84339 kg/h T = 30 °C P = 1 atm XWAT = 0.50 XMA = 0.46 XGLY = 0.003 XHEX = 0.03 XIL = 0.007

Qc = 57.22 MW

COND

T = 63 °C P = 1 atm

T = 154 °C P = 1 atm VF = 1

DIST-2 F = 42070 kg/h T = 63 °C XWAT = 0.04 XMA = 0.90 XHEX = 0.06

COMP-1 2–5

T = 169 °C P = 3.5 atm

6–11

1–4

Hc = 9.6 m Dc = 4.4 m

DWC 11

T = 146 °C P = 3.5 atm COMP-2 VF = 1

T = 154 °C P = 12.25 atm VF = 1

Qb = 51.47 MW

5

T = 257 °C P = 12.25 atm HX-1 T = 137 °C P = 1 atm VF = 0

T = 138 °C P = 1 atm VF = 0.01 Q = 0.9 MW

BOT-1 F = 473 kg/h T = 137 °C XWAT = 0.1 XGLY = 0.56 XIL = 0.34

BOT-2 F = 41796 kg/h T = 100 °C XWAT = 0.98 XMA = 0.02

HX-2 Q = 4.7 MW

T = 144 °C P = 1 atm VF = 0.11

Figure 9.3 Dividing-wall column with vapor recompression; see the caption of Figure 9.1 for the significance of symbols.

9.3.3

Dividing-Wall Column with Mechanical Vapor Recompression

To improve the performance of DWC, MVR was added. Figure 9.3 shows the schematic of DWC–MVR system, wherein the four sections of DWC are identical to those in Figure 9.2. The main difference is the use of MVR to compress (part of) overhead vapor to increase its temperature for use as the heating medium in the reboiler, thus reducing the utility steam required for DWC operation. Based on the results reported by Shrikhande et al. (2021), the economical configuration of DWC–MVR is to compress 50% of the overhead vapor (based on preliminary sensitivity analysis); a higher vapor load on the compressor reduces the hot utility consumption but increases the overall cost. Therefore, in this study, the overhead vapor from DWC is split via a flow splitter, with 50% of the vapor going to compressor 1 (COMP-1 in Figure 9.3) and the rest going to the condenser (COND in Figure 9.3). COMP-1 is operated at a pressure ratio of 3.5, and its outlet vapor is sent to pre-heat the bottom stream of DWC in the heat exchanger, HX-1. The cooled vapor is then sent to the second compressor (COMP-2), operating with the same pressure ratio of 3.5. The recompressed vapor at a pressure of 12.25 atm from COMP-2 is used to vaporize the pre-heated bottom stream in another heat exchanger, HX-2. After exiting HX-2, the bottom stream is sent to a heater (REB), which vaporizes the remaining liquid, and REB’s outlet is returned to DWC as boil-up. The condensed vapor from HX-2 is passed through a throttling valve, which drops its pressure to 1 atm, which results in some vaporization, and then sent to the condenser (COND in Figure 9.3). The condensed vapor is sent to the reflux drum, from where the distillate of 90 wt%

243

244

9 Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices

methanol is taken out at the specified flowrate of 42,070 kg/h, and the remaining liquid is returned to DWC as the reflux. Water is removed from the 11th stage of DWC (BOT-2 in Figure 9.3), and a mixture of glycerol and IL is obtained in BOT-1 from bottom stream of DWC.

9.4 Selection of Safety Indices As presented in Section 9.2, there are many indices in the literature for the assessment of inherent safety of chemical processes. Several factors need to be considered for shortlisting a few of them for evaluating inherent safety of any process during the conceptual design phase; they include risks considered in the index and their relevance to the process and/or purpose of assessment, as well as the data required for the calculation of that index and its availability during the design stage. Although easy to use, ISI cannot be applied during the design phase due to the unavailability of certain key data. For example, ISI incorporates Dow’s Fire and Explosion Index, which can only be applied to an already established process (AIChE and Dow Chemical Company, 1994). The next index is the i-Safe, which, as stated in Section 9.2, is a comprehensive index that accounts for physical and chemical safety parameters such as temperature, pressure, and heat of reaction. However, CSC, DWC, and DWC–MVR systems (Figures 9.1–9.3, respectively) do not involve any reactions; also, they do not have many hazardous compounds to show enough variation in the i-safe index. Calculation of I2SI for each unit operation involves DI, ISI, and PHCI. Of these three sub-indices, PHCI and ISI rely on the expertise of process specialists; they also, cannot be computed based on the information available in the design stage. IR is a relatively newer index and signifies the death probability by considering a range of hazardous scenarios. Moreover, it can be calculated for any process in the design stage. Both PRI and PSI depend on the stream data and not on the equipment characteristics; further, estimating PRI and PSI is generally simple as the required stream data can be obtained from process simulation. However, there is a notable difference between PRI and PSI in calculating the inherent safety. PRI is calculated for a process as a whole. On the other hand, PSI is calculated for each stream, which means it estimates relative risk of streams in a process. Hence, PSI is not appropriate for comparing safety of alternative processes. Upon careful evaluation of the above factors in the calculation of several indices and their usage in the design phase, three indices, namely, DI, IR, and PRI, were shortlisted for the study in this chapter. The calculation procedures for DI, IR, and PRI were adapted from Sharma et al. (2013), Deshpande et al. (2022b), and Leong and Shariff (2009), respectively, for application to the three distillation systems with minor assumptions in DWC and DWC–MVR cases, which are mentioned in the subsequent sections.

9.5 Results and Discussion

9.5 Results and Discussion 9.5.1

Conventional Sequence of Columns

Figure 9.4 shows the temperature and composition profiles in FRAC-3 and FRAC-4 columns of CSC system in Figure 9.1. Glycerol is obtained with IL as the bottom stream of FRAC-3, and a mixture of methanol, water, and hexane is obtained in the distillate. Similarly, it is evident that clear separation of water–methanol mixture is achieved in FRAC-4. Temperature in FRAC-3 varies from 57 ∘ C in the condenser to 135 ∘ C in the reboiler, whereas the corresponding values in FRAC-4 are 63 and 100 ∘ C. The total reboiler duty and condenser duty for CSC system are 75.38 and 73.38 MW, respectively. The height and diameter of FRAC-3 are 3.81 and 4.8 m, respectively, whereas those of FRAC-4 are 3.9 and 9.6 m, respectively. Note that the column diameter was obtained from Aspen Plus simulation, and the column height was estimated using the procedure proposed by Luyben (2002). As stated earlier, the calculation procedures for DI, IR, and PRI used in this study were adapted from the literature. DI calculation is dependent on some key values (e.g. continuous release considering either feed or products (whichever has greater damage potential), volume of equipment, chemicals, and operating conditions). For IR calculation, based on the height and diameter of the column, column volume is estimated; it is 54.8 and 77 m3 , respectively, for FRAC-3 and FRAC-4. Thereafter, the liquid volume on the trays is obtained from: 𝜋 ( 2) Dc × wh × ns (9.1) 4 Here, Dc is the column diameter, wh is the weir height (0.06667 m in this study), and ns is the number of ideal stages. The liquid hold up in the sump is assumed to be about 40% of the sump volume, where the sump volume is determined assuming a sump height of 1.8 m. Thus, the liquid volume in FRAC-3 and FRAC-4 was found to be 13.32 and 12.41 m3 (including the hold up in the column sump), respectively, and the vapor volume of the columns is 41.48 and 64.59 m3 , respectively, for CSC. IR categorizes the outcomes based on either continuous or instantaneous releases. Mass released because of an instantaneous release is assumed as the total quantity of material inside the equipment. Mass inside the distillation column (mcolumn ) is determined based on volume fraction (f ) of liquid and vapor inside the column and their respective average density (𝜌) using: mcolumn = (𝜌liquid × fliquid + 𝜌vapor × fvapor )Vcolumn

(9.2a)

To calculate the material in the reflux drum and reboiler, their inventory is assumed to be between 12 and 6 minutes of the column feed flowrate (F) (as in Contreras-Zarazúa et al., 2019). Hence, mass in the reflux drum and reboiler is respectively given by: mreflux drum = 720 × F

(9.2b)

mreboiler = 360 × F

(9.2c)

245

9 Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices

Composition

1.00 MEOH Water Glycerol Hexane Temperature IL

0.75 0.50 0.25

Temperature (C)

0.00 138 115 92 69

2

4

6

No. of stages

(a)

Composition

1.00 0.75 Water Hexane

0.50

Temperature MEOH

0.25 0.00

Temperature (C)

246

99 88 77 66 2

(b)

4

6

8

10

12

No. of stages

Figure 9.4 (a) Temperature and composition profiles in FRAC-3 (left plot) and (b) FRAC-4 (right plot).

Finally, mass released Q* (in kg) due to instantaneous release can be estimated by (Contreras-Zarazúa et al., 2019): Q∗ = (mcolumn + mreflux drum + mreboiler )

(9.2d)

For FRAC-3 and FRAC-4, Q* was calculated to be 12525 and 12,455 kg, respectively.

9.5 Results and Discussion

Mass due to continuous release (mr ) from process equipment or pipe can be obtained from the source model (Eq. 9.3) (AIChE, 2000). This is assuming an average leak from an opening of 2.5 cm (diameter) in both liquid (Eq. 9.3a) and vapor (Eq. 9.3b) phases. √ ( ) gc Pg + ghL mliquid = 𝜌ACD 2 (9.3a) 𝜌 √ kgc M ( 2 )(k+1)∕(k−1) mvapor = CD AP1 (9.3b) Rg T1 k + 1 mliquid + mvapor mr = (9.3c) 2 Here, mliquid and mvapor are mass of liquid and vapor released (kg/s), respectively; 𝜌, A, CD , gc , Pg , g, and hL are density of liquid, cross sectional area of the leak (cm2 ), discharge coefficient (0.61 for liquid and 1 for vapor), gravitational constant (1), upstream pressure (1 atm), acceleration due to gravity (9.81 m/s2 ) and liquid head (0 m; based on Deshpande et al., 2022b; and Medina-Herrera et al., 2014), respectively; and P1 , M, Rg , T 1 , and k are upstream absolute pressure (1 bar), average molecular weight, gas constant (8314 J/kg-mol/K), upstream temperature (in K), and ratio of heat capacity at constant pressure to that at constant volume, respectively, (1.32). Using Eq. (9.3), mr of 6.03 × 10−3 kg/s was calculated for both FRAC-3 and FRAC-4. Then, effects of instantaneous and continuous releases (Q* and mr , respectively) were calculated; they are: BLEVE, UVCE, flash fire, and toxic release for instantaneous release, and jet fire, flash fire, and toxic release for continuous release. Detailed calculations for each of these events are available in Deshpande et al. (2022b). In these calculations, IL was not included due to unavailability of its toxicity parameters (Table 9.2). For accidents such as BLEVE, jet fire, and flash fire, thermal radiation (Er) is taken as the causative variable. The overpressure (po ) is considered as the causative variable in the case of UVCE. Probit functions are used to determine the probability of death due to toxic substances and exposure to heat radiation. In this work, the death of a person is considered as damage caused by the respective accident. Table 9.2 Lower flammability limit (LFL), upper flammability limit (UFL), and LC50 (i.e. lethal concentration, which refers to the concentration of the chemical that has potential to kill 50% of occupants) of methanol, glycerol, and hexane. S. No.

Component

LFL (%)

UFL (%)

LC50 (ppm)

1

Methanol

6.0

36

64,000

3

Glycerol

2.7

19

570

4

Hexane

1.20

7.7

48,000

Source: Material Safety Data Sheet; https://www.osha.gov, 2023/U.S. DEPARTMENT OF LABOR/public_domain.

247

248

9 Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices

Table 9.3

Probit model parameters.

Incidents

k1

k2

V

Thermal radiation

−14.9

2.56

te ∗

Overpressure (UVCE)

−77.1

6.91

Po

4∕3

Er

104

Source: Adapted from Crowl and Louvar (2002).

Eq. (9.4) relates the death due to thermal radiation and overpressure due to explosions (p∘ ) with probit variable (Y ). In this equation, K1 and K2 are the constants for the respective event, and their values are given in Table 9.3, where te and Er for calculating V are the duration of the event and energy released in that particular event, respectively. The probability of affectation (Pa ) for incident i is calculated from the causative variables of respective consequence and probit function using Eq. 9.5 (De Haag and Ale, 2005). Y = k1 + k2 ln V [ ( Pai = 0.5

1 + erf

|Y − 5| √ 2

Finally, IR is estimated by: ∑ IR = fi Pai

)]

(9.4) (9.5)

(9.6)

Here f i and Pai are the occurrence frequency of incident i (for chemical processes, in general) and the probability of injury/death due to incident i, respectively (AIChE, 2000). The summation in Eq. 9.6 covers all incidents; here, total number of incidents is 7. Table 9.4 shows the individual death probability for each of the 7 scenarios considered in the IR calculation, and their corresponding frequencies per year, for the CSC system. In this study, for the calculation of IR, energy released is evaluated by quantitative risk analysis of each equipment, and then the probability of death is calculated for all accident scenarios. Only energy emitted by the BLEVE is more than that required for individual death/injury. Hence, only BLEVE shows probability of death in Table 9.4. The UFL and LFL of the mixture was calculated as 93.93 and 17.39, respectively, along with the heat of combustion of the mixture as 8010.56 kJ/kg. If the concentration lies between UFL and LFL values stated in Table 9.2, the probability of death would be 100% (Medina-Herrera et al., 2014). As water and methanol are the major components in the system, the overall influence of other components such as glycerol and hexane, is not much. The probability of death in the IR calculation for different events for CSC (and later for DWC and DWC–MVR) is 0 because the concentration of the particular component in the column does not lie between UFL and LFL for UVCE, flash fire (both instantaneous and continuous), and jet fire, and the concentration of the particular component in the column is not beyond LC50 for toxic release, whose values for individual components are given in Table 9.2.

9.5 Results and Discussion

Table 9.4

IR event table for CSC system.

S. No.

Event

FRAC-3a)

FRAC-4a)

fi

IR

1

BLEVE

1.71 × 10−9

1.48 × 10−9

0.00000575

1.833 × 10−14

2

UVCE

0

0

0.00000776

0

3

Flash Fire (I)b)

0

0

0.00000776

0

0

0

0.000248

0

0

0

0.0000376

0

0

0

0.00000155

0

0

0

0.0000826

0

4

Flash Fire

5

Jet Fire

(C)b) (I)b)

6

Toxic Release

7

Toxic Release (C)b)

1.833 × 10−14

Total a) Including reboiler, condenser, and reflux drum. b) Here, (I) and (C) are for instantaneous and continuous releases.

The calculation of DI is based on the continuous release, mr , and volume of the unit. It involves fire and explosion damage radius (FEDR) and toxic damage radius (TDR). For the evaluation of FEDR for equipment involving phase change and separation, there are three energy factors and six penalties. Energy factor F1 accounts for the energy of hazardous content, whereas F2 and F3 factors depend on the physical condition of hazardous content within the equipment. These three factors are given by: F1 = 0.1mrHc K

(9.7a)

F2 = 1.304 × 10−3 PP V

(9.7b)

F3 =

1.0 × 10−3 (P − VP )2 V T + 273 P

(9.7c)

In Eq. (9.7a), heat of combustion, H c is in kJ/mol, and K is a constant (i.e. 3.148). In Eqs. 9.7b and 9.7c, PP (kPa) and V P (kPa) are process pressure and vapor pressure at process temperature T (∘ C), respectively, and V stands for the volume of equipment (m3 ). In addition to the three energy factors, penalties have been applied to consider the effects of various parameters on the total damage capability, which is finally converted to FEDI; see Sharma et al. (2013) for these details. TDR is the radius of the location (in meters) that is affected lethally by toxic loading with a 50% chance of producing mortality. It is calculated utilizing transport phenomena and an empirical approach relying on the number of chemicals inside the unit, their physical state, toxicity, operating conditions, and site characteristics. It is assumed that the dispersion occurs in fairly stable atmospheric conditions. TDR is calculated by estimating the G factor as well as seven penalties (Sharma et al., 2013). The G factor takes into consideration the quantity of chemical emitted and releasing conditions. It can be calculated as: G=S×m

(9.8)

249

250

9 Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices

Here, m is the material release in kg/s and S is the release condition calculated based on the procedure given in Khan et al. (2001). TDR of the process equipment is calculated by combining G factor and pnr1 to pnr7 as per the following (Khan et al., 2001): TDR = 25.35 (G × pnr1 × pnr2 × pnr3 × pnr4 × pnr5 × pnr6 × pnr7 )0.425 (9.9) Once FEDR and TDR are calculated, FEDI and TDI (i.e. toxic damage index) are found from an equation developed by Deshpande et al. (2022b) based on the graphs in Khan and Amyotte (2004). Same values of FEDR and TDI are used for finding the acute TDI (TDIAC ) and chronic TDI (TDICH ) from different graphs. In certain cases where graph has bound limitations, extrapolations are made using data fitting. Then, overall TDI is calculated as follows (Sharma et al., 2013) [ ]1∕2 TDI = (TDICH )2 + (TDIAC )2 (9.10) Finally, DI for that process equipment is evaluated by: [( )2 ]1∕2 )2 (∑ ∑ TDI DI = FEDI +

(9.11)

The above procedure is followed for the evaluation of FEDI and TDI for each equipment in the system, i.e. column, reflux drum, condenser, and reboiler. For the evaluation of DI for reflux drum and reboiler, the volume in consideration is calculated by assuming residence time of 12 and 6 minutes, respectively (similar to those assumed in IR calculation). Then, DI was evaluated based on all the inlet and outlet stream flowrates (kg/s) for each of all equipment, i.e. heat exchangers, compressors, condensers, and reboilers; and the largest value was selected for that equipment. A similar methodology was used earlier by Sharma et al. (2013). Table 9.5 shows the calculated values of DI for CSC system as 514.74. As given earlier in this section, volume of FRAC-3 and FRAC-4 are 54.8 and 77 m3 , respectively. As the operating conditions of both the columns are similar (evident from the temperature profiles in Figure 9.4), it may be concluded that a higher volume translates to a higher risk. Similar observation can be made for the reboiler; the load on the reboiler of FRAC-3 is higher in comparison to the reboiler of FRAC-4, with higher temperature (Figure 9.4) and lower bottoms flowrate (Figure 9.1). Table 9.6 shows the calculation of PRI for CSC system; it is calculated to be 1.225. Here, for each stream, mass heating value, density and pressure are taken from the simulation results of the process, and ΔFLmix was computed by taking the sum of difference between UFL and LFL for each component multiplied by its mass fraction. It can be inferred from ΔFLmix values that FEED, DIST-1 and DIST-2 are more explosive than the other streams in the system. BOT-2 that contains 98% water, is the least explosive whereas DIST-2 has the highest ΔFLmix value of 26.76, due to the presence of hexane and methanol mixture. Due to recovery of IL, and Glycerol in BOT-1, the average density of the BOT-1 is considerably higher than that of the other streams that predominantly contain water and methanol.

9.5 Results and Discussion

Table 9.5

DI calculation for the CSC system in Figure 9.1. FEDI

TDI

DI

Column

77.4

34

[(437.01)2 + (272)2 ]1/2 = 514.74

Reflux Drum

31.79

12

Condenser

70.51

21

Reboiler

89.21

66

84.92

25

Reflux Drum

3.99

16

Condenser

38.98

30

Reboiler

40.21

68

Total

437.01

272

FRAC-3

FRAC-4 Column

Table 9.6

PRI Calculation for the CSC system in Figure 9.1

Mass heating value (kJ/kg) 3

Feed

Dist-1

Bot-1

Dist-2

Bot-2

Average of Streams

11448.21

11425.50

6763.37

7235.55

15397

10453.93

Density (kg/m )

852.47

836.55

1280.86

743.72

914.93

925.71

ΔFLmix

13.67

13.69

8.65

26.76

0.54

12.66

Pressure (atm)

1.00

1.00

1.00

1.00

1.00

1.0

PRI = (10453.93 × 925.71 × 1.0 × 12.66)/108 = 1.225

9.5.2

Dividing-Wall Column

Figure 9.5 shows the composition and temperature profiles in the left and right sections (including common top and bottom sections) of DWC. These profiles are somewhat like the profiles of FRAC-3 and FRAC-4 in CSC system (Figure 9.4); due to the integration of two columns, a slight difference in the profiles of CSC and DWC is observed. Stages 1–4 on the left side and stages 6–11 on the right side have analogous temperature profiles. Approximately 5 ∘ C difference between the left and right sides of the wall is observed, which indicates there is little or no heat transfer across the wall. It is evident from Figure 9.5 that DWC system provides a similar degree of separation as the CSC system, with 90 wt% methanol obtained in the distillate and water is removed from the 11th stage on the right side of DWC (Figure 9.5a), and with temperature ranging from 63 ∘ C in the condenser to 135 ∘ C in the reboiler. The DWC system shows a 23.71% reduction in reboiler duty and about 25% reduction in

251

9 Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices

Composition

1.00 MEOH Water Glycerol Hexane Temperature IL

0.75 0.50 0.25

Temperature (C)

0.00 138 115 92 69 2

4

6

No. of stages

(a)

Composition

1.00 0.75 Water Hexane Temperature MEOH

0.50 0.25 0.00

Temperature (C)

252

99 88 77 66 2

(b)

4

6

8

10

No. of stages

Figure 9.5 Profiles of liquid mass fractions and temperature: (a) left side of DWC, and (b) right side of DWC.

condenser duty. The IR calculation for DWC is similar to the method described for CSC system in the previous section. The DWC column is 4.4 m in diameter and 9.6 m in height, as shown in Figure 9.2, and has a volume of 145.9 m3 . The instantaneous (Q* ) and continuous (mr ) mass release for DWC case are estimated to be 25182.4 kg and 6.03 × 10−3 kg/s, respectively. Table 9.7 shows the IR event table for the DWC system. The overall IR value is found to be 1.025 × 10−12 , which is higher than that of the CSC system (1.833 × 10−14 ) by two orders of magnitude. Upon integration of the FRAC-3 and FRAC-4 into the DWC, the overall material inventory in the DWC (25182.4 kg) is more compared to that in CSC system (24979.09 kg, which is split between FRAC-3 and FRAC-4); it is all in a single column, which increases the risk. As explained in the previous section, DI calculation involves the estimation of the damage radius for fire and explosion as well as toxic release. Table 9.8 shows the results of DI for DWC system. Owing to the reduced number of streams (because of

9.5 Results and Discussion

Table 9.7

IR event table for DWC system. Events

DWCa)

Fi

IR

BLEVE

1.78 × 10−7

0.00000575

1.025 × 10−12

2

UVCE

0

0.00000776

0

3

Flash Fire (I)b)

0

0.00000776

0

0

0.000248

0

0

0.0000376

0

0

0.00000155

0

0

0.0000826

0

1

4

Flash Fire

5

Jet Fire

(C)b) (I)b)

6

Toxic Release

7

Toxic Release (C)b)

1.025 × 10−12

Total a) Including reboiler, condenser, and reflux drum. b) Here, (I) and (C) are for instantaneous and continuous releases.

Table 9.8

DI calculation for DWC system. FEDI

TDI

DI

Column

104.75

34

[(262.77)2 + (202)2 ]1/2 = 331.44

Reflux drum

29.32

20

Condenser

92.99

46

Reboiler

35.71

102

Total

262.77

202

the integration of FRAC-3 and FRAC-4 into DWC) and milder conditions, a significant decrease in DI is observed in comparison to the DI obtained in the CSC system (i.e. from 514.74 to 331.44 or ≈36 % reduction). Table 9.9 shows PRI calculation for DWC system as 1.232. Even with fewer streams than CSC system, DWC system has a slightly higher PRI. Upon comparison of PRI of CSC and DWC systems (Tables 9.6 and 9.9), it is evident that the average density has increased by a noticeable extent (3.95%) while the average heating value and average ΔFLmix (i.e. summation of difference between UFL and LFL for each component in a stream multiplied by its mass fraction) are reduced by 1.7%. The density of stream ‘Bot-1’ in DWC is 1344.62 kg/m3 (Table 9.9 and Figure 9.2), while, for the stream ‘Bot-1’ in CSC, it is 925.51 kg/m3 (Table 9.6 and Figure 9.1). The reason behind this increase is a lower water mass fraction in the DWC system, i.e. 0.1 in the Bot-1 stream of DWC as compared to that in CSC of 0.13). Since the other components are considerably denser, the density of the stream has increased, which translates to higher PRI.

9.5.3

Dividing-Wall Column with Mechanical Vapor Recompression

After achieving a significant reduction in heating utility requirement via the DWC system, the application of MVR in DWC (Figure 9.3) resulted in 31.72% reduction in

253

254

9 Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices

Table 9.9

PRI calculation for the DWC system.

Feed

Bot-1

Dist-2

Bot-2

Average of streams

Mass heating value (kJ/kg)

11448.20

7052.19

7256.17

15354.87

10277.85

Density (kg/cum)

852.47

1344.62

744.00

913.83

963.73

ΔFLmix

13.67

8.72

26.68

0.67

12.44

Pressure (atm)

1

1

1

1

1.00

PRI = (10277.85 × 963.73 × 12.44 × 1)/108 = 1.232

Table 9.10

1

IR event table for DWC-MVR. Events

DWC-MVRa)

Fi

IR

BLEVE

5.58 × 10−7

0.00000575

3.21 × 10−12

2

UVCE

0

0.00000776

0

3

Flash Fire (I)b)

0

0.00000776

0

0

0.000248

0

0

0.0000376

0

0

0.00000155

0

0

0.0000826

0

4

Flash Fire

5

Jet Fire

Ib) (I)b)

6

Toxic Release

7

Toxic Release (C)b)

Total

3.21 × 10−12

a) Including reboiler, condenser, and reflux drum. b) Here, (I) and (C) are for instantaneous and continuous releases.

reboiler duty in comparison to CSC system, however, only 19.41% reduction in the condenser duty is observed, which is lower than what DWC system offers (25% compared to CSC). These results are in agreement with those found by Shrikhande et al. (2021). This can be explained by the fact that the streams entering the condenser (COND in Figure 9.3) are at different temperatures (154 ∘ C from valve and 63 ∘ C for the stream coming from the vapor splitter) and have a lot of thermal energy. Table 9.10 shows the IR event table for DWC–MVR. There is one key assumption for IR calculation for DWC–MVR system. In contrast to the CSC and DWC systems, the DWC–MVR system employs two additional heat exchangers apart from the reboiler in the column. To evaluate their effect together, capacity of the reboiler is assumed as 10 minutes of feed flowrate (compared to 6 minutes for the reboilers in CSC and DWC systems). This assumption resulted in instantaneous release (Q* ) of 30773.4 kg (compared to 24979.09 kg for CSC, and 25182.4 kg for DWC) and consequently IR of 3.21 × 10−12 (compared to 1.833 × 10−14 for CSC and 1.025 × 10−12 for DWC).

9.5 Results and Discussion

Table 9.11

DI calculation for DWC-MVR. FEDI

TDI

DI

Column

104.75

34

[(15243.12)2 + (609)2 ]1/2 = 15255.28

RD

32.63

17

HX-1

452.33

89

HX-2

6836.87

100

COMP-1

452.33

89

COMP-2

6836.87

100

Cond

167.23

50

Reb

360.11

130

Total

15243.12

609

Table 9.11 shows the DI for DWC–MVR system. The column, being the same as in DWC system, has the same FEDI and TDI. Despite having lower heating and cooling utility requirements, the condenser and reboiler have a higher FEDI and TDI because of their operating conditions. As shown in the DWC–MVR system (Figure 9.3), there are two inlet streams to the condenser: one at 63 ∘ C coming from the vapor splitter and another at 154 ∘ C coming from HX-2. The higher inlet stream temperature increases the risk associated with the condenser operation, and therefore has a higher FEDI and TDI. HX-2 and COMP-2, which operate at extreme conditions (for instance, COMP-2 outlet is at 12.25 atm and 257∘ C) contribute significantly to the DI. Consequently, DI for DWC–MVR system is 15255.28, which is considerably higher than that for CSC and DWC. Table 9.12 shows that DWC–MVR system has a PRI of 1.273, which is 3.92% higher than that of CSC system and 3.32% higher than that of DWC system. The average pressure for DWC–MVR has increased to 2.25 due to the presence of multiple high-pressure streams (for example, those leaving compressors and flowing through heat exchangers), while average density has reduced due to several vapor streams. Another interesting observation is that the average heating value of DWC–MVR system is significantly lower than that of CSC and DWC systems (7567.28 vs. 10453.93 kJ/kg and 10277.85 kJ/kg).

9.5.4

Comparative Analysis

Table 9.13 summarizes the three distillation systems and their safety/risk based on different indices. IR for DWC (i.e. 1.025 × 10−12 ) and DWC–MVR (i.e. 3.21 × 10−12 ) is significantly greater than that for CSC (i.e. 1.833 × 10−14 ). These IR values show that DWC and DWC–MVR systems pose significantly greater risk compared to CSC. Interestingly, DI is the lowest for DWC (i.e. 331.44) while it is the highest for DWC-MVR (i.e. 15255.28), suggesting that DWC is a safer alternative as

255

Table 9.12

PRI Calculation for DWC-MVR system.

Feed

Bot-1

Bot-2

HX1-CI

HX1-HI

HX1-HO

HX2-CI

HX2-CO

HX2-HI

Dist-2

TO-COMP1

TO-COND

Average of Streams

Mass Heating Value (kJ/kg)

11448.24

7054.27

15360.71

7054.27

5993.28

6032.42

7039.95

6970.12

5828.29

7252.66

6160.63

6160.63

7567.28

Density (kg/cum)

852.47

1344.34

913.97

1344.34

3.14

3.31

114.58

21.98

9.18

743.95

1.17

1.17

411.89

ΔFLmix

13.67

8.73

0.65

8.73

26.70

26.70

8.73

8.73

26.70

26.70

26.70

26.70

18.16

Pressure (atm)

1.00

1.00

1.00

1.00

3.50

3.50

1.00

100

12.25

1.00

1.00

1.00

2.25

PRI = (7567.28 × 411.89 × 18.16 × 2.25) = 1.273

9.6 Survey of Engineers and Discussion of their Responses

Table 9.13 IR, DI, and PRI for CSC, DWC, and DWC-MVR systems (values in brackets is the % difference compared to CSC system). Quantity/Case

CSC

DWC

DWC-MVR

IR

1.833 × 10−14

1.025 × 10−12 (5492)

3.21 × 10−12 (17412)

DI

514.74

331.44 (−36)

15255.28 (2864)

PRI

1.225

1.232 (0.57)

1.273 (3.92)

compared to the other alternatives. PRI for CSC, DWC, and DWC–MVR is 1.225, 1.232, and 1.273, respectively, indicating that CSC is slightly safer than DWC and DWC–MVR is the most unsafe. Thus, values of both IR and PRI indicate that CSC is the safest alternative, whereas DI values indicate that DWC is the safest alternative. In summary, values/inconsistencies of different safety indices are due to the individual attributes/differences in both the processes and safety indices. IR calculates the probability of the death/injury due to some fatal scenarios that are likely to happen in a chemical process plant. On the other hand, DI calculates the damage radius with a 100% death probability within that radius. So, as these indices investigate different aspects of process safety, their values differ for alternative systems having different number of equipment and operating conditions. PRI assesses process safety solely on the basis of stream data (heating value, pressure, temperature, and density), unlike IR and DI, where inventory also has a role to play. As we move from CSC system to the DWC system, the number of streams reduces, whereas it increases from DWC to DWC-MVR. Also, the operating conditions of certain streams are severe (e.g. the stream leaving the compressors, comp-1 and comp-2, have both high pressure and temperature), which results in higher PRI for DWC-MVR when compared to the DWC and base case.

9.6 Survey of Engineers and Discussion of their Responses In the previous sections of this chapter, safety of the three distillation processes, namely, CSC, DWC, and DWC–MVR systems for the separation of methanol, water, glycerol, and BBAIL catalyst, was assessed using the selected safety indices from the literature. In this section, safety of these three processes is assessed from the perspective of practitioners. For this, we sought the views of experienced chemical engineers known to us through a short survey, which was voluntary and without any compensation. Eight of them responded with their inputs and comments. We are grateful to all the respondents for sparing their valuable time to participate in our survey. We acknowledge half of them at the end of this chapter, as others preferred to remain anonymous. Each of the eight respondents has at least 10 years

257

9 Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices

of experience in manufacturing and/or design/engineering of chemical and related processes, and a few of them have more than 30 years of experience. Question 1 in the Survey In the survey, the three systems are outlined as three cases, their process flowsheets (e.g. Figures 9.1, 9.2, and 9.3) are given, and then two multiple-choice questions are presented. The respondent is requested to respond to each of them and to provide justification/reasons for their choice. The first question asks the respondent to rank the three processes: CSC, DWC, and DWC–MVR, based on their safety/risk compared to each other. In this relative ranking, the respondent is asked to choose only one process for each of the high, intermediate, and low-risk processes. Relative ranking of the three processes by the eight respondents is summarized in Figure 9.6. All the eight respondents consider that DWC-MVR has the highest risk. Their reasons are as follows. ●

●

●

●

●

● ●

DWC-MVR with compressors is expensive to install and maintain. Typically, compressors do not have a spare/standby due to cost considerations. So, when they fail, the whole unit operation stops or worse there can be unsafe incidents. Compressor seals can leak, release flammable material into the atmosphere and cause fire and/or explosions. Liquid entry into compressors can cause mechanical damage (e.g. mechanical explosion of compressor), especially if reciprocating compressors are used. This can also cause fire and/or explosions. To prevent liquid entry into compressors, liquid knockout separators need to be installed at the suction of each compressor, which increases the number of equipment and inventory of hazardous materials. Here, DWC-MVR uses three reboilers (Figure 9.3), which significantly increase pressure relief required for DWC in cases of reboiler tube rupture and excessive heat input to the DWC. Any breakdown or malfunction of the compressors and heat exchangers can affect the safety parameters of the whole process. A rotating machine in the process always creates higher risk. Vapor compression increases the pressure–volume energy of the stream, which is significant compared to the others. Figure 9.6 Ranking of CSC, DWC, and DWC-MVR for their safety/risk by eight experienced engineers.

8 Number of respondents

258

6 4 2 0 CSC

DWC DWC-MVR

Highest risk

Medium risk

Lowest risk

9.6 Survey of Engineers and Discussion of their Responses ●

DWC–MVR has more equipment and flanges, larger inventory of hazardous fluids, and higher pressure. Hence, it will have more leak points, higher release volume (of liquid and vapor), and exposure to personnel.

Despite the above valid reasons, MVR is employed in industries, particularly when the energy efficiency of processes, including some distillation columns, can be improved significantly. For example, see Table 1.1 in Reddy and Rangaiah (2022). As shown in Figure 9.6, it is interesting that half of the respondents chose CSC as having the lowest risk, whereas the other half of respondents assessed DWC as having the lowest risk. Recall that DWC is a relatively new technology that has been adopted in the last several decades, whereas CSC has been employed for many decades. Let us study the reasons stated by the respondents for their choice of the process with the lowest risk. Reasons given by the respondents favoring CSC over DWC are as follows: ●

●

●

CSC is the conventional one and will be safer. However, it will contain some amount of risk during intermediate product movement from one column to another, such as leak, contamination, and fire. On the other hand, DWC in a single column enhances the chances of leakage and thereby mixing of incompatible chemicals in certain cases (not specific to this study), which may lead to untoward incidents. From a maintenance and reliability perspective, a simple unit is easier to operate, support, and train operators. Unless there is a big on-going efficiency incentive, the simplest configuration is preferred. DWC is okay until it works. When products go off-spec, it can be very challenging to troubleshoot. So, CSC is the best from a practical perspective. The respondents, who gave the lowest risk for DWC, stated the following reasons:

●

●

●

DWC with a minimum number of columns, reboilers, and condensers will have a minimum number of flange connections as well have a minimum inventory. Hence, this option will result in the minimum safety risks compared to the other two options. DWC has effectively reduced the number of pieces of equipment. Each piece of equipment means more potential for failure, leakages (flanges), safety equipment, etc. So, this is good from the safety point of view. Reboiler and condenser duties of DWC are lower than those for CSC.

Given the above reasons for CSC or DWC, it is not surprising that DWC is now finding increasing acceptance and applications for separating mixtures of three or more components (e.g. see Olujic et al., 2009). Question 2 in the Survey The purpose of the second question in the survey is to learn about the acceptance and use of IR, DI, and PRI described and employed in the earlier sections of this chapter. Before the question, we provided the following background: We are assessing the safety/risk potentials of the three processes using three indices, namely, individual risk (IR, which indicates the probability of a person dying in the plant due to any

259

260

9 Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices

incident), damage index (DI, which depicts the risk ranking of the process units in terms of a damage radius in case of an accident), and process route index (PRI, which estimates hazard depending on the process/synthesis route and uses streams instead of units/equipment). These indices are available and used in academic research. The second question itself is as follows: which among IR, DI, and/or PRI, in your opinion, would represent the practical scenario of the safety risks associated with the three cases? You can choose one or more of IR, DI, and PRI. Please give an explanation or reasons for your reply. The choices in this question are IR, DI, PRI, none of the above, and not aware of IR, DI, and PRI. The choices of the eight experienced engineers are presented in Figure 9.7. No one chose the last option of not being aware of IR, DI, and PRI, which indicates that the respondents are aware of these indices. Further, no one chose more than one option although, it is allowed. As shown in Figure 9.7, out of the eight respondents, IR, DI, PRI, and none of the above were chosen by 1, 3, 3, and 1 respondents, respectively. Before discussing, let us review the reasons given below. ●

●

●

●

●

●

From a practical and most representative (based on historical and judgmental frequencies) approach, DI and PRI can be selected with an understanding that an individual fatality is remote; also, probability of operator’s presence, duration, and direction of exposure are limited. There are always IRs, and these are usually mitigated through design, engineering controls, procedures, etc. DI is perhaps more applicable here. The temperatures are low, and pressure is atmospheric. Damage due to overpressure is somewhat low. What is more plausible is for a leak/vapor from a pool of hydrocarbons to find an ignition source and catch fire, leading to an explosion. This damage could be significant. Both IR and DI are good indicators of risk. IR is risk to a particular individual, while DI will provide societal risk, accounting for number of people affected. To find out how an incident/accident would affect manpower or surrounding population, generally IR and DI are used. However, for evaluating the safety hazards and risks associated with different processes, PRI will be better as it evaluates the whole process including the raw materials and the methods employed, for risk estimation. PRI assesses the safety level of the process by accounting for average values of stream parameters such as density, pressure, energy, and combustibility. Inherent safety of the process increases as PRI value decreases. Higher the inherent safety of the process, lower will be the chance of fatality or asset damage. Being low-pressure operations involving mainly methanol and water, these three processes do not have high risks.

The above reasons given by the respondents provide insight on their respective choices for IR, DI, and/or PRI. It seems the respondents’ choice is influenced by the three processes under consideration and the purpose of safety analysis. For example,

9.6 Survey of Engineers and Discussion of their Responses

12.5%

0.0%

12.5%

IR DI PRI 37.5%

37.5%

None of the above Not aware of IR, DI, PRI

Figure 9.7 Responses from the eight experienced engineers on the applicability of IR, DI, and/or PRI in the practical scenario of safety risks.

the last of the above reasons is probably the reason for the choice of none of the above by one respondent. As discussed in Section 9.5.4, the IR for DWC (i.e. 1.025 × 10−12 ) and DWC–MVR (i.e. 3.21 × 10−12 ) is significantly greater than that for CSC (i.e. 1.833 × 10−14 ), suggesting greater risk with the former alternatives. This is in partial agreement with the experts’ opinions, as half of the experts suggested the CSC and other half suggested the DWC as the safest among the alternatives, while all of them opined that DWC–MVR is the riskiest. DI is the lowest for DWC (i.e. 331.44) while it is the highest for DWC-MVR (i.e. 15255.28), suggesting that DWC is the safest/least damaging as compared to the other alternatives. Findings for DI are in agreement with what the expert engineers have suggested, viz. DWC-MVR is riskier. PRI for CSC, DWC, and DWC–MVR is 1.225, 1.232, and 1.273, indicating that CSC is safer, DWC is slightly riskier (compared to CSC), and DWC-MVR is the most unsafe alternative. All three indices (i.e. IR, DI, and PRI) suggest that DWC–MVR is the riskiest among CSC, DWC, and DWC-MVR systems. However, IR and PRI suggest that CSC is the safest alternative, whereas DI suggests that DWC is the safest alternative. It is important to note that both IR and DI calculations are complicated and are based on inventory, properties of materials, and operating conditions. On the other hand, calculations for PRI are simple (using the process simulation results) and are based on the influence of temperature and pressure on explosiveness without taking into account the material inventory/equipment size. Calculations of both IR and DI are complicated; further, there are difficulties in using IR and DI for quantitative risk analysis in the design phase (Yang et al., 2023); on the other hand, PRI does not consider the material inventory. Hence, we propose the following simple index, termed improved PRI (iPRI) by incorporating suitable modification in PRI for evaluating process safety considering both inventory and stream conditions.

261

262

9 Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices

9.7 Improved PRI The reasons given for the response to the first question in the survey bring out possible limitations of PRI. The quantities considered in this index are the average values of stream parameters, namely, density, pressure, energy, and combustibility. Thus, PRI does not consider the number of equipment with inventory and interconnections among them. As noted in the respondents’ reasons, increased number of equipment leads to more flanges and leakage points. Hence, without considering number of equipment, use of PRI has a limitation for comparing the alternative processes, which are unlikely to have identical number/type of equipment, or risky operations. In the processes simulated and employed in the earlier sections (Figures 9.1 to 9.3), equipment such as knockout drum/separator for removing the liquid before each stage of compressor is not included because they do not affect the results of process simulation. However, they are important from the point of safety. Distillation column is simulated as one unit, but it consists of the column itself, reboiler, condenser, reflux drum, and reflux pump. (Often, for reliable operation, there will be two reflux pumps, with one operating and another on standby.) There will be interconnections and flanges among these equipment. From the point of distillation simulation, many of these need not be considered, but they are important for safety analysis. Moreover, the risk associated with a set of operations (e.g. pump, atmospheric storage/intermediate tank, knockout drum, etc.) is likely to be lower than that associated with other operations (such as highly exothermic reactors, compressors, operating at high temperature/pressure, distillation/absorption columns, and high pressure/vacuum tanks). Recently, Ather et al. (2019) proposed an extended PRI that incorporates process, chemical, and equipment aspects to define the safety level of various process routes via relative ranking. Here, we propose a simple modification of PRI for comparing alternative processes; it will make the safety analysis comprehensive and increase the ability of PRI to differentiate processes being compared. It is based on the equipment factor defined as: Equipment factor =

n ∑ (pf1 × pf2)i

(9.12)

i=1

Here, n is the number of all equipment in the process; each term in the summation is the product of two penalty factors: pf1 for type of operation and pf2 for its inventory, which are crucial for safety. These penalty factors are proposed as follows: (a) pf1 = 2 for highly exothermic reactors, compressors, equipment operating at high temperature/pressure, distillation/absorption columns, and high pressure/vacuum storage tanks; and pf1 = 1 for pumps, knockout drums, and tanks with moderate conditions; and (b) pf2 = 2 for inventory >10 tons and pf2 = 1 for inventory ≤10 tons. The decision to set pf2 depending on the inventory is based on the existing (DI calculation, where a penalty factor pn3 is determined only if the chemical inventory is greater than 10 tons; no penalty is considered for inventory less than 10 tons (Khan and Amyotte, 2004).

Acknowledgments

Equipment ratio of a process alternative (case) is then determined using: Equipment ratio =

(Equipment factor)for any case (Equipment factor)for the case with minimum equipment factor (9.13)

Finally, multiply PRI by the equipment ratio to obtain iPRI of a process alternative: iPRI = equipment ratio × PRI

(9.14)

Calculations of iPRI for CSC, DWC, and DWC-MVR systems are presented in Table 9.14, which shows that iPRI for CSC, DWC, and DWC-MVR is 2.07, 1.232, and 2.253, respectively. The ranking of the three processes (i.e. DWC is the safest followed by CSC and DWC–MVR) by the proposed index, i.e. iPRI, agrees with the ranking by DI (Table 9.14). On the other hand, IR and PRI assess CSC as the safest followed by DWC and DWC–MVR. As shown in Figure 9.6, survey respondents are equally divided, with half choosing CSC as having the lowest risk and the other half selecting DWC as having the lowest risk. These findings are due to the different scope/principle of indices and varied opinions of experts, and they indicate the need for more in-depth safety analysis of CSC and DWC systems.

9.8 Conclusions This chapter examines the safety of three process alternatives, namely CSC, DWC, and DWC-MVR, using different safety indices: IR, DI, and PRI. Subsequently, safety of the three processes is assessed from the perspective of practitioners by conducting a short survey. Applicability and limitations of the existing safety indices are discussed, and a suitable modification to PRI is then proposed to make it comprehensive and useful for analyzing and comparing the safety of alternative processes, both intensified and conventional. All four indices: IR, DI, PRI, and iPRI suggest that DWC-MVR has the highest risk associated with it, which is consistent with the experts’ assessment. IR and PRI values indicate CSC is the safest, whereas DI and iPRI values indicate that DWC has the least risk, which is similar to experts’ views (i.e. half of them opine that CSC has the least risk and other half think that DWC has the least risk). These findings suggest the need for further assessment of safety of CSC and DWC systems. In general, experts suggested that both DI and PRI are more applicable for depicting the safety/risk of processes. In this study, the drawback of PRI is eliminated by incorporating an equipment factor, resulting in iPRI, which is comprehensive despite being simple to calculate. Hence, iPRI has potential for realistic assessment of safety of chemical processes.

Acknowledgments We are grateful to the eight experienced engineers for their kind participation in the safety survey and for providing their inputs summarized in Section 9.5. Their

263

264

9 Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices

Table 9.14 IR, DI, PRI, and iPRI of CSC, DWC, and DWC-MVR; bold face identifies the best value for each index.

IR Process [A]

DI [B]

PRI [C]

Number of equipment [D]

Equipment factor [E]

Equipment Ratio [F]

iPRI [G] = [C] × [F]

1.69

2.070

CSC

1.833 × 514.74 10−14

1.225

2 columns, 2 reboilers, 2 condensers, 2 reflux drums, and 2 reflux pumps Total: 10

(2 × 2 × 2 + 2 ×2×1+2× 2×1+2×1× 2 + 2 × 1 × 1) = 22

DWC

1.025 × 331.44 10−12

1.232

1 column, 1 reboiler, 1 condenser, 1 reflux drum, and 1 reflux pump Total: 5

(1 × 2 × 2 + 1 1 ×2×2+1× 2×1+1×1 × 2 + 1 × 1 × 1) = 13

DWCMVR

3.21 × 10−12

1 column, 1 reboiler, 1 condenser, 1 reflux drum, 1 reflux pump, 2 knock-out drums, 2 compressor stages, and 2 additional reboilers Total: 11

(1 × 2 × 2 +1×2×2+ 1×2×1+1 ×1×2+1× 1×1+2×1× 1+2×2×1 + 2 × 2 × 1) = 23

15255.28 1.273

1.77

1.232

2.253

responses, based on their expertise and practical knowledge, to the questions in the survey brought interesting and important insights pertaining to the safety indices. We are pleased to acknowledge the following respondents to the survey: Satendra Singh, CCS Reddy, Anwesh Das, and Balajee Raman, as well as the other four respondents who preferred to remain anonymous.

References Ahmad, S.I., Hashim, H., and Hassim, M.H. (2014). Numerical descriptive inherent safety technique (NuDIST) for inherent safety assessment in petrochemical industry. Process Safety and Environmental Protection 92 (5): 379–389. AIChE (2000). Guidelines for Chemical Process Quantitative Risk Analysis, 2e. AIChE ISBN: 978-0-8169-0720-5.

References

AIChE & Dow Chemical Company (1994). Fire & Explosion Index: Hazard Classification Guide. American Institute of Chemical Engineers https://www.aiche.org/resources/ publications/books/dows-fire-explosion-index-hazard-classification-guide-7thedition; accessed July 2023. Argoti, A., Orjuela, A., and Narváez, P.C. (2019). Challenges and opportunities in assessing sustainability during chemical process design. Current Opinion in Chemical Engineering 26: 96–103. Ather, M., Shariff, A.M., Buang, A., and Hermansyah, H. (2019). Equipment-based route index of inherent safety. Process Safety Progress, AIChE https://doi.org/10.1002/ prs.12108. Bruinsma, O.S.L. and Spoelstra, S. (2010). Heat pumps in distillation. In Presented at the Distillation & Absorption Conference, vol. 12, p. 15. Contreras-Zarazúa, G., Sánchez-Ramírez, E., Vázquez-Castillo, J.A. et al. (2019). Inherently safer design and optimization of intensified separation processes for furfural production. Industrial and Engineering Chemistry Research 58: 6105–6120. Crowl, D.A. and Louvar, J.F. (2002). Chemical process safety fundamental with application, 2e. Prentice Hall International Series. ´ I., Matijaševic, ´ L., and Olujic, ´ Ž. (2010). Dividing wall column—a Dejanovic, breakthrough towards sustainable distilling. Chemical Engineering and Processing: Process Intensification 49 (6): 559–580. Deshpande, G., Shrikhande, S., Sawarkar, A.N., and Patle, D.S. (2022a). Multiobjective optimization of ultrasound intensified and ionic liquid catalyzed in situ algal biodiesel production considering economic, environmental and safety indicators. Chemical Engineering Research and Design 180: 134–152. Deshpande, G., Shrikhande, S., Patle, D.S., and Sawarkar, A.N. (2022b). Simultaneous optimization of economic, environmental and safety criteria for algal biodiesel process retrofitted using dividing wall column and multistage vapor recompression. Process Safety and Environmental Protection 164: 1–14. Gao, X., Raman, A.A.A., Hizaddin, H.F. et al. (2021). Review on the inherently safer design for chemical processes: past, present and future. Journal of Cleaner Production 305: 127154. De Haag, P.U. and Ale, B.J.M. (2005). Guidelines for quantitative risk assessment. Netherlands: Committee for the Prevention of Disasters https://content .publicatiereeksgevaarlijkestoffen.nl/documents/PGS3/PGS3-1999-v0.1quantitative-risk-assessment.pdf (accessed January 2023). Heikkilä, A.M. (1999). Inherent Safety in Process Plant Design: An Index-based Approach. VTT Technical Research Centre of Finland. Khan, F.I. and Amyotte, P.R. (2004). Integrated inherent safety index (I2SI): a tool for inherent safety evaluation. Process Safety Progress 23 (2): 136–148. Khan, F.I., Husain, T., and Abbasi, S.A. (2001). Safety weighted hazard index (SWeHI) a new, user-friendly too; for swift yet comprehensive hazard identification and safety evaluation in chemical process industries. Transactions of the Institution of Chemical Engineers 79: 65–80. Kletz, T.A. and Amyotte, P. (2010). Process Plants: A Handbook for Inherently Safer Design. CRC Press.

265

266

9 Safety Analysis of Intensified Distillation Processes Using Existing and Modified Safety Indices

Koller, G., Fischer, U., and Hungerbühler, K. (2000). Assessing safety, health, and environmental impact early during process development. Industrial & Engineering Chemistry Research 39 (4): 960–972. Leong, C.T. and Shariff, A.M. (2009). Process route index (PRI) to assess level of explosiveness for inherent safety quantification. Journal of Loss Prevention in the Process Industries 22 (2): 216–221. Luyben, W.L. (2002). Plantwide Dynamic Simulators in Chemical Processing and Control. New York: Marcel Dekker Chapter 3. Medina-Herrera, N., Jiménez-Gutiérrez, A., and Mannan, M.S. (2014). Development of inherently safer distillation systems. Journal of Loss Prevention in the Process Industries 29: 225–239. Olujic, O., Jodecke, M., Shilkin, A. et al. (2009). Equipment improvement trends in distillation. Chemical Engineering and Processing: Process Intensification 48: 1089–1104. Palaniappan, C., Srinivasan, R., and Tan, R. (2002a). Expert system for the design of inherently safer processes. 1. Route selection stage. Industrial & Engineering Chemistry Research 41 (26): 6698–6710. Palaniappan, C., Srinivasan, R., and Tan, R.B. (2002b). Expert system for the design of inherently safer processes. 2. Flowsheet development stage. Industrial & Engineering Chemistry Research 41 (26): 6711–6722. Premkumar, R. and Rangaiah, G.P. (2009). Retrofitting conventional column systems to dividing-wall columns. Chemical Engineering Research and Design 87: 47–60. Reddy, C.C.S. and Rangaiah, G.P. (2022). Waste Heat Recovery: Principles and Industrial Applications. World Scientific. Segovia-Hernandez, J.G., Sanchez-Ramirez, E., Flores-Cordero, E. et al. (2021). Safety and control analysis of hybrid liquid-liquid extraction and divided wall column for biobutanol purification. Chemical Engineering Transactions 86: 1021–1026. Shariff, A.M. and Zaini, D. (2010). Toxic release consequence analysis tool (TORCAT) for inherently safer design plant. Journal of Hazardous Materials 182 (1–3): 394–402. Shariff, A.M., Leong, C.T., and Zaini, D. (2012). Using process stream index (PSI) to assess inherent safety level during preliminary design stage. Safety Science 50 (4): 1098–1103. Sharma, S., Chao Lim, Z., and Rangaiah, G.P. (2013). Process design for economic, environmental and safety objectives with an application to the cumene process. In: Multi-Objective Optimization in Chemical Engineering: Developments and Applications, 449–477. Wiley. Shrikhande, S., Deshpande, G., Sawarkar, A.N. et al. (2021). Design and retrofitting of ultrasound intensified and ionic liquid catalyzed in situ algal biodiesel production. Chemical Engineering Research and Design 171: 168–185. Sun, S., Fu, L., Yang, A., and Shen, W. (2023). An intensified energy-saving architecture for side-stream extractive distillation of four-azeotrope mixtures considering economic, environmental and safety criteria simultaneously. Separation and Purification Technology 310: 123132. Tyler, B.J. (1985). Using the Mond Index to measure inherent hazards. Plant/Operations Progress 4 (3): 172–175.

References

Vazquez-Castillo, J.A., Contreras-Zarazua, G., and Segovia-Hernandez, J.G. (2019). Heat-integrated reactive distillation processes to produce ethyl levulinate: design and optimization including environmental, safety and economics aspects. In: Computer Aided Chemical Engineering, vol. 46, 355–360. Elsevier. Yang, A., Ernawati, L., Wang, M. et al. (2023). Multi-objective optimization of the intensified extractive distillation with side-reboiler for the recovery of ethyl acetate and methanol from wastewater. Separation and Purification Technology 310: 123131. Zhai, J., Sun, X., Huang, S. et al. (2023). Economic, thermodynamic, environmental, and inherent safety investigation of heat pump-assisted extractive dividing wall column for separating binary azeotrope. Process Safety and Environmental Protection 173: 202–214.

267

269

10 Dynamic Safety Analysis of Intensified Extractive Distillation Processes with Independent Protection Layers Chengtian Cui 1 and Meng Qi 2 1

Davidson School of Chemical Engineering, Purdue University, West Lafayette, IN 47907, United States College of Chemistry and Chemical Engineering, China University of Petroleum (East China), Qingdao, 266580, P.R. China 2

10.1 Introduction Ensuring safety in industries that deal with hazardous substances like oil and gas, chemicals, and petrochemicals is crucial for their sustainable development (Khan et al., 2015). Safety is a fundamental consideration throughout the entire life cycle of a process in such industries, from conception and design to decommissioning. Nonetheless, the design stage provides the most favorable opportunity to create a safer chemical process, as it presents more options and is cost-effective to implement safety measures (Hendershot, 1997). To design a safer chemical process, safety performance must be assessed meticulously to identify any potential hazards and risks (Ortiz-Espinoza et al., 2017). This entails examining the process from various perspectives to assess the likelihood and severity of accidents or incidents that may occur. Assessing the safety level of a chemical process during the conceptual design stage is a challenging task due to the process’s dynamic nature and the limited information available for hazard and risk evaluation. The safety features of a process involve intricate interactions between the process, environment, protective devices, and human factors (Cui et al., 2023). Moreover, modern chemical processes have become increasingly complex and intensified, with numerous connections established between different types of unit operations. This complexity makes conventional safety methods, which strongly rely on expert experience and process understanding, inadequate for precisely assessing process hazards or hazardous situations (Leveson and Stephanopoulos, 2014). To address this challenge, process simulation, a widely used technique for chemical process design and optimization, can provide valuable information for process safety analysis and aid in understanding the behavior of these complicated, highly integrated, or intensified processes

Control and Safety Analysis of Intensified Chemical Processes, First Edition. Edited by Dipesh Shikchand Patle and Gade Pandu Rangaiah. © 2024 WILEY-VCH GmbH. Published 2024 by WILEY-VCH GmbH.

270

10 Dynamic Safety Analysis of Intensified Extractive Distillation Processes

(Kummer and Varga, 2019). Dynamic process simulation, in particular, can consider these interactions and accurately simulate complex dynamic behaviors, making it highly beneficial for evaluating process safety during the early design stage. A dynamic simulation can probe the actual process response under a hazardous scenario with the consideration of independent protections, such as the basic process control system (BPCS), critical alarms and human intervention (CA&HI), safety instrumented systems (SISs), and pressure safety valves (PSVs). In summary, dynamic process simulation has found numerous applications in process safety analysis, including hazard and operability study (HAZOP) (Janošovský et al., 2019), risk assessment (Berdouzi et al., 2018), layer of protection analysis (LOPA) (Cui et al., 2023), process deviation propagation analysis (Carlos et al., 2018), and human reliability quantification (Zhu et al., 2020). Engineering professionals are increasingly utilizing process simulation to improve standard safety analysis and hazard identification methods. However, the safety engineering community has not fully embraced this approach due to concerns over the validity of the simulation model. The accuracy of safety analysis and hazard identification relies heavily on the quality of the simulation model as well as the appropriateness of its settings and parameters. This is especially crucial when considering the effectiveness of various independent protections in preventing accidents under different hazardous scenarios (Labovská et al., 2014). While dynamic simulation has been employed for process safety analysis, there have been limited studies that analyze the impact of process control and other independent protections and their effectiveness in accident prevention. This chapter aims to address this gap by incorporating these protections into dynamic simulations. A scenario-based methodology is presented that assesses process safety performance and determines the associated safety time, considering different independent protections. A process with several intensified variants is studied to demonstrate the applicability of the methodology in capturing complex process dynamics and evaluating safety performance. Through this chapter, readers will learn how to enhance the accuracy of safety analysis by incorporating independent protections into dynamic simulations and assessing safety performance in complex processes. The chapter presents an exploration of process intensification measures and dynamic safety analysis applied to extractive distillation (ED) with preconcentration. A simple process-synthesis approach is proposed to integrate these measures into the conventional ED process. To obtain accurate process data for dynamic control and safety analysis, a steady-state design simulation is conducted using Aspen Plus software (Cui et al., 2023). Subsequently, Aspen Dynamics is employed to analyze the dynamics and control strategies of the intensified ED processes, and several control structures with temperature controllers are proposed to effectively manage disturbances in throughput and feed composition. These control structures can serve as the BPCS layer of protection during safety analysis. Finally, a safety

10.2 Preliminary

analysis is performed based on the LOPA theory using the developed dynamic models, taking into account independent protections such as the BPCS, CA&HI, SISs, and PSVs.

10.2 Preliminary In this chapter, scenario-based safety analysis is performed through a dynamic simulation of a specific hazardous scenario. The scenario-based safety analysis requires the theory of LOPA to assess the consequences of a hazardous scenario by considering the process design and the effectiveness of independent protections. The basic premise of LOPA is that multiple independent protection layers (IPLs) can be implemented to prevent a hazardous event from occurring or to mitigate its consequences. These layers of protection can be physical, procedural, or human-based and are designed to act independently or in combination to prevent accidents. Typical layers of protection against a possible accident are shown in Figure 10.1. If a process deviation occurs, protection layers act as barriers to prevent the deviation from escalating into a major accident. Specifically, the BPCS ensures that the process is operated within safe limits by controlling critical process variables such as temperature, pressure, and flow. It helps to correct process deviations at early stages. Alarms can be set up to trigger when certain process variables exceed safe limits or when a certain threshold is exceeded. Alarms can provide early warning to operators; therefore operators who are properly trained and equipped with the right tools can identify and correct process deviations manually. SISs are designed to take automatic corrective action once the above protection layers fail. SISs, including sensors, logic solvers, and final control elements, are activated to shut down the process when the process variables reach their activation limit. If the SIS fails to act, PSVs designed to release pressure are open to prevent equipment from overpressure

Community emergency response Plant emergency response Post-release physical protection (Dikes)

Automatic action SIS or ESD Critical alarms, human intervention Basic process control systems (BPCSs) Process design

Figure 10.1

Accident Process parameter (e.g. temperature, pressure)

Physical protection (Relief devices)

0

Equipment endurance limit

Safety valve for relief/venting SIS for process shutdown

SIS set value

Alarm for operators Normal operation Process design

BPCS

Initiating event

Alarm set value Critical alarms and human intervention Time

SIS

Physical protection

Normal process parameter

Accident

Typical layers of protection against a possible accident in chemical processes.

271

272

10 Dynamic Safety Analysis of Intensified Extractive Distillation Processes

and potential damage or rupture. By implementing multiple layers of protection and ensuring they are properly designed, installed, and maintained, the safety risk of a specific hazardous event in the process can be comprehensively managed to prevent accidents and improve process safety performance. Process dynamic simulation with the consideration of layers of protection is critical to rigorous safety analysis of intensified processes combining specific hazardous events. Specifically, this work considers column operating pressure as the key process parameter that affects the overall process safety.

10.3 Process Studied 10.3.1 Process Intensification Measures Typically, a common arrangement for ED involves two columns: an extractive distillation column (EDC) and a solvent recovery column (SRC). This setup is commonly used in most cases of ED (Luyben and Chien, 2011). However, in situations where the goal is to separate binary minimum-boiling azeotropes and the composition of the feed significantly differs from that of the azeotrope, it might be advantageous to include a preconcentration distillation column (PDC). The primary purpose of the PDC is to minimize the amount of entrainer that must be recycled, thereby reducing energy consumption (Luyben, 2006). Figure 10.2 visually represents this process. To improve the efficiency of the three-column ED system and decrease overall energy expenditures, process intensification techniques can be employed. These methods, as described by Cui et al. (2022) and Zhang et al. (2020) and depicted in Figure 10.3, involve the implementation of various process intensification measures.

A

S

Figure 10.2 Distillation configuration: (a) two-column ED and (b) three-column ED system.

B

AB S

BS (a)

EDC

SRC

A

S

B

AB Az.AB

(b)

PDC

S

BS

B EDC

SRC

10.3 Process Studied

S

1

A

3

B

5

AB

S

A

3

5

B

2

AB Az.AB 4

2 #1

6 BS

B

Az.AB 4

Step 1 S

#2

6 BS

B

S

Step 2 AB

S 2

B

Az.AB 5

BS

3

A

AB

4 Step 3

6 #4

Figure 10.3

S 2

B #3

S

Az.AB

3

A

4

5 BS

6 S

Process intensification steps for a three-column ED system.

The fundamental process for generating intensified processes is process #1. The subsequent process intensification measures can be utilized: Step 1: Transformation of the primary PDC into a stripping column (process #1 to process #2) Researchers have discovered that rectifying Section 10.1 has a relatively low reflux ratio and stage number (Arifin and Chien, 2007; Chang et al., 2012; You et al., 2018). Therefore, substituting the PDC with a stripping column would not significantly compromise the purification efficiency. Furthermore, rather than being condensed and removed, the latent energy carried by the overhead vapor can be preserved, thereby intensifying the basic process #1. However, a small compressor might be required to compensate for the pressure difference. Step 2: Combination of PDC and SRC (process #2 to process #3) The PDC bottoms and SRC distillate are identical products, allowing Sections 10.2 and 10.5 to be combined for further process intensification and energy savings. A side reboiler or condenser can be used to compensate for the energy mismatch (An et al., 2015; Liang et al., 2014). Step 3: Utilizing thermal coupling (TC) to save energy (process #3 to process #4) The reboiler can be eliminated with a (TC configuration, building upon process #3. This approach reduces the remixing effect, resulting in energy savings.

10.3.2 Steady-state Process Design The objective of this study is to explore the separation of the minimum-boiling azeotrope consisting of acetonitrile (ACN) and water, using ethylene glycol (EG) as the entrainer. The process simulation and safety analysis are conducted using Aspen Plus software. The feed mixture contains 20 mol% ACN and 80 mol% water, with a flow rate of 500 kmol/h at a temperature of 50 ∘ C. The NRTL physical property package is employed, aiming for product purities of 99.99 mol% for ACN,

273

274

10 Dynamic Safety Analysis of Intensified Extractive Distillation Processes

99.99 mol% for water, and 99.999 mol% for EG. The optimal steady-state simulation results are presented in Figure 10.4. Table 10.1 provides a comparison of four processes in terms of energy and economic costs. The simulation results reveal that the intensified processes (#2, #3, and #4) can achieve energy cost savings exceeding 30% compared to the baseline process (#1). Among these options, process #3 emerges as the most favorable from an economic standpoint. By implementing process #3, it is expected to reduce the total annual cost (TAC) by approximately 20% with a payback period of 3 years.

10.3.3 Process Intensification Analysis The temperature and liquid composition profiles depicted in Figure 10.5 offer insights into potential energy-saving opportunities in intensified processes. When an ACN-rich azeotrope liquid feed is introduced in the EDC of basic process #1, it results in a notable energy penalty due to the interruption of the decreasing ACN composition trend. Moreover, the use of a low-temperature feed causes a

Cooling water 1486 kW

0.456 bar 55.78 °C 500 kmol/hr ACN: 0.2 Water: 0.8

1C

2

80.34 °C 0.442 bar 57.11 °C

4 147.79 kmol/hr ACN: 0.6764 Water: 0.3236

8

Cooling water 1540 kW

LP steam 1671 kW 0.506 bar 81.57 °C

151.01 kmol/hr ACN: trace Water: 0.3167 EG: 0.6833

42R

LP steam 1785 kW

500 kmol/hr ACN: 0.2 Water: 0.8

8 9R

76.00 °C

Figure 10.4

RR = 0.789

355.85 kmol/hr ACN: 0.0001 Water: 0.9999

0.199 bar 60.14 °C

99.97 kmol/hr ACN: 0.9999 Water: 9.13e-5 EG: 8.69e-6

2

149.41 kmol/hr ACN: 7.69e-6 Water: 0.2957 EG: 0.7043

34

1C 44.19 kmol/hr ACN: 2.6e-5 Water: 0.9999 EG: 7.4e-5

11 12R

43R

LP steam 464 kW

Cooling water 635 kW

RR = 0.186

5

42

LP steam 1547 kW 0.403 bar

Makeup EG: 1

Cooling water 1637 kW 1C

4 144.15 kmol/hr ACN: 0.6935 Water: 0.3065

16R

60.98 °C

0.381 bar 53.15 °C 2

1

15

162.85 °C Water: 1e-5 EG: 0.99999

Cooling water 464 kW

Electrical power 88 kW

47.82 kmol/hr ACN: trace RR = 0.137 Water: 0.9999 EG: 0.0001 1C

103.187 kmol/hr MP steam 781 kW 0.329 bar ACN: trace

0.725 bar 123.30 °C

Process #2 0.346 bar 48.96 °C

2

8 34 41

352.21 kmol/hr ACN: 0.0001 Water: 0.9999

Cooling water 654 kW

0.230 bar 63.73 °C

99.97 kmol/hr 1C ACN: 0.9999 RR = 0.696 Water: 8.4e-5 EG: 1.6e-5

2

RR = 0.018

9R

Makeup EG: 1

Cooling water 399 kW

Process #1

0.672 bar 122.89 °C

MP steam 740 kW 0.270 bar

157.45 °C

105.22 kmol/hr ACN: trace Water: 1e-5 EG: 0.99999

Steady-state simulation results of extractive distillation processes.

10.3 Process Studied Makeup EG: 1

Cooling water 466 kW

Process #3

75.50 °C 0.354 bar 49.50 °C 500 kmol/hr ACN: 0.2 Water: 0.8

400.03 kmol/hr ACN: 9.638e-5 Water: 0.9999 EG: 3.62e-6

Electrical power 61 kW

1C

2

1

6 144.46 kmol/hr ACN: 0.6921 Water: 0.3079 EG: trace

0.404 bar 75.42 °C 10 LP steam 905 kW

Cooling water 1723 kW

0.304 bar 47.23 °C

RR = 0.862

99.97 kmol/hr ACN: 0.9999 Water: 0.0001 EG: trace

33 37

14 20

0.496 bar 174.60 °C

21R

144.48 kmol/hr ACN: 0.0004 Water: 0.3079 EG: 0.6917

0.559 bar 115.95 °C

38R

99.93 kmol/hr ACN: trace Water: 1e-5 EG: 0.99999

LP steam 438 kW

HP steam 911 kW Makeup EG: 1

Cooling water 567 kW

Process #4

0.304 bar 45.83 °C 500 kmol/hr ACN: 0.2 Water: 0.8

Electrical power 124 kW

142.50 kmol/hr ACN: 0.7016 Water: 0.2984 EG: trace

8

Electrical power 56 kW 19 26

0.488 bar 174.15 °C

27R

36.90 kmol/hr (TC-vap) ACN: 0.0001 Water: 0.9719 EG: 0.0279

Cooling water 1683 kW

0.402 bar 54.58 °C

2 5

1

0.354 bar 400.03 kmol/hr 72.89 °C ACN: 0.0001 Water: 0.9999 EG: trace LP steam 920 kW

52.42 °C

1C RR = 0.784

99.97 kmol/hr ACN: 0.9999 Water: 0.0001 EG: trace

40 45 180.36 kmol/hr (TC-liq ) ACN: 0.0001 Water: 0.4346 EG: 0.5653

100.93 kmol/hr ACN: trace Water: 1e-5 EG: 0.99999

HP steam 1263 kW

Figure 10.4

(Continued)

significant drop in temperature. In the stripping section near the bottom position, a common remixing phenomenon is observed, which can be eliminated through process intensification measures. To address the remixing effect in Step 1, the introduction of a vapor-phase feed proves effective. However, the majority of the energy savings in this step are attributed to the utilization of latent energy carried by the vapor-phase feed. Moving to Step 2, the combination of PDC and SRC further reduces total energy consumption by making use of the latent heat from the SRC overhead vapor. Although some

275

276

10 Dynamic Safety Analysis of Intensified Extractive Distillation Processes

Table 10.1

Comparison of four processes in terms of energy and TAC.

Process

#1

#2

#3

#4

Cooling water consumption (kW)

4079

2736

2189

2250

Steam utility consumption (kW)

4237

2751

2254

2183

Electricity consumption (kW)

0

88

61

180

Energy cost (10 US$/a)

1.0008

0.6962

0.6120

0.6755

Energy cost reduction

—

30.43%

38.85%

32.50%

TAC (10 US$/a)

1.3925

1.1071

0.9808

1.1375

TAC reduction

—

20.49%

29.57%

18.32%

6

6

remixing still occurs in the EDC stripping section, it can be eliminated through the implementation of TC. However, the impact of remixing on energy consumption is relatively small compared to the additional power required for vapor transfer in TC. In Step 3, the introduction of TC completely eliminates the undesired remixing. The EDC vapor is now transferred from the SRC bottom reboiler using high-pressure steam instead of low-pressure steam. This improvement in steam quality results in increased energy costs and diminishes the economic competitiveness of #4. Similar findings have been observed by Wu et al. (2013). Consequently, TC may not offer significant economic advantages.

10.4 Dynamics and Control In this section, the focus is on the development of BPCSs for processes #1, #2, and #3. These systems are designed to handle large ±20% disturbances in throughput and feed composition while maintaining satisfactory dynamic performance. Notably, the BPCSs will utilize inferential composition control based on temperature measurements alone. This approach is preferred in many industrial applications as it avoids the need for online composition measurements.

10.4.1 Control Basis To transition the steady-state model to dynamic mode, it is essential to determine equipment sizes. In this study, the Aspen software’s built-in tray sizing tool is utilized to calculate column diameters. The assumptions made for tray spacing and weir height are 0.6096 and 0.05 m, respectively. To determine the volume of each column base and reflux drum, a holdup time of 10 minutes is chosen, with a liquid level set at 50%. However, for the SRC column base, the base level is typically controlled by the makeup solvent flow rate, allowing it to fluctuate with changes in the solvent flow rate. To accommodate dynamic transients and provide additional surge capacity, a holdup time of 20 minutes is employed. To ensure reasonable plumbing and

10.4 Dynamics and Control 130

85

Process #1 Selected 6th stage

70 65 60 55

100 90 80 70 60

50 2

3 4 5 6 7 Stage of PDC

40

9

100 80

Crude feed 2nd stage ACN Water

0.8

0.6 0.4 0.2

1

2

3 4 5 6 7 Stage of PDC

40

5 10 15 20 25 30 35 40 Stage of EDC

8

80

Solvent feed 4th stage

ACN Water EG

Remixing effect az. AB feed 34th stage

0.4

0.0

9

Remixing caused by feed mismatch

Temperature (°C)

Selected 6th stage

65 60 55

100

1

2

3 4 5 6 7 Stage of PDC

8

Crude feed 1st stage

ACN Water

Selected 27th stage

60

Mole composition

Mole composition

0.4 0.2 0.0

0.6 0.4

2

3 4 5 6 7 Stage of PDC

Figure 10.5

8

9

0.0

6 8 10 12 14 16 Stage of SRC

80

Selected 4th stage

60 40

2

4 6 8 10 Stage of SRC

12

1.0 ACN Water EG

Solvent feed 4th stage az. AB feed 34th stage

0.2

1

4

100

5 10 15 20 25 30 35 40 Stage of EDC

0.8

0.6

2

Selected

1.0

0.8

0.2

120 7th stage

ACN Water EG

0.8 Mole composition

1.0

BS feed 8th stage

0.4

140

Selected 36th stage

80

40

9

ACN Water EG

160

50 45

6 8 10 12 14 16 Stage of SRC

0.6

0.0

5 10 15 20 25 30 35 40 Stage of EDC

120

70

4

0.8

Process #2

75

2

1.0

0.6

0.2

Selected 4th stage

60

1.0

Mole composition

Mole composition

8

Selected 11th stage

120

Mole composition

1

0.8

Temperature (°C)

140

50

1.0

0.0

Large temperature drop due to the feed mismatch

160

Temperature (°C)

45

Selected 35th stage

110

Temperature (°C)

75

Temperature (°C)

Temperature (°C)

80

180

120

0.6 0.4 BS feed 5th stage

0.2

5 10 15 20 25 30 35 40 Stage of EDC

0.0

2

4 6 8 10 Stage of SRC

12

Steady-state temperature and liquid composition profiles.

account for pump heads and control valve pressure drops, specific values are specified. For the control scheme, a conventional P-only controller is utilized for level control loops. The controller’s gain K C = 2 and integral time 𝜏 I = 9999 min. For flows, pressures, and temperatures, PI controllers are employed with different parameter settings. The controllers for flows have parameters of K C = 0.5/𝜏 I = 0.3 minutes, while the controllers for pressure have parameters of KC = 20/𝜏 I = 12 minutes.

277

10 Dynamic Safety Analysis of Intensified Extractive Distillation Processes

180

120

Process #3 Selected 16th stage

140

Temperature (°C)

Temperature (°C)

160

120 100 80

Selected 6th stage

Selected 13th stage

60

100

Mole composition

5

20

ACN Water EG

30

35

0.4

ACN Water EG

0.8

0.2

Solvent feed 6th stage

0.6

az. AB feed 33rd stage

0.4 0.2

0.0

5 10 15 Stage of combined PDC and SRC

0.0

20

5

10

15 20 25 Stage of EDC

30

35

120

180

110

Process #4 Temperature (°C)

Temperature (°C)

15 20 25 Stage of EDC

BS feed 14th stage

0.6

140 120 100 80

100

60

90 80 70 60 50

40

5

10

15

20 25 30 Stage of EDC

35

40

45

1.0

Crude feed 1st stage

ACN Water EG

0.8

40

10 15 20 25 Stage of combined PDC and SRC

Water product

BS feed 19th stage

0.6 0.4 0.2

ACN Water EG

0.8 Mole composition

5 1.0

10

1.0

Water product

Mole composition

Crude feed 1st stage

0.8

160

Selected 34th stage

60 40

5 10 15 Stage of combined PDC and SRC 1.0

Selected 27th stage

80

40

Mole composition

278

0.6

Solvent feed 5th stage az. AB feed 40th stage

0.4 0.2

0.0

0.0 5

Figure 10.5

10 15 20 25 Stage of combined PDC and SRC

10

20 30 Stage of EDC

40

(Continued)

A crucial aspect of establishing a robust control scheme is identifying the tray locations that require temperature control. This selection process is accomplished using open-loop sensitivity and the slope criterion. For each temperature controller, a 1 minute dead time is incorporated into the loop. To determine the ultimate gains and periods, relay-feedback tests are conducted. Subsequently, the K C and 𝜏 I values are determined using the Tyreus–Luyben tuning rules.

10.4 Dynamics and Control

10.4.2 BPCS #1 Process #1 comprises nine inventory variables that require control, including three pressures and six liquid levels in reflux drums and bases. Figure 10.5 illustrates the marked temperature-sensitive stages, which are selected based on the slope of the curve with the highest value. Among these stages, the 6th stage exhibits the highest temperature sensitivity in the PDC. For the PDC, a single-end control strategy is employed since only the bottom purity is strictly regulated. In the subsequent EDC, temperature control on the 35th stage is achieved solely through adjusting the reboiler duty (QR2). This approach is adopted because any variations in the reflux ratio (RR2) can significantly impact the response time of this particular temperature control point. In the case of the SRC, a dual-end strategy is implemented. The overall BPCS #1 configuration is given in Figure 10.6. The closed-loop system experiences continuous disturbances starting from 0.5 hour into the test, and the test concludes after 10 hours. The system’s responses to a 20% step change in feed flow rate are depicted in Figure 10.7a–f. The temperature controllers exhibit fast responses, effectively maintaining the target product purity close to their required specifications. Additionally, the responses for feed composition disturbances, specifically a change from 20 to 24 mol% (+20%) ACN and from 20 to 16 mol% (−20%) ACN, are shown in Figure 10.7g–l. In all of these cases, stable regulatory control is achieved even for significant composition disturbances.

10.4.3 BPCS #2 Process #2 has undergone slight modifications compared to Process #1, resulting in similar BPCS configurations. The primary distinction lies in inventory control, where the top pressure of the PDC is regulated by manipulating the compressor power. The temperature profile in Figure 10.5 reveals that the EDC exhibits a larger temperature-sensitive stage at the 27th tray, along with a smaller one at the 36th tray. These two temperature control points can be effectively managed by adjusting the reflux ratio and reboiler duty. Upon comparing the single-end and dual-end strategies, it was observed that the latter performed better, possibly due to the vapor feed flowing upward. When using reflux flow to control a temperature in TC5

Q-Sol

FT

FC FC PC

PC

Sol/F

2

2

FT

LC

RR1 TC1

4

TC2

8

FT

RR3 TC3

ACN

FC

Water

FT

BPCS of process #1.

Water

15 LC

16R

42R FT FC

Figure 10.6

FT

11

LC

9R

FT

FC FT

34 35 41

LC FC

TC4

8

RR2

FT FT

LC

4 FT

FC

6

1C

2 LC

FT

FC

PC

1C

1C

EG makeup

EG

279

10 Dynamic Safety Analysis of Intensified Extractive Distillation Processes

the upper section of the column, hydraulic lags are typically experienced as the liquid flows down the column from tray to tray, with each tray contributing a lag of 3–6 s (Luyben and Chien, 2011). Consequently, controlling a temperature that is 26 trays down the column would introduce a lag of 78–156 s into the control loop, thereby adversely affecting load rejection performance. The overall BPCS #2 is presented in Figure 10.8. Furthermore, Figure 10.9 illustrates the results obtained for ±20% throughput and composition disturbances. It is evident that, compared to other temperature control loops, the TC2 loop responds relatively slowly due to the hydraulic lag. However, the target purities generally return to their desired values with minor deviations. This case demonstrates that even a slight modification of base Process #1 can yield significant energy-saving advantages without compromising dynamic performance. The safety implications of replacing the condenser with a compressor will be explored further in the subsequent safety analysis section. +20% F

–20% F

1

0.99996

0.999

XB1, Water

0.9985

XD2, ACN XD3, Water

0.998

XB3, EG

0.9975 0.997

Mole fraction

Mole fraction

0.9995

0

(a)

2

4 6 Time (h)

0.9999 0.99984

8

10

0.9996

(b)

Pressure (bar)

Pressure (bar)

PDC

0.2

EDC SRC

0.1

(c)

4 6 Time (h)

8

Temperature (°C)

120 100 80

20 0

0

(e)

2

4 6 Time (h)

PDC

SRC-11th

EDC

SRC-4th

4 6 Time (h)

8

10

8

10

8

10

8

10

–20% F

0.3 0.2

PDC EDC SRC

0.1 0

(d)

+20% F

40

2

0.4

0

10

140

60

XB3, EG 0

0.5

0.3

2

XD3, Water

0.99966

0.4

0

XD2, ACN

0.99972

0.5

0

XB1, Water

0.99978

+20% F

0.6

Temperature (°C)

280

180 160 140 120 100 80 60 40 20 0

2

4 6 Time (h)

–20% F

0

(f)

2

PDC

SRC-11th

EDC

SRC-4th 4 6 Time (h)

Figure 10.7 Dynamic responses of BPCS #1: mole fraction changes (a, b), pressure changes (c, d), and temperature changes (e, f) under ±20% of throughput disturbance; mole fraction changes (g, h), pressure changes (i, j), and temperature changes (k, l) under ± 20 mol% ACN composition disturbance.

10.4 Dynamics and Control +20% ACN

1

–20% ACN 1 0.99995

0.9992

XB1, Water XD2, ACN

0.9988

XD3, Water

0.9984 0.998

Mole fraction

Mole fraction

0.9996

XB3, EG 0

2

(g)

0.9999 0.99985 0.9998

XB1, Water

0.99975

XD2, ACN

0.9997

XD3, Water

0.99965

4 6 Time (h)

8

0.9996

10

0.3

0

PDC EDC SRC 0

2

(i)

4 6 Time (h)

8

Temperature (°C)

Temperature (°C)

40 20 0

0

2

(k)

Figure 10.7

PDC

SRC-11th

EDC

SRC-4th

4 6 Time (h)

8

10

10

8

10

EDC SRC 0

2

(j)

120

60

8

PDC

0.1 0

80

10

–20% ACN

0.2

10

100

8

0.3

+20% ACN

140

4 6 Time (h)

0.4

Pressure (bar)

Pressure (bar)

0.4

0.1

2

0.5

0.5

0.2

0

(h)

+20% ACN

0.6

XB3, EG

180 160 140 120 100 80 60 40 20 0

4 6 Time (h)

–20% ACN

0

PDC

SRC-11th

EDC

SRC-4th

2

(l)

4 6 Time (h)

(Continued)

TC6

Q-Sol

FT

FC FC

PC PC

27 6

FT TC3

8

LC

FT

FC

Water

FT

BPCS of process #2.

FT

FC

5 TC2

RR2

FT

34 36 42

TC5

ACN

RR3

7 11

LC

12R

43R

EG

FT FC

Figure 10.8

TC4

FT

LC

9R

LC

4

LC FC

1C

2

2 4 FC

TC1

PC

1C

Sol/F

1

FT

EG makeup

FT

Water

281

10 Dynamic Safety Analysis of Intensified Extractive Distillation Processes

10.4.4 BPCS #3 The combined column has three manipulated variables: side reboiler duty (Q-side), side draw flow rate (F-side), and bottom reboiler duty (QR1). In the EDC, there are two variables: reflux ratio (RR2) and reboiler duty (QR2). According to the slope criterion, the temperature-sensitive stages in the combined column are identified as the 6th , 13th , and 16th stages. Since the temperature-sensitive 16th stage is located near bottom, QR1 is employed to control it and prevent water from flowing downward. Dynamic simulations have shown that the 6th and 13th stages should be controlled using the Q-side and F-side, respectively. Controlling the 6th stage with Q-side helps prevent the ACN from falling onto the tray of the side stream, while controlling the 13th stage allows adjustment of the side product flow rate. In the EDC, the manipulated and controlled variables are easily paired using the dual-end strategy, similar to BPCS #2. This configuration ensures effective control of the temperature-sensitive stages. The resulting BPCS #3 is depicted in Figure 10.10. +20% F

1

–20% F

0.999

Mole fraction

Mole fraction

0.9995 XB1, Water XD2, ACN

0.9985

XD3, Water

0.998 0

Pressure (bar)

(a) 0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0

2

4 6 Time (h)

8

1 0.99995 0.9999 0.99985 0.9998 0.99975 0.9997 0.99965 0.9996

10

EDC SRC 0

(c)

2

4 6 Time (h)

8

10

4 6 Time (h)

8

10

8

10

PDC EDC SRC 0

(d)

2

4 6 Time (h)

–20% F

160 140 Temperature (°C)

120 100 80 60

EDC-36th

PDC

40

EDC-27th

20 0

2

–20% F

0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0

+20% F

140

XD3, Water

0

+20% F

PDC

XB1, Water XD2, ACN

(b)

Pressure (bar)

0.9975

Temperature (°C)

282

0

(e)

2

4 6 Time (h)

SRC-4th SRC-7th 8

10

120 100 80 60

EDC-36th

40

PDC

20

EDC-27th

0

0

(f)

2

4 6 Time (h)

SRC-4th SRC-7th 8

10

Figure 10.9 Dynamic responses of BPCS #2: mole fraction changes (a, b), pressure changes (c, d), and temperature changes (e, f) under ±20% of throughput disturbance; mole fraction changes (g, h), pressure changes (i, j), and temperature changes (k, l) under ± 20 mol% ACN composition disturbance.

10.4 Dynamics and Control +20% ACN

1

0.9996

Mole fraction

Mole fraction

0.9998 XB1, Water

0.9994

XD2, ACN

0.9992

XD3, Water

0.999 0

2

0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0

4 6 Time (h)

10

PDC EDC SRC 0

2

4 6 Time (h)

8

10

0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0

0

120

EDC-36th

EDC-27th

0

0

2

(k)

Figure 10.9

2

4 6 Time (h)

2

SRC-4th SRC-7th

4 6 Time (h)

8

10

8

4 6 Time (h)

10

8

–20% ACN

160

80

20

0

EDC SRC

140

PDC

XD3, Water

PDC

100

40

XD2, ACN

–20% ACN

120

60

XB1, Water

(j)

+20% ACN

140

–20% ACN

(h)

+20% ACN

(i)

Temperature (°C)

8

Temperature (°C)

Pressure (bar)

(g)

Pressure (bar)

0.9988

1 0.99995 0.9999 0.99985 0.9998 0.99975 0.9997 0.99965 0.9996

100 80 60

0

SRC-4th SRC-7th

EDC-27th

20

10

EDC-36th

PDC

40

0

2

(l)

4 6 Time (h)

10

8

(Continued) TC6

FT

Q-Sol

FC FC

PC PC

FT

1C

Sol/F

1

2 LC

6 FC TC1

FT

FT FC

27

10

Water TC3

13 14 16

TC2

TC4

RR2

FT

TC5

ACN

LC FC

38R LC

21R

EG

FT

EG makeup

Figure 10.10

33 34 37

FT

20

FC

FT

FC

6

BPCS of process #3.

FT

283

1 0.98 0.96 0.94 0.92 0.9 0.88 0.86 0.84 0.82 0.8

+20% F

–20% F 1 Mole fraction

Mole fraction

10 Dynamic Safety Analysis of Intensified Extractive Distillation Processes

ACN Water EG

0

(a)

2

4 6 Time (h)

8

10

0.99995 0.9999 0.99985 0.9998

2

4 6 Time (h)

8

10

–20% F

0.35 Pressure (bar)

0.25 0.2 0.15

Combined

0.1

EDC

0.05 0

(c)

2

4 6 Time (h)

0.3 0.25 0.2 0.15

8

0

10

Comb-13th Comb-16th

EDC-27th EDC-34th 2

4 6 Time (h)

8

10

0

(d)

Comb-6th

0

EDC

0.05

+20% F

(e)

Combined

0.1

Temperature (°C)

Pressure (bar)

0

(b) 0.4

0.3

200 180 160 140 120 100 80 60 40 20 0

EG

0.99965 0.9996

0.35

0

ACN Water

0.99975 0.9997

+20% F

0.4

Temperature (°C)

284

180 160 140 120 100 80 60 40 20 0

2

4 6 Time (h)

8

10

–20% F Comb-6th Comb-13th Comb-16th

EDC-27th EDC-34th 0

(f)

2

4 6 Time (h)

8

10

Figure 10.11 Dynamic responses of BPCS #3: mole fraction changes (a, b), pressure changes (c, d), and temperature changes (e, f) under ±20% of throughput disturbance; mole fraction changes (g, h), pressure changes (i, j), and temperature changes (k, l) under ± 20 mol% ACN composition disturbance.

The closed-loop dynamic simulations for distillation columns with ±20% changes in throughput and composition are presented in Figure 10.11. The response of all five temperature control points is rapid, accompanied by reduced oscillations. Furthermore, all product purities quickly return to their desired values, demonstrating effective control and regulatory performance in the face of these disturbances.

10.5 Safety Analysis This chapter primarily focuses on addressing the safety issue related to pressure. The examination initiates a safety evaluation of Process #1, concentrating on two specific deviations: a potential failure in the cooling water supply to the condenser and an abrupt increase in steam flow to the reboiler. Since Processes #2 and #3 incorporate compressors, an additional safety event involving compressor power failure is also

10.5 Safety Analysis +20% ACN

–20% ACN 1 0.99995

0.9996 0.9994

Mole fraction

Mole fraction

1 0.9998 ACN Water

0.9992 0.999

EG

0.9988

0.9999 0.99985

0.9986

0.99975

0.9984

0.9997

0

(g)

2

4 6 Time (h)

8

10

Pressure (bar)

Pressure (bar)

0.34 0.33

Combined

0.32

EDC

0.31

4 6 Time (h)

8

10

–20% ACN

0.3 0.25 0.2 0.15

Combined

0.1

EDC

0.05 0

(i)

2

4 6 Time (h)

8

0

10

0

(j)

2

+20% ACN

160

Temperature (°C)

120 100 80 60 40

Comb-6th Comb-13th

20 0

(k)

Figure 10.11

2

4 6 Time (h)

4 6 Time (h)

8

10

–20% ACN

140 Temperature (°C)

2

0.35

0.35

0

0

0.4

0.36

0.3

EG

(h)

+20% ACN

0.37

ACN Water

0.9998

Comb-16th EDC-27th EDC-34th 8

10

180 160 140 120 100 80 60 40 20 0

Comb-6th Comb-13th 0

(l)

2

Comb-16th EDC-27th EDC-34th

4 6 Time (h)

8

10

(Continued)

being taken into consideration. The safety analysis for Processes #1 and #2 currently considers only the BPCS IPLs, while for the intensified Process #3, additional IPLs will be taken into account.

10.5.1 Process #1 Safety Analysis Process #1 consists of three condensers and three reboilers. Initially, our focus is on the first safety event, which involves the failure of the cooling water supply. To simulate this scenario, the system operates at a steady state for 10 seconds before the pressure controllers are manually switched, and the controller output signal is adjusted to 0% (valve fully closed). Figure 10.12 demonstrates the dynamic responses observed when the coolant supply is cut off. The figure clearly shows that cutting off the coolant supply leads to an immediate and significant increase in pressure within the respective columns, starting from their initial operating conditions. However, it is worth noting that the overpressure effect resulting from the coolant failure remains localized within the affected column and does not propagate to other

285

10 Dynamic Safety Analysis of Intensified Extractive Distillation Processes Loss of cooling water in PDC

2.5 2

PDC EDC

1.5

SRC

1

2

PDC EDC

1.5

SRC

1 0.5

0.5 0

Loss of cooling water in EDC

2.5 Pressure (bar)

3

Pressure (bar)

0

(a)

50

100

150 200 Time (s)

250

0

300

0

(b)

200

400 600 Time (s)

800

1000

Loss of cooling water in SRC

5 4 Pressure (bar)

286

3

PDC EDC

2

SRC

1 0

0

(c)

500

1000

1500 2000 Time (s)

2500

3000

Figure 10.12 Dynamic responses of process #1 when the coolant supply is cut off: (a) loss of cooling water in PDC, (b) loss of cooling water in EDC, and (c) loss of cooling water in SRC.

columns. This observation suggests that the BPCS effectively prevents the propagation of deviations. If the temperature controller is mistakenly set to manual and the steam valve is fully opened, a sudden increase in heat input to the reboilers of a column can result in overpressure. To simulate this scenario in a dynamic simulation, the system is first maintained at a steady state for 1 minute. Subsequently, the temperature controller is switched to manual mode, and the steam valve is fully opened. However, unlike the scenario involving a condenser cooling failure, an increase in reboiler duty leads to a slower and smaller deviation in column pressure, as depicted in Figure 10.13. Nevertheless, similar to the situation with cutting off the coolant supply, the BPCSs remain effective in managing the propagation of overpressure.

10.5.2 Process #2 Safety Analysis In Process #2, several deviations were considered, including compressor power off, failure of cooling water supply in EDC and SRC, and reboiler surges in three columns. The dynamic responses of pressure were depicted in Figure 10.14, taking into account the deviations related to compressors and condensers. Concerning the PDC, the deviation of compressor power-off resulted in an elevation of column pressure. However, the subsequent two columns maintained well-controlled pressure, indicating the effectiveness of the BPCS. In the case of the second EDC, the failure in the coolant supply not only caused an increase in EDC pressure but also led to an increase in PDC pressure due to the interconnected nature of the

10.5 Safety Analysis Reboiler duty surges in PDC

1.2

0.6 Pressure (bar)

1 Pressure (bar)

Reboiler duty surges in EDC

0.7

PDC

0.8

EDC

0.6

SRC

0.4

0.5 0.4 0.3 0.2

0.2

0.1

0

0

0

(a)

10

20

30 40 Time (min)

50

60

PDC EDC 0

5

(b)

SRC

10 15 Time (min)

20

25

Reboiler duty surges in SRC

1.2

Pressure (bar)

1 0.8 0.6 0.4

PDC

0.2 0

SRC

EDC 0

10

(c)

20 30 Time (min)

40

50

Figure 10.13 Dynamic responses of process #1 when surge in reboilers: (a) reboiler duty surges in PDC, (b) reboiler duty surges in EDC, and (c) reboiler duty surges in SRC. Loss of compressor power

3

PDC

2

EDC

1.5

SRC

Pressure (bar)

Pressure (bar)

2.5

1 0.5 0

0

(a)

100

200 Time (s)

300

400

2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

Failure of cooling water supply in EDC PDC EDC SRC

0

(b)

100

200 300 Time (s)

400

500

Failure of cooling water supply in SRC

1.4 Pressure (bar)

1.2 PDC

1

EDC

0.8

SRC

0.6 0.4 0.2 0

0

(c)

100

200 300 Time (s)

400

500

Figure 10.14 Dynamic responses of process #2 when the compressor power and coolant supply is cut off: (a) loss of compressor power, (b) failure of cooling water supply in EDC, and (c) failure of cooling water supply in SRC.

287

10 Dynamic Safety Analysis of Intensified Extractive Distillation Processes Reboiler duty surges in PDC

0.6 0.5

Pressure (bar)

Pressure (bar)

0.4 0.3

PDC

0.2

EDC

0.1 0

SRC 0

(a)

50 Time (min)

100

0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0

Reboiler duty surges in EDC

PDC EDC 0

50

(b)

100 Time (min)

SRC 150

200

Reboiler duty surges in SRC

0.8 0.7 Pressure (bar)

288

0.6 0.5 0.4 0.3 0.2

PDC

0.1

EDC

0

0

(c)

10

20 Time (min)

SRC 30

40

Figure 10.15 Dynamic responses of process #2 when heat duties surge in reboilers: (a) reboiler duty surges in PDC, (b) reboiler duty surges in EDC, and (c) reboiler duty surges in SRC.

two columns. Lastly, the cooling water failure in SRC did not impact the operating pressures of PDC and EDC. Figure 10.15 provides insights into the impact of deviations in reboiler duty. It is evident that when there is a sudden increase in heat input to the PDC, the pressure in each column of the system also rises. This phenomenon can be attributed to the increased upward vapor flow rate caused by the heightened PDC reboiler duty, subsequently affecting the following two columns. This observation emphasized the pressure propagation effect resulting from the integration of the two columns. On the contrary, an increase in reboiler duty in EDC and SRC does not significantly influence the pressures in the columns. This finding demonstrates the effectiveness of the BPCS in mitigating the impact of such deviations.

10.5.3 Process #3 Safety Analysis Process #3 represents the most intensified process investigated in this analysis, focusing on two deviations: loss of compressor power and coolant failure in EDC. The dynamic responses to these deviations are depicted in Figure 10.16. In the case of the loss of compressor power, it takes approximately 250 seconds for an overpressure of approximately 4 bar to develop. On the other hand, coolant failure in EDC leads to a gradual increase in column pressure over an extended period of time. Conversely, the surge in reboiler duty leads to relatively minor pressure fluctuations, as depicted in Figure 10.17. This can be attributed to the effective mitigation of this deviation by the BPCS.

4.5 4 3.5 3 2.5 2 1.5 1 0.5 0

Loss of compressor power

Failure of colling water supply in EDC

1.2 Pressure (bar)

Pressure (bar)

10.5 Safety Analysis

PDC EDC

1

PDC

0.8

EDC

0.6 0.4 0.2

0

(a)

100

200 300 Time (s)

400

0

500

0

(b)

50

100

150 200 Time (s)

250

300

Figure 10.16 Dynamic responses of process #3 when compressor power and coolant supply is cut off: (a) loss of compressor power and (b) failure of cooling water supply in EDC.

0.4

Reboiler duty surges in combined column

0.35

0.3 0.25 0.2 0.15

PDC

0.1

EDC

Pressure (bar)

Pressure (bar)

0.35

0.34 0.33 PDC

0.32

EDC

0.31

0.05 0

Reboiler duty surges in EDC

0.36

0

(a)

50

100 Time (min)

150

200

0.3

0

(b)

50

100 Time (min)

150

200

Figure 10.17 Dynamic responses of process #3 when heat duties surge in reboilers: (a) reboiler duty surges in combined column and (b) reboiler duty surges in EDC.

During the subsequent safety analysis, multiple IPLs such as CA&HI, SIS, and PSV are examined to mitigate the risk of accidents. The specific deviation of compressor power off in the combined column is selected as a case study to evaluate the implementation and effectiveness of these IPLs in preventing and mitigating potential hazards.

10.5.4 Dynamic Safety Analysis of Process #3 with IPLs In order to ensure process safety when facing potential deviations that could lead to accidents, the presence of multiple protection layers is crucial. In this subsection, we will focus on analyzing process safety using the deviation of compressor power off in the combined column as a case study. The initial pressure of the combined column is recorded as 0.354 bar, and based on Figure 10.11, we observe that the maximum pressure under various disturbances is approximately 0.36 bar. To avoid false alarms during process disturbances, it is advisable to set the alarm pressure higher than this value. Therefore, we chose an alarm pressure of 0.4 bar. For the SIS, determining an appropriate activation pressure is essential. Setting it too low would provide less time for human intervention to restore normal operation, resulting in more frequent emergency plant shutdowns and reduced plant economics. Conversely, setting it too high might lead to equipment damage before the SIS activates. Hence, determining the activation pressure is a trade-off. In this case, we set the SIS activation pressure

289

10 Dynamic Safety Analysis of Intensified Extractive Distillation Processes

at 1.5 bar. In the event of a failure of the SIS to activate during an emergency, the PSV acts as the final layer of protection against accidents. In our analysis, we set the PSV activation pressure to be 3 bars. By implementing these multiple protection layers, the risk of accidents can be significantly reduced, ensuring the safe operation of the process. In the CA&HI IPL, the estimated time available for operators to respond after an alarm is determined through dynamic simulation. After introducing the deviation of compressor power off, the column pressure reaches the first high-pressure alarm level of 0.4 bar in 5 seconds. It then takes an additional 104 seconds to reach the second high-pressure alarm level of 1.5 bar, which is the activation pressure for SIS. Hence, the operator is expected to respond within approximately 1.5 minutes. During the dynamic simulation, it is assumed that the operator can eliminate the reboiler heat duty at specific time intervals: 20/40/60/80 seconds after the high-pressure alarm. As depicted in Figure 10.18, the implementation of CA&HI effectively prevents further overpressure in the column overhead. This IPL acts as an additional layer of protection, enabling operators to intervene before the situation reaches a critical point. By providing this intervention opportunity, CA&HI reduces the risk of accidents and enhances overall process safety. In situations where the BPCS and CA&HI IPLs are unable to prevent hazardous conditions, additional interlocks can be implemented to enhance protection in chemical processes. In this chapter, the method of energy isolation through the shutdown of the steam utility is explored as a means for overpressure protection. Figure 10.19 illustrates the effectiveness of this method by demonstrating that simultaneous closure of the side reboiler and reboiler can successfully control the pressure. By implementing energy isolation, the source of energy is isolated, effectively stopping the flow of the stream to the columns. This approach provides an additional layer of protection against overpressure situations, thereby ensuring the safety of the process and the overall plant. When the IPLs, including the BPCS, CA&HI, and SIS, fail to function properly, the PSV serves as the final safety measure against overpressure accidents. To assess 1.4

Response at Response at 80 s later 60 s later Response at 40 s later Response at 20 s later

1.2 Pressure (bar)

290

1 0.8 0.6 0.4

Alarm activated at 5 s pressure of combined column at 0.4 bar

0.2 0

0

Figure 10.18

20

40 60 Time (s)

80

100

Dynamic responses of process #3 when CA&HI is introduced.

10.5 Safety Analysis

1.6

Pressure (bar)

1.4 1.2 SIS activated cut the steam

1 0.8 0.6

Pressure of combined column

0.4 0.2 0

0

Figure 10.19

50

100 Time (s)

150

200

Dynamic responses of process #3 when SIS is activated. 100

Valve position (%)

Figure 10.20 Hysteresis diagrams of the PSV.

Full lift pressure, closing (2.85, 100%)

80

Full lift pressure, opening (3.15, 100%)

60 Reset (2.85, 0)

40 20 0

Primary lift pressure (3.15, 10%)

Set pressure (3, 0) 2.8

2.9

3 Pressure (bar)

3.1

3.2

the effectiveness of the PSV, it is installed at the top of the combined column, and its hysteresis diagram with characteristic parameters is presented in Figure 10.20. For the sizing of PSV, as the backpressure is 1 bar (atmospheric pressure), the throat diameter and inlet/outlet diameter are set as 0.125 m. After operating at a steady state for 10 seconds, the introduction of a deviation, such as compressor power off, results in an immediate rise in pressure. The pressure reaches the PSV set pressure of 3 bar in approximately 180 seconds. As depicted in Figure 10.21, once the column overhead pressure exceeds the PSV set pressure, gas is released through the vent, effectively reducing the pressure. Following this event, the vent’s flow rate and the overpressure level at the top of the column stabilized. While the PSV effectively prevents the column pressure from surpassing the set pressure and avoids catastrophic accidents, relying solely on the PSV for overpressure protection is not recommended, as it is considered the last resort. It is crucial to implement multiple layers of protection to ensure process safety. Additionally, regular testing and maintenance of the PSV are necessary to ensure its effectiveness in emergency situations.

291

10 Dynamic Safety Analysis of Intensified Extractive Distillation Processes

3.5 Pressure of PDC

2.5

40

PSV activated at 3 bar

30

2 1.5 1 0.5 0

0

100

200

300

Pressure

20

Vent flowrate

10

400

Vent flowrate (kmol/h)

50

3 Pressure (bar)

292

0 500

Time (s)

Figure 10.21

Pressure responses and vent flow rates generated to compressor power off.

ACN is indeed a hazardous organic compound that can contribute to air pollution. Therefore, additional treatment is necessary for the vent flow to ensure proper handling of ACN emissions. In the final release system, the relief vent is directed toward a knockout system, which is designed to separate the liquid from the vapor. This separation process helps in removing any liquid ACN present in the vent stream. The vapor from the knockout system is then subjected to further treatment using a simple condenser. The condenser is responsible for cooling the vapor, causing it to condense back into a liquid state. This process helps in capturing ACN vapors effectively. The recovered condensates, which now contain the condensed ACN, can be recycled and sent back to the crude feed storage tank for further distillation. This recycling process ensures that the ACN is properly managed and reused, minimizing waste and reducing the environmental impact associated with ACN emissions. Overall, the additional treatment and recovery measures implemented for the ACN vent flow help in reducing air pollution and promoting sustainable practices within the process.

10.6 Conclusions This chapter demonstrates the dynamic control and safety analysis of intensified extractive processes using dynamic simulation. Process intensification measures are incorporated into ED processes to generate intensified designs, resulting in significant cost reductions compared to conventional approaches. The study examines the plantwide dynamics and control of these intensified processes, showcasing the effectiveness of control structures with temperature controllers in managing disturbances. Additionally, the assessment of safety highlights the importance of IPLs, such as the BPCS, CA&HI, SISs, and PSVs, in preventing critical issues. By appropriately configuring these protections, column overpressure problems can be effectively mitigated. Overall, this chapter provides valuable insights into enhancing process performance and safety through dynamic simulation and the implementation of process intensification measures.

References

Acknowledgments This work was supported by the Natural Science Foundation of Jiangsu Province (Grant No. BK20220348) and the National Natural Science Foundation of China (Grant No. 22208154).

References An, Y., Li, W., Li, Y. et al. (2015). Design/optimization of energy-saving extractive distillation process by combining preconcentration column and extractive distillation column. Chemical Engineering Science 135: 166–178. Arifin, S. and Chien, I.-L. (2007). Combined preconcentrator/recovery column design for isopropyl alcohol dehydration process. Industrial & Engineering Chemistry Research 46: 2535–2543. Berdouzi, F., Villemur, C., Olivier-Maget, N., and Gabas, N. (2018). Dynamic simulation for risk analysis: application to an exothermic reaction. Process Safety and Environmental Protection 113: 149–163. Carlos, M., Fatine, B., Nelly, O.-M., and Nadine, G. (2018). Deviation propagation analysis along a cumene process by using dynamic simulations. Computers & Chemical Engineering 117: 331–350. Chang, W.-T., Huang, C.-T., and Cheng, S.-H. (2012). Design and control of a complete azeotropic distillation system incorporating stripping columns for isopropyl alcohol dehydration. Industrial & Engineering Chemistry Research 51: 2997–3006. Cui, C., Qi, M., Shu, C.-M., and Liu, Y. (2023). Rigorous dynamic simulation methodology for scenario-based safety analysis of pressure-swing distillation considering independent protections. Process Safety and Environmental Protection 172: 282–304. Cui, C., Zhang, Q., Zhang, X. et al. (2022). Process synthesis and plantwide control of intensified extractive distillation with preconcentration for separating the minimum-boiling azeotropes: a case study of acetonitrile dehydration. Separation and Purification Technology 285: 120397. Hendershot, D.C. (1997). Inherently safer chemical process design. Journal of Loss Prevention in the Process Industries 10: 151–157. ˇ (2019). Software approach Janošovský, J., Danko, M., Labovský, J., and Jelemenský, L. to simulation-based hazard identification of complex industrial processes. Computers & Chemical Engineering 122: 66–79. Khan, F., Rathnayaka, S., and Ahmed, S. (2015). Methods and models in process safety and risk management: past, present and future. Process Safety and Environmental Protection 98: 116–147. Kummer, A. and Varga, T. (2019). Process simulator assisted framework to support process safety analysis. Journal of Loss Prevention in the Process Industries 58: 22–29. ˇ et al. (2014). Model-based hazard Labovská, Z., Labovský, J., Jelemenský, L. identification in multiphase chemical reactors. Journal of Loss Prevention in the Process Industries 29: 155–162.

293

294

10 Dynamic Safety Analysis of Intensified Extractive Distillation Processes

Leveson, N.G. and Stephanopoulos, G. (2014). A system-theoretic, control-inspired view and approach to process safety. AICHE Journal 60: 2–14. Liang, K., Li, W., Luo, H. et al. (2014). Energy-efficient extractive distillation process by combining preconcentration column and entrainer recovery column. Industrial & Engineering Chemistry Research 53: 7121–7131. Luyben, W.L. (2006). Plantwide control of an isopropyl alcohol dehydration process. AICHE Journal 52: 2290–2296. Luyben, W.L. and Chien, I.-L. (2011). Design and Control of Distillation Systems for Separating Azeotropes. John Wiley & Sons. Ortiz-Espinoza, A.P., Jiménez-Gutiérrez, A., and El-Halwagi, M.M. (2017). Including inherent safety in the design of chemical processes. Industrial & Engineering Chemistry Research 56: 14507–14517. Wu, Y.C., Hsu, P.H.-C., and Chien, I.-L. (2013). Critical assessment of the energy-saving potential of an extractive dividing-wall column. Industrial & Engineering Chemistry Research 52: 5384–5399. You, X., Gu, J., Gerbaud, V. et al. (2018). Optimization of pre-concentration, entrainer recycle and pressure selection for the extractive distillation of acetonitrile-water with ethylene glycol. Chemical Engineering Science 177: 354–368. Zhang, X., He, J., Cui, C., and Sun, J. (2020). A systematic process synthesis method towards sustainable extractive distillation processes with pre-concentration for separating the binary minimum azeotropes. Chemical Engineering Science 227: 115932. Zhu, C., Qi, M., and Jiang, J. (2020). Quantifying human error probability in independent protection layers for a batch reactor system using dynamic simulations. Process Safety and Environmental Protection 133: 243–258.

295

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts Juan G. Segovia-Hernández 1 , César Ramírez-Márquez 2 , Gabriel Contreras-Zarazúa 3 , Eduardo Sánchez-Ramírez 1 , and Juan J. Quiroz-Ramírez 3 1

Departamento de Ingeniería Química, Universidad de Guanajuato, Guanajuato, Guanajuato, México de Ingeniería Química, Universidad Michoacana de San Nicolás de Hidalgo, Morelia 58060, Michoacán, Mexico 3 CONACyT – CIATEC A.C. Centro de Innovación Aplicada en Tecnologías Competitivas, León 37545, Guanajuato, México 2 Departamento

11.1 Introduction The development of societies has brought a broad demand for clean, lower-cost, and safe processes to achieve sustainable development. This pattern results from the reality that society’s biggest problems today are pollution, resource scarcity, and global warming (Huang and Peng, 2014). Different raw materials, such as biomass and sugar-rich sources, are considered. The use of this sugar source for producing valuable products that do not depend on a fossil source has attracted the attention of different researchers focused on developing more efficient processes. In fermentation processes, sugars such as pentoses and hexoses are transformed to obtain bioproducts. Bioprocesses are a current and innovative solution that is continuously advancing in terms of the number of applications, responding to more industrial needs and opportunities. The benefits of these new technologies (bioproduction, biosynthesis, and biocatalysis) can be exploited from the cosmetic or pharmaceutical industry, through packaging and capital goods, to the agricultural and food sectors. However, significant challenges must be met to obtain highly efficient bioprocesses. The high amounts of water needed for the culture medium either to decrease the concentration of sugars or the dilution of the products generated, both of which are harmful in most cases for the microorganisms in charge of converting sugars into valuable products. The design of bioprocesses continues to be a tremendous challenge for process designers. The large quantities of materials, azeotropes, energy intensity, and above all, the design of these processes in a sustainable way have brought the consideration of new indices that in traditional processes had not been incorporated, such Control and Safety Analysis of Intensified Chemical Processes, First Edition. Edited by Dipesh Shikchand Patle and Gade Pandu Rangaiah. © 2024 WILEY-VCH GmbH. Published 2024 by WILEY-VCH GmbH.

296

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

as controllability and safety. The identification and understanding of the problems related to security and controllability in industries have made significant advances in the last 10 years. Many processes and products are being created with environmental and safety concerns in mind, aiming for safe and clean engineering practices to be widely and systematically adopted (Huang and Peng, 2014). In line with this objective, the development of bioprocesses and sustainable processes has gained significant importance due to the growing concerns about the sustainability and environmental impact of industries. In any project, acknowledging the importance of risk in design is crucial, as it involves the identification, evaluation, and development of suitable strategies to mitigate potential hazards. The design process extends beyond the initial stage and encompasses all subsequent phases, including downstream processes. It is imperative to evaluate and manage the risks at each stage to ensure the overall process’s safety and efficiency. Excluding risks associated with these stages may result in an incomplete risk assessment, thereby amplifying the general process risk. Consequently, conducting a comprehensive risk evaluation in all design stages, including downstream processes, is pivotal for effective risk management and successful project outcomes (Leveson, 2016; Mannan et al., 2016; CCPS, 2010). Bioprocesses and bioproducts offer a promising alternative to traditional processes based on fossil fuels by utilizing living organisms or their components to produce a wide range of products. This shift toward bio-based approaches is driven by the desire to reduce dependence on non-renewable resources and mitigate environmental damage. In the early stages of bioprocess design, various factors need to be carefully considered, such as the selection of microorganisms or cells, the choice of substrates and nutrients, and the optimization of cultivation conditions. Adequate design during this initial phase sets the foundation for an efficient and sustainable process. Continuous process improvement is another key aspect of bioprocess development. As more experience is gained and operational data is collected, opportunities to optimize and enhance process performance can be identified. This involves implementing advanced control strategies, optimizing cultivation parameters, and adopting new technologies that enable higher efficiency and productivity. Sustainable process design is essential for ensuring efficiency and minimizing environmental impact. It encompasses reducing waste generation and maximizing the yield of desired products. Furthermore, it aims to improve process safety and controllability by implementing preventive measures and risk management protocols. Conceptual design plays a pivotal role in the development of bioprocesses and sustainable processes. This stage involves exploring new alternatives, evaluating different technologies, and analyzing economic and environmental aspects. The objective is to develop novel and innovative processes that can overcome the limitations of conventional approaches. In this regard, it is crucial to explore alternatives that reduce waste generated during the process and maximize product yield. This can be achieved through the

11.1 Introduction

Figure 11.1 Development of bioprocesses and sustainable processes.

Conseptual design

Process improvement

Sustaniable process design

New, novel and inovative process alternatives

Bioprocess bioproducts Early-stage design

Reduce waste, increase yield, improve environmental, security and controllability

implementation of recycling and reutilization technologies, the optimization of material and energy flows, and the adoption of cleaner production approaches. In addition to the environmental benefits, sustainable processes can also bring economic advantages to companies. Waste reduction and improved performance can translate into significant cost savings and increased competitiveness in the market. The development of bioprocesses and sustainable processes entails the design and continuous improvement of systems that utilize living organisms to produce a wide range of products. This includes seeking new, novel, and innovative alternatives with the goal of reducing waste, increasing yield, and improving environmental sustainability, safety, and controllability. By addressing these challenges, companies can contribute to a more sustainable and resilient future (Figure 11.1). Obtaining highly efficient bioprocesses can be challenging due to several factors. In addition to the challenges mentioned earlier, another important aspect to consider is the operability and safety considerations of intensified structures for the purification of bioproducts. Some of the challenges in this area include: 1. Process Integration: Intensified structures for the purification of bioproducts often involve the integration of multiple unit operations, such as chromatography, filtration, and centrifugation, into a compact and continuous process. Designing and optimizing the integration of these unit operations while maintaining efficient and seamless operations can be complex (D’Souza et al., 2013). 2. Process Robustness: Ensuring the robustness of intensified purification processes is crucial for consistent and reliable operation. Factors such as fouling, variability in feedstock quality, and process disturbances can affect the performance and efficiency of the process. Developing strategies to handle and mitigate these challenges is essential for maintaining process efficiency (Becker et al., 2023). 3. Process Safety: Safety considerations are paramount in bioprocessing, particularly in intensified purification structures where higher pressures and flow rates may be involved. Adequate measures should be implemented to prevent accidents, ensure equipment integrity, and protect personnel. Risk assessments, safety protocols, and equipment design that comply with safety standards are essential to address these challenges (Govasmark et al., 2011; Amyotte et al., 2023.

297

298

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

4. Control and Automation: Intensified purification structures often require advanced control and automation systems to ensure efficient and precise operation. Implementing robust control strategies that can handle the complexity of intensified processes, monitor critical process parameters, and adjust operations in real time is essential for achieving high process efficiency (Thomassen et al., 2019). 5. Equipment Design and Scalability: Intensified purification structures may involve the use of specialized equipment and technologies, such as continuous chromatography columns, membrane filtration modules, or high-speed centrifuges. Designing these systems to be scalable, cost-effective, and compatible with different process conditions can be challenging. Ensuring that equipment can be easily integrated into existing facilities or scaled up for commercial production is crucial (Ruiz-Ruiz et al., 2013). 6. Validation and Regulatory Compliance: Intensified purification structures must meet regulatory requirements and undergo validation processes to ensure product quality, safety, and consistency. Developing appropriate validation protocols, conducting thorough process validation, and complying with regulatory guidelines are essential challenges that need to be addressed (Narayanan et al., 2020). Addressing these challenges requires a comprehensive understanding of process engineering principles, safety considerations, and regulatory requirements. Collaboration between process engineers, bioprocess scientists, and regulatory experts is crucial to developing efficient and safe intensified structures for the purification of bioproducts. Inherent safety has been used as a general idea rather than a specific technique to shape inherently environmentally friendly processes. Its proactive thinking and consistent application principles pave the way for maximizing resource use and risk management efficiency with fewer corrective actions. When analyzing and choosing process options during the initial process design phase, a protocol with careful considerations can be provided to better incorporate the intrinsic notions of green chemistry, sustainable development, and clean technology. The conceptual design phase, where many important design decisions are still being made, can help create intrinsic safety mechanisms. It is crucial for safety engineers and researchers to develop compatible tools and measurements that are suitable for the nature of conceptual design and can be easily utilized by process engineers. Nonetheless, safety experts should conduct research to establish requirements, specifications, and safety feature quantification for intensification, substitution, mitigation, and simplification, using the available level of detail during the conceptual design phase. Inherent safety, economics, controllability, and greenhouse gas emissions, according to Jiménez-González and Constable (2011), are the indices that contribute most to the ideas of sustainability and green chemistry. In the context of process safety, it is important to distinguish between intrinsic safety and inherent safety. Intrinsic safety is a design approach specifically aimed at preventing the ignition or explosion of hazardous materials in potentially explosive atmospheres. This is achieved through the design of electrical and electronic

11.1 Introduction

equipment with low energy levels, ensuring that no sparks or heat are generated that could potentially ignite flammable substances. By employing this design strategy, the equipment is inherently safe and poses no threat in hazardous areas (Mannan, 2011). On the other hand, inherent safety is a broader concept that encompasses the elimination or reduction of hazards and risks throughout the entire design of a process or system. It involves selecting inherently safer materials, processes, and equipment to prevent or mitigate the consequences of accidents or incidents. Inherent safety focuses on designing the system to be intrinsically safe, thereby reducing or eliminating the need for additional safety measures or protective systems. This comprehensive approach takes into account various factors, such as the choice of materials, process conditions, and equipment design, to achieve a higher level of safety in the overall system (Gao et al., 2020). The issue of intrinsic safety, which is of great importance, has been consistently overlooked or not given adequate attention after the design stage, as noted in the works of Govasmark et al. (2011). Khan et al. (2021) have previously conducted a risk assessment of chemical processes that were already developed, while Hassim and Hurme (2010) examined the safety of a chemical facility that included distillation columns but failed to consider the inherent hazards during the design stage. Additionally, Martinez-Gomez et al. (2016, 2017) discovered that the estimation of intrinsic safety for biobutanol and silane production schemes was conducted only after the design and optimization processes were completed. With this conventional approach, separation options can be created, but significant safety risks would be associated with handling high operating pressures, volatile liquids, column sizing, etc. Therefore, excluding downstream process-related risks may result in a risk assessment. That is inaccurate. A practical alternative would be to consider the inherent environmental, economic, and safety aspects during the early design stages, ensuring a process with minimal negative environmental and economic impacts while considering each alternative’s risks. This concept will address inherent risk at an early stage, along with financial and environmental concerns. One way to enhance the inherent safety of processes, as well as other metrics such as environmental and energy impact, is through process intensification (PI). PI is a concept in chemical engineering that involves the development and implementation of innovative technologies to increase the efficiency, productivity, and sustainability of chemical processes. The goal of PI is to achieve more with less, i.e. to achieve higher yields and productivity while using fewer resources, generating less waste, and minimizing the environmental impact of chemical processes. The creation of intensified processes, considering sustainability metrics during their design, can even improve safety, controllability, and environmental impact. Green metrics, also known as sustainability metrics or environmental performance indicators, refer to quantitative measures used to assess and evaluate the environmental impact and sustainability of processes, products, or systems. These metrics provide a scientific and objective framework to monitor and analyze various aspects related to sustainability, including resource consumption, waste generation, greenhouse gas emissions, energy efficiency, water usage, and ecological footprint. Green metrics

299

300

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

play a crucial role in supporting sustainable development by providing a standardized and measurable approach to assess the environmental performance of different activities. They help organizations and policymakers make informed decisions, set goals, track progress, and implement strategies to minimize negative environmental impacts and promote sustainability. Considering sustainability at an early stage of improved process design is crucial to distinguish between processes that are simple to operate and those that are challenging. To achieve the general environmental sustainability goal, it is essential to include “green metrics” in the development of an improved process, as highlighted by Jiménez-González et al. (2012). By incorporating green metrics, emphasis is placed on the environmental, health, safety, and process control aspects. This aligns with the eleventh principle of Green Chemistry, which underlines the importance of real-time process analysis and monitoring to detect process deviations and prevent waste and safety issues. Achieving this necessitates the use of real-time analysis and process control. The timely adjustment of process parameters can prevent quality or safety issues from arising during production. This necessitates the use of real-time analysis and process control. However, when using PI, which involves reducing equipment and altering system topology, control characteristics and dynamic performance may differ from those of non-intensified systems, according to Ponce-Ortega et al. (2012). When it comes to managing hazardous chemicals in the storage, inventory, operation, and transportation divisions, the chemical industry has experienced remarkable growth in plant size, location, and process complexity (Khan and Abbasi, 1999; Srinivasan et al., 2019). Intrinsic safety is considered by some studies to be more important for the sustainability and green chemistry profile of a process than for its economics and operability. Athar et al. (2019) state that intrinsic safety is a commonly used risk assessment technique that improves the sustainability of existing process plants or the design of new facilities. However, traditional risk assessments are typically carried out sequentially, after the steady-state design has been established. Therefore, it is recommended to consider inherent safety from the beginning of the process design to create a more sustainable process design, primarily due to competing design objectives for dynamic and steady-state processes. To improve safety standards, various corrective measures are often introduced, which may come too late or be impossible to change due to tragedies or necessary remedial efforts, as stated by Kidam et al. (2016). The inherent risk of extractive distillation schemes to recover ethanol and improved extractive distillation systems was evaluated (Medina-Herrera et al., 2014a,b). Their findings suggest that the physical characteristics of the chemicals and the amount of inventory in the columns significantly impact the process’s inherent safety. Martinez-Gomez et al. (2016) investigated several enhanced distillation methods to purify biobutanol and assessed the risk of each method (which serves as a measure of inherent safety), taking into account economic and environmental factors. Their findings show that intrinsic safety can be affected by process topology. One of the best-known examples of PI, the Martinez-Gomez et al. (2017)

11.1 Introduction

investigation of a reactive distillation (RD) process to create silane, included a safety analysis. In the early stages of process design, techno-economic criteria have always been the main objectives. Typically, safety is taken into account throughout the detailed design phases. Most design degrees of freedom, including those related to technology and configuration, have already been established. Recent examples of the use of safety measures and principles in process design show how important and popular this issue is becoming among researchers. While some work has integrated direct safety measures that improve economic metrics as part of the design challenge formulation, others have used intrinsic safety principles. Various methods and indicators have been proposed in different areas, such as chemical and physical properties of materials, process conditions, reaction properties, equipment, types of activities/operations, and consequences. Approximately 59% of these tools use computer-assisted techniques to manage vast amounts of data efficiently, allowing professionals to access and manipulate data using software that minimizes the time and energy required to gather information. Additionally, experts can process data using software, which reduces the time and effort required to collect indicator data. For example, combining data from process simulators such as Aspen Plus or HYSYS with visual basic for application (VBA) language can significantly reduce time and improve the safety analysis of process data. Recently, process controllability has been considered during the early stages of process design, and academia and business have emphasized the importance of taking into account dynamic aspects to obtain processes that maximize profits and have a lower environmental impact. However, since many accidents occur under abnormal process conditions, investigating strange situations with a dynamic simulator would allow process designers to reduce the likelihood of probable adverse outcomes. Incorporating sustainability into intensive process design can help distinguish between easy-to-operate and complex processes. To create an environmentally sustainable process, it is important to consider “green metrics,” as recommended by Jiménez-González et al. (2012) and Thomassen et al. (2019). The 11th principle of green chemistry, Principle #11, highlights the significance of conducting real-time analysis and monitoring throughout the process to prevent waste and safety hazards. By detecting any process deviations as they occur, control characteristics are established to reverse any excursion that may negatively impact the quality or safety of the final product. Achieving this necessitates the use of real-time analysis and process control. Processes are initially conceptualized using economic and sustainable steady-state calculations, followed by the synthesis of a control structure based on heuristic controllability measures. Therefore, control system design only begins once the critical aspects of the process have been identified, which can result in an iterative process between process and control system design. However, inadequate dynamic operability can arise due to disturbances and uncertainty. The literature reviews by Yuan et al. (2012) suggest that process design decisions can significantly influence process dynamics and control system capabilities.

301

302

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

Despite this, an enclosed control system can be employed to overcome negative factors such as external disturbances and uncertainties surrounding parameters or models that may affect the chemical process. In such situations, improving the dynamic performance and functions of control systems is crucial. Therefore, exploring the interdependence among process design, process control design, and process operability during the initial design phase is essential to enhance the dynamic performance of chemical processes. The study of fundamental concepts, theories, and techniques is one aspect of simultaneous optimization-based design and control; the study of particular application domains is the other. The interaction between these two sides is the power of concurrent optimization-based design and management (Cabrera-Ruiz et al., 2017). This work presents some case studies where the separation and purification of methyl ethyl ketone, furfural, and lactic acid, important bioproducts in the industrial sector, are analyzed using conventional and intensified processes. The intensification of these processes has been previously reported by Segovia-Hernández et al. (2021). However, this paper analyzes the effect of inherent safety and dynamic behavior due to the PI used in the purification of those bioproducts.

11.2 Methodology It is imperative to make a shift toward more sustainable social and technical systems, as the planet’s life-support systems are under increasing threat from environmental problems like resource depletion, air, water, and soil pollution, biodiversity loss, and excessive land use. Economic challenges such as flawed incentive structures, deregulated markets, problematic ownership structures, and supply risk result in recurrent financial and economic instability for entire economies and individual companies. Sustainability can be viewed as a framework with objectives that target economic and environmental indicators. Early consideration of sustainability issues during the design of intensified processes can help differentiate between easily operable processes and those that are challenging to operate. According to Jiménez-González et al. (2012), it is highly recommended to incorporate the so-called “green metrics” in the design of intensified processes with the aim of achieving environmental sustainability. These “green metrics” are indicators and measures that assess the environmental impact of industrial processes and make it possible to quantify and monitor key aspects such as energy consumption, waste generation, and greenhouse gas emissions. By including these green metrics in the design of intensified processes, it seeks to encourage more sustainable and environmentally friendly industrial practices. By considering factors such as energy efficiency, waste minimization, and emission reduction, it is possible to obtain benefits both economically and environmentally. Incorporating these green metrics into the design of enhanced processes can guide engineers and professionals in making more environmentally conscious decisions. By identifying areas for improvement and opportunities to optimize processes, more sustainable strategies can be implemented, such as the use of clean technologies, the reuse of resources, and the adoption of recycling practices.

11.2 Methodology

In summary, the inclusion of “green metrics” in the design of intensified processes not only helps to achieve environmental sustainability objectives but also drives innovation and the development of more eco-efficient solutions in the industry. The green metrics should focus on environmental, health, safety, and process control aspects. Green Chemistry Principle #11 highlights the importance of real-time analysis and monitoring of processes, with the aim of preventing the generation of waste and safety problems by immediately identifying deviations in the process as they occur. This practice is based on the premise that early detection of irregularities and timely corrective measures can prevent the appearance of undesirable situations, such as the accumulation of hazardous waste or uncontrolled reactions. By incorporating real-time analysis and monitoring, constant monitoring of key process parameters and variables is achieved, allowing early detection of any deviation or anomaly. This facilitates the taking of immediate corrective actions to restore optimal operating conditions and avoid the formation of unwanted by-products. In addition, this real-time monitoring also makes it possible to identify opportunities to improve the efficiency of the process and minimize the generation of waste. The implementation of real-time monitoring and analysis systems in chemical processes provides significant benefits both in terms of safety and environmental sustainability. By quickly detecting and addressing any deviations or undesired conditions, the risks associated with the production of hazardous substances or the release of harmful chemicals into the environment are reduced. In addition, by avoiding the generation of unnecessary waste, efficiency and optimization of resources are promoted, thus contributing to the minimization of environmental impact. Similarly, PI, which involves reducing equipment numbers and changing system topology, can alter control properties and dynamic performance compared to non-intensified systems. With the aim of fulfilling this objective, the present investigation carries out risk quantification using the approach of individual risk (IR). IR denotes the magnitude of hazards that a person encounters depending on their location, factoring in the chances of an accident happening, the probability of being injured or killed, and the rate of incidence. The calculation of IR entails multiplying the frequency of accidents (f i ) with the probability of being impacted while occupying a particular position (Px,y ), as expressed in the subsequent formula: ∑ IR = fi Px,y (11.1) Quantitative risk analysis (QRA) is an efficient technique for ascertaining the probabilities and frequencies of accidents and incidents, along with evaluating the resulting aftermath and harm. The first step in QRA involves identifying plausible incidents, which pertain to the discharge of matter or energy from the process, with distillation columns susceptible to two potential incidents: continuous release and instantaneous release. The occurrence of matter leakage from process equipment due to pipeline rupture is known as a continuous release, whereas a catastrophic vessel rupture resulting in complete matter loss from the process equipment is referred to as an instantaneous release. These incidents were identified through

303

304

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

Instantaneous release Immediate ignition P1 = 0.25

2.3 × 10–5/yr

Delayed ignition P2 = 0.9

P3 = 0.5

P3 = 0.5

No immediate ignition No ignition P1 = 0.75 P2 = 0.1

fi

BLEVE

5.75 × 10–6/yr

UVCE

7.76 × 10–6/yr

Flash fire

7.76 × 10–6/yr

Toxic release

1.55 × 10–6/yr

Continuous release Immediate ignition P1 = 0.1

3.67 × 10–4/yr

No immediate ignition P1 = 0.9

Figure 11.2

Delayed ignition P2 = 0.75

No ignition P2 = 0.25

Jet fire

3.67 × 10–5/yr

Flash fire

2.48 × 10–4/yr

Toxic release

8.26 × 10–5/yr

Events tree diagrams for any process equipment.

a hazard and operability study (HAZOP). Figure 11.2 shows the event trees for potential accidents and their corresponding frequencies (f i ), estimated based on the recommended values of the American Institute of Chemical Engineers (Kumar, 1996). The event tree demonstrates that an instantaneous release can result in accidents such as boiling liquid expanding vapor explosion (BLEVE), unconfined vapor cloud explosion (UVCE), flash fire, and toxic release, while a continuous release can lead to accidents such as jet fire, flash fire, and toxic release. Identifying the variables responsible for causing injuries or fatalities in each accident is the next step once probable accidents have been identified. Thermal radiation (Er ) received by a person is the causative variable for fires such as jet fire and flash fire, whereas for UVCE, the main cause of fatalities is overpressure (Po ), and for toxic release, the concentration (C) is the causative variable. In the context of inherent safety, the main objective is to design and operate systems and processes in a way that minimizes or eliminates the possibility of dangerous events. In the case of BLEVE, it seeks to prevent the occurrence of this type of explosion, as it can have devastating consequences in terms of property damage, injury, and loss of life. To achieve inherent security against BLEVE, different preventive and protective measures are applied(Kletz and Amyotte, 2019). These include: (a) Adequate Container Design: Containers are used that are capable of withstanding the pressure and temperature of the liquid contained in them, thus minimizing the risk of structural failures and ruptures.

11.2 Methodology

(b) Pressure Control: Pressure relief devices are installed in the vessels to ensure that the internal pressure does not exceed safe limits. These devices allow for a controlled release of pressure in the event of a surge, preventing excessive energy buildup and therefore the possibility of a BLEVE. (c) Cooling: Cooling systems are implemented to maintain the temperature of the liquid within safe ranges. This helps prevent the rapid, explosive vaporization that can trigger a BLEVE. (d) Safety Distances: Adequate distances are established between containers and populated areas or critical facilities to reduce the risks associated with a possible BLEVE. These distances make it possible to limit the scope of the explosion and minimize the effects on people and assets. (e) Emergency Planning: Emergency plans are developed that include evacuation measures, response to incidents, and mitigation of consequences in the event of a BLEVE. These plans seek to minimize the impact and guarantee the safety of people and the environment. The calculations for determining the causative variables for each accident have been demonstrated by several authors (Medina-Herrera et al., 2014a). The distance required for the process safety studies depends on dispersion and explosion modeling. This distance was chosen based on the significant consequences of all accidents shown in Figure 11.2. Additionally, a calculation must be performed to determine the correct distance required to reduce the probability of fatalities to zero, according to what is reported in the literature, and the computation of the causative variable begins with determining the amount of material released (Medina-Herrera et al., 2014a). PI, which involves reducing inventories, is an effective inherent safety strategy. After computing the causative variables, the next step is to determine the likelihood of impact, or the probability of injury or death, through a consequence assessment using probit models. A probit model is an equation that establishes a relationship between the response of the affected person, such as the probability of injury or death, and the dose received from a specific exposure, like heat, pressure, or radiation. A probit function is used to model the likelihood of death resulting from overpressure and third-degree burns. The parameters for Eq. (11.2) are shown in Table 11.1. The probability of damage is calculated by substituting probit values in Eq. (11.3). Y = k1 + k2 ln V Table 11.1

(11.2)

Probit parameters. K1

K2

Thermal radiation

−14.9

2.56

Overpressure

−77.1

6.91

V

po

305

306

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

[ Px,y = 0.5 1 + erf

)]

( Y −5 √ 2

(11.3)

where te is thermal radiation. Finally, the result obtained by Eq. (11.3) is replaced inside Eq. (11.1) to obtain the IR.

11.2.1 Control Behavior Analysis RD columns are challenging to control due to their high multivariable and non-linear nature, as well as various constraints and disturbances. Therefore, determining the degrees of freedom for the process is essential for effective control. Degrees of freedom represent the number of variables that can or must be controlled, and it is crucial to be aware of this number to avoid undercontrol or overcontrol. Energy balance considerations are a well-established approach for determining the degrees of freedom for distillation columns. This involves using the reflux flowrate L and the vapor boil-up rate V, which are directly influenced by the heat duty supplied to the reboiler, as control variables to manipulate the distillate and bottom output compositions, as shown in Häggblom and Waller (1992). 11.2.1.1 Singular Value Decomposition

To ensure that chemical plants produce the desired product specifications, it is important for each process unit to operate under constant conditions for extended periods of time. Any deviation from these conditions can result in the final product not meeting the required specifications. Therefore, it is valuable to adjust the operating variables in the event of irregularities in the plant to return the system to its steady-state conditions. A theoretical tool proposed for analyzing multivariate control systems is the Singular Value Decomposition (SVD) (Moore, 1986). SVD involves expressing the steady-state profit matrix K as the product of three matrices: ∑ (11.4) K=U VT SVD is a valuable theoretical tool for analyzing multivariate control systems (Moore, 1986). SVD decomposes the steady-state profit matrix K into three matrices: the matrix of left singular vectors U, the matrix of right singular vectors V, and the diagonal matrix sigma, whose entries correspond to the singular values of K. The condition number is the relationship between the maximum singular value 𝜎 max and the minimum singular value 𝜎 min . 𝜎 (11.5) Condition number = max 𝜎min The condition number is a measure of the sensitivity of a system to errors, where a higher value indicates higher sensitivity to disturbances, while a lower value implies better dynamic performance under control loops. This has been discussed by Cabrera-Ruiz et al. (2017). The minimum singular value can evaluate the invertibility of the system, which reflects potential control problems. On the other hand, the condition number

11.3 Methyl Ethyl Ketone

determines the system’s sensitivity to uncertainties in process parameters and modeling errors, enabling the assessment of the dynamic properties of a design. The aim is to identify designs with higher minimum singular values and lower condition numbers, which are expected to exhibit better dynamic performance under feedback control. The SVD test is advantageous because it is independent of the type of controller used. The basic idea is that the controllability properties of the system are limited or imposed by its inner dynamic structure. A complete frequency range analysis is required to fully apply the SVD method and obtain complete coverage of minimum singular values and condition numbers, as demonstrated by Hernández and Jiménez (1999).

11.3 Methyl Ethyl Ketone Methyl ethyl ketone (MEK) is a ketone with a chemical formula of C4 H8 O7 that has both a methyl and an ethyl group. It’s a clear liquid that is not water-soluble but can dissolve in alcohol, ether, and benzene. MEK is highly flammable and has a low boiling point, which makes it volatile. It is used as a fuel source and has a CAS registry number 78-93-3. Its general characteristics are summarized in Table 11.2. MEK is a versatile chemical used in various industrial applications, including paints, coatings, adhesives, printing inks, and surface cleaners. The MEK market is expected to reach $3.26 billion, with paints and coatings being the leading applications. The Asia Pacific region is the largest producer and consumer of MEK Table 11.2

General properties of MEK in comparison with other fuels.

Component

Sum formula

Gasoline

Ethanol

2-Methylfuran

MEK

Various

C2 H6 O

C5 H6 O

C4 H8 O

Carbon mass fraction

%

83.48

52.14

73.15

66.63

Hydrogen mass fraction

%

13.24

13.13

7.37

11.18

Oxygen mass fraction Density (25 ∘ C)

%

3.25

34.73

19.49

22.19

741

787

Boiling temperature

kg/m ∘C

Vapour pressure (20 ∘ C)

kPa

3

907.5

799

35.8–190.4 78

64

80

—

10.8

5.8

13.9

Specific enthalpy of vaporization kJ/kg air —

101.6

35.52

46.1

Stoichiometric air requirement

l

8.98

10.08

10.52

Lower heating value

MJ/kg

41.56

26.84

30.37

31.45

Lower heating value

MJ/l

30.78

21.09

27.63

25.16

13.97

Research octane number

I

96.9

109

101.7

117

Motor octane number

I

86.4

89.7

82.4

107

Purity

%

—

>99

>99

>99

Source: Adapted from Hoppe et al. (2016).

307

308

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

due to the growth of the printing ink market. However, strict regulations and the impact of crude oil prices may affect the market’s growth. MEK is toxic, and exposure to high concentrations can cause health effects. MEK can be produced using refined oil products or microorganisms and renewable raw materials.

11.3.1 Methyl Ethyl Ketone Production Through a Conventional Process 11.3.1.1 MEK Production from Non-renewable Sources

1,3-Butadiene (1,3-BD) can be used as an intermediate compound in the production of MEK through steam cracking. MEK production typically involves butylene hydration, secondary butyl alcohol (SBA) production, and dehydration of SBA. The butylene used is obtained from petroleum cracking, and SBA is created through acid-catalysis at around 250 ∘ C. The processed SBA is then dehydrated to produce MEK (Figure 11.3). 2,3-Butanediol (2,3-BD) can be used to produce fine chemicals, antifreezes, and food products. MEK can be produced by direct dehydration of 2,3-BD. BDO dehydration using an amorphous calcium phosphate catalyst results in the production of BD, MEK, and by-products. Research focuses on improving the yield of target products using catalysts and optimizing reaction conditions. Song (2016) proposed a new method for developing and purifying MEK from 2,3-BD using a reactor-based approach with distillation columns. The general characteristics of the scheme in Figure 11.4, reported by Song et al. (2017), are shown in Table 11.3.

11.3.2 Purification of MEK Through Process-Intensified Schemes Although the use of MEK as an industrial chemical is well established, its potential as a biofuel must be competitive with existing biofuels. However, the production of MEK from bio-based sources is not yet fully developed. Various studies have suggested using biotechnological methods to convert pure sugar into MEK. This ideal scenario represents the maximum potential for MEK production, but the reported yields vary significantly. While it is possible to directly ferment sugar to produce MEK, the production efficiency is quite low, with yields of approximately 0.004 gMEK /gglucose (Penner et al. 2017). Another bio-based approach involves producing MEK through the decarboxylation of levulinic acid, which can be derived from lignocellulosic materials. However, the reported yields are not significant, even though acetic acid and acetone are produced as byproducts. A more promising alternative for MEK production is to utilize 2,3-BD as an intermediate. This method offers relatively high yields of 2,3-BD through fermentation, approaching the Olefins from petroleum

Figure 11.3

C4 fractionator

Acidcatalyzed

2-Butanol

Conventional process to manufacture MEK.

Dehydration

MEK

11.3 Methyl Ethyl Ketone

2,3-BD purification Column V300

Column V301

Column V302

Scrubber V200

Reaction unit Reactor 3-phase separator

Decanter

Quencher V100

Column V500

Column V400

Column V502

Column V501 MEK purification

Figure 11.4 (2017). Table 11.3

Production of MEK in a two-step process. Source: Adapted from Song et al.

General parameters of process route of Figure 11.4.

Description

V100 V200 V300

Number of stages

5

5

Pressure (kg/Sqcmg) Feed temperature (∘ C)

1.4

3.9

20 3.5

V301 V302 V400

V500 V501

80

30

3.8

56

10

3.5

1.5

1.1

62

V502

16

−0.35 −0.3

180

90

38.5

38

44.4

42.2

56.9

Overhead temperature (∘ C) Bottom temperature (∘ C)

51

38

40.6

42.1

40.8

94.1

85

52.8

69

80

128

44.4

44.1

131.1 106

72.2

175

Heat duty (MMkcal/h)

—

—

138 0.76

2.16

0.93

1.72

2.25

106

2.67

72.2

0.65

theoretical limit of 0.5 g2,3-BD /gglucose . Furthermore, the direct dehydration of 2,3-BD yields over 95% efficiency (Penner et al. 2017). The downstream process of producing MEK from 2,3-BD fermentation and dehydration has not been explored thoroughly, and by-products such as isobutyraldehyde (IBA) and 2,3-BD have market value. The mixture of MEK/IBA/2,3-BD/water is more thermodynamically complex, and current purification methods for MEK/water are not applicable. Penner et al. (2017) (see Figure 11.5) proposed four alternatives for purifying the MEK/2-Methylpropanal (2-MPL)/2,3-Butanediol (2,3-BD)/water mixture, but a comprehensive design strategy is lacking. Distillation remains the primary option for this difficult separation. Sánchez-Ramírez et al. (2021) proposed an improved method for purifying MEK from a mixture containing 65% MEK and other compounds (see Figure 11.6). Their approach involves using a liquid–liquid extraction column as the first step in the process, using p-xylene as the solvent to separate the azeotropes formed by the various compounds more

309

Water (IBA)

Water IBA MEK IBA

C4

C3

Water IBA MEK IBA

C5

C1 Water (IBA, MEK) C3

Water (IBA) C1

IBA

C2

IBA Water

MEK

2, 3-BD

C2

S1

S2

C4 Water (MEK)

(a)

(b)

2,3-BD

Water IBA MEK IBA

Water (IBA)

Water IBA MEK IBA

C4

C3

C1

Water

MEK

MEK

Water (IBA)

C1

C5

Water (IBA, MEK)

C3

IBA Water

IBA

C4

C2

S3

S4

C2 2, 3-BD

2,3-BD

(c)

Figure 11.5

(d)

Pure distillation alternatives for MEK purification. (a) Scheme S1, (b) scheme S2, (c) scheme S3, and (d) scheme S4.

MEK

IBA = 823.529 kg/h MEK = 7647.06 kg/h Water = 2117.65 kg/h 2-3 BD = 1176.47 kg/h

C4

C5

C3

MEK 7489.97 kg/h Water = 3.8 kg/h IBA = 1.1 kg/h p-Xylene = 36 998.1 kg/h

p-Xylene = 36 994.2 kg/h 2,3-BD, Water (traces)

C2

Make Up p-Xylene = 3.9 kg/h

Figure 11.6

IBA = 99.691 kg/h Water, MEK (traces)

IBA = 718.027 kg/h Water = 28.4 kg/h

Water = 72.67 kg/h IBA and MEK traces

Water = 2012.67 kg/h 2,3-BD = 19.91 kg/h 2,3-BD = 1146.94 kg/h Water, p-Xylene (traces)

Intensified alternatives for the MEK purification.

C6

MEK = 155.885 kg/h IBA (traces)

312

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

efficiently. Hybrid systems combining liquid–liquid extraction and distillation columns have been shown to reduce energy consumption in the downstream process by breaking down thermodynamic interactions between the components. Sánchez-Ramírez et al. (2021) evaluated different options for purifying MEK and used a hybrid stochastic optimization algorithm to optimize the study case. They considered four objective functions: total annual cost (TAC), eco-indicator 99, inherent process safety, and condition number. The results showed that most of the schemes were not feasible, except for scheme S2, which produced designs with high purities and high recoveries. The study also identified the placement of decanters as a crucial design parameter, and a Pareto front analysis was conducted to minimize the low feasibility of conventional schemes by adjusting the design parameters. The main design parameters and objective functions are outlined in Tables 11.4 and 11.5. Sánchez-Ramírez et al. (2021) found that the LLE-based hybrid process was the most effective solution for purifying MEK. This approach showed improved energy Table 11.4

Design parameters and performance indexes for the intensified scheme. LLE

C2

C3

C4

C5

10

33

45

45

54

3.483

0.529

16.636

5.01

Feed stage

1, 10

4

27

5

23

Column diameter (m)

1.455

1.285

1.407

1.544

1.098

Operative pressure (kPa)

101.353

101.353

101.353

101.353

101.353

Distillate flowrate (kmol/h)

111.997

123.297

19.193

12.9292

Condenser duty (kW)

5776

1693

3191

694

6354

4125

3202

727

Number of stages Reflux ratio

Reboiler duty (kW) H (%)

32.56

Total annual cost ($/yr)

7 903 251

Eco-indicator (points/yr)

1 338 593

Condition number

88 121

IR (probability/yr)

0.001 408 7

Table 11.5

Objective function for pure distillation schemes.

Tac ($/yr)

S1

S2

S3

S4

Hybridintensified

104 719 750

31 011 553

153 136 510

4 435 273

7 903 251

Eco-Ind (points/yr)

2 993 581 413

14 669 116

16 200 579

891 801 275

1 338 593

Condition number

3.8

3.99

4.78

147.5

88121

IR (probability/yr)

0.001 671 56

0.001 332 3

0.001 334 14

0.001 665 87

0.001 408 7

11.4 Intensification of Alcohol-to-Jet Fuel Process

use, energy efficiency, and thermodynamic performance compared to the distillation method. The results showed significant energy savings and met existing economic, environmental, controllability, and safety requirements. The relationship between the different design alternatives and objective functions was established, enabling the design to exploit the potential of the separation alternatives in different regions.

11.4 Intensification of Alcohol-to-Jet Fuel Process In recent years, there has been a growing concern over the environmental impact of aviation, which is one of the fastest-growing sources of greenhouse gas emissions in the transport sector. As a result, there has been a significant push to develop sustainable alternatives to traditional fossil fuels for use in aviation. One promising approach is the production of biojet fuels, which are derived from renewable biomass sources and have the potential to significantly reduce carbon emissions (IATA, 2022; IRENA, 2021). One promising technology for the production of biojet fuels is the alcohol-to-jet (ATJ) process, which converts alcohols such as ethanol and butanol into high-quality aviation fuels. The ATJ process offers several advantages over other biojet fuel production methods, including its ability to use a wide range of feedstocks, its compatibility with using several types of alcohols, which provides flexibility, and finally, the technologies used to convert the alcohols into jet fuel are well known with existing aviation infrastructure and its ability to produce fuels with high energy density and low emissions (Romero-Izquierdo et al., 2021). Despite the numerous advantages of the ATJ process, there are also some challenges associated with this technology that need to be addressed to ensure its economic and environmental viability. One significant disadvantage of the ATJ process is that it is currently more expensive than traditional jet fuel production methods. The cost of producing biojet fuels via the ATJ process is mainly due to the high cost of feedstock, as the process requires large quantities of alcohol. Additionally, the ATJ process requires additional processing steps compared to traditional jet fuel production, which further adds to the cost (Romero-Izquierdo et al., 2021). With this in mind, the implementation of PI in the ATJ process can be useful to reduce energy consumption and costs. The best example of an intensified process is RD, which combines chemical reactions and distillation in a single unit. RD offers several advantages over traditional reaction and separation processes, including reduced capital and operating costs, increased productivity, and better product quality (Stankiewicz and Moulijn, 2000). However, traditionally, the RD process is more complex compared to conventional alternatives, which can affect controllability properties and generate unsafe situations (Contreras-Zarazúa et al., 2017). Currently, there is no study that relates how RD column design affects control properties and safety properties. Therefore, in this work, the design, control, and safety properties of a catalytic column that integrates the oligomerization and hydrogenation stages of the ATJ process are studied. The reactive column was designed and optimized simultaneously using differential evolution with the tabu

313

314

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

list method. A multi-objective optimization problem was considered to design the column, where the TAC was the first objective and the condition number was the second objective function to evaluate the controllability properties. It is important to highlight that the safety properties were evaluated for different design solutions. The IR index was used as the metric to quantify the safety properties.

11.4.1 Process Modeling and Optimization Figure 11.7 shows a comparison between the RD column and the conventional process. Please note that in this case, ethylene, which is a product of the ethanol dehydration stage, is fed to the oligomerization stage to produce heavier alkene, preferentially in the range of C10–C15. Subsequently, the alkenes are fed to a hydrogenation stage where the double bond is broken to form linear hydrocarbons, which are separated to produce naphtha, jet fuel, and diesel. As can be seen, this reactive column combines the stages of oligomerization, hydrogenation, and separation stages in a single equipment. In this case, the oligomerization stages should be located near the condenser because it is an exothermic reaction, whereas the hydrogenation should be located near the bottom because this set of reactions is endothermic. This RD was modeled in the software ASPEN PLUS using the RADFRAC block. An ethylene feed flow rate of 2100 kg/h is considered. The thermodynamic model used to simulate and design the column is NRTL-RK since conditions and compounds are involved in the system. This thermodynamic model was chosen according to Carlson’s algorithm. The reactions performed in this process and their respective kinetic parameters were taken from Sánchez-Ramírez et al. (2022). Guthrie’s method was used to estimate the TAC, considering a payback period of 10 years. Stainless steel is considered as construction material; additionally, trays type sieve with a spacing of 0.61 m are considered. The utilities and their respective costs are cooling water ($0.355/GJ), electricity ($16.8/kWh), and fired heat ($20.92/GJ) (Turton, 2001). The operative costs were calculated considering 8500 yearly operations. The controllability properties and safety of the column were determined using the condition number and IR indexes, respectively. Please remember that in this work, the RD column was designed and optimized simultaneously, employing the differential evolution with tabu list (DELT) optimization method and considering only the TAC and condition number as objective functions. Therefore, the optimization problem used to design the column can be expressed as follows: min Z = [TAC, 𝛾] Subject to

TBP∗ ≥ yi ≥ TBP∗ wi ≥ 1000

(11.6)

where TBP* and TBP* represent the maximum and minimum boiling points that must fulfill the biojet fuel, these boiling points correspond to 300 and 220 ∘ C, respectively, according to the regulation established by ASTM. In addition, the optimization problem is also subject to reach a minimum jet fuel production (wi ) of 1000 kg/h.

Reactive zone 1 oligomerization Light compounds recycle Light gases

Reactive zone 2 hydrogenation Light gases

Ethylene

Ethylene

Hydrogen

Oligomers

Oleofins

Jet fuel

Hydrogen

Reactive zone 1 oligomerization Reactive zone 2 hydrogenation Jet fuel

Diesel Diesel

Figure 11.7

Intensification of ATJ process using catalytic distillation.

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

The decision variables used to design the column are the number of stages, feed stages, number of reactive stages, hold-up values, reflux ratio, and the reboiler duty. The simultaneous design and optimization of this reactive column were carried out using a hybrid platform that links Excel with ASPEN PLUS. The whole optimization method and the objective functions were implemented in Visual Basic Macro and an Excel spreadsheet, respectively. On the other hand, ASPEN PLUS models and their thermodynamic data were used to rigorously simulate the equipment. The parameters considered for the DETL method correspond to a population size of 120, a number of generations of 1250, tabu list size of 60, a crossover factor of 0.8, a mutation factor of 0.3, and a tabu radius of 0.01 (Contreras-Zarazúa et al., 2017).

11.4.2 Results In this subsection, the results obtained for the simultaneous design and optimization of the reactive column are presented. The optimization was performed on a computer with a Ryzen 5500 CPU and 32 GB RAM. The utopian point procedure was employed to select the solution with the best trade-off between the objectives. Additionally, extreme solutions that correspond to designs with high cost or condition numbers were considered in order to evaluate safety as a function of these two indicators. The points selected for this study are shown in Figure 11.8. MAXCON represents the solution with the minimum TAC but the worst controllability properties (highest condition number), while EP corresponds to the equilibrium point, which represents the solution with the best trade-off between both objectives. Finally, MAXTAC represents the solution with the best control properties but the highest costs. Some important design parameters for these three processes are shown in Table 11.6.

MAXCON 2.000E+07 1.800E+07 1.600E+07

Condition number

316

1.400E+07

EP

1.200E+07 1.000E+07 8.000E+06

MAXTAC

6.000E+06 4.000E+06 2.000E+06 0.000E+00

100 000 200 000 300 000 400 000 500 000 600 000 700 000 800 000 900 000

TAC ($/yr)

Figure 11.8

Pareto front CN vs. TAC for the reactive distillation column.

11.4 Intensification of Alcohol-to-Jet Fuel Process

Table 11.6

Design parameters for different reactive distillation processes.

Design variable

MAXCON process

EP process

MAXTAC process

Number of stages

29

29

29

Ethylene feed stage

22

21

23

Hydrogen feed stage

27

26

26

Jet-fuel side stream stage

26

26

25

Reactive stages (Oligomerization zone)

11–17

12–16

14–18

Reactive stages (Hydrogenation zone)

18–21

17–25

19–26

Reflux ratio

72.13

51.55

51.62

Reboiler duty (kW)

105.1

117.64

140.82

Hydrogen mass flowrate (kg/h)

20.69

20.64

20.25

Diameter (m)

1.2

1.39

2.06

Condition number

17 307 396

82.58

23.19

215 553

183 619.4177

Total annual cost ($/yr)

183 619.42

As can be seen in Table 11.6, the solutions obtained by the optimization method are quite similar. The optimization results indicate that the number of stages for all three solutions is 29, and the feed stages and side stream locations are also similar. However, the most noticeable difference among these processes resides in the reboiler duties and column diameters. Note that there is a tendency for control properties to worsen when the thermal load and diameter are smaller. This tendency can be explained by the fact that larger diameters imply that the equipment is bigger, and there is more contact area between the liquid and vapor phases, which helps to dampen a disturbance. On the other hand, hydrogenation is endothermic, so a decrease in reboiler duty represents a drop in temperature, which reduces the production of biojet fuel and worsens the control properties by imposing more restricted operating conditions. The results of the safety properties are presented in Figure 11.9. In this case, the IR index quantifies the process safety by calculating the risk due to instantaneous and continuous release incidents. Note that for this intensified equipment, the process improves when the cost is reduced. On the other hand, contrary to popular belief, the process is safer when you have worse control properties. Please note that IR is an inherent safety index, which means that the topology of the process, equipment size, and internal flows have an important effect on safety. For this reason, smaller equipment sizes and lower energy consumptions will tend to generate safer processes due to the reduction of the inventory of inflammable compounds (internal flows) inside the column, whereas this reduction in size and energy generates smaller equipment with more a constrained operation bound, which deteriorates the control properties of the process. This can be easily corroborated by analyzing the risk values for continuous and instantaneous incidents. In this case, the numerical value of this risk depends largely on the amount of material released, so smaller equipment will tend to release less material, reducing the likelihood of damage.

317

318

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts MAXCOND

EP

MAXTAC

3.50E-05 3.00E-05 2.50E-05 2.00E-05 1.50E-05 1.00E-05 5.00E-06 0.00E+00 Instantaneous risk

Figure 11.9

Continous risk

Total risk

Safety results for the intensified processes.

11.5 New Processes for Furfural and Co-products The use of renewable biomass as a source for developing chemicals has garnered significant research interest in recent years. Particularly, novel renewable building blocks like furfural have been a focal point. The U.S. Department of Energy has compiled a list of the top 30 building block chemicals derived from biomass that can serve as competitors to petroleum-derived chemicals. Notably, furfural and two of its derivatives, furan dicarboxylic acid and levulinic acid, are ranked in the top 10 of that list (Contreras-Zarazúa et al., 2018). Furfural finds diverse industrial applications and serves as a raw material for the production of various chemicals, including hexamethylenediamine (an intermediate for nylon 6-6 production) and phenol-furfural resins. Additionally, it has been demonstrated that furfural can be converted into value-added chemicals such as furfuryl alcohol, tetrahydrofurfuryl alcohol, and furan (Contreras-Zarazúa et al., 2018). Typically, furfural is produced from biomass that is rich in pentosan, such as sugar cane bagasse, corncobs, oat hulls, and sunflower husks. In 1922, the Quaker Oats Company pioneered an industrial-scale process for furfural production using oat hulls, sulfuric acid, and steam. Although this process is relatively simple to implement, it incurs high purification costs and exhibits low furfural conversion rates. Despite minimal changes over the years, the Quaker Oats process remains responsible for nearly 80% of the world’s furfural production (Contreras-Zarazúa et al., 2018). This chapter explores new intensified processes that have the potential to be more efficient, environmentally friendly, and offer better control properties compared to the conventional Quaker Oats process. These processes are based on design intensifications proposed by Nhien et al. (2017). The aim of the separation processes was

11.5 New Processes for Furfural and Co-products

0.1

0.9

0.2

0.2

0.7

0.3

Water

Acetic

0.4

0.8

0.3 0.2

0.8 0.1

0.1

0.9

97.79 °C

0.3

0.4

0.4

0.7

0.2

0.9

Methanol

0.5

0.6

0.3

0.2

0.6

0.5

0.7

0.1

0.7

0.4

Water

0.5

0.6

0.8

0.3

0.6

0.4

0.5

(a)

0.9

0.1 0.8

97.79 °C

0.5

0.6

Furfural

0.7

0.8

0.9

0.1

(b)

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Furfural

Figure 11.10 Ternary diagrams for mixtures: (a) water-furfural-acetic acid, (b) water-furfural-methanol.

to purify a feed flow rate consisting of 90% water, 6% furfural, 2% methanol, and 2% acetic acid, which corresponds to the average composition of furfural reactors. The chosen flow rate was intended to meet the estimated global furfural demand. These processes were simulated in Aspen Plus using the RADFRAC module. The NRTL-HOC equation of state was selected to model vapor–liquid–liquid equilibrium for this mixture due to its ability to handle the heterogeneous azeotrope, dimerization, and solvation of mixtures with carboxylic acids. Ternary diagrams for the ternary mixtures using mass fraction as the basis are shown in Figure 11.10. In Figure 11.11, various new separation processes are presented for the purification of furfural and co-products. These processes utilize liquid–liquid extraction columns in conjunction with distillation. Butyl chloride is used as the solvent for the extraction column as it has selectivity for recovering furfural from the mixture. In the distillation column C1, a solvent- and furfural-rich stream are separated. The raffinate stream generated from the extraction column contains water, methanol, and traces of solvent, which are separated using a set of distillation columns (see Figure 11.11). The separation processes are named according to the arrangement used to separate compounds of the raffinate stream. Various intensified processes are implemented in this zone, based on a direct and indirect scheme to purify the raffinate stream. These intensified arrangements are generated using the methodology proposed by Errico et al. (2009). The generation of new separation alternatives can be achieved by creating direct and indirect thermally coupled (ITC) schemes, where the reboiler and condenser associated with non-product streams are substituted with vapor and liquid interconnection streams. This involves dividing the columns into rectification and stripping sections and interchanging these sections between the columns to create thermally coupled equivalent sequences. During this stage, the direct thermodynamically equivalent scheme (DTES) and the indirect thermodynamically equivalent scheme (ITES) are generated. To intensify the process further, unnecessary sections are eliminated from the thermodynamic equivalent schemes and replaced by a side stream. The resulting alternatives are the

319

320

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

Entrainer

Entrainer

C1 C1

Furfural

Feed Fresh Entrainer

Furfural Feed

Entrainer C2

E1

Fresh Entrainer

Solvent

C3

Entrainer

I C3

II

I

Methanol

III

Solvent III

C2

E1

IV

Methanol

II IV

Wastewater Wastewater

Direct scheme to purify the raffinate stream

Indirect scheme to purify the raffinate stream

Entrainer

Entrainer

C1

C1

Furfural

Furfural

Feed Fresh Entrainer

Feed Fresh Entrainer

Entrainer C2

E1

C3

Entrainer

Solvent

III

Solvent

E1

IV

I C2 C3

Methanol

II

I

III

Methanol

II IV

Wastewater Wastewater

Indirect thermally coupled scheme (ITC)

Direct thermally coupled schemes (DTC)

Entrainer

Entrainer C1

C1

Furfural

Furfural

Feed

Feed Fresh Entrainer

Fresh Entrainer

Entrainer

Entrainer

E1

C2

E1

Solvent I

Methanol

C2

Solvent III

C3 III

I

II

C3

II

IV

IV

Wastewater Wastewater

Direct thermodynamic equivalent (DTES)

Entrainer

Entrainer C1

C1

Furfural

Furfural

Feed

Feed Fresh Entrainer

Entrainer

C2 E1

Methanol

Indirect thermodynamic equivalent scheme (ITES)

I

Solvent

Fresh Entrainer

Entrainer

C2 E1

Solvent III

Methanol II

I

Methanol IV II

Wastewater

Direct intensified scheme (DIS)

Figure 11.11

Wastewater

Indirect intensified scheme (IIS).

Separation alternatives for furfural and methanol purification.

direct intensified scheme (DIS) and the indirect intensified scheme (IIS). It should be noted that conventional direct and indirect schemes were utilized as benchmarks to assess the enhancements achieved by the intensified alternatives since all of these new processes were created from conventional schemes. It should also be noted that the generation of intensified processes is entirely dependent on the design of conventional ones. This is due to the complexity of the mixture being separated,

11.5 New Processes for Furfural and Co-products

which creates a lack of methodology for designing conventional processes. The conventional processes were designed using sensitivity analysis, which included the extractive column and distillation column C1. The objective of the sensitivity study was to achieve high purity levels for furfural, methanol, and water and to minimize solvent usage and associated costs.

11.5.1 Results This section presents the design results and evaluation of indexes for all separation alternatives. It is noteworthy that the extractive column and distillation column C1 are common equipment for all processes to process the raffinate stream and can therefore be designed separately from the set of columns to recover methanol. Sensitivity results and specifications for the final design of the extractive column and distillation column C1 are shown in Figure 11.12. Reductions in solvent amounts are

5.50E+06

TAC ($/yr)

5.00E+06

4.50E+06

4.00E+06

3.50E+06

3.00E+06 45 000

50 000

55 000

60 000

65 000

70 000

Entrainer mass flow (kg/h)

Reflux ratio = 0.46 P = 1 atm

Entrainer 46 845 kg/h 99.5% wt

1

2100 kg/h Methanol 2100 kg/Acetic acid 6300 kg/h Furfural 94 500 kg/h Water

1

15

E1 C1 Entrainer Q = 7800 kW

47 000 kg/h

18 18

Furfural 6297 kg/h 99.9% wt Aqueous phase 48640.7 kg/h

Figure 11.12

Design of extractive column and distillation column C1.

321

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

shown to result in cost savings, with a dot in Figure 11.12 representing an individual design for the extractive column and distillation column. The results indicate that the minimum TAC is achieved when 47 000 kg/h of solvent flow is used, and below this flow, the required purity of furfural cannot be achieved. This reduction in solvent flowrate represents a saving of approximately 25% compared to the value reported by Nhien et al. (2017). Furthermore, the reduced solvent flowrate leads to savings in energy, with the energy required for furfural purification being 7800 kW, representing a 7% saving compared to the value reported by Nhien et al. (2017). In this section, the design of conventional processes for methanol recovery was carried out using sensitivity analysis, similar to the extraction column section. Once the number of stages and location of feed stages were determined, the methodology reported by Errico et al. (2009) was applied. Additionally, an extra sensitivity study was performed to find the suitable mass flow rate that would reduce the energy consumption of the thermally coupled and thermally equivalent schemes. Figure 11.13 presents the results of the TAC and eco-indicator 99. The results reveal that the TAC is directly proportional to the environmental impact, implying that an increase in costs also results in an increase in environmental impact. This is expected, since these processes require high energy consumption for purification, resulting in high costs and environmental impact, owing to greenhouse gas emissions, resource depletion, and climate change. The conventional processes, both direct and indirect, have the largest TAC and EI99 values when compared to the intensified processes, signifying that intensified processes are more energy efficient. It is noteworthy that both thermally coupled (DTC and ITC) and thermally equivalent (TCED and TCEI) processes exhibit almost the same TAC and EI99 values. TAC

E199 7.00E+06

1.40E+07

6.00E+06

1.20E+07

8.00E+06

4.00E+06

6.00E+06

3.00E+06

4.00E+06

2.00E+06

2.00E+06

1.00E+06

0.00E+00

0.00E+00 Direct

Figure 11.13

Indirect

DTC

ITC

TCED

TCEI

IIS

Design of extractive columns ac distillation column C1.

EI99 (Eco-points/yr)

5.00E+06

1.00E+07

TAC ($/yr)

322

11.5 New Processes for Furfural and Co-products

This is due to the fact that these processes are thermodynamically equivalent, meaning that their energy consumption and efficiency are practically the same. For intensified alternatives, please note that only the IIS is displayed in Figure 11.13, as the direct alternative was not feasible. The ISS alternative was found to be the best option in terms of cost and environmental impact, owing to its lower number of equipment and energy consumption, when compared to other processes. The results of TAC and EI99 indicate that intensified processes have better prospects in terms of energy consumption, cost-effectiveness, and environmental impact. However, the selection of the best option becomes less clear when considering other parameters such as safety or control. Safety results for the separation options are shown in Figure 11.14. To evaluate safety, the IR index was used, which takes into account factors such as operational conditions, concentrations of organic compounds, and equipment size. Since methanol and solvent are separated at high purity and can be flammable and toxic at high concentrations, it is important to quantify safety. The intensified processes showed better energy efficiency, cost-effectiveness, and environmental impact, according to TAC and EI99 results. However, when considering other factors such as safety and control, the selection of the best option becomes less clear. The safety results for the separation options are shown in Figure 11.14. The study revealed that the thermally coupled and intensified processes did not show significant improvements in terms of safety compared to conventional processes. It is worth noting that some intensified options, such as the ITC alternative, exhibited worse safety values than their corresponding conventional processes. The IR index serves as a measure for quantifying the likelihood of injury or fatality resulting from accidents, and it is dependent on process topology, equipment size, and operating conditions. The safety index for intensified and conventional processes is similar as they perform the same separation, and the index is influenced more by equipment size 60

50

IR (1/yr) × 106

40

30

20

10

0 Direct

Figure 11.14

Indirect

DTC

Safety results.

ITC

TCED

TCEI

IIS

323

324

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

Table 11.7 Comparison of condition number for the different distillation processes. Configuration

Condition number (𝜸)

Direct

18.21

Indirect

40.46

DTC

48 403

ITC

546.49

TCDE

148.60

TCEI

1351

IIS

79.99

and separation topology. The ITC and indirect schemes have higher IR values because they concentrate methanol and solvent in column C3. Despite having a similar equipment size as its respective conventional process, the ITC process has a worse safety index, indicating that its topology promotes larger amounts and higher concentrations of methanol and solvent in column C3. The IIS configuration, which performs the separation of solvent, methanol, and water in a single column, showed the lowest IR value among all the studied alternatives. The reduced equipment count in this configuration results in a lower probability of failure or accidents, leading to improved safety. Additionally, the separation of methanol and solvent occurs in a column with a significant amount of water, decreasing the chances of explosions, toxic releases, and fires. Therefore, the IIS configuration offers the best safety characteristics among all the evaluated options. In conclusion, the condition number was examined as the final index in this study, and Table 11.7 displays the condition numbers for all distillation sequences. This number indicates how sensitive a system is to disturbance. A condition number close to 1 indicates that the process can easily handle the disturbance, whereas a value far from 1 suggests that the process may have difficulty handling the disturbance. The results in Table 11.7 reveal that the conventional sequences have superior control properties. Although interconnection streams usually improve the control properties of thermally coupled processes, in this case, even a small disturbance could result in significant changes in solvent and methanol compositions. Notably, the intensified option has similar control properties to the conventional alternatives due to its use of a single shell with a large amount of water to mitigate disturbances. Based on the best environmental impact, cost, safety, and control properties, the IIS alternative is selected as the optimal method for recovering methanol and solvent.

11.6 Lactic Acid Lactic acid (2-hydroxy propionic acid) is one of the large-scale chemicals produced via fermentation. In 2021, the global market for lactic acid ascended to 1.39 million

11.6 Lactic Acid

metric tons, which is projected to grow to approximately 2.65 million metric tons by the end of 2020s (Fernández, 2022). Lactic acid fermentation is one of the oldest fermentations, dating back to the 1880s. However, lactic acid can also be commercially produced by chemical synthesis. The main applications of lactic acid are in the chemical, pharmaceutical, food, cosmetic, and plastic industries (Prado-Rubio et al., 2020). A chemo-catalyzed approach is a production route of lactic acid characterized by producing a racemic mixture of D/L-lactic acid. On the other hand, the biological pathway allows for the production of either L-lactic acid or D-lactic acid independently, originating from selecting the correct microorganism in the fermentation process (E¸s et al., 2018). Today, more than 90% of the total production of lactic acid is by bacterial fermentation. The separation and purification of lactic acid at high purities (post-fermentation) are challenging due to its high affinity for water and tendency to decompose at high temperatures (Komesu et al., 2017a). Overall, the main parameters that determine the cost of lactic acid production are rate and yield in fermentation and recovery. One point to consider is that the high energy requirement results in high greenhouse gas emissions, which makes commercialization and viability a concern. Approximately 70% of the total cost of lactic acid production is the separation and purification stage (Komesu et al., 2017b). The recovery of non-volatile lactic acid from the fermentation broth is a difficult task due to the closeness of the boiling points. Thus, the use of PI, specifically RD, is a promising process for solving this problem (Cockrem and Johnson, 1993).

11.6.1 Lactic Acid Production by Reactive Distillation Within PI, various alternatives can be used to produce lactic acid. Hydrolysis with methanol, esterification, and oligomerization of lactic acid are some options. Kim et al. (2017) were the first to propose the feasibility of RD for lactic acid production by means of said options. In their work, Kim et al. (2017) show the real possibility of their implementation. The feasibility is presented by addressing several requirements, which are: One, the presence of more than one product. Two, that the reaction and separation temperatures coincide (the reaction is near 50 ∘ C and the boiling point of both reactants, CH3 OH and C3 H6 O3 , is 64 and 122 ∘ C, respectively). And three, that the operating conditions, such as pressure and temperature, are close to the critical region. Therefore, the reaction consumed in both columns is as follows: kf ,1

CH3 CH(OH)COOH + CH3 OH ←−−−−→ CH3 CH(OH)COOCH3 + H2 O ⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟ ⏟⏞⏟⏞⏟ kr,1 ⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟ ⏟⏟⏟ Lactic acid

Methanol

Metillactate kf ,2

Water

CH3 CH(OH)COOH + CH3 CH(OH)COOH ←−−−−→ C6 H10 O5 + H2 O ⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟ ⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟ kr,2 ⏟⏞⏟⏞⏟ ⏟⏟⏟ Lactic acid

Lactic acid kf ,3

Dilactate

C6 H10 O5 + CH3 CH(OH)COOH ←−−−−→ C9 H14 O7 + H2 O ⏟⏞⏟⏞⏟ ⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏞⏟ kr,3 ⏟⏞⏟⏞⏟ ⏟⏟⏟ Dilactate

Lactic acid

Trilactate

Water

(11.7)

(11.8)

Water

(11.9)

325

326

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

Water Methanol Methanol Feed

I

II

III

IV

Water Succinic acid

Figure 11.15

Lactic acid

Base case for the separation and purification of lactic acid (Conventional).

It is important to acknowledge that if the dome temperature is subjected to values above 80 ∘ C oligomerization may occur. Therefore, the operating pressure is important to prevent the oligomerization reaction. In addition, it is considered that the holding volume is occupied at 50% by the solid catalyst and that the catalyst density is 770 kg/m3 (Kim et al., 2017). A UNIQUAC-HOC thermodynamic model was used due to the high non-ideality and strong phase interactions (Mavalal and Moodley, 2020). The fermentation broth conditions to feed the RD column are 50 kmol/h, a pressure of 2 atm, a temperature of 35 ∘ C, and percentages of 8.4, 90.5, and 1.1 mol%, of lactic acid, water, and succinic acid, respectively (Kim et al., 2017). The base case is established by conventional distillation columns (see Figure 11.15).

11.6.2 Design and Synthesis of Intensified Processes For the synthesis and design of the RD columns, the methodology of Errico et al. (2009) was followed. The stages are as follows: Stage 1 consists of the specification of the reference configuration (see Figure 11.15). Stage 2 consists of the generation of a mutation of the reference configuration, changing the reboiler and/or condenser for interconnecting (liquid–vapor) streams. Figure 11.16a,b shows the replacement of the equipment by interconnection streams. Stage 3 considers ideal mixtures to establish thermodynamically equivalent configurations by moving column sections. In the case of azeotropic mixtures, it is not a straightforward task. Figure 11.16c,d shows the recombination of sections. Stage 4 consists of the construction of the side-stream configuration from the thermodynamically equivalent configurations (Figure 11.16e,f).

11.6.3 Optimization The optimization took into consideration the economic, environmental, safety, and control aspects. The evolutionary method used was DETL (Bonilla-Petriciolet et al., 2010). The optimization for the separation and purification sequences of lactic acid

11.6 Lactic Acid

I

II

III

IV

I

II

III

IV

(c)

(a)

I

II

III

IV

I

II

III IV

(d)

(b)

II

I

III

IV

(e)

I

II

III IV

(f)

Figure 11.16 Synthesis and design of the reactive distillation columns, (a) TC-I, (b) TC-II, (c) Petlyuk and (d) DWC, (e) Petlyuk-I, and (f) DWC-I.

were restricted by the purities (yi ) used industrially. Therefore, the objective function is subject to: Min (TAC, Eco99, IR, CN)

(11.10)

Subject to: yi ≥ xi

11.6.4 Results and Discussion It is important to mention that purity restrictions of a weight of 90% were established for lactic acid, a weight of 99.9% for succinic acid, and a weight of 99.9% for methanol. This is due to the different qualities demanded by the market. The food industry demands 80%, and the pharmaceutical and plastic industries demand 90%, for example. The results in Table 11.8 show that the DWC-I sequence has the lowest economic cost (lowest TAC) and the lowest environmental impact; it is important to highlight that this sequence is the second-best sequence in terms of inherent risk and one of the best in terms of operability. An important aspect to note is that intensification by means of thermal couplings (first stage) generates a significant improvement in most indicators. However, the

327

328

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

Table 11.8

Optimal results of each sequence. TAC (M$/yr)

EI99 (MEco-points/yr)

IR (probability/yr)

CN

Conventional

1.27

0.40

0.000 818 08

8.25E+24

TC-I

1.07

0.30

0.000 818 02

—

TC-II

1.13

0.37

0.000 823 00

9039.28

DWC

1.08

0.31

0.000 734 82

13 654.40

DWC-I

1.06

0.30

0.000 736 67

7887.28

difference between the elimination of a reboiler and a condenser is notorious. There is an 11% decrease in the TAC if columns I and II are thermally coupled, and with respect to the replacement of the condenser in columns III and IV, the TAC decreases by 17%. This reduction is generated by the large amount of energy that is mitigated. In the case of the TC-II sequence, an opposite phenomenon occurs, with a 76% increase in heating services compared to the conventional scheme. The TC-I scheme has an energy saving of 14% in the reboiler of column IV, and its behavior is similar to that of environmental impact (Eco99). An aspect to consider in this chapter is the visualization of safety behavior in intensified structures. In this regard, it can be observed that both column I and column II have a higher risk probability because there is a greater heating service that eliminates excess water. However, in the sequence TC-I, a different behavior occurs. The methodology that calculates the IR exhibits a great dependence on the amount of material associated with each equipment for the increment of the indicator. It is because of this that by intensifying the first two columns a continuous exchange of material is generated, thus producing a minimum increase in the inherent risk of the TC-I sequence (1%). In the DWC sequence, there are also economic savings regarding the conventional scheme. Even though the elimination of the reboiler leads to savings in capital cost, it is not enough to generate substantial savings in the TAC. Regarding the environmental impact, there is an improvement compared to the conventional sequence, but if compared to the TC-I sequence, there is only an increase of 3%. In practice, the existence of an inlet and outlet steam/liquid stream in the pre-fractionator subdues the net balance of matter present in the pre-fractionator; therefore, the IR is reduced by about 10%. The DWC-I sequence does not show any economic savings or a decrease in environmental impact compared to the DWC scheme. In contrast, an increase in the inherent risk in the separation scheme was caused. Operability was evaluated with study control. Table 11.8 shows that the dynamic properties of a conventional sequence in relation to an intensified one are poor. In other words, as safety improves in the intensification process, controllability improves as well. It is therefore apparent that for the sequences studied, the greater the number of couplings, the better the control properties. However, not all thermal couplings improve this indicator as a whole. Examining the topological differences between different alternatives, it should be noted that the thermal coupling

11.8 Conclusions

between the first two columns has the greatest influence on the dynamic properties of the alternatives. Thus, by considering a single thermal coupling, sustainability is increased.

11.7 Future and Perspectives In the highly competitive bioproduct market, traditional PI offers an excellent opportunity to witness improvements in economic aspects and in its environmental impact. Process systems engineering (PSE) is a tool that has contributed to the development of new and intensified technologies. However, there are several aspects on which to improve upon, such as: guaranteeing operable and functional intensified processes; estimating the feasibility of intensified processes in terms of safety; and systematically obtaining intensified designs. Therefore, new proposals to solve these issues have emerged. One of them is the synthetization of processes from phenomenological levels. This goes beyond the traditional concepts of unitary functioning and aims to propose new intensification schemes from a lower level of aggregation. With recent advances in PSE in operability/safety/controllability, attempts should be made to ensure the best operational performance of intensified technologies. Furthermore, as proposed by López-Guajardo et al. (2022), Process Intensification 4.0 defined as PI4.0, will help to improve equipment design in terms of safety and control. The PI4.0 perspective is to make use of artificial intelligence tools, particularly machine learning to accelerate the design of safe and operable intensified equipment. With more systematic tools, it will soon become possible to foresee the next generation of bioproduct intensification technologies that will usher in a new era.

11.8 Conclusions In the present chapter, an analysis was made to demonstrate certain considerations of operability and safety in intensified structures for the purification of bioproducts (methyl ethyl ketone, furfural, and lactic acid). It was observed that by intensifying processes, it is possible to obtain sequences with performance in terms of operability and safety equal to or better than conventional schemes. In addition, economic aspects and the environmental impact were evaluated, where the intensified schemes generally show gains or benefits. In general terms, the intensification of processes helps to reduce the size of the equipment and the external services (vapor/cooling water). Consequently, inherent safety benefits by reducing the area where accidents might occur and the probability of accidents. It should be considered that the IR values depend to a great extent on the amount of matter within the equipment, which is related to the size of the columns. As the interior matter increases, the IR increases. Operability showed some important points: alternatives with higher topological simplicity do not have the best control properties, and it is possible to generate better controllability in intensified schemes.

329

330

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

However, it should be mentioned that controllability is associated with higher costs and greater environmental impact. In other words, these indicators tend to compete with each other, and the lower the monetary cost we desire, the more we tend to sacrifice in terms of environmental impact and process operability.

Acknowledgments The authors acknowledge CONACyT (Mexico).

Acronyms BLEVE DETL EI99 HAZOP IR MEK PI QRA RD SVD TAC UVCE

boiling liquid expanding vapor explosion differential evolution with tabu list Eco-Indicator 99 hazard and operability study individual risk methyl ethyl ketone process intensification quantitative risk analysis reactive distillation singular value decomposition total annual cost unconfined vapor cloud explosion

Nomenclature F fi FM P Px,y T U V 𝜎 max 𝜎 min

flow rate (kg/s) frequency of the accident molar flow (mol/s) pressure (atm) probability of affectation in a specific position Temperature (o C) matrix of left singular vectors matrix of right singular vectors maximum singular value minimum singular value

References Amyotte, P.R., Brown, K.R., Khan, F.I., and Alauddin, M. (2023). Future of inherently safer design. In: Methods in Chemical Process Safety, vol. 7, 379–388. Elsevier.

References

Athar, M., Shariff, A.M., and Buang, A. (2019). A review of inherent assessment for sustainable process design. Journal of Cleaner Production 233: 242–263. Becker, L., Sturm, J., Eiden, F., and Holtmann, D. (2023). Analyzing and understanding the robustness of bioprocesses. Trends in Biotechnology 41 (8): 1013–1026. Bonilla-Petriciolet, A., Rangaiah, G.P., and Segovia-Hernández, J.G. (2010). Evaluation of stochastic global optimization methods for modeling vapor–liquid equilibrium data. Fluid Phase Equilibria 287 (2): 111–125. Cabrera-Ruiz, J., Santaella, M.A., Alcántara-Ávila, J.R. et al. (2017). Open-loop based controllability criterion applied to stochastic global optimization for intensified distillation sequences. Chemical Engineering Research and Design 123: 165–179. Center for Chemical Process Safety (CCPS) (2010). Guidelines for Risk Based Process Safety. Wiley. Cockrem, M.C. and Johnson, P.D. (1993). Recovery of lactate esters and lactic acid from fermentation broth. US Patent No. 5,210,296. Washington, DC: U.S. Patent and Trademark Office. Contreras-Zarazúa, G., Vázquez-Castillo, J.A., Ramírez-Márquez, C. et al. (2017). Multi-objective optimization involving cost and control properties in reactive distillation processes to produce diphenyl carbonate. Computers & Chemical Engineering 105: 185–196. Contreras-Zarazúa, G., Sánchez-Ramírez, E., Vazquez-Castillo, J.A. et al. (2018). Inherently safer design and optimization of intensified separation processes for furfural production. Industrial & Engineering Chemistry Research 58 (15): 6105–6120. D’Souza, R.N., Azevedo, A.M., Aires Barros, M.R., et al. (2013). Emerging technologies for the integration and intensification of downstream bioprocesses. Errico, M., Rong, B.G., Tola, G., and Turunen, I. (2009). A method for systematic synthesis of multicomponent distillation systems with less than N-1 columns. Chemical Engineering and Processing: Process Intensification 48 (4): 907–920. E¸s, I., Khaneghah, A.M., Barba, F.J. et al. (2018). Recent advancements in lactic acid production – a review. Food Research International 107: 763–770. Fernández, L. (2022). Global lactic acid market value 2015–2029. Statista. statista.com/ statistics/1310499/lactic-acid-market-value-worldwide/ (accessed 01 November 2022). Gao, X., Raman, A.A.A., Hizaddin, H.F., and Bello, M.M. (2020). Systematic review on the implementation methodologies of inherent safety in chemical process. Journal of Loss Prevention in the Process Industries 65: 104092. Govasmark, E., Stäb, J., Holen, B. et al. (2011). Chemical and microbiological hazards associated with recycling of anaerobic digested residue intended for agricultural use. Waste Management 31 (12): 2577–2583. Häggblom, K.E. and Waller, K.V. (1992). Control structures, consistency, and transformations. In: Practical Distillation Control (ed. W.L. Luyben), 192–228. New York, NY: Springer. Hassim, M.H. and Hurme, M. (2010). Inherent occupational health assessment during process research and development stage. Journal of Loss Prevention in the Process Industries 23 (1): 127–138.

331

332

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

Hernández, S. and Jiménez, A. (1999). Design of energy-efficient Petlyuk systems. Computers & Chemical Engineering 23 (8): 1005–1010. Hoppe, F., Burke, U., Thewes, M. et al. (2016). Tailor-made fuels from biomass: potentials of 2-butanone and 2-methylfuran in direct injection spark ignition engines. Fuel 167: 106–117. Huang, Y. and Peng, X. (2014). Sustainable chemical product and process engineering. ACS Sustainable Chemistry & Engineering 2 (1): 1–2. IATA (2022). Air passenger numbers to recover in 2024. https://www.iata.org/en/ pressroom/2022-releases/2022-03-01-01/ (accessed 22 January 2023). IRENA (2021). Reaching zero with renewbles. https://www.irena.org/-/media/Files/ IRENA/Agency/Publication/2020/Sep/IRENA_Reaching_zero_2020.pdf (accessed 17 January 2023). Jiménez-González, C. and Constable, D.J. (2011). Green Chemistry and Engineering: A Practical Design Approach. Wiley. Jiménez-González, C., Constable, D.J., and Ponder, C.S. (2012). Evaluating the “Greenness” of chemical processes and products in the pharmaceutical industry—a green metrics primer. Chemical Society Reviews 41 (4): 1485–1498. Khan, F.I. and Abbasi, S.A. (1999). Major accidents in process industries and an analysis of causes and consequences. Journal of Loss Prevention in the Process Industries 12 (5): 361–378. Khan, S., Naushad, M., Lima, E.C. et al. (2021). Global soil pollution by toxic elements: current status and future perspectives on the risk assessment and remediation strategies – a review. Journal of Hazardous Materials 417: 126039. Kidam, K., Sahak, H.A., Hassim, M.H. et al. (2016). Inherently safer design review and their timing during chemical process development and design. Journal of Loss Prevention in the Process Industries 42: 47–58. Kim, S.Y., Kim, D.M., and Lee, B. (2017). Process simulation for the recovery of lactic acid using thermally coupled distillation columns to mitigate the remixing effect. Korean Journal of Chemical Engineering 34 (5): 1310–1318. Kletz, T. and Amyotte, P. (2019). What Went Wrong?: Case Histories of Process Plant Disasters and How They Could Have Been Avoided. Butterworth-Heinemann. Komesu, A., Maciel, M.R.W., and Maciel Filho, R. (2017b). Separation and purification technologies for lactic acid – a brief review. BioResources 12 (3): 6885–6901. Komesu, A., Martins Martinez, P.F., Lunelli, B.H. et al. (2017). Study of lactic acid thermal behavior using thermoanalytical techniques. Journal of Chemistry 2017: 1–7. Kumar, A. (1996). Guidelines for Evaluating the Characteristics of Vapor Cloud Explosions, Flash Fires, and Bleves. Center for Chemical Process Safety (CCPS) of the AIChE, Published by the American Institute of Chemical Engineers, New York, NY (1994), 387 pages, [ISBN: 0-8169-0474-X], US List Price: $150. Leveson, N.G. (2016). Engineering a Safer World: Systems Thinking Applied to Safety, 560. The MIT Press. López-Guajardo, E.A., Delgado-Licona, F., Álvarez, A.J. et al. (2022). Process intensification 4.0: a new approach for attaining new, sustainable and circular processes enabled by machine learning. Chemical Engineering and Processing Process Intensification 180: 108671.

References

Mannan, M.S., Reyes-Valdes, O., Jain, P. et al. (2016). The evolution of process safety: current status and future direction. Annual Review of Chemical and Biomolecular Engineering 7: 135–162. Mannan, S. (2011). Lees’ Loss Prevention in the Process Industries: Hazard Identification, Assessment and Control, vol. 3. Elsevier. Martinez-Gomez, J., Ramírez-Márquez, C., Alcántara-Ávila, J.R. et al. (2017). Intensification for the silane production involving economic and safety objectives. Industrial & Engineering Chemistry Research 56 (1): 261–269. Martinez-Gomez, J., Sánchez-Ramírez, E., Quiroz-Ramírez, J.J. et al. (2016). Involving economic, environmental and safety issues in the optimal purification of biobutanol. Process Safety and Environmental Protection 103: 365–376. Mavalal, S. and Moodley, K. (2020). Isothermal vapour-liquid equilibrium measurements for the water + butane-1,4-diol/butane-2,3-diol system within 353.1–373.2 K. Fluid Phase Equilibria 512: 112518. Medina-Herrera, N., Grossmann, I.E., Mannan, M.S., and Jiménez-Gutiérrez, A. (2014a). An approach for solvent selection in extractive distillation systems including safety considerations. Industrial & Engineering Chemistry Research 53 (30): 12023–12031. Medina-Herrera, N., Jiménez-Gutiérrez, A., and Mannan, M.S. (2014b). Development of inherently safer distillation systems. Journal of Loss Prevention in the Process Industries 29: 225–239. Moore, C. (1986). Application of singular value decomposition to the design, analysis, and control of industrial processes. In: 1986 American Control Conference, 643–650. IEEE. Narayanan, H., Luna, M.F., von Stosch, M. et al. (2020). Bioprocessing in the digital age: the role of process models. Biotechnology Journal 15 (1): 1900172. Nhien, L.C., Long, N.V.D., Kim, S., and Lee, M. (2017). Techno-economic assessment of hybrid extraction and distillation processes for furfural production from lignocellulosic biomass. Biotechnology for Biofuels 10 (1): 1–12. Penner, D., Redepenning, C., Mitsos, A., and Viell, J. (2017). Conceptual design of methyl ethyl ketone production via 2,3-butanediol for fuels and chemicals. Industrial & Engineering Chemistry Research 56 (14): 3947–3957. Ponce-Ortega, J.M., Al-Thubaiti, M.M., and El-Halwagi, M.M. (2012). Process intensification: new understanding and systematic approach. Chemical Engineering and Processing: Process Intensification 53: 63–75. Prado-Rubio, O.A., Gasca-González, R., Fontalvo, J. et al. (2020). Design and evaluation of intensified downstream technologies towards feasible lactic acid bioproduction. Chemical Engineering and Processing Process Intensification 158: 108174. Romero-Izquierdo, A.G., Gómez-Castro, F.I., Gutierrez-Antonio, C. et al. (2021). Intensification of the alcohol-to-jet process to produce renewable aviation fuel. Chemical Engineering and Processing Process Intensification 160: 108270. Ruiz-Ruiz, F., Benavides, J., and Rito-Palomares, M. (2013). Scaling-up of a B-phycoerythrin production and purification bioprocess involving aqueous two-phase systems: practical experiences. Process Biochemistry 48 (4): 738–745.

333

334

11 Operability and Safety Considerations in Intensified Structures for Purification of Bioproducts

Sánchez-Ramírez, E., Huerta-Rosas, B., Quiroz-Ramírez, J.J. et al. (2022). Optimization-based framework for modeling and kinetic parameter estimation. Chemical Engineering Research and Design 186: 647–660. Sánchez-Ramírez, E., Quiroz-Ramírez, J.J., Contreras-Zarazúa, G. et al. (2021). Intensified alternative to purify methyl-ethyl ketone in a framework of green process. Energy 220: 119641. Segovia-Hernández, J.G., Sánchez Ramírez, E., Ramírez Márquez, C., and Contreras Zarazúa, G. (2021). Improvements in Bio-Based Building Blocks Production Through Process Intensification and Sustainability Concepts, 1e. Elsevier. Song, D. (2016). Kinetic model development for dehydration of 2,3-butanediol to 1,3-butadiene and methyl ethyl ketone over an amorphous calcium phosphate catalyst. Industrial & Engineering Chemistry Research 55 (45): 11664–11671. Song, D., Yoon, Y.G., and Lee, C.J. (2017). Conceptual design for the recovery of 1,3-butadiene and methyl ethyl ketone via a 2,3-butanediol-dehydration process. Chemical Engineering Research and Design 123: 268–276. Srinivasan, M., Venkatesan, M., Arumugam, V. et al. (2019). Green synthesis and characterization of titanium dioxide nanoparticles (TiO2 NPs) using Sesbania grandiflora and evaluation of toxicity in zebrafish embryos. Process Biochemistry 80: 197–202. Stankiewicz, A.I. and Moulijn, J.A. (2000). Process intensification: transforming chemical engineering. Chemical Engineering Progress 96 (1): 22–34. Thomassen, G., Van Dael, M., Van Passel, S., and You, F. (2019). How to assess the potential of emerging green technologies? Towards a prospective environmental and techno-economic assessment framework. Green Chemistry 21 (18): 4868–4886. Turton, R. (2001). Analysis, Synthesis and Design of Chemical Process, 4e, vol. 40, 163–200. Prentice Hall. Yuan, Z., Chen, B., Sin, G., and Gani, R. (2012). State-of-the-art and progress in the optimization-based simultaneous design and control for chemical processes. AICHE Journal 58 (6): 1640–1659.

335

12 Analysis of Safety and Economic Objectives for Intensified Algal Biodiesel Process Gunavant Deshpande, Ashish N. Sawarkar, and Dipesh Shikchand Patle Department Chemical Engineering, Motilal Nehru National Institute of Technology Allahabad, Prayagraj 211 004, Uttar Pradesh, India

12.1 Introduction Due to the depletion of conventional natural resources, growing emissions of combustion-generated contaminants, and increasing prices, renewable sources are becoming increasingly alluring alternatives (Ahmed et al., 2021; Athar and Zaidi, 2020; Goh et al., 2019). Petroleum-based fuel reserves are limited and centered in some regions of the world (Sawarkar, 2019a). These sources are approaching the end of their lifespans (de Jesus et al., 2020; Makareviciene and Skorupskaite, 2019; Sawarkar, 2019b). Because recognized petroleum reserves are limited, renewable sources of energy will become increasingly attractive. The replacement of conventional petro-fuels (coal, natural gas, and petroleum oil) with sustainable alternatives (such as biofuels) is highly desired because of the exhaustion of fossil fuels and their negative ecological influence (Chisti, 2008). The primary benefits of biofuels include reduced greenhouse gas emissions, improved sustainability, and increased energy stability. Considering many drawbacks of biodiesel production from first-/secondgeneration feedstocks (such as edible oil, inedible oil, waste cooking oil, etc.), microalgae-based biodiesel production (referred to as third-generation biofuel) is an attractive option because microalgae can be grown on non-fertile land, has a fast production rate, and has higher lipid productivity. Further, microalgae have been considered as feasible feedstock for biodiesel production as it does not compete with food (Patle et al., 2020a). According to reports, numerous algal species with higher lipid content for biodiesel production have been identified, such as Chlorella sp., Chlorella vulgaris, Chlorella sorokiniana, Chlorella protothecoides, Spirulina platensis, Nannochloropsis gaditana, and Nannochloropsis salina. Unfortunately, the cost of producing biodiesel from algae feedstock is recognized as prohibitory for commercial viability (Kim et al., 2019). To avoid the two-stage process (i.e. lipid extraction followed by transesterification reaction), in situ transesterification was invented as a “one-pot” technique (Ahmed et al., 2021). Control and Safety Analysis of Intensified Chemical Processes, First Edition. Edited by Dipesh Shikchand Patle and Gade Pandu Rangaiah. © 2024 WILEY-VCH GmbH. Published 2024 by WILEY-VCH GmbH.

336

12 Analysis of Safety and Economic Objectives for Intensified Algal Biodiesel Process

In situ, transesterification is beneficial in producing biodiesel because it reduces the lipid extraction and solvent recovery stages while also lowering the investment and overhead costs. Large-scale biodiesel production from microalgae usually necessitates longer extraction and transesterification times (Ganesan et al., 2021). Several methods for reducing these times have been described in the literature (Ahmed et al., 2021), including thermochemical methods and the use of supercritical solvents. However, these processes have several drawbacks, such as high cost and safety concerns. In situ processing makes the biodiesel production process easier and reduces industrial waste, time, and associated costs. In general, the existing chemical processes need to be intensified to improve their efficiency in consideration of profit and greenhouse gas emissions. Distillation is one of the important processes that is highly energy-intensive. Reactive distillation (RD) and divided wall column (DWC) have been sufficiently explored by researchers. However, the combination of the above two processes with the vapor recompression technique still requires investigation in terms of its impact pertaining to feasibility, economics, and safety. In our previous study (Shrikhande et al., 2021), the biodiesel process was intensified by DWC and DWC with multi-stage vapor recompression (i.e. DWC-MVR). The DWC-MVR process was found to be the most economical and the least polluting in terms of carbon emissions. To boost the efficiency, costs, profitability, safety, and durability of industrial systems, chemical plants must be optimized in terms of design and operating conditions. Optimization is the process of identifying and comparing viable solutions until no better solution is found. Optimization problems in chemical engineering and other disciplines often have multiple objective functions associated with performance, economics, safety, and reliability that must be optimized concurrently because these objective functions may be either entirely or partially conflicting over the range of interest (Fouladvand et al., 2021; Rangaiah et al., 2020; Sharma et al., 2013). Multi-objective optimization (MOO) is concerned with the solution of such problems. The MOO of a problem is expected to yield a set of equally good solutions, known as the Pareto-optimal front or solutions. Decision-makers can choose one of these solutions, for implementation (Rangaiah et al., 2020). Besides process economics, process safety is one of the main priorities in the chemical process industries. The number of events attributed to hazards in the process equipment is increasing, resulting in loss of containment, contamination of the environment, and human damage (Khan and Amyotte, 2004). As a result, it is important to take into account quantitative process safety indices as objectives to evaluate the safety performance of the processes; such indices include the damage index (DI) (Khan and Amyotte, 2005) and individual risk (IR) (De Haag and Ale, 2005; Sánchez-Ramírez et al., 2019). Contreras-Vargas et al. (2019) proposed the process alternatives (i.e. hybrid sequence, thermally coupled sequence, thermodynamically equivalent sequence, and DWC intensified sequence) for the separation of acetone, butanol, and ethanol mixture. After the optimization of these processes, the DWC intensified process was found to be the most promising as it reduced the safety risk by 66%.

12.2 Process Development

Sánchez-Ramírez et al. (2019) studied the optimization of the process alternatives for the production of 2,3-butanediol. It was reported that intensified alternative processes improved the economic, environmental, and safety indicators by 15%, 14%, and 50%, respectively, compared to the conventional distillation system. Mondal and Jana (2019) verified the techno-economic feasibility of biodiesel production through the RD route. They reported 53% savings in total annual cost, 40% reduction in CO2 emissions, and 43% savings in energy consumption for the RD-intensified process compared to the conventional multiunit system. Sánchez-Ramírez et al. (2020) evaluated the quaternary DWC for energy requirement, environmental impact, safety, and controllability with respect to conventional multi-unit distillation systems. The quaternary DWC process improved the considered objectives and might be considered as a viable alternative. This study investigates the effect of DWC-MVR intensification on the economics and safety of the process. The biodiesel process developed by Shrikhande et al. (2021), which was further modified by Deshpande et al. (2022a) by replacing the two flash columns with two distillation columns, is used in this study as alternative 1. Deshpande et al. (2022a) optimized the algal biodiesel process with variable capacity for the objectives pertaining to economics (i.e. total annual cost), safety (i.e. IR), and environment (i.e. carbon emission and organic waste). Optimization of this process proved beneficial in terms of all objectives compared to those of the non-optimized process. The process reported in Deshpande et al. (2022a) is optimized in this study for the simultaneous minimization of IR and break-even cost (BEC) using the elitist non-dominated sorting genetic algorithm (NSGA-II) for the fixed capacity. Another process (i.e. DWC-MVR-based biodiesel process) proposed by Shrikhande et al. (2021), which was later modified by replacing two flash columns with the two distillation columns and optimized for IR and BEC by Deshpande et al. (2022b), is considered in this study as alternative 2 (i.e. intensified case). Deshpande et al. (2022b) conducted the MOO of the process with fixed plant capacity and reported that optimization of DWC-MVR intensified biodiesel process improved the economics (i.e. BEC), environmental indicators (i.e. Eco-indicator 99) and safety indicators (i.e. IR). In this chapter, both process alternatives are compared to understand the effect of process intensification on process economy and safety. A detailed comparison between both alternatives is conducted for BEC (as an economic indicator) and IR (as a safety indicator).

12.2 Process Development 12.2.1 Process Development of Alternative 1 The process is designed to process 20 mt of lipids per year, as shown in Figure 12.1. The process is simulated in Aspen Plus V10 using a non-random two-liquid (NRTL) model to estimate the thermodynamic properties. The properties of the BBAIL (i.e. benzimidazolium based Brønsted acid ionic liquid) catalyst are generated in Aspen Plus by providing the structure, as the BBAIL catalyst is not available in the

337

338

12 Analysis of Safety and Economic Objectives for Intensified Algal Biodiesel Process

Aspen Plus library (Shrikhande et al., 2021). Wet microalgae, i.e. Scenedesmus sp., used in this study, contains 83.4 wt% of water and 16.4 wt% microalgae. Microalgae contains ∼30 wt% of lipid on a dry basis, while the remaining composition consists of proteins and carbohydrates. The reaction parameters, i.e. activation energy and pre-exponential factor, were taken from Gautam et al. (2021). The experimental validation is done by comparing the conversion/yield from Aspen Plus with the experimental results of our research group, obtained earlier. Details about the complete process development are given by Deshpande et al. (2022a). Microalgal biomass contains a variety of triglycerides (TGs) as lipids. However, for simplicity, lipids in the feedstock are considered triolein. Many researchers (e.g. Piemonte et al., 2016; Sharma and Rangaiah, 2012; Patle et al., 2020b) have made a similar assumption. Biodiesel is produced as per the following reaction, where TGs react with alcohol (i.e. methanol) to produce methyl esters (i.e. biodiesel) and glycerol (GL) TG + 3CH3 OH ↔ 3R′ COOCH3 + GL

(12.1)

Here, R′ denotes the alkyl group in TG. In the biodiesel process alternative 1 (Figure 12.1), wet microalgae is fed to the reactor “RTRANS” (RCSTR unit in Aspen Plus) maintained at 45 o C and 1 atm at a mass flow rate of 50 322 kg/h. Fresh methanol and BBAIL catalysts are sent to the RTRANS at a flow rate of 1038.8 and 200 kg/h, respectively. Methanol is recovered partially, and the BBAIL catalyst is recovered completely from the process and reused. Biodiesel and GL are produced in RTRANS and are processed further. RTRANS outlet mixture contains unreacted lipids, oleic acid methyl ester (i.e. FAME or biodiesel), carbohydrates, proteins, chlorophyll, BBAIL, GL, hexane, and methanol. Carbohydrates, proteins, and chlorophyll from the reactor outlet are separated using component splitter “SEP-1”. Unreacted lipids, biodiesel, GL, BBAIL, methanol, and water are sent to the phase separator “PS-1” and then to “PS-2” (DECANTER block in Aspen Plus), operated at 1 atm and 45 o C. In PS-1 and PS-2, a calculated amount of hexane is added as an extracting solvent. In PS-1 and PS-2, organic and aqueous phases are separated. Organic phases from both separators containing biodiesel, hexane, and TG are sent to the first distillation column FRAC-1 (RADFRAC block). In FRAC-1, hexane is distilled out from the column at −12 ∘ C, which is recycled back to the phase separator after exchanging heat through HX-2, HX-3, and HX-4 in order to heat hexane stream to about 40 o C before entering PS-1 and PS-2. The bottom product from FRAC-1, containing biodiesel and TG, is sent to FRAC-2 (RADFRAC block). In FRAC-2, 98% pure biodiesel is obtained at the top after washing it with water to remove the traces of impurity. Unreacted TG from the bottom is recycled back to the RTRANS. The aqueous phase from the phase separator “PS-2” containing water, GL, methanol, and BBAIL catalyst is fed to the FRAC-3 (RADFRAC block) column, where GL and BBAIL catalyst are separated from the bottom. The methanol and water mixture are further sent to FRAC-4 (RADFRAC block). In FRAC-4, methanol is obtained from the top and recycled back to the reactor RTRANS, whereas water is obtained from the bottom. GL and BBAIL catalyst obtained from the bottom of the FRAC-3 are fed to R-ACET (RSTOIC block), where GL reacts with acetic acid to produce triacetin (TAG) as a

IL F = 0 kg/h T = 25 °C XIL = 1

FRESH ALGAE F = 50322 kg/h T = 30 °C XTG = 0.05, XWAT = 0.816

M-1

Qc = –10.2 MW Qb = 7 MW

0.05 atm

SEP - 1 1

atm

PRO - CARB

BIODIESEL F = 2509 kg/hr T = 25 °C XFAME = 0.98

WAT - 1 F = 29.1 kg/h T = 30 °C XWAT = 1

RTRANS P-1

4

WASH - 1

M-2

45 °C

7

WATER OUT

8

H-1 F = 0.05 kg/h T = 30 °C XHE = 1

FRAC - 1

M-3

Qc = - 0.35 MW Qb = 0.84 MW 0.05 atm

HX - 1

SPLIT

P-2

PS - 1 M-4

PS - 2 Qc = –46 MW Qb = 47.2 MW

Qc = –25.3 MW Qb = 28.2 MW

1 atm

1 atm 6

5 FRAC - 3 6

7

HEX - REC F = 88380 kg/h T = 57 °C XHE = 0.967

FRAC - 2 HX - 3

MA - REC F = 42070 kg/h T = 63 °C XMA = 0.9

15

P-4

WATER F = 41796 kg/h T = 30 °C XWAT = 0.98

HX - 4

12

M-5

WAT – VAP - 2

AA F = 612 kg/h T = 30 °C AA XAA = 1

R - ACET

M-6

1 atm

Catalyst Reuse

80 °C

Qc = –0.44 MW Qb = 0.44 MW 1 atm

SEP - 2 0.05 atm FRAC - 5

AA - REC F = 905.9 kg/h T = 102 °C XAA = 0.79 XWAT = 0.209

7

1 atm

Qc = –0.72 MW Qb = 0.84 MW 13

DIST - 5 F = 1219.7 kg/h T = 100.4 °C XAA = 0.69 XWAT = 0.309

6

FRAC - 6

DIST - 6 F = 313.81 kg/h T = 99 °C XWAT = 0.5974 XAA = 0.4024 WAT - 2 F = 540.5 kg/h T = 30 °C XWAT = 1 WASH - 2

TG - REC F = 1533 kg/h T = 25 °C XTG = 0.93

16

FRAC - 4 11

HX - 2

P-3

8

HT - 1

0.05 atm

MA F = 1039 kg/h T = 25 °C XMA = 1

FLASH - 1

V-1

WAT - VAP- 1

0.05 atm HT - 2

V-2

12

P-5

FLASH - 2

TAG F = 572 kg/h T = 30 °C XTAG = 0.97 XDAG = 0.02

IL - REC F = 176.5 kg/h T = 100 °C XIL = 0.86

P - 6 CL - 1

Figure 12.1 Biodiesel synthesis from wet microalgae with conversion of glycerol to triacetin. Source: Reproduced with permission from Elsevier, Deshpande et al. (2022a).

340

12 Analysis of Safety and Economic Objectives for Intensified Algal Biodiesel Process

value-added product as per the following reaction (12.2a–12.2c), where monoacetin (MAG) and diacetin (DAG) are the intermediate products during the formation of TAG. Glycerol (Gly) + Acetic Acid (AA) ↔ Monoacetin (MAG) + Water

(12.2a)

Monoacetin (MAG) + Acetic Acid (AA) ↔ Diacetin (DAG) + Water (12.2b) Diacetin (DAG) + Acetic Acid (AA) ↔ Triacetin (TAG) + Water

(12.2c)

TAG, unreacted acetic acid, and BBAIL catalyst are sent to the distillation columns FRAC-5 and FRAC-6 to recover the acetic acid and mixture of BBAIL catalyst and TAG. The catalyst and TAG are separated in the wash column to obtain 97% TAG and 86% catalyst. The catalyst is reused in the transesterification reactor “RTRANS.”

12.2.2 Process Development of Alternative 2 Shrikhande et al. (2021) reported that about 70% of the total steam consumption of the process is attributed to FRAC-3 and FRAC-4. They combined FRAC-3 and FRAC-4 into a DWC as the composition and temperature profiles were favorable. Figure 12.2 shows the process flowsheet for alternative 2. DWC has a vertical wall placed at the center of the column. Accordingly, it has four sections left section (FRAC-3), right section (FRAC-4), stripping section (including a reboiler), and rectifying section (including a condenser). The heavier phase from PS-2 is fed to the 3rd stage of the left section. Reflux from the top is distributed on each side of the wall on 2nd plate in the left section and on 5th plate in the right section. Similarly, vapors from the reboiler are split into two: vapor streams enter the left section at the 5th plate (i.e. FRAC-3) and on the 11th plate in the right section (i.e. FRAC-4). The temperature difference between the top and bottom of the DWC is 73 ∘ C, indicating a wide boiling mixture. As a result, a two-stage recompression strategy is used (Shrikhande et al., 2021). The overhead vapor generated from the DWC is divided into two streams for the implementation of vapor recompression. One stream goes to the condenser (CL4), whereas the remaining portion goes to the compressor (COMP-1). Then, the vapor is compressed using a compressor that increases the temperature of the vapor, and the heat gained due to compression is utilized in HX-5, which partly replaces the reboiler. After exchanging heat in HX-5, the vapor is sent to compressor COMP-2. Again, a vapor is compressed, and the heat gained due to compression is utilized in HX-6, which also partly replaces the reboiler. After exchanging the heat in HX-6, vapor returns to the condenser through the throttling valve. After the condensation of vapor, the liquid is collected in a reflux drum (DRUM); a part of it goes to DWC as reflux, and the remaining part (with 96% methanol) is taken out as distillate and recycled back to the reactor. Water is withdrawn from the 11th plate of the right section of the DWC. A mixture of GL and BBAIL catalyst is removed from the bottom of the DWC. The remaining process for algal biodiesel production, except for DWC-MVR, is similar to alternative 1 (explained in Section 2.1). Preliminary values of vapor and liquid splits are determined based on simple sensitivity analysis prior to the MOO.

IL F = 0 kg/h T = 25 °C XIL = 1

FRESH ALGAE F = 50322 kg/h T = 30 °C XTG = 0.05, XWAT = 0.816

M-1

Qc = –10.2 MW Qb = 7 MW

0.05 atm

SEP - 1 1

atm

PRO - CARB

BIODIESEL F = 2509 kg/hr T = 25 °C XFAME = 0.98

WAT - 1 F = 29.1 kg/h T = 30 °C XWAT = 1

RTRANS P-1

4

WASH - 1

M-2

45 ° C

7

WATER OUT

8

H-1 F = 8.5 kg/h T = 30 °C XHE = 1

FRAC - 1

M-4

M-3

SPLIT

Qc = - 0.35 MW Qb = 0.84 MW 0.05 atm

HX - 1 P-2

PS - 1 CL - 4

COMP - 2

1-4 6-11 DWC 5

R - ACET

1 atm

1 atm 80 ° C

SEP - 2 0.05 atm

1 atm FRAC - 5

7 12

13

DIST - 5 F = 1219.7 kg/h T = 100.4 °C XAA = 0.69 XWAT = 0.309

6

FRAC - 6 11

M-5

HX - 6 DIST - 6 F = 313.81 kg/h T = 99 °C XWAT = 0.5974 XAA = 0.4024

Catalyst Reuse

MA - REC F = 42070 kg/h T = 63 °C XMA = 0.9

WATER F = 41796 kg/h T = 30 °C XWAT = 0.98

HX - 4

11

HT - 6

M-6

P-4

TG - REC F = 1533 kg/h T = 25 °C XTG = 0.93

16

2-5

HX - 5

AA - REC F = 905.9 kg/h T = 102 °C XAA = 0.79 XWAT = 0.209

15

DRUM

COMP - 1

Qc = –17.4 MW Qb = 11.3 MW

AA F = 612 kg/h T = 30 °C AA XAA = 1

HX - 2

FRAC - 2 HX - 3

PS - 2

HEX - REC F = 88380 kg/h T = 57 °C XHE = 0.967

P-3

8

WAT - 2 F = 540.5 kg/h T = 30 °C XWAT = 1 WASH - 2

HT - 1

MA F = 1039 kg/h T = 25 °C XMA = 1

0.05 atm FLASH - 1

V-1

WAT – VAP - 1

0.05 atm FLASH - 2 V-2

12

HT - 2

TAG F = 572 kg/h T = 30 °C XTAG = 0.97 XDAG = 0.02

IL - REC F = 176.5 kg/h T = 100 °C XIL = 0.86

P-5 P-6

CL - 1

Figure 12.2 Biodiesel production from micralgae having DWC-MVR and conversion of glycerol to triacetin. Source: Reproduced with permission from Elsevier, Deshpande et al. (2022b).

342

12 Analysis of Safety and Economic Objectives for Intensified Algal Biodiesel Process

12.3 Multi-Objective Optimization Industrial chemical plants must have their design and operating conditions optimized to perform effectively. For optimization, it is required to first define suitable objective functions, which are quantitative measures of the system’s effectiveness. Each of these objectives could be any number or combined effect of (decision) variables representing the system. The goal of optimization is to find the values of variables that optimize the objective functions, which are frequently restricted or constrained in a certain way. The best results obtained through MOO of conflicting objectives are known as the Pareto-optimal fronts, which are made up of many non-dominated or equally good solutions in terms of the objectives. Pareto-optimal solutions are those in which any advancement in one objective leads to a deterioration in at least one other (Rangaiah et al., 2020). Evolutionary algorithms, such as NSGA-II, are common methods for solving MOO problems to obtain Pareto-optimal solutions. Due to its ability to quickly sort populations based on Pareto dominance and crowding distance, NSGA-II is superior to earlier genetic algorithms. More details about NSGA-II can be found in Deb et al. (2002). The NSGA-II has been widely used, as shown by its numerous applications in biotechnology, food technology, pharmaceutical manufacturing, refinery, and chemical process design and operation (Ahmed et al., 2022). Sharma and Rangaiah (2012) created an Excel-based MOO (EMOO) program that used binary-coded NSGA-II. Later, Wong et al. (2016) added the real-coded NSGA-II and two progress-based termination criteria. In this program, Excel worksheets are utilized for the program interface, input algorithm parameters, objective functions and constraints calculations, and result display. Visual basic for application (VBA) is used for the execution of NSGA-II and interfacing between the Aspen Plus simulator and Excel. This EMOO program was used in the present study. In this MOO study, the aim is to minimize the BEC and IR of the biodiesel production process. minimize (BEC, IR) = min (F(x), G(x))

(12.3a)

subject to: xL ≤ x ≤ xU

(12.3b)

g(x) ≤ 0

(12.3c)

where x is the decision variables vector, and xL and xU are the lower and upper bounds of x, respectively. g(x) is the set of inequality constraints. For more information about mathematical optimization approaches along with equality and inequality constraints, readers may refer to Section 3.4.1. In this study, BEC and IR are the objectives, and temperature and residence time of the transesterification reactor and feed tray of all distillation columns are considered as the decision variables. The residence time and temperature of the transesterification reactor affect the capital and operating costs as they affect the conversion/product formation and heat requirements. The position of the feed tray in the distillation columns influences the steam requirement in the reboiler. The goal of the biodiesel plant design is to maintain the purity of the product, byproduct,

12.3 Multi-Objective Optimization

343

and recycle streams; therefore, biodiesel purity (mass fraction > 0.965), TAG purity (mass fraction > 0.96), methanol purity in methanol recycle stream (mass fraction > 0.88), and hexane purity in hexane recycle stream (mass fraction > 0.95) are considered as the constraints in this study. The maximum temperature of all distillation columns is also considered a constraint to avoid the decomposition of products in the respective column. Also, the maximum temperature at the outlet of compressor 2 in process alternative 2 is considered a constraint owing to process safety. The ranges of the decision variables and limits of the constraints are given in Table 12.1.

Table 12.1 Decision variables, their bounds, and constraints for MOO of both process alternatives. Decision variables (x) Alternative 1

Alternative 2

Constraints

35 ≤ Temperature of RTRANS ≤ 60 o C 55 ≤ Residence time of RTRANS ≤ 85 min

35 ≤ Temperature of RTRANS ≤ 60 o C 55 ≤ Residence time of RTRANS ≤ 85 min 2 ≤ Feed tray of FRAC-1 ≤ 7 2 ≤ Feed tray of FRAC-2 ≤ 15 2 ≤ Feed tray of DWC ≤ 6 2 ≤ Feed tray of FRAC-5 ≤ 12 2 ≤ Feed tray of FRAC-6 ≤ 11 290 ≤ Vapor split in DWC ≤ 440 2000 ≤ Liquid split in DWC ≤ 2900 2.2 ≤ CR in compressor-2 ≤ 4.2 0.4 ≤ Overhead vapor split ≤ 0.8

Mass purity of biodiesel ≥ 0.965

4 ≤ Feed tray of FRAC-1 ≤ 6 6 ≤ Feed tray of FRAC-2 ≤ 9 3 ≤ Feed tray of FRAC-3 ≤ 6 5 ≤ Feed tray of FRAC-4 ≤ 9 4 ≤ Feed tray of FRAC-5 ≤ 8 4 ≤ Feed tray of FRAC-6 ≤ 8

Mass purity of TAG ≥ 0.96 Recovery of methanol-hexane from FRAC-4 ≥ 0.88 Recovery of hexane from FRAC-1 ≥ 0.95 Maximum temperature in FRAC-1 ≤ 250 o C Maximum temperature in FRAC-2 ≤ 290 o C Maximum temperature in FRAC-3 ≤ 150 o C Maximum temperature in FRAC-4 ≤ 150 o C Maximum temperature in FRAC-5 ≤ 150 o C Maximum temperature in FRAC-6 ≤ 150 o C Maximum temperature out of compressor-2 (T COMP2_OUT ) ≤ 300 o C (For alternative 2)

344

12 Analysis of Safety and Economic Objectives for Intensified Algal Biodiesel Process

Transfer of decision variables from MS Excel to Aspen Plus

Transfer of decision variables from VBA to MS Excel

MS Excel Biodiesel plant in aspen plus

VBA Interface

Final results

Last generation

MOO algorithm in VBA

Calculate objectives and constraints Transfer of required data from Aspen Plus to MS excel

Input/Run MOO parameters Decision variable, constraints, population size and maximum number of generation, etc.

Figure 12.3

Transfer of required data from MS excel to VBA

MOO framework for biodiesel production from wet microalgae.

Figure 12.3 depicts the MOO framework for optimizing the algal biodiesel production process. NSGA-II, implemented in VBA, generates the decision variables and provides these as input to the Aspen Plus simulation. After the completion of the simulation run in Aspen Plus, the data required for the evaluation of objectives and constraints are supplied to MS Excel from the simulator via the VBA interface. This continues until the maximum number of generations is reached. NSGA-II algorithm parameters for this study are maximum number of generations = 100, population size = 100, mutation probability = 0.1, cross-over probability = 0.85, and random seed = 0.5. Similar values for NSGA-II parameters were used by earlier optimization studies for biodiesel production (Deshpande et al., 2022a; Patle et al., 2014).

12.3.1 Objective Functions 12.3.1.1 Break-Even Cost

BEC is selected as an economic indicator (i.e. an objective) to understand the economics of the biodiesel process. The BEC of the process is determined by setting the biodiesel cost such that the cost of manufacturing (COM) (Eq. (12.5)), which incorporates raw material cost, capital cost, utility cost, and labor cost (Turton et al., 2008), equals the revenue thus generated by selling the products. In other words, it is the point at which the profit is zero. Equipment is first sized, and the cost is determined as per the procedure reported in our earlier article (Deshpande et al., 2022b). Length-to-diameter aspect ratio is considered as 2 : 1 for pressure vessels such as reactors, reflux drums, etc. The reactor volume is calculated based on the required

12.3 Multi-Objective Optimization

residence time in the reactor. The diameter of the distillation columns is calculated using Aspen Plus. The number of actual trays in each of all distillation columns is taken as the number of ideal trays (i.e. 100% efficiency, as the tray efficiency is 90% using the O’Connell correlation). Column height is calculated from the number of trays having 0.8 m tray spacing, and assuming the free space at both ends as 20% of the calculated height. The volume of phase separators is determined using the viscosity of the mixture and the density of the light and heavy phases. For a heat exchanger, the area is found based on heat duty obtained from the simulator, and its cost is estimated using the method described in Turton et al. (2008). The cost of the biodiesel process having DWC-MVR (i.e. alternative 2) is as reported in our earlier article (Deshpande et al., 2022b). In this study, the Chemical Engineering Plant Cost Index (CEPCI) of 607.5 is used to take inflation into consideration. The bare module cost (BMC) is determined using a bare module factor for every equipment. Subsequently, total module cost (TMC) is determined using: TMC = 1.18BMC

(12.4)

COM is found using the following equation in Turton et al. (2008). COM = 0.28TMC + 2.73 (labor cost) + 1.23(utility cost + raw materials cost) (12.5) For calculating utility cost, the heat duty of heat-exchanging equipment is fetched from Aspen Plus. Shaft power was also taken from Aspen Plus to calculate the electricity requirements of pumps and compressors. A year is assumed to have 8000 hours of operation. More details about the cost can be found in Deshpande et al. (2022b). 12.3.1.2 Individual Risk (IR)

As reported by De Haag and Ale (2005) and Contreras-Zarazúa et al. (2019), the IR indicates the probability of an individual’s casualty/injury due to loss of containment events. It is used to determine the extent of safety (Sánchez-Ramírez et al., 2019). IR represents the threat that personnel encounter based on their location, taking into account the occurrence of incidents and the possibility of death/injuries caused by an event (Contreras-Zarazúa et al., 2019). It is determined as: ∑ IR = fi Pai (12.6) where f i and Pai are the occurrence frequency of incident i and the probability of injury/death due to incident i, respectively (AIChE, 2000). Quantitative risk analysis can estimate the frequency and likelihood of an event. Its methodology enables the identification of possible hazards and the assessment of their consequences and harm. Possible accidents that can occur in chemical process industries and their frequencies of occurrence are shown in Figure 12.4 (De Haag and Ale, 2005). The accidents considered in the calculation for IR are divided into two types: instantaneous release such as catastrophic rupture in equipment, and continuous release

345

346

12 Analysis of Safety and Economic Objectives for Intensified Algal Biodiesel Process

such as rupture in a pipeline carrying liquid and/or vapor (Medina-Herrera et al., 2014; Contreras-Zarazúa et al., 2021). The total quantity of the content within the equipment is assumed to be the mass that escapes because of an instantaneous release. The content in the columns is calculated from the volume fraction of liquid and vapor within the column and their corresponding densities. The detailed procedure to calculate the amount of substance that escaped from the column is evaluated by following the procedure in Medina-Herrera et al. (2014). The quantity of material escaped as a result of continuous discharge from the operation can be calculated using the source model. Substance escaped is considered as an average leak of 2.5 cm in diameter by presuming a supersonic stream (AIChE, 2000). The detailed procedure to calculate the material release during the continuous release is given in Medina-Herrera et al. (2014). The Pasquille Gifford dispersion model, which translates the source term release to concentration in fields downwind from the source, is used in this study (Lee et al., 2019; Crowl and Louvar, 2002). This dispersion model assumes passive dispersion. The assumed conditions are stability class F (i.e. very stable) and a wind velocity of 1.5 m/s. The concentration of released material at ground level is determined from the dispersion model; these concentrations are used for the flash fire and toxic release estimation (AIChE, 2000). Figure 12.4 presents the event tree of potential accidents and corresponding frequencies (f i ) (AIChE, 2000). A hazard and operability (HAZOP) study can help to identify the likelihood of instantaneous/continuous occurrences in the process. Accidents such as unconfined vapor cloud explosion (UVCE), boiling liquid expanding vapor explosion (BLEVE), flash fire, and toxic release are classified under instantaneous release. On the other hand, accidents such as flash fires, jet fires, and toxic releases are classified as continuous releases. For the calculation of flash fire Instantaneous release Frequency (fi)

Immediate ignition P1 = 0.25 Delayed ignition P2 = 0.9

2.3 × 10–5/year

P3 = 0.5

BLEVE

5.76 × 10–6/year

UVCE

7.76 × 10–6/year

Flash fire

7.76 × 10–6/year

Toxic release

1.55 × 10–6/year

P3 = 0.5 No immediate ignition P1 = 0.75

No ignition P2 = 0.1

Continuous release Immediate ignition P1 = 0.25 3.67 × 10–5/year

Flash fire

2.48 × 10–4/year

Delayed ignition P2 = 0.9

3.67 × 10–5/year

No immediate ignition P1 = 0.75

Figure 12.4

Jet fire

No ignition P2 = 0.1

–5 Toxic release 8.26 × 10 /year

Event tree of potential accidents and corresponding frequencies.

12.4 Results and Discussion

and toxic release, the worst-case situation with an air velocity of 1.5 m/s is assumed. The concentration of explosive materials is calculated by the dispersion model. When the calculated concentration is within the lower and upper flammability limits, the thermal radiation generated by a flash fire is assumed to be adequate to pose a hazard (Medina-Herrera et al., 2014). Similarly, if the concentration calculated by the dispersion model is more than the lethal concentration LC50 , then the persons present at that location are assumed to face fatal injury/death in the case of toxic release. Once the possible accidents are recognized, probit function Y for individual events is calculated based on the thermal radiation and over-pressurization for BLEVE, UVCE, and jet fire. The procedure to calculate the probit function Y is given in Crowl and Louvar (2002). The constant used for the calculation of Y for thermal radiation and over-pressurization is taken from Medina-Herrera et al. (2014). Then, the probability of affection is calculated. IR is finally calculated using Eq. (12.6) based on the obtained probability of affection and the frequency of each event given in Figure 12.4. Detailed procedures, including the lower and upper flammability limits and LC50 values for all chemicals used in this study, are given in Deshpande et al. (2022b).

12.3.2 Simple Additive Weighting (SAW) Method MOO gives a variety of optimal solutions, i.e. Pareto-optimal front. A decision-maker faces the challenge of choosing one of these solutions for implementation. For this decision-making, there are many methods, of which Simple Additive Weighting (SAW) is one of the several methods recommended by Wang and Rangaiah (2017). Details of this method as well as an MS Excel program for SAW and nine other methods are available in Wang and Rangaiah (2017). The steps in the SAW method are (i) normalization of objective values of all Pareto-optimal solutions; (ii) multiplying the normalized objective values with the respective weight of each objective; (iii) finding the sum of weighted objectives of each of the Pareto-optimal solutions; and (iv) finding the solution having the largest sum of weighted objectives; this solution is the recommended optimal solution. Thus, SAW requires weights for objectives (i.e. the relative importance of each objective). In the present study, an equal weight of 0.5 is assumed for BEC and IR.

12.4 Results and Discussion In this study, the ultrasound-intensified algal biodiesel production process (i.e. alternative 1) and the same process further intensified by DWC-MVR (i.e. alternative 2) are optimized and discussed for economic and safety performance by considering BEC and IR as objectives. First, each of these processes is discussed for performance in terms of BEC and IR. Then, the first rank-optimized solution obtained by the SAW method for each of the two alternative processes is compared and analyzed for their merits.

347

348

12 Analysis of Safety and Economic Objectives for Intensified Algal Biodiesel Process

12.4.1 Minimization of BEC and IR for Alternative 1 The MOO results of process alternative 1 with BEC and IR as objectives are presented in this section. IR of the process is affected by the quantity of explosive/ flammable/toxic contents in the plant. It is also influenced by the internal flow within the process, like reflux, recycling, etc. Hence, the higher the inventory of hazardous/ toxic material and internal flow, the higher will be the IR of the process. Ideally, minimum BEC and minimum IR are desired for the process. The trade-off between BEC and IR is shown in Figure 12.5. Every point on the optimal front reflects an individual set of decision variables for the minimization of BEC and IR. From Figure 12.5a, it is clear that BEC and IR are conflicting in nature. The BEC of the biodiesel process decreases from $3.4 to $1.78 per kg at the expense of an increase in the IR from 0.8 × 10−5 to 1.64 × 10−5 per year. The BEC of the process increases with the decrease in temperature of the RTRANS (Figure 12.5b) because lipid conversion is reduced at lower temperatures, which has a negative effect on the BEC. Generally, the optimal reactor temperature is in the lower part of the plot, around 35 o C (Figure 12.5b). Less conversion of the lipid has a positive effect on the IR because unreacted lipid is less harmful than biodiesel. Both objectives, i.e. BEC and IR, also depend on the residence time of the reactor. The BEC of the process is negatively affected by a decrease in the residence time of the reactor as a result of lesser conversion. At the same time, it is positively affected as a result of the smaller volume of the reactor, which leads to lesser capital investment. Lower biodiesel conversion also increases the separation cost of the biodiesel from the other chemicals downstream of the process. Lower biodiesel production favors the IR of the process, as unreacted lipid is less toxic than biodiesel. Figure 12.5c depicts the optimal residence time of the transesterification reactor. The optimal residence time of the reactor is largely in the lower part of the plot. At lower residence time, less lipid conversion occurs, which results in lesser biodiesel production, which leads to increased BEC while decreasing the IR of the process. Also, with a lowerbiodiesel concentration, a smaller quantity of hexane is required for the separation in phase separators. Hexane has the highest heat of combustion among the chemicals involved in the process. However, the increase in the residence time of the transesterification reactor increases the material content of the reactor. Also, an increase in biodiesel conversion increases the toxicity and flammability of chemicals in the process, which has a negative effect on the IR. The optimal feed trays of all distillation columns are shown in Figure 12.5d–i. From these plots, it is clear that all optimal feed plates are toward the lower bound of the plot (i.e. near the reboiler of the column), except for FRAC-3, whose optimal feed tray is scattered between the upper and lower bounds. The optimal feed tray near the reboiler depicts the minimum heating duty in the reboiler, which affects the operating cost of the distillation column. It also results in an optimal reflux ratio, which in turn reduces the internal flow in the column and reduces the material content of the plant, which favors the IR of the process.

1.8 1.6 1.4 1.2 1.0 0.8 0.6

(T)RTRANS (°C)

IR (1/year) 10–5

12.4 Results and Discussion

2.8

60 55

8 7

(e)

3.2

6 3.0

3.2

(Feed stage)FRAC-6

5.0 3.0

3.2

3.4

BEC ($/kg)

6 5 4 3 2.8

3.0

3.2

3.4

BEC ($/kg)

7 6 5

3.4

2.8

3.0

(h)

BEC ($/kg)

3.2

3.4

BEC ($/kg)

8.0 7.5 7.0 2.8

(i)

5.5

2.8

(Feed stage)FRAC-5

7

3.4

6.0

(f)

8

3.2

BEC ($/kg)

3.4

BEC ($/kg)

2.8

3.0

(d)

9

3.0

2.8

3.4

(Feed stage)FRAC-3

(Feed stage)FRAC-2

3.2

BEC ($/kg)

2.8

(Feed stage)FRAC-4

3.0

40

(b)

65

2.8

50

30

3.4

BEC ($/kg)

(c)

(g)

3.2

(Feed stage)FRAC-1

(Residance time)RTRANS (min)

(a)

3.0

60

3.0

3.2

3.4

BEC ($/kg)

Figure 12.5 MOO of process alternative 1 with BEC and IR as objectives (a) IR vs. BEC, (b) temperature of RTRANS vs. BEC, (c) residence time of RTRANS vs. BEC, (d) feed stage of FRAC-1 vs. BEC, (e) feed stage of FRAC-2 vs. BEC, (f) feed stage of FRAC-3 vs. BEC, (g) feed stage of FRAC-4 vs. BEC, (h) feed stage of FRAC-5 vs. BEC, (i) feed stage of FRAC-6 vs. BEC.

349

350

12 Analysis of Safety and Economic Objectives for Intensified Algal Biodiesel Process

12.4.2 Minimization of BEC and IR for Alternative 2 In our earlier article (Deshpande et al., 2022b), simultaneous minimization of BEC and IR for the intensified biodiesel process (referred to as alternative 2 in this study) using multi-objective differential evolution coupled with dynamic local search was presented. Results from our earlier study are reproduced here to compare the economic and safety performance of the two alternatives considered in this study. Figure 12.6a depicts the obvious trade-off for BEC and IR of process alternative 2: one objective improves with the loss in another objective. It can be seen that the BEC of the process increases from $2.19 to $4.12 per kg for the reduction in the IR to 1.26 × 10−5 from 2.01 × 10−5 per year. The IR of the process increases with the increase in toxic, flammable, and explosive materials in the reactor, phase separators, and distillation columns. Higher temperature and residence time in the reactor increase biodiesel production and favor the BEC; however, they affect the IR negatively. The optimal reactor temperature is found to be distributed between the bounds (Figure 12.6b). This is because lower temperature decreases biodiesel production, which increases the BEC of the process, whereas it reduces the IR because biodiesel produced is more toxic and flammable than the algal lipid. Figure 12.6c shows that the optimal residence time of the reactor is in the upper part of the plot (i.e. 65–75 minutes). Higher residence time in the reactor increases the conversion of lipids to biodiesel, favoring the BEC of the process, whereas it increases the toxic, flammable, and explosive content that increases the IR of the process. Figure 12.6d–h presents the optimal location of the feed tray for all six distillation columns. Optimal feed plates are generally scattered in the upper portion (i.e. near the reboiler) of the plot for all distillation columns. Note that optimal feed tray location affects the dimensions and utility consumption of the distillation columns. The size of the column influences the fixed cost, and utility consumption affects the operation cost of the process, affecting the BEC of the process. The optimum feed tray of the distillation column minimizes the internal flow in the column, thereby potentially facilitating lower IR. From Figure 12.6g, it is observed that the optimal LIQ-SPLIT-F4 (i.e. the liquid split across the DWC having a range of 300 and 400 kmol/h) is near the original (i.e. base case) value at the middle of the plot. Similarly, from Figure 12.6h, it is clear that optimal VAP-SPLIT-F3 (which splits the vapor after reboiler across the DWC in the range of 2200–2600 kmol/h) is near the original value (i.e. base case) of 2460 kmol/h (i.e. middle of the plot). Note that an appropriate split of the vapor and liquid is essential for optimal operation of DWC. The compression ratio (CR) of compressor 2 (i.e. COMP2-CR) and overhead vapor split (i.e. OH-VAP-SPLIT) also need to be optimized for the optimization of both objectives. Both these are near the lower bound of the plot. As COMP2-CR increases, the capital cost and electrical consumption in the compressor increase, thereby increasing the BEC of the process. Optimal OH-VAP-SPLIT is toward the lower bound of the plot to minimize the BEC and IR.

12.4 Results and Discussion

(T)RTRANS (°C)

IR (1/year) 10–5

2.25 2.00 1.75 1.50 1.25 1.00

2

5

80 70 60 2

3

4

5

2

3 4 BEC ($/kg)

5

10 8 6 4 2

2

(g)

3 4 BEC ($/kg)

5

VAP SPLIT F3

LIQ SPLIT F4

2

3

4

5

BEC ($/kg) 7 6 5 4 3 2

2

3

4

5

BEC ($/kg) 6 5 4 3 2

2

3 4 BEC ($/kg)

5

2

3 4 BEC ($/kg)

5

2

3

5

12 10 8 6 4 2

350 300

2800 2600 2400 2200 2000

2

3

4

5

(j)

BEC ($/kg)

4

BEC ($/kg)

4.5 OH VAP SPLIT

COMP. RATIO (COMP2)

35

3000

400

(i)

(k)

40

(h)

450

250

45

(f) (Feed stage)FRAC–6

(e)

50

(d)

BEC ($/kg) 14 12 10 8 6 4 2

55

(b)

(Feed stage)FRAC–3

(Feed stage)FRAC–2

4

BEC ($/kg)

(c)

(Feed stage)FRAC–5

3

(Feed stage)FRAC–1

(Residence time)RTRANS (min)

(a)

60

4.0 3.5 3.0 2

3

4

BEC ($/kg)

5

(l)

0.8 0.6 0.4 1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

BEC ($/kg)

Figure 12.6 Simultaneous optimization of BEC and IR for alternative 2 (a) IR vs. BEC, (b) temperature of RTRANS vs. BEC, (c) residence time of RTRANS vs. BEC, (d) feed stage of FRAC-1 vs. BEC, (e) feed stage of FRAC-2 vs. BEC, (f) feed stage of FRAC-3 vs. BEC, (g) feed stage of FRAC-5 vs. BEC, (h) feed stage of FRAC-6 vs. BEC, (i) liquid split in FRAC-4 vs. BEC, (j) vapor split in FRAC-3 vs. BEC, (k) compression ratio in COMP2 vs. BEC, (l) overhead vapor split in DWC vs. BEC. Source: Reproduced with permission from Elsevier (Deshpande et al., 2022b).

351

352

12 Analysis of Safety and Economic Objectives for Intensified Algal Biodiesel Process

12.5 Comparative Analysis To investigate the performance of both process alternatives and to understand the significance of intensification, the best optimal solution obtained using the SAW method from the Pareto-optimal fronts (marked as a red Δ in Figures 12.5a and 12.6a) is compared against the base case (i.e. initial without optimization) of each alternative. BEC and IR for the base and optimized cases for each alternative are presented in Table 12.2. For alternative 1, the BEC decreased slightly from $3.016 to $2.98 per kg (by 1.1%) after optimization, whereas risk to the individual (i.e. IR) is improved from 1.48 × 10−5 to 8.9 × 10−6 per year (by 39.9%). For alternative 2, BEC is reduced from $3.07 to $2.74 per kg (by 10.7%) and IR is reduced from 1.82 × 10−5 to 1.44 × 10−5 per year (by 20.8%), upon optimization. For the base cases of both alternatives, it is observed that IR of the alternative 2 (i.e. 1.82 × 10−5 ) is greater than that of alternative 1 (i.e. 1.48 × 10−5 ) whereas there is a negligible difference in the BEC. On the other hand, for the optimized cases of both alternatives (i.e. solutions represented by red Δ in Figures 12.5a and 12.6a), the BEC of alternative 2 is reduced by about 8% (from $2.98 to $2.74 per kg), whereas the IR is increased by about 62% (from 8.9 × 10−6 to 1.44 × 10−5 ). This change in IR is attributed to greater material holdup in alternative 2 as a result of DWC-MVR on account of increased internal flow. In summary, alternative 2 is found to be economically beneficial, whereas alternative 1 is significantly better in terms of safety. Improvement in terms of BEC and IR of both biodiesel process alternatives after optimization is due to the optimized decision variables. Table 12.3 presents the decision variables for the base cases and optimal cases (for a solution obtained by the SAW method). It can be seen from this table that the reactor temperature after optimization is 36.5 and 38.19 ∘ C in alternative 1 and alternative 2, respectively, which are below 45 ∘ C in both the base cases. These indicate that the lower temperature of the RTRANS reactor in both cases is favorable. The optimal residence time is 55.15 and 75.4 minutes for alternative 1 and alternative 2, respectively, whereas the residence time for the base cases is 60 minutes. The residence time affects the conversion in the reactor, which affects both economics and safety. Optimal feed plates for all distillation columns in both cases are found to be generally closer to the reboiler. This results in lower steam consumption in this Table 12.2

Comparative analysis of both process alternatives.

Alternative 1 (present study)

Objective

Base case

Optimized case (% savings w.r.t. base case)

BEC ($/kg)

3.016

2.98 (1.1%)

IR (1/year)

−5

1.48 × 10

8.9 × 10

−6

(39.86%)

Base case

Optimized case (% savings w.r.t. base case)

% difference between the optimized solutions of both alternatives

3.07

2.74 (10.74%)

8.05%

Alternative 2 (Deshpande et al., 2022b)

−5

1.82 × 10

1.44 × 10

−5

(20.8%)

−61.8%

12.6 Conclusions

Table 12.3

Values of decision variables of base cases and optimized cases. Alternative 1

Alternative 2

Base case

Optimized case

Base case

Optimized case

Temperature of reactor (∘ C)

45

36.49

45

38.19

Residence time of reactor (min)

60

55.15

60

74.42

Feed tray of FRAC-1

4

6

4

5

Feed tray of FRAC-2

8

7

8

9

Feed tray of FRAC-3

5

5

—

—

Feed tray of FRAC-4

6

8

—

—

Feed tray of DWC (i.e. left section, FRAC-3)

—

—

5

6

Feed tray of FRAC-5

7

6

7

6

Feed tray of FRAC-6

6

8

6

8

VAP-SPLIT-F3a)

—

—

366

326.7

LIQ-SPLIT-F4a)

—

—

2460

2430

COMP2-CRa)

—

—

3.5

3.16

OH-VAP-SPLITa)

—

—

0.5

0.48

a) These are for alternative 2 only.

process and reduces the BEC as well as IR. Other optimal decision variables in alternative 2, such as overhead vapor split (OH-VAP-SPLIT), vapor split from FRAC-3 (VAP-SPLIT-F3), liquid split from FRAC-4 (LIQ-SPLIT-F4), and compression ratio in compressor 2 (COMP2-CR), are closer to the base case values (Table 12.3). Appropriate liquid flow and vapor flow on each side of the wall in alternative 2 are required for optimal operation of DWC. Similarly, the optimal amount of vapor flow through compressors is required to obtain sufficient heat by means of vapor recompression.

12.6 Conclusions In this study, ultrasound-assisted and ionic liquid-catalyzed in situ biodiesel production from wet microalgae is optimized using MS Excel-based NSGA-II algorithm, considering the BEC and IR as objectives. One of the Pareto-optimal solutions is chosen using the SAW method. For process alternative 1, the BEC of the optimized process is reduced from $3.016 to $2.98 per kg, and the IR of the process is improved from 1.48 × 10−5 to 8.9 × 10−6 per year, compared to the non-optimized base case. Although the saving in BEC is marginal (∼1%), the improvement in the IR is noteworthy (∼39%). Later, the results of the present study are compared against the results reported by Deshpande et al. (2022b) for the biodiesel process intensified with DWC-MVR (i.e. alternative 2). For non-optimized base cases, the

353

354

12 Analysis of Safety and Economic Objectives for Intensified Algal Biodiesel Process

IR of the alternative 2 (1.82 × 10−5 ) is greater than that of alternative 1 (1.48 × 10−5 ), whereas there is a negligible difference in the BEC. Comparison between BEC and IR of optimized cases for both alternatives shows that the BEC of alternative 2 is reduced by ∼8% (from $2.98 to $2.74 per kg), whereas the IR is increased by ∼62% (from 8.9 × 10−6 to 1.44 × 10−5 ). The change in IR is due to the greater material holdup in alternative 2 as a result of DWC and MVR with increased internal flows. Overall, it is observed that alternative 2 is economically favorable considering BEC, while alternative 1 is superior in terms of safety, as indicated by IR.

References Ahmed, M., Abdullah, A., Laskar, A. et al. (2022). Process simulation and stochastic multiobjective optimisation of homogeneously acid-catalysed microalgal in-situ biodiesel production considering economic and environmental criteria. Fuel 327: 125165. https://doi.org/10.1016/j.fuel.2022.125165. Ahmed, M., Abdullah, A., Patle, D.S. et al. (2021). Feedstocks, catalysts, process variables and techniques for biodiesel production by one-pot extraction-transesterification: a review. Environmental Chemistry Letters. Springer International Publishing 20: 335–378. https://doi.org/10.1007/s10311-021-01358-w. AIChE (2000). Guidelines for Chemical Process Quantitative Risk Analysis, 2e. AIChE. Athar, M. and Zaidi, S. (2020). A review of the feedstocks, catalysts, and intensification techniques for sustainable biodiesel production. Journal of Environmental Chemical Engineering 8: 104523. https://doi.org/10.1016/j.jece.2020.104523. Chisti, Y. (2008). Biodiesel from microalgae beats bioethanol. Biotechnology Advances 25: 294–306. https://doi.org/10.1016/j.tibtech.2007.12.002. Contreras-Zarazúa, G., Jasso-Villegas, M.E., Ramírez-Márquez, C. et al. (2021). Design and intensification of distillation processes for furfural and co-products purification considering economic, environmental, safety and control issues. Chemical Engineering and Processing Process Intensification 159: 108218. https://doi.org/10 .1016/j.cep.2020.108218. Contreras-Zarazúa, G., Sánchez-Ramírez, E., Vázquez-Castillo, J.A. et al. (2019). Inherently safer design and optimization of intensified separation processes for furfural production. Industrial and Engineering Chemistry Research 58: 6105–6120. https://doi.org/10.1021/acs.iecr.8b03646. Contreras-Vargas, C.A., Gómez-Castro, F.I., Sánchez-Ramírez, E. et al. (2019). Alternatives for the purification of the blend butanol/ethanol from an acetone/butanol/ethanol fermentation effluent. Chemical Engineering and Technology 42: 1088–1100. https://doi.org/10.1002/ceat.201800641. Crowl, D.A. and Louvar, J.F. (2002). Chemical Process Safety Fundamental with Application, 2e. Prentice Hall International Series. de Jesus, S.S., Ferreira, G.F., Moreira, L.S., and Filho, R.M. (2020). Biodiesel production from microalgae by direct transesterification using green solvents. Renewable Energy 160: 1283–1294. https://doi.org/10.1016/j.renene.2020.07.056.

References

Deb, K., Pratap, A., Agarwal, S., and Meyarivan, T. (2002). A fast and elitist multiobjective genetic algorithm. IEEE Transactions on Evolutionary Computation 6: 182–197. Deshpande, G., Shrikhande, S., Sawarkar, A.N., and Patle, D.S. (2022a). Multiobjective optimization of ultrasound intensified and ionic liquid catalyzed in situ algal biodiesel production considering economic, environmental and safety indicators. Chemical Engineering Research and Design 180: 134–152. https://doi.org/10.1016/j .cherd.2022.02.011. Deshpande, G., Shrikhande, S., Patle, D.S., and Sawarkar, A.N. (2022b). Simultaneous optimization of economic, environmental and safety criteria for algal biodiesel process retrofitted using dividing wall column and multistage vapor recompression. Process Safety and Environment Protection 164: 1–14. https://doi.org/10.1016/j.psep .2022.05.059. Fouladvand, M.T., Asadi, J., and Lotfollahi, M.N. (2021). Simulation and optimization of aromatic extraction from lube oil cuts by liquid-liquid extraction. Chemical Engineering Research and Design 165: 118–128. https://doi.org/10.1016/j.cherd.2020 .10.026. Ganesan, R., Manigandan, S., Shanmugam, S. et al. (2021). A detailed scrutinize on panorama of catalysts in biodiesel synthesis. Science of the Total Environment 777: 145683. https://doi.org/10.1016/j.scitotenv.2021.145683. Gautam, A., Pandey, A., Balinge, K.R., et al. (2021). Optimization, kinetics and thermodynamics of ultrasound assisted and ionic liquid catalyzed in situ biodiesel synthesis from wet microalgae. Proceedings of the 14th AUN/SEED-Net Regional Conference on Environmental Engineering 2021 (RCEnvE2021) in conjunction with the 7th International Conference on Environment 2021 (ICENV2021-RCEnvE2021), USM Malaysia (6–7 October 2021), AIP Conference Proceedings. Goh, B.H.H., Ong, H.C., Cheah, M.Y. et al. (2019). Sustainability of direct biodiesel synthesis from microalgae biomass: a critical review. Renewable and Sustainable Energy Reviews 107: 59–74. https://doi.org/10.1016/j.rser.2019.02.012. De Haag, U. and Ale, B.J. (2005). Guidelines for Quantitative Risk Assessment, 2e. Delft, Netherlands: Netherlands Organ https://content.publicatiereeksgevaarlijkestoffen .nl/documents/PGS3/PGS3-1999-v0.1-quantitative-risk-assessment.pdf (accessed January 2023). Khan, F.I. and Amyotte, P.R. (2005). I2SI: a comprehensive quantitative tool for inherent safety and cost evaluation. Journal of Loss Prevention in the Process Industries 18: 310–326. https://doi.org/10.1016/j.jlp.2005.06.022. Khan, F.I. and Amyotte, P.R. (2004). Integrated inherent safety index (I2SI): a tool for inherent safety evaluation. Process Safety Progress 23: 136–148. https://doi.org/10 .1002/prs.10015. Kim, B., Heo, H.Y., Son, J. et al. (2019). Simplifying biodiesel production from microalgae via wet in situ transesterification: a review in current research and future prospects. Algal Research 41: 101557. https://doi.org/10.1016/j.algal.2019.101557. Lee, Y., Kim, J., Ahmed, U. et al. (2019). Multi-objective optimization of Organic Rankine Cycle (ORC) design considering exergy efficiency and inherent safety for

355

356

12 Analysis of Safety and Economic Objectives for Intensified Algal Biodiesel Process

LNG cold energy utilization. Journal of Loss Prevention in the Process Industries 58: 90–101. https://doi.org/10.1016/j.jlp.2019.01.006. Makareviciene, V. and Skorupskaite, V. (2019). Transesterification of microalgae for biodiesel production. In: Second and Third Generation of Feedstocks: The Evolution of Biofuels (ed. A. Basile and F. Dalena), 469–510. Elsevier https://doi.org/10.1016/ B978-0-12-815162-4.00017-3. Medina-Herrera, N., Jiménez-Gutiérrez, A., and Mannan, M.S. (2014). Development of inherently safer distillation systems. Journal of Loss Prevention in the Process Industries 29: 225–239. https://doi.org/10.1016/j.jlp.2014.03.004. Mondal, B. and Jana, A.K. (2019). Techno-economic feasibility of reactive distillation for biodiesel production from algal oil: comparing with a conventional multiunit system. Industrial and Engineering Chemistry Research 58: 12028–12040. Patle, D.S., Sharma, S., Ahmad, Z., and Rangaiah, G.P. (2014). Multi-objective optimization of two alkali catalyzed processes for biodiesel from waste cooking oil. Energy Conversion and Management 85: 361–372. https://doi.org/10.1016/j .enconman.2014.05.034. Patle, D.S., Pandey, A., Srivastava, S. et al. (2020a). Ultrasound intensified biodiesel production from algal biomass: a review of current advancements, issues, and future directions. Environmental Chemistry Letters 19: 209–229. Patle, D.S., Shrikhande, S., and Rangaiah, G.P. (2020b). Process development, design and analysis of microalgal biodiesel production aided by microwave and ultrasonication. In: Process Systems Engineering for Biofuels Development (ed. A. Bonilla-Petriciolet and G.P. Rangaiah), 259–284. Wiley https://doi.org/10 .1002/9781119582694.ch10. Piemonte, V., Di Paola, L., Iaquaniello, G., and Prisciandaro, M. (2016). Biodiesel production from microalgae: Ionic liquid process simulation. Journal of Cleaner Production 111: 62–68. https://doi.org/10.1016/j.jclepro.2015.07.089. Rangaiah, G.P., Feng, Z., and Hoadley, A.F. (2020). Multi-objective optimization applications in chemical process engineering: tutorial and review. Processes 8 (5): 508. Sánchez-Ramírez, E., Quiroz-Ramírez, J.J., and Segovia-Hernandez, J.G. (2019). Synthesis, design and optimization of alternatives to purify 2,3-butanediol considering economic, environmental and safety issues. Sustainable Production and Consumption 17: 157–162. https://doi.org/10.1016/B978-0-12-818634-3.50027-8. Sánchez-Ramírez, E., Quiroz-Ramírez, J.J., and Segovia-Hernández, J.G. (2020). Intensified purification alternative for methyl ethyl ketone production: economic, environmental, safety and control issues. In: Process Systems Engineering for Biofuels Development (ed. Bonilla-Petriciolet, A. & Rangaiah, G.P.), Wiley, 311–339. Sawarkar, A.N. (2019a). Cavitation induced upgrading of heavy oil and bottom-of-the barrel: a review. Ultrasonics Sonochemistry 58: 104690. https://doi.org/10.1016/j .ultsonch.2019.104690. Sawarkar, A.N. (2019b). Upgrading of Mumbai High vacuum residue. Petroleum Science and Technology 37: 1090–1098. https://doi.org/10.1080/10916466.2019.1575875. Sharma, S., Chao Lim, Z., and Rangaiah, G.P. (2013). Process design for economic, environmental and safety objectives with an application to the cumene process. In: Multi-Objective Optimization in Chemical Engineering: Developments and

References

Applications (ed. Rangaiah, G.P. & Bonilla-Petriciolet, A.), Wiley, 449–477. https://doi .org/10.1002/9781118341704.ch16. Sharma, S. and Rangaiah, G.P. (2012). Modeling and optimization of a fermentation process integrated with cell recycling and pervaporation for multiple objectives. Industrial and Engineering Chemistry Research 51: 5542–5551. https://doi.org/10 .1021/ie202205h. Shrikhande, S., Deshpande, G., Sawarkar, A.N. et al. (2021). Design and retrofitting of ultrasound intensified and ionic liquid catalyzed in situ algal biodiesel production. Chemical Engineering Research and Design 171: 168–185. https://doi.org/10.1016/j .cherd.2021.05.010. Turton, R., Bailie, R.C., Whiting, W.B., and Shaeiwitz, J.A. (2008). Analysis, Design and Synthesis of Chemical Processes, 3e. Upper Saddle River, N.J., USA: Prentice Hall. Wang, Z. and Rangaiah, G.P. (2017). Application and analysis of methods for selecting an optimal solution from the pareto-optimal front obtained by multiobjective optimization. Industrial and Engineering Chemistry Research 56: 560–574. https://doi .org/10.1021/acs.iecr.6b03453. Wong, J.Y.Q., Sharma, S., and Rangaiah, G.P. (2016). Design of shell-and-tube heat exchangers for multiple objectives using elitist non-dominated sorting genetic algorithm with termination criteria. Applied Thermal Engineering 93: 888–899.

357

359

Index a acidic ion-exchange (IEX) resins 185 Aspen Custom Modeler (ACM) (www.aspentech.com) 90 air preheater (APH) 26 alcohol-to-jet (ATJ) process 313 American Institute of Chemical Engineers (AIChE) 134 Aspen Dynamics 84, 90, 270 Aspen HYSYS 52 Aspen Plus 52 Aspen Plus Dynamics 161 attenuation ratio 97 ATV (Auto-Tune Variation) method 202

b bare module cost (BMC) 345 basic process control system (BPCS) 270, 271 control basis 276–284 safety analysis 284–292 batch and semi-batch processes, design of 88 batch distillation column 209 benzimidazolium based Brønsted acid ionic liquid (BBAIL) catalyst 337 biobutanol 299 biodiesel plant design, goal of 342–343 biofuels, benefits of 335 bioprocesses 295 bioproducts purification control and automation 298 control behavior analysis 306

equipment design and scalability 298 green metrics 299 inherent safety 298 intensification of alcohol to jet fuel process 313–314 intrinsic safety 298 lactic acid 324–325 methodology 302–307 methyl ethyl ketone 307–313 new processes for furfural and co-products 318–324 process controllability 301 process integration 297 process intensification 299 process robustness 297 process safety 297 validation and regulatory compliance 298 boiling liquid expanding vapor explosion (BLEVE) 241 break-even cost (BEC) base and optimized cases 352–353 minimization of 348–351

c carbon capture 30–31 Center for Chemical Process Safety (CCPS) 134 checklist analysis 129–130 CHEMCAD 52 Chemical Engineering Plant Cost Index (CEPCI) 345

Control and Safety Analysis of Intensified Chemical Processes, First Edition. Edited by Dipesh Shikchand Patle and Gade Pandu Rangaiah. © 2024 WILEY-VCH GmbH. Published 2024 by WILEY-VCH GmbH.

360

Index

chemical engineers, survey and responses 257, 259–261 chemical safety index 239 circular chemistry 5 climate change 4 closed-loop dynamic simulations, for distillation columns 284 conventional middle vessel batch distillation (CMVBD) vs. controlled CMVBD & single-stage vapor recompression in middle-vessel batch distillation (SiVRMVBD) 230–233 CMVBD vs. SiVRMVBD 225–226 COCO Simulator 53 commercial process simulator 52–23 compact heat exchangers (CHEs) 27 compact membrane deaerators/contactors 33 component object model (COM) technologies 62 composition control 210 conceptual design 296 constant composition control 226–229 continuous improvement 135 continuous process improvement 296 control behavior analysis 306–307 control degrees of freedom (CDOF) 101–102 conventional middle vessel batch distillation (CMVBD) constant composition control 216 dynamic composition profiles 219–222 systematic simulation approach of 212–216 conventional sequence of columns (CSC) 238, 245–251 critical alarms and human intervention (CA & HI) 270 CSC systems, comparative analysis 257 CycloneCCTM CO2 capture solution 31

d damage index (DI) 240 calculation for

CSC system 251 DWC-MVR system 255 DWC system 253 decision variables, of base and optimized cases 353 dimethyl sulfoxide (DMSO) 158 distillation systems 21–24, 237, 241 conventional sequence of columns 241 dividing-wall column 241–242 dividing-wall column with mechanical vapor recompression 243–244 safety analysis 237–264 dividing-wall column (DWC) 84, 237–238, 251–253 dividing-wall column with mechanical vapor recompression (DWC-MVR) 253–255 dividing-wall column with MVR (DWC-MVR) 238 Dow Fire and Explosion index 240 DWC-MVR intensification effect 337 DWC-MVR systems, comparative analysis 257 DWC systems, comparative analysis 257 DWSIM 53, 90 dynamic matrix control (DMC) 188 dynamic process simulation 270 dynamic safety analysis, of process with independent protection layers (IPLs) 289–292 dynamic simulation of chemical processes applications 87–88 case study analysis of dynamic simulation results 112–120 control structure design 107–112 preparation/initialization 103–107 steady state simulation and optimization 103 tuning of controller parameters 112 dynamic simulation and control procedure 91–98 environmental impact assessment 83 intensified chemical processes 98–100 process control 100–102

Index

process control and optimization 83 process design and scale-up 83 process safety and risk assessment 83 software 88–91 understanding 84–87 DYNSIM Dynamic Simulation 90

e economic evaluation criterion 71–75 effective dangerous property (EDP) 240 environmental performance indicators 299 equation oriented method 93 Excel-based MOO (EMOO) program 342 extractive distillation (ED) 270 process intensification analysis 274–276 process intensification measures 272–273 steady-state process design 273–274

f failure modes and effects analysis (FMEA) 127, 129 feedforward control structure 172–177 fire and explosion damage radius (FEDR) 248–250 flue gas heat recovery 26 fossil fuels 296 free process simulators 53 furfural 318–324

g gain scheduling proportional integral controller (GSPI) 216 generalized modular representation framework 150 generalized predictive control (GPC) 188 glycerol LC50 247 lower flammability limit 247 upper flammability limit 247 gPROMS 53, 90 gravity type L-L-G separator 21

greenhouse gas (CO2 ) emissions 4, 209, 219, 298 green metrics 299 Guthrie’s method 314

h hazard and operability (HAZOP) study 127, 129, 304, 346 hazard identification 128–130 heat exchangers 26–27 heat-integrated distillation column (HIDiC) 210 heat integration schemes 210 heat transfer enhancement (HTE) techniques 27 heterogeneous catalysis 184 RD processes 185 hexane LC50 247 lower flammability limit 247 upper flammability limit 247 hexoses 295 high-gravity (HiGee) technologies 31, 142 homogeneous catalysis 184 human factors and safety culture 132–134 hybrid reactive-extractive distillation (RED) control performance evaluation 177–178 dynamic simulation setup 161 feedforward control structure 3 172–177 inventory control setup 162 quality control structures simple temperature control 165–168 triple point temperature control 168–170 triple point temperature control using SVD analysis 170–172 sensitivity analysis 163–165 steady-state design of 160–161 4-hydroxybutyl acrylate (HBA) 184, 192

361

362

Index

i individual risk (IR) 303 base and optimized cases 352 event table for CSC system 249 for DWC-MVR 254 for DWC system 253 minimization of 348–351 Industry 4.0 5 inherently safer design (ISD) 33–35, 131–132 inherent safety 298, 299 inherent safety index (ISI) 239, 240 in situ transesterification 335 integral of absolute error (IAE) 177 integral of error criteria 98 integrated inherent safety index (I2SI) 240 intensification of alcohol to jet fuel process 313 intensified algal biodiesel process, development 337–340 intensified chemical processes absence of intensified unit in the simulator 65 availability of basic data for simulation 65 basic process flow design 57 case study economic evaluation criterion 71–73 model construction 69–70 problem analysis 66–67 process flow design 67–69 process optimization 73–75 process simulation 70–71 results and analysis 75–78 commercial process simulator 52–53 computational methods for process simulation 53–56 convergence difficulties 65–66 dynamic simulation of chemical processes 98–100 free process simulators 53

mathematical optimization methods 59–62 MATLAB 62–63 model construction 57–59 need for experimental validation 66 problem analysis 56–57 process intensification and integration 57 Python 63–65 result analysis 59 simulation and convergence 59 usefulness of process simulation 50–52 intensified distillation systems 237 International Organization for Standardization (ISO) 134 intrinsic safety 298 i-safe method 240

k Kyoto Protocol

210

l lactic acid 324–329 design and synthesis of intensified processes 326 optimization 326–327 production by reactive distillation 325–326 results and discussion 327–329 layer of protection analysis (LOPA) 127, 270, 271 liquid jet ejector (LJE) 29

m management of change (MOC) 135 optimization methods 59–62 MATLAB 62–63 methanol LC50 247 lower flammability limit 247 upper flammability limit 247 methyl acetate hydrolysis 183 methyl ethyl ketone (MEK) 307–313

Index

production from non-renewable sources 308 through process intensified schemes 308–313 microalgae based biodiesel production 335 middle vessel batch distillation energy savings 218 greenhouse gas (CO2 ) emissions 219 total annual cost 218–219 Modelica 90 model predictive controllers (MPC) 188 moderation 143 Mond Index 240 multi-component batch distillation 209 multi-objective optimization (MOO) 336, 342 decision variables, bounds and constraints 343 framework for algal biodiesel production process 344 objective functions break even cost 344–345 individual risk 345–347 simple additive weighting method 347

n non-dominated sorting genetic algorithm (NSGA-II) 337, 342 non-random two liquid (NRTL) model 337 non-safety instrumented solutions 37–39 NSGA-II see non-dominated sorting genetic algorithm (NSGA-II) numerical descriptive inherent safety technique (NuDIST) 240

o open-loop sensitivity analysis 163 operator training 88 optimization 59–65, 336 overshoot 98

p Pareto-optimal front 342 Pareto-optimal solutions 60, 342 partial differential equations (PDEs) 86 Pasquille Gifford dispersion model 346 pentoses 295 potential of danger (POD) 240 pressure-driven simulations 161 pressure relief systems non-safety instrumented solutions 37–39 safety instrumented system (SIS) solutions 39–40 pressure safety valves (PSVs) 270, 271 probability of affectation 248 probit model parameters 248, 305 process and hazard control index (PHCI) 240 process control 87–88, 100–102 process controllability 301 process design 87 process dynamic simulation 272 process hazard analysis (PHA) 129 process intensification (PI) 125, 299, 303 approaches 4 benefits of 16–17 carbon capture 30–31 challenges of 43–44 circular chemistry 5 climate change 4 definition 3 distillation 21–24 economical and sustainable technologies 3 goals of 4 greenhouse emissions 4 heating compact heat exchangers (CHEs) 27 flue gas heat recovery 26 heat exchangers 26–27 sonic horn 27 steam/electric heaters 25 steam injection heater 24–25 impacts of 136–137

363

364

Index

process intensification (PI) (contd.) inherently safer design (ISD) 33–35 pressure relief systems non-safety instrumented solutions 37–39 safety instrumented system (SIS) solutions 39–40 principles 141–144 safety analysis 137–138 need for control 5–7 studies on control 7–9 separation vessels 18–21 static mixer 17–18 steam compression 27–30 sustainable development of chemical process plants 15 vacuum systems 31–32 wastewater recovery 41–43 water deaeration 33 process intensification 4.0 strategy 5 process route index (PRI) 240 calculation for CSC system 250, 251 for DWC system 254 improved 262–263 process safety 336 process safety analysis (PSA) 87, 126, 128 process safety management (PSM) standard 134 process stream index (PSI) 240 PRO/II 52 proportional-integral-derivative (PID) controller 94 Python 63–65

q quadratic dynamic matrix control (QDMC) 188 quantitative risk analysis (QRA) 303 quaternary DWC process 337

r reactive distillation (RD) coupled with a distillation–reactor system and recycle

basis of design and basic data 192–197 discussion 204 process control 201–204 process design 197–200 reactive distillation (RD) 313 in recycle systems case study 192–204 control of 188–192 design of 184–188 reflux flow rate 210 regulatory framework and compliance 134–135 retrofitting/revamping of existing plants 87 robust multivariable predictive control technology (RMPCT) 188

s safety, health and environment (SHE) index 240 safety analysis in chemical process industry case studies 149–151 tools 128–135 safety analysis tools hazard identification 128–130 human factors and safety culture 132–134 inherently safer design 131–132 monitoring and continuous improvement 135 regulatory framework and compliance 134–135 risk assessment 130–131 safety instrumented systems 132 safety indices for process safety assessment 239–241 selection of 244 safety instrumented functions (SIF) 132 safety instrumented system (SIS) 39, 132, 270, 271 safety integrity level (SIL) 132 safety level assessment, of a chemical process 269 safety management systems (SMS) 144–146

Index

safety performance indicators 135 safety training and competency development for intensified processes 146–149 scenario-based safety analysis 271 sensitivity analysis 163–165 separation vessels 18–21 Seveso Directive 134 shell and tube heat exchanger (STHE) 24 silane production schemes 299 simple additive weighting (SAW) method 347 simple temperature control 165–168 simplification 143 single objective approach (SOA) 61 single objective optimization (SOO) 60 single stage vapor recompression in middle vessel batch distillation (SiVRMVBD) 216–218, 222 singular value decomposition (SVD) 306–307 SiVRMVBD-GSPI 229 sonic horn 27 start-up and shutdown of a process 88 state variables 85 static mixer (SM) 17–18 steady state deviation (offset) 98 steam compression 27–30 steam/electric heaters 25–26 steam injection heater 24–25 steam jet ejector (SJE) 29 substitution 142 Sulzer KatapakTM -SP11 199 sustainability metrics 299 sustainable development 295 sustainable process design 296

t techno-economic feasibility, of biodiesel production 337 tetrahydrofuran (THF) 157 total module cost (TMC) 345 toxic damage radius (TDR) 249–250 toxic release and consequence analysis tool (TORCAT) 240 triacetin 338 triple point temperature control 168–170 triple point temperature control using SVD analysis 170–172 tuning of a controller 94 Tyreus-Luyben tuning rules 278

u ultrasound intensified algal biodiesel production process BEC and IR minimization 348–351 comparative analysis 352–353 unconfined vapor cloud explosion (UVCE) 241

v vacuum systems 31–32 vapor recompression column (VRC) method 210 visual basic for application (VBA) 342

w waste reduction 297 wastewater recovery 41–43 water deaeration 33 Wegstein’s method 55 what-if analysis 129

365