Thermal Spraying for Power Generation Components 3527313370, 9783527313372

Citation preview

I

K. E. Schneider, V. Belashchenko, M. Dratwinski, S. Siegmann, A. Zagorski Thermal Spraying for Power Generation Components

II

III

Klaus Erich Schneider, Vladimir Belashchenko, Marian Dratwinski, Stephan Siegmann, Alexander Zagorski

Thermal Spraying for Power Generation Components

IV

Authors Klaus Erich Schneider Kuessaberg, Germany e-mail: [email protected] Vladimir Belashchenko Concord, NH, USA e-mail: [email protected]

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. Library of Congress Card No.: applied for

Marian Dratwinski Stein, Switzerland e-mail: [email protected] Stephan Siegmann EMPA Thun, Switzerland e-mail: [email protected] Alexander Zagorski ALSTOM Baden, Switzerland e-mail: [email protected]

Cover Simulated Spray Pattern, ALSTOM

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 in the Internet at http://dnb.d-nb.de. © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 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. Composition Manuela Treindl, Laaber Printing Strauss GmbH, Mörlenbach Binding Litges & Dopf Buchbinderei GmbH, Heppenheim Printed in the Federal Republic of Germany Printed on acid-free paper ISBN-13: 978-3-527-31337-2 ISBN-10: 3-527-31337-0

V

Preface Coatings constitute an intrinsic part of the power-generation hardware. Thousands of patents, papers and conference presentations address new coating types, new hardware and software, new process developments, new chemical compositions. A huge unpublished knowledge is stored in manufacturers “know-how”. However, sometimes coatings are still considered as an “art” and there are fair reasons for that. The thermal spray is still not a “plug and play” tool and the product quality largely depends on the deep understanding of process physics and hardware features, accumulated experience, engineer’s intuition and operator’s training. This book now deals with questions that are essential for a good performance of this “art”: x Is there a given process stability? What is the ratio of deterministic and stochastic in the coating process? x Is there an inherent process capability for a given specification that cannot be improved? x What is the right preventive maintenance strategy? x Is there a chance to end up with coating-process capabilities in the order of other manufacturing processes? x What can be predicted and designed a priori by physical modeling and offline programming and what can be achieved by trial and error only? x What can be done to describe and control quality? This book is not a pure scientific book. It is of most value for the engineer involved in design, processing and application of thermally sprayed coatings: To understand the capability and limitations of thermal spraying, to understand deposition efficiency – and the importance of maintenance and spare parts for quick changeover of worn equipment, to use offline programming and real equipment in an optimum mix to end up with stable processes in production after the shortest development time and in the end to achieve the final target in production: Process stability at minimum total cost Klaus Schneider

VI

Preface

Acknowledgement

The authors would like to thank the following companies and institutions for supplying valuable material – published and unpublished – for this book. ALSTOM, Sulzer Metco, Turbocoating, CENIT, EMPA Material Science and Technology, Praxair, HC Starck, Siemens, ASM, Elsevier Publ., Stellite Coatings, Progressive Technologies, National Research Council Canada. And personal acknowledgements to F. Stadelmaier (TACR), P. Ryan, P. Holmes, J. E. Bertilsson (ALSTOM), A. Scrivani (Turbocoating), A. Sickinger (ASA, California, USA), K. Matty (former AETC). In particular, I would like to express my gratitude to the management and my colleagues at ALSTOM for the assistance and valuable discussions during all the years that enabled me to start this book. The production experience with offline programming and monitoring was only possible together with the erection and start-up of the ALSTOM coating shop in Birr, Switzerland.

VII

The Authors of this Book Klaus E. Schneider received a degree in Physics and Materials Science and a PhD in Materials Science and Technology from the University of Erlangen, Germany. He has three decades of experience in manufacturing and materials technology in power and turbine engineering, mechanical engineering. During his professional carreer at BBC, ABB, ALSTOM Mannheim, Germany, and Baden, Switzerland (1974–2004), he worked in several leading positions in materials, supply management and manufacturing. He was responsible for national and international R&D programmes and for erecting new manufacturing facilities. Since 2004 he is active as a consultant for materials and manufacturing technology.

Vladimir Belashchenko has a PhD in Physics and Chemistry of Plasma Technology and a ScD in Materials Science. He has over 30 years of experience in research, development and implementation of thermal spray equipment, materials and technologies. In 1992, he obtained the ASM International Award, in 2004 the R&D 100 Award.

Marian Dratwinski is a process development engineer with a very wide range of technical knowledge and experiences. In his current post, he is responsible for Coating Applications Development at Sulzer Metco AG in Switzerland.

VIII

The Authors of this Book

Stephan Siegmann received his degree in Physics from the University of Basel, Switzerland. After completing the PhD in the field of Thermal Spraying, he was working as Vice Manager Research at MGC Plasma Company at Muttenz, Switzerland, in the field of waste treatment by thermal plasma at 1.2 MW. In the year 1994 he changed back to his former field of Thermal Spraying and built up a position at the Swiss Federal Institute for Materials Science and Technology (EMPA), where he is now responsible for all Thermal Spray activities.

Alexander Zagorski received his degree in Mechanical Engineering from the Novosibirsk State Technical University and his PhD in Hydromechanics and Plasma from the Institute of Theoretical and Applied Mechanics in Novosibirsk, Russia. After having worked in Research and Development for eighteen years, he is now the Expert Engineer at the ALSTOM Customer Service Development in Switzerland.

IX

Contents Preface V The Authors of this Book VII

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8

1 Requirements for Materials and Coatings in Powerplants 1 Examples of Coatings in Gas Turbines 2 Definition of Thermal Spraying (THSP) 5 Thermal-Spraying Systems 5 Coatings for Power-Generation Components 6 The Complete Manufacturing and Coating Process 7 Coating-Process Development 12 Tasks for “Target” Readers 15

2

Practical Experience Today

1

2.1 2.2 2.3 2.3.1 2.3.2 2.3.2.1 2.3.2.2 2.3.2.3 2.3.2.4 2.4 2.4.1 2.4.2 2.4.3 2.4.3.1 2.4.3.2 2.4.3.3 2.4.4 2.4.4.1

Introduction

17 Coating Processes 17 Basics of Thermal Spraying 21 Feedstock 23 Wire 23 Powder 24 Powder Types 24 Powder-Production Processes and Morphologies 27 Powder Characterization 33 Powders for Power-Generation Applications 36 Thermal-Spraying Equipment 40 Example of a Low-Pressure Plasma-Coating System 41 Flame and Arc Spray Torches 43 HVOF Process 45 Comparison of HVOF Fuels 47 A Brief Overview of the Major Existing HVOF Systems 48 Possible Improvements of HVOF Systems 51 Plasma Process 54 A Brief Overview of Plasma Torches 58

X

Contents

2.4.4.2 2.5 2.5.1 2.5.1.1 2.5.1.2 2.5.1.3 2.5.1.4 2.5.1.5 2.5.2 2.5.3 2.5.4 2.5.5 2.5.6 2.6 2.7 2.7.1 2.7.1.1 2.7.2 2.7.2.1 2.7.3 2.7.4 2.7.4.1 2.7.4.2 2.7.4.3 2.8 2.8.1 2.8.2 2.8.3 2.8.4

Possible Improvements of Plasma Systems 63 Work Flow and Important Coating Hardware 65 Powder Preparation and Powder-Delivery System 68 Powder Preparation 68 Powder Delivery and Injection System 68 Powder Injection and Plasma/Hot Gas Jet 73 Injector Plugging and “Spitting” 75 Powder Buildup at the Front Nozzle Wall 77 Cooling System 77 Power-Supply System 79 Gas Supply and Distribution System 80 Manipulation Systems 81 Fixtures and Masking 83 Examples of Coated Power-Generation Components 84 Production Experience 86 Surface Preparation 87 Internal Plasma and Transferred Arc 89 Process and Systems 91 The Programming of the Coating Process 94 Finishing 95 Repair of Turbine Parts 95 Coating Removal, Stripping 97 Restoration of the Base Materials 98 Refurbishing, Recoating 98 Commercial 99 General 99 Surface Preparation 103 Coating Equipment 103 Finishing 104

3

105 Quality Assurance 105 Sources of Process Variations 105 Special Causes of Coating-Process Variation 107 Stochastic Nature of a Spray Process 108 Arc and Jet Pulsations 108 Powder-Size Distribution 109 Powder Injection 110 Powder Shape 110 Particle Bonding 110 Gun and Component Motion and Positioning 110 Drifting 111 Stability of the Quality Control 112 Process Capability and Stable Process 115 Definition of Process Capability 115

3.1 3.2 3.2.1 3.2.2 3.2.2.1 3.2.2.2 3.2.2.3 3.2.2.4 3.2.2.5 3.2.2.6 3.2.3 3.2.4 3.3 3.3.1

Quality and Process Capability

Contents

3.3.2 3.3.3 3.3.4 3.3.5 3.3.6 3.3.6.1 3.3.6.2 3.3.6.3 3.3.6.4 3.3.6.5 3.3.6.6 3.3.6.7 3.3.6.8 3.4

Definition of a Stable Coating Process 117 Operational Window 118 What Process Capability is Required? 122 Additional Factors that Affect the Process Capability 124 Case Study: Achievable Process Capability 125 Part Complexity 125 Mutual Position of the Gun and Component Fixtures 125 Powder Quality 125 Torch Pulsations and Drifting 126 Instability of the Quality-Control Process 128 Surface Preparation and the Part Temperature 128 Conditions of the Powder-Injection System 129 Process Capability 129 Maintenance 130

4

133 Coating Formation from Separate Particles: Particle Impact, Spreading and Bonding 133 Physics of Plasma Torches 138 Plasma Properties 139 Gas Dynamics of Plasma Torch 145 Energy Balance of the Plasma Gun 147 Major Trends 149 Variation of the Gun Power; the Gas Flow Rates and Composition Unchanged 149 Variation of the Plasma Composition at the Same Specific Plasma Enthalpy 149 Variation of the Plasma Flow Rate at Unchanged Gun Power and Gas Composition 150 Effect of Nozzle Diameter 151 Plasma Swirl 151 Structure of Plasma Jets 151 APS Jet 151 Structure of LPPS Jet 153 Particles in Plasma 155 Particles at APS 156 Particle at LPPS 158 Particle Acceleration and Heating in the LPPS Free Jet 158 Particle Acceleration and Heating Inside the Nozzle 160 Spray Footprint (Spray Pattern) 161 Influence of Particles on Plasma Flow 164 Substrate Surface Temperature 165 Formation of the Coating Layer 167 Use of Different Plasma Gases 168 Some Distinguishing Features of HVOF Physics 169

4.1 4.2 4.2.1 4.2.2 4.2.3 4.2.4 4.2.4.1 4.2.4.2 4.2.4.3 4.2.4.4 4.2.5 4.3 4.3.1 4.3.2 4.4 4.4.1 4.4.2 4.4.2.1 4.4.2.2 4.5 4.6 4.7 4.8 4.9 4.10

Theory and Physical Trends

XI

XII

Contents

5.7.1.5 5.7.1.6

Offline Simulation of a Thermal-Spray Process 171 Simulation in Production 171 Physical Background of Simulation Package 175 Viscoplasticity Model of a Splat and Particle Bonding 175 Thermodynamic and Transport Properties of Argon/Hydrogen Mixtures 176 Modeling of the Plasma Gun 176 Modeling of the Plasma Jets 176 APS Jet 177 LPPS Jet 177 Acceleration and Heating of Particles in Plasma 179 Surface Thermal Conditions 180 Spray Pattern 182 Calibration of the Bonding Model and Sensitivity of a Spray Pattern to the Process Parameters, Spray Angle and Bonding Model 182 Coating Porosity and Roughness 185 Modeling of Turbine Blades 187 Coating Thickness Optimization and Stochastic Modeling Tools 189 Simulation of HVOF Process 195 Use of Offline Simulation in Coating Development 199 Application Areas of Modeling in the Coating Process 199 Coating Definition and Design for Coating 199 Coating-Process Development 199 Part Development 200 Physical Modeling and Offline Simulation as Process-Diagnostic Tools 201 Simulation as a Numerical Experiment 201 When the Offline Simulation Should Be Used 202

6

Standards and Training

5

5.1 5.2 5.2.1 5.2.2 5.2.3 5.2.4 5.2.4.1 5.2.4.2 5.2.5 5.2.6 5.3 5.3.1 5.3.2 5.4 5.5 5.6 5.7 5.7.1 5.7.1.1 5.7.1.2 5.7.1.3 5.7.1.4

6.1 6.1.1 6.1.2 6.1.3 6.2 6.2.1 6.2.2 6.2.3

205 Standards, Codes 205 Introduction to Standards 205 Quality Requirements for Thermally Sprayed Structures and Coating Shops 206 Qualification and Education of Spraying Personnel 209 Special Case: Spraying for Power-Generation Components 211 Coating-Process Development 212 Coating Production 213 General Requirements for Coating-Shop Personnel 213

7

Monitoring, Shopfloor Experience and Manufacturing Process Development 215

7.1 7.1.1 7.1.2

Monitoring, Sensing 215 Introduction of Monitoring 215 Particle-Monitoring Devices 217

Contents

7.1.3

7.3.3 7.3.3.1 7.3.3.2 7.3.3.3 7.3.3.4 7.3.3.5 7.3.3.6 7.3.4 7.3.4.1 7.3.4.2 7.3.5

Influence of Spray Parameters on Particle Speed and Temperature 218 Influence of Particle Velocity and Temperature on Microstructure 219 How to Use Monitoring for Process Control 222 Monitoring, Sensing from a Job Shop Point of View 222 Vision for Future Coating Control and Monitoring 224 Manufacturing Coating Development 228 Coating Development Process 229 Coating Definition and Coating Specification; Design for Coating 230 Process Development 233 Powder Selection 234 Torch Parameters 234 Spray Pattern and Standoff Distance 234 Coating Mono-Layer; Powder Feed Rate and Traverse Gun Speed 235 Spray Trials and Coating Qualification 235 Sensitivity Checks 236 Part Development 236 Coating Program 236 Process Qualification and Preserial Release 237 Serial Release 239

8

Outlook, Summary

7.1.4 7.2 7.2.1 7.2.2 7.3 7.3.1 7.3.2

8.1 8.2

241 Thermal Spray Torches 242 Future Offline Programming and Monitoring in Process Development and Production 244

References 245 Subject Index 261

XIII

XIV

Disclaimer While every precaution has been taken in the preparation of this book, the publisher and the authors assume no responsibility for errors or omissions, or for damages resulting from the use of the information contained herein. As far as the authors of this book specify products of third parties they merely provide a description pertaining to this book. They do not want to promote or advertise any product or are liable for specific qualities of these products. In no event the authors are liable for damages suffered or personal injury including every kind of damages, especially consequential damages, arising out of the use or the inability to use these products. The same applies to the facts and information taken from foreign authority. The authors do not guarantee that the provided facts and information is state of the art, correct, complete or the quality thereof.

1

1 Introduction 1.1 Requirements for Materials and Coatings in Powerplants

We do not want to write another book on thermal spraying, plasma spraying, HVOF (see Section 2.4.3) and other spraying processes. We will not repeat what is already written in excellent books, reviews, and journals. Many general descriptions of thermal spraying can be found today in the Internet on web pages of equipment suppliers, material and gas suppliers, coating shops and research facilities. Our intentions are to show some ways how to achieve a stable reliable coating production for power-generation equipment within reasonable time and at optimum cost. We will address how to identify problems and mistakes in advance. We will show how to minimize development effort and to improve product quality. First, we will try to simplify and summarize the topic of this book: Electric power generation today and in the future is using and will use steam turbines, gas turbines and turbogenerators, steel tubing and heat exchangers and boilers. Components consist of many parts that are welded, brazed or assembled. Each part has a specific function within the powerplant. The original equipment manufacturers (OEMs) and the powerplant customers like utilities or other power producers consider as the most important parameters of a powerplant: x x x x

Investment cost Operation cost Long-term reliability Availability and scheduled, short maintenance

These parameters translate into requirements for components like material cost, optimized fuel cost, high operation temperatures and long operation times without in-operation control possibilities. Today, all powerplant hardware is coated wherever no affordable and reliable structural material can be found that resists the operation environment. For simplification we start with the view of a metallurgist: Metallurgists select materials for specific applications or for a variety of applications. A powerplant is basically built from metals. The structural materials and functional materials are metals and metallic alloys:

2

1 Introduction

x x x x

Steel, low alloyed up to high chromium steels Nickel alloys Cobalt alloys Copper and brass

In some rare cases titanium alloys are used. This picture is completely different from aero engines where the weight of a part is important. In power-generation parts weight is only important as material cost and for rotating parts if weight causes mechanical stresses. Designers select materials for operational conditions like: x x x x x x

Mechanical stresses, loadings, strains Operational temperatures Temperature changes Environment, atmosphere Design lifetime (times and cycles) Expected safe operation times

and of course for cost reasons. If a material class is not able to withstand the operational temperatures cooling is required by available cooling media that are mainly air, steam, or water. In closed-cycle cooling other media like hydrogen are being used. In many cases a division of material properties for a variety of tasks is required. Base metal has to have the required strength. Coatings withstand the environmental attack or add additional properties like wear resistance. In cooled components thermal-barrier coatings reduce the temperature gradient within the structural material. The designer selects the structural material and the coating by iterating the loading, component thicknesses and cost.

1.2 Examples of Coatings in Gas Turbines

We promised to address powerplant components. However, when we look more closely we find the following situation: In steam turbines thermal-spray coatings are not in standard use. In certain cases erosion damages are solved by replacing missing material by a thermal spray overlay of erosion-resistant material containing tungsen carbide or chromium carbide. Large-scale application is found in boilers where the tubes are coated by wire spray. For example, FeCrAl and FeCrAlY coatings are used generally as high-temperature oxidation protection to resist corrosive gases in boiler atmospheres. The more complex applications are found in gas turbines, especially at higher temperatures. Therefore we will concentrate on examples from industrial gas turbines.

1.2 Examples of Coatings in Gas Turbines

Fig. 1 Siemens Westinghouse gas turbine (courtesy of Siemens).

Basically there are three types of components: x Large single structural components like casings x Multiple medium-sized components with plane or slightly curved surfaces, like combustor parts x Multiple complex-shaped components, like turbine vanes and blades The following example of a stationary gas turbine illustrates the situation (Fig. 1): The air intake (1) is a steel construction most probably painted with a zinc-rich paint. The compressor blades and vanes (2) are made out of Cr steels where in certain operation regimes aqueous corrosion, pitting corrosion might end up in corrosion fatigue or stress corrosion conditions. Here the OEM will decide to use higher alloyed steels, titanium alloys or protection of the parts by coating. For clearance-control purposes the counterparts of the rotating compressor blades might be coated with so-called abradable coatings. The hot section parts in the combustor (3) and turbine (4) are made out of nickel- or cobalt-based alloys. In some cases ceramics are used. If oxidation and hot corrosion becomes important coatings are also used. In some cases for aircooled components the cooling is assisted by ceramic thermal-barrier coatings that reduce the operational temperature of the structural material the part is made of. The exhaust (5 and 6) again is made out of zinc-plated or zinc-sprayed steel. Rotor and stator casings are steel components, sometimes coated. For certain operation conditions nickel-based alloys are used for rotor disks. Wherever parts are rubbing against each other in operation or in order to control gaps between components wear-resistant coatings or so-called abradables are being used. Years ago it was already noted that in aero engine components up to 80% of all components are coated by thermal spraying. Today, in stationary gas turbines probably 50% of components are coated. In earlier days galvanic processes like

3

4

1 Introduction

chrome plating, chemical vapor deposition methods (CVD) or pack processes (explained later in Section 2.1) had been used. Today many of them are replaced by thermal-spray processes. Table 1 shows examples of coated components, materials, coatings and the basic requirements for the coating application. Details of coating compositions and requirement for feedstock will be found later in the Section 2.2.

Table 1 Components of stationary gas turbines, ther base and coating materials. Component

Base metal

Coating

Coating process

Coating requirement

Air intake

Steel

Zinc, epoxy

Painting

Oxidation, aqueous corrosion, erosion

Compressor blading

12%, high Cr steel, Aluminum, ceramic, PTFE TiAl6V4

Painting

Aqueous corrosion, erosion, stress corrosion, corrosion fatigue

Compressor leakage control

12%, high Cr steel, Abradables: metal matrix, TiAl6V4 solid lubricant, and polyester

Plasma spraying

Leakage reduction

Assembled structures

Steel castings, Ni base castings and sheets

CrC, WC + Ni,Co

Wire spray, APS, HVOF

Wear, friction welding

Casings

Low alloyed castings

NiCrxx, NiAlxx

HVOF

Oxidation

Combustor parts

Ni, Co superalloy, Ni-based sheet

NiAl, MCrAlY

APS, HVOF

Hot corrosion, oxidation, bond coat

Cooled combustor parts

Ni, Co superalloy, Ni-based sheet

ZrO2 8Y2O3 1 1)

APS, HVOF

Thermal barrier, surface temperature reduction

Gas turbine blades and vanes

Ni, Co superalloy

Cr, Al

CVD, Aluminizing, Chromizing

Hot corrosion, oxidation, bond coat

M(Ni,Co)CrAlY (+ Re,Ta.)

LPPS, HVOF Hot corrosion, (+2nd process) oxidation, bond coat

PtAl

Galvanic Pt Aluminizing

AlSi

Slurry painting + sintering

ZrO2 8Y2O3

APS, HVOF

1) Yttria Stabilized Zirconia (YSZ).

Thermal barrier, surface temperature reduction

1.4 Thermal-Spraying Systems

1.3 Definition of Thermal Spraying (THSP)

We will use the definition of thermal spraying as given by ASM2): “A group of processes in which finely divided metallic or nonmetallic surfacing materials are deposited in a molten or semimolten condition on a substrate to form a spray deposit. The surfacing material may be in the form of powder, rod, cord, or wire” [1]. Another detailed description is found in the US patent classification [2]. Subclass 446 – sprays coating utilizing flame or plasma heat (e.g., flame spraying, etc.): Processes wherein (1) a gaseous flame is used to heat and project a coating material toward a substrate or (2) a coating material is converted to or engulfed by a highly ionized gas composed of ions, electrons and neutral particles in which the positive ions and negative electrons are roughly equal in number, and projected on to a substrate In addition, the following notes are included: (1) Torch spraying is considered a form of flame spraying and is included in this and indented subclasses. (2) Electric-arc metal spraying is properly classified in this and indented subclasses. (3) Explosive or detonation spray vaporization, wherein the vaporized coating is applied in the form of a spray is properly classified in this and indented subclasses. (4) Thermal spraying is properly classified in this and indented subclasses. In short: Thermal spraying are all coating processes that coat surfaces with heated particles that are deposited by a high enthalpy kinetic gas stream. The feedstock used could be wire (if the material can be drawn as wire) or powder.

1.4 Thermal-Spraying Systems

Thermal-spray equipment can be classified according to the energy source needed to heat and accelerate the particles. In the European standards EN 657 [3] as well as in the equivalent international standard ISO 14917 [4] the different systems are described. A typical overview of thermal-spraying processes is shown in Fig. 2. For power-generation components thermal spraying by gas and electric arc discharge spraying are applied.

2) ASM = American Society of Materials.

5

6

1 Introduction

Fig. 2 Overview of the different thermal-spray processes in analogy to EN 657 [3].

1.5 Coatings for Power-Generation Components

What is specific in coatings, especially in thermal spraying for power-generation components? Why do we need another book on the subject thermal spraying? There are so many excellent reviews around. When looking for thermal spraying in the Internet search engines like Google will show millions of web pages. Thermal spraying has been used for decades for applying coatings on components of industrial structures in order to protect them against corrosive attack or wear. The first applications go back to the year 1909. A Swiss patent was applied for by Dr. M. U. Schoop for using flame-spray techniques [5]. In order to answer the question “what is specific in coatings for power-generation components?” let us start with the design requirements shown earlier and apply them to coatings: x x x x x x x

Mechanical stresses, loadings, strains Operational temperatures Temperature changes Environment, atmosphere, chemical attacks Design lifetime (times and cycles) Expected safe operation times Cost

1.6 The Complete Manufacturing and Coating Process

The metallurgist translates these requirements into: x Coating chemistry x Coating microstructure, e.g. phases, oxides, grain size, porosity x Coating thickness For production and purchasing people these requirements have to be put into specifications for manufacturing and purchasing. The specification and the corresponding quality-assurance procedure have to ensure that the coating will meet the requirements of the powerplant operator: x x x x

Investment cost Operation cost Long-term reliability Availability and scheduled, short maintenance

The specifications for manufacturing and purchasing will address: x x x x x x x

Repeatable manufacturing process with defined process parameters Defined coating material, e.g. powder specification Required coating thickness and tolerance Required coating microstructure Allowable coating defects and microstructure Defined coating substrate interface and tolerances of bonding defects3) Defined coating surface, e.g. roughness, oxide layer, residual stress and tolerances

The answer to the question why this book is written is: We found a lack in combination of several disciplines that make a reliable, affordable coating. Only the teamwork of design, manufacturing and supplier is able to provide the right product. We will show as a thread running through this book that only the intelligent combination of process physics, accumulated experience and operator training can supply coatings with the required quality. Finally, by complying with such manufacturing and purchasing specifications the OEM or the overhaul shop will guarantee the reliable operation of the coated part in powerplant service.

1.6 The Complete Manufacturing and Coating Process

Coating never is a standalone process within manufacturing, repair or refurbishment of a component. Let us take the example of a turbine blade. Figure 3 shows a typical manufacturing chain for a new component. 3) Bonding defects are details in the interface coating substrate that are not allowed according to specification.

7

8

1 Introduction

Fig. 3 Gas turbine blade manufacturing process.

Before the investment casting takes place alloy has to be procured. Ceramic cores shaping the interior of the cooled blade have to be injected and fired to provide stability during casting at temperatures in the order of 1500 °C. Wax is injected around the core and a shell mold is applied. By removing the wax the cavity in the shape of the cooled blade is formed. Vacuum casting, finishing and heat treatment provide an airfoil that later will be coated. Other processes like machining, electro discharge machining (EDM) will follow before coating. It is evident that certain processes have to take place before coating and others will follow the coating process. The latter processes have to be done in such a way that the coating is not damaged by these operations. The coating process is not independent of the other processes. In more detail every coating process consists of 3 steps: x Surface preparation x Coating application x Finishing/post treatment All thermal-spraying processes require these 3 steps as well. When concentrating on the coating application we find the following situation: Coating by thermal spraying can be divided into 3 topics shown in Fig. 4 as the example of low-pressure plasma spraying (LPPS):

Fig. 4 Major parameters of influence of plasma spraying (courtesy of ASA).

1.6 The Complete Manufacturing and Coating Process

Fig. 5 LPPS process and system (courtesy of Sulzer Metco).

All three influencing parameters have a specific effect on the coating quality. The spraying equipment provides the coating thickness and microstructure, fixture and masking influence the coating thickness distribution. The powder forms the coating microstructure by chemical composition and grain-size distribution. Of course, this representation may be rather schematic and does not reflect the whole complexity of internal structures and cross-links between the topics. Details of process and system are given in Fig. 5. It can be clearly seen that the number of influencing parameters increases. There are not only the spraying equipment and handling system together with the control equipment that determine the coating quality. There are the outside factors such as gases, electrical power and cooling water that enter the system. All these parameters can be controlled within the production facility. However, the powder quality is controlled by the powder supplier. A more detailed view of additional parameters is given in Fig. 6. It shows that gas supply, power source, controller and cooling features represent important factors for coating quality. When analyzing the coating process many process parameters (without powder material) can be found. A system analysis divides each parameter into more subparameters. Each of the subparameters will influence the coating quality in a specific way. In addition, some of the parameters are not independent. They will influence each other. Another look at the coating process from a shop floor perspective, i.e. from practical experience is given in Fig. 7. Even more parameters are shown that can be adjusted or occur during coating production. All the examples show that there are a high number of parameters to be considered in order to produce a high-quality coating in serial production.

9

Fig. 6 System analysis for plasma spraying (courtesy of ALSTOM).

10

1 Introduction

1.6 The Complete Manufacturing and Coating Process

Fig. 7 Factors influencing the thermal-spraying process.

The excellent review on plasma spraying [6] estimates that 50 to 60 parameters have to be considered. When looking through the literature and conferences one gets the feeling that everything is addressed and already resolved. Many technical universities seem to have an activity in “plasma spraying” or “thermal spraying” in order to evaluate spraying parameters and their influence on coating properties. However, experience in production and procurement of powerplant equipment shows that always the same or new mistakes are made. Unknown coating defects

11

12

1 Introduction

arise. Changes in personnel result in a new learning curve. Deviation of established working parameters results in changes in coating quality and in a number of improvement actions.

1.7 Coating-Process Development

The basic principle for coating of power-generation parts is: When a new coating process is to be established a process development has to take place. This process development has to result in reliable, stable production. The main task is to find the operational window, i.e. the manufacturing regime where small deviations in process parameters have negligible effect on the product quality. A factorial test matrix will result in a huge number of tests required, which is already restricted by the fact what kind of power-generation parts have to be coated. Either they are single pieces, like one casing per turbine, or when they come in larger quantities like turbine blades they are very expensive easily summing up to thousands of Euro per destroyed part. Table 2 shows an example of a coating-development matrix for coatings for a turbine blade and the expected correlation with the coating specification. Let us use an example: take Table 2 and make crosses in each field where an experiment is required. If a turbine blade has to be coated by all three processes like APS, HVOF and LPPS, it becomes evident how many tests are required. In addition the table shows how important “in process control”4) is. This is especially necessary because in many cases there are no nondestructive methods available for controlling online the coating quality. Process development must result in a repeatable stable manufacturing and quality-assurance process. Every coating process shows a scatter in quality results specified in the coating requirements. The results can be measured by applying a 6-sigma routine and determining process capabilities. A 6-sigma process assures that only 3–4 defects per million [7] are allowed. A 4-sigma process exhibits already 6210 defects per million. Just to show an example: Assume a gas turbine with 400 coated blades. If the coating is a 4-sigma process then you will find 2–3 blades in the turbine with defective coating. The coating life determines the maintenance interval of the whole powerplants. Therefore the process capability is important and has to be measured. This process capability requires a good interaction between design and manufacturing. The result is a product that can be manufactured by a defined and released manufacturing process.

4) “In process control” means controlling the established process parameters during coating.

Cost

Coating quality Roughness

Bonding Bonding Bonding Bonding Porosity Porosity Bonding

Influence

Post treatment x Heat treatment x Surface treatment

HVOF

Porosity Cost

LVPS

Thickness Porosity

APS

Coating quality Coating quality Coating quality Coating quality Coating quality Coating quality Coating quality Porosity

HVOF

Coating parameters In process control

Thickness Thickness Thickness Thickness Thickness Thickness Thickness Stability Coating quality

LVPS

Equipment

Spraying Parameters x Current/power x Powder x Powder feed rate x Spraying gun x Tooling x Deposition efficiency x Diagnostics x Gas flow x Vacuum Relative movement x Speed x Angle

Precoating Surface quality x Cleanliness x Roughness x Oxidation x Preheating x Cooling x Transferred arc cleaning

APS

Table 2 Coating-development matrix.

Cost

Cost Stability

Stability

Cost Cost Cost 1.7 Coating-Process Development 13

14

1 Introduction

Because not all conditions can be tested for time or cost reasons and theoretical simulations alone are not accurate enough (accuracy of input data, calibration requirement) there is a lot of practical experience involved. Therefore pragmatic simulation and modeling technology are required. These tools have to be calibrated, optimized and validated by practical experience and know-how. Process development is always followed by a learning curve. In all cases the number of coated pieces is small. Therefore the development has to result in stable processes by using experience or modeling and simulation. In the past, the experienced operator or plasma sprayer was the only resource. Simulation and modeling is a difficult task with many parameters to be considered. But in principle it could be easy: The starting point for simulation in thermal spraying is x Input: Powder material, gas, energy. The result is x Output: Coated component with a specified coating quality. However, when you look closer you will identify complications. The thermal spray powder consists of millions of single particles of different sizes and masses. Each particle has a slightly different chemistry and morphology. The coating is produced on the component by layers of splats formed by a moving spray pattern across the surface. The spray equipment itself has fluctuations in arc discharge, gas regulation and powder supply. In the past, people applied coatings developed through the “trial and error” and the know-how of experienced people. This has led to the following situation: a “natural” scatter of coating properties like thickness or porosity. Because coating is sometimes considered an art, coating manufacturers are rated as average, excellent or low performers even when they use identical equipment, identical subsuppliers and apply identical manufacturing specifications. Questions come up and have to be answered: x Is there a given process stability? Is there an inherent process capability for a given specification that cannot be improved? What is the right preventive maintenance strategy? x Is there a chance to end up with coating processes with process capabilities of the order of other manufacturing processes, e.g. like milling. x How does stability of equipment (e.g. power source) influence the coating quality? x Is gun-cooling variation a source for spitting? x What is the effect of carrier gas, etc.? x What is the influence of the coating powder-manufacturing process? Is there an optimum?

1.8 Tasks for “Target” Readers

These considerations have to be made when applying both “trial and error” and simulation approaches. The final answer must be a manufacturing process that uses modeling to a certain extent and builds some kind of knowledge base to be used for new developments and manufacturing and in-service feedback.

1.8 Tasks for “Target” Readers

This book will help to establish the grounds of the knowledge base. This book is not a scientific book. Its target community is the practice, the application: x Designer: He has to understand the capability and limitations of thermal spraying. It is a line-of-sight process. Spray angle, speed and distance of the spraying gun determine the coating quality. Therefore there are certain optimum operational windows. x Financial controller: He will understand deposition efficiency (waste of powder) and the importance of maintenance and spare parts for quick changeover of worn equipment. x Coating manufacturing engineer: He will use offline programming and real equipment in an optimum mix in order to end up with stable processes in production after the shortest development time. x Production operators and quality engineers: They have to understand that a frozen process5) together with monitoring of a number of process parameters is the only way to ensure quality. x Operators of the equipment: They understand every variation and its influence on the frozen process. x Supplier of material to be coated, e.g. casting houses and machinists: They understand the requirements of the coater in respect of geometry, surface quality, gas content of castings, cleanliness from machining fluids. They all have to understand the final target in production: x Achieve process stability at minimum total cost x Introduce measuring tools like 6-sigma x Total Quality management (TQM), ISO 9000x We will show in this book that only the combination of theory and practical experience will result in a reliable process. It will be evident that the main issue is to keep a drifting process within the required operational window. 5) A frozen manufacturing process is established by process development. All process parameters and the equipment must not be changed after process development. It is assumed that such a frozen process always yields an identical result within a defined tolerance (process capability).

15

16

1 Introduction

In order to guide the reader through the book we selected the following structure: In Chapter 2 “Practical Experience Today” we will describe the experience today in coating processes, material supply, hardware and processes. It will show what the important parameters for establishing the operational window are. The following chapter shows the factors influencing the quality and process capability. It addresses the importance of equipment maintenance as well. Theory of plasma and HVOF spraying is addressed in Chapter 4. Chapter 5 explains an approach to the offline simulation of thermal spraying. Standards and qualification of personnel are given in Chapter 6. Based on all the information supplied Section 7.2 shows how monitoring can be applied for controlling the coating process in production. The importance of coating-process development for reliable production is given in Section 7.3, followed by an outlook in Chapter 8.

17

2 Practical Experience Today This chapter will provide some basic definitions for thermal spraying (Section 2.2) after a short description of coating processes (Section 2.1). As explained in the introduction (see Fig. 4) the feedstock – wire or powder – is one of the three major parameters. Therefore we will explain in Section 2.3 the types and manufacturing routes of powders and their characterization. Section 2.4 addresses thermalspraying equipment with emphasis on torches. Especially HVOF and plasma torches will be explained. Sections 2.4 and 2.5 describe components of the system and an example of a plasma-coating system is found in Section 2.4.1. In Section 2.6 we show examples of coated power-generation components. Section 2.7 provides some aspects of production experience. In Section 2.7.4 we will address repair and refurbishment of coated turbine components. We conclude the chapter with Section 2.8: “Commercial”.

2.1 Coating Processes

We will use the requirements of the stakeholders, e.g. designer, production manager and financial controller to assess available coating processes for powergeneration components and parts: x x x x x x x x x

Repeatable manufacturing process with defined process parameters Defined coating material, e.g. powder specification Required coating thickness and tolerance Required coating microstructure Allowable coating defects and microstructure Defined coating substrate interface and tolerances of bonding defects Defined coating surface, e.g. roughness, oxide layer Acceptable residual stress and tolerances Cost of coating, e.g. investment, consumables, process efficiency, maintenance

Additional important issues are: x What feedback information on the long-term behavior in powerplant service is available?

18

2 Practical Experience Today

Fig. 8 Questionnaire for coating-technology selection.

x How does the coating process fit into the total manufacturing chain? x Is it possible to apply the coating process on site or is it restricted to the shop floor environment. x What are the environmental restrictions? Is recycling of waste powder possible? A typical questionnaire for coating technology selection is shown in Fig. 8. Basically three major topics have to be considered: Quality and technology, process and cost, expressed in investment and operating costs. For power-generation components a robust technology with a potential for further development is very important. In addition, throughput time of the process is crucial for selection of coating processes. A final decision for any process is determined by initial and operating cost and by considering the workforce involved in it. Let us have a look at available coating technologies. There is an interesting chapter in an US EPA Capsule report [8] called: “Alternative Surface Finishing Processes and Coatings”. Here, amongst others, all surface treatments are summarized today used in the industry, especially for power-generation equipment: x x x x x x x x

Electroplating and electroless coatings Organic coatings Vapor deposition (physical (PVD), chemical (CVD)) Thermal spray Hard facing Metal cladding and bonding Alternative base metal Surface treatments like hardening, shot peening

Table 3 brings examples of coating process applications and certain features that are important for application. Painting and electrochemical processes like hard chromizing and zinc galvanizing are well known in the industry. Physical vapor deposition (PVD) and chemical vapor deposition (CVD) are used in many applications. Cladding is a process where mainly thick wear-resistant surfaces are applied. It may also be used for component forming. Thermal spraying has the advantage that multicomponent coatings can easily be applied.

2.1 Coating Processes Table 3 Coating processes and typical properties (courtesy of Sulzer Metco). Coating process

Typical coating thickness (μm)

Coating material

Properties

Examples

PVD

1–5

Ti (C, N)

Wear resistant

Tools

CVD

1–50

SiC

Wear resistant

Fibre coating

Paint

1–10

Polymere

Corrosion resistant esthetical

Automotive

Thermal spraying

40–3000

Ceramic and metallic alloys

Wear resistant corrosion resistant thermal protection

Bearing Turbine blades

Hard chromizing

10–100

Chromium

Wear resistant

Rolls

Weld cladding

0.5–5

Steel, stellite

Wear resistant

Valves

Zinc galvanize

1–5

Zinc

Corrosion resistant

Steel plates

Braze cladding

10–100

NiCrBSi alloy

Hard and dense

Shafts

Power-generation equipment is coated with processes most suitable for the specific component. For instance thermal spraying is a line-of-sight process. Therefore certain features that cannot be directly seen by the spray gun cannot be coated with these processes in contrast to gas phase technologies. For most powerplant components the temperature during coating is important for the mechanical and physical properties of the base material. Figure 9 gives an overview of the temperatures involved in coating processes and thicknesses that can be easily achieved by the respective processes. Typical thicknesses for powergeneration component coatings are in the order of 100 Pm to 1 mm.

Fig. 9 Qualitative comparison of different coating processes and their typical substrate temperature (courtesy of Sulzer Metco).

19

20

2 Practical Experience Today

On first glance one could conclude that thermal spraying is a process where one can adjust the component temperature to a low level where no influence on materials properties is found. However in metallic coatings very often spraying at high temperature is preferred in order to minimize residual stresses during cooling down of the solidified sprayed particles. Ceramic coatings sometimes are applied on components that are cooled during spraying in order to achieve a certain required porosity level. The comparison of coating processes has to take into account several decision criteria: At first we will discuss technical issues: Remember the requirement of the operator of a powerplant: x x x x

Investment cost Operation cost Long-term reliability Availability and scheduled, short maintenance

Long-term reliability and availability in powerplants very often require known processes. So in steam turbines metal weld cladding against droplet erosion is standard and can be applied easily on site. Wire spraying of steam generator tubing is a standard process. On the other hand an electroplating line cannot be easily fitted into a product line of a mechanic shop and requires extensive waste-disposal infrastructure. If the deposition of several elements or compounds is required thermal spraying is the optimum process because nearly all materials that have a liquid phase and do not segregate can be processed. CVD and PVD processes are very sensitive if several elements have to be deposited in the right composition. CVD and PVD processes are widely used for coating relatively small parts, e.g. blades for aero-engines, due to the fact that they can provide very good coating thickness uniformity, including in locations with limited accessibility. Investment, operation costs and throughput time usually are quite high, but the possibility of simultaneous (batch) coating of sets helps to mitigate this issue. Both of these advantages vanish to a great extent when large parts (compared to the size of coating chamber) for land-based gas turbines are to be coated and thermal spraying becomes more attractive due to a shorter manufacturing cycle and lower costs. Which method is to be used depends on the technical requirements laid down in the coating specification. Every company involved in thermal spraying has an overview of the methods and tries to classify them according to different coating properties. An example of such classification is shown in Table 4. All parameters shown are important for compliance with the coating specifications for powerplant components. The assessment of the available technologies can be started with checking bond strength, porosity and thickness. Therefore when a minimum porosity is required, HVOF and low-pressure plasma spraying (LPPS) are required. Ceramic powders and carbides should be applied by atmospheric plasma spraying and HVOF, respectively.

2.2 Basics and Definitions of Thermal Spraying Table 4 Classification of thermal-spray processes with regard to different coating properties (courtesy of Sulzer Metco). Flame spraying

HVOF spraying

Gas temperature [°C]

3000

3000

4000

4000

12 000– 16 000

12 000– 16 000

Particle velocity [m/s]

40

800

600–1000

100

200–400

300–600

Spray distance [mm]

100…200

120..350

50..400

80..200

80–250

300–500

Bond strength [N/mm]

8

> 70

> 70

12

60–80

> 70

Content of oxygen [%]

10..15

1…5

1…5

10…20

2…3

ppm range

Porosity [%]

10…15

1…2

1…2

10

2…5

< 0.5

2…6

1…9

1

10…25

2…10

3…15

Spray capacity [kg/h]

Detonation Electric-arc Atmosphespraying spraying ric plasma spraying

Vacuum plasma spraying

Considering all advantages and disadvantages thermal spraying offers often the best solution for power-generation components. This statement applies for new components as well as for overhaul. All thermal-spraying processes applied in air can be used on site as well.

2.2 Basics of Thermal Spraying

The coating results from the impact of accelerated particles (jet) on a surface that is called a substrate. The acceleration of the particles is achieved with a device called a gun or torch (Fig. 10). During thermal coating the particles are usually molten or at least softened (semi-molten). If they are not, the process is called “cold gas spraying” that virtually belongs to thermal coating. A jet of molten/softened and accelerated particles generated in the gun hits the substrate. The first mono-layer develops. If the gun rapidly moves to the spray point and rests there for some time, the resulting coating spot is called a spray pattern. A spray pattern is the major source of information about the entire process. It is like a first print of a brush that is later used to paint the surface. One can see the shape, size and position of the “brush” (see the book cover!). It is even possible to measure the weight of the layer. By combining this data with the known spraying time and the amount of powder used, one is able to determine the effectiveness of the coating process. This figure is called the deposition efficiency (DE). The definition

21

22

2 Practical Experience Today

Fig. 10 Basics of thermal spraying showing the heating and acceleration of the particles in the heat bath (flame, plasma, etc.) and impact at the surface (top), including the temperature versus time evolution (bottom).

of deposition efficiency is given in the standard EN ISO 17836 [9] for comparing different processes and spray materials. In our book we will differentiate between two definitions of the DE: x Process deposition efficiency, which is defined as the ratio between the weight of the spray pattern deposited on a big flat plate and the weight of powder injected. This is a characteristic of the process, i.e. it mostly depends on the gun design and settings, the powder properties as well as the distance between the gun and substrate (which is called the stand-off), substrate temperature and the spray angle, which is the angle between the plate and the torch axis. x Target deposition efficiency, which is defined for the component to be sprayed. This parameter also depends on the component shape and size and on the way the torch movement is programmed. To coat (“paint”) a surface with a specified spray pattern (“brush”), the latter has to be moved. By moving it along a certain line, the coating strip is produced. By “painting” many strips next to each other, the whole surface will be covered. The distance between two lines is called the line offset. If the lines of the second coating layer don’t match the position of the lines of the first one, this is called the layer offset. Further on, the movement speed of the gun is called the traverse speed. We will later discuss the physics and some elements of kinematics of the thermal spray.

2.3 Feedstock

2.3 Feedstock

The most important parameters, influencing the coating quality can be addressed using the 3 following process steps: x Surface preparation x Coating application: process and system, powder/wire, masking and fixture x Finishing We need a repeatable manufacturing process with well-defined process parameters. This statement applies for the feedstock material particularly. There are standards available: The powders are classified and specified according to the European standard EN 1274 [10] or equivalent ISO 14232 [11] and the wires, rods and cords for flame and arc spraying according to EN ISO 14919 [12]. 2.3.1 Wire

Every thermal-spraying process relies heavily on the materials used. Wire is used mainly for processes where the coating specification allows higher porosity and higher oxygen content. Table 5 shows examples of feedstock wire used for wire flame spraying of boiler tubes. Table 5 Feedstock wire (courtesy of Kanthal). Alloy

Kanthal Kanthal Kanthal Kanthal Kanthal Kanthal Kanthal Kanthal SW 010 SW 100 SW 030 SW 200 SW 210 SW 230 SW 806 Sw 782

Composition

FeCrAl

FeCrAlY FeCrAl

NiCr

NiCrFe

NiCrFe

NiAl

NiFe

Ni

–

–

–

Bal

60

35

Bal

52

Al

5.7

5.4

4.8

–

–

–

5

–

Cr

21

21

21

20

16

20

–

–

Fe

Bal

Bal

Bal

–

Bal

Bal

–

Bal

Others

–

Y

–

–

–

–

–

–

Melting, °C

1500

1500

1500

1400

1390

1390

1400

1435

Resistivity, : mm2 m–1

1.45

1.39

1.35

1.09

1.11

1.04

0.42

0.37

Density, g/cm3

7.10

7.15

7.25

8.3

8.20

7.9

8

8.2

23

24

2 Practical Experience Today

2.3.2 Powder

After the short excursion to wire feedstock we will now concentrate on powder, which is basically defined by chemistry and tolerances, powder size and form distribution and phases [10, 13]. For details please refer to the standard EN 1274 [10] Thermal spraying – Powders – Composition and technical supply conditions or its equivalent ISO 14232 [11]. Also Refs. [14, 15] could be referred. In thermal spraying it is assumed that the powder has an identical chemical composition to the sprayed coating. This fact means that while spraying, no selective vaporization takes place during heating of the powder. On the other hand, melting or at least sufficient softening of the powder must take place in the heat source in order to assure a uniform coating microstructure. A smooth powder-feeding process must be secured with a defined powder feed rate. These requirements are best fulfilled by powders with spherical shape and a narrow particle-size distribution. Wider tolerances in particle-size distribution, particle shape and chemistry reduce the powder costs considerably; however, this leads to an increase in scattering of the coating process and finally a reduction of coating quality. 2.3.2.1

Powder Types

Table 6 shows the material types used for thermal spraying. Evidently nearly every material can be produced in powder form. The types of powder to be used depend very much on the application and the planned spray system. Table 7 shows a variety of applications and examples of powder materials. The first three applications x Wear protection x Corrosion protection x Thermal protection are especially important for power-generation components. It is evident that every specific application requires a specific powder. It is the know-how of the coating user to select the right coating. For power-generation applications feedback from powerplant services is important. Therefore the coating behavior is only known after several years of powerplant operation. This feedback could influence the coating specification, e.g. the tolerances of the chemical composition of a coating powder.

Pure metals Ti Nb Ta Cr Mo W Ni Cu Al Si Zn …

Self-fluxing alloys

NiBSi NiCrBSi CoCrNiWBSi …

Table 6 Thermal-spray material.

NiCr al. NiAl al. MCrAlY al. CoCrW al. CuAl, CuZn al. Al al. Ni-graphite High alloy steel …

Alloys TiC WC Cr3C2 WCrC BC WB CrB MoB TiN SiN …

Carbides, borides, nitrides

Powder for thermal spraying Polymers PEEK PE PTFE …

Oxides Al2O3 Al2O3–TiO2 Cr2O3 Cr2O3–TiO2 ZrO2 ZrO2–Y2O3, CaO, MgO Ca5(PO4)3(OH) …

Mo + NiCrBSi NiCrBSi + WC Cu + W Al + Mo Cu + Cr Metals + ceramics Metals + polymers Ceramics + polymers

Mixtures – pseudo-alloys

2.3 Feedstock 25

Atmospheric corrosion Intermediate corrosion Oxidation Sulfidation

Coating requirements

Corrosion resistance Low porosity Good bonding Same heat expansion as the substrate

Abrasion Adhesion Erosion “Fretting”

Coating requirements

Wear resistance Good bonding Good cohesion Homogeneous

Material ZrO2–Y2O3 ZrO2–MgO

Material

MCrAlY NiCr Mo Hastelloy Inconell

CuNiIn WC-Co Al2O3–TiO2 Cr2O3 AlSi–Polyester

Thermal conductivity Thermal stability Good mechanical properties Controlled porosity

Coating requirements

Thermal Insulation Reflection Absorption

Thermal functions

Material

Low porosity High hardness

Corrosion protection

Wear protection

Table 7 Powder types and functions (courtesy of Sulzer Metco).

Al2O3 YBa2Cu3O7

Material

Good electrical insolating Low porosity No microcracks No oxides

Coating requirements

Isolation Conductivity High-temperature supercondition

Electrical functions

HA Ti TiO2 Al2 O3MgO2 NiAl

Material

Rough surface Good bonding Chemical stability

Coating requirements

Bioactive Reparations Free standard coatings Sensors Catalytic functions

Special functions

26

2 Practical Experience Today

2.3 Feedstock

Fig. 11 Powder production steps (courtesy of Praxair).

2.3.2.2

Powder-Production Processes and Morphologies

There are many routes for powder production. Beside chemical precipitation processes, mechanical crushing/milling or thermal treatments (atomizing/ sintering) can be used. In general there are several steps (Fig. 11) involved in powder production: the coating applicant requires a defined morphology (usually spherical) and a well defined powder-size distribution. Depending on the capacity of the powder-producing equipment blending might be required. This step sometimes causes discussion between the powder supplier and the coating applicant because sampling for quality control has to be clearly specified. The powder inspection will be described in more detail in Section 2.3.2.3. Packaging and shipping has to be controlled carefully in order to keep the powder dry. After mixing the starting components and melting there are very often several proprietary steps to form the right chemical composition. The following step in powder production determines the powder morphology and microstructure. In fact, it is a sequence of several processes described in the following paragraphs. A summary is given in Table 8. Only processes 1 (gas, water atomized), 4 (spherodized) and 6 (agglomerated, sintered and spherodized) deliver the required spherical particles. The following sections (courtesy of Sulzer Metco and HC Starck) describe the powder-manufacturing processes in more detail. Mechanical Techniques

Crushing and Milling (Fig. 12): This technique is used primarily for ceramics and brittle metals. Fine material will be melted using electric-arc furnaces. The melt is cooled down. Large cast blocks or rods of raw material are crushed by mechanical force producing angular, irregular shaped, blocky and dense particles. Crushed particles (–44 + 10 Pm) may be screened for use or milled to smaller sizes

27

28

2 Practical Experience Today Table 8 Different powder-production processes used for different spray powders leading to different morphologies (courtesy of Sulzer Metco). Production process

Material

Spray powder

Particle morphology

1

Gas-, wateratomized

Metals, alloys

Cu, Ni, Co, Ni50Cr, NiCrAlY, Zr, Ti, Ni20Cr

Gas atomized: more round, more expensive Water-atomized: coarse, cheaper, dense particles

2

Fused and crushed

Oxides

Al2O3, Al2O3–TiO2, Blocked, dense particles Cr2O3, ZrO2–Y2O3

3

Sintered and crushed

Oxide, carbides

Al2O3–TiO2, Cr2O3, WC–Co, WC–10Co4Cr

Blocked, porous particles

4

Spherodized Metals, (with a plasma gun) carbides

Mo, WC–17Co

Changed from blocked to round Porous particles

5

Agglomerated (spray dried)

Oxide

ZrO2–Y2O3

Porous particles

6

Agglomerated, sintered, spherodized

Metals, oxides, carbides, HA

Cr3C2–NiCr, ZrO2–Y2O3, Mo, WC–Co, HA

Porous particles

7

Cladded

Alloys, carbides NiMoAl, Ni5Al, WCNi

Fig. 12 Crushing and milling.

Dense particles

2.3 Feedstock

(< 0.1 Pm). Milling may be done by rod, ball, or fluidized-bed methods. When using rod or ball milling, care must be taken so that the grinding media does not contaminate the finished product. Examples are Cr2O3, Al2O3, Al2O3, TiO2, TiO2, ZrO2, Y2O3. Atomization (Fig. 13): The atomization process is used in the production of various metallic thermal spray powders. Water atomization produces larger (–100 + 75 Pm) irregular-shaped, oxidized, dense particles at high rates, which require dewatering, drying, and milling after production. Metals are combined and melted (e.g. induction furnace), fed through a nozzle into the atomization tank forming fine droplets, cooled by high-pressure fluids (gasses or water) and collected in a water-cooled chamber. Examples are NiCr, NiAl. Gas atomization produces smaller (–100 + 25 Pm), spherical, low oxide particles at lower rates than water atomization. Particles are dense and of high purity. Particles may be screened and packaged or milled to smaller sizes. These powders are typically free flowing. Examples are MCrAlYs, NiCr, NiAl, Inconels, Stellites. Other atomization methods are: inert gas or vacuum, centrifugal or ultrasonic atomization.

Fig. 13 Gas-atomized powders.

29

30

2 Practical Experience Today

Fig. 14 Blend of two metals/alloys.

Blending (Fig. 14): Mechanical blending is a simple, low-cost method used to produce multicomponent powders. All kinds of powders can be blended together using appropriate mixers. Raw materials are placed in a large end-over-end tumbling vessel and mechanically blended until the raw materials are randomly dispersed. Specialized vessels (e.g. a double-cone blender or V-blender) must contain sufficient empty space for movement of the powder within. If the particles are poorly bonded, it is possible for the materials to segregate while spraying. The size, density, and shape of the raw materials must be chosen carefully to minimize the amount of segregation. Agglomeration: Agglomeration is the binding of two or more particles to form one composite particle. Agglomeration can be achieved: x x x x

Mechanically Chemically By sintering (thermal granulation) By using of a binder (cladding, fluidized bed, spray drying etc.)

The particles cannot be easily separated after sintering or cladding. Cladding (Fig. 15): This process uses an organic binder (e.g. polyvinyl alcohol or carboxy-methyl cellulose) to “glue” small particles to a larger core particle. Slurries of binder, core material, cladding component and water are mixed in a heated chamber. Typical examples are Ni-Al, NiCr-Al, and Ni-Mo-Al. Spray Drying (Fig. 16): A sol or slurry of one or several fine materials of similar size (1–10 Pm) are atomized through a fine nozzle and sprayed into a heated (400 °C/750 °F) chamber. The dried particles are spherical and porous, ranging from –250 + 5 Pm.

2.3 Feedstock

Fig. 15 Porous-coated (cladded) powders.

Fig. 16 Spray-dried agglomerated powders.

Fine powders (–10 + 1 Pm) can be clad or agglomerated by using a binder and spray drying. This creates a relatively inexpensive, free-flowing, spherical composite particle. The spray-dried powders are then classified by screening and/or cyclonic separation. Sintering (Fig. 17): Sintering is a process where heat and pressure cause the particles to fuse to each other, typically WC/Co-type materials. Examples are WC–Co, WC–Co–Cr, Cr2C3–NiCr, Mo, Al2O3–TiO2, ZrO2–Y2O3. Plasma Densification (Fig. 18): Plasma densification is a process whereby an agglomerated or sintered powder is injected into high-energy plasma and melted. The molten droplets are sprayed into a water-cooled chamber and collected for classification. Plasma densification alloys the agglomerated particles into a dense, spherical and homogenous powder.

31

32

2 Practical Experience Today

Fig. 17 Spray-dried and sintered powders.

Fig. 18 Plasma-densified (spherodized) powders.

The HOSP™ process is similar to plasma densification, except that it is used specifically with ceramic particles and produces hollow, spherical particles. These densified powders produce premium coatings needed in the turbine industry. Examples are ZrO2–Y2O3, Cr2C3–NiCr, WC–Co.

2.3 Feedstock

Chemical Techniques

Sol–Gel – Freeze Drying: A sol or slurry is stirred into liquid hexane at –30 °C (–86 °F). The rapidly frozen droplets are then filtered out of the liquid hexane. Any residual solvents are evaporated at low pressures and temperatures using a solid-to-gas process that bypasses the liquid state (sublimation). 2.3.2.3

Powder Characterization

Powder characterization is the tool in quality assurance to determine the consistent quality of a powder used for production. Typical parameters are: 1. Particle size and particle-size distribution 2. Particle shape and morphology 3. Chemical/physical analysis 4. Flowability 5. Density 6. Specific surface It is difficult to overstate the importance of fulfillment of specification requirements on the powder quality. All the aforementioned parameters are directly reflected in the coating quality: x Particle size influences the particle trajectory in the plasma/hot gas plume and, consequently, the level of particle treatment and the shape of the coating spot, which determine the coating microstructure, the process deposition efficiency and, finally, the coating thickness. x Particle shape and morphology affect the feeding conditions (by influencing the apparent density) as well as injection conditions and in-flight parameters since they influence the aerodynamic forces and heat fluxes in the gas flow. x Powder chemical composition and certain physical characteristics (for instance, crystalline phase structure) determine the coating functionally. x The powder flowability strongly affects the powder feeding and injection conditions. x Powder density influences the deposition rate (in the case of volumetric powder metering) with an obvious effect on the coating thickness. x The powder specific surface is an important characteristic that influences particle in-flight treatment and density of voids and microcracks in the coating. The stability of the powder quality depends on the selected type of powder, on the supplier and on the incoming powder control procedures. In this chapter we will not go into details. We will only address the names and measuring methods. Specific details can be found in the literature, e.g. [16–18]. 1. Particle Size and Particle-Size Distribution

The range of particle sizes in a sample of powder can be determined by various methods: sieving, wet screening, sedimentation, microscopy and laser light measurements.

33

34

2 Practical Experience Today

Fig. 19 Typical particle-size distribution (Vol.%) (left axis) and cumulative results (right axis) measured using the laser-scattering method (size distribution: –38 + 8 μm, d50 = 19 μm).

The results are typically represented by a graph of weight per cent or volume per cent versus particle diameter. Other methods of presentation use the cumulative results (Fig. 19). Different measurement devices may deliver different results depending on the physical methods used. Particles with the same chemical composition but different particle sizes may not behave in the same way when thermally sprayed. The micrometer (or micron) is a millionth of a meter and is the typical unit to measure the particle size of thermal spray powders (abbreviated “Pm”). The smallest particle size that can be seen by the unaided human eye is about 40 Pm (e.g. a pencil dot). Table 9 Correlation of mesh size and sieve opening. Mesh size

Sieve opening [μm]

30 40 50 60 80 100 120 140 200 230 325

600 425 300 250 180 150 125 106 75 63 44

2.3 Feedstock

Sieve Analysis: Passing a powder through a screen of known size or a series of screens with decreasing opening diameters to classify the particle size [19, 20]. The greater the sieve number (mesh size), the smaller the openings in the screen (Table 9). Mesh size is the number of openings per square inch. The sieves have to be re-inspected to insure no displacement of the distances of the metal wires [21]. Light Scattering: The particle-size distribution is determined by a laser-scattering method (MicroTrac™). The laser scattering principle involves measuring the amount of light that is diffracted by the powder particles being sampled. Other Methods: For completeness, the following methods for particle size characterization are mentioned: x Sedimentation x Microscopy and image analysis x Fisher subsieve sizer (FSSS) 2. Particle Shape and Morphology (Production Process)

Powder morphology indicates how the material was manufactured. Scanning electron microscopy (SEM) is often used to gain closer insight into particle shape and surface roughness as well as showing small satellites (if any). The SEM displays the topography of the sample by measuring the various signals emitted by the sample after electron beam contact. 3. Chemical/Physical Analysis

Basically, all available methods are used: x x x x x x x x

Wet-chemical analysis Atomic absorption spectroscopy (AAS) Optical emission spectroscopy (OES) X-ray fluorescent analysis (XRF) Energy dispersive X-ray analysis (EDX) Secondary ion mass spectroscopy (SIMS) Auger electron spectroscopy (AES) Electron spectroscopy for chemical analysis (ESLA, XPS, ESCA)

4. Flowability

Rate of flow measures the time it takes a specific weight or volume of powder to flow through a funnel, called Hall flowmeter (ISO 4490, ASTM B212). Free-flowing powder will form a consistent “repose angle” to be used for measuring the flowability (ISO 4324: Surface active agents; powders and granules; measurement of the angle of repose).

35

36

2 Practical Experience Today

5. Density

Several methods are used and defined to determine the density of powders: x Absolute Density: The mass per unit volume of a solid material expressed in grams per cubic centimetres. x Bulk Density: Powder in a container or bin expressed in mass unit per volume x Apparent Density: The weight of a unit volume of powder, usually expressed as grams per cubic centimetres, determined by a specific method x Tap Density: The density of a powder when the volume receptacle is tapped or vibrated under specified conditions while being loaded. 6. Specific Surface (BET)

This method determines metal powder specific surface area by physical gas adsorption. 2.3.2.4

Powders for Power-Generation Applications

In power generation, thermal spraying is used for protection against wear, corrosion and oxidation and for the application of thermal-barrier coatings. For leakage control, abradables are used. In addition, for restoration of worn geometries thermal spraying is used, e.g. for superalloy components. Which powders a spray shop uses depends on the background. Shops working for OEM customers receive specifications that very often contain proprietary

Table 10 Powders used for wear protection (courtesy of Praxair). Composition

Powder name

Size

Cr3C2 (Sintered)

CRC-105

–325 mesh/+5 Pm –45 Pm/+5 Pm

Cr3C2 (Sintered)

CRC-107

–140/+325 mesh –106 Pm/+45 Pm

Cr3C2–7NiCr (Blended)

CRC-184

–325 mesh/+10 Pm –45 Pm/+10 Pm

Cr3C2–25NiCr (Blended)

CRC-106

–325 mesh/+5 Pm –45 Pm/+5 Pm

Cr3C2–25NiCr (Blended)

CRC-108

–140 mesh/+20 Pm –106 Pm/+20 Pm

Cr3C2–25NiCr (Agglomerated & Sintered)

1375VM

–325 mesh/+15 Pm –45 Pm/+15 Pm

Cr3C2–25NiCr (Agglomerated & Sintered)

1375VF

–400 mesh/+10 Pm –38 Pm/+10 Pm

Cr3C2–25NiCr (Densified)

1376T

–270 mesh/+20 Pm –53 Pm/+20 Pm

2.3 Feedstock

coatings. Non-OEM manufacturing and repair shops often use powders that are available on the open market. Through experience the spray shops derive their own purchasing specs from the OEM specification or the open-market specification. Experience is required to select the right powder and the right powder supplier. It is up to the spray shop to understand which variations are allowed without leaving the allowable operational window. This understanding is the spray shops knowhow and is never found in the open literature. Today apparently every powder supplier comes up with a distribution of powder sizes with generally different centering. Every spray shop wants to have a repeatable powder size with the same centering and standard deviation. However, this information, as a rule, isn’t specified by powder manufacturers. Reduction of the amount of fine powder

Table 11 Powders used for protection against hot corrosion and oxidation (only nonproprietary shown, courtesy of Praxair). Composition

Powder name

Size

CoNiCrAlY (Ni32Cr21Al8Y0.5CoRem.) (Atomized)

CO-127

–325 mesh/+5 Pm –45 Pm/+5 Pm

CoNiCrAlY (Ni32Cr21Al8Y0.5CoRem.) (Atomized)

CO-159

–200 mesh/+400 mesh –75 Pm/+38 Pm

CoNiCrAlY (Ni32Cr21Al8Y0.5CoRem.) (Atomized)

CO-210-1

–325 mesh/+10 Pm –45 Pm/+10 Pm

CoNiCrAlY (Ni32Cr21Al8Y0.5CoRem.) (Atomized)

CO-210-24

–325 mesh/+20 Pm –45 Pm/+20 Pm

CoNiCrAlY (Ni32Cr21Al8Y0.5CoRem.) (Atomized)

CO-211

–170/+325 mesh –90 Pm/+45 Pm

CoNiCrAlY (Ni32Cr21Al8Y0.5CoRem.) (Atomized)

CO-211-3

–100/+320 mesh –150Pm/+63 Pm

Ni–22Cr–10Al–1Y (Atomized)

NI-164/NI-211

–140/+270 mesh –106 Pm/+53 Pm

Ni–22Cr–10Al–1Y (Atomized)

NI-164-2

–200/+325 mesh –75 Pm/+45 Pm

Ni–22Cr–10Al–1Y (Atomized)

NI-343

–325 mesh/+10 Pm –45 Pm/+10 Pm

Ni–31Cr–11Al–0.1Y (Atomized)

NI-246-3

–325 mesh/+5 Pm –45 Pm/+5 Pm

Ni–31Cr–11Al–0.1Y (Atomized)

NI-246-4

–170/+400 mesh –90 Pm/+38 Pm

Ni–23Cr–6Al–0.5Y (Atomized)

NI-278

–170/+325 mesh –90 Pm/+45 Pm

Ni–20Cr–9Al–0.2Y (Atomized)

NI-292

–200/+325 mesh

37

38

2 Practical Experience Today Table 12 Powders used for thermal-barrier coatings (courtesy of Praxair). Composition

Powder name

Size

ZrO2–8Y2O3 (Agglomerated and sintered)

ZRO-271

Custom sizes available

ZrO2–8Y2O3 (Agglomerated and sintered)

ZRO-182

–140/+325 mesh –106 Pm/+45 Pm

ZrO2–12Y2O3 (Agglomerated and sintered)

Al-1078

–120 mesh/+15 Pm –125 Pm/+15 Pm

sizes below 10 μm is also desirable. The particles should be spherical in order to guarantee a smooth particle flow. Typical information on powders used for power-generation parts are shown in Tables 10 to 12. Powders for wear-resistant coatings contain hard phases like e.g. Cr2C3 in a metallic NiCr matrix. The carbides provide the required hardness against wear; Cr delivers a certain oxidation resistance. Metallic Co or Ni alloys contain Al and Cr for protection against oxidation and hot corrosion. Chromium forms chromium oxide or a NiCr spinel; Al forms aluminum oxide. These oxides slow down the oxidation process at the surface. The oxides grow to a certain thickness and spall off during cyclic operation of hot components. Therefore these coatings are consumed during powerplant operation. Additions of Y or other elements stabilize the oxides and prolong the coating life. These so-called MCrAlY coatings, where M = Co and/or Ni, are used as bond coats under thermal-barrier coatings. Typical representatives of thermal-barrier coatings are ZrO2 8% Y2O3, a material with a low thermal conductivity of approximately 2–2.5 W/mK as a bulk material and even lower as thermally sprayed coating due to designed porosity and high density of microcracks. Applied on cooled components it reduces the temperature gradient of the load-carrying structure. Many of these powders had first been used in aero engines. Later, operational conditions in stationary gas turbines led to modifications in order to cope with hot corrosion attack. For manufacturing and purchasing, the powders have to be specified including tolerances [10, 11]. The specification contains all information the coating applicant requires in order to repeat the coating process in production. The frozen production process requires a powder with a repeating quality. Typical material data sheets of powder suppliers are shown in Table 13.

2.3 Feedstock Table 13 Typical material data sheets for thermal spraying powder (courtesy of HC Starck). AMPERIT® 827 Chemical Formula

ZrO2–Y2O3 93-7

Chemical Name

Zirconium Oxide – Yttrium Oxide 93-7

Description of Product

Agglomerated, Sintered

Grades Available

Product Designation AMPERIT® 827.7 AMPERIT® 827.064 AMPERIT® 827.6

90/16 Pm 45/10 Pm 125/45 Pm

Chemical Characteristics (Mass fraction in % [cg/g]; ppm [Pg/g]) Y2O3 HfO2 SiO2 TiO2 Al2O3 Fe2O3 ZrO2

7–9% max. 2.0% max. 0.5% max. 0.4% max. 0.2% max. 0.3% balance

Physical Characteristics

827.7

176 Pm 88 Pm

100%

D 90% D 50% D 10%

85–99 Pm 47–57 Pm 21–30 Pm

53–63 Pm 31–39 Pm 17–21 Pm

107–125 Pm 68– 77 Pm 39– 49 Pm

Apparent Density acc. ASTM B 212

2.2–2.5 g/cm3

2.2–2.6 g/cm3

2.2–2.4 g/cm3

827.054

min. 100% 100%

AMPERIT® 410 Chemical Formula

NiCoCrAlY

Description of Product

Gas Atomized

Grades Available

Product Designation AMPERIT® 410.1

Chemical Characteristics (Mass fraction in % [cg/g]; ppm [Pg/g]) Co Cr Al Y Ni

20–26% 15–19% 11.5–13.5% 0.2–0.7% remainder

Physical Characteristics Particle Size Distribution

827.6

410.1 +45 Pm max. 7% –22 Pm max. 15%

45/22 Pm

39

40

2 Practical Experience Today

2.4 Thermal-Spraying Equipment

Thermal-spraying equipment is available from several suppliers. Standard suppliers supply equipment for surface preparation and finishing. Stability of the production process is of utmost importance. The equipment must be able to operate within the regime of the operational window. Restriction in selection of optimum process parameters will lead to lower-quality products. In nearly all cases, new powerplant components have already seen a major portion of value added before coating. In certain cases of powerplant service there is no replacement possible in the given maintenance period. Therefore, the coating has to fulfill the specification. First pass yield is required. Replacement of components, recoating or a too low coating life in powerplant operation will generate high costs. Stability means that a released manufacturing process can be repeated using identical parameter sets. The manufacturing results have to stay within the given manufacturing tolerances. Drifting is to be controlled. The issue of stability will be discussed in detail in section 3 of this book. Therefore repeatability is required and a strict monitoring of the equipment status. In addition, in production the layout of the process is very important as are all the handling activities in between. Transfer between stations must be quick. Productivity of equipment has to be adjusted and aligned in order to achieve the shortest possible lead-time and avoid bottlenecks or adverse work in progress (WIP)6) build up.

Fig. 20 Structure of thermal-spraying equipment within three levels (courtesy of Sulzer Metco). 6) WIP, work in progress is the hardware and its value added during the manufacturing sequence. Example: A casing value before coating amounts 100 000 €. This cost is in the books of the owner as WIP and he has to pay interest on it.

2.4 Thermal-Spraying Equipment

The equipment must allow a one-piece flow or the smallest possible batch size. On the other hand, there are health and safety and environmental restrictions. Noise and spread powders require housing. Fine powders can lead to explosions and health risks. There is an absolute need of protection! The major components of thermal-spraying equipment are shown in Fig. 20. 2.4.1 Example of a Low-Pressure Plasma-Coating System

Based on Fig. 20 we will show the components of an integrated LPPS system. Figure 21 shows a real system. It shows the size of a system. The photograph shows a complete LPPS system with all features installed, including controlling/monitoring device, power supply, vacuum chamber (center), and send in/take out stingers (left and right of chamber). The 3D model (Fig. 22) shows more clearly the components of such a system. According to Fig. 20 the core and key components are the manipulation systems and the spray gun with all their auxiliaries. Figure 23 shows the manipulation systems. On the left and right sides of the main central coating chamber are located the preheating chambers. They are alternatively loaded with the part to be coated. The parts can move back and forth and they can rotate. From the rear side the sting for the gun manipulation comes into the chamber. This sting again carries out translational and rotational movements. Inside the chamber the gun can be tilted. This allows a five-axis manipulation for the whole system.

Fig. 21 A typical layout of the LPPS process (courtesy of Sulzer Metco).

41

42

2 Practical Experience Today

Fig. 22 3D model of an integrated LPPS system (courtesy of Sulzer Metco).

Fig. 23 LPPS manipulation systems (courtesy of Sulzer Metco).

A schematic self-explaining view of the main parts of the chamber is given in Fig. 24. The key and core components like spray gun or torch and, powder feeder and the manipulation systems mostly determine the quality of the coating. Therefore stability is required and a defined maintenance strategy (see Section 3.4). The peripheral system partially has the task to protect the environment, the people on the shop floor. This applies also for the soundproof cabin. Other targets are controlled atmosphere, like a vacuum vessel. Filter, exhaust system, power and cooling water supply require stability as failure of any of them may disrupt the ongoing coating process.

2.4 Thermal-Spraying Equipment

Fig. 24 Schematic view of plasma spraying in a chamber.

The following sections address the spraying torches, powder delivery and injection, the manipulation systems, fixtures, masking and auxiliaries. Because of their importance, descriptions that are more detailed are done for HVOF and plasma systems. 2.4.2 Flame and Arc Spray Torches

Principle of flame spraying using powders (Fig. 25): The spray material in powder form is fed continually into a fuel gas-oxygen flame where it is typically molten by heat of combustion. A powder feed carrier gas transports the powder particles into the combustion flame, and the mixed gases transport the material towards the prepared workpiece (Fig. 25).

Fig. 25 Principle of flame spraying using powders.

43

44

2 Practical Experience Today

Fig. 26 Principle of flame spraying using wire.

Principle of flame spraying using wire (Fig. 26): The spray material in wire form is fed continually into a fuel gas–oxygen flame where it is melted by the heat of combustion. Compressed air surrounds the flame and atomizes the molten tip of the wire. This accelerates the spray of molten particles towards the prepared workpiece surface. Principle of electric-arc spraying (Fig. 27): Wire arc spray uses two metallic wires as the coating material. They are electrically charged and fed into the arc gun. When the wires are brought together at the nozzle, the opposing charges on the wires create an arc and atomize the now molten material and compressed air jet accelerates it onto the workpiece surface.

Fig. 27 Principle of electric-arc spraying.

2.4 Thermal-Spraying Equipment

In Table 4 the particle velocity of all thermal-spraying processes is compared. Evidently, the flame and arc spray processes have the lowest velocities. Therefore, the bond strength is low and the porosity is high. Applications are restricted towards components with lower cyclic stresses. However, the spray efficiency is medium to high compared with other thermal-spraying processes. 2.4.3 HVOF Process

High-velocity oxygen-fuel (HVOF) (Figs. 28 and 29), high velocity air-fuel (HVAF) and similar processes allow spraying of coatings with advanced and, often unique properties. Therefore, the processes have been used in numerous advanced applications and have become very popular within the thermal-spray industry over the last two decades. HVOF processes have found a wide acceptance in power generation as well, allowing high-quality carbide and metallic coatings to be

Fig. 28 Principle of high-velocity oxygen-fuel (HVOF) spraying using gas for fuel. An oxygen–fuel gas mixture is ejected from a nozzle and ignited externally of the gun. The ignited gases form a flame. The powder exits the gun through the flame and is propelled to the workpiece.

Fig. 29 Principle of high-velocity oxygen-fuel (HVOF) spraying using liquid fuel. An oxygen–kerosene mixture is injected into a combustion chamber of the gun. Either a pilot hydrogen flame or a spark plug will ignite the flame. The powder exits the gun through its barrel and is propelled to the workpiece.

45

46

2 Practical Experience Today

sprayed. Examples are Cr3C2-NiCr coatings sprayed on a tip shroud of blades as well as MCrAlY coatings sprayed as an overlay or a bond layer for TBCs. It is necessary to note that HVOF competes with LPPS in MCrAlY-related applications (Figs. 30 and 31) especially when relatively small blades and vanes are to be sprayed (see also Section 2.8). Besides, HVOF is very effective when large parts like combustors have to be sprayed by MCrAlY when LPPS cannot be used because of the size. In case of hard metal spraying, HVOF shows less decarburization and building of sub- stoichiometric phases compared to the hotter APS or LPPS. The coating quality in Figs. 30 and 31 is very similar. There seem to be more bonding defects and slightly higher porosity in the HVOF coating. The surface roughness is different because the LPPS coating was smoothed after coating.

Fig. 30 Micrographic cross section of LPPS MCrAlY coating (courtesy of Turbocoating).

Fig. 31 Microscopic cross section of HVOF sprayed MCrAlY coating (courtesy of Turbocoating).

2.4 Thermal-Spraying Equipment

Altogether, the HVOF coating is within the coating specification. Therefore, HVOF may substitute LPPS applications in gas turbines applications. The first industrial HVOF system was developed by James A. Browning [23]. The system was commercialized by Deloro Stellite Corp., getting the name Jet Kote®. A variety of different HVOF/HVAF systems has been developed and commercialized in thermal spraying since then. However, presently the most popular and widely accepted by the marketplace systems are Jet Kote® manufactured by Stellite Coatings, Diamond Jet Hybrid manufactured by Sulzer Metco, JP-5000® and similar liquid fuel systems (e.g. WOKA Star™-600) manufactured by Praxair – Tafa, Sulzer Metco and several other smaller companies. The performance of HVOF processes generally depends on the type of fuel and the stoichiometry ratio, combustion pressure, as well as specific features of torches design, like powder injector position, nozzle profile, combustion efficiency. 2.4.3.1

Comparison of HVOF Fuels

In Table 14 general information regarding the fuels commonly used in HVOF systems is presented [24]. In the Table 14 Tc is combustion temperature at atmospheric pressure (0.1 MPa); Tcc is combustion temperature at 0.5 MPa combustion pressure; Qg is a heat value per 1 kg of fuel; Qcg is a heat value per 1 kg of combustion products at a stoichiometry ratio D = 1; Wne, Une and Tne are, respectively, velocity, density and temperature of combustion products combusted at 0.5 MPa and exiting a profiled nozzle; O is thermal conductivity of combustion products; (UW 2 )ne is the kinetic energy of combustion products exiting the nozzle; (O Tne) is a parameter characterizing a heat transfer from combustion products to spraying particles. Table 14 shows that there are minor differences in combustion temperature, kinetic energy and capability of heat transfer when hydrocarbon fuels are used in Table 14 A comparison of different fuels combusted with oxygen at stoichiometry ratio = 1. Flame parameter Tc, K (0.1 MPa)

Hydrogen Propane Propylene Methane C3H8 C3H6 CH4 H2 3076

3094

3150

3052

Mapp C3H4

Kerosene-1 AlcoholC9H20 96% C2H5OH

3195

3109

2944

Tcc, K (0.5 MPa)

3296

3310

3380

3261

3435

3329

3128

Qg, MJ/kg

120.5

50.45

49

63.6

59.454

42.9

27.3

Qgc, MJ/kg

13.48

10.9

11.08

12.75

13.76

9.55

9.25

O, W/(m · K)

0.46

0.3

0.2

0.321

0.282

0.289

0.289

Wne, m/s

2287

1849

1840

1892

1845

1828

1772

Tne, K

2634

2632

2671

2602

2701

2642

2513

Une, kg/m3

0.07

0.108

0.11

0.103

0.985

0.111

0.117

369230

372416

368705

335296

370915

367378

789.6

774.59

835.242

761.682

763.538

726.257

(U W 2 )ne, kg/98 s 366125 OTne, W/m

1211.64

47

48

2 Practical Experience Today

HVOF processes. Consistency of the fuels composition, cost, availability, safety related issues and possible ash/carbon formation are the major factors to be considered in this case to choose a fuel. At the same time, combustion of hydrogen provides combustion products having approximately 50% better capability of heat transfer to particles in comparison with the conventional hydrocarbon fuels. This fact determines the necessity to use hydrogen for some of HVOF applications when heat transfer to particles, their softening and melting are desirable to provide coating formation. However, hydrogen applications are limited because of safety issues, relatively high operating cost as well as a limited availability in some regions. Hydrocarbon fuels have a significant preference in comparison with hydrogen when high efficiency of heat transfer and related high temperatures of spraying particles are not desirable as they may cause phase transformation, tensile stresses formation, particle splashing and related voids in a coating. Presently, kerosene is, probably, the most popular fuel because of its low cost, relatively good consistency, especially, minimum safety problems dealing with kerosene feeding under high pressure. Natural gas may be another inexpensive option. However, possible inconsistency of natural gas composition and contamination of the natural gas by byproducts may create some problems dealing with torch performance and consistency of spraying coatings beside the difficulties dealing with availability of high pressure delivery. Ethyl alcohol or ethanol may become a popular fuel for HVOF processes in the near future. Ethanol has a slightly lower combustion temperature in comparison with kerosene. However, ethanol consistency and low price should make it a competitive fuel. For example, in October 2005 wholesale ethanol prices were $1.20 per gallon, compared to approximately $2.50 for wholesale kerosene and gasoline. Using more ethanol would be also preferable from the environmental point of view. Burning ethanol results in 30 per cent less toxic emissions than burning petroleum fuel [25]. 2.4.3.2

A Brief Overview of the Major Existing HVOF Systems

Generally, the relative positions of different HVOF systems may be characterized by particles temperature and velocity [26]. Sulzer Metco has proposed the diagram shown in Fig. 32 illustrating positions of different HVOF systems. Jet Kote® (JK)

JK (Fig. 33) uses axial powder injection and operates at approximately 550– 700 slpm hydrogen flow rate and 220–300 slpm oxygen flow rate. Fuel-rich combustion mixtures are recommended as a rule by Stellite for hydrogen/oxygen providing lower oxidation of particles. Oxygen-rich combustion mixtures are recommended for propylene and natural gas providing relatively low formation of ash and related deposits inside the torch. The estimated heat power of JK is around 90–110 kW and operating combustion pressure is in a range of 0.25–0.45 MPa. JK uses cylindrical nozzles providing subsonic gas velocity or so-called “choked” flow inside a nozzle. Axial powder injection and features of JK nozzle design results

2.4 Thermal-Spraying Equipment

Fig. 32 Schematic illustration of ranges of particles temperature and velocity produced by different HVOF systems as well as “cold gas-dynamic” spraying (CDS) [26] (courtesy of Sulzer Metco).

in very efficient powder heating and particle temperature becomes very close to the temperature of the combustion products in the very beginning of the nozzle if the spraying material has a relatively high thermal conductivity. The temperature of JK combustion products when hydrogen is a fuel is around 2400–2600 K, which exceeds the melting point of the most types of metals, metallic alloys as well as metallic binders used in composite carbide powders. Therefore, generally JK sprays metallic coatings from molten or semimolten particles accelerated by a combustion jet.

Fig. 33 Jet Kote® (courtesy of Stellite Coatings)

49

50

2 Practical Experience Today

Formation of coatings from molten particles has advantages related to DE and to some independence in variation of spraying parameters like spray angle and distance. From another viewpoint, the formation of coatings from molten particles has disadvantages related to stresses formed during particle shrinkage on a substrate, relatively low bond strength for some materials, possible phase transformation and decomposition because of high temperature, particles splashing during collision with formation of related voids, etc. This is one reason why JK coatings have a thickness limit of about 380–510 μm for WC-based coatings and 510–630 μm for alloys like T-800. However, a low oxide content, clean and dense coatings make JK still competitive on the marketplace. Diamond Jet® (DJ)

Diamond Jet® and Diamond Jet® Hybrid (DJ 2600 for H2 option) (Fig. 34) also use the axial powder injection and similar hydrogen and oxygen flow rates to JK. However, DJ may have a lower temperature in the convergent part of the nozzle because of still incomplete combustion in this area as well as an additional 350– 500 slpm of cold air or nitrogen that is fed into the convergent part of the nozzle for cooling purposes (see Fig. 32). Diamond Jet Hybrid uses a profiled nozzle and the operating combustion pressure may be estimated on 0.3–0.5 MPa level. Therefore, combustion products have a supersonic velocity. Finally, Diamond Jet Hybrid design features may provide lower temperature and higher velocity of particles in comparison with JK. Hydrogen-rich combustion mixtures are also recommended to minimize oxidation of particles. Liquid-Fuel HVOF Systems

Praxair Tafa and Sulzer Metco licensed liquid-fuel-related HVOF technologies from James Browning. Presently, JP-5000® manufactured by Praxair Tafa, WokaJet™400 and WokaStar™-600 manufactured by Sulzer Metco and similar liquid-fuel torches generally use kerosene as a fuel with an operating combustion pressure in the range of 0.5–0.85 MPa. WokaStar™-600 can provide a higher combustion

Fig. 34 Schematic layout of the Diamond Jet® Hybrid (courtesy od Sulzer Metco).

2.4 Thermal-Spraying Equipment

Fig. 35 Illustrations of WokaStar™-600 (a) and Operating JP-5000® (b) (courtesy of Sulzer Metco and Praxair-Tafa).

pressure because of a slightly higher combustion efficiency (Fig. 35). All liquid-fuel HVOF torches have profiled nozzles, providing supersonic velocity of combustion products. Radial powder injectors are located downstream of a throat in a zone with a relatively low pressure of about 0.15–0.2 MPa. Therefore, this type of HVOF torches doesn’t need a pressurized powder feeder, which is of advantage for practical use. Kerosene flow rates are within 0.3–0.5 l/min and oxygen flow rates are within approximately 750–1000 slpm. The estimated equivalent heat power is around 250 kW. Oxygen-rich combustion mixtures are generally recommended by manufacturers for minimizing ash formation and related carbon “spitting”. Very high oxygen and fuel flow rates and related power create disadvantages dealing with relatively high operating cost and possible overheating of spraying surfaces. The thermal-spray industry would definitely welcome liquid-fuel HVOF torches with lower flow rates and operating cost. The design features and operating parameters of JP-5000® and similar HVOF torches generally provide particles with temperature below their melting point. Moreover, molten particles may result in a build up at the end of a barrel and related malfunctioning of torches. Particle velocity may achieve 700–800 m/s, which is significantly higher in comparison with JK and DJ. Deposition efficiency for this type of HVOF torches is usually at 35–50% level, which is lower in comparison with Jet Kote® and Diamond Jet®. However, the possibility to control stresses in sprayed coatings, the high bond strength and density, the possibility to spray very thick coatings, the low oxides content and the minimum or no decomposition of coating materials have made them popular and widely accepted HVOF tools. 2.4.3.3

Possible Improvements of HVOF Systems

Existing HVOF systems have satisfied some needs of the power-generation industry. Better coating quality as well as lower operating cost is still needed to provide the market with advanced coatings and more economical and efficient HVOF tools. Improvements of HVOF system performance may be dealing with potential optimizations of combustion module, nozzle/barrel geometry and dimensions as well as powder-injection conditions.

51

52

2 Practical Experience Today

Generally, different types of spraying materials need different kinetic and thermal energies to form desirable coatings. For example, copper, aluminum, nickel, zinc and some other materials have relatively low melting point and a yield strength of about 30–200 MPa at room temperature. This type of material doesn’t need high temperatures above 200–300 °C as only a high kinetic energy is sufficient for a coating formation in this case. Cold spraying is a confirmation of that. Materials with higher yield strength and/or melting point need higher particle temperature. Thus, the operating window for HVOF processes may be generally determined as the maximum available temperature and velocity of combustion products as well as the range of their possible control and adjustments satisfying different requirements related to optimum kinetic and thermal energy of spraying particles for each particular type of material, like, for example, metallic alloys with low yield strength, superalloys with high yield strength, composite carbidebased materials having extremely high yield strength of carbides and metallic binder with relatively low yield strength, oxides with high melting point and yield strength. Control of particles’ temperature and velocity may be done, for example, by controlling the combustion pressure and nozzle expansion. The temperature of combustion products depends on the type of fuel (see Table 14), efficiency of combustion, combustion pressure as well as on the oxygen/ fuel ratio. The existing HVOF systems have relatively low thermal efficiency losing approximately 25–50% of the available combustion energy into the cooling water. For example, heat losses in liquid-fuel combustion chambers are at the 20% level. And there are unavoidable additional heat losses in the nozzle/barrel. Table 15 [24] illustrates that decreasing the heat losses results in increasing the temperature of combustion products as well as improving heat transfer to particles. Decreasing heat losses also result in the increased amount of dissociated molecules. Further recombination of the dissociated molecules finally results in better control and homogeneity of particle temperature and velocity. Thus, one of the possible ways to improve HVOF performance is the development of high-efficiency combustion modules. In Table 15 D is the amount of dissociated molecules; Tcc is combustion temperature; O is thermal conductivity of combustion products; (O Tcc) is a

Table 15 Influence of heat losses in a combustion module on parameters of combustion products (calculated for kerosene, combustion pressure 0.5 MPa and stoichiometry ratio D = 1). Heat losses, %

0

5

10

15

20

30

40

3329

3294

3231

3187

3129

3015

2870

D (H2O), %

29

27

24

22

18.8

14

10

D (CO2), %

58

55

51

48

43.7

36

27

Tcc , K

¼, W/(m K) ¼ · Tcc , W/m

0.277 924

0.273 899

0.266 848

0.263 847

0.255 804

0.247 745

0.235 674

2.4 Thermal-Spraying Equipment Table 16 Influence of combustion pressure on parameters of combustion products (calculated for kerosene, heat losses of 20% and stoichiometry ratio D = 1). Pcc , MPa Tcc , K

0.1 2944

0.3 3069

0.5 3129

0.8 3189

1.1 3223

1.4 3253

2 3297

D (H2O),%

21

19.6

18.8

18.2

17.7

17.3

16.8

D (CO2),%

46.7

44.7

43.7

42.7

42

41.5

40.6

¼, W/(m K) ¼ · Tcc , W/m

0.247 727

0.254 779

0.255 804

0.26 829

0.262 844

0.263 855

0.266 877

parameter characterizing a heat transfer from combustion products to spraying particles at Tcc and 0.5 MPa pressure. There is a definite trend in thermal spraying to use higher velocity and relatively low-temperature particles to spray advanced coatings [27]. Higher particle velocity needs higher combustion pressure and liquid fuels like kerosene are the most attractive in this case. Table 16 illustrates combustion parameters of kerosene at elevated combustion pressures. The data in Table 16 show that increasing combustion pressure results in a higher combustion temperature and better heat transfer from combustion products to particles that may partially compensate the negative effects arising from decreasing particle dwell time because of their higher velocity. Thus, another possible way of HVOF performance improvements is the application of elevated combustion pressures. The further increase in pressure in JP-5000®, WokaJet™-400, WokaStar™-600 and similar existing liquid-fuel HVOF systems would result practically in a proportional increase in fuel/oxygen flow rates and the related significant increase in operating cost and put a limit on economically affordable high-pressure HVOF applications. Therefore, high-pressure HVOF torches need to be developed with smaller nozzle throat diameter thus still using economical flow rates. However, the data regarding ST-4000™ [28] that is just a downscaled version of JP-5000® showed the particle temperature at least 200–300 °C lower in comparison with JP-5000®. Lower temperatures resulted in a significant decrease of deposition efficiency. This fact may be explained by the higher influence of boundary layers inside a barrel and related heat losses on particle temperature and velocity when a smaller-diameter nozzle/barrel is used. Thus, a development of highpressure economical HVOF torches needs a thorough optimization of the nozzle/ barrel geometry. Besides, the development should be based on high-efficiency combustion modules that allow for compensating of some heat losses. Shock-wave generators inside the nozzle/barrel may also be included into the development to allow an improvement in powder injection and increase in the heat exchange between products of combustion and particles [29, 30]. Application of different HVOF nozzle attachments is another potentially promising option to provide better control of parameters of sprayed particles [31–33].

53

54

2 Practical Experience Today

2.4.4 Plasma Process

By generating an ignition between an electrical plus pole (anode) and an electrical minus pole (cathode) an electrical arc is formed. Through the pressure of the gas, the temperature of the arc increases, the gas ionizes and plasma develops. At the same time powder is injected into the plasma jet, the powder particles melt and reach the surface of the workpiece at high velocity. Plasma spraying (see Fig. 36) is used in power generation for production of thermal barrier, anticorrosion, wear-resistant, abradables and some other coatings extending the life of elements and improving the performance of turbines. Powder morphology and consistency, plasma gases composition as well as features of plasma systems design and performance have the major input in coating quality and consistency, deposition efficiency and process repeatability.

Fig. 36 Principle of plasma spray gun.

Generally, a “perfect” plasma system should satisfy at least the following major requirements: 1. Long-term stability of plasma parameters. 2. Minimum or no pulsing of plasma parameters in the range below 3000–4000 Hz. 3. Minimum erosion of electrodes. 4. Consistent powder feeding and injection. 5. High deposition efficiency. 6. Process capability. 7. Relative simplicity, easy to maintain. 8. Lower investment and maintenance costs. Long-term stability of plasma parameters is needed during production providing long-term reproducibility of coating properties and keeping constant deposition efficiency.

2.4 Thermal-Spraying Equipment

Fig. 37 Illustration of F-4 anode wear during a study of long-term stability of plasma spraying [34, 35].

Drifting of voltage is explained by erosion and wear of electrodes as well as an arc trend to minimize its length. Figure 37 illustrates anode wear for F4 plasma torch after different operation times. Figures 38 and 39 [34, 35] illustrate drifting of the DC voltage and related evolution of the torch power plasma net energy as well as in-flight particle temperature and velocity during 55 hours of spraying. Long-term instability of some plasma systems may result in significant amount of rejected sprayed parts. Coating weight gain control on spraying parts or witness samples are highly recommended to determine a moment when electrodes are to be changed because of the drifting. A control of spray pattern shape may be also used for the same purpose. Plasma parameter fluctuations with relatively low frequency below 3000–4000 Hz are a prime source of statistical variance in the in-flight particle velocity and temperature. For example, particle temperature variations as large as 600 °C and particle velocity variations as large as 200 m/s were observed for ceramic particles sprayed by F4 and MB type plasma torch [36]. Thus, a particle travelling on a peak of plasma energy may be overheated, while another particle travelling at minimum plasma energy may have a temperature below the melting point, thus forming “unmelts” and other related defects in a coating or being just bounced off the surface. Figure 40 illustrates the time dependence and corresponding spectra on the voltage fluctuation [37]. The data show that maximum intensity of fluctuation

55

56

2 Practical Experience Today

Fig. 38 Evolution of the plasma-gun parameters during 55 h of spraying using the nominal operating conditions (constant arc current). Arc current (A) and DC voltage (B) are used in the calculation of the gun power (C) and plasma net energy (D) (courtesy of ASM) [35].

Fig. 39 Evolution of the plasma gun power (A) and net energy (B), in-flight particle temperature (C) and velocity (D) during 55 h of spraying. Nominal operating conditions (circles), constant power (squares) and constant particle state (triangles) (courtesy of ASM) [35].

2.4 Thermal-Spraying Equipment

Fig. 40 Time dependence (A) and corresponding power spectra (B) on the voltage fluctuation for new (–) and used (…) electrodes and for two gas mixtures. Plasma parameters, 50 slmp Ar or 4/50 slmp H2/Ar, 500 A, 3 Ps sampling time (courtesy of ASM) [35].

was observed around 4000 Hz. Argon plasma has a minimum level of the voltage fluctuation. However, application of Ar/H2 plasmas results in a significant increase of the fluctuations. The increase of fluctuations was also observed when worn electrodes were used during the experiments. Requirement of minimum erosion of electrodes is caused by several important reasons. Firstly, a minimized erosion of electrodes results in longer stability and less pulsation of plasma parameters as discussed above. Besides, products of electrode erosion may contaminate a coating and cause the coating failure. An example may be the TBC coating below contaminated by copper inclusion. Such inclusions may cause premature failure of the TBC coating during thermal cycling.

Fig. 41 Illustration of TBC coating contaminated with copper eroded from F4 anode.

57

58

2 Practical Experience Today

2.4.4.1

A Brief Overview of Plasma Torches

Plasma systems and torches presently used in power generation may be subdivided into the following groups using their design and performance features: 1. Conventional plasma torches operating at a relatively low voltage of about 40–90 V and a high current of about 400–800 A and even more. 2. Conventional plasma torches operating at a high voltage exceeding 200–250 V and relatively low current that, as a rule, below 400–500 A. 3. Multi-torch plasma systems (e.g. triple torch). 4. Cascaded and segmented plasma torches. Conventional Low-Voltage Plasma Torches

A major part of design and development of this type of torches was done in the 1960s and 1970s. A significant fraction of coatings in power generation is still sprayed by conventional low-voltage plasma torches like F4, 7MB, 9MB manufactured by Sulzer Metco, SG-100 manufactured by Praxair-Tafa and similar torches manufactured by several other companies. However, this type of plasma torches has major disadvantages related to the plasma-parameter longterm instability, plasma-parameter pulsing, high current and related short life of electrodes as well as significant limitations in process capability. Generally, these torches operate at relatively high plasma flow rates and generate turbulent plasma jets. Disadvantages of the conventional plasma torches have been well investigated and reported [34–37]. Significantly, more stable and efficient plasma torches are needed to satisfy the present requirements to plasma spraying of coatings in power generation and similar industries. Conventional High-Voltage Plasma Torches

Development of high-voltage conventional plasma torches started in the second half of the 1980s as an attempt to design more stable and efficient plasma torches for thermal spraying. At the end of the 1980s – beginning of the 1990s James A. Browning developed and introduced the first plasma torches of this type [38–40]. The torch developed by Browning has been manufactured and marketed by Praxair – Tafa under the name of PLAZJET®. Standard operating parameters are the following: arc voltage is of approximately 400 V; arc current is 500 A or lower; recommended plasma gas is nitrogen/hydrogen mixture; nitrogen flow rate is approximately 230–240 slpm, hydrogen flow rate is 90–100 slpm. Thus, the total flow rate of plasma gases is 320–340 slpm. The thermal efficiency of the torch is 60–70%, which is high in comparison with low-voltage conventional torches. PLAZJET® may operate with cylindrical, converging/conical and stepped anode nozzles. The stepped nozzle [39] creates a discontinuity in the anode passage, thus providing an additional source for arc stabilization. The stepped nozzle exit diameter may vary within 12–14 mm. Plasma torch type HE100™ manufactured by Progressive may be attributed to the second generation of the conventional high-voltage plasma torches. The initial concept of the torch was developed by Lucian Bogdan Delcia in 2000 [41]. The major idea of the invention is to stabilize the anode arc attachment by 3 rings

2.4 Thermal-Spraying Equipment

Fig. 42 Illustration of HE100™ design features (a) and supersonic plasma plume (b) (courtesy of Progressive Technologies, Inc.).

and grooves included into the anode design. Figure 42 illustrates some features of the HE100™ design. HE100™ also has an option providing an axial powder feeding into the plasma jet [42, 43]. HE100™, as a rule, operates at the 210–250-V level and currents below 350 A. Relatively low current and design features result in long life of electrodes. An argon/nitrogen/hydrogen mixture is recommended as a plasma gas with Ar flow rate within 75–175 slpm (160–370 scfh), N2 flow rate within 47–57 slpm (100–120 scfh) and H2 flow rate within 28–57 slpm (60–120 scfh). In particular, recommended flow rates, powder feed rate and gun power for MCrAlY and YSZ coatings are listed in Table 17. Table 17 Parameters recommended by Progressive for He100™ plasma torch.

MCrAlY YSZ (dense) YSZ (~13% porosity)

Ar flow, L/min (scfh)

N2 flow, L/min (scfh)

H2 flow, L/min (scfh)

Powder feed rate g/min

Gun power kW

118 (250)

47.2 (100)

28.3 (60)

100

70

85 (180)

55.6 (120)

56.6 (120)

50

95

75.5 (160)

56.6 (120)

47.2 (100)

100

85

59

60

2 Practical Experience Today

Fig. 43 Illustration of a long-term stability of HE100™ plasma parameters (a) as well as a long-term stability of velocity and temperature of spraying particles (b) (courtesy of Progressive Technologies, Inc.).

High-voltage conventional plasma torches provide significantly better longterm stability of plasma parameters and related long-term stability of velocity and temperature of spraying particles, which is illustrated by Fig. 43. This is the reason why this type of plasma torches has found acceptance in power generation and has been used for several applications. Multi-Plasma System

The Axial III plasma spray system may be attributed to multi-plasma systems. The Axial III was originally developed at the University of British Columbia by Doug Ross in 1990 [44]. The patent was exclusively licensed to Northwest Mettech Corp. Since then, other patents [45, 46] have been developed based on the Axial III electrode design and plasma-converging system. The Axial III concept consists of application of three plasma torches generating three plasma jets and having a common exit nozzle passage and axial powder feeding in the exit nozzle passage. Axial powder feeding may result in a very high level of deposition efficiency, coating quality and consistency for some spraying materials. However, complexity of the design and an excessive number of variables related to simultaneous generation

2.4 Thermal-Spraying Equipment

and control of the three separate plasma jets have resulted in a very limited number of Axial III applications in power generation. Cascaded Plasma Torches

Figure 44 depicts a general cutaway view of a cascade plasma torch, which has the single cathode and anode separated by one or more electrically insulated segments. USSR Certificate of Invention No. 125323 of December 18, 1958 [48] may be considered as one of the first references regarding a cascade torch application for plasma treatment of materials. The cascade plasma approach is very promising for plasma spraying as theoretically it may allow generating plasmas using a high voltage – low current approach with very good long-term stability, minimum or no pulsing of plasma parameters at frequencies below 3000–4000 Hz, long life of electrodes, possibility of using different plasma gases with or without minimum adjustments to the plasma gun, and an ability to minimize plasma gas flow rate, thus producing high-enthalpy plasmas. The potential advantages of the cascade plasmas caused a significant number of investigations and studies of different cascade-based plasma torches in the USSR in the 1970s and 1980s. Some results of the studies were summarized in [49, 50]. However, the cascade technology wasn’t really commercialized in USSR because of non-technical reasons. The first attempt to commercialize the cascade approach was done by Metco at the end of the 1980s [51]. The gun was marketed under the name of the APG™ plasma gun. However, that attempt did not result in a penetration of the cascade approach into the power generation. The major reason for this was probably related to the complexity of the design of adjustable cathode module and of some other APG™ elements.

Fig. 44 Cutaway view of a cascade plasma torch. (1) swirl nut; (2) cathode; (3) pilot insert; (4) neutral inserts; (5) anode; (6) electrically insulated nozzle (optional), (7–9) power source.

61

62

2 Practical Experience Today

The second attempt Sulzer Metco started in the middle of the 1990s based on Klaus Landes’ developments [52, 53]. The major idea of the developments is the application of a plurality of cathode elements in the cascade type of plasma torches. In the beginning of the development, an axial powder injection was included into the design [52]. However, the axial powder injection into a plasma channel created some challenges related to stability and performance of the plasma torches. Therefore, the axial powder injection was replaced by an external radial powder injection in the later designs [53, 54]. Three cathode elements have been used by Sulzer Metco in a family of plasma torches marketed under the trade name Triplex®. Three well-balanced power sources or a single power source with three identical outputs are needed to support the three parallel burning arcs with three anode arc roots having fixed positions. If the three currents of the Triplex® torch were merged into a single arc current, a fixation of the anodic arc root would destroy the anode rapidly because of thermal loads [55]. The first Triplex® system with an electrical power of 25 kW was put into operation in 1997. The second generation of Triplex® with the maximum electrical power of 55 kW was introduced in the industrial production environment in 1998. Several more Triplex®-based designs have been developed since then and the most recent TriplexProTM-200 providing maximum electrical power of 65 kW is shown in Fig. 45. Two major powder-feeding modes are recommended because of the asymmetrical cross section of a plasma plume caused by three anode arc roots [54]. The first mode is based on a use of a single powder injector and is recommended for applications, which require the high deposition efficiency. The second mode is using a triple injector system aligned in accordance with the positions of three anode arc roots. This mode is recommended for the spraying on large components. Generally speaking, the fixed position of powder injectors and their proper alignment with anode arc roots may create disadvantages for some applications where an adjustable spray pattern is required. Recent models of the Triplex® family of plasma torches operate, as a rule, using argon/helium mixture as a plasma gas and the following parameters:

Fig. 45 General view (a) and cutaway view (b) of a TriplexProTM-200 plasma torch (courtesy of Sulzer Metco).

2.4 Thermal-Spraying Equipment

x x x x x

Torch (console) voltage: 82–98 V (101–115 V) Current: 460–540 A Torch (Console) power: 44–52 kW (54–62 kW) Argon flow rate: 20–30 slpm Helium flow rate: 30–35 slpm

Within this range TriplexProTM-200 has demonstrated a long-term stability and minimum oscillations of plasma parameters as well as long life of electrodes, which exceeded 200 h [54–56]. The Triplex® specific total plasma gas flow rate (calculated per kW of the torch power) may be significantly lower in comparison with HE100™, PLAZJET® and conventional plasma torches. It means that Triplex® may provide significantly higher plasma enthalpy. Besides, low flow rates of plasma gases combined with 9–11 mm diameter of standard and high-enthalpy nozzles result in low Re numbers of a plasma plume, relatively low particle velocity and long dwell time that is especially attractive for high-efficiency spraying of low thermal conductivity/ high melting point ceramic coatings. At the same time, Triplex® may generate a supersonic plasma plume using the same or higher flow rates when 5–6.5-mm diameter high-velocity nozzles are used [56]. 2.4.4.2

Possible Improvements of Plasma Systems

Torch types HE100™ and Triplex® may be presently considered as the most advanced plasma guns providing stable plasmas, long life of electrodes and some other advantages. However, their penetration in power generation and similar industries is very slow even though there is an obvious need for better and higher efficiency plasma torches and systems. This contradiction may be probably explained as follows. The Triplex® design features allow use of low flow rates of plasma gases, thus generating very high enthalpy and low Reynolds number plasmas in the anode area in the vicinity of anode arc roots. Application of different exit nozzles may allow an increase in plasma velocity and Re number. However, in our opinion application of higher plasma gases flow rates to adjust enthalpy to a level similar to conventional plasmas is limited by the Triplex® design. Thus, Triplex® may have a very specific operating window with minimum operating window related to conventional plasma systems. On the contrary, the HE100™ design is based on very high flow rates and specific compositions of plasma gases based on argon/nitrogen/hydrogen mixtures allowing expanding of the arc and achieving high voltage. Thus, HE100™ may generate plasmas with relatively low to medium enthalpy and high Reynolds numbers. Therefore, HE100™ may also operate in a very specific operating window having only some overlapping with conventional plasma systems. A variety of different plasma-sprayed coatings has been developed in power generation and similar industries based on conventional plasma systems like F4, SG-100, 7MB, 9MB, etc. In our opinion slow penetration of the presently available advanced plasma systems into this existing segment of plasma spraying may be explained by prohibitively expensive R&D and approval phases needed to

63

64

2 Practical Experience Today

Fig. 46 Diagram illustrating relative range of operating windows related to conventional and cascade plasma systems.

switch from well-established but obsolete technologies to advanced technologies, as those technologies may have significantly different operating windows, spray patterns, etc. For example, optimum parameters may be developed for TriplexProTM-200 plasma torch to spray efficiently a high-quality TBC coating in accordance with one of the approved specifications. However, plasma generated by TriplexProTM-200 may have a specific swirl component. Therefore, the Triplex spray pattern may have significant difference from the spray patterns achievable by conventional plasma torches having swirl stabilization of plasma. Thus, all robot programs have to be modified for each particular blade, vane or other parts to be sprayed. This activity together with further approvals may be a big reason of slow implementation of the existing advanced plasma systems like HE100™ and TriplexPro™-200, even though the power generation industry desperately needs more stable and efficient systems for plasma spraying. Generally, a plasma spray torch consists of three major modules: a plasma generator providing a plasma with desirable parameters, an exit nozzle controlling the plasma velocity, and a powder-feeding/injection module. The exit nozzle may be a part of an anode or may be also electrically insulated from an anode. It may be concluded that acceptance and penetration of a new advanced plasma system into the existing segments of plasma spraying may be significantly faster in case if a new system would include a stable and efficient plasma generator capable of reproducing the approved plasma parameters. That generator should be coupled

2.5 Work Flow and Important Coating Hardware

to the powder-injection module and an exit nozzle, which have been used for the existing coatings. In this case only very minor development and related tests would be required to confirm the identity of the coating properties and spray patterns. Recent developments in the cascade area have shown that single cathode/single anode cascade plasma generators (SCPG) specifically designed for plasma spraying may satisfy the major part or all requirements to a “perfect” plasma system. SCPG can operate using argon and nitrogen as primary plasma gases as well as their mixtures with hydrogen and helium. Plasma gases flow rates are controlled within a very wide range, thus providing reproducibility of plasma parameters presently generating by conventional as well as advanced plasma torches. Figure 46 illustrates the relative position of the operating windows related to conventional plasma and cascade plasma systems. SCPG design has also demonstrated the following advantages: x x x x

Long-term stability of plasma parameters. Minimum or no oscillations of plasma parameters. Minimum erosion of electrodes due to high voltage – low current approach. SCPG allows adjustment of the swirl intensity and can be coupled with the existing powder-injection modules, thus reproducing the existing powderinjection conditions and related spray patterns. x SCPG has an extremely wide operating window allowing generation of quasilaminar, transient as well as turbulent plasmas. A wide operating window determines a unique process capability. x Design is relative simple, easy to maintain. The advanced performance of the SCPG may allow recommendation of this approach for the future consideration in power-generation and similar industries. However, it would take some time to transfer SCPG from the R&D phase to industrial applications.

2.5 Work Flow and Important Coating Hardware

Thermal spraying for power-generation components requires component-specific production strategies. A production strategy is determined by 1. 2. 3. 4. 5. 6. 7. 8. 9.

Type and size of component. Number of components to be coated in a period of time. Site of coating, being shop floor or on site in the powerplant. Location of the coating site. Standalone component or assembled component. Schedule available for coating. Available equipment and well-known spraying process. Available skilled people. Available untrained people.

65

66

2 Practical Experience Today

The production strategy goes along with a maintenance strategy and a qualityassurance strategy. In this chapter we will identify the classes of production strategies. In addition we will identify which rules are common to all strategies and where standardization is required. The input for every production is a specification and a drawing of a component. The specification defines the coating requirements in: x x x x x

Thickness Hardness Defects Microstructure Surface quality

Eventually, in addition, the following are defined: x x x x

Processes Process requirements Additional operations, like heat treatment Forbidden, excluded processes and operations

The drawing and 3D model show the areas to be coated, where coating is forbidden and how the transition is to be made [58]. During process development these requirements are transformed into operation sheets, tooling and fixtures and masks to protect areas not to be coated. The operation sheets contain the steps: x Before coating, i.e. surface preparation x Coating x After coating, i.e. finishing including the required quality assurance A typical manufacturing sequence is shown in Fig. 47 as an example: Coating of a set of 100 gas turbine blades with a MCrAlY coating. It is assumed that all operations take place in one shop floor. Therefore the transport time is short. This cycle takes about 2 weeks. The following sections will address important parts of the coating system.

67

Fig. 47 Coating manufacturing sequence.

2.5 Work Flow and Important Coating Hardware

68

2 Practical Experience Today

2.5.1 Powder Preparation and Powder-Delivery System

Stable optimized powder feeding and injection is needed to reach a stable coating quality, deposition efficiency and coating thickness distribution [59, 60]. Powder feeders presently manufactured by market leaders and some other companies may provide consistent powder feeding. However, there are several reasons that may cause inconsistency of powder feeding and injection. They may be related to the powder structure, hardware functioning and process arrangement. The following sections describe the situation in more detail. 2.5.1.1

Powder Preparation

Spraying powder is a specified mass product, whose conditions may affect the process quality and stability. The manufacturer of the powder describes the chemical components, technology used in production, size of the particles and some other powder properties. But there are characteristics, which are not specified but still have to be controlled. Humidity is one of the critical powder parameters. Excessive humidity can be caused by several sources. For example, the powder can become humid during storage. A drying oven with air exchange solves this problem, but depending on the subsequent application, specific drying processes have to be used. Possibilities are virtually unlimited. Whether drying it in the original containers, filling it into a shallow container to maximize the surface for evaporation or filling it in special devices where the powder is mixed and dried at the same time, there are many possibilities. Powder homogeneity is another critical feature to be addressed. Powder segregation during transportation and storing may be the major source of quality problems. The powder has to be dry and homogenous. In factories, where quality is a major goal, investments into the appropriate powder storing are fully justified. Depending on the size and importance of the production an air-conditioned room with constant low humidity would be the best arrangement. This should be combined with the use of drying and mixing devices. The powder should be taken from the storing facilities only when it is needed in production and only in amounts needed for filling the powder feeders. 2.5.1.2

Powder Delivery and Injection System

The powder-feeding system is expected to deliver a required stable amount of powder into the coating jet. There are many potential sources of problems: electrostatic charging, effect of erosion wear, unstable feed properties of powder and many others. Some of them are well understood and can be solved by the person who controls the coating process, some are usually not recognized, some are ignored and some are not even noticed or addressed too late, when the reduction of product quality has already happened. The main rule is that components of the system have to

2.5 Work Flow and Important Coating Hardware

be regularly inspected, maintained in faultless condition and properly assembled after inspections and maintenance. A typical powder-supply system consists of the following components: 1. Powder storage device (may be heated). 2. Powder-dosing device – powder feeder working on different principles like mass or volume per time. 3. Hoses, through which the powder is transported from the powder feeder to the injector using the carrier gas. 4. An injection block, which comprises one or several injectors (powder ports). 5. A powder carrier-gas supply system. A function of the powder feeder is to deliver the powder from the hopper into the injection hoses. Most of them use a gas as the powder carrier. They could be of mechanical and fluidized-bed types. Mechanical devices are more stable than the latter ones with respect to vibrations, gas pulsations, etc. The fluidized-bed feeders are advantageous from the viewpoint of maintenance since they have no moving and, therefore, wearing parts. From the point of view of powder metering the volumetric and gravimetric feeder types have to be differentiated. The gravimetric feeders are the preferred option when low scatter of the coating thickness is the critical target. The volumetric devices are simpler and cheaper. An example of the mechanical powder feeder is shown in Fig. 48. A powder insert is the major part of the mechanical system, which actually performs the function of powder volumetric metering and delivery into the powder hoses. From the operational point of view there exists a range for the gas flow rate between the minimum level, which is determined by the powder jamming in the system, and the maximum one, which can not be exceeding because of risk of gas pulsations. Due to the very strong interaction of fine particles with the walls as well as due to an increased probability of agglomeration, feeding of the powder fraction below 10 Pm is typically a challenge for the conventional feeder designs. From the standpoint of process stability the major issues that should be taken care of, are: 1. The powder could be segregated in the hopper due to influence of vibrations; this leads to the variation of the actual powder-size distribution with time. A passive vibration damping could be a solution for this problem. 2. The tightness of seals must be assured; this is especially critical for the LPPS powder-injection systems, which operate at the reduced pressure – in this case, not only the coating thickness but also its chemical composition and microstructure may be affected by leakages. 3. The worn mechanical parts must be regularly replaced. 4. The system must be regularly cleaned. 5. Preventive measures against accumulation of static electricity should be taken (e.g. by using electrical conductive hose and proper electrical grounding of the systems (same level as the spray gun).

69

70

2 Practical Experience Today

Fig. 48 A typical mechanical powder feeder (a) and the powder insert (b) (courtesy of Sulzer Metco).

2.5 Work Flow and Important Coating Hardware

Particles are delivered from the feeder into the injection port(s) by the carrier gas through the powder hoses. An average carrier-gas speed in the hose or injector is g R Tcg , where g is the volumetric carrier gas flow determined as u inj = 2 pinj 0.786 S dinj rate, R and T – its gas constant and temperature, p – pressure at the injector exit and d – injector/hose diameter. In Fig. 49 the mean gas velocity is shown depending on the carrier gas flow rate and the diameter of the hose or injector. The radial gas-velocity profile in the injection pipe is not flat nor is the radial particle velocity distribution. In the absence of particles one could expect a parabolic Poiseuille radial velocity distribution with the maximum speed up to two times its average value. In [61] a distribution of particle injection speed was directly measured. Results showed that the average and maximum values of the particle speed were approx. 80% and 180% of the average gas speed, which points out the situation of velocity equilibrium between the gas and particles. There exist several physical mechanisms that can affect this distribution. First, the flow with higher speed at the pipe centerline looses more momentum to accelerate the particles. Secondly, due to the shear lift force [62], particles tend to drift to the area of higher gas speed, i.e. towards the pipe axis. Correspondingly, the particle concentration increases at the axis, which results in increased gas momentum losses. These two effects flatten the speed profile and reduce the powder velocity scatter. Another reason is that due to the collisions with the hose wall, a particle may gain a radial velocity component, which would result in a conical shape of the particle flow at the exit [63].

Fig. 49 Gas velocity depending on the gas flow rate and diameter of the hose or injector.

71

72

2 Practical Experience Today

Fig. 50 A simple layout of APS powder feeding showing effects of siphon and of the hose-injector transition.

The powder hoses should be laid in the shortest possible way to the gun in order to avoid formation of siphons, which could provoke accumulation and jamming of the powder as is illustrated in Fig. 50. In the injection hose particles are accelerated by the viscous drag. A distance of speed relaxation can be estimated ugj Up Dp2 as x rel ∝ , where ug, Pg are the gas speed and viscosity, Dp, Up are the 36 Pg particle diameter and material density. For the typical injection parameters and the powder size of 20–50 Pm a relaxation distance is of the order of several centimeters, which is much more than any envisaged diameter of powder hoses and injector nozzles. From this observation, several practical conclusions immediately follow: x In the injection lines particles hit the pipe wall at every elbow and partly lose their speed. Therefore, the particle speed at the injector exit is determined to a great extent by the length of the straight parts of the pipeline and the number of elbows and bends. This effect depends on the hose diameter. x Injected particles always hit the walls of hoses or injectors when the gun moves with high acceleration. This is important for the understanding of such effects as the injector plugging and so-called “spitting”. x If the injector is located shortly after the last elbow the large particles do not have enough distance to be accelerated again and the different powder size fractions will have different injection speeds. x Frequently the powder hose and injector have different diameters; transition between them could be another reason for scatter of the particle injection speed as is illustrated in Fig. 50. Case Study

With a gas flow of 2 slpm in a hose of inner diameter of 4 mm, the gas velocity is approximately 2.65 m/s. Passing into the injector with a diameter of 1.8 mm,

2.5 Work Flow and Important Coating Hardware Table 18 Influence of powder-injection conditions on coating uniformity on a test sample. Powder hoses (diameter, length, looping)

Feeder position powder injection

Thickness standard deviation, micrometers

4 mm, 5m long, loops

down

24.3

4 mm, 3 m long, loops

down

25.8

4 mm, 3 m long, no loops

down

18.5

2.7 mm, long, no loops

down

19.7

2.7 mm, long, no loops

up

22.9

2.7 mm, short, no loops

up

17.2

the gas is accelerated up to 13.1 m/s. This causes a stratification of the powder injection speed due to different particle size, shape and weight. Therefore, a dispersion of the particles increases. This effect is aggravated if the injector is shortened and the distance becomes too small for the heavier powder particles to be accelerated to the gas velocity. The problem mainly emerges if powder with a broad spectrum of particle sizes or HOSP powder is used. The use of a hose with a diameter of 2.7 mm increases the gas velocity to 6.5 m/s at the same flow rate, which makes the difference between the gas speeds in the hose and injector much less pronounced. A dependence of the powder-feeding conditions on the coating quality has been observed experimentally. In Table 18 results from the APS spray tests with the different injector position, lengths of powder hoses and hose fixations are presented. The carrier gas flow rate was 1.7 slpm. Nonuniformity of the coating thickness on the flat test plate was the monitored parameter. It could be seen that the hose shape and dimensions can be sources of increased variation of the coating thickness. The standard deviation of coating thickness for coatings sprayed with the reduced diameter of houses is smaller. The same is valid for the hoses without loops. 2.5.1.3

Powder Injection and Plasma/Hot Gas Jet

There exist three general methods of particle injection into the plasma flow: external radial, internal radial and internal axial. Depending on the desirable particle trajectory, the radial injection can be realized with certain axial and azimuthal angles. Axial internal injection may provide the best conditions for the particle treatment in the plasma but is extremely difficult in technical realization (for examples see Section 2.4.4.1). In external injection, the powder injectors are positioned at a certain distance from the nozzle exit outside the plasma jet. Usually, this distance does not exceed 10 mm, but in some cases it can reach several centimeters (for instance, when the material should not be overheated as is required when polymers are sprayed). The process of injection is determined by the particle size and mass, injector diameter and the carrier gas flow rate. Due to the very low dynamic pressure of the carrier

73

74

2 Practical Experience Today

gas compared with the one of plasma jet, the former does not penetrate into the jet and is deflected by it. Therefore the role of carrier gas is to provide the particles with the velocity, which on the one hand would be sufficient for the particle to penetrate into the plasma and not to be deflected with the gas and, on the other hand, should not be too high in order not to have particle shooting through the jet instead being dragged by it. Particles penetrate into the plasma jet due to their inertia. Before actually entering the jet, the particles have to pass through the relatively cold and, therefore, dense boundary layer surrounding the jet. It has the radial component of the gas velocity, which is directed from the jet axis (the jet is always slightly expanding due to viscous losses). In magnitude it can reach several per cent of the jet speed (i.e. several tens of meters per second in the case of APS and more than 50 m/s at LPPS). This reduces the particle radial velocity and obstructs powder penetration into the jet core. One can expect this effect to be relatively weak close to the gun nozzle and increasing with the distance downstream. Therefore, the boundary layer plays the role of “wind”, which slows down the radial particle motion towards the jet. In the case of low injection velocities (below 2–4 m/s) and rapidly moving spray, gun particle trajectories in APS could be affected by the airflow, which would make a spray process sensitive to the direction of gun motion with respect to the injector axis. In one case particles can be dragged aside by the “wind”, in another they can be shaded by the injector holder. An effect of shear lift due to the radial velocity gradient in the jet boundary layer also should be mentioned with respect to the particle entrainment into the plasma. At the subsonic APS conditions this effect actually helps particles to enter the jet core. In the case of the supersonic LPPS flow, the shear lift has the opposite direction and creates a force that repels particles from the jet. This effect, together with the boundary layer “wind”, makes particle injection into the supersonic flow very difficult if not impossible. Effect of particle dragging by the LPPS jet boundary layer is shown in Fig. 51. In DPV2000 measurements, which were carried out during that experiment, no particles inside the jet were detected. In [64] a particle injection into the HVOF jet was ascertained in the specially arranged series of shock waves, which create local areas of the subsonic flow.

Fig. 51 External injection of the particles into the LPPS jet; effect of boundary layer “wind”.

2.5 Work Flow and Important Coating Hardware

In the case of internal radial injection, particles enter the jet together with the carrier gas slightly upstream of the exit of the APS nozzle or before the critical cross section (beginning of expansion cone) in the LPPS gun. In both cases the flow is either subsonic or trans-sonic and does not generate any forces that could obstruct particles entering the jet. It is worth noting that the internal powder injection increases the axial pressure gradient in the nozzle and, therefore, may cause the back-influence of the injection parameters on the torch operation. 2.5.1.4

Injector Plugging and “Spitting”

Due to intense heat fluxes from the plasma jet the tips of external injectors and exit areas of the internal powder feeders can be heated to the high temperature, which could reach 700–1000 °C depending on the injector material and cooling arrangement. In this case a particle that hits the wall close to the injector exit could be stuck on the hot wall due to adhesion forces, which exponentially depend on the contact temperature. This particle becomes an obstacle for the next ones and finally a particle compound is formed, which could block a significant part of the injector cross section as is shown in Fig. 52. It is heated by the plasma radiation and, in the case of internal injection, by the heat conductivity from the anode wall. In the case of particles with a high melting point, for instance zirconia or alumina, this compound is sintered and can be removed only mechanically. In the case of metallic materials with a relatively low melting temperature it melts and after reaching a certain size the droplet of liquid metal starts moving and falls into the main plasma stream. The latter phenomenon is typical for LPPS and APS with internal injection and in manufacturing is called “spitting”. Both plugging up and spitting are not acceptable in the coating process. The former effect changes the flow speed and direction at the injector exit, which could result in significant variation of particle trajectories and of the spray spot. The latter is detrimental for the coating quality due to the appearance of solid “bumps” on the coated surface, with sizes as much as 1 mm. There are many factors that determine the appearance of these effects. Let us consider some of them.

Fig. 52 Clean APS powder injector (a) and injector plugged by zirconia powder (b).

75

76

2 Practical Experience Today

1. Experience shows that the spitting and plugging frequently are related to the gun motion. In one of the case studies, the rate of spitting in LPPS spray of turbine blades was as much as 1–3 events per hour of spray, whereas no single spitting event happened during the 6 h of test spray with the immovable gun at the same torch and injection parameters. The reason for this is quite obvious. As was mentioned earlier, particles hit the injector wall during drastic lateral accelerations, which are inevitable at the actual spray with the moving gun. 2. Spitting and plugging are provoked by fine particles. Small particles are easily stuck at the surface due to the higher surface to volume ratio. In addition, they are “sunk” deeper in the near-wall boundary layer and can hardly be removed from the injector wall by the carrier gas “wind”. The practical recommendation is obvious: there must be strict control of the powder specifications and the presence of fines in the powder should be avoided unless it is necessary in order to create a certain coating structure. Experience shows that from the standpoint of spitting, a fraction of metallic powder smaller than 15–20 Pm is critical, whereas for the APS application of alumina and zirconia a risk of plugging increases with the presence of a significant fraction of particles below 25 Pm. 3. The risk of plugging and spitting dramatically increases with the temperature of the injector walls. It was observed that the spitting rate in LPPS with internal injection could be much higher if the anode with the tungsten insert is used compared to the plain copper nozzle. The heat conductivity of tungsten is much lower than that of copper. Respectively, the temperature of the insert inner surface and of the injection channel, which is drilled through it, is much higher and can reach 1200–1500 °C. Metal particles, for instance, an MCrAlY powder are easily stuck and melted at the channel exit. In the case of a pure copper nozzle the wall temperature usually does not exceed 800–900 °C and the probability of particle sticking is much lower. Also, it was observed that the rate of spitting increases with the worsening of cooling-water quality as well as due to the manufacturing deviations, which affect the cooling efficiency of the anode. In the case of external injection a forced cooling of injectors could significantly reduce the risk of plugging. 4. The risk of injector blockage is reduced with the increase of carrier gas speed – this “wind” helps particles to detach from the injector wall surface. 5. Increased roughness of the injector inner surface, which is usually caused by the erosion of wall material due to particle impact, can produce these effects. This imposes an “erosion” limit of the injector lifetime. Also any manufacturing defects such as scratches and cavities are extremely undesirable from this point of view. As far as prevention of injector plugging and spitting is concerned, a protective approach can be summarized as follows: 1. Better cooling, no fines and avoidance of unnecessary robot accelerations. 2. Regular injector inspections and mechanical cleaning. 3. Adaptation of parameters between injector and plasma conditions.

2.5 Work Flow and Important Coating Hardware

2.5.1.5

Powder Buildup at the Front Nozzle Wall

The affect of powder deposition at the front surface of the APS gun also should be mentioned with regard to powder injection. The plasma and powder carrier gas jets being discharged into the atmosphere create a complex hydrodynamic structure in the vicinity of the nozzle exit, where the appearance of recirculation areas is possible. Such flows may deliver a certain fraction of the injected powder towards the nozzle, where so-called powder “lumps” could build-up [65]. These lumps can detach from the gun surface and appear inside the coating, which is absolutely unacceptable from the standpoint of quality. In order to prevent such a scenario, usually the powder is blown off by the jet from the special gas port mounted on the gun. The blowing sequence is embedded into the normal coating program. The disadvantage of this approach is that an intense cold gas jet has a strong upstream influence on the nozzle operation and can irreversibly change the dynamics of the arc attachment and, therefore, affect the coating quality. Another approach would be to modify the gun design in order to prevent powder from flowing upstream as is described, for instance, in [65]. 2.5.2 Cooling System

None of the existing electrode materials can survive a long time in contact with the plasma without intensive cooling. Therefore, both anode and cathode have to be cooled by an appropriate cooling agent. Deionized water in a closed circulating loop is typically used due to its superior properties as a cooling agent and relatively low costs. Major requirements to the cooling water are the following: 1. It should not contain air bubbles, which could cause a local loss of cooling inside the torch (hot spot) and rapid damage of the electrodes, respectively. 2. Low electrical conductivity (below 10–20 μS): This requirement is necessary to avoid electrolytic chemical reactions, which result in rejection of the gas from the water and formation of gas bubbles, as well as power losses due to leakage currents. 3. Sufficiently high purity water in order to avoid an intense deposition of salts (including lime) inside the torch cooling system and related decrease of cooling efficiency. 4. Low inlet temperature; this requirement allows to keep the water outlet temperature at the acceptable level (increase in temperature intensifies release of the air dissolved in water). Usually, the temperature difference between inlet and outlet varies between 15 and 30 K, depending on the plasma operating power and water flow rate. In extreme conditions it could reach 40–45 K. Higher temperature increase inside the gun usually is an indicator of faulty operation of the cooling system or of its improper design. The typical coating-cell cooling system consists of: 1. The torch-cooling system, whose function is to assure the removal of heat from the anode and cathode.

77

78

2 Practical Experience Today

2. Water-supply pipes. 3. Water pump(s), which develops pressure required to pump the necessary amount of water through the entire hydraulic network. 4. Water chiller/heat exchanger with a tank to store the cooling water. 5. Water settling and conditioning tank, where the water is freed from air bubbles and chemically treated in order to remove the dissolved oxygen, organic compounds and salts. The latter measure helps to reduce oxidation of the copper electrode surfaces and to prevent contamination of the inner surfaces, which would decrease the thermal conductivity, and would lead to the final clogging of the cooling system. 6. Elbows and other hydraulic fittings. The number of these elements should be kept as low as possible in order to avoid excessive hydraulic losses in the line and corresponding reduction of the cooling-water flow rate. Malfunctioning of the coating-cell cooling system leads to the overheating of electrodes and reduction of their lifetime. Increased anode temperature could also stimulate such effects as powder build-up inside the anode and spitting (see Section 2.5.1.4). A significant overheating of the anode thermal regime could result in a change of the dynamics of the electric-arc root and by doing that could affect the coating properties. In Fig. 53 an effect of malfunctioning of the cooling system on the external surface of the anode of F4 gun is shown. The dark oxide layer reduces the heat transfer between electrode surface and cooling water and causes significant reduction of anode lifetime.

Fig. 53 F4 Nozzles after 40 h of operation, before and after improvement of the cooling system.

The typical maintenance tasks for the water-cooling systems are: 1. Assure a constant and low-temperature cooling-water supply. 2. Maintaining the optimal water quantity and quality in the system. 3. Control of the electrical conductivity of cooling water; if it exceeds a specified level, the system has to be drained, cleaned and refilled with deionized water and possibly enriched with antifouling chemicals. 4. Pressure in the cooling system should be maintained as high as allowed by the supplier; this assures the maximum possible water flow rate and better cooling conditions, respectively.

2.5 Work Flow and Important Coating Hardware

HVOF guns usually are also water-cooled (air-cooled high-velocity air fuel guns – HVAF – with the heat recuperation described in [66, 67] are the rare exceptions). Unlike for the plasma torches, the requirements on the water quality in HVOF are less stringent. This is because there exists no electric field inside and as a result the rate of salt deposition in the cooling channels is much lower. Usually, the time when one can think of cooling deterioration due to deposits, well exceeds the lifetime of gun parts to be cooled. 2.5.3 Power-Supply System

The power supply has to supply the system with high electrical energy and because of this it can be a critical component of the system. There are three power sources types available on the hardware market (courtesy Sulzer Metco). The work principles were responsible for their names: thyristor power source, primary and secondary chopped power source. The coating shops are accustomed to thyristors. This is an old proven technology but without any improvement potential for the future. The factories are afraid of changes because of lack of knowledge regarding new power-sources technology. Additional costs for tuning or redeveloping of coating parameters also cause new systems to be rejected. Sometimes it may be profitable to switch to the high-frequency chopped power sources since that would increase a deposit efficiency, for instance, for the APS spraying of porous TBC. But even in this case the break-even could be achieved only in a long term prospective. The advantages and disadvantages of different systems are listed below. Thyristor Sources (Fig. 54)

Advantages:

robust, suitable for high current and high voltage, simple control, simple wiring, low costs of production. Disadvantages: dimensions and weight, low effectiveness, interference of supply net, high ripple, low dynamics control. Primary Chopped Power Sources (Fig. 55)

Advantages:

small inductive components necessary by high work frequency, low weight, low ripple, high dynamics control for better control of process, possibility of modulation, high efficiency. Disadvantages: possible faults caused by fast-switching operations, complex and complicated wiring, high costs of production. Secondary Chopped Power Sources (Fig. 56)

Advantages:

low ripple, high dynamics control (better control of process with possibility of close loop control), possibility of modulation, high effectiveness, parallel (multiple) connections possible. Disadvantages: possible faults caused by fast switching operations, complex and complicated wiring, high costs of production.

79

80

2 Practical Experience Today

Fig. 54 Principle of thyristor power source.

Fig. 55 Principle of primary chopped power source.

Fig. 56 Principle of secondary chopped power source.

Independent on the type the power sources do not absorb too much maintenance resources. Keep them away from overheating, ensure the suitable cross section of cable’s connections and there are no problems to expect. Nevertheless caution has to be paid for not releasing any harmonic distortions back to the supplier net. 2.5.4 Gas Supply and Distribution System

This section is deliberately very short. The gases can be supplied by any of the gas supplier in the required quality. All the necessary data sheets can be found on the web pages of the gas suppliers. The required quality is determined by the expected

2.5 Work Flow and Important Coating Hardware

lifetime of the spray gun. Gases, their features, distribution, and connectors can be a source of serious quality problems in coatings. However, in nearly all cases the quality of delivered gas satisfies our expectation. It is the gas-distribution system that has to be optimized. Certain rules apply: x Gas metering and control devices have to be selected according to the required accuracy (class A r 1.5% or class B r 5%). They must be inspected and calibrated on a regular basis. Violation of this rule can become a major source of variations of the torch and particle parameters. x All devices and each connection must be absolutely tight; no leakages can be tolerated, since they lead to drastic changes in the coating chemistry, microstructure and thermo-mechanical properties. x Flexible connections should be avoided. x Purity for plasma gases should typically be > 4.7 (i.e. > 99.997%). 2.5.5 Manipulation Systems

The spray gun and work piece can be held and moved in relation to each other in different manners. The simplest method is the human hand, which holds the spray gun and passes repeatedly row-by-row over the surface to be coated, until the expected coating thickness is reached. This was popular in the past and is even today not replaceable in many complex and single-part coatings and in restoration applications. The coating quality and evenness depends strongly on the education and skills of the thermal sprayer. In applications where repeated coatings have to be applied on similar parts automation should be considered. Technical progress also delivered new possibilities in the field of handling systems: Simple electromechanical movement units, CNC-steered manipulators and robots with heavy loading capacity are nowadays prerequisites for constant and reproducible coating results in an industrial environment. Each manipulation system has its advantages and disadvantages. The advantages are often obvious, whereas the disadvantages force us to look for new and better solutions. For power generation components coated by LPPS vacuum technology, the CNC-steered units are usually used; for atmospherically applied coatings robots are the typical choice. We will use the neutral term manipulator for all the different moving methods. Due to the spreading of the spray spot there is in general no need for manipulator systems for thermal spraying to show accuracy of positioning which is required for tooling for cutting, milling, drilling machines. Depending on the spray gun and nozzle type and on the distance from the work piece the spray spot may be in the diameter range of approximately 10 to 50 mm. Instead of standard machine robots, X-Y-Z gun manipulators are often used. These manipulators exhibit worse repeatability and travel speed compared to robots. Nevertheless, X-Y-Z gun manipulator systems serve normal spray jobs. For

81

82

2 Practical Experience Today

Fig. 57 (a) Rotational handling system with a transverse sledge and accompanying exhaust system (back side) for flame and HVOF spraying of cylindrical work pieces. (b) 6-axis robot with plasma spray gun mounted in a vacuum plasma chamber.

coating application on drums and rolls a component-rotating manipulator with a transverse sledge will be the right choice. The rotating system typically runs at a higher surface speed compared to the transverse gun motion. Care has to be taken not to spray a helix on the cylinder surface, when the transverse speed is too high. Enough overlap of the spray spot (approx. 25%) after each rotation has to be guaranteed. There are helpful tools to calculate the rotation and transverse speed from a given cylinder diameter and requested relative surface velocity (see, e.g., Fig. 58). The technical demands on handling systems and acceptance/inspection of such systems are described in the European standard EN 1395-6 [68]. In the case of a rotational handling system with a transverse sledge, the verification of the leveling in longitudinal and transverse directions as well as the straightness and parallelism of spindle axis with the carriage movement should be inspected. In the case of multiaxis robots the movement range as well as the repeatability to reach an arbitrary position in the 3D space may be considered as a test criteria for reproducibility.

2.5 Work Flow and Important Coating Hardware

Fig. 58 Graph showing the relation between rotational (bottom axis) and traverse speed (top axis) for a given cylinder diameter (left axis) and requested surface velocity (courtesy of Sulzer Metco).

2.5.6 Fixtures and Masking

For stable repeatability of production only fixtures and masking should be used, which are produced in and for conditions of serial production, not for prototype or development purposes. These components have a major influence on the final quality and the amount of subsequent finishing treatment.

83

84

2 Practical Experience Today

In practical environment single use flexible tapes made out of woven glass fibers or silicon tapes, as well as metallic shields are in use. The accuracy and stability of position is especially important within the limits shown in Section 2.5.5. Masking provides a transition coating – uncoated areas (see design for manufacturing in Section 7.3). The definition of the transition – coated to uncoated surface – must be made in respect to the limits of thermal spraying. Sharp or too short transitions are not possible to coat. In LPPS units deflection of the plasma jet and shunting it to the masking during transferred arc cleaning may take place. Mask and fixture design today is still partially know-how, i.e. an art of the expert. Fixture design has to take into account high temperatures, thermal expansion. Damping of masks and fixtures is required; vibrations due to plasma jet pulsations have to be avoided.

2.6 Examples of Coated Power-Generation Components

This section shows examples of coatings, their main requirements and their critical features. Today in power generation, most of the components are protected by thermalspray coatings. Important parameters are always long/reliable lifetime and reasonable cost. Cost becomes more important when the to-be-coated surfaces become larger. This fact especially applies to coatings in boilers or in general in heat-exchanging components. Wear-resistant coatings have to withstand normal and shear stresses. Hot corrosion and oxidation resistant coatings are “eaten up”. Therefore, the reaction speed is the important parameter. Thermal-barrier coatings normally do not react or react only slowly. However, they are not stable at higher temperatures for long times. Deliberate defects like porosity or cracking might sinter together and decrease the cyclic behavior. We will show a few examples of coated power-generation components (Figs. 59 to 63). These components (Figs. 59 and 60) require a NiAl bond coat applied by APS. On top of this a thermal-barrier coating (ZrO2 8Y2O3) is sprayed also by APS. Both coating layers are applied by HVOF as well. The manipulation of the gun is performed by a robot. Only this automation assures consistent quality of the coating. Flat or slightly curved, medium-size components (Fig. 61) are coated by APS or HVOF using bond coat and TBC. If a complete ring assembly is possible, the segments are coated together as a ring in a tool. Such processes yield very high deposition efficiency and lower production cost. Turbine blades (Fig. 62) are coated with several coatings: Oxidation-resistant MCrAlY coating (sometimes also aluminized which is used as internal coating as well) by LPPS or HVOF, ZrO2 8Y2O3 thermal-barrier coating by APS or HVOF,

2.6 Examples of Coated Power-Generation Components

Fig. 59 Combustor front segment of a gas turbine coated with TBC (courtesy of Turbocoating).

Fig. 60 Combustor can of a gas turbine internally coated with TBC (courtesy of Turbocoating).

Fig. 61 Burner, combustor segment (courtesy of Sulzer Metco).

Fig. 62 Different turbine blades in shape and sizes (courtesy of Sulzer Metco).

85

86

2 Practical Experience Today

Fig. 63 Coating of boiler tubes by wire flame spraying (courtesy of Sulzer Metco).

wear-resistant Cr3C2-CoCr coatings by shrouded HVOF. In addition, the MCrAlY coating requires a diffusion heat treatment. If the component has to be assembled by brazing, the heat treatment cycles have to be optimized. Vane segments are difficult to spray because of nonoptimal spray angles especially in radii of tip shrouds. Boiler tubes (Fig. 63) are large structures that are wire flame sprayed either manually or using fully automated equipment. In principle, there are three types of components in powerplants, which have to be treated differently: x A single piece like a large casting x A medium-size structural part with simple geometry and more than one coating x Complex parts like turbine blades

2.7 Production Experience

The basis of this section is the workflow given in Fig. 47. It is evident that logistics and quality assurance play an important role in production. We will address this later in Section 2.8. However, we have to point out as well that handling is a critical issue. On the one hand shop floor people have to assure that they do not damage expensive parts, on the other hand lead time is important because of the cost of WIP. Therefore, only experienced, trained people are able to work fast without damaging parts and coatings. In this section, we will show some examples and will address critical parameters for production. Stability means that a released manufacturing process can be repeated using an identical parameter set. The manufacturing result has to stay within the given

2.7 Production Experience

manufacturing tolerances specified in the manufacturing order. Drifting is to be controlled. For thermal-spray coatings the manufacturing chain of all three processes: x Surface preparation x Coating consisting of process and system, fixture and masking, powder x Finishing has to work together in such a way that all tolerances and changes during the production process yield a result within the given tolerance band. This result is only possible when the operational window is known. Quality-assurance equipment is important as well. However, in some cases it cannot fulfil the requirements of specification, e.g. coating thickness tolerances, roughness measurement, bonding/adhesion, porosity, cracks, etc. Therefore the right adjustment between specifications and manufacturing, i.e. design for coating has to be achieved (see Section 7.3). Changes in equipment performance caused by wear and drifting have to be detected and controlled. The equipment should be able to operate around the clock. However, maintenance is required to exchange worn components, remove dust, clean and replace masks and fixtures. In Chapter 3 we will address these issues. In the following sections we will concentrate on the three steps of coating process. 2.7.1 Surface Preparation

The assumption that surface preparation is less important than the actual coating process is wrong. Especially for thermal-sprayed components the surface preparation step is very important [69]. Roughening of the surface by sand blasting is the standard process. The process has to be standardized and repeatable. Therefore, a certain degree of automation is required. Surface preparation has two intentions. First, it has the task to clean and activate the surface in order to avoid bonding defects in the coating/base material interface. Therefore, the time between grit blasting and coating should be kept as short as possible. The second task is to provide the right surface roughness for optimum coating adherence. Table 19 summarizes the activities required. Nozzle wear and distance nozzle-workpiece have to be controlled during grit blasting. An important issue in sandblasting is the angle under which the surface is to be treated. If the angle is too small, the roughening effect is too low. If the angle is too large, grit particle might be imbedded and could cause bonding defects. For surface preparation the knowledge of the surface treatment in the manufacturing step before coating is important. If the communication between the two steps (surface preparation and following coating process) is not sufficient, critical adhesion problems may occur. In some cases it is impossible to get coatings

87

88

2 Practical Experience Today Table 19 Surface preparation (courtesy of Sulzer Metco). Surface activation summary 1.

Make sure the surface to be prepared is free of grease and other foreign elements that can contaminate the grit.

2.

Always use air which is clean and dry at a regulated pressure.

3.

Always use the most appropriate grit for the surface preparation required.

4.

Ensure the grit size is kept within close limits.

5.

If the grit is recycled, always ensure its size is maintained and it is free of dust.

6.

Always ensure that optimum performance is obtained from the blasting equipment.

according to specification, because the supplier of the prematerial cleans the surface with grit blasting not suitable for the following coating process. Residues of this process sometimes already exceed the allowed range of bonding defects (number of inclusion per square area). In case of LPPS an improper surface preparation may cause serious problems with transferred arc cleaning. In the part and coating specifications the required roughness of the surface is prescribed. Roughness Measurements

The texture of a surface can be described and measured using different kinds of surface roughness parameters and measurement devices. The standard EN ISO 4287 [70] describes the terms; definitions and surface texture parameters, whereas EN ISO 4288 [71] describes the rules and procedures for the assessment of surface texture. Typically, people often use profilometry (stylus or laser instrument methods) and rely on the so-called average roughness, which is defined as: Ra =

1 N

N

∑ Zi i =1

The main surface topography influencing factors are the hardness of the substrates, the gas pressure in the case of high-pressure blasting, the standoff distance between the nozzle exit and the piece, the impact angle, the number of passes, the grit material and shape, particle size and size distribution [72].

Fig. 64 Definition of average roughness Ra out of a scanned line profile and Rt as maximum peak to valley difference within a measured length l.

2.7 Production Experience

Fig. 65 Polished surfaces of X5 CrNiMo 18–10 substrates after 1× (a), 5× (b), and 10× (c) passes of grit blasting (scanned area by stylus profilometry: 500 × 500 μm).

When multilayer coatings are applied, e.g. bond coat and topcoat, the surface of the bond coat can no longer be treated. Here, the required roughness on the surface of the bond coat must be achieved by spraying with the right powder-size distribution. There is a special case of surface preparation in low-pressure plasma spraying LPPS, the so-called “Transferred arc cleaning” (see next section). 2.7.1.1

Internal Plasma and Transferred Arc

Besides the internal plasma arc circuit between the cathode and anodic nozzle, a secondary electrical circuit may be supported in low-pressure plasma spraying using the conductive plasma as a “current lead” [73]. In the plasma core, the free ions and electrons contribute to the electrical conductivity. This, of course, is a function of plasma temperature and plasma gas mixture (see Section 4.2.2). This circuit is attached on one side to the electrically floating nozzle of the anode and uses the conductive plasma as the electrical coupling to the working part, which is connected to the other end of the power supply. The electric conductive work piece can either be connected to the positive or negative side of the secondary power supply (see Fig. 66). If the substrate is connected to the negative pole (so-called “negative transferred arc”; position 1 in Fig. 66), the current will flow from the anodic nozzle of the torch through the plasma to the work piece. The work piece itself plays a role of the electron emitter. The arc roots at the work piece are split into the multiple local spots continuously moving over the surface. The electrically conductive piece will be heated by the ion bombardment form the cathodic layer as well as by Joule heating in the surface spots. As soon as the work function of the metal oxide is reached, which is usually lower than that of pure metals, the arc roots tend to concentrate at the edges of oxide film, which is then melted and evaporated or

89

90

2 Practical Experience Today

Fig. 66 Cross section of the plasma-torch components and their electrical connections to the internal, plasma-supporting power supply (bottom left) and the transferred arc power supply (bottom right) between anode and work piece.

peeled off by thermomechanical stresses. Also, due to high stresses in the arc spots, the surface is activated, i.e. the density of dislocations drastically increases that improves the adhesion of the first coating layer. In this configuration a cleaning of the metallic surface from oxides and other impurities before coating application can be attained in the vacuum plasma spray process [74]. It was reported that the previous sputter cleaning by reversed transferred arc can increase the bond strength of the coatings considerably [75]. In order to ascertain cleaning and activation of the entire surface, frequently it is intentionally preoxidized in the preheating chamber in the air or oxygen. Typical secondary currents are in the rage of a few tens of Amps up to a few hundred A. The treatment of locations with reduced accessibility like corners, cavities and so on, is a major challenge of transferred arc (TA) technology – the arc is shunted to the adjacent walls and edges and the problematic location remains untreated. The microcracks initiated at the surface in the arc spots can be another problem. If the work piece is connected to the positive pole (“positive transferred arc”, position 2 in Fig. 66), the current will flow against the plasma gas flow. The arc attachment remains diffused at the surface and does not damage it. The position of the arc root can be well controlled. The work piece is heated by the electron emission energy released during recombination at the surface and electron kinetic energy accumulated in the anode layer. This makes the positive transferred arc an effective means to control the part temperature during the vacuum spray. Overheating and erosion of the torch anode could become a problem if the positive TA is continuously used. A general overview about plasma spraying is given in the book of Heimann “Plasma-Spray Coating – Principles and Applications” [76].

2.7 Production Experience

2.7.2 Process and Systems

The overview of thermal-spraying process parameters (Figs. 4–7) applies to every technology. During production, defined process parameters have to be identified and controlled within the limits given by the operational window (see Section 3.3.3). Figures 67–69 show typical parameter sets as examples for APS, LPPS and HVOF for the spraying of different powders used in gas turbine applications.

Fig. 67 Typical APS parameter set for ZrO2 coatings.

91

92

2 Practical Experience Today

Fig. 68 Typical LPPS parameter set for CoNiCrAl coatings.

2.7 Production Experience

Fig. 69 Typical HVOF parameter set for Cr3C2-NiCr wear-resistant coatings.

Important parameters have to be established according to the respective coating specifications: A hard face coating (Fig. 69) – very often NiCr–Cr2C3 is used – requires Vickers hardness values higher than 1000 and a tight thickness tolerance, because it is applied in areas where parts are assembled. A hot corrosion or oxidation resistant coating (Fig. 68) must show a low porosity or (even better) a porosity close to zero. A thermal-barrier coating (Fig. 67) has to have a low thermal conductivity of less than 2 W/mK and a defined high porosity in the range of 10–20 Vol-% or microcracking giving advantage in thermal cycling.

93

94

2 Practical Experience Today

During process development these properties have to be established. The development is finished when the process yield approaches 100% in serial production, i.e. zero defect production or when the process achieves a capability expressed in a high sigma > 4. 2.7.2.1

The Programming of the Coating Process

There are different methods to create a coating program for a manipulation system. The oldest one is moving the manipulator manually in defined positions and saving the coordinates. It is helpful to use a laser (point or pattern) attached to the gun to simulate the coating jet and to show us the position of the points. This method functions well with simple geometries, which do not need many shapes for creating a program. Given the fact there are coating objects whose geometrical figure and their required coating technology need programs with several hundreds or even thousands of points, it is not possible to accomplish this work with manual teaching without blocking the coating devices for several weeks thus making any other production processes impossible. The role of movement is often underestimated. Figure 70 shows the levels of complexity. The simplest method of coating requires a line-like movement of the spraying point on the surface of the substrate. The lines are parallel and the movement speed is constant. This is easy to reach if the objects to be coated are rotating and cylindrical or it is a nonrotating flat or simply shaped surface. In nonrotative (called translative) spraying, the difficulty increases with the growing complexity of the object. An airfoil of a turbine with two platforms, for example, is a big challenge. Offline programming is the magic word for many coating manufacturers.

Fig. 70 Two classes of thermal-spraying complexity.

2.7 Production Experience

Offline programming can significantly reduce the use of production facilities for process development of new coated components. The machine has to be used during process development only to spray test components and can immediately return to production after test spraying. There are several programming software packages available, which help the user to solve this time-consuming task [77, 78]. Many of these are still under development and have not yet reached an industrial level. The reason for this is that teamwork between programmers and those performing the coating is lacking. Adequate training with regards to optimizing process development, with and without the use of offline tools, is also missing. The software should be able to generate movement directions which would be understood by every standard robot or CNC controller. The optimum target of these tools is combined programming and coating simulation software. The first steps are done. One of the forerunners in this field was ALSTOM, whose coating department has developed such a system. We will discuss the status of offline programming in Chapter 5. Hopefully the coating process and equipment suppliers will in the future offer such systems as additions to their technologies. 2.7.3 Finishing

Every coating specification contains requirements for final surface quality. Polishing or tumbling might be necessary. A heat treatment might be required in order to reduce porosity or improve coating adherence. Unwanted overspray must be removed from machined surfaces; holes and slots have to be protected in advance and/or afterwards cleaned. 2.7.4 Repair of Turbine Parts

Thermally sprayed coatings do not last forever and most components of gas turbines have a limited or expected lifetime in the turbine. Specially selected and optimized thermal-spray coatings do extend the lifespan of the base materials due to protection from wear loss or thermal load. But, after a period of severe service, thermally sprayed coatings may be worn out or fail by spallation from grown cracks. Some of the damages will be obviously seen by regular examination, whereas more hidden failures have to be examined by nondestructive examination (NDE) techniques like fluorescent penetrant, eddy current, thermal imaging or ultrasonic inspection. After a few thousand operational hours owners and operators are therefore faced with the questions regarding component replacement and/or repair. Thermal machines, like land-based or aircraft turbines are heavily loaded systems.

95

96

2 Practical Experience Today

Nevertheless, overall efficiencies are expected to increase, which is mainly achieved by increasing the input temperatures from the combustion chamber to the first vanes in such a way that Ni-based unidirectionally solidified alloys have to be applied as blade materials and even Ni-based single-crystal alloys have to be considered. Oxidation resistance and thermal shielding are therefore the main purposes of the thermal-spray coatings on turbine parts. Bond coats prevent the base material from oxidation by formation of thermally grown oxide layers in the bond/coat interface regime. Typical bond-coating alloys contain CoNiCr and aluminum and yttrium, as oxygen-attracting elements. On top of them the yttria-stabilized zirconia (YSZ) thermal-barrier coatings (TBCs) with low thermal conductivities prevent the underlying components from the high heat loads of the surface. They are mainly used in gas turbines in the region of the transition ducts, burners, rotating blades, stationary vanes, and chamber tiles. The lifetime of such TBCs is limited by the sintering and progressive destabilization of the zirconia matrix. During service the coatings are subjected to high mechanical and thermal stresses, as well as the formation of thermally grown oxides (TGO) at the ceramic/bond coat interface. As these oxides are brittle and tend to crack, especially during heating up and cooling, this effect may be used intentionally for coating removal. Such diffusion layers between the coating and the base material should usually be removed during reconditioning and before recoating. Coating removal and restoration are advantageous and economically reasonable, if the base material is very costly and still in good shape. With maintenance and coating repair the minimum acceptance standard can be reached even after excessive use, e.g. for cooled blades and vanes (see Fig. 71).

Fig. 71 Typical maintenance approach for cooled turbine blades and vanes based on EOH7)

7) EOH (Equivalent Operating Hours) – duration of turbine operation with correction factors for starts, loading and fuel conditions.

2.7 Production Experience

Fig. 72 Turbine component life flowchart.

Critical design and life-enhancing features have to be checked and rechecked before recoating. The flowchart (Fig. 72) shows the gas turbine component life diagram and the most influencing key factors. 2.7.4.1

Coating Removal, Stripping

The removing of the coating can be done using different methods depending on the chemical composition of the coating and base material or the difference in their physical properties. Special attention has to be paid not to destroy the base material, when the coating only should be removed. Different physical and chemical routes do exist to remove effectively a worn or defective coating. It may be necessary to remove the whole coating or – especially when applying coatings on new component – only to do local repair. In principle the removing or stripping techniques can be divided into the following groups: x Mechanical methods (grit blasting, water-jet erosion, grinding, turning, etc.). x Thermal methods (by extreme cooling or heating and stress cracking due to the difference in thermal expansion coefficients between coating and base material). x Chemical and electrochemical methods (dissolving of the coating without destroying the base component). For brittle TBC ceramic top layers the bombardment of ice pellets or high-pressure water jet stripping showed good results concerning removal efficiency [79–84]. Grit blasting acts at different angles in different ways, i.e. more abrasively for flat impact angles and more erosive at steep impact angles. It is widely and effectively used for manual coating removal on complex shaped parts, where programming of a robot would be complex and/or too expensive for treating a few parts.

97

98

2 Practical Experience Today

Other mechanical methods are turning and grinding, which are often used in mechanical shops to remove coatings from cylindrical components. Other coatings like bond coats can be removed, e.g. by wet-chemical methods. Scrivani et al. [85] described the mechanisms and performances of chemical stripping for the removal of NiCrAlY thermal-spray coatings on turbine blades and vanes. Hard metal coatings like WC-Co or WC-10Co4Cr are typically applied by HVOF spraying. An attempt at stripping hard metal coatings by an electrolytic process using different chemical solutions was done by Menini et al. [86]. He investigated five different stripping solutions namely 1. NaOH (75 g/l) (Cr stripper) 2. NaOH (52 g/l) + Na carbonate (68 g/l) (Cr stripper) 3. Meta-nitrobenzaene sulfonate (60 g/l) + NaOH + Na cyanide (90 g/l) (Ni-stripper) 4. Na carbonate + Na tartrate (Rochelle salts) 5. Commercial Ni stripper without damaging the three different base materials (4340, 300M, Aermet100) The highest stripping rate (SR) to remove WC-10Co4Cr coatings of SR = 152 μm h–1 was observed for solution 3) on flat plates of dimension of 2.54 × 10.16 × 0.64 mm3 without attacking the three above-mentioned base materials. In general, chemical methods can be quite effective to remove coatings from components within a few seconds at high temperatures up to some days at room temperature. Nevertheless they leave expensive residues, which have to be dumped. After the chemical stripping there might be certain problems with the transferred arc (TA) cleaning in the case of the LPPS process – the surface is too clean without any “natural” oxides. 2.7.4.2

Restoration of the Base Materials

The restoration of the base materials for eliminating the missing dimensions of the component can either be done by local alloying [87–89], laser methods [90] or thermal spraying [91]. In all cases the restored components have probably to be remachined to fulfill the dimensional conditions. Afterwards, the new coatings can be applied on the prepared surfaces, like on new parts. 2.7.4.3

Refurbishing, Recoating

After the dimensional restoration and surface preparation, the new coatings can be reapplied. If not all of the old coatings are removed, a smooth transition between the old and the new coatings should be manufactured. This insures a better adaptation of the possibly imposed stresses by the removal methods and therefore a better expected lifetime of the intersection.

2.8 Commercial

One of the problems that can occur due to repair may be the reduced lifetime of the base components by fatigue after coating removal and grit blasting. There may be some potential material solutions against that, which at least seemed to have improved the fatigue strength of substrate of CMSX-4 after applying CoNiCrAlY, or HaynessC22 coatings, respectively [92]. In any case, refurbished components have to pass a requalification before being released for further applications [93]. It is mostly accepted that the quality requirements for recoated components may be reduced sometimes. The reason for this is that frequently the reconditioned component is supposed to retire after the next service interval and the function of the coating is to assure its survival only but not the next reconditioning – this is the major difference from the coating on unused parts.

2.8 Commercial

The next four paragraphs will cover the commercial aspects. 2.8.1 General

Every coating process is a sequence of operations and quality-assurance steps: x Surface preparation x Coating x Finishing Every step can be charged with a value in €. This value depends on the actual piece, the utilization of the coating shop, the specific equipment, personnel, overhead, deposition efficiency, consumables cost, tooling, etc. Even if Fig. 47 gives only an indication of involved times it can be used to elaborate the cost drivers: In value-added steps the longest times are the coating cycle and the heat treatment. In both cases expensive equipment is involved. Therefore this equipment never should be idle and the manipulation times for the part and the torch have to be optimized. In quality assurance the laboratory investigation in the metallography requires 2 days in the example. Even if the laboratory cost is medium, the influence on WIP waiting in the shop on delivery is important. On the other hand, metallography is a destructive method that has to be minimized by a thorough process control. A further important factor is the yield of a process. In thermal spraying for powerplant components 100% yield is expected. Sometimes 100% yield is required because rework is not possible (for example, film-cooled turbine blades).

99

100

2 Practical Experience Today

Fig. 73 Cost of coating depending on coating process.

Therefore, a statement cannot be made saying this coating process is less costly than another one. This statement can only be made on a comparable basis that is, e.g. new equipment, identical labor cost, identical raw material cost. Therefore it is more reliable for comparison to look at coating times, deposition efficiency, equipment time, personnel cost. The following example based on Fig. 73 shows the influence of part geometry on cost. The difference in cost is mainly due to different deposition efficiencies of each scenario. The conclusion on the commercial side is that coating can be rated qualitatively according to production costs (including amortization) in increasing magnitude as: x x x x

Wire, flame spraying HVOF APS LPPS

However, technical requirements like restrictions in porosity levels might cause a process to be obsolete or certain processes cannot melt the ingoing material properly. In addition, component sizes, spraying times and consumables, like gases, might change the sequence. The following comparison between LPPS and HVOF shows the complexity of the situation (Tables 20 and 21).

2.8 Commercial Table 20 Economies of MCrAlY coatings produced by LPPS and HVOF (part 1) (courtesy of Sulzer Metco). LPPS Low Pressure Plasma Spraying

HVOF High Velocity Oxy-Fuel Spraying

Features Pros

x x x x

Cons

x High initial investment x Maintenance of system

Oxide free coatings Fast processing time High quality yield Preheat of parts

x No electric power x Lower spray system cost x No vacuum x x x x

High flow dust removal High demand on motion device Special powders required Longer process time

Spray pattern x Large

x Small

Spray capacity x High with 2 stings

x Low because of spraying time

The conclusion is that operating costs and component sizes are the most important factors influencing the economics for making decisions. As stated before, the costs of a thermal-spraying process can be divided into: x Value-added costs x Nonvalue-added costs The task in process development and production is to find the sequence that minimizes nonvalue-added activities, e.g. overspray removal, opening of holes that are closed by the coating. In addition, the position of the coating process in the manufacturing chain has a big influence on cost. Let us look at 2 sequences of a cast part: Sequence 1: Casting o Machining o Coating Sequence 2: Casting o Coating o Machining Obviously, sequence 1 requires masking of machined surfaces, whereas sequence 2 requires handling of coated components. Masking is a nonvalue-added operation. Sequence 1 has higher cost and lower risk. Sequence 2 takes out masking, disassembly of masks and overspray removal. This is the ideal process from a cost point of view if handling in machining and special fixtures can be found that do not increase the risk of destroying the coating. A high-risk process will never be stable and will not end up with 100% yield. We will discuss now the cost of the coating steps.

101

2 Practical Experience Today Table 21 Economies of MCrAlY coatings produced by LPPS and HVOF (part 2) (courtesy of Sulzer Metco). Cost comparison LPPS vs. HVOF Scenario x x x x

Standard work year 3120 hours (2 shifts, 5 days, 75% availability) Amortization 3 years Gas turbine part (14 kg) Similar handling methods

Capital cost

Operating cost/part 2500000,00

180,00 160,00

Annual depr. Cost Investment

Operating cost/part HVOF Operating cost/part LPPS

2000000,00

140,00 120,00

Cost($)

Cost($)

102

100,00 80,00

1500000,00

1000000,00

60,00 40,00

500000,00

20,00 0,00

Wear Operator Gas parts cost cost

Power Powder Total cost cost

0,00

HVOF

LPPS

Process

Cost Item

Total cost/part 200,00 180,00 160,00 140,00 120,00

Cost ($) or %

100,00

HVOF LPPS

80,00 60,00 40,00 20,00

LPPS

0,00

Total cost/part ($)

HVOF

Operating cost % Capital cost/part %

Cost item($) or %

Conclusion x x x x

Spray time is key factor Spray time depends on size and shape of spray pattern Capital costs are less important than operating costs Total costs depend on number of parts to be coated

Process

2.8 Commercial

2.8.2 Surface Preparation

Surface preparation uses low-cost equipment, and is labor intensive. Nevertheless, it must be 100% successful. Apparently, it is considered know-how to judge if a surface is clean, has the right roughness and no particle inclusions. Surfaces produced by earlier operations might contain inclusions. Therefore a good communication within the manufacturing chain is required in order to achieve 100% yield and to avoid rework costs. 2.8.3 Coating Equipment

The most expensive equipment in the coating sequence is the thermal-spraying equipment including all handling and tooling devices. We have to look at thermal spraying as just another manufacturing process like machining, casting, forging. Therefore thermal spraying has to target towards: x x x x x x x x

High productivity Flexibility Use single-piece flow instead of batch processes Short changeover times, quickly changeover tools Correct sizing of machines Value creating activities for specific products Interest in value flow Value defined by customer

Cost of thermal spraying is determined by: x x x x x x x x x x x x

Coating process Spraying times Spray efficiency Maintenance Consumables, powder, gases, cooling Power (electricity, gases) People Overtime Planning Schedule Investment and infrastructure cost (calculation) Location (hourly rate, productivity)

It is important to look carefully at the utilization of the equipment. Value added time is only during coating. The following Figs. 74 and 75 compare cycle times of batch and continous processes. It can be clearly seen that a continuous process provides better utilization of the spraying equipment.

103

104

2 Practical Experience Today

Fig. 74 Cycle time batch process VPS showing typical operation pressure levels and the required time (courtesy of Sulzer Metco).

Fig. 75 Cycle time of continuous process LPPS.

2.8.4 Finishing

The optimum finishing process from the cost point of view is “no finishing required”, i.e. use the as-sprayed component. For certain applications a diffusion heat treatment is required in order to improve coating bonding. Sometimes the as-coated surface roughness is not sufficiently low and needs post-grinding [94]. These operations are technically necessary steps and need e.g. in case of hard metal coatings the use of diamond tools. Other operations, like removal of coatings that are too thick, are unnecessary, add costs and must be avoided.

105

3 Quality and Process Capability 3.1 Quality Assurance

Good quality in terms of coating means that the requirements of the specification are reached. Specified are, e.g., bonding defects at the interface, content of oxides and blasting material on interface, oxides and cracks in the coating, unmolten particles, coating thickness and hardness, gas permeability, porosity and several other factors that are relevant for the coated component. Coating in power generation requires the passing of a thermal shock test of TBC coatings. Due to the increasing quota of abradables there are abrasion tests to be accomplished, too. Some specifications set a limit from both sides, some only in terms of a maximal or minimal value. Variations in the coating process make it difficult for us to match these specifications. Because it is virtually impossible to eliminate these variations we are forced to learn about them and try to minimize their influence on the coating process. Quality assurance addresses several aspects: 1. 2. 3. 4. 5. 6. 7.

Quality assurance of supplied material (powder, gases) Maintenance of equipment Maintenance of tooling and fixtures Repeated training, testing of personnel Process control Nondestructive testing of coatings Destructive testing of coatings

3.2 Sources of Process Variations

As was mentioned earlier in this book, there exist more than fifty parameters that determine the coating properties. Most of them are not completely stationary during the process. In addition, they can not be controlled with absolute accuracy. These parameters form a field of potential process and product variations. The most important of them are described in Table 22.

106

3 Quality and Process Capability Table 22 Major sources of plasma coating-process variations. System

Element

Sources of process instability

Influence on process

Electrodes

Anode

Wear of the wall

Drift of voltage leads to variation of particle parameters

Cathode

Erosion

Possible coating contamination with eroded material, change of pulsation characteristics o change of particle parameters

Flow-control devices

Unstable gas flow rate

Variation of particle parameters

Gas ring

Improper selection: too high gas swirl and particle dispersion

Unstable coating quality at the periphery of spray pattern (see Section 4.5)

Incoming powder control

Deviation of powder-size distribution, fines in powder, variation of apparent density and flowability

Fluctuations of coating thickness and microstructure, clogging of injectors, spitting

Powder preparation

Humidity of powder, segregation

Oxides in coating, variation of coating thickness

Powder feeder

Powder segregation Nonuniform powder loading

Variation of coating thickness

Powder hoses

Vibrations, loops

Change of particle parameters

Injection block

Plugging and spitting, wear

Change of particle parameters, premature maintenance

Gun movement

Robot, CNC

Too high acceleration, which affects stability of injection, misalignment, backlashes and insufficient stiffness of the gun-holding system

Variation of coating thickness Change of particle parameters

Power supply

Stabilization schemes

Arc pulsations

Variation of particle parameters

Masking and fixtures

Part holder masks

Backlashes, plays, thermal deformations

Variation of coating thickness and transitions

Gassupply system

Powdersupply system

All sources of the coating-process variations can be categorized into one of the following four groups: 1. Special causes. Usually they are attributed to the malfunctioning and maladjustment of the equipment and some external influences. An excellent example of the special cause can be the influence of a steam hammer, which was operated 100 meters from the coating machine at the construction site, on the porosity of TBC on turbine blades. It took some time for production engineers to understand the cause–effect chain, which appeared to be closed

3.2 Sources of Process Variations

via stratification of the powder in the powder feeders due to an excessive level of the floor vibrations. 2. Common causes. Those are purely random phenomena, which are known to be present in the process, but their effect can not be identified by measurements taken from the one specific coated part or from the single production set. A contribution of the common cause factor into the overall process uncertainty in general can be quantified by a series of special tests, which involve programmed variations of the parameters, which affect the phenomenon. Also, a modeling could be helpful in defining the sensitivity coefficients and expected variation ranges. 3. Parameter drifting. This type of process variation is quite specific for the thermal spray since certain variations of the process parameters during coating of a batch of parts can not be avoided. 4. Unstable quality-control procedures may provoke faulty requests for the process adjustments and de facto may cause variations of the product parameters. Let us consider these four types of process variations in more detail. 3.2.1 Special Causes of Coating-Process Variation

It is a commonly accepted point that the special causes of operation-process variations can not be tolerated in a stable manufacturing. They must be identified and eliminated. The most typical special causes of coating variations are (see also Fig. 6): 1. 2. 3. 4. 5. 6. 7. 8.

Fine fraction in powder Plugging of injectors and spitting Gas leakages in the gun and powder hoses Cathode melting and spitting Incorrect fixtures and masking Gas impurity Unstable power supply Unstable cooling-water supply and contamination of cooling water

In the manufacturing process under statistical quality control the presence of special cases can be detected by the appearance of statistically unlikely events, e.g. the quality parameters beyond the allowed control limits or systematic trends within the limits, which can not be attributed to chance (formal approaches could be found, for instance, in [95]). The only way to identify these special causes is a cause–effect analysis if there are instabilities in coating quality. A risk analysis is a preferred method to identify special causes. The mitigation plans must be part of the maintenance strategy and personnel training strategy. The appearance of special cases can be prevented by regular maintenance, which includes: x Preventive replacement of the consumable/worn parts x Calibration and adjustment of the control devices, fixtures and masking

107

108

3 Quality and Process Capability

x Periodic usage of additional diagnostic tools, which usually are not part of a regular process control In more detail the maintenance strategies are described in Section 3.4. 3.2.2 Stochastic Nature of a Spray Process

An intrinsic scatter of the process parameters and of the characteristics of the resulting coatings is a distinctive feature of thermal spray technology. This is caused by the enormous quantity and variety of particles participating in the process as well as by the complexity of physical processes that determine the hardware operation. Therefore, there exists a certain intrinsic limitation for the achievable process repeatability, which must be understood and taken into account in manufacturing industry. Further in this section we will consider the major sources of thermal-spray stochastic behavior. 3.2.2.1

Arc and Jet Pulsations

Motion of the plasma arc attachments has essentially a stochastic nature. Despite the fact that significant efforts have been made to understand this phenomenon in recent decades, it still can not be predicted with the accuracy required for practical purposes. It is known that due to the stochastic jerks of the anode arc attachment, the plasma voltage oscillates with a frequency of 3–10 kHz [96]. The magnitude of these oscillations can reach as much as 40 per cent of the average gun voltage [97]. Also important is that the arc never occupies the entire nozzle cross section. Therefore, beside strong temporal variations plasma parameters in the nozzle exhibit a significant circumferential nonuniformity. A swirl, which increases when the flow passes through the converging part of the anode due to the conservation of angular momentum, is a source of instability and of the occurrence of the radial flows in the nozzle. All this leads to the strong temporal variations of the axial, radial and swirling components of the forces acting on the particles both inside the nozzle (in case of internal injection) and in the free jet. Respectively, particle trajectories are randomly scattered in the plasma plume. This scatter strongly depends on the swirl intensity, which in its turn is determined by the design of the gas ring of the gun. For instance, a change of the tangential gas injection angle at the gas ring from 90 to 45 degrees could lead to the reduction of the LPPS spray spot size from 50–70 to 30–40 mm at a spray distance of 300 mm at comparable gas flow rates. A dispersion of particle trajectories results in the corresponding variations of the particle impact temperatures and velocities. It is worth noting that a typical particle dwell time in the plasma (1–2 ms) is larger but still comparable to the period of pulsations (0.1–0.3 ms) and to the period of swirl in the anode channel (0.1–0.2 ms). This means that depending on the moment of injection, the number of plasma pulses that a particle experiences could be different. This results in additional dispersion of the particle parameters.

3.2 Sources of Process Variations

Particles flying close to the jet boundary and/or crossing it are extremely sensitive to the jet oscillations. Gas parameters around such particles can oscillate from the cold ambient up to the conditions close to the plasma core. Stochastic sensitivity of the process to the pulsation characteristics of the plasma torch explains certain effects that are known to the coating developers: x Deposition efficiency and coating properties can be sensitive to the characteristics of power supply. Authors experienced a situation, when replacement of the APS power source to the one with formally identical characteristics, but produced by a different supplier, resulted in an increased TBC deposition efficiency by more than 10 per cent and coating porosity dramatically reduced at the same process settings. In the end, the whole process had to be requalified. x Coating quality can react to oscillations of the shop electrical grid (which is an extremely rare situation in the research labs but quite typical for large industrial sites). x Coating parameters may be sensitive to the torch-cooling conditions, which affect the arc motion, e.g. purity (i.e. electrical conductivity) and temperature of the cooling water. x Minor changes of the chemical composition of the anode material and variations of the anode manufacturing process may also result in different arc dynamics. x Electrical buses may influence the electrical circuit as well and, therefore, stability and mean power delivered to the torch. x Air in the gas could cause oxide films in coatings due to plume pulsations [98–106]. Stability of the arc essentially depends on the process parameters such as the torch current, plasma gas composition and flow rate. Arcs with the higher hydrogen content exhibit higher level of pulsations. Usually, the arc becomes more stable with the increase of gas flow rate and current and vice versa, reduction of the latter parameters below a certain level, which depends on the torch design and on the stabilization schemes implemented in the power source, makes a process unstable and the coating properties irreproducible. Estimations and computer simulations show that arc pulsations and the flow oscillations caused by them are the major contributors to the dispersion of the particles and formation of the shape of the spray pattern. 3.2.2.2

Powder-Size Distribution

Different fractions of powder follow different trajectories in the plasma. Respectively, variations of the powder-size distribution, for instance, due to segregation in the feeder or due to quality variation between different powder batches, produce variations of the spray patters and, respectively, a scatter of the coating physical properties and thickness. This effect may be enhanced by external radial powder injection ports, where a “wind sieving” and separation of particles by size/mass takes place.

109

110

3 Quality and Process Capability

3.2.2.3

Powder Injection

Specific features of the powder injection in thermal spray were already discussed in Section 2.5.1.3. Particle injection conditions are affected by numerous factors such as gun motion, powder humidity, vibrations, injector plugging, etc., which are not fully controlled during the coating application and by acting together make the powder injection stochastic. 3.2.2.4

Powder Shape

The initial shape of the powder particles is important because, first of all, it affects the powder mixing, the transport and injection properties. In most cases spherical particles are the best option for thermal spraying. A nonspheroidized particle keeps its irregular shape until it reaches the melting point. A flow around an asymmetrical particle creates not only the drag force but also a lift, which makes the particle drift across the flow. If the particles stay solid along a significant part of their trajectories, an aerodynamic dispersion could noticeably contribute to the size and shape of a spray pattern. Depending on the particle shape, a magnitude of the lift could reach 5–10 % of the drag force that therefore results in the radial particle dispersion of 5–10 % of the spray distance. 3.2.2.5

Particle Bonding

A brief analysis of bonding mechanisms can be found in Chapter 4. The probability of a particle sticking at the surface depends on many factors. Particle velocity and temperature, surface roughness and temperature, material properties and a spray angle are the major ones. In the case of LPPS spraying this probability could be close to 100 % in the hot center of the spray spot. It drops down to 10–20 % in the peripheral areas of the spray pattern that makes those areas extremely sensitive to process fluctuations. If the coating process is stable the shape and properties of spay-spot periphery may be reproducible in the spray trials, for instance at the test plates, but they still remain rather unpredictable (at least quantitatively) when the real 3D component is sprayed in a serial production. The same is applied to the spray of such unstable structures like, for instance, a highly porous TBC. A bonding probability of yttria-stabilized ZrO2 particles in APS TBC process, which is directly reflected in a deposition efficiency, usually does not exceed 30–40 %. Actually, the coating is created at the limit of particle bonding. In such a situation, sensitivity of the bonding process is a factor that significantly amplifies the process uncertainty. 3.2.2.6

Gun and Component Motion and Positioning

Due to misalignment, backlashes and plays in the gun holding system and component fixation as well as their thermal deformations, their mutual positions can be never defined with absolute accuracy and should also be considered to a great extent random. The magnitude of scatter caused by this effect depends on the hardware design, gun trajectory as well as the gun traverse speed and component rotation. Thermal deformation of the component masking is another factor that is difficult, if not impossible, to predict for the real coating cell.

3.2 Sources of Process Variations

Typically, the influences of, e.g., flow rates of powder feed gas are much more influencing the position of the spray spot, than the reproducibility of the robot system itself. Therefore, the requirements on positioning and reproducibility of the handling systems for thermal spraying are much less restricted than for other conventional machine tools. The acceptance criteria for handling systems can be consulted in the standard EN 1395-6 [107]. 3.2.3 Drifting

As was already discussed in previous chapters, certain variations of the process parameters during the coating of the batch of parts or even of the single part are inevitable in thermal spraying. Let us outline the most typical sources of drifting. x Electrode erosion. A typical appearance of the worm cathode is shown in Fig. 76. Decrease of the plasma gun voltage due to the wear of electrodes is the most pronounced drifting phenomenon and the most important one. Some experimental results related to this effect could be found in [35, 36, 37, 108]. Voltage drift (decrease) results in the reduction of the plasma net power. Respectively, particle temperature and velocities also decrease. Usually, it leads to the reduction of the deposition efficiency and changes in the coating microstructure, for instance, the level of porosity could significantly increase in the case of APS-sprayed TBC. What is also important, the anode wear could lead to a significant increase of the amplitude of the arc pulsations and change their frequency. Thus, it could essentially broaden the stochastic distribution of particle parameters.

Fig. 76 Wear of electrodes: (a) tip of a new electrode in contrast to worn electrode (b) side view; (c): top view.

111

112

3 Quality and Process Capability

x Variation of the substrate temperature. Depending on the component size and the gun power, a surface temperature could change during the coating application. For instance, a part that is sprayed by APS or HVOF usually becomes hotter by the end of the process, whereas the component that is preheated by transferred arc in LPPS loses its temperature during the spray, making the coating properties varying between the different coating layers. Also due to this effect, the same set of spray parameters could result in different coating quality on the parts of different size. x Drift of the plasma and carrier gas flow rates due to de-adjustment and wear of flow rate control devices or drifts in temperature of the cooling water. Generally speaking, there are other factors that are suspected to result in a gradual change of process parameters, but their influence can hardly be quantified. These are, for instance, contamination of the gun cooling water, daily variation of the ambient temperature and others, which, in principle could affect an operation of the coating hardware. 3.2.4 Stability of the Quality Control

Importance of the accuracy and reproducibility of the quality measurements for the overall process capability cannot be overstated. In the ideal case, any influence of the measuring tolerances on the overall uncertainty of the measured characteristic should be minimized or even eliminated. In the case of thermal-spray coating this cannot always be achieved. Let us consider two examples. Control of coating thickness. Today, there are many methods available for measuring the thickness of a coating. The international standard ISO 3882 [109] gives an overview on the different methods applicable. Usually, the coating thickness is measured destructively by cutting the control piece [110] or nondestructively with the use of eddy-current sensors [111]. The latter are very efficient for the electrically nonconductive coatings (e.g. ceramics) and with some difficulties could be used for measurements of metallic ones. In the case of destructive measurements the following two factors are the major contributors to the measuring error: 1. Roughness and texture of the coating and substrate surfaces. The magnitude of the roughness (peak to valley) of metallic coatings could be as high as 30–70 Pm, depending on the powder size and process parameters (and roughness specification of course, if applicable), whereas for the ceramic highly porous TBC it could achieve 100–150 Pm. The long-wave (200–400 Pm) variations (waviness: see EN ISO 4287) of the coating thickness could be as high as 100–200 Pm [70]. The decision as to which points are to be taken as the references remains either with the lab specialist or is generated by the special image-processing software. Therefore, the result could significantly depend on the human factor or on the software settings.

3.2 Sources of Process Variations

Fig. 77 Simulated coating thickness depending on the position of virtual cuts at the transition from blade airfoil to the platform; cross section (CS) separation 3 mm (courtesy of ALSTOM).

2. For the destructive tests the component is cut using a special template. An accuracy of the cut positioning is 1–2 mm if the template fixations are done properly. After cutting the sample, it is ground and polished and this procedure could consume another 2–3 mm of the sample bulk. If after polishing the sample surface still has some defects, like scratches or cavities, it is repolished or even reground. Thus, as a result of the preparation procedure, the actually measured location could be 2–4 or even 5–8 mm displaced from the one that is prescribed by the coating specification. This would not be critical for such locations as airfoils (if the cuts are done along the chord). But at the radii and transitions with the sharp variations of the thickness, the difference could be dramatic, as is shown in Fig. 77 generated by the coating simulation of a real turbine blade. It is obvious that the locations that are difficult to coat are usually also difficult to measure and the level of uncertainties described above also increases there. There exists a general dilemma in definition of the quality-control positions. The designers want, of course, the most critical locations (also from the standpoint of coating quality) to be controlled. At the same time, additional process uncertainty could be introduced. One of the compromises could be to try to make the cut perpendicular to the direction of the least thickness variation to make the measurement more robust. Porosity measurements. For the destructive porosity measurements, a test component is cut. The metallic samples are polished and the fraction of the voids is measured by optical microscopy with the use of image analysis. In the case of brittle ceramic coatings prior to polishing the sample is impregnated with the epoxy. Experience showed that the resulting porosity figures significantly depend on the polishing procedures and the type of epoxy used [112]. The influence of the

113

114

3 Quality and Process Capability

Fig. 78 Influence of 8 different embedding media and steps of preparation (10 μm, 1 μm, OPU = SiO2 suspension) on the interpretation of the obvious porosity of the same TBC sample (courtesy of ASM) [112].

viscosity of the different epoxies and the degree of infiltration can be visually seen from the different gray levels in Fig. 78. Three different stages of the preparation are shown for each embedding media for polishing steps ending with diamond of 10 μm, 1 μm and the final stage (suspension of amorphous SiO2, pH ~10). The apparent coating qualities, as seen in Fig. 78, show totally different amount of pores and coating/substrate interfaces depending on the different type of epoxies. Obvious large breakouts as well as apparent bad bonding of the coating to the substrates can easily lead to wrong interpretation of the real structure. Also, it has been found that the image software settings significantly affect the result of analysis. It is quite obvious that conclusions that are based on such a scattered quality control could result in the underestimating of the actual process capability or even worse – in the certain adjustments of the process, which would result in the uncontrolled deviation of the coatings from the specified quality. In order to avoid or, at least minimize, the influence of such factors on the manufacturing the quality-control process should be treated as part of the manufacturing, which means:

3.3 Process Capability and Stable Process

1. Equipment used is regularly checked and calibrated. 2. Sample-preparation procedures, hardware, instrumentation and software settings are explicitly defined and documented. 3. Special templates for sample fixation, cutting, polishing, etc. should be used. 4. Personnel should have appropriate training. 5. Quality control of incoming consumables, like, for instance, a polishing powder, should be in place. 6. The capability of the quality-control procedures should be quantified based on special statistical tests, for instance, comparative analyses preformed by different specialists and/or at different preparation machines, “blind” tests, etc. 7. Definition of the controlled locations should be done in a way that results in more robust measurements.

3.3 Process Capability and Stable Process 3.3.1 Definition of Process Capability

Generally speaking, a process capability is usually understood as a capability to manufacture a product with the quality, quantity and costs that would be satisfactory for the customer, manufacturer and controlling authorities (if any are involved). As far as statistical definition of the process capabilities is concerned, it is determined by the correspondence between the actual statistical distribution of the measured product characteristics and the specified allowable ranges of those characteristics. There are several formal definitions of the process capability [113], which involve all or some of the following four process values: x Lower and upper specification limits (LSL and USL) of the characteristic, which is defined as a quality-control parameter. x Standard deviation V and the average value X of the parameter, which are calculated based on the actual manufacturing results. A Cpk criterion of process capability is one of the most commonly used. It is defined as a distance between an average value of the coating quality parameter and the closest specification limit related to its tripled standard deviation, i.e.: CPK =

MIN ( X − LSL, USL − X ) 3V

Graphically, the definition of Cpk is shown in Fig. 79. This parameter considers both a process scatter and a shift of the process average with respect to the specified quality band. It implicitly defines the probability of the quality parameter considered to fall within the specification band. For instance, in the case of centered distribution Cpk of 1 corresponds to the approach of “three sigma”. It is

115

116

3 Quality and Process Capability

Fig. 79 Definition of process capability.

quite obvious that there exist two basic ways to improve the process capability: to reduce the process scatter V and position the process average in the center of the quality band. According to [113] the process is being considered under the statistical control and, therefore, can be described in capability terms only when all the special causes of variations have been eliminated and only the common causes are active. This cannot be fully applied to the thermal-spray process due to such phenomena as drifting of parameters which is typical for the majority of plasma torches used in production (see Section 2.4.4). In essence, drifting is a special case that cannot be eliminated. In such a situation a process capability depends not only on the stability of hardware, quality of consumables and on intrinsic process stochastics, but also on the process duration and number of parts in the batch to be sprayed. The influence of the parameter drifting on the process capability is illustrated by Fig. 80. Let us assume that due to allowable drifting of the control parameters the resulting coating property shifts by da. So does a statistical quality distribution. The initial (curve 1) and final (2) instant positions of the distribution functions are shown in Fig. 80. Distribution 3 is a time-averaged distribution of the quality parameter. The average distribution function is wider than the original ones. The first reason for this is the superposition of shifted distributions. The second one may result from the increased process stochastic variation as in the case of the gun voltage drifting (see Fig. 38). Apart from the process scatter, drifting shifts the distribution function and may bring it closer to the USL or LSL. Thus, in order to minimize the effect of drifting on the process capability, it does make sense to position the initial process target asymmetrically with respect to the band center, i.e. to shift it by half of the allowed drift in the opposite direction. Another practical point: if because of cost reasons a larger drift is to be allowed, a corresponding reduction of the process scatter from all other sources has to be achieved in order to maintain the overall statistical process quality. And last, but not the least observation: with the drifting phenomena taken into account the “instant” and “average” process capability criteria can be differentiated. What is important is that the reduced “instant” Cpk does not necessarily mean an unacceptable average value (if, of course, the process drifts in a right direction).

3.3 Process Capability and Stable Process

Fig. 80 Effect of drifting on the process capability.

An important remark should be made with regard to long-term and short-term process capabilities. There may exist sources of variations that are not reflected by the capability studies, which usually have limited timeframes. It is a common practice in industry to apply higher process capability requirements (for instance, it is required to have Cpk 1.33 for the “3-sigma” process instead of 1.00) to account for long-term drifts and deviations. Another way is to assume a shifted process average while calculating Cpk. Usually this shift is ±1.5 sigma (in this sense the qualified “6-sigma” process would be considered a “4.5 sigma” process in a longterm prospective). As discussed above, in a thermal spray process a certain drift has to be tolerated even in the short term. Therefore, it must be reflected by the usual capability studies. On the other hand, process drift leads to reduced maintenance intervals, which reduces the impact of long-term factors on the production quality and the corresponding corrective shift may be set lower. 3.3.2 Definition of a Stable Coating Process

A process-capability analysis is a statistical quality-control tool, which can be applied to the stable process only. What does the term “stable process” mean with regard to coating manufacturing? A stable coating process should meet three groups of requirements. Technical requirements: 1. Reasonable variations of the process parameters should not dramatically impair the coating properties. This requirement is critical, since, as was already discussed, process variations constitute an intrinsic part of the thermal spraying. 2. The process should be based on the scheduled maintenance procedures; unplanned hardware checks, adjustments and part replacements as well as emergency troubleshooting should be exceptional events. It should be understood that any unplanned interventions into the stochastic and drifting

117

118

3 Quality and Process Capability

process are, in essence, additional sources of process variation. For instance, if the initial targeted value of the coating parameters is shifted against the center of the quality band to compensate for a drift, a premature readjustment would lead to the reduction of the process capability instead of improving it! Another point is that the process “stabilization” by means of excessively frequent replacement of worn parts could suppress the effect of drifting and create a false impression of stability, which may disappear as soon as the maintenance schedule is normalized. Quality requirements: 1. The process must assure a specified coating quality. 2. The amount of rejected parts should not exceed the predefined limit. 3. All sources of special-cause deviations are known, documented and eliminated. Corresponding online “alarms” and preventing control procedures are active. Economy requirements: 1. Process yield must be assured. 2. Quality assurance should obey the principle: as much as necessary, as little as possible. A principle “quality at any cost” usually does not work in a serial production. 3. The process supervision, maintenance procedures and quality control should be oriented on the personnel with the reasonable level of qualification (see also Section 6.1.3). Formally speaking, cost considerations are not the factor that directly affects the coating quality. But in real life the product quality frequently becomes a victim of a suffered process economy. 3.3.3 Operational Window

The term “operational window” (OW) can be defined as a range of input parameter variations that is acceptable from the standpoint of fulfillment of the coating specification and achievement of required process capability. We will differentiate a parameter operational window, which is related to the selected determining process parameter, and a process operational window, which is a multidimensional set of all relevant parameter operational windows. Formally speaking, all determining parameters that have been mentioned in Table 22, contribute to the process operational window, but only for a limited number of them the parameter operational windows can be defined. A typical list of such parameters looks as follows: x x x x x

Gun current or voltage Main gases flow rates Carrier gas flow rate Standoff distance Powder properties

3.3 Process Capability and Stable Process

Fig. 81 Example of the parameter OW; Sensitivity of the TBC coating weight on a given component to Ar/H2 flow rates (courtesy of ALSTOM).

An operational window is always defined with regard to the measurable coating properties that form the coating quality specification. Further, we will call such properties “quality parameters”. From the practical point of view the operational window should satisfy two requirements: x Sensitivity of the quality parameters to the variation of process parameters is small within the operational window. This allows a manufacturer to run the production for a long time without readjustment of the equipment despite the drifts. x When the process parameters change the coating properties should vary monotonously. This gives the possibility to adjust the process. An example of a two-parameter operational window is presented in Fig. 81. The sensitivity of the weight of the TBC applied by APS process to variations of the flow rates of argon and hydrogen has been studied in ALSTOM’s coating shop. One can see that the range Ar (20–24) × H2 (5.5–6.5) fulfils the criteria of OW. The sensitivity of the process to the variations of powder carrier gas has also been studied and the results are shown in Fig. 82 for two series of tests performed with the time intervals of several months. It is seen that nonmonotonous variations of the coating weight, which reach almost 20% of the average value within the quite small range of the carrier gas flow rate settings do not leave any room for the stable process. In such a situation the same test has been repeated for other powders: Powder 2 was just another cut of the initial powder with an increased lower boundary of size specification and Powder 3 was one of a different size and was provided by a different supplier. It appeared that despite the reduced deposition efficiency compared with Powder 1, Powder 2 provides a much better operational window. Of course, as soon as a different powder is accepted for further development sensitivity checks for other parameters have to be repeated.

119

120

3 Quality and Process Capability

Fig. 82 Example of powder carrier gas OW; (a) initial powder, (b) comparison of different powders (courtesy of ALSTOM).

An example of definition of the operational window for the gun voltage is presented in Fig. 83. A 24-h drifting test has been performed; every hour a test coupon has been sprayed with porous yttria-stabilized zirconia (YSZ) TBC and analyzed. A sensitivity of the deposition efficiency to the mean voltage drift of ca. 25%/V has been derived from the weight data. Based on this, a range of allowable voltage deviation can be defined. When defining the parameter operational window, the parameter control tolerances, possible parameter drifts as well as stochastic variations and drifting of the coating properties due to variations of other parameters should be considered as is schematically shown in Fig. 84. Without the influence of stochastic mechanisms and parameter drifting, the relationship between the process parameter Pi and a coating quality parameter Cj would represent a line. But in reality, this dependence is represented by a “band” with a width that is determined by the coating sensitivity to the stochastic scattering and drifting of all the other relevant process parameters.

3.3 Process Capability and Stable Process

Fig. 83 Sensitivity of APS TBC weight on the voltage drift at constant current.

The parameter operational window could be broadened by the online process control and/or by the a priori shift in settings applied to other parameters. Such a control shifts the correlation band with respect to the parameter operational window in an appropriate direction in order to compensate the parameter drift and keep the coating properties within the specified range. Of course, a stochastic scatter of the parameters cannot be compensated by any control. Let us make several examples of process corrections: x Gradual reduction of the gun voltage and its influence on the net plasma power can be corrected by the corresponding increase of the gun current or by adding H2 to keep the net power constant. Respectively, a voltage operational window under such a control may become wider. Nevertheless, it is still limited. x Influence of certain deviations in the powder characteristics can be minimized by corrections of the power level and plasma and/or carrier gas flow rates. Usually, such technologies constitute know-how of the coating manufacturer. It is important to understand that the definition of the operation window is an iterative process. First, the process parameters are defined at the stage of process development as will be described in more detail in Section 7.3. Secondly, sensitivity tests should be carried out and the operational window, which corresponds to the given set of process parameters (process “recipe”), should be defined. Then the decision is made whether this window is “wide” enough to assure the necessary process capability. If the decision is negative, then another process operational point should be looked at and another iteration is to be made. It is also frequently the case that the operational window is not unique and the different coating supplies could fulfil the same specification and guarantee the required process capability operating within the quite different parameter ranges.

121

122

3 Quality and Process Capability

Fig. 84 Schematic definition of the parameter operational window.

Apart from the process capability there is another important reason why the operational window should not be too “narrow”. It might happen that in the course of turbine development or as a result of analysis of field experience the coating specification may be changed. In such a situation the absence of quality margins could necessitate a completely new process development and requalification. 3.3.4 What Process Capability is Required?

The number of “sigmas” that is required to fulfill the business objectives varies from industry to industry. Twenty- up to thirty years ago the coating on the turbine part was a “nice to have” feature. Failure of the coating meant reduction in the number of service cycles, additional costs and increased scrap rate in reconditioning but not the engine failure. Currently, due to the significantly increased firing level state-of-the-art gas turbines have become more and more socalled coating reliant, i.e. without the oxidation/corrosion protection and/or TBC the turbine part or parts might not survive even until the scheduled replacement. Quality requirements for the gas turbine components are determined by the end

3.3 Process Capability and Stable Process

user, i.e. by the power plant. From the point of view of the turbine owner a good coating is the one that helps to prevent the part failure and to ascertain its suitability for reconditioning after the service interval. The former function is usually the most critical one. The required reliability of the gas turbine should be at least 99.5% (this is a current industrial standard). Keeping in mind that in the gas turbine, apart from the blading, there exist many factors that could cause the unscheduled outage, the probabilistic duration of the outage caused by major turbine blade or vane failure should not exceed 0.1–0.2%. A typical service interval of GT is ca. 3–4 years. Statistically speaking this means that the unscheduled out-of-grid time during the service interval caused by such a failure cannot exceed 1–2 days. In the case of failure or major damage of the turbine part, the outage duration, which is required to organize the emergency stock of parts, to open the machine, to replace damaged components and to return the engine into operation, usually exceeds 30 days. From this it follows that a tolerable probability of unplanned outage during the inspection interval due to damage of blades/vanes cannot exceed ca. 3–5%. Arbitrary attributing half of that to the damage resulting from the coating being out of specification or from its destruction, we can consider the “tolerable” probability of premature outage due to coating poor quality to be lower, i.e. PCF = 2%. It is necessary to emphasize that the fact of deviation from the coating specification or even the premature coating failure do not necessarily mean damage of the part. According to the design practice such events usually mean a violation of the mechanical integrity safety criteria that with certain probability PC–P (usually, of the order of 1%) could lead to failure of the whole part. It is also worth noting that the violation of the specification could lead to the part damage only if it occurs in certain critical surface areas. The probability of the part failure due to violation of coating specification and/or premature coating failure can be estimated as: PCF =

PC PC − P N Acontr Acrit

where: x PCF is the probability of failure of at least one part in the engine due to coating. In the gas turbine service practice the premature outage can result from the destruction of one or several turbine parts or due to the discovery of cracks, deformations or burnt areas during the intermediate hot-gas-path inspections. As discussed above this probability should not exceed 0.02. x PC is the allowed probability of the coating being beyond the specification band in the locations used for the statistical control. x PC-P is the probability of the part damage because of violation of mechanical integrity criteria; let us adopt a value of 0.01. x N is the number of turbine blades/vanes that are coating reliant. In the state-ofthe-art gas turbines these are usually 1st and 2nd stages, which means 200–300 parts per machine.

123

124

3 Quality and Process Capability

x

Acontr is the ratio between the area, which is covered by the regular quality Acrit control in production, and the critical surface area. Usually only 5–10% of the critical location areas are controlled.

The tolerable probability of coating out-of-spec in the controlled areas can be 0.02 ⋅ (0.05 − 0.1) = 3 ⋅ 10 −4 − 10 −3 , which corresponds to the estimated as PC = (200 − 300) ⋅ 0.01 “3–3.5 sigma” process. With the correction for the long-term process variations, which in the case of thermal spray could be between 0 and 1.5 sigma depending on the shop maintenance strategy, the short-term process capability must be not less than 1.2. A value of Cpk = 1.5–1.6 (4.5–5 sigma) should be the target. The estimates shown in this paragraph are rather optimistic and they do not reflect the additional quality requirements resulting from the cost implication of the turbine failure. Frequently, failure of one blade causes destruction of several rows of blades/vanes or even of the entire gas turbine. Also financial implications related to the loss of business and violation of contractual obligations may be extremely high. It may be envisaged that the increasing focus on the cost issues will bring the quality requirements to the coating process up to the level of 5–6 sigma (Cpk of 1.6–2). 3.3.5 Additional Factors that Affect the Process Capability

Among the factors that make the coating process more or less prone to all sorts of variations, the following should be mentioned: 1. Coating material. Different coating materials have different sensitivity to the deviations in the process. 2. Coating specification: – Coating microstructure; for instance, the highly porous YSZ TBC is extremely sensitive to smallest variations of process parameters (see Fig. 83). – Coating thickness distribution. Of course, the more stringent the specification requirements, the lower the process capability that can be expected. 3. Size of the parts to be sprayed: – Increased coating time leads to the increased probability of variations happening during the spray. – Size of the spray spot and the powder feed rate do affect the achievable coating quality and stability. This will be discussed in more detail in Section 4. – Weight of the part does affect the stability and accuracy of part positioning in the fixtures with respect to the spray gun. – Required power of the torch depends on the component size. Size and weight of the spray gun may influence the accuracy of its positioning with respect to the sprayed part. – Temperature regime is more difficult to maintain with big components.

3.3 Process Capability and Stable Process

4. Complexity of the part – elements that are difficult to access are usually less stable from the standpoint of coating quality. 5. Position of the coating process in the manufacturing chain. Certain preceding manufacturing operations (for instance, drilling of cooling holes) can make the part more difficult to coat. 6. Coating programming approach; usually, the more sophisticated the mutual gun-part motion, the higher the sensitivity of the resulting coating thickness to the gun position, fixation plays, etc. 3.3.6 Case Study: Achievable Process Capability

Let us try to estimate a coating-process capability by the example of MCrAlY applied by LPPS on the turbine blade/vane. A typical coating specification required a coating thickness to stay within the 200–600 Pm band. The LSL in this case is determined by the necessity to assure the coating oxidation/corrosion lifetime of 20 000 –30 000 operational hours, the USL value can result from the requirement to avoid a reduction of the component cyclic (TMF) lifetime. The weight can also become an issue if the coating is too thick. Let us consider the most relevant sources of the coating deviation from the specification. 3.3.6.1

Part Complexity

Certain elements of the blade/vane design, like the transitions between the airfoil and the platform, tip shroud, leading and trailing edges, are extremely difficult to coat due to their complex shapes. Frequently, the uniformity of the coating thickness suffers at such locations. This effect will be discussed in more detail in Chapter 7. A typical level of coating variation on such surface elements is of order of 100–200 Pm. Generally speaking, this is not a stochastic variation – it leads to a shift of the thickness average value towards one of the specification limits. Nevertheless a coating sensitivity to the stochastic mechanisms like, for instance, a robot/CNC deflection or the deposition efficiency at the shallow spray angle is also higher. Providing that the efforts are made during programming to center the coating thickness, a value of 50–100 Pm can be taken for the shift of thickness average from the center of the specification band. 3.3.6.2

Mutual Position of the Gun and Component Fixtures

According to the estimate of Section 5.5 this random effect could be assessed as 5–10% of the coating thickness, i.e. 20–40 Pm (a standard deviation). 3.3.6.3

Powder Quality

A powder-size distribution, powder flowability and apparent density are usually specified for the coating process. The stability of the size-distribution affects the inflight parameters of the particle ensemble and, therefore, the deposition efficiency. Apparent density critically affects the actual feed rate (especially, if the volumetric

125

126

3 Quality and Process Capability

feeding is used). The flowability influences the powder-injection conditions and the deposition efficiency, respectively. The size distribution is specified by sieve or Microtrac™ measurements as the set of the weight fractions falling into certain size bands, i.e. a weight percentile of the particles not exceeding a certain size. Variations of distribution to the most extent are related to the following two factors: 1. Whether the particles penetrate through the sieve or are captured by it depends not only on the particle size but also on its shape and surface morphology, which are usually not specified. 2. A size distribution within the size band is not specified. Both those factors together can produce a scatter of 2–4% of the deposit rate (10–20 Pm). In its nature, the apparent powder density is the particle packing density in the bulk. For instance, a theoretical maximum relative density of the bulk created from the monosized ideal solid spheres is 0.74 (when each sphere touches exactly 12 others), whereas another way of dense packing (cubic – with 6 neighbors) assures the density of 0.52 only. Depending on the conditions of powder transportation, storage and filling of the powder feeder, the actual powder density during spray can significantly differ from that determined by the standard test (e.g. ASTM B212). The ability of the powder particles to reach one or another degree of packing density also critically depends on the particle morphology. Rather optimistically, the uncertainty, resulting from the apparent powder density could be estimated as 5–10% of the feed rate, which leads to ca. 20–40 Pm of the thickness variation. It should be emphasized that in case of inappropriate powder handling (e.g. increased humidity, segregation, etc.) this uncertainty can be significantly higher. 3.3.6.4

Torch Pulsations and Drifting

From the point of view of the arc root motion two stages of anode wear should be differentiated. In the first phase, the arc creates more and more surface defects, which become new attraction points for the arc attachment. Due to this, the arc tends to become shorter and the voltage decreases (drifts). After a critical amount of anode damage is accumulated the arc root fixes itself at certain points of the anode surface and the arc stops swirling. This leads to drastic (usually, abrupt) changes in the plasma and particle parameters and in the coating properties, respectively. Such a development makes the second phase of anode wear absolutely unacceptable for the spraying. In terms of quality assurance it should be considered a special cause and must be avoided. How long the first wear phase lasts before it turns to the fixed arc (a wear operational window) depends on the torch design and selected process parameters. All our considerations about the process capability are valid only within this operational window. Wear of the torch electrodes has a double influence on the particle parameters. First, the average gun power drops and the particle parameters drift accordingly. Secondly, as the nozzle wear increases the amplitude of arc-voltage fluctuations increases and this has also a negative influence on the deposition efficiency. Let us consider these two effects separately.

3.3 Process Capability and Stable Process

Shift of the Coating Average Thickness Due to Drifting

For the YSZ TBC a sensitivity of the APS deposition efficiency to the voltage drift varies between 5% [114] and 25% per volt (see Fig. 83 of the present book). This gradient dramatically depends on the targeted coating microstructure. Highly porous TBC are much more sensitive to the torch power than the dense ones. LPPS MCrAlY coating process usually has relatively high process deposition efficiency up to 70–80%. Therefore, one can expect lower sensitivity of this parameter to the voltage variation. Let us assume a value of 3% (or 10–15 Pm) per V (which is also confirmed by theoretical assessments carried out with the simulation package described in Chapter 5). It should be emphasized that this drift could be compensated by the appropriate increase of the current in order to keep the same torch power or the net plasma power depending on the process control philosophy. Increase in Stochastic Thickness Scatter Due to Electrode Wear

Let us refer to the experimental data [35, 115, 116] on the long-term behavior of the torch voltage and related change in the scatter of particle in-flight characteristics. In [35] it was shown that during a 55-h test due to nozzle wear the voltage drift was as high as 10 V and the amplitude of voltage fluctuations was doubled. At the moment no quantitative theory could correlate the arc-voltage fluctuations to the deposition efficiency. For the purpose of a rough estimate let us adopt the following model. Let us assume that the particle impact velocities and temperatures distributed statistically at the impact. Let us also assume that there exist a combination P(V,T) of those two, which determines the particle bonding at the substrate. This parameter is also statistically distributed and the lower “wing” of the distribution function is below the bonding threshold, i.e. particles having the lower values of parameter P are bounced off the substrate as it is shown in Fig. 85. This figure shows temporal oscillations of particle-average bonding parameter P(V,T) and its time-average probability density function p.d.f. for spraying with new and worn anodes. In this case the process deposition efficiency (specifically, its part, which is determined by the torch pulsations) is equal to the area of right part of the distribution function. In [35, 115] it was shown that the scatter of the particle parameters correlate to the amplitude of the voltage oscillations. Let us assume that P is distributed normally and its standard deviation is proportional to the amplitude of arc pulsations. Effect of doubling of the pulsation amplitude on the deposition efficiency in this case depends on its initial value with the new nozzle as is shown in Table 23. One can see that the relative change of the deposition efficiency can be as much as 10–20%. This estimate gives the order of magnitude of the effect considered. In [115] the samples Table 23 Estimate of influence of the electrode wear on the deposition efficiency (increase level of voltage pulsation-double magnitude). Deposition efficiency with the new nozzle

0.9

0.8

0.7

0.6

0.5

Deposition efficiency with the double pulsation magnitude (worn nozzle)

0.74

0.66

0.6

0.55

0.5

127

128

3 Quality and Process Capability

Fig. 85 Influence of arc pulsations on the deposition efficiency – probabilistic interpretation.

were sprayed at two sets of spray parameters that were selected so as to have very close average values of the particle temperatures and velocities, whereas the levels of torch pulsations were very different due to different hydrogen content in the gas mixtures. Deposit efficiencies for those regimes were 73 and 48%, respectively with the lower figure corresponding to the higher level of torch pulsations. This gives a sensitivity of 35% of DE change at doubled pulsation level. The estimates above have been carried out for the ca. 10 V of the voltage drift. Therefore, assuming that 2 V drift (probably, with the power compensation) is allowed and an uncertainty in the coating thickness of 15–30 Pm should be considered. It should be emphasized that this effect has a stochastic nature (it can not be controlled during the process) and, unlike the shift of the average, it cannot be compensated by the adjustment of current. 3.3.6.5

Instability of the Quality-Control Process

Based on the discussion of Section 3.2.4 let us estimate this effect as 3–4% of the additional scatter of the coating thickness, i.e. 10–15 Pm. 3.3.6.6

Surface Preparation and the Part Temperature

Surface preparation and part temperature are two factors critically important for the formation of the first coating layer. Usually, a thickness of this layer is of order of 5–7% of the total coating thickens, i.e. 20–30 Pm. Let us estimate the

3.3 Process Capability and Stable Process

uncertainty resulting from the variation of component temperature and surface pretreatment as 10–15 Pm. 3.3.6.7

Conditions of the Powder-Injection System

As it was shown in Section 2.5.1 conditions of powder hoses and orientation of the injectors could be responsible for the thickness scatter of 5–10Pm in case of APS. Let us adopt this figure for LPPS also. From the examples of Sections 3.2.3 and 5.3.1 for APS and LPPS it follows that the sensitivity of the coating thickness can be as high as 0.5–1% per 1% of the carrier gas flow rate. From the geometrical considerations it is clear that already 1–2 mono-layers of the particles deposited on the injector wall can block 2–5% of the injector cross section and accordingly change the carrier gas speed. Since the effect of the gas acceleration on the powder particle velocities is similar to the variation of the gas flow rate, we can estimate a scatter of the deposition rate resulting from such effect like injector plugging as 2–5% of the coating thickness, i.e. 10–20 Pm. 3.3.6.8

Process Capability

Estimates of process variations are summarized in Table 24, where statistical uncertainties are interpreted as “one-sigma” scatters. It is worth noting that these Table 24 Estimated coating thickness variations at LPPS. Source of variation

Standard deviation, V (μm) – “one sigma” for stochastic phenomena

Part complexity

Shift of average ' (μm)

Remarks

50–100

Mutual position of the gun and part

20–40

Powder quality

30–60

Electrode erosion, voltage drift

20–30

Electrode erosion, increased pulsations

15–30

Instability of quality control

10–15

Surface preparation and part temperature

10–15

Conditions of powderinjection system

10–20

Total

∑ V2

Assuming 2 V drift allowed; effect could be eliminated by the current adjustments Assuming 2 V drift allowed

= 43 − 85

70–130

129

130

3 Quality and Process Capability

estimates can be considered only as indications of the level of importance rather than exact quantities. It is also important that all the assessments were made from quite moderate assumptions about parameter variations, i.e. no extreme cases were considered. Assuming the LSL = 200 Pm, USL = 600 Pm, targeted average 400 Pm, the process is well centered and all stochastic mechanisms acting together the estimate 400 − ∆ − LSL for the process capability looks like C pk = ≈ 0.3…1.0 for the 3V “pessimistic” and “optimistic” scenarios. By eliminating the shift related to the voltage drifting, these figures can be increased up to 0.4 and 1.2, respectively. For rather “simple” locations, where the effect of part complexity is not significant, the process capability could reach the level of 0.8–1.5. One can see that without taking care of the stochastic sources of process variations achievement of the required “3.5–5 sigma” target seems unrealistic. It is quite obvious that the torch stability, powder quality and powder-injection system should be the primary candidates for improvement.

3.4 Maintenance

With a consequent, intelligent and non-overstated maintenance system we can avoid many of the unexpected problems. If we eliminate the most unforeseen events during spraying we will win the time needed to understand other deep problems of thermal coating. “Put all machinery in the best possible condition, keep it that way, and insist on absolute cleanliness everywhere in order that a man may learn to respect his tools, his surroundings and himself.” Henry Ford in his corporate leadership “Today and Tomorrow” published in 1926.

At the beginning of human civilization maintenance had only one form. Corrective maintenance was the oldest form of locally centered maintenance and consisted of tasks to correct failures of equipment. The tasks may consist of repairing or replacing broken components. This “end-of-life” replacement maintenance mentality has survived several thousand years and is still present. Many small companies still work in such a way; they only repair parts that are actually broken. This is only possible if the production is not intensive and the quality requirements are very low. The move towards progressive maintenance happened when people realized that the consequence of a failure could be more “expensive” than the cost of trying to prevent it. Another kind of maintenance was developed. Today, maintenance is a complex working process. It can be classified into several working fields like maintenance management, operations and maintenance,

3.4 Maintenance

preventive maintenance, equipment knowledge, CMMS (computerized maintenance management system), planning and scheduling, engineering, safety, RCFA (root cause failure analysis) and many more. Only the word “maintenance” has different names and a variety of different meanings. Breakdown maintenance, condition-based maintenance, corrective maintenance, emergency maintenance, planned maintenance, predictive maintenance, preventive maintenance, proactive maintenance, reliability-centered maintenance, shutdown maintenance and total productive maintenance are only examples picked out from the currently used nomenclature. Intelligent maintenance should help us to increase our overall production efficiency and to decrease our total manufacturing costs. Goals like life prolongation of equipment can be achieved by operations like fundamental or essential care. This process prevents failures caused by cleaning, misalignment, wrong operating procedures, adjustment and installation. Equipment with a known lifetime is periodically replaced in fixed time intervals. This fixed-time maintenance doesn’t extend the lifetime but prevents failures from occurring by fundamental care. Monitoring the conditions provides assistance. Subjective monitoring like visual or acoustical inspection when supported by human senses is not reliable but often shouldn’t be underestimated. Objective monitoring with the use of advanced sensor technology supported by standard equipment allows us to monitor and log physical values that help keep our machines in the best possible condition and detect any deviations. Fundamental care with fixed-time maintenance combined with subjective and objective monitoring are the main ideas of preventive maintenance in thermal spraying. Most spraying devices can be separated into individual sections (see Fig. 20). The typical spray booth consists of core components, key components and peripheral equipment called the process environment. Different components require a different structure of preventive maintenance (Fig. 86).

Fig. 86 Estimated structure of preventive maintenance for components of thermal-spray system.

131

132

3 Quality and Process Capability

The parts of the core components, like the spray controller, spray gun, power supply, powder feeder and water chillers, are the most critical components of the system because these parts are directly responsible for the spray jet. If we assume that the gun sometimes works close to its physical limits, we can understand the importance of maintenance. The key components, like the movement or manipulation systems, and peripheral equipment, like a soundproof cabin, controlled atmosphere or vacuum vessel, filter or exhaust system, electrical power and gas distribution units, are not vital parts of the system, but there is no system that can work correctly without its peripherals. Every section of the thermal-spray system needs and uses the mentioned main parts of preventive maintenance in thermal spraying but the requirement of each section is different.

133

4 Theory and Physical Trends This chapter addresses the theory of thermal spraying. From the physical point of view a thermal-spray process consists of four subsystems: x x x x

Substrate Particles Plasma/hot gas jet Plasma/combustion torch

In order to apply a good-quality coating on a substrate with required physical, mechanical and chemical properties those four systems must be developed and interact in a way that results in: x Appropriate combination of substrate and particle parameters, which determines the necessary coating quality. x Appropriate combination of plasma/hot gas parameters and powder properties (spray recipe), which assures the required particle in-flight treatment. x Appropriate design of the plasma/HVOF gun (including supply systems like gas, electrical power, etc.) in order to produce the plasma/hot gas with the required parameters as well as to assure the necessary powder-injection conditions. During the last 25 years several thousand papers and tens of books have been published in the field of thermal spraying. Depending on the goals of their authors and the targeted readers, they dealt with different aspects of thermal-spray physics and used different theoretical approaches. Coating-process capability and stability are the targeted subjects of this book. Therefore, understanding of the sensitivity of coating subsystems to the variation of process parameters should be a focus of a theoretical introduction. In this chapter we will attempt to derive the major physical trends of thermal spray based on principles of thermodynamics, fluid mechanics and heat transfer.

4.1 Coating Formation from Separate Particles: Particle Impact, Spreading and Bonding

In thermal spraying the coating is built up from the enormous quantity of particles. Typical number of particles, which form the envisaged 500-Pm thick TBC coating

134

4 Theory and Physical Trends

layer, could be as much as 1010–1011 per square meter. Whereas coating chemical properties, e.g. corrosion-protection behavior, depend mostly on the composition of coating material, its thermal-mechanical behavior depends not only on the corresponding properties of the bulk material but to a great extent on the size and shape of resulting grains and interparticle interfaces and on the character of particle bonding at the contact surfaces. Therefore, in order to have a coating with the required chemical, physical and mechanical properties the coating developer must select the right coating material and develop a process that assures the right shape and size of particle splats on the surface and the necessary quality of the intersplat and splat-to-substrate bonding. Particles impacting the surface will form splats of different shapes, most typical of them are shown in Fig. 87. The shape of splats realized and the splat adhesion to the substrate depend strongly on the particle and substrate material properties, surface morphology of the base material or previous coating layers as well as on three critical process physical parameters [117–121], which are: x Particle temperature x Particle velocity or, more precisely, its component normal to the substrate x Substrate temperature In general, an impacting particle can be solid, semiliquid (softened) or liquid. An impact process comprises two stages. During the first one, which has a duration of 10–9–10–7 s, a particle bulk is affected by a series of shock waves initiated at the particle–substrate contact surface. The level of mechanical stresses at this stage is about V v Up cp up, where Up, cp are the density and sound speed of the coating material and up is the particle normal impact velocity. Whether the particle starts “sagging” or disintegrates into pieces depends on whether the material is in a

Fig. 87 Typical shapes of particle splats at thermal spray and shapes of particles that can create certain types of splats [122].

4.1 Coating Formation from Separate Particles: Particle Impact, Spreading and Bonding

liquid, solid ductile or a brittle state. In order not to be disintegrated a solid particle must have a temperature exceeding the brittle-to-ductile transition point. The intensity of the shock waves and probability of particle to be destroyed, respectively, depend also on the “softness” or pliability of the substrate material. Therefore, the substrate temperature may also be an important factor for this phase. If the particle survives the first stage of impact and the irreversible deformation starts, a second, so-called hydrodynamic phase of impact takes place. At this stage the stress level is of the order of particle specific kinetic energy V v Up up2. The degree of particle flattening depends on the ability of material to resist the mechanical stress, which is determined by the yield stress and solid-state viscosity for solid particles and by the viscosity only for liquid ones. In order to create a splat at the substrate a particle must either be liquid or sufficiently soft. The term “soft” in this sense means that the particle kinetic energy should suffice to overcome its yield stress, which drastically depends on the material temperature. Only for the limited number of materials, e.g. copper, pure nickel, aluminum, can the latter condition be fulfilled without significant heating, for instance, under conditions of a cold spray [123]. All coatings used in power generation can be sprayed only if the particle is heated above the softening temperature, which corresponds approximately to 60–80% of the melting point. A flattening ratio at impact can be estimated from the basic mechanical considerations. An overview of the models of liquid-splat formation can be found in [125]. Theories of the liquid particle impact for the conditions of thermal spray d [124, 125] produce the trend s ∝ Re0.2 , where ds, Dp are the splat and particle D p u p Up Dp is the Reynolds number, calculated for the particle diameters, Re = Pp parameters. In [126] a splat formation from the solid viscous-plastic particle has d  ρ Vp2 been considered, which resulted in the relationship ln  s  ∝ , where Y(Tp)  Dp  Y (Tp ) is the yield stress of the particle material, Tp is the particle impact temperature. The material temperature during the particle flattening depends on the particle temperature and speed at impact as well as on the substrate temperature. Two competing mechanisms are responsible for the change of particle temperature – heat release due to dissipation of the particle kinetic energy and conductive heat losses into the substrate. Overall, as soon as a splat formation is concerned, particle and substrate temperatures determine the particle mechanical resistance and particle integrity (when it is solid), whereas a particle impact velocity determines the level of mechanical stresses that actually deform the particle. Coating adhesion to the substrate and intersplat cohesion can be stipulated by either mechanical interlocking (which depends on the splat shapes and substrate roughness), nonchemical interaction forces (Van der Waals forces) which depend on the quality of contact or chemical bonding (covalent or metallic bonding). Particle bonding at the impact is a probabilistic event. In [117] the number of bonds created during impact is considered proportional to the Arrhenius

135

136

4 Theory and Physical Trends

Fig. 88 Substrate microstructure underneath the removed splat; a central spot and concentric areas where the splat was “welded” to the substrate are visible; substrate and powder material – silver (courtesy of Prof. V. V. Kudinov).

 −E a  exponent exp  where Tcon is the contact temperature, which lies between  k Tcon  the particle and substrate temperatures. A value of activation energy Ea under normal conditions is equal approximately to half of the material evaporation heat calculated per atom and it decreases with increasing contact pressure. One can expect particle bonding as well as coating adhesion and cohesion properties, respectively, to depend dramatically on the particle and substrate temperatures. A significant pressure pc | 0.5 Up cp up [127] of the order of a few thousand bars can develop within the contact area, which has a diameter comparable to the size of the impacting particle. At the periphery of the splat this pressure is much lower, since the spreading material in that area has a vanishing normal velocity component. Therefore, particle bonding takes place mainly in the area of the size of impacting particle as is seen in Fig. 88, where a surface structure under the detached splat is shown. One could expect this effect to be more pronounced on the rough surface, where the spreading splat does not necessarily follow the roughness profile and its periphery can be physically discontinued from the substrate. It is worth noting that the combination of the particle and surface temperatures and of the impact pressure [128] also stipulates a level and a character of residual stresses in the coating. As one can see in Fig. 87 a fingered edge structure is a characteristic feature of most splat types. It develops due to the Rayleigh–Taylor hydrodynamic instability of the decelerating flow [129, 130]. This deceleration is a result of droplet expansion (that follows from the radial mass flow conservation), viscous forces, contact wetting (which depends on the material and the contact temperature) as well as of the influence of the surface morphology underneath the splat. In Fig. 89 [131] an effect of different coating roughness on the splashing behavior and splat formation of 40-μm diameter alumina droplets after solidification is illustrated.

4.1 Coating Formation from Separate Particles: Particle Impact, Spreading and Bonding

Fig. 89 Computer-simulated images of 40-μm diameter alumina droplets, initially at 2055 ºC, impinging with an impact velocity of 65 m/s onto smooth and different rough alumina substrates initially at 25 °C (courtesy of ASM) [131].

Quarters of the final splats are shown for a smooth surface and with dot profiles, representing different roughness levels of 1 μm, 2 μm, and 3 μm. The surface roughness influences locally the contact pressure, material velocity and heat transfer and thus affects the shape of the splat. Coating roughness depends on the process parameters and on the powder morphology, e.g. on the powder size. It should not be forgotten that the splat shape also depends on the particle impact angle. Deflection from the normal impact makes the splat asymmetrical. A normal velocity component decreases and, therefore, so does the contact pressure. As a result, the bonding suffers. Overall, it can be stated that from the physical point of view the whole coating development is essentially about the particle material, powder morphology and about the appropriate combination of the three factors discussed in this section: x Particle temperature x Surface temperature x Particle velocity In the further discussion we will assume that the coating material has already been defined by the metallurgists and handed over to the developers and the task now is to develop the right coating process. And whatever feature of the coating formation is discussed in this book, the other three aforementioned “whales” are behind the issue.

137

138

4 Theory and Physical Trends

4.2 Physics of Plasma Torches

Detailed analysis of the plasma torch operation is beyond the scope of the present book. One can find excellent overviews in [76, 132–135]. But fundamental principles of plasma guns and major trends of their operation can be illustrated with the use of rather simple theoretical estimates. We will consider two types of plasma torches, which are used in the atmospheric pressure (APS) and lowpressure plasma spray (LPPS) technologies correspondingly. A principle scheme of the typical plasma torch is presented in Section 2.4.4 of this book and the schematic layouts of the conventional torch nozzles are shown in Fig. 90. As one can see their designs are almost identical, except for the presence of the nozzle diverging part in the LPPS design, which forms the supersonic De Laval nozzle. In fact, some of the industrial plasma torches, for instance, F4 are able to operate at both conditions using the different nozzles.

Fig. 90 Schematic layouts of APS (a) and LPPS (b) nozzles.

4.2 Physics of Plasma Torches

4.2.1 Plasma Properties

First, the working medium of the torch has to be described. Argon, hydrogen, helium, nitrogen as well as their mixtures, are the typical working gases in the plasma spray. Argon is the most frequently used gas due to the excellent stability of the electrical discharge. In our discussions we will assume that the Ar/H2 mixture is used as a plasma-creating gas. In order to understand the torch behavior one needs to know a set of thermodynamic plasma parameters and the plasma velocity. For our purposes a conventional consideration of the plasma as a two-parameter medium is sufficient. We will use the plasma enthalpy and pressure as determining parameters. Usually, they can be estimated from the standard process control and monitoring data. All other plasma properties can be calculated with the use of corresponding equations of state for thermodynamic parameters and theoretical correlations for the transport properties. Detailed theories of hot gas and plasma thermodynamic and transport properties could be found in the fundamental monographs [136–139] as well as other numerous papers on plasma theory. Furthermore, we will use the following plasma parameters: T U P a h h* V P Q = PU O

– – – – – – – – – –

Temperature [K] Density [kg/m3] Pressure [Pa, mbar] Sound speed [m/s] Static enthalpy [J/kg, MJ/kg] Total (stagnation) enthalpy [J/kg, MJ/kg] Electrical conductivity [1/Ohm/m] Dynamic viscosity [kg/m/s] Kinematical viscosity [m2/s] Heat conductivity [W/m/K] and T

Φ=

∫ O(T ) dT

a heat flux potential [W/m]

T0

D

The latter parameter is more convenient to use than O when heat fluxes in the areas of high temperature gradients are considered; ∆T ∆Φ the heat flux is expressed as q = − O =− , where x is the ∆x ∆x coordinate – Degree of ionization, which characterizes a fraction of the ionized atoms in the gas

Typical calculated enthalpy dependencies of properties of Ar/10%H2 mixture are shown in Figs. 91–96. The data have been obtained with the use of software, which was developed for the offline simulation package described in Chapter 5. Also data [138] could be referred. Specific gas enthalpy in plasma spray applications ranges from 15–30 MJ/kg, which means a range of average temperature from 8000–12 000 K for the argon/hydrogen plasmas depending on the pressure and H2 content. The lower temperature boundary is set by the ionization properties of the

139

140

4 Theory and Physical Trends

gas mixture and corresponds to the ionization degree of 10–4 to 10–3. According to Le Chatelier’s principle, ionization starts earlier at the reduced pressure. A typical plasma spray regime corresponds to the plasma ionization between 10 and 20%. A sharp change of the Ar/H2 enthalpy between 2000 and 4000 K corresponds to the dissociation of hydrogen, whereas the drastic increase at 8000–10 000 K corresponds to the ionization of the gas. For the mixture the ionization starts a little earlier than for pure argon because of the lower ionization energy of hydrogen. It is important that for all thermal-spray regimes the molecular fraction of the gas is fully dissociated and the plasma can be described as a mixture of monatomic gases. Thermal conductivity peaks at the beginning of hydrogen dissociation and the second time at the ionization level of 1–10%. Viscosity as a momentum characteristic is determined mostly by the heavier component and depends rather weakly on the hydrogen content. Below 8000–10 000 K (6–9 MJ/kg) it increases with the temperature and then drops with the increasing level of ionization due to the very high collision cross sections of ions. It should be mentioned that plasma parameters are not homogeneous over the nozzle cross section and the temperature at the APS anode axis could be higher by 1000–3000 K than its section average value (for reduced-pressure plasmas at LPPS this effect is less pronounced). The higher limit of plasma temperature in thermal spray torches is determined by the radiative losses, which dramatically increase with the temperature.

Fig. 91 Enthalpy of argon-hydrogen mixture as function of temperature.

4.2 Physics of Plasma Torches

Fig. 92 Heat conductivity (a) and heat flux potential (b) of argon-hydrogen plasma at different pressures.

141

142

4 Theory and Physical Trends

Fig. 93 Density of argon-hydrogen plasmas at different pressures.

Fig. 94 Sound speed of argon-hydrogen plasma at different pressures.

4.2 Physics of Plasma Torches

Fig. 95 Dynamic (a) and kinematic (b) viscosity of argon-hydrogen plasma at different pressures.

143

144

4 Theory and Physical Trends

Fig. 96 Ionization (a) and electrical conductivity (b) of argon-hydrogen plasma at different pressures.

4.2 Physics of Plasma Torches

4.2.2 Gas Dynamics of Plasma Torch

Major plasma-torch parameters and trends can be estimated with the use of basic gas dynamic relationships. Let us introduce a set of nondimensional parameters: u UD x Reynolds number Re = P , where D is the characteristic size (e.g. nozzle diameter or length, particle diameter etc. depending on the object considered) x Mach number M = u a DD x Nusselt number Nu = , where D is the heat transfer coefficient O M x Knudsen number Kn ∝ which describes continuity of the gas phase Re A gas-dynamic behavior of the torch is determined by the exit nozzle cross section in the case of APS and by the critical section in case of LPPS, which is located at the end of the cylinder section (see Fig. 90). The properties of the LPPS plasma jet are strongly affected by the shape and dimensions of the nozzle diverging part. One should have three relationships in order to determine the two thermodynamic parameters and the plasma speed at the certain nozzle cross section. Furthermore, the section average parameters are considered. The first two relationships are provided by the conservations of energy and mass. The total plasma enthalpy is calculated as the net plasma power per unit of the plasma gas mass: h* =

KI V G

(1)

where K is the torch thermal efficiency, I – electric current, V – gun voltage, G – gas mixture mass flow rate. From the mass conservation one has: Uu =

G S

(2)

where U, u are the gas density and speed and S is the nozzle cross-sectional area. In the thermal-spray applications specific mass flux U u ranges from 10 to 30 kg/m2/s. In the case of APS a nozzle exit pressure is known that provides us with the required third relationship: pe = pa

(3)

where pa is the atmospheric pressure. At LPPS, a supersonic flow no longer “feels” the exit pressure downstream the nozzle and in a general case pe z pa, a condition of the sonic flow at the critical nozzle cross section should be applied in this case: M cr = 1

(4)

145

146

4 Theory and Physical Trends

The plasma parameters at the exit of the nozzle expansion and at the critical cross section are connected through the conventional gas dynamic relationships for the polytropic gas [140]: 1 J +1

Scr  J + 1  2 J −1 = M ex   2  Sex

J−1 2   M ex  1 +  2

J

pex J−1 2   J + 1  J −1  M ex  =  1+    2    pcr 2

−

−

1 J +1 2 J −1

(5)

J J −1

where for the polytropic ratio of partly ionized plasma a value of 1.1 can be adopted. Let us also introduce two equations, which follow from Eq. (2) and general thermodynamic relationships and will be used later in this chapter: J pM = Uu a =

Ga S

(6)

and J p M 2 = U u2

(7)

Based on Eqs. (1) to (5) five typical APS and LPPS regimes have been calculated and results are presented in Table 25. Table 25 Typical parameters of APS and LPPS for ZrO2 and NiCrAlY, respectively. Regime

F4 APS

F4 APS

F4 APS

03CP LPPS

F4 LPPS

1

2

3

4

5

10.5 N/a 55/5 750 46 34.5 43

7 N/a 35/12 600 68 40.8 55

8 N/a 24 500 56 28 43

19 12.7 90/8 1500 53 79.5 60

14 10.5 55/5 750 46 34.5 60

N/a N/a N/a

N/a N/a N/a

260 11570 2340

220 10600 2170

1000 12100 1150 0.42

1000 11900 520 0.2

67 9730 3490 1.6

73 9350 3200 1.5

Torch parameters Exit nozzle D, mm Critical nozzle D, mm Flow rate Ar/H2, slpm Current, A Voltage, V Power, kW Thermal efficiency, %

Average plasma parameters in critical cross section Pressure, mbar Temperature, K Speed, m/s

N/a N/a N/a

Average exit plasma parameters Pressure, mbar Temperature, K Speed, m/s Mach number

1000 10800 460 0.2

4.2 Physics of Plasma Torches

4.2.3 Energy Balance of the Plasma Gun

A heat input into the plasma is determined by the resistive heat realized in the electric arc. Neglecting Joule losses in the electrodes, the heat input is calculated as QJ = IV. The thermal energy is partly transformed into the plasma enthalpy and kinetic energy and partly transferred back to the electrodes and removed with the cooling water due to the near-electrode voltage drop as well as radiative and conductive heat losses. Near-electrode energy losses due the electron flux into the anode are determined by the anode voltage drop and the work function of the anode material. The overall anode energy balance is approximately 10–12 eV per electron for most plasma gases and electrode materials [137]. Estimations of the cathode heat balance are more difficult due to the complexity of electron-emission mechanisms and uncertainty in the magnitudes of ion and electron currents in the near-cathode layer. Experiments show a large scatter in measured data. According to [137] the volt equivalent of the heat flux is ca. 1–2 V depending on the current that is much less than that at the anode. For estimates, the overall near electrode heat losses could be assumed to be Q e = 12 I [W]

(8)

which constitute 15–25% of the overall arc heat balance. For the typical spray parameters an approach of the “grey” body can be used to estimate the magnitude of the radiative heat flux from the plasma bulk. According to this approach, radiative heat losses from the nozzle can be assessed as Q r = H(T , p ) V SB T 4 S D L

(9)

where D and L are nozzle diameter and length and VSB is Stefan–Boltzmann’s constant. The maximum value of the “degree of greyness” H is 1, which corresponds to the “absolute blackbody” radiation. It is worth noting that for the typical plasmaspray parameters H is much less than 1 and rapidly grows with the temperature and pressure. In fact, the actual temperature dependence of the radiation flux could be as strong as T 8–T 9 or even T 10–T 12. For our estimates values for H the semispherical radiative layer of argon plasma of 1 cm thickness were taken from −8 2 −3 [141] and interpolated as H = 10 −8.1 ⋅ 10 T + 2 ⋅ 10 T − 14.3 p [bar]1.45. Accurate calculations of conductive and convective heat fluxes into the nozzle wall are extremely difficult or even impossible due to the complexity of the flow structure caused by stochastic motion of the electric arc. Estimated Reynolds u UD numbers Re = for the typical spray regimes range between 500 and 2000 P that formally could be attributed to the laminar flow. But, on the other hand, based on current understanding of the arc root motion inside the anode, one could expect intensive convective flows in the nozzle. For a very rough estimate, a boundary layer approximation for the cylinder tube could be used:

147

148

4 Theory and Physical Trends

Q c = Nu Φ S D D  Nu = 1.86  RePr   L

(10) 1/3

 K   K  w

0.14

where subscript w denotes parameters at the anode wall temperature. Estimated heat balance of APS and LPPS plasma guns is shown in Fig. 97 for the following parameters: Ar/H2 35/15 slpm mixture, nozzle diameter (exit-APS, critical – LPPS) 6 mm, voltage 60 V, current 600 A, nozzle length 40 mm. Thermal efficiency has been determined from the iterative procedure to fulfil the energy conservation with heat losses defined according to Eqs. (8) to (10), which resulted in values of 60% and 43% for LPPS and APS regimes, respectively. In the LPPS nozzle a significant part of the plasma thermal energy is transformed into the kinetic one. In the critical nozzle section a specific kinetic energy is of order of a2/2 = 2.5 MJ/kg that constitutes 10–15 per cent of the plasma enthalpy. At the exit of the nozzle diverging part the kinetic component reaches as much as 20–40 per cent of the total jet energy. This is why the plasma at LPPS conditions is

Fig. 97 Typical energy balance of APS and LPPS plasma torches, with thermal efficiency of 43% and 60%.

4.2 Physics of Plasma Torches

colder at comparable levels of the net flow energy. LPPS torch has higher thermal efficiency due to the vanishing role of the plasma radiation, which results from the reduced plasma temperature and, especially, from the lower pressure in the cylinder nozzle part. Another important effect should be mentioned with regard to the torch heat balance. As was discussed in Sections 2.4.4 and 3.2.3 during operation plasma parameters tend to drift, which is usually reflected in the reduction of gun voltage. This happens due to the erosion of anode walls and movement of the arc attachment towards the cathode, which makes the arc shorter. Usually, this is corrected by increasing the current in order to get the same electric gun power. But different loss mechanisms react differently to the increase of current. For instance, at the higher currents, the role of the electrode voltage drop in the overall heat balance increases. Therefore, the net plasma power should be the target of process stabilization rather than the gun electrical power. This could be easily implemented in the manufacturing process since the water in and water out temperature measurements as well as the water flow rates are the standard outputs of any commercial hardware control system. 4.2.4 Major Trends

Based on model Eqs. (1) to (5) and plasma properties (Figs. 91–96) it is possible to estimate the most important reactions of the plasma torches to the variations of process parameters without detailed calculations. 4.2.4.1

Variation of the Gun Power; the Gas Flow Rates and Composition Unchanged

Such a variation leads to the increase of plasma specific enthalpy and temperature. The Reynolds number increases due to a reduction of viscosity, the flow-rate factor UuD stays constant. Radiative heat losses increase due to the higher temperature. So do the convective losses, since heat flux potential and the Nusselt number increase. Therefore, the torch thermal efficiency drops. In the case of APS the exit pressure stays unchanged, consequently the plasma density drops and the velocity increases. So do the sound speed and the plasma speed at LPPS critical nozzle section, where the sonic condition of M = 1 remains. In this case, the plasma pressure increases proportionally to the sound speed as follows from Eq. (6). 4.2.4.2

Variation of the Plasma Composition at the Same Specific Plasma Enthalpy

An increase of the hydrogen content causes a reduction of the plasma electrical conductivity (Fig. 96), which results in an increased torch voltage. Respectively, the gun power also increases. This effect is widely used to achieve the higher plasma enthalpy at the same current and vice versa, the same plasma specific enthalpy can be achieved at the reduced current. Let us consider the latter option. From Fig. 91 it follows that the gas temperature is reduced with the increase of hydrogen content in the plasma. This leads to the significantly decreased radiative

149

150

4 Theory and Physical Trends

losses (see Eq. 9). It is worth mentioning that the heat conductivity and heat flux potential of the argon-hydrogen mixture is higher than the one of pure argon at the same temperature due to higher mobility of hydrogen atoms. But if compared at the same specific enthalpy, the result is opposite, as is seen from Fig. 92. This is due to the lower temperature of the mixture and lower concentration of electrons. This is why the convective heat losses in the presence of hydrogen in the working medium are also lower at the same enthalpy. Since the current, which is required to maintain the same plasma enthalpy with H2 is lower than for the pure argon arc, the contribution of the electrode voltage drop into the heat balance is also lower. All these three factors make the thermal efficiency of the plasma torch significantly higher at comparable power levels when operating with hydrogen. The difference in thermal efficiency could be as high as 5–15 per cent depending on the hydrogen content. Also the electrode lifetime is higher at constant electrical power due to the reduced current. Hydrogen at LPPS also carries another important role – it binds the residual oxygen, prevents the material oxidation and may lead to a chemical reduction of the coating. It is worth noting that due to very high diffusivity of hydrogen an electric discharge in Ar/H2 mixture is more prone to restrikes and its stability, respectively, is lower. Generally, conventional torches with more than 25–30% of hydrogen in the mixture are extremely unstable. 4.2.4.3

Variation of the Plasma Flow Rate at Unchanged Gun Power and Gas Composition

Increase of the gas flow rate at the same gun power leads to the reduction of plasma enthalpy, temperature and heat conductivity: h ∝ P /G

(11)

Therefore, the heat exchange with the electrode walls decreases (even though the Reynolds and Nusselt numbers increase – effect of the higher heat flux potential is dominating) and torch efficiency increases. A velocity variation is defined by two opposite trends in Eq. (2): increased flow rate and increased density due to lower temperature. Nevertheless, from Eqs. (2) and (11) one obtains: u ∝

1 Uh

From Fig. 93 it can be seen that Uh decreases when the specific enthalpy is reduced at constant pressure. Therefore, at APS the exit plasma speed increases with gas flow rate. The trend is opposite at LPPS. According to Eq. (6) the pressure in the critical nozzle section goes up. So does the exit pressure. But the reduction effect of the enthalpy on the speed of sound is stronger. Therefore, the plasma speed in LPPS torch reduces with increased flow rate.

4.3 Structure of Plasma Jets

4.2.4.4

Effect of Nozzle Diameter

Reduction of the nozzle diameter while keeping the gas flow rates unchanged reduces heat losses mainly due to the smaller surface of plasma volume. From the standpoint of its influence on plasma parameters it is similar to the variation of the gas flow rate, which was considered in Section 4.2.4.3, since the area-specific mass flux increases at the same specific plasma enthalpy. The effect of the powder carrier gas, in the case of internal injection, could also be interpreted as a reduction of the effective nozzle cross section. In the case of LPPS with internal powder injection it leads to an increased nozzle pressure and, therefore, changes the torch conditions upstream. 4.2.5 Plasma Swirl

A tangential component of gas velocity inside the plasma nozzle is formed by a special swirling device, for instance, by the gas ring installed at the entrance of the plasma gas into the torch interelectrode chamber. The ring has a set of holes drilled at a certain angle to the radii. The major idea of a swirling gas flow is to enforce the movement of the anode arc attachment (swirl-stabilized plasma arc). This makes the gas heating more homogeneous and drastically increases the anode lifetime. Depending on the gas flow rate, size and quantity of holes and the gas injection angle tangential speed can be as high as 50–100 m/s. Because of conservation of swirl momentum this rotation is accelerated inversely proportional to the radius at the transition to the cylinder anode channel where the swirl could be as high as 200–300 m/s. Moreover, the rotational speed increases towards the anode axis. Such a speed distribution is a prerequisite of the Taylor-type flow instability [142], which leads to the radial flow circulation and intensification of heat transfer into the anode wall. Respectively, torch heat losses increase. Swirling plasma creates an aerodynamic force, which drags injected particles in the tangential direction. Due to circumferential inhomogeneity of the plasma this force is oscillating, which constitutes a powerful source of particle dispersion perpendicular to the direction of powder injection.

4.3 Structure of Plasma Jets 4.3.1 APS Jet

A typical APS plasma jet is shown in Fig. 98. At the torch outlet an isobaric plasma jet is formed. Typical values of Reynolds numbers at the exit of APS plasma gun calculated using the nozzle diameter of 8 or 10.5 mm range from 700 to 2000 depending on the gun size and operational regime. This is close to but still below the threshold value of ca. 2300, which according to the classical fluid dynamics

151

152

4 Theory and Physical Trends

Fig. 98 A typical picture of the structure of an APS jet. The initial quasilaminar part transitioning into the bright converging cone is clearly seen. Injected particles visualize the distant region of the jet.

corresponds to the transition of laminar to turbulent flow. Nonetheless, due to intense disturbances the flow still cannot be considered laminar in the classical meaning of this term. A term “quasilaminar” can be used to describe such a situation. Several physical factors determine the turbulent transition at the contact interface between the plasma and surrounding atmosphere. The high-speed jet entrains the air into the motion and creates a transitional boundary layer with thickness increasing downstream. At a certain distance this layer becomes turbulent, which is followed by intensive momentum and mass transfer through the jet boundary and finally by the complete turbulization of the jet. As a rough estimate a condition of the critical Reynolds number for the turbulent transition at the flat plate [143] can be used, which gives Re xCr =

u UX GX = ≈ 3 ⋅ 105 − 106 P SP

(12)

depending on the initial level of flow disturbances. Using the typical air properties 5 − 15 ≈ 5 − 15 mm . This testifies that at atmospheric pressure one obtains X ≈ u shortly after the nozzle exit the APS jet boundary becomes turbulent (for instance, measurements [144] could be a reference). This turbulent zone expands in both directions inside and outside of the jet. The angles of expansion depend on the gas plasma composition and, to a less extent, on the initial level of disturbances. This process shapes a so-called “bright plasma cone”, outside of which the plasma is already partly mixed with the cold air and, respectively, radiates significantly less. After the point where the turbulent layer reaches the axis (a “tip” of the cone), the flow is fully turbulent and intensively mixes with the surrounding air. In fact, it is no longer a plasma. A typical length of the bright cone is from 4 to 6 exit nozzle diameters for the guns like F4 operating at atmospheric conditions. After the turbulent transition is over, the jet expands at an approximately linear rate with an expansion angle of 10–15 degrees, as is seen in Fig. 99, where results on jet radii vs. standoff distance derived from the direct measurements of heat fluxes into the substrate [145] are shown. By reducing the initial level of flow disturbances as well as by reducing the flow speed it is possible to some extent to control the turbulent transition and

4.3 Structure of Plasma Jets

Fig. 99 Evolution of plasma jets with the axial distance from the gun nozzle [145].

the jet expansion angle [145]. From Fig. 99 it follows that the expansion rate of the hydrogen-nitrogen jet from PlazJet gun is visibly lower than that of F4 and the jet enthalpy, respectively higher at the comparable distances. This effect could be attributed to the reduced level of arc pulsations in the high-voltage PlazJet torch. 4.3.2 Structure of LPPS Jet

A supersonic LPPS jet generally consists of two parts, which are clearly seen in Fig. 100: an initial shock wave structure (so-called “shock diamonds”) and an equilibrium isobaric slightly expanding jet, which ends up with the region of intense turbulent mixing. A “diamond” structure occurs when the plasma pressure at the exit of the diverging part of the nozzle differs from that in the spray chamber (pressure or dynamic imbalance) and/or when the nozzle does not form the fully axial flow and the plasma possesses a significant radial velocity component at the exit (a kinematic imbalance). In the case of underexpanded jet, i.e. the exit nozzle pressure is higher than the ambient one pa, the flow tends to expand immediately after the nozzle, which results in overexpansion and the plasma pressure at the axis can become much lower than pa. Then the flow is contracted and the pressure increases in a series of shock waves. The first “diamond” ends up with the straight shock wave. Then the oscillating process repeats several times with decreasing magnitude. If the exit pressure is lower than pa, (overexpanded jet) the flow is initially compressed and then the process follows a scenario similar to that of an underexpanded jet. Usually, one can see up to 5–6 “diamonds”, but only one or two of them are clearly pronounced as in Fig. 100. The shock structure does not exist in case of the profiled De Laval [146, 147] nozzle with the bell-shaped diverging part when the design pressure ratio matches the chamber pressure. After the series of “diamonds”, the jet becomes isobaric with the pressure equal to pa and its parameters can be estimated from the basic gas-dynamics relationships [148]. For the plasma conditions, when the effective specific heat ratio is close to 1 (as already mentioned, a value of 1.1 could be a good approximation), the jet

153

154

4 Theory and Physical Trends

Fig. 100 A typical LPPS supersonic jet. Shock diamonds are distinguishable at the exit of the nozzle.

Mach number after the “ideal” Laval nozzle can be calculated with good accuracy p 2 as M ≈ ln cr + 1 . In the case of the imbalance of exit conditions the Mach Jp pc number of the equilibrium part of the jet after the shock diamonds is lower than it would be in the case of an equilibrium nozzle. Corresponding reduction factors for several off-design regimes as well as for the conical nozzles with different expansion angles are shown in Fig. 101. One can see that the kinematic imbalance has a stronger influence on the parameter of the jet than the exit pressure mismatch. The intensity of the shock waves and diameter of the shock “diamonds” also increase with the increased imbalance of exit nozzle conditions. A subsonic boundary layer at the nozzle wall still “feels” the pressure outside the nozzle. A significant overexpansion could cause a separation of the boundary layer in the diverging part of the nozzle and, therefore, reduction of the effective nozzle expansion. Respectively, the Mach number in such a case would be lower than predicted by the nonviscous quasi-1D gas dynamics [147, 148].

Fig. 101 Influence of exit-pressure imbalance and exit nozzle angle on jet max number Ma; Mid – envisaged Mach number in case of equilibrium profiled De Laval nozzle with an axial exit; a/2 – exit nozzle half-angle; Pex and Pa – exit nozzle and chamber pressures, respectively; J = 1.1, Mid = 2.4.

4.4 Particles in Plasma

A typical jet Reynolds number calculated using the plasma properties and jet diameter D is usually of the order of 200–500 at chamber pressures of 30–50 mbar and it decreases with the pressure, i.e. the jet is quasilaminar. As far as stability of the jet boundary is concerned, one can apply the same considerations as for the APS case. From Eq. (12) one can see that the critical jet length increases proportionally to the inverse density and, respectively, to the inverse pressure. For instance, in the case of a chamber pressure of 30 mbar and a jet speed of 3500–4000 m/s, which corresponds approximately to a Mach number of 2, the turbulent transition would happen at a distance of the order of 0.5 m. A spray distance usually does not exceed 300–400 mm. Laminar viscosity and heat conductivity also contribute to the decay of the plasma jet. Viscous and thermal transiton layers are initiated at the nozzle exit and propagate towards the flow axis downstream the jet. They completely occupy the jet cross sections G G at distances, which could be estimated as Lvis ≈ 0.1 and Ltherm ≈ 0.1 Pr , P P respectively. These mechanisms may determine the length of laminar plasma jet at low pressures and/or at low plasma gas flow rates. It can be seen that LPPS jets become longer with the increase of plasma enthalpy and corresponding decrease of viscosity. From the mass conservation a diameter of the equilibrium part of the jet can be 4 Ga estimated as D ≈ . It increases with the reduction of pressure. S J pM For the typical LPPS regimes the plasma jet is quasilaminar and the effect of mixing with the surrounding gas could be neglected in a first approach. Also radiative losses are negligible due to the low pressure and the jet can be considered adiabatic. A slight jet expansion with the distance occurs because of its deceleration due to the influence of viscous forces and momentum losses to powder acceleration.

4.4 Particles in Plasma

Mechanical interaction between particles and the hot gas or plasma is determined by various forces such as: x x x x x x

Aerodynamic drag, which is directed against the relative particle motion Aerodynamic lift, which is applied to the asymmetrical particles Pressure gradient (Archimedes force) Added mass force related to the particle acceleration Basset history term Shear lift, which is caused by the gas lateral velocity gradient and produces particle drifting across the flow x Forces caused by particle rotation (Magnus force) x Jet force in the case of evaporating particle

155

156

4 Theory and Physical Trends

The following physical mechanisms determine the heat exchange between particles and the plasma: x x x x

Thermal conductivity Atom/ion/electron collisions in the case of rarefied plasma flow Plasma and particle radiation Electron emission from the particles at high temperatures (> 2000 °C) and related floating electrical charge

For our subject, the aerodynamic drag, atom and ion fluxes and the thermal conductivity are the most important physical processes. Various theories are used to describe them for the different flow regimes. One can find an overview of the theoretical models in [149]. Let us make some general estimates. 4.4.1 Particles at APS

In-flight particle conditions at APS both inside the gun and in the free jet correspond to the regime with M < 1, Re = 0.5–10, Kn < 0.1. This means that the Stokes model of the drag force is applicable, i.e. dVp

≈ 3 S P (u − u p ) D ∝ Q ( U u ) D ∝ Q G D (13) dt where the kinematic viscosity Q depends only on plasma properties, whereas the term Uu is proportional to the mass flow rate of the gas. Particle heating in this case is determined by the plasma heat flux potential and by the particle diameter: mp

dTp

≈ Nu [Φ(T ) − Φ(Tp )] S D ∝ Φ(T ) D (14) dt From Eq. (14) and Fig. 92 it follows that the particle temperature always increases with the increase of specific plasma enthalpy. From Eqs. (2) and (13) it follows that the particle acceleration is proportional to G Q. From these two observations the following trends can be derived: mp c

x If the flow rate is kept unchanged the particle acceleration is proportional to the plasma kinematic viscosity. The latter plasma characteristic at atmospheric pressure is increasing up to the enthalpy of 7–10 MJ/kg (see Fig. 95b). Above this level it changes very little with the power. Therefore, one can expect the particle velocity to be rather insensitive to the torch energy at small distances from the nozzle exit, where the plasma enthalpy is still high. At the same time an increase in power causes an increase of particle speed in the area of colder plasma – at higher standoffs. These trends are illustrated by the data from DPV2000 measurements of zirconia powder inflight parameters at APS presented in Fig. 102. Whereas an increase of particle temperatures is clearly seen for all standoffs, a similar velocity trend becomes pronounced only at distances from the nozzle exceeding 100 mm.

4.4 Particles in Plasma

Fig. 102 Influence of the torch power and standoff on the particle parameters at APS. Nozzle diameter 6 mm; Ar/H2 flow rate 40/8 slpm, squares – torch current 800 A, rhombuses – 600 A; measurements performed with DPV2000 device; different locations correspond to the scanned area of 20 × 20 mm2; measurements carried out at standoffs of 75, 100, 150 and 200 mm (from the top to the bottom). The scanning started at the centre of the area forming a rectangular spiral with 5 mm location spacing.

157

158

4 Theory and Physical Trends

x Particle velocity increases with the increased flow rate. Increase of the flow rate without adjustment of the torch power to keep the specific enthalpy at least at the same level would lead to the controversial trends from the point of view of coating quality: higher velocity at lower particle temperature. It is why the process control (e.g. compensation of the voltage drift) by means of adjustment of the gas flow rate can bring rather unpredictable effect on the coating quality. x From the point of view of particle behavior variations of hydrogen content in the plasma affect particles in a similar way as the variation of total gas flow rate. In order to have a definite trend in the coating quality simultaneous “smart adjustments of the power” by means of current variation would be required. Also from Figs. 92 and 95 one can see that at the low levels of plasma enthalpy (5–10 MJ/kg) the effect of the hydrogen content can be uncertain. All this makes such a control difficult to use in the manufacturing. x With the net power input, gas flow rate and composition fixed a change of the nozzle diameter would affect the plasma velocity, whereas its enthalpy and viscosity remain approximately the same. The plasma velocity changes as the inverse diameter squared, whereas a “bright cone” length is proportional to the nozzle diameter. Therefore, with the increase of nozzle exit diameter one can expect a reduction of the particle speed and some increase of temperature due to the effect of dwell time. Experimental confirmation of this trend can be found in [112] and in Fig. 148 of this book. x From the behavior of the kinematic viscosity and heat flux potential it is seen that the particle parameters are more sensitive to any types of variations at specific enthalpies of 5–10 MJ/kg. 4.4.2 Particle at LPPS Particle Acceleration and Heating in the LPPS Free Jet

4.4.2.1

In a free LPPS jet the typical particle parameters correspond to the free molecular flow (Kn > 1). Therefore, the particle acceleration is proportional to the plasma jet dynamic pressure, mp

dVp dt

≈ CD

U (u − u p )2 S D2 ∝ p M 2 D2 2 4

where CD = 2–3, whereas for the particle heating the following simple expression is applicable: mp c

dTp dt

∝ U u h * D2

Variation of power. As was shown in Section 4.2.4.3 an increase of energy input leads to the higher critical nozzle pressure and consequently to the increased exit Mach number, providing the chamber conditions and gun flow rates are fixed.

4.4 Particles in Plasma Table 26 Influence of torch power on particle parameters at LPPS. Torch power, kW

38.25

32.72

27.58

22.86

18.35

Particle temperature, C

2054

2014

1998

1977

1942

250

246

249

249

245

Particle speed, m/s

Obviously, the specific plasma enthalpy also increases. Therefore, both particle acceleration and heating will be enhanced. Nevertheless, due to the logarithmic dependence of the Mach number on pressure, the particle velocity trend is likely to be rather weak. For instance, an increase of plasma enthalpy by 10% would change the critical nozzle pressure by not more than 5% and the particle acceleration by only 1%, which can hardly be measured. In Table 26 an effect of torch power on the in-flight parameters of MCrAlY particles is illustrated. Experiments have been carried out with the use of DPV2000 at EMPA (Thun, Switzerland). One can see that whereas the temperature trend is obvious, the particle speed remains, in fact, unchanged. Variation of chamber pressure. In case of the ideal De Laval nozzle and with the critical nozzle pressure given, a product pcM2 that determines the acceleration of the particles in the low-pressure jet, has a maximum that corresponds to the criticalto-exit pressure ratio of approximately 1.7 (chamber pressure of 100–150 mbar). Heating intensity, determined by enthalpy flux Uuh*, monotonously decreases with the Mach number. This means that in the typical range of LPPS parameters both particle acceleration and heating rates outside the nozzle decrease if the chamber pressure is reduced. At the large Mach numbers they decrease proportionally to the chamber pressure, as do the particle acceleration and heating rates. Nevertheless, particle impact parameters depend also on the dwell time in the plasma, i.e. on the jet length, which increases with the reduction of pressure (see Section 2.3.2). Thus, reduction of chamber pressure and increase of standoff distance are complementary measures in the case of LPPS and should be considered together. They could be used, for instance, in order to enlarge a spray pattern and increase a powder spray rate, while maintaining a thickness of the coating layer applied in one gun passage. Reduction of pressure can also be necessary from the standpoint of chemical purity of the coating. Significant reduction of the LPPS operational pressure below 5 mbar [150, 151] allowed development of a thin coating technology for large-surface applications. It is worth mentioning that at such low pressures the plasma can no longer be considered equilibrium. In the presence of a relatively weak electrical field the plasma electron temperature can become significantly higher than that of heavier atoms and ions [152]. Therefore, the heat transfer to the particles also increases. Variation of gas flow rate and hydrogen content. Increase of the gas flow rate at the fixed net power input into the plasma leads to the increased nozzle pressure and to the higher Mach number in the jet, respectively, if the chamber pressure

159

160

4 Theory and Physical Trends

is kept unchanged, whereas the specific plasma enthalpy decreases. This results in a higher particle acceleration and reduced temperature. As already discussed, hydrogen in plasma makes the same specific enthalpy possible at the reduced torch current. At comparable values of plasma enthalpy the particle speeds slightly increase, whereas the particle temperatures can be slightly reduced because of decreased heat flux potential (see Fig. 92b) and the dwell time. As in the case of flow-rate variation the trends are rather controversial from the point of view of coating formation. Moreover, due to the rather low sensitivity of sound speed and heat conductivity to the hydrogen content, these trends can be shaded by associated effects, for instance, by the change of the arc-pulsation spectrum and/or an increased torch thermal efficiency. Therefore, hydrogen is not the best candidate to be used for the compensation of power drift as a process control parameter, nor is the total plasma flow rate. Influence of shock waves. As shown in Fig. 101 a strong dynamic and/or kinematic imbalance at the nozzle exit results in lower Mach numbers and gas speed due to losses in shock waves and, respectively, in higher static temperature and reduced gas density (at the given chamber pressure). This reduces both acceleration and heating rates. For instance, a velocity imbalance at the exit of the conical nozzle with a half-angle of 45 degrees leads to the reduction of acceleration rate, which is proportional to the Mach number squared, by more than 30%; the effect of pressure imbalance is less pronounced, especially for the underexpanded jet. Thus, a bell-profiled nozzle with an axial flow discharge could be advantageous within the relatively wide range of chamber pressures and torch parameters. The applicability of the profiled nozzles is discussed, for instance in [147]. Significant pressure or velocity imbalance at the nozzle exit leads to the appearance of the strongly pronounced first shock diamond with a length up to 5–8 nozzle exit diameters. In the case of an underexpanded jet the pressure inside this structure could be an order of magnitude lower than that in the chamber. This could result in the situation when a significant part of the particle trajectory is in fact lost for the acceleration and heating. Generally speaking, it is difficult to imagine a situation when a shock wave structure at LPPS would be beneficial for the particle treatment. For the particular spray process an attempt should be made to achieve a “smooth” exit from the nozzle by selecting an appropriate combination of plasma power, gas flow rates, chamber pressure and nozzle shape. Nevertheless, use of profiled nozzles, while being beneficial from the standpoint of higher particle in-flight parameters, may be unacceptable in industrial production due to difficulties with their manufacturing and costs of revalidation. 4.4.2.2

Particle Acceleration and Heating Inside the Nozzle

Usually particles at LPPS are injected several millimeters upstream of the critical nozzle cross section. Depending on the torch design and process parameters the particles gain from 10 to 50 per cent of their enthalpy and kinetic energy inside the nozzle. From the point of view of particle treatment it is obviously beneficial

4.5 Spray Footprint (Spray Pattern)

to utilize the dense plasma flow as much as possible. But the possibility to move the injection points upstream is limited by the build-up of sprayed material at the nozzle wall when particle trajectories reach it. Particle parameters in the critical cross section of the under expanded supersonic LPPS gun nozzle correspond to the regimes with M | 1, Re v 1, Kn v 1, which are called transitional between the continuous and free molecular flows. Different theories may produce quite different values of the drag coefficient CD as well as for the heat fluxes. Theoretical and semiempirical expressions (see, for instance [153]) produce CD values, which differ significantly from each other that makes estimations of trends very difficult. In such a situation it makes sense to consider both limiting trends. In the continuous flow particle acceleration is determined by the product GQ, whereas for the free-molecular flow it is proportional to U u2 v G a. Comparing Figs. 94 and 95 one can see that “continuous” and “free molecular” limiting trends show opposite trends. Despite the fact that it is impossible to say without detailed analysis, how close would the reality be to one or another trend, one could expect relatively low sensitivity of particle acceleration in the nozzle to the power input. Particle heating increases with the plasma enthalpy. As in the case of APS, an increase in mass flow rate and/or in the hydrogen content produces the opposite trends in acceleration and heating.

4.5 Spray Footprint (Spray Pattern)

At the injection point a bunch of particles has a diameter of 2–4 mm. After being affected by the stochastically distributed aerodynamic forces in the plasma jet, as well as the initial dispersion of the injection particle velocities the particle flow forms a footprint with a size that may multiply exceed the diameter of injectors. For instance, the size of the APS spray spot at the spray distance of 100–150 mm can reach 20–40 mm, whereas an LPPS single-injector footprint used in industrial applications can be as big as 50–60 mm at typical standoffs. An HVOF footprint usually is smaller. In the case of a single-injector spray, a coating pattern usually has a round or elliptical shape, as is shown in Fig. 103. In the case of the double-injector gun, a spot can have two peaks or a single common peak depending on the separation of the spots from each individual injector. The geometrical structure of the spray spot can be characterized by its position with respect to the powder injector(s), ellipticity, height of the peak(s) and size of the spot(s). When speaking about the variety of shapes of APS and LPPS spray patterns used for producing the surface coatings, one should differentiate two major cases, which with certain reservations can be called “small” and “large” spray spots. A spot size is defined with respect to the diameter of the hot jet, which carries the particles. The term “small” spot denominates the situation when the absolute majority of the particles fly inside the jet until they hit the substrate. In contrast, the outer

161

162

4 Theory and Physical Trends

Fig. 103 LPPS (left) and APS (right) spray patterns (plate height 70 mm).

area of a “large” pattern is created by particles that leave the jet soon after the exit from the gun nozzle and fly a significant part of their trajectories outside the hot flow. The former spot is fully geometrically aligned with the plasma jet and has a clear advantage due to the fact that all the particles have sufficient dwell time to be fully treated in the plasma. It is typically used in experiments in research labs. However, the need to assure the high production rate leads to increased spray patterns. A large spray pattern consists of two regions: the internal spot (hot area) created by the well-treated particles and the peripheral area that is created by the particles, which left the plasma some distance before the impact. Such a structure determines several features of the coating: x Coating quality in the spot periphery could be much worse than in the hot area. In Fig. 104 a typical large LPPS spray pattern of a MCrAlY coating is shown. Porosity distribution has been measured by microscopy image processing of polished cuts. One can see that the porosity level, being below 1% in the middle of the spot, could reach as much as 6–10% at the periphery. Examples of the corresponding coating structures are presented in Figs. 105 and 106. x Deposition efficiency at the periphery is relatively low. Because of this, the peripheral areas are extremely sensitive to the spay angle. x Particles that form the outskirts of the spot, leave the hot jet at shallow angles. This makes their dwell time in the plasma and, therefore, their impact parameters very sensitive to the variations of injection conditions and gun parameters. Also sensitive are the properties of the spot periphery. This effect can also make a “paintbrush” sensitive to the gun acceleration, since the latter affects the powder-injection conditions. It is worth noting that even in the “small” spot particle parameters are distributed in-homogeneously. The plasma temperature and velocity reduce towards the jet boundary and the particles still flying inside the jet, but in its outer part are usually colder and slower.

4.5 Spray Footprint (Spray Pattern)

Fig. 104 A typical porosity distribution in a MCrAlY LPPS spray spot on a test plate; a double-injector spray spot has been cut for the metallurgical analysis (plate height 120 mm) (courtesy of ALSTOM).

Fig. 105 Typical quality of the MCrAlY coating from the “hot” area of the spray spot (courtesy of ALSTOM).

Fig. 106 A typical coating quality at the peripheral area of the large spray pattern (courtesy of ALSTOM).

163

164

4 Theory and Physical Trends

The size of the spray pattern is selected based on the following considerations: x Size of the sprayed components. This determines the target deposition efficiency; with a too wide spray pattern a large fraction of powder will be wasted. x Diameter of plasma jet and requirements for particle temperatures and speed in the peripheral areas; if a certain reduction of the coating quality due to effect of spot margins is tolerable, an operation with the large spot allows a significant increase of the spray rate. Several examples of the sensitivity of the LPPS spray pattern to the process parameters will be presented in Section 5.3.1.

4.6 Influence of Particles on Plasma Flow

The effect of plasma deceleration can be easily assessed from the conservation of the total axial momentum: ∆u = −

ɺp m G

Vp

ɺp m

ɺ p is the powder load in [kg/s], G is gas mass flow in where m G [kg/s] can be as high as 1–2 in industrial plasma spray applications. Therefore, The ratio

the plasma can be slowed down by 200–400 m/s. Internal injection of powder causes a slight increase of the pressure inside the torch nozzle. Depending on the material properties a heat equivalent of 1–1.5 MJ/kg must be transferred to the powder in order to have it melted in the plasma flow. Therefore, ɺp m for ∝ 1 − 2 up to 20% of the jet enthalpy is to be spent for particle heating. G

Fig. 107 Influence of the powder feed rate on the coating weight gain.

4.7 Substrate Surface Temperature

Reduction of plasma enthalpy and momentum because of powder loading causes a corresponding decrease in particle acceleration and heating rates. This effect is important, for instance, in the case of spraying of porous TBCs. This process is extremely sensitive to particle parameters. Jet deceleration and cooling due to powder overloading are the likely explanations of the fact, that at a certain point the coating weight stops increasing or even decreases with the increased powder feed rate, as is shown in Fig. 107, where the dependence of the weight gain at the test coupon for APS coated with ZrO2/Y2O3 vs. the powder feed rate is depicted.

4.7 Substrate Surface Temperature

As was already discussed the surface temperature of the substrate is one of the critical parameters for particle spreading and cooling, which determine coating properties. The local substrate temperature in the hot spot is determined by two factors. The first is the average part surface temperature, which depends on the process parameters and history, specifically on the gun power and standoff distance, duration of spray and component preparation. For the LPPS spray of MCrAlY coatings with component transferred arc cleaning and preheating this temperature can be as much as 800–900 °C in order to have the thinner splats (to reduce the coating porosity) and reduced residual stresses due to the effect of annealing. Heat transfer from the plasma jet may not be sufficient to compensate for radiative heat losses at such a temperature level (especially for large parts) and during thermal spraying, the component is cooled down. For the APS application of porous zirconia TBC usually the temperature is kept not higher than 150–200 °C with the opposite aim – to have higher porosity. To avoid the component overheating and the corresponding reduction of porosity a forced cooling from special air nozzles can be used. The second factor is the local reaction of the surface temperature to the passing hot spot. This depends on the heat flux from the plasma jet and on the traverse speed. When the hot spot arrives at this point, the surface temperature starts to increase. It reaches its maximum slightly after the moment of arrival of the jet center. This delay is caused by heat conductivity into the material. Then the temperature decreases. A typical temperature distribution at the symmetry plane of the hot spot is shown in Fig. 108. The amount of heat that is transferred from the plasma to substrate, depends on the plasma composition and the structure of the jet. In [145] some results from direct measurement of the heat fluxes have been presented for the different APS plasma guns and HVOF and a wide range of torch parameters, gas compositions and standoff distances. It was found that the heat flux into the substrate changes proportionally to the torch power and inversely proportional to the standoff distance squared (as it should be for the linear expanding jet). In Fig. 109 the jet and torch heat-transfer efficiency (the heat flux into the substrate related to

165

166

4 Theory and Physical Trends

Fig. 108 Calculated temporal temperature variation in the fixed substrate point under the moving spray spot. Parameters used: torch traverse speed 1 m/s; effective diameter of coating spot (a standard deviation of the Gaussian distribution) 7 mm; effective diameter of the hot spot – 15 mm; torch power 120 kW; particle enthalpy = 2.4 MJ/kg; jet plasma transfer efficiency 10%. The hot and spray spots are aligned. Time = 0 corresponds to the moment when the spot center passes the considered surface point.

the net plasma power and torch power, respectively) is presented for two types of plasma torches. It is seen that for the same gun it weakly depends on the gas composition. The heat-transfer efficiency of the PlasJet torch is higher than that of the F4, which should be attributed to the lower jet turbulence and expansion rate, respectively, as was already discussed in Section 4.3.1. The heat flux from the impinging supersonic vacuum jet depends much less on the standoff distance than in the case of APS. This is because of the quasilaminar character of the flow, due to which the jet has a lengthy part of almost constant cross section. Under low-pressure conditions a dissipation rate of the plasma energy in the surrounding atmosphere is much lower that at normal pressure. From the results of [145] one can draw an interesting observation: APS and HVOF coatings are typically sprayed at the distances that correspond to the jet heat-transfer efficiency of 10–15% (or 5–10% of the torch efficiency). The general trends of the surface thermal behavior follow from the basic principles of heat transfer: x Increase of the traverse gun speed leads to the reduction of the temperature peak x The magnitude of temperature increase in the hot spot is proportional to the heat flux from the jet x Width of the temperature peak is proportional to the jet width Loading of the plasma with powder causes redistribution of the heat fluxes. As was already mentioned up to 20% of plasma energy is transferred to the powder. What fraction of this energy is then released at the substrate surface depends on the deposition efficiency and powder-size distribution (one should remember that smaller particles are heated better due to the higher surface-to-volume ratio and,

4.8 Formation of the Coating Layer

Fig. 109 Efficiency of heat transfer into the substrate at APS process for F4 (lines 1–4) and PlazJet (line 5) guns with different plasma compositions and torch power: (1) Ar = 40, H2 = 8, 40 kW, (2) Ar = 30, H2 = 6, P = 30 kW, (3) Ar = 33, H2 = 3, P = 30 kW, (4) Ar = 36, H2 = 0, P = 27 kW, (5) PlazJet, N2 = 235, H2 = 94, P = 180 kW; gas flow rates are in slpm; (a) absorbed heat related to the gun electric power (Torch heat efficiency), (b) related to the plasma enthalpy (plasma heat efficiency) [145].

therefore, carry a disproportionately larger portion of heat). Usually, this leads to the increased surface temperature, but if the deposition efficiency is very low the effect may be opposite.

4.8 Formation of the Coating Layer

A typical metallic or ceramic coating on the turbine blade can consist of 10–15 subsequently applied layers (if the coating quality is good, they are usually indistinguishable at the cross-section). The definition of a mono-layer is a critical task of coating development. The thickness and microstructure of the coating layer depend on the shape of the spray pattern, distribution of heat fluxes in the spray hot spot, gun traverse speed and the powder feed rate. Let us discuss formation of the coating layer at the given surface point.

167

168

4 Theory and Physical Trends

When the hot spot arrives at this point, the surface temperature starts increasing. As was already mentioned in a previous paragraph the temperature reaches its maximum value slightly after the moment of arrival of the center of the hot spot. Afterwards the temperature starts decreasing. When the moving spray pattern arrives, the first come the peripheral particles of its front (leading) edge. As it was discussed in Section 4.5 particles at the periphery of the spray pattern have lower speeds and temperatures than those in the spot center. Also, the particles of the front edge meet the surface, which has not been fully heated by the upcoming plasma hot spot. Next come the particles from the plume center, which have the maximum temperatures and speeds, and finally come particles from the trailing edge of the spot that have parameters similar to ones of the front edge but they hit the surface, which has been fully preheated by the outcoming hot spot. Therefore, the layer consists of three sublayers with the gradual transitions between them. The first one has the poorest quality of all three, the second one has the maximum density and the third one is between these two. Also, other mechanical properties and residual stresses can be very different across the layer. If the traverse speed is increased the contributions of the second and third sublayers into the coating may also increase, so may the coating quality. But if the gun moves too fast the temperature effect of the hot spot vanishes and the coating quality can suffer. If the powder feed rate is too high the thickness of the first (worst) layer can become critically high, which could cause a failure of the entire coating [154]. Also in such a case a coating stratification becomes visible that is usually unacceptable in manufacturing. Another practical conclusion: in some situations, for instance, if the coating has to be sprayed at angles much less than 90 degrees due to geometrical restrictions, it is better to shift the spray pattern and make it trail the hot spot in the motion. From all these considerations it follows that there exists an optimum combination of the shape of spray patter, traverse velocity and the powder feed rate. These three parameters together unambiguously determine the layer thickness. Today it is impossible to make a comprehensive theoretical analysis of the interaction of these parameters. But it can be parameterized, for instance, with the use of a dimensionless parameter B = W U h Vp / G [154], where h is a layer thickness, G – powder feed rate, W – torch speed (relative motion or surface speed), U – powder density and VP – spray spot width. In a series of experiments the optimum coating microstructure has been achieved at B | 1.

4.9 Use of Different Plasma Gases

Helium and nitrogen are also frequently used in a thermal spray. The presence of helium, in binary Ar–He and ternary Ar–H2–He mixtures [155] leads to increased plasma heat conductivity and to the faster particle heating. Similar to hydrogen, helium increases the arc voltage and a plasma enthalpy, respectively. Arcs with He have higher temperatures due to its high ionization potential, but also higher plasma velocity. Use of nitrogen (for instance, in PlazJet£ spray gun)

4.10 Some Distinguishing Features of HVOF Physics

brings advantages of higher plasma enthalpy (due to high energy expenditures for dissociation and ionization) and reduced heat losses similar to hydrogen due to lower temperature, but also it is beneficial due to low cost of the gas. Air, pure oxygen and carbon dioxide as plasma gases are not widely used due to reduced lifetime of the electrodes. Use of water-stabilized plasma torches with consumable cathodes (graphite) and rotating, water cooled anodes is also very limited to high throughput ceramics.

4.10 Some Distinguishing Features of HVOF Physics

The temperatures of combustion products used in the HVOF gun are much lower than for the typical thermal-spray plasmas, whereas gas densities are much higher. These features determine the major difference between HVOF and plasma spraying: lower particle temperatures at the comparable torch heat power (usually between 1000–2500 °C) and higher speeds (up to 600–800 m/s). Therefore: 1. Transformation of the kinetic energy into the thermal one during particle impact and corresponding particle heating becomes an important mechanism of splat formation in case of HVOF, HVAF or cold gas spraying. 2. Significant fraction of particles in the case of spraying of metals is still solid (but softened!) at impact. 3. Particle velocity component normal to the surface is a primary factor of splat formation; therefore the coating is very sensitive to the spray angle. 4. Due to the high velocity the particle dwell time in the jet is smaller than in APS or LPPS. An HVOF gun generates a supersonic gas jet with a high dynamic pressure, which is quite stable and is able to propagate to large distances without destruction and mixing with the ambient air. Therefore, the efficiency of the heat transfer into the substrate is usually higher than for the APS at the comparable standoff distances [145]. Due to this a temperature in the hot spot can reach quite high levels (up to 800–1000 °C), whereas the hot gas jet is hardly visible (see Fig. 110). The particle flow is well centered in the jet. Correspondingly, a spray spot is relatively small and the coating build-up rate in the spot is high. A spray pattern is aligned with the hot spot.

Fig. 110 Typical “cold” HVOF jet with characteristic shock diamonds (from left) and spot of impacting particles (right).

169

171

5 Offline Simulation of a Thermal-Spray Process 5.1 Simulation in Production

At the moment, theoretical and computer models have been developed for most of the elements of thermal-spray processes, whether HVOF or plasma. Of course, the bulk of simulation work has been dedicated to the simulation of plasma/HVOF torches and to the particle-in-flight modeling [149, 156–160]. These models were based on the in-house and commercial CFD software packages of different levels of complexity and on the theoretical correlations for particle heating and acceleration in plasma flows. The goal of particle-in-flight modeling was to create tools that would be able to predict velocities, temperatures and trajectories of single particles injected into the plasma plume. Some of the models have been developed up to the level of software products and can be used as development tools for thermal spray engineers and operators [63, 161, 162, etc.]. Some of the organizations have developed their models up to the level of multiparticle simulation, often with the elements of stochastic modeling [126, 163]. From the comparison of theoretical predictions with the experimental results, which can be found in the literature, a conclusion can be drawn that at the current level of understanding of the thermal-spray physics and physical gas dynamics the models are capable of predicting particle-in-flight temperature and velocity within the wide range of process parameters. Of course, these models require a thorough experimental calibration and adjustment of empirical and semiempirical coefficients (which constitute an intrinsic part of any complex physical simulation). Also, numerous works have been dedicated to the simulation of a splat formation [164–172]. They qualitatively predict splat dimensions after the impact and describe such delicate physical phenomena as splashing and cracking, In some works [128, 161, 173–175] attempts to simulate a formation of the coating bulk and to calculate effective physical and mechanical properties have been made. Also, some coating features like residual stresses, coating porosity and coating texture were estimated [126, 128, 176, 177]. However, such a critical element of the coating process like particle bonding, still has no adequate theoretical models, nor reliable experimental description. Theoretical analysis of the particle bonding based on fundamental principles of surface kinetics, dislocation theory and surface activation energy can be found in

172

5 Offline Simulation of a Thermal-Spray Process

[117]. Significant advances in experimental investigations of the impact behavior of separate particles with the connection to their in-flight parameters should be mentioned [178–182]. Accumulation of such information at some point will definitely result in the appearance of a consistent theory. But till now no descriptive models can be derived that would predict a probability of a splat bonding to the surface depending on its impact speed and temperature. Nor can the intersplat cohesion be predicted. Developers of the full-scale simulation tools have to involve certain assumptions based on the empirical information (which is always hardware and powder specific) or even on physical intuition. The huge progress in available computational resources made possible development of the models, which allow not only predictions of the coating properties and the shape of spray pattern but also programming and visualization of the real 3D process [183–188]. By combining the physical torch model, particle-in-flight, splat and statistical coating formation models with the CAD and CNC/robot programming software, the full-scale offline process simulation tool can be created. In this chapter an example of a plasma spray modeling package is described, which has been developed at ALSTOM Switzerland in cooperation with CENIT. This package is currently used in ALSTOM coating manufacturing as a process development tool.

Fig. 111 Structure of the offline coating simulation package showing the different modules.

5.1 Simulation in Production

A structure of the package is schematically shown in Fig. 111. It comprises three major modules: 1. A physical module [126], which generates a 3D geometry of a spray pattern as well as a set of calculated parameters describing a coating quality. 2. A programming module, which has been developed based on the IGRIP software package. It includes: x A “painting” program that uses a spray pattern from the physical module as a paintbrush. The coating on a 3D surface is modeled by sequential overlapping of the spray footprints moving according to the control program. x Special tools that allow creation and usage of the CAD design models of sprayed blades and spray cells. x CNC/robot programming tools that transfer the “painting” algorithm into the files, which can be readily uploaded into the control system of the coating machine. Also, the existing manipulator program can be downloaded into the programming module and visualized in the IGRIP shell. A structure of the IGRIP programming module is presented in Fig. 112. An interface between programming and physical modules is realized by means of tables that contain matrices of calculated thickness and porosity distributions in the coating pattern.

Fig. 112 Structure of the IGRIP programming module (courtesy ALSTOM and Cenit).

173

174

5 Offline Simulation of a Thermal-Spray Process

3. A set of calibration models that are used for comparison of simulated and actual coating distributions. When the new process recipes and/or new manipulator programs are introduced, these models are programmed, downloaded into the machine and the corresponding shapes are sprayed and analyzed in the laboratory. All the modules are covered with the graphical user interface, which makes them suitable for users that are not aware of the internal structure of the package. Process engineers and development technicians are the intended users of the software. Besides the major modules, there exist so-called satellite programs, which are not included in the software installed in the shop. They allow development engineers and researchers to perform rapid studies of separate process elements. For instance, a separate 2D coating optimization tool can be used for modeling of the blade airfoil coating with the given spray spot shape and the given optimization function (e.g. uniformity of coating thickness). Such accompanying tools are necessary for the new technology developments, whereas they are not required for the daily manufacturing work and would overload the package in case of being fully integrated into the software. The package is based on the following philosophy: 1. The physical model must include as much physics as necessary and as little as possible. That means: x Everything that is measured during the process is not calculated in the software. Measured values are used as input parameter for the models, e.g. torch current, voltage and thermal efficiency. x Effects of physical mechanisms, which can not be accurately calculated or a priori measured (e.g. plasma turbulence or arc pulsations), must be described in a form that allows their calibration and tuning based on simple tests under the shop conditions. For instance, all stochastic mechanisms, which affect the particle motion, are described by a single two-parametric stochastic distribution function with the average and standard deviation derived from processing of a series of test spray patterns. 2. There should not be any physical models integrated into the painting program nor in CNC/robot programming tools. The latter should be responsible solely for the geometrical modeling. The painting program uses extensive multidimensional data tables, which describe a coating build-up rate and the porosity of the spray spot versus standoff distance, spray angle and the twocoordinate position at the spray plane, Each specific data point is determined by the linear interpolation from the table. Tables are generated in the physical module for each specific set of physical spray parameters (spray recipe). The painting program operates with a library of such “paintbrushes”. Later in this chapter the basic physical principles behind the software will be briefly outlined to show the possible direction of further coating-quality prediction tools.

5.2 Physical Background of Simulation Package

5.2 Physical Background of Simulation Package 5.2.1 Viscoplasticity Model of a Splat and Particle Bonding

During the impact on a substrate the particle is typically totally deformed, rapidly cooled and solidified. It is a commonly accepted approach to consider the particle deformation only until the moment of complete solidification. But quite often, especially for HVOF and similar processes, particle deformation continues after the particle becomes solid. A typical value of the specific kinetic energy of the particle is of the order of 100 MPa (MJ/m3), whereas the plasticity limit at a temperature slightly below the melting point can be much lower. In this situation viscoplastic flow is realized. For instance, due to the low values of Prandtl numbers metal particles are cooled down and solidified before viscous forces can dissipate the kinetic energy. Moreover, usually there is a large quantity of partially melted and unmelted but softened hot particles in the flow, which are still able to form the coating. Thus, at least for spraying of metals, a plastic deformation of the splat after solidification can be crucially important for splat formation. An approximate viscoplastic splat model [126] has been integrated into the software package. Theoretical flattening ratios of MCrAlY particles are shown in Fig. 113 for different particle and substrate temperatures. As was discussed in Section 4.1 particle bonding is a random event with the probability depending on: x Particle impact temperature and pressure; the latter depends on the particle velocity, impact angle and material density. x Surface conditions such as surface temperature, roughness and a surface crystalline structure, e.g. density of dislocations. An average probability of particle bonding determines the process deposition efficiency for the given set of spray parameters and the given powder. In the offline simulation tool, a one-parametric model of particle bonding is adopted.

Fig. 113 Relative diameter of NiCrAlY splat vs. impact speed for different initial particle and substrate temperatures; initial particle diameter 35 Pm (courtesy of ASM).

175

176

5 Offline Simulation of a Thermal-Spray Process

It is based on the assumption that the probability of bonding correlates to the particle flattening ratio since both of them depend on the particle impact velocity and temperature. It is postulated that there exists a critical flattening ratio, which is a threshold of bonding. If particle impact parameters suffice to assure this critical flattening or higher, the particle is stuck on the surface and vice versa. It is impossible to derive a practically usable value of this criterion quantitatively from any of the existing theories. It has to be obtained by adjusting a spray model to the experimental results. A range of its applicability can be determined by comparing the trends predicted by the model with the actual spray experience. One example of such validation will be presented in Section 5.3.1. 5.2.2 Thermodynamic and Transport Properties of Argon/Hydrogen Mixtures

Based on well-known theoretical models [136–139] a software module has been developed that calculates properties of the working medium. Argon and hydrogen specific physical atomic and molecular parameters have been taken from [189–191]. Radiative data have been interpolated from tables [141]. With the gas mixture composition as input parameter detail tables of all relevant properties are calculated prior to the CFD calculations. Plasma internal energy and density are used as independent variables. In the CFD algorithm the necessary thermodynamic and transport parameters are linearly interpolated based on the tabulated data. Such an approach dramatically reduces the overall computational time. 5.2.3 Modeling of the Plasma Gun

Electrical parameters of the plasma torch are not calculated in the model but taken from the measured data. Plasma parameters are calculated with the use of energy and mass conservation laws with the given pressure (APS nozzle exit) or the sonic flow condition (LPPS gun critical section) as closing relationships. A diverging part of the LPPS nozzle is modeled with the use of quasi-1D gas dynamic equations [140]. Plasma swirl is calculated from the momentum conservation law based on the gas parameters and geometry of the distribution system in the torch chamber. 5.2.4 Modeling of the Plasma Jets

Usually modeling of plasma jet is carried out with the use of sophisticated commercial or in-house CFD tools. The major idea of the offline simulation package is to bring the modeling into the shop. Therefore, no software can be used, which requires a specific knowledge of fluid dynamic or numerical techniques from the shop user. A software input should contain only the process parameters and minimum necessary geometrical data.

5.2 Physical Background of Simulation Package

5.2.4.1

APS Jet

An approach of the isobaric turbulent jet has been used in the simulation of an APS plasma plume. A simplified semiempirical method [192] was realized based on the theories of boundary layer and free turbulence. Two parts of the jet were modeled: the initial one, which contains the bright quasilaminar cone, and the main one, which was considered fully turbulent. In the bright cone all components of the plasma velocity were assumed to be constant, whereas the temperature decreased due to radiative losses. A layer between the cone and the ambient atmosphere was considered according to the theory of the initial part of turbulent jet [192]. This layer is bounded by the converging and diverging conical surfaces. The main part was assumed to expand linearly with the expansion angle taken from the measurements [145]. For the APS Ar/H2 jet from F4 gun this angle appeared to be approximately 15 degrees [145] that is typical for conventional turbulent jets. An empirical law “power 3/2” [143] for the velocity and temperature profiles in the jet has been used. With the use of those profiles, mass energy and momentum conservation equations have been solved. 5.2.4.2

LPPS Jet

Two approaches have been used for simulation of a supersonic LPPS jet. The first one was a simplified model, which considered a supersonic flow comprising the first shock diamond and the main equilibrium isobaric part of the jet. Parameters of the first shock diamond were calculated using an approximate solution [148]. Parameters of the equilibrium jet were determined from the momentum and energy conservation laws. Criterion (12; Section 4.3.1) of the laminar-to-turbulent transition has been used. Downstream of the point of turbulent transition a model of the turbulent isobaric jet has been used. Such an approach allowed rapid estimations of all physical parameters required for the particle-in-flight simulation. The second approach included numerical solution of the complete set of 3D axisymmetrical hydrodynamic equations. Unlike a simplified approach, the full CFD model allows the flow structure and particle interaction with the shock waves to be understood in detail. Such modeling becomes extremely useful when the shock structure has a significant influence on the particle acceleration and heating. Nonetheless, it has serious disadvantages: x CFD simulation of the jet destruction is extremely difficult and would require implementation of the sophisticated turbulence modeling. Validation and calibration of such models is a serious challenge for the researchers, especially keeping in mind a critical lack of the experimental data on the turbulent transition in plasmas. x Such models usually contain some parameters that have no physical meaning; nonetheless they are required to ascertain stability and convergence of numerical algorithms. Therefore, the model cannot be fully controlled by the production personnel and requires supervision from the developers or qualified users. This would violate an original idea of the offline production tool for the shop.

177

178

5 Offline Simulation of a Thermal-Spray Process

x Computational time is significantly higher than that for the simplified model.8) This is a typical dilemma between simplicity of obtaining the major process parameters and the deep understanding of the process details (see, for instance, [193]). In our case a compromise solution was to use the simplified model for the process simulation and to employ the full CFD algorithm for double checking and adjustment of the results when significant changes in the process parameters took place. The ability of the CFD algorithm to model a sophisticated shock-wave structure has been checked by comparing with available experimental data [194]. In Fig. 114 a pressure distribution downstream of the supersonic nozzle is compared with the measurements.

Fig. 114 Pressure distribution in “shock diamonds” of the supersonic jet; M = 2, exit pressure ratio 2.5; comparison with experiment [194] (courtesy of ASM).

Fig. 115 Typical LPPS plasma parameters in the plasma plume impinging a flat surface; chamber pressure 3.5 kPa, specific plasma enthalpy 15.5 MJ/kg; Ar/H2 mixture (courtesy of ASM). 8) It is worth noting that the computational grid cannot be controlled by the production personnel. In such a situation, the simple way to assure robustness of the numerical algorithm to all imaginable initial conditions and model parameters is to allocate a computational domain and generate a grid with the multifold excess to the minimum necessary requirements. But this approach is extremely consuming from the standpoint of computational resources.

5.2 Physical Background of Simulation Package

Typical temperature and speed distributions in the LPPS plasma jet impinging the wall are presented in Fig. 115. The working mixture was 90%Ar/10%H2, the heat input into the plasma was 15.5 MJ/kg, the working pressure in the chamber was 3.5 kPa. The calculated pressure in the nozzle critical section was about 30 kPa, and the plasma speed at the gun exit corresponded to M = 2. 5.2.5 Acceleration and Heating of Particles in Plasma

A multiparticle approach has been used. The whole ensemble of particles has been split into groups according to their diameter, injection speed and injection angle. A population of each “cell” of such a 3D matrix was calculated according to the particle distribution functions. A size distribution has been given by the actual powder specification. A mean value of injection speed distribution has been taken based on the calibration experiments with the spray patterns. A scatter of injection speed was assumed to range from zero up to the double mean value. A maximum size of the particle injection “cone” was assumed to be 20 degrees. Trajectories and heating have been calculated for each group of particles separately. As was discussed in Chapter 4 a plasma/particle momentum and heat exchange in the free LPPS jet has to be modeled using an approach of rarefied gas. Values of particle drag coefficients [195] calculated with the effects of drag reduction under the plasma conditions taken into account are used. Heat fluxes are calculated with the use of a semiempirical theory [196] for rarefied supersonic flows. For the APS modeling correlations [197] and [153] have been used. In order to verify the model a series of experiments has been carried out in the LPPS facility of EMPA Laboratory (Switzerland) with the use of a DPV 2000 measuring device. NiCrAlY powders of 30 Pm mean diameter were used. The working pressure varied from 2 to 14 kPa and the plasma gases were injected into the torch chamber with a tangential swirl of 45 degrees. Comparative results are presented in Figs. 116 and 117 [126]. It is seen that the model gives a good agreement with experiment on particle speeds and temperatures. In the experiments, clusters of particles were well aligned in the plasma plume and most of the particles were heated close to the boiling point (approximately 2200–2300 °C at the reduced pressure) before the sensor was able to detect them. Random variations of the particle in-flight velocity contribute to the scatter of their trajectories and to the shape and size of a spray pattern, respectively. In order to take into account stochastic fluctuations of the aerodynamic forces they have been considered random variables with the mean values given by the aforementioned correlations. A scatter (standard deviation) was a parameter to be adjusted to the calibration tests. It should be emphasized that the latter variable is not a real physical characteristic of the process. It reflects contributions from, at least, three stochastic mechanisms such as: plasma fluctuations due to arc pulsations, plasma turbulence and uncertainty of the aerodynamic forces due to the variation of powder shape. Since all these mechanisms affect the particle motion in a similar way and no accurate theory is available for any of them, it is

179

180

5 Offline Simulation of a Thermal-Spray Process

Fig. 116 Model validation and calibration – particle temperature in LPPS for different working pressures; solid lines – calculations, markers – experiment (courtesy of ASM) [126].

Fig. 117 Model validation and calibration: particle speed in LPPS for different working pressures; solid lines – calculations, markers – experiment (courtesy of ASM) [126].

impossible to separate their contributions. Nor it is necessary for the purpose of the modeling. A stochastic scatter of the particle temperature due to plasma fluctuation contributes to the variations of probability of particle bonding to the substrate. This mechanism has been already taken into account by introducing such a parameter as an average probability of the particle bonding, which is to be adjusted to the test spray results. Thus, no stochastic modeling of heat fluxes was required. 5.2.6 Surface Thermal Conditions

For LPPS a heat flux from the plasma jet into the substrate has been calculated with the use of a theory of heat transfer in impinging jets [198, 199]. In the case of APS, experimental data have been used in the 3D heat conductivity calculations as boundary conditions as was described in [145]. In order to verify the model, temperature field measurements have been performed on a 50 × 100 mm2 steel

5.2 Physical Background of Simulation Package

Fig. 118 Surface temperature distributions from thermal paint measurements (a) and from 3D calculations (b); temperature isolines and threshold paint discoloration temperatures are shown; an arrow indicates the position of the spray gun and direction of its movement. Temperature in °C, dimensions of plate size are 100 × 50 × 2 mm3 (courtesy of ASM).

plate of 2 mm thickness at one pass of the F4 torch. The plate has been marked with thermal paint stripes that change their colors after exceeding a threshold of the specified temperature. The torch electrical power was 27 kW, standoff distance 100 mm, robot speed 20 mm/s. From the discoloration of stripes due to plasma heating, as is presented in Fig. 118a, one can deduce some information about the maximum surface temperature during the process. In Fig. 118b the calculated temperature field for the same setup is presented. Comparing Figs. 118a and b one can see that the major features of the temperature distribution are well reproduced by the calculations both qualitatively and quantitatively. Use of experimental

181

182

5 Offline Simulation of a Thermal-Spray Process

data of heat transfer allows for the calculations of surface temperature at APS conditions without incorporating uncertainties resulting from the modeling of plasma turbulence and radiation. Knowledge of heat flux distribution allows the temperature increase in the hot spot to be calculated. In general, the average part surface temperature should be modeled as well. This would require additional models of surface radiation for LPPS and convective cooling for APS. In our case an average part temperature was known from measurements and this information was used directly in the calculations.

5.3 Spray Pattern 5.3.1 Calibration of the Bonding Model and Sensitivity of a Spray Pattern to the Process Parameters, Spray Angle and Bonding Model

A rotating plate has been sprayed with MCrAlY powder with the use of the twopeak spray pattern generated by double injectors mounted each at opposite sites at the exit of the LPPS spray gun. In Fig. 119 a measured thickness distribution is compared with the theoretical ones. Two models of particle bonding have been used in calculations. In the first one a probability of particle sticking at the surface was assumed to be the same for all particles and was adjusted to the overall powder deposition efficiency, which was approximately 80% at the flat immovable plate.

Fig. 119 Comparison of the measured coating distribution (cross section perpendicular to the rotation axix) with simulation results for two bonding models; flat rotating plate sprayed by LPPS with double powder injection (location 17 = 150 mm).

5.3 Spray Pattern

Fig. 120 One-factor-a-time sensitivity of spray spot to process parameters; (a) base shape, spray angle 90 degrees, (b) carrier gas flow rate increased, (c) gas swirl increased, (d) flow pulsations reduced, (e) spray angle 45 degrees, (f) powder melting temperature increased.

In the second one a criterion of the critical splat flattening ratio was used. The value of the criterion was adjusted to the overall deposition rate and appeared to be approximately 4.5. One can see that the averaged model is not able to explain the rapid decrease of the coating thickness towards the edges of the coating, nor can the shape of a “saddle” be represented, whereas the “critical flattening” hypothesis fits very well into the picture. Let us illustrate some capabilities of a model by the example of spray spot variations. In Fig. 120 the sensitivity of the spray pattern to the one-factor-atime variation of the major LPPS parameters is shown. The following cases are considered:

183

184

5 Offline Simulation of a Thermal-Spray Process

a) The basic shape, which was created by the double-injector LPPS spraying, injector positions (looking in the flow direction) and direction of swirl are shown. The pattern consists of two overlapping spots, twisted clockwise. The lower spot of footprint corresponds to the powder flow discharged from the right injector. b) Increase in carrier gas flow rate leads to the rotation of the pattern in the clockwise direction and broadening of the spots; c) Increase of gas swirl in the nozzle leads to the increase in separation between the spots and a reduction of spot overlapping; therefore, the peak height decreases. d) Flow pulsations are responsible for the size of the spot and the sharpness of peaks correspondingly. Reduction of simulated stochastic flow oscillations leads to “focusing” of particle trajectories and smaller separated spray spots. e) Reduction of the spray angle affects the periphery of the pattern; deposition efficiency in the central part of the pattern stays almost unchanged, whereas the marginal areas disappear; peak heights reduce as the sine of the spray angle. f) Increase of the powder melting temperature leads to the shrinkage of the spot periphery due to reduction of deposition efficiency. Particles that fly close to the jet axis, have still sufficient dwell time to be heated and the height of the pattern hot area stays almost unchanged. An example of the model validation is presented in Fig. 121 where theoretical trends of the coating spot thickness vs. the carrier gas flow rate are compared with the test results for two spray recipes.

Fig. 121 Thickness of LPPS spray spot vs. carrier gas flow rate for two sets of process parameters.

5.3 Spray Pattern

5.3.2 Coating Porosity and Roughness

A statistical model of the coating formation from separate particles was used. In the model, splat dimensions were obtained from the viscoplastic model. A kinematic approach was adopted to describe particle spreading over the rough surface with no detailed dynamic aspects taken into account. Basic kinematic assumptions are similar to [173]. Porosity formation is the most difficult issue to model. A review of possible mechanisms of porosity could be found in [172]. In the present model a splat detachment at the edges of underlaying particles is considered the major source of porosity. Such porosity is assessed assuming that a void has two dimensions of the order of the splat thickness and the third one of the splat circumference. For the average porosity this gives an estimate of p ∝ 5/ D3 where D is the mean degree of splat flattening (the final splat diameter related to the initial particle diameter). For ceramics ( D ≈ 4.5 ) this gives a value of about 5 per cent (compare, for instance with [200]). Submicrometer intersplat flaws and intrasplat cracks were not considered in the simulation. In Fig. 122 estimated porosity distributions in the LPPS MCrAlY and APS ZrO2 spray patters are compared with the experimental values measured under a microscope at the longitudinal cuts. One can see that qualitatively the agreement is quite good, whereas absolute figures are different, especially at the periphery. Simulation of the coating surface structure was not part of the manufacturing software package. It was used as a satellite tool in order to determine a process operational window with regard to spray angles and to determine the conditions of formation of so-called overspray structures. In Figs. 123 and 124 experimental and numerical results on the coating formation for two values of the spray angle are presented. It can be seen that the model gives correct predictions of the surface texture. Statistical scattering of splat impact positions is a major driving force of surface roughness. The roughness increases approximately as the square root of coating thickness. Within the impacting angle range of 60 to 90 degrees the roughness has little sensitivity to the spray angle. Below 60 degrees it grows quite rapidly and further decrease in the angle leads to the appearance of an overspray structure similar to one shown in Fig. 124. The cause is the shading effect, which provokes instability of the formation of the coating surface. There is a threshold angle of instability that varies between 35 and 45 degrees depending on the splat flattening. Thicker splats provoke earlier instability. Deflection of “protuberances” from the vertical is due to the horizontal component of the particle velocity. It has been found that the coating has a “memory” and disturbances such as substrate roughness or overspray are traced in the structure of the overlaying coating. In should be emphasized that the models of porosity and surface texture described above are just advanced estimates, not the detailed theories. In general, results can give directions on how to improve the process, but no exact figures can be expected. The same is valid for the models of other coating characteristics, like oxidation or mechanical properties.

185

186

5 Offline Simulation of a Thermal-Spray Process

Fig. 122 Comparison of modeled and measured porosity distributions in LPPS (a) and APS (b) spray spots.

Fig. 123 Real (top) and simulated (bottom) coating microstructure at a spray angle of 90 degrees (perpendicular particle impact).

5.4 Modeling of Turbine Blades

Fig. 124 Real (top) and simulated (bottom) coating structure at a spray angle of 30 degrees (flat impact angle) (courtesy of ASM).

5.4 Modeling of Turbine Blades

The simulation package has been calibrated by comparing calculated and sprayed coatings with a set of simplified shapes, which include all geometrical elements critical for spraying. An example of such a model is shown in Fig. 125. A 120-degree angled plate has been used to develop an algorithm for coating of platform–airfoil transitions. After calibration of all the elements of the simulation package, it has been validated in the coating shop with standard test configurations as well as with real turbine components. CAD component models have been taken from the design office. Coating-cell models have been generated using the CATIA® package.

Fig. 125 A typical model shape for validation and adjustment of coating software; yellow color – maximum coating thickness, dark blue color – minimum thickness; a plate has been virtually “sprayed” with the double-injector LPPS system (courtesy of ALSTOM).

187

188

5 Offline Simulation of a Thermal-Spray Process

The CNC programs have been generated with the use of IGRIP software and after debugging uploaded to the coating machine control system. The programs have been checked element-by-element with the real turbine blades installed in the machines and the parts sprayed. A typical calculated coating distribution on the pressure side of turbine blade is presented in Fig. 126. The blade has been “virtually sprayed” by combined rotation of the component and translation movement of the torch. Comparisons between the simulated thickness and measured distributions are shown in Fig. 127. In the lab the destructively measured coating thickness has been defined as a distance between the surface of blade base material and bottom of the coating roughness, whereas the model produced the mass-average coating thickness. This is why measured average half-heights of roughness are also shown on the simulated curves. Agreement between the theoretical and actual distributions with the roughness taken into account is quite good at the airfoil and worse at the transition to the tip shroud. There are two reasons for the disagreement in the latter case. Location 3 in Fig. 127b corresponds to the transition radius. A local increase of the coating thickness there was caused by overlapping of the two gun passages, the former of which was to coat the airfoil and the latter targeted the shroud inner surface. Also, because of the computational limitations the mesh at that location was too coarse and a thickness variation was not properly reflected in the computed results. In the locations 4 and 5 the measured coating thickness was lower than expected due to the increased roughness, which was caused by the effect of an unfavorable spray angle. Apart from these two sources of disagreement, it was found that the modeling error was generally higher at the blade trailing and leading edges. Also, the comparison was not stable at the locations of the high thickness gradients, mostly due to the uncertainty of the cut positioning in the destructive metallurgical evaluation. This effect is discussed in more detail in Section 3.2.4. At the moment several types of turbine blades and vanes have been “coated” offline prior to the real coating cell tests that allowed reduction of the overall development effort by at least 50–70%.

Fig. 126 MCrAlY coating thickness distribution on the pressure side of a turbine blade; part of the tip shroud has been masked; red – maximum thickness, blue – minimum. Dashed lines show positions of the control cuts (courtesy of ALSTOM).

5.5 Coating Thickness Optimization and Stochastic Modeling Tools

Fig. 127 Simulated and measured coating thickness distributions at the blade air foil (a) (locations (1, 13) trailing edge, (6, 7) leading edge, (1–6) pressure side, (7–13) suction side) and at the airfoil/tip shroud transition (b) (locations (1, 2) airfoil, (3) transition radius, (4, 5) tip shroud). Positions of the cuts are shown in Fig. 126 (courtesy of ALSTOM).

5.5 Coating Thickness Optimization and Stochastic Modeling Tools

Certain conceptual elements of the coating development can be simulated with the use of simplified models. 2D-thickness simulation and stochastic modeling tools comprise programs that calculate a coating distribution over 2D profiles (e.g. airfoils, cylinders, corner configurations, etc.) using simplified assumptions about the shape of the spray pattern and movement of torch. An optimization function is applied that characterizes the quality of the thickness distribution. In our examples it will be an RMS (root mean square) deviation of the thickness from its average value. In the stochastic modeling random deviations of the spray torch positions and/or of the spray intensity from the given control law are assumed. Let us illustrate a usage of those tools by the example of APS and LPPS of the typical turbine blade airfoil. Three methods of torch positioning already mentioned in Section 2.8.1 were simulated (Fig. 128):

189

190

5 Offline Simulation of a Thermal-Spray Process

a)

b)

Fig. 128 (a) Methods of coating application and (b) shapes of spray patterns for a blade in atmospheric and vacuum plasma spraying at the spraying standoff distance.

1. “Painting”, which assumes a subsequential positioning of the relatively small (compared with the profile chord) spray pattern around the airfoil with a constant powder feed rate. This method is most frequently used in APS; a spray-pattern offset is an optimization parameter. 2. Rotation spray with the broad spray pattern, which is typical for LPPS. In this case the torch is positioned stationary at a certain standoff distance (in our examples it is equal to 4 chord widths of the airfoil), whereas a profile

5.5 Coating Thickness Optimization and Stochastic Modeling Tools

Fig. 129 APS relative coating-thickness distribution according to the spots described in the insert for different offsets of spraying passages.

is rotating. The number of torch stationary points was varied from 1 to 4. The parameters to optimize were the positions of each point with respect to the airfoil chord and the spray intensity at each point. Single- and double-injector spray patterns were considered. 3. A combination of the rotation spray with several spot corrections (i.e. 2D projections of the translational torch motion along the airfoil span). In this case the positions of stationary spots around an immovable airfoil and their spray intensities are the additional optimization parameters. APS coating distributions simulated for three values of spray passage offsets are shown in Fig. 129. One can see that the thickness is rather homogeneous but in the areas of high curvature it is quite sensitive to the variation of offset. Close to the trailing and leading edges as well as on the pressure side there are areas of significant thickness variation. In practice, this effect is usually corrected by adding or eliminating several gun passages from the robot program. Analysis of sensitivity to the mutual positioning of the gun and component is necessary to define the requirements of the gun fixation on the robot/CNC and to the part fixtures. Sensitivity of the coating thickness to the misalignment of a spray gun has been studied. In APS stochastic modeling it was assumed that the spray gun could randomly deviate from the prescribed trajectory within the range of r1 mm. The corresponding scatter of the thickness is shown in Fig. 130. It is seen that a misalignment of 1 mm could result in a deviation of the coating thickness by more than 10 per cent. In the locations of high sensitivity, like a concave area at the pressure side, this scatter could be even higher. An extensive optimization and sensitivity analysis has been carried out for the LPPS process with the large spray pattern. Calculated distributions of the coating thickness from the nonoptimized rotation spray with the single injector powder feeding are shown in Fig. 131 for the different positions of the spray gun with respect to the profile chord. One can see that the thickness variation can reach

191

192

5 Offline Simulation of a Thermal-Spray Process

Fig. 130 Sensitivity of the APS relative coating thickness to the torch misalignment of 1 mm; thickness distribution for the “ideal” case of perfectly positioned gun, local thickness variations due to stochastic gun deflection and the statistical average values are shown.

Fig. 131 Coating distribution after the single injector nonoptimized rotation LPPS spray for the gun position above 33, 50 and 67% of the chord. Coating thickness related to its average value.

Fig. 132 Optimized LPPS coating showing the influence of rotation and number of injection ports. Coating thickness related to its average value.

5.5 Coating Thickness Optimization and Stochastic Modeling Tools

Fig. 133 Spray positions and fractions of the coating material supplied for the optimized combined LPPS double-injector spray.

60% of the average value. An optimization has been carried out for the different spray methods. The number of gun stationary points and their positions with respect to the airfoil as well as spray intensity at each gun positions were varied. Results for the three-positional combined rotational/translational spray with three positions of spot corrections are presented in Fig. 132. Corresponding locations of the torch and distribution of spray intensities are shown in Fig. 133. A 3+3 position combined spray program results in a very uniform coating with the thickness deviation less than 10%. Single and double injector coating patterns produce coatings with comparable thickness quality. A random position deviation of maximum of r5 mm has been simulated. In Fig. 134 scatters of 1- and 3-position single injector rotational sprays are shown. In fact, a coating from the 3-position spray is more sensitive to misalignment due to the effect of overlapping. But the double-injector spraying gun and the combined spray allow reducing significantly the thickness scatter, as is shown in Fig. 135.

Fig. 134 Sensitivity of the 1- and 3-position rotation spray to the 5-mm random deviation of the gun position. Coating thickness related to its average value.

193

194

5 Offline Simulation of a Thermal-Spray Process

Fig. 135 Process scatters of the single- and double-injector rotation and combined sprays. Coating thickness related to its average value.

An optimization diagram is a convenient tool to visualize simultaneously an achievable level of thickness uniformity and a coating sensitivity to misalignment. An example of such diagram is shown in Fig. 136. Each X/Y point of the chart corresponds to the fixed values of the gun position and powder spray rate at the first stationary location. All other parameters (for instance, gun position and spray rate at the second location) are optimized in order to achieve a minimum possible RMS (root mean square) deviation of the thickness. The latter is represented on the diagram with the use of a color scale. The depth and the size of the “pit” characterize the level of achievable coating quality and its sensitivity to the positioning and spray rate variation. From comparison of dimensions of the green areas in Fig. 136 it is seen that the sensitivity of a single-injector spray is higher

Fig. 136 Optimization diagrams of the single- and double-injector rotational spray.

5.6 Simulation of HVOF Process

compared to that with the double-injector pattern, whereas the achievable levels of the coating uniformity are similar. Overall, using the stochastic coating modeling it is possible to define the process that would be optimal both from the standpoints of coating quality and its robustness to the process variations.

5.6 Simulation of HVOF Process

In general, HVOF process modeling is very similar to the LPPS simulation. A subsonic flow inside the gun and a supersonic free jet loaded with the particles are described by the system of two-phase fluid dynamic equations, which is closed with an appropriate equation of state for the working medium. Nonetheless, there are some differences to be mentioned: 1. Internal powder injection with the relatively long particle dwell time inside the torch chambers and long nozzles and/or barrels are distinctive features of the HVOF technology, which impose increased requirements on the modeling of the internal gun gas dynamics and particle motion and heating. Depending on the geometry of the gun interior, a 2D-axisymmetric or even 3D CFD modeling may be necessary. Calculation of particle transition from the subsonic to supersonic parts of the flow may create additional computational problems. 2. Unlike the plasma torch, heat losses in the HVOF gun can not be considered concentrated in a single location because of the presence of long nozzles and barrels. More detailed analysis of the gun heat balance is required in order to calculate accurately particle heating at the different parts of their trajectories. Particle motion and heating in the high-pressure supersonic jet should be considered within the model of continuous medium (unlike the LPPS jets), typical Reynolds numbers with respect to particles vary from 101 to 102 in the subsonic part of the gun up to 103 in the supersonic barrel and free jet. Therefore, models of particle acceleration and heating are different from those used in plasma spray. One can find corresponding correlations in [201–209]. 3. The thermodynamic behavior of the combustion products is, usually, simpler than in the case of plasmas since no effects of ionization are to be considered (but vaporization in case of kerosene and dissociation of combustion products may be still important). Gas properties frequently can be calculated using conventional expressions for the ideal gas. Usually, energetic contributions of chemical reactions may be neglected as well. But in the case of nonstoichiometric fuel-rich mixtures (reducing atmospheres), consideration of the afterburning in the free jet may become necessary. 4. The spray spot is relatively small and the spray pattern is well aligned with the hot spot on the substrate. This simplifies selection and analysis of the spray

195

196

5 Offline Simulation of a Thermal-Spray Process

pattern. Also, issues of surface shading, overspray and overlapping of passages at radii and corners are not that pronounced as in the case of large LPPS patterns. This may make unnecessary usage of the integrated offline software, since the gun trajectories can be easily generated by most commercial offline robot programming tools. On the other hand, due to the smaller spray spot, the coating becomes very sensitive to the gun misalignment and the role of the stochastic sensitivity modeling increases. Examples of the HVOF modeling packages can be found in [64, 210–218]. The main assumptions and features of the model [64] are the following: x x x x x

Axisymmetric gas flow Gas assumed to be ideal Chemical transformations of the combustion products are negligible Fragmentation and agglomeration of particles are negligible There are no interactions between particles and barrel walls

The software has the following main input parameters related to the spraying process: type of fuel, fuel and oxygen flow rates, the torch thermal efficiency, properties and size of the spraying particles, spraying distance, carrier gas type and its flow rate, and powder feed rate. Besides, the software also has input parameters related to the features of HVOF torches design: throat diameter, barrel length and diameter, geometry of the barrel expansion, position and diameter of powder injectors, etc. Thus, the simulation package allows not only the optimization of spraying parameters but also the optimization of HVOF torches design. A software package consists of three major modules: 1. A numerical core comprising: – Navier–Stokes equations solver – Particle-momentum and heat-transfer solver – Models for particle and jet interactions with the substrate 2. Database module comprising thermodynamic properties of combustion gases and sprayed materials as well as determining features of the HVOF torch designs. 3. User interface module. A typical geometry of HVOF torch, which was analyzed with the use of the model [64] is shown in Fig. 137. A radial method of powder injection into the low-pressure zone of the nozzle was chosen to avoid disadvantages of the high-pressure powder feeders, which are used in case of axial powder feeding. The torch geometrical parameters were the following: straight burning chamber, throat diameter Dt = 6 mm; diameter of the barrel exit Dbe = 10 mm; barrel length Lb = 100 mm. Liquid fuel (kerosene) was considered as fuel, with the oxygen/kerosene mass flow ratio mo/mk = 3.5. In Fig. 138 the spatial distributions of gas axial velocities and temperatures are shown for two values of combustion pressure. A shock diamond structure is clearly pronounced, which becomes more intensive with the increased pressure. Typical radial distributions of particle velocity up and

5.6 Simulation of HVOF Process

Fig. 137 Geometry of HVOF torch used for calculations.

Fig. 138 HVOF flow structure: (top) gas axial velocity and (bottom) temperature distribution for combustion chamber pressures of a) 0.7 MPa and b) 1.3 MPa, respectively (courtesy of ASM).

197

198

5 Offline Simulation of a Thermal-Spray Process

Fig. 139 Typical radial distributions of particle velocity and temperature (Dp = 30 Pm); 1) at the barrel exit 2) at standoff of 125 mm 3) at 250 mm (courtesy of ASM).

Fig. 140 Effect of powder injection velocity on mean values of particles velocity and temperature at the barrel exit (courtesy of ASM).

5.7 Use of Offline Simulation in Coating Development

temperature Tp at the barrel exit as well as 125 and 250 mm downstream of the barrel exit are illustrated in Fig. 139. Calculations were done for NiCr particles of average diameter of 30 μm. As an example of sensitivity simulation an effect of particle injection conditions on their exit parameters is shown in Fig. 140.

5.7 Use of Offline Simulation in Coating Development 5.7.1 Application Areas of Modeling in the Coating Process

Let us consider possible applications of different coating-process simulation tools. 5.7.1.1

Coating Definition and Design for Coating

x Feasibility of achievement of specified coating quality distribution at the designed part can be checked at an earlier design phase by means of the offline simulation tool. Based on this, an optimization of the specification can be carried out. Also changes in the design (for instance geometry of the shroud, modification of the blade cooling, etc.) can be proposed. x Stochastic modeling of the coating thickness distribution can give a good idea about the robustness of the proposed coating to the process variations and about the expected process capability. x Realistic assumptions about the coating distribution can be made for the mechanical integrity analysis. x Full-scope modeling can help to estimate a priori the coating costs and, therefore, can affect a selection of the technology, hardware and supplier. Being carried out at the vendor’s site, such a modeling can help with the realistic quotation of coating offers. 5.7.1.2

Coating-Process Development

It is difficult to overstate a potential contribution of physical spray models in the selection of coating parameters. Information obtained from such modeling can be used in: x Powder size selection x Definition of the torch parameters like flow rates, powder and nozzle geometry x Selection of the chamber pressure in case of LPPS process x Selection of the powder feed rate x Definition of a coating mono-layer and a spray pattern x Definition of the test trial matrix x Modification of the process parameters based on the test trials x Robustness check of the selected set of coating parameters to the possible process variations

199

200

5 Offline Simulation of a Thermal-Spray Process

5.7.1.3

Part Development

The step of coating development is the most resource consuming. Therefore, benefits from the introduction of process simulation tools are also most pronounced. Generation of the program for the coating machine based on the predicted coating thickness distribution is the major goal of the offline simulation. It is difficult to overstate the benefits of taking the part development offline: x Operational time of the coating cell, required for the part development online, could reach as much as 10–15 work shifts if all necessary iteration loops are considered. By using the offline programming and coating visualization, this time could be reduced by factor of 2 or 3. Money wise, this could result in 50 000 to 100 000 US$ of reduction of direct development costs for one new turbine blade. x Reduction of out-of-production time and related loss of business brings significant savings. x Usually, a part development and the regular production run in parallel on the same machine, which required additional readjustments of the hardware and controls. Generally speaking, this is in principle a contradiction to the philosophy of a statistical process control and of the six-sigma approach, which requires all intervention into the process to be ceased as soon as the process capability is established and the process is frozen. x A multivariant optimization of the coating algorithm becomes possible offline. Quantitative comparison of different coating programs in terms of deposition efficiency, minimization of spray time and accelerations, etc., is much easier with the coating simulation program. x Visualization of the local spray angles and quantitative estimations of the certain coating properties (e.g. porosity, roughness) with the subsequent program adjustments could lead to improved coating quality. x A parallel development of the coating algorithm and designing of part masking and fixtures also becomes feasible. x Offline tools allow more informative comparison of trial results with the simulation that leads to more efficient tuning of the coating programs and a reduction in the number of development iterations. A priori coating analysis and visualization of the gun movement help to avoid the gross mistakes in programming, such as collisions, coating flaws, etc. and to eliminate completely wasted coating trials. x Offline simulation allows for analysis of the coating sensitivity to the tooling and masking tolerances, gun misalignment, effects of part temperature change, variation of powder feed rate, etc. x Analysis of robustness of the quality-control procedures is possible, for instance, an influence of location of the control points on the overall process uncertainty can be understood. Applicability of the full-scale offline simulation, of course, essentially depends on the spray technology. For instance, visualization of the coating thickness on the real complex component is absolutely critical for the LPPS process with the large

5.7 Use of Offline Simulation in Coating Development

spray pattern comparable with the size of the part to be coated. This is necessary to take into account the effects of shading, overlapping and overspray. At the same time the coating thickness from APS or HVOF with the small spray spots is much more predictable intuitively and experienced developers might not be interested in usage of sophisticated and expensive software tools for visualization and may constrain themselves with particle-in-flight or spray-pattern models only. 5.7.1.4

Physical Modeling and Offline Simulation as Process-Diagnostic Tools

Simulation tools of all levels, especially, the full-scale offline software, can be invaluable for the process understanding, diagnostics, debugging and monitoring. Complex physical models have a huge advantage – they allow one-parameter-a-time variations and creation of cause/effect diagrams. Using such capabilities, it becomes much easier to reveal and eliminate special cases of process deviations that are caused by hardware and control failures or by major maladjustments. Potential causes of such events can be analyzed before carrying out any experiments, which makes the debugging procedures much faster and less costly. When making the capability analysis, contributions of the different sources of variations can be modeled and potential candidates for process improvement ranked. This could dramatically reduce a number of tests, required for the process capability to be understood. With theoretical trends known, more efficient test matrices can be developed using the statistical “design of experiment” (DOE) tools. Analysis of data from online process monitoring, like plume spectrometer, in-flight particle fluxes, speed and temperature measurements, component thermography, etc., is another very important application of the process simulation and feedback control. The value of the information obtained by online sensors can be drastically enhanced, if there is a clear theoretical background behind them (see Chapter 7). 5.7.1.5

Simulation as a Numerical Experiment

What can the models of this type deliver and what are their limitations? A fullscale simulation tool for modeling of such complex processes as a thermal spray includes numerous physical models and empirical and semiempirical parameters. As a result such tool constitutes a big system with a sequence of superimposed uncertainties. Some of them are due to the theoretical approximations and extrapolations of theory beyond the experimentally calibrated parameter window. But the statistical nature of the process remains the major contributor to the uncertainty “pie”. Also major differences in process tolerances and stability between the laboratory facilities, which are used for model validation and adjustments, and the shop floor should not be forgotten. Such a big simulation system is not deterministic with regard to the process, which it is supposed to describe. In fact it is intrinsically stochastic. Results from it should be treated to some extent as experimental. They possess a certain mathematical expectation and a stochastic scatter, which can be expressed in terms of “n-sigma”. Developers and users of the process simulation tools must refrain themselves from attempting to

201

202

5 Offline Simulation of a Thermal-Spray Process

achieve a “perfect” agreement between the “calculated” and measured coatings on the real components. It is nearly impossible due to the complexity of the process to be modeled. In this case a question arises, what kind of prediction accuracy can be expected from the simulation tool and what are the major requirements it must fulfill to become usable? They must be adjustable by tuning the physically reasonable parameters (e.g. injection speed, turbulence factors, gas swirl, etc.) to a reasonable level of agreement with the results of standard trials (which are defined by the coating developers). Such trials can include, for instance, spraying of the stationary spray patterns and uniformly coated plates, translational or rotation spraying of certain typical configurations like corner plates or bent rods. The term “reasonable” must be quantified by trial and error in several iterations between spraying the test coupons and actual components. Ideally, accuracy of predictions at the test coupons should be reproducible at similar elements of actual parts. The model calibrated and adjusted in the limited parameter area must produce the correct trends throughout the whole range of expected process parameters. This means that several control experiments should be carried out at the operational points, which are far enough from ones used for the model adjustments. It is worth noting here that the well-known saying that states that the model with two or three free parameters can be adjusted to any set of experiments is, of course, true, but to a certain extent only. Experience with the APS and LPPS process modeling shows that an improperly adjusted model (by tuning the wrong parameters and/or of the right ones but beyond the reasonable range) may lead to the significant disagreement between predictions and test results if the trial window is sufficiently wide. Frequently, such a “brutal” adjustment leads to the completely unphysical results, when the process parameters change slightly. For instance, an elliptical shape of the LPPS spot of the typical 50–60 mm size at a standoff of 300 mm cannot be explained by any “reasonable” variation of the powder-injection parameters. After all model adjustments and testing it must produce predictions that are consistently close to the spray results on different real components. It is quite difficult to define the term “close”. For instance, as far as the coating thickness is concerned, it means that the simulated distributions must, firstly, exhibit the right trends at all the checked locations, and secondly, they must not deviate from the actual thickness by more than a magnitude of coating roughness in terms of surface average values. For such parameters like the porosity or surface roughness the simplified models are capable of producing reasonable trends versus the spray parameters, distance and angles, but not the absolute figures. 5.7.1.6

When the Offline Simulation Should Be Used

Whether to involve offline simulation tools or should the coating development solely rely on the personnel experience and trial and error – the decision should be stipulated by the business considerations. The question is, what are the prerequisites for the investments in simulations to achieve the break-even and

5.7 Use of Offline Simulation in Coating Development

produce a profit. For instance, it may be reasonable to consider a full-scale modeling in the following production cases: x Large-scale serial production, when even minor improvements of the coating quality can bring significant returns. x Low-number serial production (or even a single-piece manufacturing) with the high cost of poor quality and high costs of preserial trials. Usually, turbine-blade manufacturing falls into this category. An offline simulation tool in the consulting business can be considered as part of a know-how kit, which an external expert can bring into the company. The importance of the modeling, especially its usefulness in multiparametric studies and process visualization, cannot be overstated as far as personnel training is concerned. The decision on which level of simulation tools should be used depends mostly on the size of the coating manufacturer. Simple particle-in-flight models, which are available on the market (like, for instance, CASPSP® – a product of Paton Welding Institute, Kiev, The Ukraine) are quite stable, relatively inexpensive and can be used even in a small-size business. Experience with full-scale offline simulation tools is currently quite limited and they can be considered experimental developments or beta version (e.g. ALSTOM/CENIT coating package) rather than fully commercialized tools. Therefore, there is a certain risk of using such software related to its stability and long-term support. Thus, it would be a reasonable assumption that only quite large companies are able to invest in such packages.

203

205

6 Standards and Training This chapter presents an outline of standards and recommendations available for thermal spraying and the people involved. Application of these standards and training rules is a precondition of producing coatings of sufficient quality. However, they are not sufficient because standardization is always a compromise of the interests of the parties involved. Therefore, every OEM, overhaul shop and coating applicant develops his own rules in addition to the general accepted standards.

6.1 Standards, Codes

We checked if we should include other standards, like US American or British UK standards as well. The plan was to select the optimum set of standards. However, there are so many national standards available, that we decided to concentrate on those ISO and EN standards that anyhow are taken over to national standards within EU (AFNOR, BSI). Some specialties are ASTM standards and not published standards of the turbine and aero engine industry, which are normally not available to the public. 6.1.1 Introduction to Standards

The requirements on quality of thermally sprayed coatings and their reliability in powerplant application are steadily increasing in order to ensure reliable operation and guaranteed maintenance intervals. This high level can only be guaranteed if all production steps from planning, preparation, coating application to post finishing and quality assurance are carried out properly. Beside the empirical knowledge of all specialists, it is a necessity to document all steps for being able to trace back at any time, if necessary. A typical OEM specification requires traceability of quality documentation during the whole life of the powerplant. Reliability and safety in the turbine business (stationary or flight) are of special importance, since costly machine downtime or even human life may be affected.

206

6 Standards and Training

In any larger coating job shop, and especially in those dealing with turbine components or other highly demanding applications, such quality-assurance systems are essential and required by their customer. During the last decade, an international quality-assurance system was established within the thermal spraying business, dealing in more detail with specific cases than the general ISO 9000-series could do. The ISO 9000 series “Qualitymanagement systems” can only comprise the general “roof”, but does not contain any process- or product-related specifications. These standards only describe the evidence of preventive management tasks as well as analysis and documentation of all the processing steps in a company. Basically, the ISO 9000 requires from the user that he manages the quality of his product in a systematic and documented way. ISO 9000 certification is a strict requirement for every party involved in the design and production of a powerplant. The surface-protecting processes are pointed out as special processes, because they often can not be controlled by nondestructive testing during the production to guarantee the fulfillment of the required quality. The quality of thermally sprayed coatings can not be improved afterwards. Therefore, it was necessary to establish a product-oriented quality-assurance system for thermally sprayed parts, like EN ISO 14922 Part 1–4 [219–222] where an adequate quality-management system is described. The new quality-assurance system describes the whole field of thermal spraying with four most important “M’s”, which are covering the x x x x

Materials for spraying (consumables) Machines (acceptance inspection of thermal-spraying equipment) Methods of testing (coating and interface characterization) and Man/women carrying out the work (education and qualification)

The first three topics are covered by European (EN) and international (ISO) standards mainly elaborated within the technical committee of CEN/TC 240 “Thermal spraying and thermally sprayed coatings” (see later). The fourth point dealing with education and qualification was worked out within the European Federation for Welding, Joining and Cutting (EWF) and will be taken over by the International Institute of Welding (IIW). These requirements for education and qualification have now become a worldwide acceptance standard in the thermal-spray business. 6.1.2 Quality Requirements for Thermally Sprayed Structures and Coating Shops

The installation of an effective quality-assurance system is initially expensive, but guarantees afterwards a high quality of products at reasonable costs. A coating job shop working according to the quality standards needs a documented quality-management system, which is described in the quality-management handbook (QMH) and in process and operating procedures. The demonstration of the transformation of this resolution by internal and external audits is very important.

6.1 Standards, Codes

The international standard ISO 14922 Part 1–4 gives general “quality requirements of thermally sprayed structures” (see Fig. 141). x ISO 14922-1: Quality requirements of thermally sprayed structures – Part 1: Guidance for selection and use Part 1 of ISO 14922 is the framework document of the new standard and gives the users general guidelines for the selection and use of the different quality levels. The standard comprises three levels of quality demands, namely: ISO 14922-2 Comprehensive quality requirements, ISO 14922-3 standard quality requirements, and ISO 14922-4 elementary quality requirements. x ISO 14922-2: Part 2: Comprehensive quality requirements The comprehensive quality requirements have to be selected, where coatings and components are subjected to the highest demands, like in the air turbine or stationary turbine business or any other sectors where failing of the coating or the components would have dramatic financial, environmental or public consequences. For comprehensive quality requirements it is compulsory to have qualified supervising personnel according to the standard EN 13214 [223] or equivalent ISO 17833 [224] (Thermal spray coordination – Tasks and responsibilities). They are responsible for the quality of the sprayed components. The spraying process itself has to be executed by qualified sprayers according to ISO 14918 [225] (Approval testing of thermal sprayers), whether qualified for flame, arc, plasma, or HVOF spraying. If the comprehensive quality requirements apply, all quality-management elements from contract checking up to quality recording have to be fulfilled completely. x ISO 14922-3: Standard quality requirements Standard quality requirements need, except for a few exceptions, less restricted claims than ISO 14922-2, e.g. in maintenance of equipment and procedure specifications. x ISO 14922-4: Elementary quality requirements The last part 4 of ISO 14922 defines the “elementary quality requirements”. This quality level will be selected if no specific quality-management system is required and the property of the final product is less demanding or the consequences of failure are less drastic. Besides the full requirements on spray consumables and health as well as environmental aspects, the claims on personnel are only to be present and instructed in thermal spraying. The following Figs. 141 and 142 are showing how the ISO quality standards described above and the personnel qualification specified within the EWFguidelines are integrated in the total quality management system for thermal spraying. The following paragraph describes more detailed the different levels of qualification and education of personnel for thermal spraying.

207

208

6 Standards and Training

Fig. 141 ISO standards integrated in the total quality-management system for thermal spraying (courtesy Boehme, Ohliger, SLV München) [226].

Fig. 142 Relation between EN/ISO standards integrated in the total quality-management system for thermal spraying and EWF guidelines (courtesy Boehme, Ohliger, SLV München) [226].

6.1 Standards, Codes

6.1.3 Qualification and Education of Spraying Personnel

The education in thermal spraying is built up in analogy to the field of welding, to which thermal spray technology belongs. There are three levels of personnel qualifications (Fig. 143), namely: European Thermal Sprayer (ETS) European Thermal Spray Practitioner (ETSP) European Thermal Spray Specialist (ETSS)

EWF Doc. 507-01 EWF Doc. 592-01 EWF Doc. 459-01

Depending on the type of spray work that has to be done, an educated thermal sprayer and supervisor is required. The specific parts of education are open to transitions from the bottom up (sprayer) to the top level of the thermal spray specialist. These requirements can be harmonized within Europe and should be taken over by the International Institute of Welding (IIW) to become a part of the worldwide quality-assurance system in the field of thermal spraying. Table 27 summarizes the quality requirements concerning the selection and personnel. Beside those personnel concerned standards, the three other M’s, namely the Materials for spraying (consumables), Machines (acceptance inspection of thermal-spraying equipment), and Methods of testing (coating and interface characterization) contribute to the total quality-management system. These standards are harmonized within Europe (EN) and are usually taken over by ISO (International Standards Committee) due to the Vienna agreement.

Fig. 143 Structure of ETS-education: Entrance and completing scenarios (Ref. DVS/SLV) (courtesy Boehme, Ohliger, SLV München).

209

210

6 Standards and Training Table 27 Quality requirements in thermally spraying business (in general). ISO 14922-1:

Thermal spraying – Quality requirements of thermally sprayed structures – Part 1: Guidance for selection and use,

ISO 14922-2:

Thermal spraying – Quality requirements of thermally sprayed structures – Part 2: Comprehensive quality requirements,

ISO 14922-3:

Thermal spraying – Quality requirements of thermally sprayed structures – Part 3: Standard quality requirements,

ISO 14922-4:

Thermal spraying – Quality requirements of thermally sprayed structures – Part 4: Elementary quality requirements,

ISO 17833:

Thermal spraying – Thermal spray coordination – Tasks and responsibilities,

ISO 14918:

Thermal spraying – Approval testing of thermal sprayers.

Figure 143 shows different entrance levels with required expertise (left column) and completing scenarios (right column). The required expertise is given in percentage of theory (T) and practical work (P). The certificates (middle column) on ETS and ETSP levels are given according to their specific knowledge: -F (flame), -A (arc), -P (plasma) and -H (HVOF). The EN standards have been elaborated at the CEN/TC 240 committee and can be divided into the following categories: Thermal spraying (general) x Terminology, classification [227, 228], recommendations for thermal spraying [229] x Procedures for the application of thermally sprayed coatings for engineering components [230] Materials, their characterization, and application related standards x Powders – Composition – Technical supply conditions [10] x Wires, rods and cords for flame and arc spraying – Classification – Technical supply conditions [12, 231] x Metallic and other inorganic coatings; thermal spraying; zinc, aluminum and their alloys [232] x Coatings for protection against corrosion and oxidation at elevated temperatures [233] x Methods of test for ceramic powders – Part 5: Determination of the particle-size distribution [234] x Metallic powders; determination of flowability by means of a calibrated funnel (Hall flowmeter) [235] Pre- and posttreatments x Pretreatment of surfaces of metallic parts and components for thermal spraying [236, 237] x Posttreatment and finishing of thermally sprayed coatings [238, 239] x Spraying and fusing of thermally sprayed coatings of self-fluxing alloys [240]

6.2 Special Case: Spraying for Power-Generation Components

Machine and spraying equipments x Acceptance inspection of thermal-spraying equipment [241, 242] (Remark: The above-mentioned standard is now in revision. It will be divided soon in substandards (part 1 to 6) incl. handling systems and powder feeder) x Determination of the deposition efficiency for thermal spaying [243] Specifications of parts, components and tests x Component design (DVS guideline) [244] (EN standard in preparation!) x Thermal sprayed coatings – Symbolic representations on drawings [245] x Characterization and testing of thermally sprayed coatings [246] x Determination of tensile adhesive strength [247, 248] x Technical specifications of thermally sprayed components for aerospace application [249] Of course, there are many other national standards available, among them probably the most from ASTM (American Society for Testing and Materials). This book only intends to draw the reader’s attention to these interests and likes to refer to the actual applicable standards available (including revised versions). These international standards are usually for common applications and have to be completed with application-field- and company-specific internal standards. They contain the exact information regarding e.g. coating uniformity, tolerances, and special testing procedures, etc.

6.2 Special Case: Spraying for Power-Generation Components

Thermal spraying for power-generation components requires people with certain skills. The final target of a reliable stable coating process is achieved by process development, process release, freeze, and documentation. The production has to use the released and frozen process only. Changes require a new release. Process release requires documentation proving the process capability. In all cases, an ISO 9000x and ISO 14000 certification is required. The coating facility proves and assures that it has a management system and works in a systematic and documented way, as described above. Every coating facility for power-generation components has to have the expertise in: x Coating-process development x Coating production (trained personnel and adequate equipment) x Feedback of production results or powerplant operation results into coatingprocess development We will now describe in more detail the specific tasks.

211

212

6 Standards and Training

6.2.1 Coating-Process Development

The task “coating-process development” starts usually with a technical specification, containing: x x x x

Coating material Thickness distribution on the component Allowable defects and tolerances Coating microstructure (porosity, portion of molten/unmolten particles, etc.)

Besides the pure production and scientific knowledge, the coating development engineer has to keep in mind the commercial targets, like x Costs x Lead time x WIP It must be assumed that the thermal-spraying process results in 100% yield in serial production (see Section 7.3). Therefore, the coating development engineer has to understand the quality requirements in the technical specification and the commercial parameters of the production process. He must be able to discuss with his counterpart on the customer side the coating-process capability and in many cases design for thermal-spraying issues. Within the coating facility, he has to understand x the process steps and equipments involved, i.e. – substrate preparation: i.e. degreasing, sand blasting, etc., – robot programming and coating application (parameters, their influences, etc.), – post-treatment/finishing and potential heat treatment, as well as x the supply chain, i.e. powder, power, gases, etc. In most cases, the relative motion of component and spray gun has to be performed by a CNC controller or by a robot. Therefore, programming of controllers is required. There are certain applications, very often in the repair business, where manual operations have to be carried out. The right combination of coating trials and simulation will deliver the information about the process stability in production. The design of experiments (DOE) can help to reduce the costly experiments and coating characterizations giving access to a ranking of the most important parameters and their influence on the required coating properties or more sophisticated simulation tools (see Section 5.7.1). Destructive testing of components has to be minimized or is even impossible, if only one component is available.

6.2 Special Case: Spraying for Power-Generation Components

The coating development engineer selects a coating process suitable for routine production. He works out the quality-assurance procedure. Altogether, his deliveries are documented in the manufacturing process plan. This manufacturing process plan will have to be used in production. All functions involved release the manufacturing process plan (first article inspection, serial production release). Together with known powder/coating deposition efficiency (which is dependent on the shape complexity of the component), the process capability and overall equipment effectiveness (OEE) the coating costs can be calculated. The coating costs may differ strongly according to the spraying equipment and powders used. 6.2.2 Coating Production

Production operators have to accept the manufacturing process plan including all equipment parameters, manufacturing times and material consumption. They have to operate with the required personnel, coat the components, perform the quality assurance and measure all items required to control the production process. They have to assure the material flow on the shop floor and have to minimize WIP. Production operators have to understand and maintain the equipment used. They have to know the operational window. They might be given the right to adjust settings within the limits of the manufacturing process plan. They have to understand the influence of changes in settings. 6.2.3 General Requirements for Coating-Shop Personnel

All people involved should give their feedback of observations and collected results into the coating-process development. This feedback might influence: x x x x

Raw material specification Substrate/work piece preparation Coating process parameter settings Finishing operations

All the requirements of ISO 9000x have to be addressed in the coating facility in order to come up with a reliable stable thermal-spraying process. Depending on the size and volume of a coating facility, all the tasks described have to be done by one or more persons. Often the same person does the process development and afterwards the coating production. In such a case, the understanding is important, that only a frozen process will deliver the right results. In such process development typically several disciplines are involved, among them x Design x Purchasing

213

214

6 Standards and Training

x x x x

R&D Production Quality assurance Cost controlling, etc.

In addition, the coating application is often only one part of the manufacturing chain. Therefore, communication between the disciplines is very important.

215

7 Monitoring, Shopfloor Experience and Manufacturing Process Development In this chapter we will add assessment of monitoring and sensing in thermal spraying. Ideas on the usage of monitoring in real production will follow. Finally, we will combine everything we identified in earlier chapters into the most important step of coating production: Development of a coating process that enables stable serial production of coatings of powerplant components.

7.1 Monitoring, Sensing 7.1.1 Introduction of Monitoring

The knowledge about the influence of thermal-spray process parameters on the final coating properties is a crucial part in coatings design [112, 250–258]. Particle temperatures and velocities during the moment of impact on the substrate are believed to be the main factors influencing splashing and spreading of the droplets, beside the influence of substrate roughness and substrate temperature itself. In the early 1970s Hantzsche described already the possibility of “temperature measurement of powder particles in a plasma jet” [259] and in 1988 Fincke and Jeffery the possibility for “simultaneous in-flight measurement of particle size, velocity and temperature” [260]. Nowadays, many sophisticated online particle diagnostic systems are available on the market. Based on different physical methods such systems allow for the simultaneous measurement of particle speed, temperature, and size distributions as a function of time and place. Among the first equipments available for particle velocity measurements have been Laser Doppler Anemometers [261–267], LaserStrobe [268], or high-speed (quotient) pyrometers for temperature measurements of particles in-flight [269–274]. Laser Doppler Anemometers work according to the principle of superposition of two laser beams crossing a common volume at a given angle (2 D). This superposition leads to interference fringes. Any object passing this volume

216

7 Monitoring, Shopfloor Experience and Manufacturing Process Development

Fig. 144 Principle of Laser Doppler Anemometry (LDA) (in analogy to [262]).

will cause a scattering fringe that can be collected by a photo detector. Out of the evaluated signal the velocity and direction of particle flight can be deduced (Fig. 144). The dual-wavelength pyrometers calculate the surface temperature of a passing particle from the radiation intensity measured at two different wavelengths. The temperature can be derived from the quotient of the two signals at two different colors without exactly knowing the emission factor of the species envisaged. The time response of such pyrometers has to be fast in order to detect bypassing particles in the diameter range of 10 to 60 μm with velocities up to 600 m/s. Both types of particle-monitoring systems require exact positioning either of the two laser beams or the volume to be detected in the case of the quotient pyrometer. A combination of these two systems in one equipment, which allows simultaneous access to temperature and velocity of each particle and does not need careful adjustment, is historically the DPV-2000 [275]. Newer CCD-camera-based devices are able to deliver likewise information on temperature and speed of the particles as well as on the whole jet and its position [276–281]. In this steadily growing market such CCD-based systems are nowadays commercially available as Accuraspray® [282, 283], PlumeSpector® [282], SprayWatch® [284, 285], IPP-2010® [286], or ThermaViz® [287], to name the most popular of them today. Laser- and DPV-based systems extract typically local information out of a few hundred μm3, which have afterwards to be averaged (i.e. scanned) for the whole jet to get representative particle distributions in the plume. In contrast, CCDcamera-based devices offer the advantage of a larger view field (dm2), but typically a lower local resolution.

7.1 Monitoring, Sensing

In the following example, the optimization of atmospherically plasma-sprayed (APS) coatings, especially highly porous thermal-barrier coatings (TBC), will be shown using particle-monitoring devices. This study aims to elucidate the correlation between the basic plasma-spraying parameters (like electrical current, plasma gas (composition and amount), standoff distance between plasma generator outlet and substrate, etc.) and the resulting particle properties (temperatures, velocities) and the final microstructure and thermomechanical coating properties. 7.1.2 Particle-Monitoring Devices

As an example for spray-parameter optimization, the portable in-flight particle diagnostics system DPV-2000 of TECNAR Automation Ltd. (Canada) was used. The equipment consists of an optical sensing device based on a patented technology developed by the National Research Council of Canada. The system uses infrared pyrometry along with a dual-slit optical device in order to perform real time in-flight particle temperature, velocity and diameter measurements (Fig. 145). Because of the slit mask, a particle passing in front of the sensor will generate a two-peak signal (Fig. 146). From the known optical geometry (magnification factor), the speed of the particles can be calculated. The signal is further split in two different wavelengths (F, F1) and the temperature of this particle is analyzed by quotient pyrometry (PD-1, PD-2). Assuming spherical particles, the amplitude varies proportionally to the square of the particle’s diameter, which leads to the indication of the diameter distribution.

Fig. 145 Principle of DVP 2000 system with sensing head and optical beam splitting for quotient pyrometry (F, F1) and velocity measurements (courtesy of Tecnar).

217

218

7 Monitoring, Shopfloor Experience and Manufacturing Process Development

Fig. 146 Signal (intensity versus time) generated from a particle passing in front of the dual slit for speed measurements (courtesy of Tecnar).

Table 28 Technical specifications (according to the manufacturer). Diameter measurement range

1 to 300 μm (depending on photomask used)

Velocity measurement range

10 to 400 and 400 to 1200 m/s (depending on photomask used)

Temperature measurement range

1100 to 4500 °C (depending on photomask used and emissivity of the powder)

Analysis rate

up to 800 particles per second

Calibration module for

temperatures between 1800–2500 °C and speed between 40–90 m/s

However, because of signal filtering, a better estimate of the diameter is obtained from the time integral of the complete signal normalized for (i.e. multiplied by) velocity. Table 28 shows some typical specifications. 7.1.3 Influence of Spray Parameters on Particle Speed and Temperature

For different positions downstream the nozzle exit (75–250 mm) of an atmospheric plasma-spray system, particle velocities and temperatures were measured and afterwards samples were sprayed on substrates using the same spraying conditions. As an example, Fig. 147 shows the typical particle speed and temperature distributions in the cross section (horizontal and vertical axes) of the plasma jet at a distance of 100 mm from the nozzle exit ensuring homogeneous and symmetrical particle loading. The measurements were done for different plasma gas compositions and amounts, different plasma current levels and finally nozzle diameters. For two specific nozzle diameters, any stable process parameters (including extreme and critical settings) were experienced regardless of the resulting coating formation (efficiency, adhesion, etc.). Figure 148 shows two distinct regimes of the cumulative results of the following variations with stable plasma conditions (current: 600–800 A, plasma gases Ar: 24–37 l/min, H2: 3–7 l/min and standoff distances: 75–200 mm) for two different nozzle diameters, namely 6 and 8 mm.

7.1 Monitoring, Sensing

Fig. 147 Examples of (top) velocity and (bottom) temperature distributions of YSZ-particles in cross section perpendicular to plasma jet measured at a distanced of 100 mm from the nozzle exit (coordinate system (0/0) = torch center line) using 26 kW of electrical input power (courtesy of ASM International) [122].

There are distinct particle speed and temperature regimes for the two nozzle types. The main influence caused by the different nozzle diameters can be seen for the particle speed (reduced cross section), whereas the mean temperatures of the particles do not differ much. The slope between temperature and speed is steeper for the 8-mm nozzle compared to the 6-mm nozzle. 7.1.4 Influence of Particle Velocity and Temperature on Microstructure

The structural differences of the coatings sprayed at vastly different parameters can be examined, e.g. using image analysis and mechanical testing. Fig. 149 shows the dependencies of particle temperature and speed on the total porosity of samples measured by image analysis. It can be seen that for decreasing particle speed and temperature the porosity increases. It can also be seen, that the pronounced trend

219

220

7 Monitoring, Shopfloor Experience and Manufacturing Process Development

Fig. 148 Influence of nozzle diameter on the distribution of particle speed and temperatures for all parameter variations tested (courtesy of ASM International) [112].

Fig. 149 Comparison of differently sprayed YSZ coatings and their total porosity (mean value measured at 5 different places) compared to corresponding particle temperature and velocity (courtesy of ASM International) [112].

7.1 Monitoring, Sensing

Fig. 150 Coating microstructure sprayed with high particle velocity (Vmax = 307 m/s) and high temperature (Tmax = 2925 °C) showing an overall porosity of about 26% and vertical cracks with a crack density of 2.7 cracks/mm in length (courtesy of ASM International) [112].

Fig. 151 Coating microstructure sprayed with low particle velocity (Vmax = 115 m/s) and low temperature (Tmax = 2520 °C) showing higher porosity level (about 44%) than Fig. 149, but no cracks (courtesy of ASM International) [112].

of porosity increase with the lowering of particle in-flight parameters vanishes, when the mean particle temperature drops below the melting point (approximately 2600 °C for YSZ). The particle sticking efficiency dramatically decreases, because slow impacting particles below their melting point can no longer adhere. Micrographs from coating cross sections reveal entirely different microstructures for the various parameter ranges depicted in Fig. 149. Two candidates from the “extreme” positions (i.e. range (1) “hot/fast” and range (2) “cold/slow”) are shown, respectively, in Figs. 150 and 151. The sample sprayed with high velocity (Vmax = 307 m/s) and high temperature (Tmax = 2925 °C) shows less porosity (approximately 26%), but a high level of vertical cracks with a density of 2.7 cracks/mm in length compared to the coating sprayed with low velocity (Vmax = 115 m/s) and low temperature (Tmax = 2520 °C) showing approximately 44% porosity, but no visible cracks. The higher the particle speed and temperatures, the higher the degree of melting and spreading will be, which results in a higher coating density. However, due to larger residual stresses, the coatings tend to crack, as is known from the literature [288, 289]. In contrast, the lower the temperature and the particle speed, the more porous the coatings will be until no more adhesion occurs and “grit blasting with cold particles” will take place.

221

222

7 Monitoring, Shopfloor Experience and Manufacturing Process Development

The gun electric input power (EIP), argon and hydrogen flow rates varied for each nozzle diameter and corresponding changes in the coating porosity can be analyzed. As far as power is concerned, results of the experiments followed the wellknown trend of general increase in particle speed and temperature with the power and decrease of the coating porosity [173, 175, 290–295]. However, variations of the total gas flow rate and of the relative content of the gas components when keeping the input power approximately constant have not revealed any clear trends in the particle temperatures. Nor was there any clear correlation between those parameters and the coating microstructure. Generally, an increase in the flow rate of one or both gas components led to higher particles velocities, while the temperature could decrease due to the reduction of particle dwell time and specific plasma enthalpy. All statements are related to coatings in the “as-sprayed state” only. However, creating tendencies concerning the sensitivity to cracking may be partially transferred to the behavior under thermal cycling. The future goal will be to use such online monitoring devices as feedback for intelligent process controlling. Knowing all, or at least the most important input parameter dependencies, such systems would allow making self-adjustments of the spraying input parameters to keep the particle speed and temperature, and thus the coating quality constant (see Section 7.2). This is an important task, because controlling of, e.g., the current and gas flow rates in plasma spraying are not enough to detect any ageing of the electrodes or any other alterations (e.g. clogging of the spray nozzle or the powder port, etc.) influencing the final coating quality.

7.2 How to Use Monitoring for Process Control

This section shows the potential and limitations of using different stages of process control from a coating shop point of view. 7.2.1 Monitoring, Sensing from a Job Shop Point of View

The overall target is to achieve stable, reliable coating production for powergeneration equipment, reasonable lead-time and optimum cost. We consider three types of coatings in power generation for the protection of the basic material from the influences of the harsh environment (heat, corrosion and wear). The surfaces of the coated components contain the characteristics of the spraying material and extend thereby their lifespan. They can even make the components working under conditions, where uncoated elements will hardly survive and fail in a short time period.

7.2 How to Use Monitoring for Process Control

The traditional control of the initial parameters by a process controller is not always sufficient to reach the necessary characteristics of the layer. The constantly growing requirements on process stability and coating quality require innovations. Finally, we understand that the heat/particle jet is the actual tool that needs special attention. To illustrate this comparison the following examples from cutting and shaping processes are given: x We should try to image a milling procedure, whose tool geometry (diameter, length) is unknown before its start. The result surely doesn’t correspond with our expectations. In this bad scenario we may watch a crash including the damage of the machine and the destruction of the work piece. x In the second example we try to manage the same task with a correctly defined tool. However with a scuff we thereby change the characteristics of our tool during machining. In this case the monitoring equipment of the machine will discover the problem, before something worse happens. The machine possesses sensors that can prevent such cases by the measurements of the amperage, vibrations, pressing force and even by optical observations. Although most problems of the drilling and milling machines are directly visible on the surface of the work piece after the working process, the machines are supplied with many control systems. All this together saves time and costs, makes realistic planning possible and increases the productivity of the process and the quality of the produced parts. Problems of thermal-sprayed coatings typically occur inside the layer or at the bond between the coating and substrate. They mostly become only visible after the preparation in the laboratory or by premature coating failure. There is so far no reliable online method available to control the growth of the coating. The representative laboratory tests have to be accomplished on a suitably chosen work piece or accompanying specimen. The selected components (see Section 2.6) are x Large casing (single piece) x Combustor segments x Turbine blades Such parts are typically complex in shape and need special attraction concerning the coating application. Typically turbine blades with different cross sections need careful planning of the torch and plasma trajectory, respectively. An overheating of thin-walled parts or coatings on surfaces above cooling channels may be critical. Care has to been taken in such cases and observation of surface temperature would be helpful. Online coating inspection becomes difficult as these components often have bent surfaces. We have to consider different types of thermal-spray processes, different components, which require different monitoring, control and closed loop procedures.

223

224

7 Monitoring, Shopfloor Experience and Manufacturing Process Development

General requirements for monitoring in production are: x Measurement of plasma and particle properties when torch is at low or zero movement, after certain time intervals. x Put manipulation into work piece, keep torch motionless. 7.2.2 Vision for Future Coating Control and Monitoring

Gun and component have to be representative for the whole amount of produced pieces. It must be guaranteed that the whole production shows the same coating characteristics like the piece of the laboratory. Due to this it is very important to take constantly care that the coating jet does not change its characteristics during the entire working process. On the other hand, we know from practice that the jet changes during operation. There are many factors influencing the jet behavior. Some of them are predictable and based on physics, simply imaginable, understandable and easy to measure. The others are also physically based but not easy to measure and control. Only a small group of experts and selected specialists will understand them. Some of the effects are purely stochastic problems. We try to get a group of influencing factors under control, which are easy to understand and to measure. In the process circle – human, powder, machine – the human is the directly responsible factor for success or failure. x Human – learning, understanding, training, exercising and a great motivation. x Powder – first it will be examined with different methods. When it is confirmed that its characteristics match our specifications it is ready for use. x Machine – this topic has been much neglected in the past. Fortunately, the situation has now changed and we talk more and more about the intelligent systems using monitoring and closed-loop process control. The development of thermal-spray coating systems can be divided into several phases, depending on the increasing complexity of the control mechanisms, but with the intention of finally entire process control. (1) Simple controlling of the initial parameters (Fig. 152): This is completely sufficient for many low-end applications. Unfortunately, it is not sufficient for coatings with high quality requirements. The operator may influence the process by changing one or more input values. This can only be done based on measurements of a feedback signal which is usually done after stopping or finishing the coating process. The coated piece can, e.g., be weighed to determine the weight of the coating. The coating thickness can be determined with the help of the additionally coated samples or in certain cases by the “eddy-current method” [296]. The feedback information of the coating quality is only available with a certain time delay and therefore only of limited use for process control.

7.2 How to Use Monitoring for Process Control

Fig. 152 Simple process with fixed input parameters.

A drift in process parameters influencing the x powder velocity, temperature and their distributions, or x component temperature, etc. will not be detected until the x thickness distribution or component weight can be measured at the end of the spray session. It will be too late for corrections and quality and reproducibility can be strongly affected. Often, such systems use flow controllers, but output changes in the plasma or flames, respectively, will not be detected directly. (2) It is possible and practicable to refit a simple spraying system with a jet-imaging tool to capture some process data from the most important area, where particle and heat exchange takes place. This information from the plume has to be understood by operators and helps him by manual process adjustments or allows the coating process to be stopped by detecting deviations rising above the given threshold values. The feedback is carried out manually according to the observations on the screen (Fig. 153). This type of process control insures the detection of changes caused by component aging or powder feeder clogging.

Fig. 153 Process control by operator using the aid of plasma-plume imaging.

(3) The next step in complexity and the first step to closed-loop optimized process control is the adjustment of coating processes using data delivered from different sensors. The collected data sent to a computerized control unit will be calculated and converted to different control signals and sent to several actuators that can affect and compensate a changing or drifting process. Plasma-beam data allows a computer-based controller to change the input parameters within well-defined limits (Fig. 154).

225

226

7 Monitoring, Shopfloor Experience and Manufacturing Process Development

Fig. 154 Process control using a simple beam observation system and feedback to a computer-based controller.

First trials of such procedures have been done based on the knowledge of the dependencies of the process parameters. Nevertheless, no information on the quality of the coating itself is known. (4) The ultimate dream of all technicians and engineers in the thermal-spray business would be an intelligent, fully automated spray process. The extreme rapidly growing electronic industry puts hardware tools at our disposal to start the further evolvement steps of closed-loop spraying systems. We need both good ideas and people who understand the whole process. The experts should be able to determine the processes operational windows and to create several procedures needed for the programming or teaching of such systems, using, e.g., fuzzy logic or neural-network concepts [297, 298] to predict coating qualities. Either the dependencies are known explicitly, then a programming of the process can be done, or the dependencies are known implicitly, then statistical methods have to be applied. Fuzzy logic or neural networks can be “taught” and will deliver dependencies without the need to know all physical relations in detail. The advanced, closed-loop-controlled coating process aims to manage the formation of perfectly sprayed coatings. Online measurements of sprayed weight (or instant weight increase), coating thickness (increased or absolute thickness), particle velocity and particle temperatures, sprayed layer thermal imaging for detecting bonding defect, are no longer science fiction. Such intelligent multimonitoring systems taking into account all kind of process measurement quantities are becoming a reality (Fig. 155). The most complex tasks are the analysis of all monitoring sources and tracing back finding the right consequences just in time. Such a system equipped with several monitoring tools gives power generation an efficient and cost-saving coating application of turbine components satisfying the highest quality requirements nowadays. (5) The last step of total quality control will be spraying equipments using a comprehensive set of monitoring tools (Fig. 156). With a beam control we receive information like a particle velocity, particle temperature, beam location, intensity and shape, but we don’t use the thermal properties of our process.

7.2 How to Use Monitoring for Process Control

Fig. 155 Multimonitoring system with intelligent feedback control.

A very important factor is monitoring of the gun power (net energy). The values of voltage, current, plasma gas flow, cooling-water temperature and cooling-water flow deliver data describing the gun efficiency and the plasma properties. If we measure the temperature of the substrate before and after impact, calculate the thermal energy released to the substrate, measure the heat transfer, calculate the energy absorbed by cooling air, use gun efficiency, we can get additional data helpful for total process control. The approaching technology will make it possible to spend initially more money for process control, but less money on wasted parts afterwards. In the case of costly turbine blades, such expenses will be quickly reimbursed.

Fig. 156 Process control of thermal-spraying system using total monitoring tools.

227

228

7 Monitoring, Shopfloor Experience and Manufacturing Process Development

The multiprocessor controller should be able to separate the interactions of all the influencing parameters and will maintain the coating quality within the predefined operational spray window. The “total monitoring concept” is not the way to replace the operator. It is only a vision of an intelligent and reliable tool, able to operate in a rough production environment while maintaining the coating quality constant, independent of the substrate properties. It will recognize immediately any alteration and changes from the preset parameters leading to a change in the operational window. It will adapt itself within the given boundaries and therefore reach the highest and constant quality and productivity (low rejection rate) and so far the best coating price.

7.3 Manufacturing Coating Development

Thermally sprayed coatings possess properties that distinguish them remarkably from the same solid materials due to the specific microstructure caused by the different production techniques. Therefore the appropriate structural basic design and the proper base material have to bear most of the load. As thermally sprayed coatings are typically thin compared to the underlying substrates, they do not increase the mechanical strength values of the parent metal. Nevertheless, the coatings represent a part of the whole construction design. It is wrong to design a part without having the chances/limitations of thermal spraying in mind. It will, e.g., not be possible to coat small inner diameters, nor cavities without proper access, as well as rectangular corners. Thermal spraying is a “ line-of-sight process”. Only particles reaching the surface to be coated under an appropriate angle of approximately 90 degrees will give a good coating quality, whereas particles impinging on the substrates at flat angles, will lead to poor coating qualities. Besides special mechanical preparation and pretreatments, like smoothening the transitions at the edge of the coatings to reduce edge stresses, the components have to fulfill the requirements to be “coatable” (see Fig. 157). Design instructions for specific applications are given in the standards EN ISO 14921 [230] showing procedures for the application of thermally sprayed coatings for engineering components and EN ISO 17834 [233] and for protection against corrosion and oxidation at elevated temperatures in EN ISO 12944-3. Besides the constructional design aids, the proper symbolic representations of coated parts on drawings should be used according to EN 14665 [299]. These harmonized guidelines are important for clarifying the situation between the contractor and the coating shop, where the coatings have to be applied and up to which boundary and thickness limits. As the coatings are part of the whole design and construction process, they have to be included in the considerations from the beginning. Correct design and construction will lead to a better performance and prolonged lifetime of the whole component.

7.3 Manufacturing Coating Development

Fig. 157 Considerations for basic structural design (courtesy of Sulzer Metco).

7.3.1 Coating Development Process

In this section we will outline the steps required to bring the coating into serial part production. A general structure of the coating development process is presented in Fig. 158. It consists of three major blocks: x Part design phase, when the coating material and technology are selected and all required coating properties specified. x Process development, when the process parameters are defined, which ascertain required coating properties. x Part coating development, when the coating machine is programmed and the coating process is qualified for the serial production with the required process capability. Let us consider major elements of the coating development in more detail.

229

230

7 Monitoring, Shopfloor Experience and Manufacturing Process Development

Fig. 158 A layout of the manufacturing coating development.

7.3.2 Coating Definition and Coating Specification; Design for Coating

During the design the new turbine parts are always checked for all relevant failure modes. For turbine blades and vanes these are low cycle fatigue (LCF) or thermomechanical fatigue (TMF), creep, oxidation and, sometimes, corrosion. Depending on the specific company’s design practice and methods corresponding lifetime figures are calculated. Creep and oxidation lifetime is expressed in terms of operational hours and LCF/TMF – in turbine startup/shutdown cycles. An oxidation protective coating is checked for its oxidation/corrosion lifetime. Frequently, a cyclic penalty factor is also defined, depending on the combination of the coating and base materials. For TBC a time to spallation is estimated based on the failure mechanism, which is due to the growth of oxides (TGO = Thermally Grown Oxide) in the interface bondcoat/TBC. At a critical thickness the TBC fail by spallation. There are two approaches to considering the coating at the design phase, which arbitrarily could be called coating for design and design for coating. The former approach presumes that the coating material and coating technologies are selected or even developed to satisfy the conditions of the new engine parts.

7.3 Manufacturing Coating Development

This makes sense, for instance, when the new generation of turbines is under development. But when the coating composition and application technology have been selected or a decision to reuse the coating from the previous developments has been made, the latter approach comes into force. The part must be designed to be coated and to have the coating surviving during the interval between periods of engine maintenance. Let us assume that the coating material and required microstructure have been selected by metallurgists. Usually this means that the coating technology is also decided. Apart from the coating material the cost issue could affect this decision, for instance, a choice of HVOF instead of LPPS for the bond coat. Sometimes the principle design decisions could make use of certain processes impossible; for instance, a decision in favor of twin- or triple-cast vanes could veto the use of thermal spray due to the line-of-site issue. Design for coating presumes two activities running in parallel: generation of the coating specification and influencing the blade design process in order to make that specification realistic. A coating specification describes all the relevant coating properties as well as the quality parameters such as: x Minimum and maximum allowed coating thickness for different surface locations. One should keep in mind that the definitions of the thickness could be different in the design office and in the metallurgical lab. Whereas the designer believes that 400 Pm of TBC means the distance between the inner and outer smooth coating surfaces, it could happen that the metallurgist would measure the distance between the tops of bond coat roughness and bottoms of the outer TBC one. And both those definitions could differ from the understanding of the developer of nondestructive testing hardware and software. Ideally, the lower and upper thickness limits should be applied only in those locations that are critical for the corresponding failure mode. This would make achievement of the coating specification more realistic, and coating development – cheaper and manufacturing yield – higher. For instance, there is not much sense in specifying the minimum oxidation coating thickness in the location, where the oxidation is not an issue at all. Nor should the maximum limit be applied where there is no threat of cyclic failure. Of course in both cases blade aerodynamics should not suffer. A general principle “as stringent as necessary and as generous as possible” should be applied to all quality requirements, not only to the thickness range. x Coating porosity and microstructure in different locations. Here also allowable deviations from the requirements in noncritical locations could be written explicitly. x Coating roughness for the bond coat to assure the TBC adhesion and for the overlay oxidation coating or TBC if required by aerodynamics. x Quality check procedures and locations including the lifetime coating tests like, for instance, thermocycling. There are not too many blade or vane elements where design could be significantly influenced by the coating requirements. Usually, mechanical compatibility,

231

232

7 Monitoring, Shopfloor Experience and Manufacturing Process Development

Fig. 159 Typical overspray structure created inside the corner transition (courtesy of ALSTOM).

aerodynamic, cooling, mechanical integrity and casting issues have priority. Nonetheless, designers should always keep some points in mind when they make their decisions, for instance: x It does make sense to check whether the part could be really coated with the specified thickness distribution, which is not always the case, especially at LPPS with a large spray pattern. Offline simulation is the only tool to do this at the design phase, when the part is not existent yet. x TBC frequently spalls off (sometime even in the shop) from the convex surfaces with small radii. x Too thick TBC has a higher risk of spallation. x Metallic and ceramic coatings frequently chip off at the sharp edges. x Coating quality usually suffers in the corner when the angle is significantly less than 90 degrees. It is almost impossible to coat both sides of such transitions at a reasonable spray angle with the thermal spray gun. At least one of them could be coated at too shallow angle and get what is usually called an “overspray” – very poor, sometimes even flaked coating as is shown in Fig. 159. Shrouded blades are the typical providers of problematic locations of this type. x Positioning of TBC close to the cooling holes may cause problems with their masking during the coating. x Too-thin elements could be damaged during the transferred arc cleaning at LPPS. x Too-big components require much time to coat and may become too cold again during LPPS. Design for coating of the big structural components like casings has certain specifics: x Usually the mechanical integrity criteria are less severe than in the case of blade and vanes, which means that LCF penalty from the coating usually is not the issue.

7.3 Manufacturing Coating Development

x Design lifetime is usually at least 4–5 times longer than that of the turbine blade. x Such components cannot be put into the coating machine. Sometimes they are coated manually. x They are usually cheaper than blade sets and have larger surfaces. Therefore the issue of powder cost and, respectively, deposition efficiency becomes more important. Summarizing the above, the coating quality to a great extent is determined by the solutions design made at the earlier stages of the part development and from the very beginning the coating specifics should be discussed between the designers and coating specialists. Overall, the designed part can be considered to be manufactured only after the coating suppliers confirm their readiness to coat it. 7.3.3 Process Development

The targets of this phase of the coating development are to create a set of process parameters (the process operational point), which would assure a coating with required physical and mechanical properties, maximum possible process deposition efficiency and maximum possible powder feed rate. The two latter goals are driven by the necessity to minimize the coating costs and the process throughput time. Let us consider this development step-by-step, but still keeping in mind that most of those steps are interconnected and decisions made at one of them frequently necessitate the developer to reconsider some of the previous elements. What are the tools that could be used at this development step? The answer essentially depends first of all on the resources available for this task. Nonetheless, let us try to list them: x Qualification and experience of the development personnel. The coating process is essentially stochastic, parameter operational windows frequently are quite narrow and even the provisional parameter screening could last forever without visible success if the deep understanding of the process is lacking. x Company’s know-how and access to know-how of other coating manufacturers. Quite frequently the coating recipes become a product on the market, which could be supplied, for instance, by the hardware manufacturer. x The role of physical simulation cannot be overstated. Physically correct and well-calibrated particle-in-plasma models provide the development with qualitative and quantitative trends, which are invaluable for the test planning and interpretation of the trial results. x Online process monitoring can be used in different levels of complexity as described in Section 7.2.2. x Design of experiments “DOE” is a powerful mathematical tool that allows one to set up efficient trial matrices and extract maximum information from the available experimental results [76, 300–302].

233

234

7 Monitoring, Shopfloor Experience and Manufacturing Process Development

7.3.3.1

Powder Selection

The basic powder selection is done according to the coating requirement and specifications (see Section 2.3.2). In addition, certain factors may influence the powder-size distribution. Figure 160 shows dependence of coating roughness from average powder size. Therefore if a smooth coating is required the average powder size should be lower. In certain cases, e.g. as a bondcoat for a TBC, the higher roughness is preferred in order to improve the bonding strength between bondcoat and top layer without the need of surface roughening before applying the top coat.

Fig. 160 Coating roughness Rt (see Fig. 64) vs. powder size (courtesy of ALSTOM).

7.3.3.2

Torch Parameters

Proper definition of the torch parameters and selection of powder are the key elements that determine the dimensions of the process operational window and finally, the process robustness and capability. Selection of the plasma gas composition and flow rate, torch current and diameter of the anode channel is driven by the powder size and material and by the particle parameters, which are required to assure the particle bonding and necessary flattening ratio. At LPPS the chamber pressure and geometry of the nozzle expansion, which determine the jet shape, are also to be defined. Some coatings like, for instance, a porous TBC sprayed by APS, have a very narrow operational window for the standoff distance, which in such cases becomes an additional physical parameter to be considered simultaneously with the torch settings. This step requires a deep understanding of the process physics. General trends of the thermal-spray process are discussed in Chapter 4 of this book. 7.3.3.3

Spray Pattern and Standoff Distance

Selection of the shape and structure of the spray spot is primarily driven by the coating quality requirements and by the powder load. Some physical phenomena related to the formation and properties of a spray pattern have been discussed in Section 4.5. From the manufacturing point of

7.3 Manufacturing Coating Development

view selection of the size and structure of a spray footprint is about the finding a compromise between the quality issues like thickness, microstructure, coating integrity, coating sensitivity to the process variations on the one hand and the economic considerations like deposition efficiency and spray time on the other. In principle, it is possible to use several spray patterns to coat a part. This gives additional flexibility in coating of problematic areas. For instance, it seems reasonable to spray the radii and corners with the relatively small spot in order to avoid an overspray at adjacent surfaces. However, as was discussed already in this book, due to drifting phenomena, the process parameters might not be reproduced after multiple cyclic switching between different recipes. The standoff distance is determined mostly by three considerations: an appropriate particle dwell time in the plasma or hot gas plume (sufficiently long for the dense coatings and not too long – for the highly porous ones), size of spray pattern and a surface thermal regime. The latter factor can be crucial for certain applications, for instance, for spraying of vertically segmented TBC [303]. Also geometrical restrictions and accessibility of certain locations are the factors, which affect the choice of a standoff in coating programming. 7.3.3.4

Coating Mono-Layer; Powder Feed Rate and Traverse Gun Speed

The total amount of coating material to be put on the part is determined by the specifications. Together with the information about the deposit efficiencies and with the given production rate this determines the powder feed rate. The definition of the thickness and, correspondingly, of the number of layers, which are needed to achieve the required coating thickness is determined mostly by the compromise between the coating quality and the deposition rate. With the given mono-layer thickness, powder feed rate, shape of spray pattern and the passage offset the gun traverse speed is determined unambiguously. Too thick sprayed layer in a single gun passage sometimes is accompanied by the too-high residual stresses in the coating and related cracking due to the increased temperature in the hot spot. If the porous heat-insulating structure is the target, this can also lead to the reduced coating porosity. Also a coating consisting of just a few thick sublayers is very sensitive to the misalignment of the torch and fixtures. On the other hand, a too thin mono-layer means too many interlayer interfaces, which could become the origin of horizontal cracks. Reduction of the layer thickness with the given powder feed rate leads to the increased traverse speed and to the increased number and magnitude of gun accelerations with all the possible consequences for the stability of powder injection. 7.3.3.5

Spray Trials and Coating Qualification

When the spray recipe, spray distance, powder feed rate and traverse velocity have been selected, a series of test pieces are sprayed to qualify the coating. They must fulfill two requirements: x They must be simple enough to make the coating process easily predictable and controllable.

235

236

7 Monitoring, Shopfloor Experience and Manufacturing Process Development

x They must contain the critical geometrical elements of the targeted real components, e.g. corners and radii. Coupons are also sprayed, which are required for the coating mechanical tests, e.g. thermo-shocks, oxidation, etc. The costs of such trials are always an issue; therefore the test pieces must be cheap. This is why the material must not necessarily be the same as that of the details to be sprayed. For instance, instead of expensive superalloys, some materials like HastX or even steels usually can be used. But still some coating features, like residual stresses or adhesion strength, cannot be reproduced if the base metal properties are very different from the ones of the test coupon material. If the coating on the test pieces fulfils the quality requirements, spraying of the real component can be programmed. 7.3.3.6

Sensitivity Checks

Definition of the process parameters, which ascertain achievement of the required coating properties, is only one target of the coating development. The second one is to assure stability of the production, i.e. these properties should not be dramatically altered by the reasonable variation of the process parameters. To achieve this, a set of sensitivity checks should be carried out at this stage of development. The goal of these checks is to define the range of parameter variation, within which the process can still be adjusted to produce the necessary coating, i.e. a so-called operational window should be determined. If this range appears to be too small and the natural process variations can move the coating out of specification, the set of process parameters should most likely be redefined. The issue of operational window is discussed in more detail in Section 3.3.3. 7.3.4 Part Development 7.3.4.1

Coating Program

Depending on the design and whether the robot or CNC is implemented, the manipulation system provides totally from 6 to 9 degrees of freedom for the process programming, counting the gun and the component movements together. Usually the gun is installed on the robot/CNC with 6 and 4 degrees correspondingly and the part to be sprayed is fixed at the turning table or rotating shaft. The latter has the possibility to move axially. Turning tables usually allow one or two rotational degrees of freedom in the part fixation. The “inverse” arrangement, when the component is installed on the robot and the gun is immovable or rotating is also possible (if the component is not too heavy for the robot, of course). Most frequently, the program comprises the gun passages along the component with certain offset between subsequent traverses (so-called “box” spray) and rotation of the sprayed component. There are numerous ways to program the process. Which one to use? The answer depends on the hardware capabilities, the part geometry, the developer’s experience and preferences. In fact, technology of the spray programming constitutes an important part of the company’s know-how.

7.3 Manufacturing Coating Development

Discussion on the specific programming features is beyond the scope of this book. Let us summarize the major requirements of the coating program. x It must be optimized from the point of view of the target deposition efficiency – this is a critical issue in the process economy. x It must be optimized from the standpoint of uniformity of coating thickness. This is discussed in more detail in Section 5.5. x Spray-gun accelerations should be minimized to reduce the stochastic scatter of gun positioning and powder-injection velocity. By doing this also the risks of spitting and injector plugging are reduced. x A spray angle should be as close to 90 degrees as possible. This requirement concerns the coating quality. x The program should be robust to the process variations. This means that the sensitivity of the coating thickness to the misalignment of the robot/CNC, fixtures and masking should be kept as low as possible. Thermal expansion of the sprayed part should be taken into account where applicable, for instance for the large components at LPPS with the preheating. Programming is an iterative process with the feedback loop closed through the test sprays of the real parts and destructive and nondestructive evaluations of the coating quality. Usually this is the most expensive part of the coating development due to high costs of the machine time, sprayed parts and laboratory analyses. The role of offline programming tools in the coating programming cannot be overstated. Automated trajectory generation for the complex 3D parts like turbine blades and vanes is not common at the moment despite the fact that software development for this purpose is ongoing. Nevertheless, even the point-by-point or pass-by-pass work offline with the graphic and numerical representation of the expected coating distribution brings the huge economical effect and dramatically reduces the component development time. One of such offline tools developed in ALSTOM in cooperation with CENIT was described in Section 5.4. 7.3.4.2

Process Qualification and Preserial Release

Part development finishes with process qualification and preserial release. Process qualification is a procedure used by OEMs and other coating applicants. The aim is to show that the total coating process meets the requirements of the given specification. This means that the processes of x Surface preparation, x Coating application, and x Finishing/post treatment are described in detail as required by the customer. Normally, this description includes all the equipment, tooling and fixtures used. For each equipment, tooling and fixture the maintenance and calibration procedure must be available. This is always the case when the coating applicator is certified according to ISO 9000x. The respective processes, operational sheets and quality-assurance procedures and tests, must be available for auditing.

237

238

7 Monitoring, Shopfloor Experience and Manufacturing Process Development

For the specific part and coating the parameter sets for coating (Figs. 67–69) have to be available. Comparable process sheets are required for grit blasting and all the other processes involved. These process sheets are part specific and have to be applied for the specific part. Only these processes are to be used to coat the specific part. Changes are only allowed with the agreement of the customer. It has to be shown that the people involved are trained and capable to perform their task. Again, this is always the case when the coating applicator is certified according to ISO 9000x. Together with the customer, an agreement has to be reached on which and how many test results have to be shown. These results have to be compliant with the specification, which normally includes: x x x x x x

Defined coating material, e.g. powder specification Required coating thickness and tolerance Required coating microstructure Allowable coating defects and microstructure Defined coating substrate interface and tolerances of bonding defects9) Defined coating surface, e.g. roughness, oxide layer, residual stress and tolerances

Depending on the component and the number of components a statistical analysis of results – e.g. coating weight, thicknesses, and porosities – has to be provided. These results are used to show the process capability. All the information is collected into a first article report. Sometimes this is called MMC-brochure (manufacturing methods and control). Here also the relevant certificates and test reports of subsuppliers have to be included. This information is checked by the customer. He might request to witness certain operations in the coating facility or at the subsupplier. The first article report contains proprietary information. Which documentation is passed on to the customer is subject to an agreement between customer and coating applicant. If the documentation and the audits give evidence that the coated part fulfils the requirement of the specification and the required process capability can be expected then the coating process is released for preserial production. After such a release it is expected that the coated part is usable for real power plant operation and has the required durability. It is important that at this point in time all the manufacturing processes in the coating shop and at the subsuppliers are frozen. This implies, for example, that the applied equipment, tooling and fixtures may not be changed without the consent of the customer, even if an identical version is used.

9) Bonding defects are details in the interface coating substrate that are not allowed according to the specification.

7.3 Manufacturing Coating Development

7.3.5 Serial Release

During preserial production there might be more thorough control by the customer. After a certain repeated production all the results and the statistical analysis are checked. The coating applicant shows that the process is capable and achieves the required process capability of 4 to 5 sigma. If this is not possible a further discussion has to be performed with the customer if the specification has to be adjusted or if the process requires further improvement. If all requirements are fulfilled serial release is given. This allows the coating applicant to continue with coating the part without strict control of the customer. The customer controls should be restricted to auditing from time to time. The described procedure seems very formal. However, it is the working version of ISO 9000x requirements. It is used by all OEMs in the power-generation and aero fields and by many other industries. Overhaul and repair have to identify workable routes with their coating applicants in order to guaranty durability of components in power-plant service.

239

241

8 Outlook, Summary The targets of the book are to show how to achieve a stable and reliable coating production for power-generation equipment within reasonable time and at optimum cost. We described the theory, practical approaches, process monitoring possibilities and quality management aspects with the target in mind to show how they help to x x x x x

Improve quality, process stability, maintenance process. Reduce cost and lead time Secure process development Reduce tolerances Produce the right properties (i.e. microstructure)

We said this book deals with questions that are essential for a good performance of thermal spraying for power generation components. We asked the question: Is this still an art or is it just another manufacturing process that can be controlled by intelligent machinery? We followed the obvious logic: A stable process comprises the appropriate hardware, rigorous but still realistic specifications, material and tools according to specifications, optimized process sequence, offline programming tools, right quality system, standardization, well-established control and monitoring and, of course, the people. When we described process stability we changed the focus on the operational window, changing parameters, drifting and monitoring. Ideally we would like to use a non drifting process. Clearly the gun development goes into this direction (see Section 2.4.4.2). When we looked at the stochastic nature of the process we concluded that we have to live with the phenomenon. The behavior of so many particles can hardly be used to describe the coating formation during the spraying process and to use this description for modeling the quality of a coating on a component. Therefore our approach is to separate the events in the torch and plume from the real painting process on the component. We use the spray pattern as the tool for offline programming the coating on the component. Of course the description of the spray pattern is derived by a combination of theoretical explanations of the development of the plume or flame and calibration by experiments. This approach is specific for equipment and group of components to be coated.

242

8 Outlook, Summary

It is shown that the process capability for coating of a specific component is achieved by using the right rules for process and part development. This is not, however, sufficient. Only trained and motivated people can guarantee the required process capability in this changing, drifting process. Special causes of disturbances of the coating process (see Chapter 3) have to be eliminated and documented. This type of influence on the process is a major part of what is considered as an “art”. Of course, the right equipment, tooling and fixture maintenance is a condition of every good process. Probably all efforts will not bring thermal spraying process capability into the order of other manufacturing processes. However sigma values of 4 to 5 seem achievable under the following conditions: Design for manufacturing must be guaranteed. This means the component must be coatable. In addition, the tolerances given in the specifications have to be consistent with coating process possibilities. However, today’s modeling and predictions have limitations. Thickness and porosity can be modeled, microstructure and bonding defects can not. Trial and error is still required in process development of a mature technology. Many published articles and standards are available, e.g. on: x Coating production parameters and their influence on microstructure, microstructure investigations and limitations (e.g. porosity measurements), heat impact in parts and components monitoring. x Quality assurance: the 4 Ms rule: materials, machines, measurement methods, man. All these pieces of information and results have to be documented in such a way that they can be used as a knowledge base for the future. This becomes even more important due to the fact that the basics of thermal spraying still rely on the know-how of experts who have been dealing with it for decades. For the future we see two aspects that we will describe in more detail in separate sections. On the equipment side, the development and introduction of new torches. On the process side, the introduction and usage of offline programming and monitoring in process development and production.

8.1 Thermal Spray Torches

Further developments of HVOF torches are expected to bring spray guns, which would assure higher deposition efficiency and production rate. A wider operational window on a temperature–velocity plane is also a requirement emerging from the trend “higher quality at lower cost”.

8.1 Thermal Spray Torches

There is a market demand to have advanced plasma spray torches. Customers are expecting the device to provide: x Lower level of plasma fluctuations, which means reduced scatter of particle parameters and higher level of deposition efficiency. x Less pronounced effects of drifting, which would directly increase the process capability. x Longer electrode lifetimes in order to reduce the number of maintenance occurrences and increase the process output. However, the thermal spray market now is strongly focused on cost issues. This is why to bring the three aforementioned advantages, or even more, is not enough for the new development to win a large market niche. Redevelopment and requalification of the whole coating shop nomenclature of turbine parts with all the setbacks related to the learning process could cost millions €. And the manufacturer simply follows the well-known saying “do not trouble trouble unless trouble troubles you”, i.e. they prefer the conservative evolutionary way of gradual improvements and accumulation of know-how with the existing equipment. In order to break into production the new hardware must be able to reproduce existing coating conditions in terms of particle parameters and, in the ideal case, to reproduce the existing spray patterns in order to avoid reprogramming. Therefore, operational windows of new equipment should include those of existing torches as subsets. The transition to the new generation of coating equipment may become easier if it is accompanied by the introduction of advanced online monitoring and control technologies. If the existing process can be mapped in terms of particle temperatures and velocities as well as in terms of surface temperatures, it is much easier to find an appropriate operational window for the equipment with significantly or even completely different control parameters like power, mass flow rates and gas compositions, which would still fulfill the existing coating specifications. It may be also expected that the equipment manufactures will package their hardware with the customer-tailored accessories like nozzle and injectors, which would make the transition to new equipment easier. An expert consulting service in redefinition of the process operational windows is likely to be another demanded product in a thermal spray market, especially in small and mediumsize businesses, which can not afford continuous employment of high-ranked development experts. Similar requirements may be imposed on other new products for thermal spray, for instance, powders or power sources. Apart from improvement in the process stability coating manufactures also expect from these new products the ability to be integrated into existing process without significant changes in the coating specifications.

243

244

8 Outlook, Summary

8.2 Future Offline Programming and Monitoring in Process Development and Production

The final target in production is to: x Achieve process stability at minimum total cost x Introduce measuring tools like 6 sigma, TQM, ISO 9000u This book makes the statement: This can only be achieved by good: x x x x

Materials: Quality, shape and size specified according to the demands Machines: Capability, service and process development Measurement methods: Aspects of coating quality Man, i.e. trained people

When considering the importance of process development it becomes evident that this step requires personnel with the right education, experience and communication skills. On the other hand production personnel have to take over such a process development, understand the important features and work independent of the process developer. Therefore production has to understand the limitations of offline programming. It will not happen that the offline program delivers a perfect process. It has to be adapted by production experience. Coating preparation by blasting or other methods has to make the step from pure testing and experimental to a certain degree of automation which means some modeling. Probably, an identical program structure can be used. Modeling must be used to describe microstructure and out of that the physical and technological properties in order to correlate them with lifetime of the coatings and the components, respectively. Therefore the future of modeling and offline programming is an integrated model which includes most of the coating development and production steps. This does not mean that all processes have to be in one model. The output of one model must be such that it can be used as input for the next step in coating or part manufacturing. Monitoring and closed-loop control shall not be independent from the modeling. A correlation has to be established. However, one has to keep in mind the conditions. Process development has to achieve such an operational window, which is not sensitive to small changes, drifts in process and material parameters. Only then should monitoring and – if reasonable – closed-loop control be applied. Nevertheless it is not expected that all parameters are adjustable and controllable. Therefore, also in the future, the spraying operator has the task: He has the prime responsibility for the coating quality. An important aspect in addition is the cost of the coating process. Because the costs depend on materials cost, maintenance cost and process flow with all the people and equipment involved there is a high responsibility on the shop floor personnel. It is expected that coating cost can be optimized by optimizing these factors.

References

References 1 (http://www.asm-intl.org/tss/glossary/ t.htm) 2 US patent classification, http://www.uspto.gov/go/classification 3 EN 657: Thermal spraying; terminology, classification. 4 ISO 14917: Thermal spraying; terminology, classification. 5 Schoop, M. U.: Verfahren zur Herstellung von dichten metallischen Schichten, CH-Patent No. 49278 (Nov. 26, 1909). 6 Fauchais, P.: Understanding plasma spraying, J. Phys. D: Appl. Phys. 37 (2004) R86–R108. 7 Pande, P. S., Neumann, R. P., Cavanagh, R. R.: The Six Sigma Way, McGraw-Hill (2000). 8 US EPA Capsule report: Approaching Zero Discharge in Surface Finishing, 09-20-2004, Chapter 5. 9 EN ISO 17836: Determination of the deposition efficiency for thermal spaying (ISO 17836 equivalent). 10 EN 1274: Thermal spraying – Powders – Composition – Technical supply conditions. 11 ISO 14232: Thermal spraying – Powders – Composition and technical supply conditions. 12 EN ISO 14919: Thermal spraying – Wires, rods and cords for flame and arc spraying – Classification – Technical supply conditions (equivalent to ISO 14919). 13 NISTIR 7014 Sept. 2003, FY 2003 Programs and Accomplishments, p. 37 ff. 14 UK DEF STAN 02-828, issue 2, Requirements for Thermal Spray Deposition of Metals and Ceramics for Engineering Purposes, Dec. 2005.

15 http://www.ts.nist.gov NIST Standard Reference Materials (SRM). 16 Lugscheider, E., Suk, H.-G., Lee, H.-K.: Vergleich verschiedener Messmethoden zur Bestimmung der Partikelgröße von Pulvern zum Plasmaspritzen. Schweißen und Schneiden 50 (1998), p. 72–731. 17 Xu, R., Di Guida, O. A.: Comparison of sizing small particles using different technologies. Powder Technology 132 (2003), p. 145–153. 18 Rawle, A.: Basic principles of particle size analysis, p. 8 (technical paper: www.malvern.co.uk). 19 EN ISO 4497: Metallic powders; determination of particle size by dry sieving. 20 ASTM B 214: Test Method for Sieve Analysis of Granular Metal Powders. 21 ISO 3310-1: Test sieves – Technical requirements and testing – Part 1: Test sieves of metal wire cloth (equivalent EN 23310-1). 22 ISO 4490: Determination of flow time by means of a calibrated funnel (Hall flowmeter). 23 US Patent 4,416,421 – 1983, James A. Browning. 24 Voronetski, A. V., Belashchenko, V.: Analysis of potential improvements of HVOF based processes. To be published in proceedings of ITSC – 2006. 25 Finnerty, L.: Ethanol Spells Price Relief, The Voice of Agriculture View, May 2 2005. 26 Höhle, H.-M.: “Gas- and Liquid-FuelHVOF-Technology Philosophy and Facts” Customer Day 7. October 2004 Eisenach.

245

246

References 27 Schmid, R. K. et al.: New HVOF Developments at Sulzer Metco. Proceedings of the 6th HVOF Conference, 27–28 November 2003, Erding, Germany, p. 149/155. 28 Legoux, J.-G. et al.: Evaluation of Four High Velocity Thermal Spray Guns using WC-10% Co-4% Cr Cermets. Journal of Thermal Spray Technology, Vol. 11, No. 1, p. 86–94. 29 International patent application No. 10/844,110 filed on 12 May 2004: Vladimir Belashchenko, Andrei Voronetski. Combustion Apparatus for High Velocity Thermal Spraying. 30 International patent application No. 60/602,253 filed on 17 August 2004: Vladimir Belashchenko, Andrei Voronetski. Method and Apparatus for Thermal Spray Coating. 31 Dolatabadi, A. Pershin, V., Mostaghimi, J.: New Attachment for Controlling Gas Flow in the HVOF Process. Journal of Thermal Spray Technology, Vol. 14 (1), March 2005, p. 91–99. 32 US Patent Application 2003/0178511 A1. A. Dolatabadi, J. Mostaghimi, V. Pershin. High Efficiency Nozzle for Thermal Spray of High Quality, Low Oxide Content Coatings. 33 US Patent Application 2005/0199739 A1. Seiji Kuroda, Takeshi Fukushima. Method of Forming Metal Coating with HVOF Spray Gun and Thermal Spray Apparatus. 34 Moreau, C., Lima, R., Marple, B.: Diagnostics for Advanced Materials Processing by Plasma Spraying Industrial Materials Institute National Research Council Canada. Presented at ISPC-16, Taormina, Sicily, Italy, 22–27 June 2003. 35 Leblanc, L., Moreau, C.: Study on the Long-term Stability of Plasma Spraying, Proceedings of the 1st International Thermal Spray Conference, 8–11 May 2000. Montreal. 36 Bisson, J. F., Gauthier, B., Moreau, C.: Effect of plasma fluctuations on inflight particle parameters, Proceedings of the International Thermal Spray Conference, 28–30 May 2001, Singapore. 37 Dorier, J.-L., Hollenstein, C., Salito, A., Loch, M., Barbezat, G.: Influence of External Parameters on Arc Fluctuations

in a F4 DC Plasma Torch Used for Thermal Spraying, Thermal Spray Surface Engineering via Applied Research. Proceedings of the 1st International Thermal Spray Conference, 8–11 May 2000. Montreal. 38 US Patent 4,788,402. November 29, 1988. James A. Browning. High power extended arc plasma spray method and apparatus. 39 US patent 4,841,114. June 20, 1989. James A. Browning. High-velocity controlled temperature plasma spray method and apparatus. 40 US patent 4,916,273. April 10, 1990. James A. Browning. High velocity controlled temperature plasma spray method. 41 US patent 6,114,649. September 5, 2000. Lucian Bogdan Delcia. Anode electrode for Plasmatron structure. 42 US patent 6,202,939. March 20, 2001. Lucian Bogdan Delcia. Sequential feedback injector for thermal spray torches. 43 US patent 6,392,189. May 21, 2002. Lucian Bogdan Delcia. Axial feedstock injector for thermal spray torches. 44 US patent 5,008,511. April 16, 1991. Douglas A. Ross. Plasma torch with axial reactant feed. 45 US patent 5,514,848. May 7, 1996. Douglas Ross, Alan Burgess. Plasma torch electrode structure. 46 US patent 5,556,558. September 17, 1996. Douglas Ross, Alan Burgess. Plasma jet converging system. 47 Boulos, M. I., Fauchais, P., Pfender, E.: Thermal Plasmas. Fundamentals and Applications. Vol. 1. Plenum Press. New York and London. 1994. 48 USSR Certificate of Invention 125323, December 18, 1958. I. D. Kulagin, A. V. Nikolaev. Method of generating of a high temperature gas jet. 49 Solonenko, O. P., Zhukov, M. F. (eds.): Thermal Plasma and New Materials Technology. Vol. 2: Investigations and Design of Thermal Plasma Technologies. Cambridge Interscience Publishing, 1995. 50 Low Temperature Plasma, Vol. 17: Electrical Arc Generators of Thermal Plasma. M. F. Zukov et al. Russian Academy of Science, “Nauka”, 1999.

References 51 US Patent 4,780,591. October 25, 1988. Bernecki, T. F. et al. Plasma gun with adjustable cathode. 52 US Patent 5,225,652. July 6, 1993. Klaus Landes. Plasma spray apparatus for spraying powdery or gaseous material. 53 US Patent 5,332,885. July 26, 1994. Klaus Landes. Plasma spray apparatus for spraying powdery or gaseous material. 54 Barbezat, G., Zierhut, J., Landes, K. D.: Triplex – a high performance plasma torch, Proceedings of United Thermal Spray Conference, Dusseldorf, March 17–19, 1999, p. 271–274. 55 Barbezat, G., Landes, K.: Plasma Technology TRIPLEX for the Deposition of Ceramic Coatings in the Industry. Proceedings of the 1st International Thermal Spray Conference. 8–11 May 2000, Montreal, Canada, p. 881–883. 56 Hawley, D., Refke, A., Landes, K. D.: Plasma plume condition forming of cascaded type plasma guns. Proceedings of International Thermal Spray Conference. Switzerland, Basel, May 2005. 57 Solonenko, O. P., Smirnov, A. V., Belashchenko, V.: Advantages of Cascade Plasma Generators for Thermal Spraying. 58 EN 14665: Thermal spraying – Thermal sprayed coatings – Symbolic representations on drawings. 59 Huang, X., Xu, L., Ma, X., Fan, B., Yin, X.: The Influence of Several Powder Feed Parameters on Feed Precision and Continuity, Proceedings of 15th International Thermal Spray Conference – Thermal Spray: Meeting the Challenges of the 21st Century 1 (1998), p. 779–784. 60 Gruner, H., Gruner, P., Moens, J.: Optimized powder feeding, a key parameter for thermal spraying, Proceedings of ITSC 2005 Thermal Spray connects: Explore its surfacing potential! (2005), p. 1558–1560. 61 Vardelle, M., Vardelle, A., Fauchais, P., Boulos, M. I.: “Particle Dynamics and Heat Transfer Under Plasma Conditions”, AIChE J., 34 (4), p. 567–573 (1988).

62 Saffmam, P. G.: The lift on a small sphere in a slow shear flow. J. Fluid. Mech, 22, 1965, p. 385–400. 63 Finke, J. R. Williamson, R. L. Chang, C. H.: Plasma spraying of functionally graded materials: measured and simulated results, In: Themal Spray: Surface Engineering via Applied Research, C. C. Berndt (Ed.) ASM thermal spray society, 2000, p. 141–148. 64 Voronetski, A. V., Suchkov, S. A., Filimonov, L. A., Belashchenko, V. E.: Modeling and Optimization of HVOF Processes, Proceedings of ITSC 2004 Thermal Spray Solutions – Advances in Technology and Applications (2004), p. 317–322. 65 Choquet, I., Bjoerklung, S., Johansson, J., Wigren, J.: Proceeding of ITSC 2005 Thermal Spray Connects: Nozzle exit geometry and lump formation in APS, ITSC 2005, p. 259–264. 66 Browning, J. A.: Maximum Combustion Energy Air Fuel Internal Burner, US Patent 5,120,582 (1992). 67 Belaschenko, V. Baranovsky, V.: Thermal Spray Systems, US Patent 5932293. 68 prEN 1395-6: Acceptance inspection of thermal spraying equipment – Part 6: Manipulator systems. 69 EN 13507: Thermal spraying – Pretreatment of surfaces of metallic parts and components for thermal spraying. 70 ISO 4287: Geometrical Product Specification (GPS) – Surface texture: Profile method – Terms, definitions and surface texture parameters. 71 EN ISO 4288: Geometrical Product Specifications (GPS) – Surface texture: Profile method – Rules and procedures for the assessment of surface texture. 72 ISO 8503-1: Preparation of steel substrates before application of paints and related products; surface roughness characteristics of blast-cleaned steel substrates; part 1: specifications and definitions for ISO surface profile comparators for the assessment of abrasive blast-cleaned surfaces. 73 Steffens, H.-D., Höhle, H.-M.: Further Development of the Thermal Spraying Technique by Low-Pressure Plasma Spraying with Transferred Arc, Proceedings of 9th International

247

248

References Thermal Spraying Conference – General Aspects of Thermal Spraying (1980), p. 420–428. 74 Itoh, A., Takeda, K., Itoh, M., Koga, M.: Pretreatment of Substrates by Using Reversed Transferred Arc in Low Pressure Plasma Spray, Proceedings of 3rd National Thermal Spray Conference – Thermal Spray: Research and Applications (1990), p. 245–251. 75 Honda, K., Chida, I., Fuse, T., Murakami, T., Kobayashi, K. F.: Effect on Reverse Transferred Arc Treating on Bonding Strength in Low Pressure Plasma Spraying, Proceedings of 9th National Thermal Spray Conference – Thermal Spray: Practical Solutions for Engineering Problems (1996), p. 405–411. 76 Heimann, R. B.: Plasma-Spray Coating – Principles and Applications (1996), p. 339. 77 Trifa, F.-I., Montavon, G.,Coddet, C.: Integrating a deposition model for off-line spray tools programming, Proceedings of ITSC 2005 Thermal Spray connects: Explore its surfacing potential! (2005), p. 280–285. 78 Deng, S., Liao, H., Zeng, C., Charles, P., Coddet, C.: development of robotic trajectories autogeneration in thermal spraying a new extended program of abb robotstudio™, Proceedings of Les Deuxièmes Rencontres Internationales sur la Projection Thermique (2005), p. 97–104. 79 Kratochvil, W. R., Sohr, J. M.: Removal of Thermal Spray Coatings with Ultra High Presssure Water Jets, Proceedings of 13th International Thermal Spray Conference – Thermal Spray: International Advances in Coatings Technology (1992), p. 285–295. 80 Blades, B., Sohr, J. M.: Water Jet Coating Removal Applications Aircraft Engines and Beyond, Proceedings of 5th National Thermal Spray Conference – Thermal Spray Coatings: Research, Design and Applications (1993), p. 275–278. 81 Kreye, H., Ahlhorn, T., Niebergall, J., Schwetzke, R.: Removal of Thermal Spray Coatings by High-Pressure Water Jets, Proceedings of 14th International Thermal Spray Conference: Thermal

Spraying – Current Status and Future Trends 1 (1995), p. 127–131. 82 van Wonderen, M.: Ultra High Pressure Waterjet Stripping of Thermal Sprayed Aircraft Engine Parts, Proceedings of 14th International Thermal Spray Conference: Thermal Spraying – Current Status and Future Trends 1 (1995), p. 157–162. 83 Mabrouki, T., Raissi, K.: Stripping process modeling: interaction between a moving waterjet and coated target, International Journal of Machine Tools and Manufacture 42 (2002), p. 1247–1258. 84 Bach, F.-W., Brüggemann, P., Louis, H., Schenk, A., Versemann, R.: Dry ice blasting and water jet processes for the removal of thermal sprayed coatings, Proceedings of ITSC 2005 Thermal Spray connects: Explore its surfacing potential! (2005), p. 1542–1548; 1549–1555. 85 Scrivani, A., Soranzo, M., Rizzi, G., Bardi, U., Carafiello, L.: On the Stripping of Turbine Blades and Vanes: Mechanisms and Performances of Chemical Stripping for the Removal of NiCrAlY Thermal Spray Coatings, Proceedings of ITSC 2003 International Thermal Spray Conference – Advancing the Science and Applying the Technology 1 (2003), p. 745–748. 86 Menini, R., Salah, N. B., Nciri, R.: Stripping Methods Studies for HVOF WC-10Co-4Cr Coating Removal, Journal of Materials Engineering 13 (2004), p. 185–194. 87 Huber, P., Kress, H., Kaiser, H.: Vacuum Plasma Spraying in Combination with Hot Isostatic Pressing for the Repair or Fabrication of Superalloy Components, Proceedings of 12th International Conference on Thermal Spraying 2 (1989), p. 51–59. 88 Burke, J.: “RECAST” – The Repair of Airfoils When Welding is Unacceptable, Proceedings of ITSC 2003 International Thermal Spray Conference – Advancing the Science and Applying the Technology 1 (2003), p. 731–733. 89 Su, C. Y., Lih, W. C., Chou, C. P., Tsai, H. C.: Activated diffusion brazed repair for IN738 hot section components

References of gas turbines, Journal of Materials Processing Technology 115 (2001), p. 326–332. 90 Williams, A. D., Humphries, J. L.: Latest Gas Turbine Repair Techniques Using Laser Technology, Proceedings of 15th International Thermal Spray Conference – Thermal Spray: Meeting the Challenges of the 21st Century 2 (1998), p. 1431–1435. 91 Suidu, T., Harada, Y., Teratani, T., Okazaki, M.: Optimization of the repair thermal spray coatings for the damaged single crystal superalloy, Proceedings of ITSC 2004 Thermal Spray Solutions – Advances in Technology and Applications (2004), p. 9–12. 92 Okazaki, M., Ohtera, I., Harada, Y.: Damage repair in CMSX-4 alloy without fatigue life reduction penalty, Metallurgical and Materials Transactions A: Physical Metallurgy and Materials 35 (2004), p. 535–542. 93 Daleo, J. A., Ellison, K. A., Boone, D. H.: Metallurgical considerations for life assessment and the safe refurbishment and requalification of gas turbine blades, Journal of Engineering for Gas Turbines and Power 124 (2002), p. 571–579. 94 ISO 14924: Thermal spraying – Pretreatment and finishing of thermally sprayed coatings. 95 Nelson, L. S.: The Shewhart control chart-tests for special cases, Journal of Quality Technology, 1984, 16(4), p. 237–239. 96 Dorier, J.-L., Hollenstein, C., Salito, A., Loch, M., Barbezat, G.: in 1st International Thermal Spray Conference – Thermal Spray: Surface Engineering via Applied Research (ed. Berndt, C. C.) 37–43 (ASM International, Materials Park, OH 44073-0002, Montréal, Québec, Canada, 2000). 97 Vardelle, A., Fauchais, P., Dussoubs, B., Themelis, N. J.: Heat Generation and Particle Injection in a Thermal Plasma Torch. Plasma Chemistry and Plasma Processing 18, 551–574 (1998). 98 Vardelle, A., Fauchais, P., Themelis, N. J.: Oxidation of Metal Droplets in Plasma Sprays, Proceedings of 8th National Thermal Spray Conference – Advances in Thermal

Spray Science and Technology (1995), p. 175–180. 99 Denoirjean, A., Lagnoux, O., Fauchais, P., Sember, V.: Oxidation Control in Atmospheric Plasma Spraying: Comparison between Ar/H2He and Ar/H2 Mixtures, Proceedings of 15th International Thermal Spray Conference – Thermal Spray: Meeting the Challenges of the 21st Century 1 (1998), p. 809–814. 100 Volenik, K., Chraska, P., Hanousek, F., Ilavsky, J. Neufuss, K.: In-Flight Oxidation of High-Alloy Steels during Plasma Spraying. Materials Science and Engineering: A 272, 199–206 (1999). 101 Palka, V., Brezovsky, M., Ivan, J., Sith, J.: “Identification of the Oxides in Plasma Sprayed APS Coating of the NiCrAlY Type” in 13th International Thermal Spray Conference – Thermal Spray: International Advances in Coatings Technology (ed. Berndt, C. C.), p. 537–542 (ASM International, Materials Park, OH 44073-0002, Orlando, Florida USA, 1992). 102 Voleník, K., Leitner, J., Hanousek, F., Dubsky, J. Kolman, B.: Oxides in Plasma-Sprayed Chromium Steel. Journal of Thermal Spray Technology 6, 327–334 (1997). 103 Wan, Y. P., Jiang, X. Y., Fincke, J. R., Herman, H., Prasad, V., Sampath, S.: Modeling of Oxidation of Molybdenum Particles during Plasma Spraying, Metallurgical and Materials Transactions B: Process Metallurgy 32 (2001), p. 475–481. 104 Fincke, J., Xiong, H., Zheng, L. L., Sampath, S.: Melting/oxidation behavior of in-flight particles in plasma spray processes, ASME Journal of Heat Transfer 372 (2002). 105 Kolman, D., Volenik, K.: Modeling of Oxidation During Plasma Spraying of Iron Particles, Plasma Chemistry and Plasma Processing 22 (2002), p. 437–450. 106 Syed, A. A., Denoirjean, A., Denoirjean, P., Labbe, J. C., Fauchais, P.: In-Flight Oxidation of Metallic Particles in Plasma Spraying, Proceedings of ITSC 2003 International Thermal Spray Conference – Advancing the Science and Applying the Technology 2 (2003), p. 985–992.

249

250

References 107 prEN 1395-6: Thermal spraying – Acceptance inspection of thermal spraying equipment – Part 6: Manipulator systems. 108 Fauchais, P., Vardelle, A., Vardelle, M., Bianchi, L., Léger, A. C.: Parameters controlling the generation and properties of plasma sprayed zirconia coatings. Plasma Chemistry and Plasma Processing, Springer 1995, Vol. 16, Suppl. 1, p. S99–S125. 109 ISO 3882: Metallic and other inorganic coatings – Review of methods of measurement of thickness. 110 ISO 1463: Metallic and oxide coatings – Measurement of coating thickness – Microscopical method. 111 ISO 2360: Non-conductive coatings on non-magnetic basis metals; Measurement of coating thickness; Eddy current method. 112 Siegmann, S., Margadant, N., Zagorski, A., Arana-Antelo, M.: Comparison of Particle In-Flight Characteristics and Coating Properties. in ITSC 2003 International Thermal Spray Conference – Advancing the Science and Applying the Technology, Vol. 2 (Marple, B. R., Moreau, C., eds.), p. 1591–1598, ASM International. 113 Pyzdek, Th.: The six sigma handbook, McGraw-Hill, New York, 2001. 114 Bisson, J. F., Moreau, C., Dorfman, M., Dambra, C., Mallon, J.: Influence of secondary hydrogen on the microstructure of plasma sprayed Yttriastabilized Zirconia, Thermal Spray 2003: Advancing the science and applying the technology, Ed. C. Moreau, B. Marple, ASM Intl, Materials Park, Ohio, USA, 2003, p. 1229–1235. 115 Moreau, C., Lima, R., Marple, B.: Diagnostics for Advanced materials processing by plasma spraying, ISPC-16, Taormina, Sicily, Italy, 22–27 June 2003. 116 Bisson, J. F., Gauthier, B., Moreau, C.: Effect of plasma fluctuations on in-flight particle parameters, ITSC 2005, Basel, 2005. 117 Kudinov, V. V.: Plasma coatings, Nauka, Moscow, 1977 (in Russian). 118 Kudinov, V. V.: Analysis of bonding mechanisms during thermal spraying, ITSC 2005.

119 Pasandideh-Fard, M., Bhola, R., Chandra, S., Mostaghimi, J.: Deposition of tin droplets on a steel plate: simulation and experiments, Int. Journal. of Heat and Mass Transfer, 41, p. 2929–2945 (1998). 120 Armster, S. O., Delplanque, J.–P., Rein, M., Lavernja, E. J.: Thermo-fluid mechanisms controlling droplet based material process, Int. Materials Review, Vol. 47, No. 6, p. 265–301, 2002. 121 Fauchais, P.: Understanding plasma spraying, J. Phys D: Appl. Phys., 37(2004) R86–108. 122 Kudinov, V. V., Pekshev, P. Yu., Belashchenko, V. E., Solonenko, O. P., Safiullin, V. A.: Coating applications by plasma, Nauka, Moscow, 1990. 123 Papyrin, A. N., Alkimov, A. P., Kosarev, V. F., Nesterovich, N. I.: “Method of Applying Coatings”; Russian Patent No. 1,618,778; 1990 (priority of the Invention: June 6, 1986). 124 Vardelle, A., Vardelle, M.: Study of the influence of particle temperature and velocity distribution within a plasma jet coating formation, p. 155–161. 125 Solonenko, O. P., Kudinov, V. V., Smirnov, A. V., Cherepanov, A. N., Popov, V. V., Mikalchenko, A. A., Kartaev, E. V.: Micro-metallurgy of splats: theory, computer simulation and experiment, JSME Int. J. B, 48, No. 3, 2005. 126 Zagorski, A., Stadelmaier, V.: Full-scale modeling of a plasma spray process, Surface and coating technology, 146–147, p. 162–167 (2001). 127 Sobolev, V., Guilemany, J. M.: Effect of droplet impact angle on flattening of splat in thermal spraying, Materials Letters 32 (1997), p. 197–201. 128 Bengtsson, P., Persson, Ch.: Modeled and measured Residual stresses in plasma sprayed thermal barrier coatings, Surface and coating technologies, 92, 1977, p. 78. 129 Kim, H.-Y., Feng, Z. C., Chun, J.-H.: Instability of the liquid jet emerging from a droplet upon collision with a solid surface, Physics of Fluids, 12, No. 3, 2000, p. 531–541. 130 Allen, R. F.: The role of surface tension in splashing, J. Colloid Interface Sci., 51, 350 (1975).

References 131 Raessi, M., Mostaghimi, J., Bussmann, M.: “Droplet Impact during the Plasma Spray Coating Process – Effect of Surface Roughness on Splat Shapes”, Proceedings of ISPC 17 Conference, Toronto (2005). 132 Handbook of Thermal Spray Technology. Ed. J. R. Davis, ASM International, 2004. 133 Pawlowski, L.: The science and engineering of thermal spray coatings, John Wiley & Sons, 1995. 134 Pfender, E.: Fundamental studies associated with the plasma spray process, Surface and Coatings Technology, 34, 1988, p. 1–14. 135 Zhukov, M. F.: Electric arc generators of thermal plasma: Review, Thermal Plasma and New Materials Technology, Vol. 1. Investigation and Design of Thermal Plasma Generators, Eds.: O. P Solonenko, M. F. Zhukov, Cambridge Interscience Publishing, Cambridge, England, 1995. p. 6–23. 136 Zeldovich, Ya. B., Raizer, Yu. P.: Physics of shock waves and high-temperature phenomena, Academic Press, New York, 1967. 137 Raizer, Y. P.: Gas Discharge Physics, Springer, Berlin, 1991. 138 Boulos, M. I., Fauchais, P., Pfender, E.: Thermal Plasmas (Advances in Experimental Medicine & Biology) (Hardcover), Plenum US, 1994. 139 Hirschfelder, J. O., Curtis, Ch. F., Bird, R. B.: The Molecular Theory of Gases and Liquids, John Wiley & Sons, 1964. 140 Crocco, L.: One-dimensional treatment of steady gas dynamics. In Fundamentals of Gas. Dynamics, ed. H. W. Emmons, p. 64–349, 1958. 141 Soloukhin, R. I.: Handbook of Radiative Heat Transfer in High-Temperature Gases, CRC Press, 1987. 142 Landau, L. D., Lifshits, E. M.: Fluid Mechanics: Vol. 6 (Course of Theoretical Physics), Butterworth-Heinemann, 1987. 143 Schlichting, H.: Boundary layer theory, New York, McGraw-Hill, 1955. 144 Spores, R., Pfender, E., Heberlein, J.: Fluid-dynamic characteristics of a plasma spray jet, in Plasma Jets in the Development of New Materials Technology, Ed. O. P. Solonenko,

A. I. Fedorchenko, VSP Publ, 1990, p. 699–716. 145 Zagorski, A., Szuecs, F., Belashchenko, V., Siegmann, S., Margadant, N., Ivanov, A.: Experimental Study of substrate thermal conditions at APS and HVOF. Proceedings of 2003 International Thermal Spray Conference, Orlando, Florida, USA, May 5–8, 2003, p. 1255–1260. 146 Schwenk, A., Gruner, H., Zimmermann, S., Landes, K., Nutsch, G. (2004): Improved Nozzle Design of de-Lavaltype Nozzles for the Atmospheric Plasma Spraying, in ITSC 2004 Thermal Spray Solutions – Advances in Technology and Applications (IIW, D. T., ed.), p. 17–22, DVS. 147 Schwenk, A., Mihm, S., Nutsch, G., Wank, A., Gruner, H.: Modified supersonic nozzles for the vacuum plasma spraying, Proceedings of ITSC 2005 Thermal Spray connects: Explore its surfacing potential! (2005), p. 363–368. 148 Abramovich, G. N.: Applied Gas Dynamics, Nauka, Moscow, 1991. 149 Lee, Y. C., Chyou, Y. P., Pfender, E.: “Particle Dynamics and Particle Heat and Mass Transfer in Thermal Plasmas – Part II: Particle Heat and Mass Transfer in Thermal Plasmas,” Plasma Chemistry and Plasma Processing, Vol. 5, No. 4, 1985, p. 391–416. 150 Refke, A., Hawley, D., Doesburg, J., Schmid, R. K.: LPPS thin film technology for the application of TBC systems, ITSC 2005, Basel, 2005. 151 Loch, M., Barbezat, G.: Characteristics and potential applications of thermally sprayed thin film coatings, Proc. 1st ITSC, Montreal, Canada, 2000, p. 1141–1144. 152 L. M. Biberman, V. S. Vorob’Ev, I. T.: Yakubov, Kinetics of Nonequilibrium Low-Temperature Plasmas, Consultants Bureau, 1987. 153 Pfender E., Lee, Y. C.: “Particle Dynamics and Particle Heat and Mass Transfer in Thermal Plasmas – Part I: The Motion of a Single Particle without Thermal Effects,” Plasma Chemistry and Plasma Processing, Vol. 5, No. 3, 1985, p. 211–237.

251

252

References 154 Belashchenko, V. E., Chernyak, Y. B.: Stochastic approach to the modeling and optimization of Thermal spray coating formation, Journal Thermal Spray Technology, 2(2), p. 159–164, 1993. 155 Janisson, S., Melliot, E., Vardelle, A., Coudert, J. F., Pateyron, B., Fauchais, P.: Plasma spraying using Ar-He-H2 gas mixtures. Proc. of 15th Int. Thermal Spray Conference, Nice, France, 1996, p. 803–808. 156 Vardelle, M., Vardelle, A., Fauchais, P., Boulos, M.: Plasma – particle momentum and heat transfer: modeling and measurements, AIChE Journal, 29, No. 2, p. 238–243, 1983. 157 Eichert, P., Coddet, C.: On the modeling approach of DC Arc plasma flows by implementation of the CFD Phoenix code, Thermal spray: Practical solutions for engineering problems. Ed. CC. Berndt, p. 563–568. ACM Intl, Materials Park, Ohio-USA, 1996. 158 Varacalle, D. J., Jr., Neiser, R. A., Smith, M. F.: Modeling of the Vacuum Plasma Spray Process, Proceedings of 13th International Thermal Spray Conference – Thermal Spray: International Advances in Coatings Technology (1992), p. 781–785. 159 Westhoff, R., Trapaga, G., Szekely, J.: Plasma-particle interaction in Plasma Spraying Systems, Metallurgical Trans, B, 23B, p. 683–693, 1992. 160 Apelian, D., Wei, D., Paliwal, M.: Particle-plasma interaction during low pressure plasma deposition, Thin Solid Films, 118, 395–407, 1984. 161 Varacalle, D. J., Jr., Barnett, K.: A Complete Predictor Methodology for the Plasma Spray Process, Proceedings of 9th National Thermal Spray Conference – Thermal Spray: Practical Solutions for Engineering Problems (1996), p. 709–715. 162 Hurevich, V., Gusarov, A., Smurov, I.: Simulation of coating profile under plasma spraying conditions, Proceedings of ITSC 2002 International Thermal Spray Conference, Vol. 1, 2002, p. 318–323. 163 Williamson, R. L., Finke, J. R., Chang, C. H.: A computational examination of the sources of statistical

variance in particle parameters during thermal plasma spraying, Plasma Chemistry and plasma Processing, 20, No. 3, p. 299–324, 2000. 164 Madejski, J.: Solidification of droplets on a cold surface, Int. J. Heat Mass Transfer, 19, 1009–1013 (1976). 165 Solonenko, O. P.: Plasma spraying coatings from oxide powders: thermo physical fundamentals, Journal of Physical Mesomechanics, 2001, Vol. 4, No. 6, p. 45–56. 166 Pasandiedeh-Fard, M., Bhola, R., Chandra, S., Mostaghimi, J.: Deposition of thin droplets on a steel plate: simulations and experiments, International Journal of Heat and Mass transfer, Vol. 41, 2929–2945, 1998. 167 Heichal, Y., M. Xue, S. Chandra, J.: Mostaghimi: Measuring thermal contact resistance between molten aluminum droplets and a steel plate, Proceedings of ITSC 2005 Thermal Spray connects: Explore its surfacing potential! (2005), p. 810–814. 168 Trapaga, G., Matthys, E. F., Valencia, J. J., Szekely, J.: Fluid flow, heat transfer and solidification of molten metal droplets impinging on substrates: comparison of numerical and experimental results, Metallurgical transactions B, 23B, p. 701–718 (1992). 169 Maruo, H., Hirata, Y., Matsumoto, Y.: Deformation and Solidification of a Molten Droplet by Impact on a Planar Substrate, Proceedings of 14th International Thermal Spray Conference: Thermal Spraying-Current Status and Future Trends, Vol. 1, 1995, p. 341–346. 170 Watanabe, T., Kuribayaashi, I., Honda, T., Kanzawa, A.: Deformation and solidification of a droplet on a cold substrate, Chemical Engineering Science, 47, p. 3059–3065, 1992. 171 Liu, H., Lavernia, E. J., Rangel, R. H.: Numerical Simulation of Impingement of Molten Titanium, Nickel, and Tungsten Droplets on a Flat Substrate, Journal of Thermal Spray Technology 2 (1993), p. 369–378. 172 H. Liu, Lavernja, E. J., Rangel, R. H., Modeling of molten droplet impingement on a non-flat surface, Acta metal. Mater. 43, No. 5, p. 2053–2072, 1995.

References 173 Cirolini, S., Harding, J. H., Jacucci, G.: Computer simulation of plasma sprayed coatings I. Coating deposition model. Surface and coating technology, 48, p. 137–145 (1991). 174 Borgerding, B., Soelter, H. J., Lugscheider, E., Simhan, K.: Thermal Spraying: Modeling of Temperature Gradients and Stress-Strain Distributions during the Plasma Spraying Process, PMI, Vol. 24, No. 4, 1992. 175 Ferrari, M., Harding, J. H., Marchese, M.: Computer simulation of plasma-sprayed coatings II. Effective bulk properties and thermal stress calculations, Surface and Coating Technology, 48, 147–154, 1991. 176 Liu, H., Lavernia, E. J., Rangel, R. H.: Numerical investigation of micropore formation during substrate impact of molten droplets in plasma spray process. Atomization and sprays, 4, p. 369–384, 1994. 177 Fukanuma, H.: An Analysis of the Porosity Producing Mechanism, Proceedings of 13th International Thermal Spray Conference – Thermal Spray: International Advances in Coatings Technology (1992), p. 767–772. 178 Solonenko, O. P., Ohmori, A., Matsuno, S., Smirnov, A. V.: Deformation and Solidification of Melt Microdroplets Impinging on Substrates – Theory and Experiment – Proceedings of 14th International Thermal Spray Conference: Thermal Spraying – Current Status and Future Trends 1 (1995), p. 359–364. 179 Solonenko, O. P., Mikhalchenko, A. A., Kartaev, E. V., Ogawa, K., Shoji, T.: Model studying zirconia droplet deposition and solidification on substrate under plasma spraying: theory and experimental verification, Proceedings of 5th ISME-KSME Fluids Engineering Conference, 2002, Nagoya, Japan, 17–21 Nov. 2002, p. 1–7. 180 Pasandideh-Fard, M., Pershin, V., Chandra, S., Mostaghimi, J.: Splat Shapes in a Thermal Spray Coating Process: Simulations and Experiments, Journal of Thermal Spray Technology 11 (2002), p. 206–217.

181 Pershin, V., Lufitha, M., Chandra, S., Mostaghimi, J.: Effect of Substrate Temperature on Adhesion Strength of Plasma-Sprayed Nickel Coatings, Journal of Thermal Spray Technology 12 (2003), p. 370–376. 182 Salimijazi, H. R., Raessi, M., Coyle, T. W., Mostaghimi, J.: Numerical Simulation and Experimental Evaluation of Solidification Behaviour of Vacuum Plasma Sprayed Ti-6Al–4V Alloy, Proceedings of 17th International Symposium on Plasma Chemistry, ISPC17 (2005), p. 6. 183 Nassenstein, K.: Progress in thermal spray processes, Proceedings of ITSC 2005 Thermal Spray connects: Explore its surfacing potential! (2005), p. 378–382. 184 Deng, S., Liao, H., Zeng, C., Charles, P., Coddet, C.: Robotic trajectory autogeneration in thermal spraying, Proceedings of ITSC 2005 Thermal Spray connects: Explore its surfacing potential! (2005), p. 481–485. 185 Trifa, F.-I., Montavon, G., Coddet, C.: Integrating a deposition model for offline spray tools programming, Proceedings of ITSC 2005 Thermal Spray connects: Explore its surfacing potential! (2005), p. 280–285. 186 Fasching, M. M., Weiss, L. E., Prinz, F. B.: Optimization of Robotic Trajectories for Thermal Spray Shape Deposition, Proceedings of 13th International Thermal Spray Conference – Thermal Spray: International Advances in Coatings Technology (1992), p. 221–226. 187 Nylen P., Wigren, J., Pedjrid, L., Hansson, M.-O.: Modelling of coating thickness, heat transferr and fluid flow and it’s correlation to the TBC mocrostructure for a plasma sprayed gas turbine application. In: Proc. 15th Int. Therm. Spray Conf., 25–29 May 1998, Nice, France, p. 361–366. 188 Hansbo A., Nylen P. E.: Modeling of spray deposition and robot motion optimization for a gas turbine application. Proceedings of the 1st International Thermal Spray Conference, 8–11 May 2000. Montreal, p. 203–210.

253

254

References 189 Smirnov, B. M., Chibisov, M. I.: Resonance charge exchange in noble gases. Zurn. Techn. Fizik, Vol. 35, No. 1, p. 117 (in Russian). 190 Neumann, W.: Beiträge aus der Plasma Physik, B2, h2, 1962, P80–115. 191 Brown, S.: Elementary Processes in Gas Discharge Plasma. MIT press, Cambridge, MA, 1959. 192 Abramovich, G. N.: The Theory of Turbulent Jets. MIT Press, 1963. 193 Mariaux, G., Vardelle, A., Fauchais, P.: Usefulness of 3D time dependent models and steady 2D models in plasma spraying engineering? Proceedings of ITSC 2005 Thermal Spray connects: Explore its surfacing potential! (2005), p. 351–356. 194 Glasnev, V. N., Sulemenov, Sh.: Gas dynamic parameters of weekly underexpanded jets, Novosibirsk, Nauka, 1980 (in Russian). 195 Gnedovets, A. G.: in O. P. Solonenko, M. F. Zhukov (eds.) Thermal plasma and new materials technology, Vol. 1, CIP, Cambridge, 1995, p 134. 196 Chambre, P. L., Schaaf, S. A.: Rarefied gas flows, In Fundamentals of Gas Dynamics, ed. H. W. Emmons, p. 64–349, 1958. 197 Xi Chen: Particle heating in thermal plasma, Proc. 8th International symposium on plasma chemistry, Tokyo, P4–12, 1987. 198 Martin, H.: Heat and mass transfer between impinging gas jet and solid surfaces, In Advance in Heat transfer, Eds. J. P. Hartnett and T. F. Irvine, Jr., Vol. 13, Academic Press, New York, 1977, p. 1–60. 199 Downs, S. J., James, E. H.: Jet Impingement Heat Transfer – A Literature Survey, In: ASME paper no 87-HT-35, The National Heat Transfer Conference, ASME, Oct 1987, p. 1–11. 200 Ilavsky, J., Allen, A. J., Long, G. G., Krueger, S., Berndt, C. C., Herman, H.: Influence of Spray Angle on the Pore and Crack Microstructure of Plasma-Sprayed Deposits, Journal of the American Ceramic Society 80 (1997), p. 733–742. 201 Oberkampf, W. L., Talpallikar, M.: Analysis of a High Velocity Oxygen-Fuel

(HVOF) Thermal Spray Torch – Part 1: Numerical Formulation, Proceedings of 6th National Thermal Spray Conference – Thermal Spray Industrial Applications (1994), p. 381–386. 202 Oberkampf, W. L., Talpallikar, M.: Analysis of a High Velocity Oxygen-Fuel (HVOF) Thermal Spray Torch – Part 2: Computational Results, Proceedings of 6th National Thermal Spray Conference – Thermal Spray Industrial Applications (1994), p. 387–392. 203 Sobolev, V. V., Guilemany, J. M., Calero, J. A.: Theoretical Investigation of the Gas-Powder Particles Transport Phenomena in HVOF Spraying, Proceedings of 6th National Thermal Spray Conference – Thermal Spray Industrial Applications (1994), p. 393–398. 204 Chang, C. H., Moore, R. L.: Simulation of Gas Particle Flow in a HVOF Torch, Proceedings of 8th National Thermal Spray Conference – Advances in Thermal Spray Science and Technology (1995), p. 207–212. 205 Chang, C. H., Moore, R. L.: Numerical Simulation of Gas and Particle Flow in a High-Velocity Oxygen-Fuel (HVOF) Torch, Journal of Thermal Spray Technology 4 (1995), p. 358–366. 206 Yang, X., Eidelman, S., Lottati, I.: Numerical Simulation of Gas and Particle Flow Field Characteristics in HVOF Guns, Proceedings of 8th National Thermal Spray Conference – Advances in Thermal Spray Science and Technology (1995), p. 213–218. 207 Tawfik, H. H., Zimmerman, F.: Mathematical Modeling on the Gas and Powder Flow in HVOF Systems, Journal of Thermal Spray Technology 6 (1997), p. 345–352. 208 Yang, Y. M., Liao, H., Coddet, C.: Numerical Modeling of Combustion and Flow Characteristics in HVOF Systems with Chemical Equilibrium Approach, Proceedings of 14th International Thermal Spray Conference: Thermal Spraying – Current Status and Future Trends 1 (1995), p. 399–404. 209 Sobolev, V. V., Guilemany, J. M.: Dynamic Processes during HighVelocity Oxyfuel Spraying, International Materials Reviews 41 (1996), p. 13–32.

References 210 Smith, E. B., Power, G. D., Barber, T. J., Chiappetta, L. M.: Application of Computational Fluid Dynamics to the HVOF Thermal Spray Gun, Proceedings of 13th International Thermal Spray Conference – Thermal Spray: International Advances in Coatings Technology (1992), p. 805–810. 211 Hassan, B., Lopez, A. R., Oberkampf, W. L.: Computational Analysis of a Three-Dimensional High-Velocity Oxygen-Fuel (HVOF) Thermal Spray Torch, Proceedings of 8th National Thermal Spray Conference – Advances in Thermal Spray Science and Technology (1995), p. 193–198. 212 Joshi, S. V.: A Prediction Model to Assist Plasma and HVOF Spraying, Materials Letters 14 (1992), p. 31–36. 213 Dolatabadi, A., Pershin, V., Mostaghimi, J.: Modeling and Design of an Attachment to the HVOF Gun, Proceedings of ITSC 2003 International Thermal Spray Conference – Advancing the Science and Applying the Technology 1 (2003), p. 779–787. 214 Dolatabadi, A., Mostaghimi, J., Ivanovic, M.: Numerical Modeling of Particle Laden Flow in HVOF Torch with Gas Shroud, Proceedings of 1st International Thermal Spray Conference – Thermal Spray: Surface Engineering via Applied Research 1 (2000), p. 105–113. 215 Knotek, O., Schnaut, U.: Process Modeling of HVOF Thermal Spraying Systems, Proceedings of 13th International Thermal Spray Conference – Thermal Spray: International Advances in Coatings Technology (1992), p. 811–816. 216 Zheng, L., Zhang, H.: Modeling of particle in-flight and deposition during plasma and HVOF spraying, Proceedings of ICECA ’2001 Int. Conf. Energy Conversion and Application (2001), p. 147–153. 217 Li, M., Christofides, P. D.: Modeling and analysis of high-velocity oxygen fuel thermal spray process accounting for powder size distribution, Chemical Engineering Science 58 (2003), p. 849–857. 218 Gu, S., Eastwick, C. N., Simmons, K. A., McCartney, D. G.: Computational

Fluid Dynamic Modeling of Gas Flow Characteristics in a High-Velocity OxyFuel Thermal Spray System, Journal of Thermal Spray Technology 10 (2001), p. 461–469. 219 EN ISO 14922-1: Thermal spraying – Quality requirements of thermally sprayed structures – Part 1: Guidance for selection and use (equivalent ISO 14922-1). 220 EN ISO 14922-2: Thermal spraying – Quality requirements of thermally sprayed structures – Part 2: Comprehensive quality requirements (equivalent ISO 14922-2). 221 EN ISO 14922-3: Thermal spraying – Quality requirements of thermally sprayed structures – Part 3: Standard quality requirements (equivalent ISO 14922-3). 222 EN ISO 14922-4: Thermal spraying – Quality requirements of thermally sprayed structures – Part 4: Elementary quality requirements (equivalent ISO 14922-4). 223 EN 13214: Thermal spraying – Thermal spray coordination – Tasks and responsibilities. 224 ISO 17833: Thermal spraying – Thermal spray coordination – Tasks and responsibilities. 225 ISO 14918: Thermal spraying – Approval testing of thermal sprayers. 226 Böhme, D., Ohliger, A.: (SLV München), Handbuch der thermischen Spritztechnik, DVS Verlag (2002). 227 EN 657: Thermal spraying; terminology, classification. 228 ISO 14917: Thermal spraying – Terminology, classification. 229 EN 14616: Thermal spraying – Recommendations for thermal spraying. 230 ISO 14921: Thermal spraying – Procedures for the application of thermally sprayed coatings for engineering components. 231 AWS C2.25-XX, Specification for Solid and Composite Wires, and Ceramic Rods for Thermal Spraying, DS1, 7-11-2001. 232 ISO 2063: Metallic and other inorganic coatings; thermal spraying; zinc, aluminium and their alloys.

255

256

References 233 EN ISO 17834: Thermal spraying – Coatings for protection against corrosion and oxidation at elevated temperatures (equivalent ISO 17834). 234 EN 725-5: Advanced technical ceramics – Methods of test for ceramic powders – Part 5: Determination of the particle size distribution. 235 ISO 4490: Metallic powders; determination of flowability by means of a calibrated funnel (Hall flowmeter). 236 EN 13507: Thermal spraying – Pretreatment of surfaces of metallic parts and components for thermal spraying. 237 ISO 17835: Thermal spraying – Pretreatment of surfaces of metallic parts and components for thermal spraying. 238 EN ISO 14924: Thermal spraying – Posttreatment and finishing of thermally sprayed coatings (equivalent ISO 14924). 239 ISO 14924: Thermal spraying – Posttreatment and finishing of thermally sprayed coatings. 240 EN ISO 14920: Thermal spraying – Spraying and fusing of thermally sprayed coatings of self-fluxing alloys (equivalent ISO 14920). 241 EN 1395: Thermal spraying – Acceptance inspection of thermal spraying equipment. 242 ISO 14231: Thermal spraying – Acceptance inspection of thermal spraying equipment. 243 EN ISO 17836: Determination of the deposition efficiency for thermal spraying (equivalent ISO 17836). 244 DVS 2308: Grundsätze zur Konstruktion von Bauteilen und Werkstücken für das thermische Spritzen. 245 EN 14665: Thermal spraying – Thermal sprayed coatings – Symbolic representations on drawings. 246 ISO 14923: Thermal spraying – Characterization and testing of thermally sprayed coatings. 247 EN 582: Thermal spraying – Determination of tensile adhesive strength. 248 ISO 14916: Thermal spraying – Determination of tensile adhesive strength. 249 DIN 65144: Aerospace; thermally sprayed components; technical specification.

250 Margadant, N., Siegmann, S., Keller, T., Wagner, W., Kulkarni, A.: Insights to Spraying Conditions, Microstructure and Properties and Their Statistical Correlation for Different Thermal Spraying Processes Using Complementary Characterization Methods, Proceedings of ITSC 2003 International Thermal Spray Conference – Advancing the Science and Applying the Technology 2 (2003), p. 1053–1061. 251 Sundararajan, G., Sivakumar, G., Srinivasa Rao, D.: The Interrelationship between Particle Velocity and Temperature, Splat Formation and Deposition Efficiency in Detonation Sprayed Alumina Coatings, Proceedings of Thermal Spray 2001 – New Surfaces for a New Millennium (2001), p. 849–858. 252 Prystay, M., Gougeon, P., Moreau, C.: Structure of Plasma-Sprayed Zirconia Coatings Tailored by Controlling the Temperature and Velocity of the Sprayed Particles, Journal of Thermal Spray Technology 10 (2001), p. 67–75. 253 Nylén, P., Friis, M., Hansbo, A., Pejryd, L.: Investigation of Particle In-flight Characteristics during Atmospheric Plasma Spraying of Yttria Stabilized ZrO2: Part 2: Modeling, Journal of Thermal Spray Technology 10 (2001), p. 359–366. 254 Friis, M., Nylén, P., Persson, C., Wigren, J.: Investigation of Particle In-Flight Characteristics during Atmospheric Plasma Spraying of YttriaStabilized ZrO2 – Part 1: Experimental, Journal of Thermal Spray Technology 10 (2001), p. 301–310. 255 Srinivasa Rao, D., Sen, D., Somaraju, K. R. C., Ravi Kumar, S., Ravi, N., Sundararajan, G.: The Influence of Powder Particle Velocity and Temperature on the Properties of Cr3C2-25NiCr Coating Obtained by Detonation-Gun, Proceedings of 15th International Thermal Spray Conference – Thermal Spray: Meeting the Challenges of the 21st Century 1 (1998), p. 385–393. 256 Ilyuschenko, A. P., Okovity, V. A., Kundas, S. P., Kuz’menkov, A. N.: Modeling and Experimental Studies of Particles Velocity and Temperature in Plasma Spraying Processes, Proceedings

References of 1st United Thermal Spray Conference – Thermal Spray: A United Forum for Scientific and Technological Advances (1997), p. 657–663. 257 Wright, R. N., Fincke, J. R., Swank, W. D., Haggard, D. C.: Particle Velocity and Temperature Influences on the Microstructure of Plasma Sprayed Nickel, Proceedings of 9th National Thermal Spray Conference – Thermal Spray: Practical Solutions for Engineering Problems (1996), p. 511–516. 258 Prystay, M., Gougeon, P., Moreau, C.: Correlation between Particle Temperature and Velocity and the Structure of Plasma Sprayed Zirconia Coatings, Proceedings of 9th National Thermal Spray Conference – Thermal Spray: Practical Solutions for Engineering Problems (1996), p. 517–523. 259 Hantzsche, H.: Temperature Measurement of Powder Particles in a Plasma Jet, Proceedings of 7th International Spraying Conference 1 (1973), p. 1–4. 260 Fincke, J. R., Jeffery, C. L.: Simultaneous In-Flight Measurement of Particle Size, Velocity and Temperature, Proceedings of 2nd National Thermal Spray Conference – Thermal Spray Technology: New Ideas and Processes (1988), p. 55–60. 261 Mayr, W.: Rapid Optimization of Spraying Parameters by Means of an Automated Laser Doppler Measuring Equipment, Proceedings of 11th International Thermal Spraying Conference – Advances in Thermal Spraying (1986), p. 221–230. 262 Mayr, W., Landes, K., Reusch, A.: Untersuchung des Partikelverhaltens im Spritzstrahl mittels Laser-DopplerAnemometrie (LDA), Proceedings of Thermische Spritzkonferenz TS 93 (1993), p. 143–147. 263 Knight, R., Smith, R. W., Xiao, Z., Hoffman, T. T.: Particle Velocity Measurements in HVOF and APS Systems, Proceedings of 6th National Thermal Spray Conference – Thermal Spray Industrial Applications (1994), p. 331–336. 264 Reusch, A., Voggenreiter, H., Huber, H., Mayr, W.: Mobile Messung

von Partikeleigenschaften beim Thermischen Spritzen durch LaserDoppler-Anemometrie, Metall 49 (1995), p. 38–44. 265 Reusch, A., Mayr, W., Landes, K.: Mobiles Laser-Doppler-Messsystem zur Untersuchung des Partikelverhaltens in thermischen Spritzstrahlen, Technisches Messen 62 (1995), p. 277–284. 266 Mayr, W., Landes, K., Reusch, A., Beyer, S., Huber, H., Voggenreiter, H.: Praktische Erfahrungen mit dem Einsatz eines mobilen, automatisierten Laser-Doppler-Messsystems beim HP/HVOF-Verfahren, Proceedings of Thermische Spritzkonferenz TS 96 175 (1996), p. 219–223. 267 Hale, D. L., Swank, W. D., Haggard, D. C.: In-Flight Particle Measurements of Twin Wire Electric Arc Sprayed Aluminum, Journal of Thermal Spray Technology 7 (1998), p. 58–63. 268 http://www.controlvisioninc.com 269 Gougeon, P., Moreau, C.: In-Flight Particle Surface Temperature Measurement: Influence of the Plasma Light Scattered by the Particles, Proceedings of 5th National Thermal Spray Conference – Thermal Spray Coatings: Research, Design and Applications (1993), p. 13–18. 270 Yu, W., Cetegen, B. M.: Particle Temperature Measurements in a HVOF Thermal Spray by Two Color Radiant Emission Pyrometry, Proceedings of MRS Fall Meeting ’97 – The Science and Technology of Thermal Spray Materials Processing (1997), p. 557–558. 271 Hollis, K., Neiser, R.: Analysis of the Nonthermal Emission Signal Present in a Molybdenum Particle-Laden PlasmaSpray Plume, Journal of Thermal Spray Technology 7 (1998), p. 383–391. 272 Hollis, K., Neiser, R.: Particle Temperature and Flux Measurement Utilizing an Nonthermal Signal Correction Process, Journal of Thermal Spray Technology 7 (1998), p. 392–402. 273 Mingard, K. P., Alexander, P. W., Langridge, S. J., Tomlinson, G. A., Cantor, B.: Direct measurement of sprayform temperatures and the effect of liquid fraction on microstructure, Acta Materialia 46 (1998), p. 3511–3521.

257

258

References 274 Fischer, A., Seemann, K., Lugscheider, E.: Investigation of atmospheric plasma spray process by inflightparticle pyrometry and thermography, Proceedings of ITSC 2002 International Thermal Spray Conference 1 (2002), p. 782–787. 275 Gougeon, P., Moreau, C., Lacasse, V., Lamontagne, M., Powell, I., Vewsher, A.: A New Sensor for On-Line Diagnostics of Particles under Thermal Spraying Conditions, Proceedings of Advances in Powder Metallurgy and Particulate Materials – 1994. Advanced Processing Techniques 6 (1994), p. 199–210. 276 Vattulainen, J., Lehtinen, T., Hernberg, R., Knuuttila, J., Mäntylä, T.: In-Flight Particle Concentration and Velocity Measurements in Thermal Spraying Using a Non-Intensified CCD Camera, Proceedings of 15th International Thermal Spray Conference – Thermal Spray: Meeting the Challenges of the 21st Century 1 (1998), p. 767–772. 277 Vattulainen, J., Hämäläinen, E., Hernberg, R., Vuoristo, P., Mäntylä, T.: Novel Method for In-Flight Particle Temperature and Velocity Measurements in Plasma Spraying Using a Single CCD Camera, Journal of Thermal Spray Technology 10 (2001), p. 94–104. 278 Ignatiev, M., Smurov, I., Bertrand, P.: Application of digital CCD camera for monitoring of particle-in-flight parameters in plasma and HVOF spraying, Proceedings of ITSC 2002 International Thermal Spray Conference 1 (2002), p. 72–77. 279 Ignatiev, M., Senchenko, V., Dozhdikov, V., Smurov, I., Bertrand, P.: Digital diagnostic system based on advanced CCD image sensor for thermal spraying monitoring, Proceedings of ITSC 2002 International Thermal Spray Conference 1 (2002), p. 1001–1006. 280 Fang, J., Xu, W.: CCD technology applied to noncontact diagnosis of plasma flame, Proceedings of ICECA ’2001 Int. Conf. Energy Conversion and Application (2001), p. 1294–1297. 281 Vuoristo, P., Ahmaniemi, S., Nuutinen, S., Mänthylä, T., Hämäläinen, E., Arola, N., Vattulainen, J.: Optimisation and Monitoring of Spray Parameters by a

CCD Camera Based Imaging Thermal Spray Monitor, Proceedings of Thermal Spray 2001 – New Surfaces for a New Millennium (2001), p. 727–736. 282 Nadeau, F., Pouliot, L., Blain, J.: Practical Applications of Spray Plume Sensors in the Aeronautic and Automotive Industries, Proceedings of 1st International Thermal Spray Conference – Thermal Spray: Surface Engineering via Applied Research 1 (2000), p. 1131–1134. 283 Bourque, G., Lamontagne, M., Moreau, C.: A New Sensor for OnLine Monitoring the Temperature and Velocity of Thermal Spray Particles, Proceedings of 1st International Thermal Spray Conference – Thermal Spray: Surface Engineering via Applied Research 1 (2000), p. 45–50. 284 Knuuttila, J., Saarenrinne, P., Hernberg, R., Lehtinen, T., Mäntylä, T.: In-Situ Measurement of Particle Concentration and Velocity Using a Non-Intensified CCD Camera, Proceedings of 1st United Thermal Spray Conference – Thermal Spray: A United Forum for Scientific and Technological Advances (1997), p. 577–582. 285 Hämäläinen, E., Vattulainen, J., Alahautala, T., Hernberg, R., Vuoristo, P., Mäntylä, T.: Imaging Diagnostics in Thermal Spraying-Spray Watch System, Proceedings of 1st International Thermal Spray Conference – Thermal Spray: Surface Engineering via Applied Research 1 (2000), p. 79–83. 286 Hanson, T. C., Hackett, C. M., Settles, G. S.: Independent Control of HVOF Particle Velocity and Temperature, Journal of Thermal Spray Technology 11 (2002), p. 75–85. 287 Craig, J. E., Biancaniello, F., Ridder, S., Lee, D. Y., Parker, R. A.: A Two-Wavelength Imaging Pyrometer for Measuring Particle Temperature, Velocity, and Size in Thermal Spray Processes, Proceedings of 38th Annual Conference of Metallurgists of CIM – Advanced Sensors for Metal Processing (1999), p. 369–380. 288 Bengtsson, P., Johannesson, T., Wigren, J.: Crack Structures in Plasma

References Sprayed Thermal Barrier Coatings as a Function of Deposition Temperature, Proceedings of 14th International Thermal Spray Conference: Thermal Spraying – Current Status and Future Trends 1 (1995), p. 513–518. 289 Nakahira, H., Harada, Y., Mifune, N., Yogoro, T., Yamane, H.: Advanced Thermal Barrier Coatings Involving Efficient Vertical Micro-Cracks, Proceedings of 13th International Thermal Spray Conference – Thermal Spray: International Advances in Coatings Technology (1992), p. 519–524. 290 Das, S., Suri, V. K., Chandra, U., Sampath, K.: One-Dimensional Mathematical Model for Selecting Plasma Spray Process Parameters, Journal of Thermal Spray Technology 4 (1995), p. 153–162. 291 Harding, J. H., Mulheran, P. A., Cirolini, S., Marchese, M., Jacucci, G.: Modeling the Deposition Process of Thermal Barrier Coatings, Journal of Thermal Spray Technology 4 (1995), p. 34–40. 292 Fukanuma, H.: A Porosity Formation and Flattening Model of an Impinging Molten Particle in Thermal Spray Coatings, Journal of Thermal Spray Technology 3 (1994), p. 33–44. 293 Mulheran, P. A., Harding, J. J., Kingswell, R., Scott, K. T.: Modeling the Structural Development in a Plasma Sprayed Coating, Proceedings of 13th International Thermal Spray Conference – Thermal Spray: International Advances in Coatings Technology (1992), p. 749–753. 294 Jacucci, G., Dirolini, S., Marchese, M., Harding, J. H.: Computer Modeling of the Manufacturing Process of Ceramic Thermal Barrier Coatings, Proceedings of Computational Plasticity: Fundamentals and Applications II (1992), p. 1683–1695. 295 Berndt, C. C., Brindley, W., Goland, A. N., Herman, H., Houck, D. L., Jones, J., Miller, R. A., Neiser, R., Riggs, W.,

Sampath, S., Smith, M., Spanne, P.: Current Problems in Plasma Spray Processing, Journal of Thermal Spray Technology 1 (1992), p. 341–356. 296 EN ISO 2360: Non-conductive coatings on non-magnetic electrically conductive basis materials – Measurement of coating thickness – Amplitude-sensitive eddy current method (equivalent ISO 2360). 297 Guessasma, S., Montavon, G., Gougeon, P., Coddet, C. (2003): Designing expert system using neural computation in view of the control of plasma spray processes. Materials and Design 24 (7), 497–502. 298 Lugscheider, E., Seemann, K. (2004): Prognose von Schichteigenschaften mittels neuronaler Netze. In Tagungsband zum 7. Werkstofftechnischen Kolloquium (Wielage, B., ed.), p. 50–59, Eigenverlag Chemnitz. 299 EN 14665: Thermal spraying – Thermal sprayed coatings – Symbolic representations on drawings. 300 Guessasma, S., Montavon, G., Coddet, C.: Neural Networks, Design of Experiments and Other Optimization Methodologies to Quantify Parameter Dependence of Atmospheric Plasma Spraying, Proceedings of ITSC 2003 International Thermal Spray Conference – Advancing the Science and Applying the Technology 2 (2003), p. 939–948. 301 Kingswell, R., Scott, K. T., Wassell, L. L.: Optimizing the Vacuum Plasma Spray Deposition of Metal, Ceramic, and Cermet Coatings Using Designed Experiments, Journal of Thermal Spray Technology 2 (1992), p. 179–186. 302 Huston, R. P.: Role of Designed Experiments to Evaluate Factors that Influence Measured Properties of Sprayed Coatings, Proceedings of 3rd National Thermal Spray Conference – Thermal Spray: Research and Applications (1990), p. 675–680. 303 Taylor, T.: Thermal barrier coating for substrates and process for producing it US Patent 5,073,433, Dec. 17, 1991.

259

Subject Index

Subject Index a acceleration and heating of particles in plasma 179 – model validation and calibration, particle speed 180 – model validation and calibration, particle temperature 180 – surface thermal conditions 180 advanced coatings 53 advanced online monitoring 243 aero engines 38 agglomeration 30 aid of plasma-plume imaging 225 air plasma spraying – spray parameters (APS) 91 – ZrO2 coatings 91 Al2O3 29 ALSTOM/CENIT 203 anode cooling system 77 anode cooling water 77 APS jet 151, 177 APS powder feeding 72 APS relative coating-thickness distribution 191 – different offsets of spraying passages 191 APS relative coating thickness to the torch misalignment 192 arc and jet pulsations 108 arc pulsations 128 argon plasma 57 atomization 29 – gas-atomized powders 29 – gas atomization 29 – water atomization 29 auditing 237 automation 244 6-axis robot with plasma spray gun 82

b basics of thermal spraying 21, 22 batch process VPS 104 batch size, one-piece flow 41 benefits of process simulation tools big structural components 232 blend of two metals/alloys 30 blending 30 bond coat 89 bond strength 20 bonding model 182 burner, combustor segment 85

200

c 3D calculations 181 calibration models 174 calibration of the bonding model 182 carrier gas flow 184 – rate 71 cascade plasma torch 61 CASPSP® 203 categories 210 – machine and spraying equipments 211 – materials, their characterization, and application related standards 210 – pre- and posttreatments 210 – specifications of parts, components and tests 211 – thermal spraying (general) 210 cathode – cooling system 77 – cooling water 77 cause–effect analysis 107 CCD-camera-based devices 216 ceramic coatings 20 chemical/physical analysis 35 chemical composition 24 chemical techniques 33 chemical vapor deposition (CVD) 18

261

262

Subject Index cladding 18, 30 – Ni-Al 30 – Ni-Mo-Al 30 – NiCr-Al 30 classes of thermal-spraying complexity 94 classification of thermal-spray processes 21 – atmospheric plasma spraying 21 – detonation spraying 21 – electric-arc spraying 21 – flame spraying 21 – HVOF spraying 21 – vacuum plasma spraying 21 closed-loop-controlled coating process 226 CMMS (computerized maintenance management system) 131 CNC – programs 188 – robot programming tools 173, 174 coated power-generation components 84 – APS 84 – Cr3C2-CoCr coatings 86 – hot corrosion 84 – HVOF 84, 86 – LPPS 84 – MCrAlY 84 – oxidation resistant coatings 84 – thermal-barrier coatings 84 – wear-resistant coatings 84 – ZrO2 8Y2O3 thermal-barrier coating 84 coating 40 – adhesion 135 – application 8 – definition 230 – – and design for coating 199 – development 137, 237 – – engineer 212 – – particle temperature 137 – – particle velocity 137 – – process 229 – – surface temperature 137 – equipment 103 – for design 230 – formation from separate particles 133 – for power-generation components 6 – – design requirements 6 – in gas turbines 2 – jet 224 – life 12, 40 – manufacturers 14 – manufacturing engineer 15 – manufacturing sequence 67 – microstructure 186, 221 – mono-layer 235 – of boiler tubes 86

– – – – – – –

– wire flame spraying 86 porosity 231 porosity and roughness 185 powder-manufacturing process 14 preparation 244 process 17, 20 – for power-generation components and parts 17 – process applications 18 – processes and typical properties 19 – production 211, 213 – program 236 – qualification 235 – quality 9, 162 – – powder-feeding conditions 73 – removal 96 – removal, stripping 97 – – chemical and electrochemical methods 97 – – mechanical methods 97 – – thermal methods 97 – requirements 66 – roughness 231, 234 – specification 24, 230, 231 – technology 231 – thickness 224 – thickness optimization 189 coating-cell – cooling system 77 – elbows and other hydraulic fittings 78 – water chiller/heat exchanger 78 – water pump(s) 78 – water settling and conditioning tank 78 – water-supply pipes 78 coating-development matrix 12, 13 coating distribution 182 – single injector nonoptimized rotation 192 coating-process development 199, 211, 212 coating-technology selection 18 cold gas-dynamic 49 combustion pressure 53 combustion products 53 combustor 85 – front segment 85 commercial 99 – value in Euro 99 company’s know-how 233, 236 comparison of HVOF fuels 47 – ethyl alcohol or ethanol 48 – hydrocarbon fuels 48 – hydrogen 48 complex parts 86 components of stationary gas turbines – coating process 4

Subject Index – coating requirement 4 comprehensive set of monitoring tools 226 computer-based controller 225 computer-simulated images of 40-μm diameter alumina droplets 137 considerations for basic structural design 229 continuous process LPPS 104 control 223 – of coating thickness 112 – technologies 243 controller 9 controlling of the initial parameters 224 conventional high-voltage plasma torches 58 conventional low-voltage plasma torches 58 cooling features 9 core and key components – manipulation systems 41 – spray gun 41 corrective maintenance 130 corrosion protection 24 cost comparison LPPS vs. HVOF 102 – capital costs 102 – spray time 102 – total costs 102 cost of the coating process 100, 244 cost of thermal spraying 103 Cpk 115 Cr2C3–NiCr 32 Cr2C3 in a metallic NiCr matrix 38 Cr2O3 29 customers 243 CVD and PVD processes 20 d definition of a stable coating process 117 – economy requirements 118 – quality requirements 118 – technical requirements 117 definition of process capability 116 definition of the thickness 231 density 36 – of argon-hydrogen plasmas at different pressures 142 deposition efficiency 21, 22, 128 – at the periphery 162 – process deposition efficiency 22 – target deposition efficiency 22 design – for coating 87, 230, 232 – for manufacturing 242

– of experiments 212, 233 designer 15, 232 destructive tests 113 Diamond Jet® (DJ) 50 Diamond Jet® Hybrid 50 different plasma gases 168 different types of thermal-spray processes 223 distribution function 116 DPV-2000 216, 217 drift of the plasma and carrier gas flow rates 112 drifting 111, 116 – process capability 117 drifting of voltage 55 – erosion 55 – wear of electrodes 55 dual-wavelength pyrometers 216 dynamic and kinematic viscosity of argon-hydrogen plasma at different pressures 143 e economies of MCrAlY coatings 102 – LPPS and HVOF 101 effect of nozzle diameter 151 efficiency of heat transfer into the substrate at APS process for F4 and PlazJet 167 electrode erosion 111 electrode wear 127 – deposition efficiency 127 electro discharge machining (EDM) 8 elementary quality requirements 207 energy balance of the plasma gun 147 EN standards 210 enthalpy of argon-hydrogen mixture 140 equipment knowledge 131 erosion of electrodes 57 ETS 209 ETSP 209 ETSS 209 EWF guidelines 208 experience in production 11 expert consulting service 243 f factors influencing the thermal-spraying process 11 failure modes 230 features of HVOF physics 169 feedback 211 – of observations 213 feedstock 23 – wire 23

263

264

Subject Index financial controller 15 finishing 95, 104 – coating specification 95 – final surface quality 95 finishing/post treatment 8 first article inspection 213 fixtures and masking 83 flame and arc spray torches 43 flow rate 158, 222 – particles velocities 222 flowability 35 formation of the coating layer 167 front nozzle wall 77 – coating quality 77 – powder buildup 77 – powder blow off 77 frozen process 38, 213 fuels 48 fully automated spray process 226 g gas dynamics of plasma torch 145 gas flow rate 158 gas supply 9 – distribution system 80 gas velocity 71 general requirements for coating-shop personnel 213 graphical user interface 174 gun and component motion and positioning 110 gun electric input power 222 gun particle trajectories 74 h handling systems 9, 82 – acceptance/inspection 82 hard face coating 93 HE100™ 63, 64 heat conductivity and heat flux potential of argon-hydrogen plasma at different pressures 141 heat flux potential 158 heat losses in a combustion module 52 high costs of the machine time 237 high particle velocity 221 high-pressure economical HVOF torches 53 high-speed pyrometers 215 high velocity oxy fuel-spray parameters (HVOF) – Cr3C2-NiCr wear-resistant coatings 93 HOSP™ process 32 hot area of the spray spot 163

hot corrosion or oxidation resistant coating 93 HVOF 46, 47, 49 – flow structure 197 – modeling package 196 – or plasma 171 – process 45, 48 hydrogen content 158 i IGRIP 173 – software 188 impact speed 175 impacting particle 134 in process control 12 in-flight particle temperature 56 Inconels 29 increase in stochastic thickness scatter due to electrode wear 127 influence of particles on plasma flow 164 – powder feed rate on the coating weight gain 164 – substrate surface temperature 165 influence of particle velocity and temperature on microstructure 219 influence of rotation 192 influence of shock waves 160 influence of spray parameters on particle speed and temperature 218 influencing factors 224 injector 71 injector plugging and “spitting” 75 – carrier gas speed 76 – coating quality 75 – cooling 76 – fine particles 76 – flow speed and direction at the injector exit 75 – gun motion 76 – injector and plasma conditions 76 – injector inspections 76 – robot accelerations 76 – roughness of the injector inner surface 76 – temperature of the injector walls 76 input for every production 66 instability in coating quality 107 instability of the quality-control process 128 internal plasma and transferred arc 89 – plasma-torch components 90 investment casting 8 – finishing 8 – heat treatment 8 – vacuum casting 8

Subject Index ionization and electrical conductivity of argon-hydrogen plasma at different pressures 144 ISO standards for thermal spraying 208 j Jet Kote® 49 JP-5000® 53 k kinematic viscosity 158 knowledge base 15 l large casting 86 Laser Doppler Anemometers 215, 216 LaserStrobe 215 layer offset 22 layout of APS and LPPS nozzles 138 layout of the LPPS process 41 Le Chatelier’s principle 140 lifetime of the coatings 244 limitations of offline programming 244 line-of-sight process 228 liquid-fuel HVOF – systems 50 – torches 51 long-term stability of HE100™ plasma parameters 60 low-pressure plasma-coating system 41 – 3D model of an integrated LPPS system 42 – LPPS manipulation systems 42 low particle velocity 221 low pressure plasma-spray parameters (LPPS) 92 – CoNiCrAl coatings 92 LPPS 46, 47 – jet 177 – MCrAlY coating 46 – process 191 – spray spot 163 – supersonic jet 154 – system 41 m machining 8 maintenance 87, 130 – approach for cooled turbine blades and vanes 96 – management 130 major existing HVOF systems 48 – Jet Kote® (JK) 48 manipulation systems 81

– CNC-steered manipulators 81 – electromechanical movement units 81 – manipulator 81 – robots 81 manufacturing and coating process 7 manufacturing chain 7 manufacturing coating development 228, 230 – part design 230 – part development 230 – process development 230 manufacturing methods and control 238 manufacturing process development 215 manufacturing process plan 213 manufacturing sequence 66 manufacturing tolerances 40 material data sheets – for thermal spraying powder 39 – of powder suppliers 38 material flow 213 material properties 2 materials and coatings, requirements 1 – functional materials 1 – structural materials 1 material temperature 135 MCrAlY 29, 46 – coating thickness distribution on the pressure side of a turbine blade 188 – coatings 38 mechanical and physical properties 19 mechanical powder 70 medium-size structures 86 mesh size and sieve opening 34 metal weld cladding 20 metallic coatings 20 metallic Co or Ni alloys 38 methods of coating application 190 microstructure 231 MMC-brochure 238 model shape for validation and adjustment of coating software 187 modeled and measured porosity distributions 186 modeling 15, 244 – in the coating process 199 – of the plasma gun 176 – of the plasma jets 176 – of turbine blades 187 models for particle and jet interactions with the substrate 196 monitoring 215 – equipment 223 – in process development and production 244

265

266

Subject Index – of the equipment 40 moving spray spot, temporal temperature variation 166 multimonitoring system with intelligent feedback control 227 multi-plasma system 60 mutual position of the gun and component fixtures 125 n national standards 211 – ASTM (American Society for Testing and Materials) 211 Navier–Stokes equations solver 196 net energy 56 NiAl 29 NiCr 29 NiCrAlY splat 175 nozzle diameter 220 number of influencing parameters 9 number of injection ports 192 o offline programming 94, 244 – tools 237 offline simulation 202 – in coating development 199 – of a thermal-spray process 171 online method 223 online particle diagnostic systems 215 online process monitorin 233 operating JP-5000® 51 operating parameters of JP-5000® 51 operating window – for HVOF processes 52 – related to conventional and cascade plasma systems 64 operational conditions 2 operational sheets 237 operational window 16, 37, 40, 87, 91, 118, 119, 121, 244 – carrier gas flow rate 118 – for the gun voltage 120 – gun current or voltage 118 – main gases flow rates 118 – new equipment 243 – parameter operational window 122 – parameter sets 91 – powder properties 118 – standoff distance 118 operations and maintenance 130 operations and quality-assurance steps 99 operation sheets 66 operators of the equipment 15

optimization and sensitivity analysis 191 optimization of atmospherically plasmasprayed (APS) coatings 217 – particle-monitoring devices 217 optimized LPPS coating 192 overspray structure 232 p parameter sets for coating 238 part complexity 125 part development 200, 236 – full-scale offline simulation 200 particle acceleration and heating – in the LPPS free jet 158 – inside the nozzle 160 particle at LPPS 158 particle bonding 110, 136, 175 particle injection 74 particle-momentum and heat-transfer solver 196 particle-monitoring devices 217 particle parameters 158 particle shape and morphology 35 particle-size distribution 34 – laser-scattering method 34 particle size and particle-size distribution 33 – light scattering 35 – other methods 35 – sieve analysis 35 particle speed 219, 220 particle velocity 158 particles at APS 156 particles in plasma 155 people 238 perfect plasma system 54 personnel 244 physical background of simulation package 175 physical model 174 physical simulation 233 physical vapor deposition (PVD) 18 physics of plasma torches 138 plasma-beam data 225 plasma densification 31 plasma-densified (spherodized) powders 32 plasma enthalpy and pressure 139 plasma-gun parameters 56 plasma gun power 56 plasma jet 74 plasma parameter fluctuations 55 plasma process 54 plasma properties 139

Subject Index plasma spraying 54 – thermal barrier, anticorrosion, wear-resistant, abradables and some other coatings 54 plasma swirl 151 plasma torches 58 plugging 76 porosity 20, 219, 220 – distribution 163 – measurements 113 porous-coated (cladded) powders 31 possible improvements – of HVOF systems 51 – of plasma systems 63 powder 21, 23, 24, 38, 224 – carrier gas OW 120 – characterization 33 – cost 233 – feeder 69 – feed rate 235 – for power-generation applications 36 – for wear-resistant coatings 38 – hoses 72 – injection 110 – injection and plasma/hot gas jet 73 – injection velocity 198 – injectors 73 – – carrier gas flow 73 – – particle size and mass 73 – insert 70 – inspection 27 – preparation and powder-delivery system 68 – – powder delivery and injection system 68 – – powder preparation 68 – production steps 27 – quality 125 – selection 234 – shape 110 – types 24 – types and functions 26 – – corrosion protection 26 – – electrical functions 26 – – special functions 26 – – thermal functions 26 – – wear protection 26 – used for – – protection against hot corrosion and oxidation 37 – – thermal-barrier coatings 38 – – wear protection 36 powder-feeding 24 powder-injection conditions 73

– feeder position 73 – powder hoses 73 powder-injection system 129 powder-production processes 28 – crushing and milling 28 – material 28 – particle morphology 28 – spray powder 28 powder-production processes and morphologies 27 – crushing and milling 27 – mechanical techniques 27 powder-size distribution 109 powder-supply system 69 power input 158 power source 9 power-generation components 21 power-generation equipment 19 power-generation parts 12 power-supply system 79 – primary chopped power sources 79 – secondary chopped power sources 79 – thyristor sources 79 powerplant components 19 powerplant operation 24 practical experience 17 preheating chamber 90 preoxidized 90 preserial production 238 preserial release 237 preventive maintenance 131 principle of electric-arc spraying 44 principle of flame spraying – using powders 43 – using wire 44 principle of high-velocity oxygen-fuel (HVOF) spraying – using gas for fuel 45 – using liquid fuel 45 probability 175 – of coating out-of-spec 124 – of the part failure 123 process 116, 237 process “recipe” 121 process and systems 91 process capability 12, 14, 16, 122, 124, 129, 239, 242 – achievable 125 – actual 114 – coating material 124 – coating specification 124 – failure of the coating 122 – size of the parts to be sprayed 124 process capability and stable process 115

267

268

Subject Index – definition 115 process control 225 – of thermal-spraying system using total monitoring tools 227 – using a simple beam observation system and feedback to a computer-based controller 226 process controller 223 process development 14, 95, 233, 244 – teamwork 95 process operational windows 243 process parameters 119, 182 process qualification 237 process release 211 process scatter 115 process scatters of the single- and doubleinjector rotation 194 process stability 14, 69, 241 process steps 23 process throughput time 233 process variations 237 production 211 production costs 100 production experience 86 – drifting 87 – handling 86 – lead time 86 – logistics 86 – quality assurance 86 – released manufacturing process 86 – trained people 86 production operators and quality engineers 15 productivity of equipment 40 programming module 173 programming of the coating process 94 progressive maintenance 130 q qualification and education of spraying personnel 209 qualification and experience of the development personnel 233 quality 16 – and process capability 105 – assurance 105 – – procedure 213, 237 – – systems 206 – at the peripheral area 163 – check procedures 231 – of the MCrAlY coating 163 – parameters 119 – requirements for thermally sprayed structures and coating shops 206

– requirements in thermally spraying business 210 – requirements of thermally sprayed structures 207 – – comprehensive quality requirements 207 – – guidance for selection and use 207 – – qualified sprayers 207 – – qualified supervising personnel 207 quality-control process 114 – part of the manufacturing 114 r random deviation of the gun position 193 RCFA (root cause failure analysis) 131 real and simulated coating structure 187 redevelopment 243 refurbished components 99 – recoating 98 – requalification 99 reliability of the gas turbine 123 reliable coating production 222 reliable stable coating process 211 repair of turbine parts 95 repeatability 40 requalification 243 requirements 7, 54 – for monitoring in production 224 – to be “coatable” 228 restoration 96 – of the base materials 98 rotation spray 193 rotational handling system 82 roughness 89 – and texture of the coating and substrate surfaces 112 – measurements 88 – – laser instrument methods 88 – – stylus 88 s scatter of coating properties 14 scattered quality 114 schematic view of plasma spraying in a chamber 43 SCPG 65 – advantages 65 selection of powder 234 sensing 215 sensitivity 193 – of APS TBC weight on the voltage drift 121 – of a spray pattern 182 – of spray spot to process parameters 183

Subject Index – of the quality parameters 119 sensitivity checks 236 serial production release 213 serial release 239 shapes of spray patterns 190 – blade in atmospheric and vacuum plasma spraying at the spraying standoff distance 190 shift of the coating average thickness due to drifting 127 shopfloor experience 215 short-term process capability 124 6-sigma process 12 simple process with fixed input parameters 225 simple spraying system with a jet-imaging tool 225 simulation 14 – and modeling technology 14 – in production 171 – of HVOF process – – heat losses in the HVOF gun 195 – – internal powder injection 195 – – spray spot 195 – – thermodynamic behavior of the combustion products 195 simulation of HVOF process 195 single- and double-injector rotational spray 194 single cathode/single anode cascade plasma generators (SCPG) 65 sintering 31 sol–gel-freeze drying 33 sound speed of argon-hydrogen plasma at different pressures 142 sources of plasma coating-process variations 106 – common causes 107 – electrodes 106 – gas-supply system 106 – gun movement 106 – masking and fixtures 106 – parameter drifting 107 – powder-supply system 106 – power supply 106 – special causes 106 – unstable quality-control procedures 107 sources of process variations 105 special case: spraying for power-generation components 211 special causes of coating-process variation 107 – prevented by regular maintenance 107 special processes 206

specification 38, 40, 66, 238 – powders 38 specifications for manufacturing and purchasing 7 specific programming features 237 specific surface 36 specified coating quality 14 spitting 76 splat at the substrate 135 splat formation 135 splat types 136 splats 134 spot periphery 162 spray angle 22, 182, 186, 187, 237 spray drying 30 spray footprint 161, 235 – deposition efficiency 235 – quality issues 235 – spray time 235 spray pattern 21, 22, 163, 182 – and standoff distance 234 spray positions and fractions of the coating material supplied 193 spray programming 236 spray trials 235 spray-dried agglomerated powders 31 spray-dried and sintered powders 32 spray-gun accelerations 237 spraying equipment 9 spraying materials 52 – kinetic energy 52 – thermal energy 52 ST-4000™ 53 stability of the quality control 112 stand-off 22 standard quality requirements 207 standards and training 205 standards, codes 205 standoff distance 235 stationary gas turbine 3 statistical control 116 steam turbines thermal-spray coatings 2 Stellites 29 stochastic modeling tools 189 stochastic nature – of a spray process 108 – of the process 241 stripping hard metal coatings 98 structure – of an APS jet 152 – of ETS-education 209 – of LPPS jet 153 – of plasma jets 151 – of preventive maintenance 131

269

270

Subject Index – of the IGRIP programming module 173 – of the offline coating simulation package 172 substrate microstructure underneath the removed splat 136 substrate temperature 22 supplier 15 surface – temperature distributions 181 – treatments 18 surface preparation 8, 87, 88, 103 – and the part temperature 128 – automation 87 – sandblasting 87 surface-protecting processes 206 system analysis for plasma spraying 10 t target deposition efficiency 237 targets of the book 241 tasks for “target” readers 15 TBC 85 technical specifications 212, 218 temperature 219, 220 – of combustion products 52 test pieces 235 test results 238 theory and physical trends 133 – particles 133 – plasma/combustion torch 133 – plasma/hot gas jet 133 – substrate 133 thermal-barrier coating 93 thermal paint measurements 181 thermal protection 24 thermal-spray coatings 87 – manufacturing chain 87 thermal-spray material 25 thermal-spray processes 6 thermal spray torches 242 thermal spraying 5, 18, 20 – definition of thermal spraying 5 – US patent classification 5 thermal-spraying equipment 40 – core components 40 – key components 40 – peripheral equipment 40 – process environment 40 thermal-spraying process 101 – nonvalue-added costs 101 – sequences of a cast part 101 – value-added costs 101 thermal-spraying systems 5 – power-generation components 5

thermodynamic and transport properties of argon/hydrogen mixtures 176 thermodynamic properties of combustion gases 196 thickness 20 – of LPPS spray spot 184 TiO2 29 torch parameters 234 torch power – and standoff on the particle parameters at APS 157 – on particle parameters at LPPS 159 torch pulsations and drifting 126 total monitoring concept 228 total quality control 226 total quality-management system 209 – for thermal spraying 208 traceability of quality documentation 205 trained and motivated people 242 transferred arc (TA) 90 – power supply 90 traverse gun speed 235 traverse speed 22 Triplex® 62, 63 TriplexPro™-200 62, 64 turbine blades 85, 86 turbine component life flowchart 97 types of components 3 typical “cold” HVOF jet with characteristic shock diamonds 169 typical energy balance of APS and LPPS plasma torches 148 typical parameters of APS and LPPS for ZrO2 and NiCrAlY, respectively 146 typical radial distributions of particle velocity 198 typical shapes of particle splats 134 u uniformity of coating thickness 237 use monitoring for process control 222 user interface module 196 utilization of the equipment 103 v value-added steps 99 variation of chamber pressure 159 variation of gas flow rate and hydrogen content 159 variation of the gun power; the gas flow rates and composition unchanged 149 variation of the plasma composition at the same specific plasma enthalpy 149

Subject Index variation of the plasma flow rate at unchanged gun power and gas composition 150 variation of the substrate temperature 112 viscoplasticity model of a splat and particle bonding 175 vision for future coating control and monitoring 224 voltage fluctuation 57 w water-cooled, HVOF 79 water-cooling systems, maintenance WC–Co 32 wear of electrodes 111 wear protection 24 WIP 213 wire 23 wire spraying 20

78

WokaJet™-400 53 WokaStar™-600 51, 53 work flow and important coating hardware 65 – maintenance strategy 66 – production strategy 65 – quality-assurance strategy 66 work in progress (WIP) 40 y Y2O3 29 yield in serial production 212 yield of a process 99 YSZ coatings 220 Yttria Stabilized Zirconia (YSZ) z ZrO2 29 ZrO2–Y2O3 32 ZrO2 8% Y2O3 38

4

271