Casting: An Analytical Approach (Engineering Materials and Processes) (Engineering Materials and Processes) [1 ed.] 1846288495, 9781846288494

Citation preview

Engineering Materials and Processes

Series Editor Professor Brian Derby, Professor of Materials Science Manchester Materials Science Centre, Grosvenor Street, Manchester, M1 7HS, UK

Other titles published in this series: Fusion Bonding of Polymer Composites C. Ageorges and L. Ye Composite Materials D.D.L. Chung Titanium G. Lütjering and J.C. Williams Corrosion of Metals H. Kaesche Corrosion and Protection E. Bardal Intelligent Macromolecules for Smart Devices L. Dai Microstructure of Steels and Cast Irons M. Durand-Charre Phase Diagrams and Heterogeneous Equilibria B. Predel, M. Hoch and M. Pool Computational Mechanics of Composite Materials ´ski M. Kamin Gallium Nitride Processing for Electronics, Sensors and Spintronics S.J. Pearton, C.R. Abernathy and F. Ren Materials for Information Technology E. Zschech, C. Whelan and T. Mikolajick Fuel Cell Technology N. Sammes Computational Quantum Mechanics for Materials Engineers L. Vitos Publication due August 2007

Alexandre Reikher and Michael R. Barkhudarov

Casting: An Analytical Approach

123

Alexandre Reikher, PhD Albany Chicago Co. 8200 100th St. Pleasant Prairie, WI 53158 USA

Michael R. Barkhudarov, PhD Flow Science, Inc. 683A Harkle Road Santa Fe, NM 87505 USA

British Library Cataloguing in Publication Data Reikher, Alexandre Casting : an analytical approach. - (Engineering materials and processes) 1. Founding I. Title II. Barkhudarov, Michael R. 671.2 ISBN-13: 9781846288494 Library of Congress Control Number: 2007928128 Engineering Materials and Processes ISSN 1619-0181 ISBN 978-1-84628-849-4 e-ISBN 978-1-84628-850-0

Printed on acid-free paper

© Springer-Verlag London Limited 2007 FAVOR™ and FLOW-3D® are trademarks and registered trademarks of Flow Science Inc., 683 Harkle Rd. Ste A, Santa Fe, NM 87505, USA. http://www.flow3d.com MATLAB® is a registered trademark of The MathWorks, Inc., 3 Apple Hill Drive, Natick, MA 01760-2098, USA. http://www.mathworks.com Visual Basic® is a registered trademark of Microsoft Corporation, One Microsoft Way, Redmond, WA 98052-6399, USA. http://www.microsoft.com The software disk accompanying this book and all material contained on it is supplied without any warranty of any kind. The publisher accepts no liability for personal injury incurred through use or misuse of the disk. Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced, stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers, or in the case of reprographic reproduction in accordance with the terms of licences issued by the Copyright Licensing Agency. Enquiries concerning reproduction outside those terms should be sent to the publishers. The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore free for general use. The publisher makes no representation, express or implied, with regard to the accuracy of the information contained in this book and cannot accept any legal responsibility or liability for any errors or omissions that may be made. 987654321 Springer Science+Business Media springer.com

Michael Barkhudarov: I dedicate this work to my family, Natasha, Sophia and Philip, and to my dear parents

Alexandre Reikher: I dedicate this work To my dear parents To my wife Marianna And to my children Daniel and Alison

Preface

This book is the result of 40 years of the combined authors’ experience in mechanical and fluid dynamics engineering. It gives an overview of product and process development from the analytical standpoint. This book has not been intended to revolutionize the casting industry. The principals of fluid dynamics and static mechanics were largely developed in the nineteenth century, but process development still largely remains a trial and error method. This book is intended to underline the principals of strength of materials and fluid dynamics that are the foundation of the casting product and process development. This book has been written as a resource and design tool for product and process engineers and designers who work with aluminium castings. It combines many aspects of product and process development, which include the basic principals of static mechanics and fluid dynamics as well as completely developed applications allowing solving problems at every stage of the development process. This book has five main parts: (1) overview of casting processes, (2) fluid dynamics, (3) strength of materials, (4) sand casting, permanent mould, and die casting process development, and (5) quality control. The unique feature of this book is a combination of real life problems, which product and process engineers face every day, with user-friendly applications written in MATLAB® and Visual Basic. The comprehensive unit conversion calculator as well as examples in the area of strength of materials is intended to serve as a reference for practicing engineers as well as for students who are beginning to study mechanical engineering. It cab also be used by process engineers, who contribute greatly to product design. Completely developed process design applications will help process engineers to take full advantage of the power that computers and software bring to engineering. Product engineers can also get inside of process development procedures as well as the governing equations that are used in casting process development. Examples in this book are solved with the help of MATLAB® functions, Visual Basic applications as well as a general purpose CFD code FLOW-3D®. All MATLAB® functions, art work and Visual Basic applications are developed by the

viii

Preface

authors. Proper references are given for commonly available equations and theories. Considerable effort has been made to avoid errors. The authors would appreciate reader’s comments and suggestions for any corrections and improvements to this book. Alexandre Reikher, Milwaukee, Wisconsin, USA, April 2007 Michael Barkhudarov, Los Alamos, New Mexico, USA, April 2007

Contents

1 Casting of Light Metals .................................................................................... 1 1.1 Casting Processes....................................................................................... 1 1.2 Sand Casting ............................................................................................. 1 1.2.1 Gating System ................................................................................ 2 1.2.2 Risers and Chills............................................................................. 2 1.3 Permanent Mould ...................................................................................... 3 1.3.1 Gravity Casting............................................................................... 4 1.3.2 Low-pressure Permanent Mould Casting ....................................... 4 1.3.3 Counterpressure Casting ................................................................. 4 1.4 Die Casting ................................................................................................ 5 1.4.1 Die-cast Process ............................................................................. 5 1.4.2 Die-cast Dies ................................................................................. 8 1.4.3 Runner System ............................................................................... 8 1.4.4 Cavity of Die-cast Die ................................................................. 11 1.4.5 Air Ventilation System ................................................................ 11 1.4.6 Ventilation Blocks ....................................................................... 12

2 Introduction to Fluid Dynamics .................................................................... 13 2.1 Basic Concepts ....................................................................................... 13 2.1.1 Pressure ...................................................................................... 13 2.1.2 Viscosity ...................................................................................... 14 2.1.3 Temperature and Enthalpy ........................................................... 16 2.2 Equations of Motion ................................................................................ 18 2.3 Boundary Conditions ............................................................................... 19 2.3.1 Velocity Boundary Conditions at Walls ....................................... 20 2.3.2 Thermal Boundary Conditions at Walls ....................................... 20 2.3.3 Free Surface Boundary Conditions............................................... 21 2.4 Useful Dimensionless Numbers............................................................... 23 2.4.1 Definitions .................................................................................... 23 2.4.2 The Reynolds number................................................................... 24 2.4.3 The Weber Number ...................................................................... 24

x

Contents

2.5 2.6

2.7

2.4.4 The Bond Number ........................................................................ 25 2.4.5 The Froude Number...................................................................... 25 The Bernoulli Equation............................................................................ 26 Compressible Flow .................................................................................. 26 2.6.1 Equation of State .......................................................................... 27 2.6.2 Equations of Motion ..................................................................... 28 2.6.3 Specific Heats............................................................................... 29 2.6.4 Adiabatic Processes ...................................................................... 31 2.6.5 Speed of Sound............................................................................. 31 2.6.6 Mach Number............................................................................... 33 2.6.7 The Bernoulli Equation for Gases ................................................ 33 Computational Fluid Dynamics ............................................................... 35 2.7.1 Computational Mesh..................................................................... 35 2.7.2 Numerical approximations ........................................................... 36 2.7.3 Representation of Geometry ......................................................... 38 2.7.4 Free Surface Tracking .................................................................. 39 2.7.5 Summary ...................................................................................... 41

3 Part Design ...................................................................................................... 43 3.1 Design of Light Metal Castings ............................................................... 43 3.2 Static Analysis ....................................................................................... 43 3.2.1 Moment of an Area....................................................................... 44 3.2.2 Moment of Inertia of an Area ....................................................... 44 3.3 Loading ................................................................................................. 46 3.4 Stress Components................................................................................... 47 3.4.1 Linear Stress ................................................................................. 48 3.4.2 Plane Stress................................................................................... 50 3.4.3 Mohr’s Circle................................................................................ 51 3.5 Hooke’s Law............................................................................................ 55 3.6 Saint-Venant’s Principle .......................................................................... 56 3.7 Criteria of Failure..................................................................................... 57 3.7.1 Ductile Material Failure Theory .................................................. 58 3.7.2 Brittle Material Failure Theory.................................................... 58 3.8 Beam Analysis ........................................................................................ 58 3.8.1 Free Body Diagram ...................................................................... 67 3.8.2 Reactions ...................................................................................... 67 3.9 Buckling ....................................................................................... 68 3.10 Bending Stresses ...................................................................................... 71 3.10.1 Normal Stresses in Bending ......................................................... 72 3.10.2 Shear Stress .................................................................................. 74 3.11 Stresses in Cylinders................................................................................ 76 3.12 Press-fit Analysis ..................................................................................... 82 3.13 Thermal Stresses ...................................................................................... 86 3.14 Torque...................................................................................................... 87 3.15 Stress Concentration ................................................................................ 93

Contents

xi

3.16 Fatigue ...................................................................................................... 94 3.17 MATLAB® Functions Used in Chapter 3 ................................................. 96

4 Process Design ................................................................................................. 97 4.1 Evaluation of Dimensionless Numbers.................................................... 97 4.1.1 Gravity Pour ................................................................................. 98 4.1.2 High-pressure Die Casting: Slow Shot ......................................... 98 4.1.3 High-pressure Die Casting: Fast Shot........................................... 99 4.2 Flow in Viscous Boundary Layer .......................................................... 100 4.3 The Bernoulli Equation.......................................................................... 102 4.3.1 Stagnation, Dynamic and Total Pressure.................................... 102 4.3.2 Gravity Controlled Flow............................................................. 103 4.3.3 Flow in the Runner System ........................................................ 104 4.3.4 Filling Rate ................................................................................. 106 4.4 Flow in a Shot Cylinder ......................................................................... 108 4.5 Gas Ventilation System ......................................................................... 116 4.6 Ventilation Blocks ................................................................................. 125 4.7 Little’s Formula ..................................................................................... 126 4.8 Poisson Process and the Exponential Distribution................................. 127 4.9 Cooling ................................................................................................. 132 4.9.1 Lumped-temperature Model ....................................................... 132 4.9.2 Heat Flow into a Semi-infinite Medium ..................................... 138 4.10 CFD Simulations.................................................................................... 141 4.10.1 High-pressure Die Casting.......................................................... 142 4.10.2 Gravity Sand Casting.................................................................. 152

5 Quality Control ............................................................................................. 157 5.1 Basic Concepts of Quality Control ........................................................ 157 5.2 Definition of Quality.............................................................................. 158 5.3 Definition of Control ............................................................................. 158 5.4 Statistical Process Control ..................................................................... 159 5.5 Tabular Summarization of Data............................................................. 160 5.6 Numerical Data Summarization............................................................. 160 5.6.1 Normal Probability Distribution ................................................. 161 5.6.2 Poisson Probability Distribution................................................. 162 Appendix ............................................................................................................ 165 A.1 Unit Conversions ................................................................................... 165 A.2 Prefixes ............................................................................................... 171 References ............................................................................................................ 173 Index .................................................................................................................... 175

1 Casting of Light Metals

1.1 Casting Processes There are several casting methods used to produce light metal parts. The most widely used are x x x

Sand casting Permanent mould casting Die casting

Usually economic considerations are the driving force in deciding which casting process can be used. The sand casting process requires the least amount of up-front investment in tooling. But parts cannot be produced with close tolerances and minimum machine stock. It will require extra machining operations, which will drive part price cost up. Permanent mould requires up-front investing in tooling. But parts can be cast with much closer tolerances and reduced machine stock. Due to intensive cooling, parts can be produced with much shorter cycle time, compared with sand casting. The die-cast process requires a large up-front investment in tooling. Due to high pressure used during the die-cast process, parts can be produced with close tolerances and minimum machine stock.

1.2 Sand Casting Sand casting is the oldest way to produce near net shape parts. Sand casting moulds (Figure 1.1) are made using green or chemically bonded sand. Green sand moulds use either a mixture of natural sand and clay or synthetic sands. A typical sand casting mould has a gating system, risers and chills.

2

Casting: An Analytical Approach

1.2.1 Gating System The primary function of the gating system is to allow metal to flow into the cavity. Good design of the runner and gating system should use basic principles of liquid flow. It should allow filling the cavity at the slowest possible fill rate to avoid air entrainment and to ensure complete fill of the thinnest areas of the casting. A typical gating system includes x x x x

A pouring basin A sprue A runner In-gates

To avoid shrink porosity, heavy sections have to be supplied with metal during solidification. That is why it is a good practice to gate into thicker sections of the part. The basic principles of gating system design are as follows: 1. 2. 3. 4.

5.

6.

The ratio of the pouring basin to sprue diameter should allow keeping the gating system full during fill time. The gating system should allow directional solidification of the casting. In gating system design, sharp turns and sudden changes in cross section should be avoided. The gating system should allow maximum flow rate at minimum velocity. This will help to avoid turbulence in metal flow and reduce gas entrainment. If a casting has multiple combinations of thick and thin sections, several sprues and gates have to be used. This also applies to large castings to reduce fill time and avoid filling the cavity with overheated metal. Avoid gating directly into the core or into the wall. It will produce splashing and result in gas porosity.

1.2.2 Risers and Chills During solidification metal density increases and the volume of the casting is decreasing. This causes a decrease in internal pressure. Differences in casting internal pressure force metal to flow from hotter sections that are still in a liquid state to cooler shrinking sections. At some point, metal reaches the state called critical solid fraction. This is state at which metal can’t flow any more to compensate for decreasing volume. At this point, shrink porosity is formed. Risers are placed in the last areas to solidify to continue to fit metal into the thicker, slow solidifying sections of the casting. Some features of the casting are isolated by a thinner section that prevents feeding metal during solidification. An example is a boss attached to a thinner wall. In this case, chills are used to accelerate solidification and minimize or eliminate formation of shrink porosity. In complex castings, multiple risers and chills can be used to force directional solidification.

Casting of Light Metals

3

Figure 1.1. Sand casting mould

1.3 Permanent Mould Permanent mould casting referred to a method of casting in which mould is not destroyed during extraction of the casting. Permanent moulds are capable of producing large number of the same casting. Castings produced in permanent moulds have finer grain structure and superior mechanical properties compared with sand castings. Castings also have almost no gas porosity, major defect of the die-castings. Permanent mould has next major components 1. 2. 3. 4.

Gating system, which direct metal into the cavity at selected rate. Feeding system, which feed metal to thicker areas of the part during solidification period. Chills, which complement feeding system by cooling thicker areas of the part. Venting system, which allows gases to escape during cavity fill process.

In general, permanent mould operationally very similar to a sand casting. It employs gravity as a feeding method. In order to ensure proper fill of the casting sufficient head has to be provided. Position of the gating system, feeders, risers, and chill has to allow direction solidification, starting from the areas of the casting away from the gate and moving into the direction of the gates and feeders. Incorrectly designed and positioned gating system will result in short fill and

4

Casting: An Analytical Approach

shrink porosity. Mistakes in design of the feeding system and chills will result in excessive shrink porosity, or longer dwell time. Incorrectly placed and sized ventilation channels will result in excessive gas porosity in the casting. There are three major processes are currently used to produce castings in permanent moulds x x x

Gravity casting Low-pressure castings Counterpressure casting

1.3.1 Gravity Casting Gravity casting is a basic casting process that uses gravity to fill the cavity of the mould. This process can be used for simply shaped parts that are not going to be used in high stress or leak free applications. 1.3.2 Low-pressure Permanent Mould Casting Low-pressure permanent mould casting is a process that uses pressure to feed metal in to the cavity. Castings produced by this method have higher density and lower gas and shrink porosity. Molten metal is fed from the bottom of the cavity through the riser tube under some pressure (0.5 – 0.8 Bar). Advantages of this method are 1.

2. 3. 4.

5.

The process can be easily automated, which allows control of metal velocity, reduces the turbulence of the metal flow and minimizes air entrainment. A hermetically sealed furnace minimizes metal oxidation and avoids unwanted inclusions. Metal is fed from the bottom of the bath which allows feeding cleaner metal into the cavity of the mould. Directional solidification to the riser allows feeding metal until the casting is completely solidified. This reduces the amount of shrink porosity. This method allows producing quality casting with thinner walls.

1.3.3 Counterpressure Casting Counterpressure casting is a method that uses low pressure to feed metal into the cavity from the bottom of the mould, similar to the low-pressure permanent mould casting method, as well as pressurized cavity. As the cavity is filled with metal, the pressure constantly increases which suppresses hydrogen precipitation. Counterpressure permanent mould casting method allows achieving the highest mechanical properties in a casting. The pressurized cavity eliminates shrink porosity without using risers.

Casting of Light Metals

5

1.4 Die Casting The earliest examples of die casting by pressure injection, as opposed to casting by gravity pressure, occurred in the mid – 1800s. A patent was awarded to Sturges in 1849 for the first manually operated machine for casting printing type. The process was limited to printer’s type for the next 20 years, but development of other shapes began to increase toward the end of the century. By 1892, commercial applications included parts for phonographs and cash registers, and mass production of many types of parts began in the early 1900s. The first die-casting alloys were various compositions of tin and lead, but their use declined with the introduction of zinc and aluminium alloys in 1914. Magnesium and copper alloys quickly followed, and by the 1930s, many of the modern alloys still in use today became available. The diecasting process has evolved from the original low-pressure injection method to techniques including high-pressure casting at forces exceeding 4500 pounds per square inch squeeze casting and semisolid die casting. These modern processes are capable of producing high integrity, near net-shape castings with excellent surface finishes. Alloys of aluminium, copper, magnesium, and zinc are most commonly used for casting. x x x x

Aluminium is a lightweight material exhibits good dimensional stability, mechanical properties, machinability, thermal and electrical conductivity. Copper alloy is a material with high strength and hardness. It has high mechanical properties, dimensional stability, and wear resistance. Magnesium is the lightest cast alloy. It about 4 times lighter than steel and 1.5 times lighter than aluminium. It has a better strength to weight ratio than some steel, iron and aluminium alloys. Zinc is the easiest alloy to cast. It can be used to produce castings with 0.5 mm wall thickness.

1.4.1 Die-cast Process High-pressure die casting is used for a wide range of applications in all major industries. Advantages of die castings are 1. 2. 3. 4. 5. 6. 7.

High mechanical properties in combination with light weight. High thermal conductivity. Good machinability. High resistance to corrosion. Parts can be produced with no or a limited amount of machining. Parts can be cast with close dimensional tolerances. Low scrap rate.

6

Casting: An Analytical Approach

Die casting is a precision manufacturing process in which molten metal is injected at high pressure and velocity into a permanent metal mould. There are two basic die-casting processes: 1. 2.

The hot chamber process. The cold chamber process.

In a hot chamber diecast machine (Figure 1.2), a metal injection system is immersed in the molten metal. Advantages of hot chamber die-cast process are 1. 2.

Cycle time kept to a minimum. Molten metal must travel only a short distance, which ensures minimum temperature loss during cycle time.

The hotchamber process can be used only for alloys with a low melting point (lead, zinc). Alloys with a higher melting point will cause degradation of the metal injection system. The hot-chamber die cast process has these steps 1. 2. 3. 4. 5.

Hydraulic cylinder applies pressure on plunger (Figure 1.2). Plunger pushes metal from the sleeve through the gating system into the cavity (Figure 1.3a). High pressure is maintained during solidification process . After solidification is complete, the die opens (Figure 1.3b). The part is ejected from the cavity (Figure 1.3c).

Figure 1.2. Hot chamber die cast machine

Casting of Light Metals

a

7

b

c Figure 1.3. Steps in the hot-chamber die-cast process: a. plunger pushes metal from the sleeve through the gating system into the cavity; b. after solidification process is complete, the die opens; c. the part is ejected from the cavity.

The cold chamber die-casting process is used for alloys with a high melting point (aluminium, brass). In a cold chamber die-casting machine (Figure 1.4), the metal is in contact with the machine injection system for only a short period of time. A typical process consists of several steps (Figure 1.5): 1. 2. 3. 4. 5. 6.

Molten metal is ladled into the shot sleeve (Figure 1.5a). Hydraulic cylinder applies pressure on plunger (Figure 1.5b). Plunger pushes metal from the sleeve through the gating system into the cavity (Figure 1.5c). High pressure is maintained during solidification process (Figure 1.5d). After solidification is complete, the die opens (Figure 1.5e). The part is ejected from the cavity (Figure 1.5f).

8

Casting: An Analytical Approach

Figure 1.4. Cold chamber die-cast machine

1.4.2 Die-cast Dies Die-cast dies (Figure 1.6) are made from alloy tool steel and must withstand multiple cooling-heating cycles. A die-cast die usually consists of two halves, the stationary (cover) side and the ejector side. The cover side of the die is mounted to a fixed platen on the die-cast machine and the ejector side is mounted to a movable platen. The cover and ejector halves are separated at the parting line to allow for removal of the casting. The molten metal enters the die through the shot hole in the cover half of the mould. The runner system of the die allows metal to flow from the shot hole into the cavity. An ejector mechanism is mounted on the ejector side of the die. Its purpose is to push the casting out of the cavity. There are several ventilation channels cut on the parting line of the die to allow air to escape. These channels are either open to the atmosphere or connected to a vacuum system. The cooling system of the die-casting die allows excess energy to be removed to cool the casting below solidification temperature.

1.4.3 Runner System The runner system of the die-cast die delivers metal from the shut cylinder of the die-cast machine into the cavity of the die with minimum losses. As in every hydraulic system, head losses in the runner system are the product of major and minor losses. hL

hMajor  hMinor .

(1.1)

Major losses that occur due to friction between metal flowing through the runner system and cavity walls can’t be avoided. That is why calculations of the runner system have to account for these losses.

Casting of Light Metals

a

b

c

d

e

f

9

Figure 1.5. Steps in the cold chamber die-cast process: a. molten metal is ladled into the shot sleeve; b. hydraulic cylinder applies pressure on plunger; c. plunger pushes metal from the sleeve through the gating system into the cavity; d. high pressure is maintained during solidification; e. after solidification is complete, the die opens; f. the part is ejected from the cavity.

10

Casting: An Analytical Approach

Figure 1.6. Die-cast die

Major losses can be described by the Darcy-Weisbach equation:1 hMajor

f

l V2 . D 2g

(1.2)

where f = coefficient of friction L = length of the flow pass D = equivalent diameter of the runner system V = velocity of the metal through the gate g = acceleration of gravity. Minor head losses that are results of dissipation of kinetic energy can be minimized or completely avoided (Figure 1.7).

1

Equation developed as a variation of the empirical equation developed by Gaspard de Prony.

Casting of Light Metals

11

Figure 1.6. Flow pattern for the sharp and rounded edge runner

1.4.4 Cavity of Die-cast Die After the first stage of the die cast process is complete and metal has been delivered to the cavity, the flow pattern through the cavity has to be designed to fill the cavity in the most efficient way. Die-cast dies use forced cooling method where water, oil or steam flows through the cooling channels. The location and effectivness of the die-cast die cooling system are limiting factors in calculating cavity fill time. Longer fill time can result in metal reaching point where an increase in metal density can significantly influence fluidity of the metal, which will result in a nonfill condition. In the die design stage a lot of consideration has to be given to a location of the gates, because they will define the flow pattern of the metal through the cavity. After runner and gate size and location have been defined, the cooling system has to be designed to allow for uniform cavity temperature. 1.4.5 Air Ventilation System The air ventilation system is a part of the die-cast die. After the die-cast die is closed and ready for the next cycle, the shot cylinder, gate system, cavity, and ventilation system are filled with air. As the molten metal is forced through the gating system into the cavity, it displaces air. Air flows to the outside or into a vacuum tank through the ventilation channels. Correctly placed and sized ventilation channels allow most of the air to escape from the cavity. Air compressibility, coefficient of friction, temperature changes in the air, and the shape of the ventilation channel all have to be taken into account to prevent pressure loss and shock wave formation in the airflow stream. Gases entrapped in the cavity form internal porosity. They are a major defect in the high-pressure die-casting process. Gas porosity can originate from many sources. A properly designed ventilation system can minimize the amount of gas entrapped in the cavity.

12

Casting: An Analytical Approach

1.4.6 Ventilation Blocks Properly positioned ventilation channels are connected to the last area of the diecast die cavity to be filled. To prevent the molten metal from escaping through the ventilation channels, ventilation blocks are used. Ventilation blocks are used for both conventional and vacuum ventilation systems. When metal flows through the ventilation blocks due to the thermal exchange between the blocks and the molten metal, its temperature is reduced until it reaches its critical point of rigidity. There are two different types of ventilation blocks - valve - valveless. Ventilation blocks with valves have larger cross-sectional areas in the gas evacuation channel. The disadvantage of these types of ventilation blocks is a mechanical system that shuts-off the ventilation channel as soon as metal reaches the valve. Periodic failure, constant maintenance requirements and low reliability of the mechanical system limit the application of this type of ventilation system.

2 Introduction to Fluid Dynamics

2.1 Basic Concepts The behaviour of metals during filling and solidification can be described by applying the laws of fluid dynamics. In the liquid state, metals behave largely like common liquids such as water or liquefied natural gas. Molecules in liquids do not form a rigid crystalloid structure and, therefore, can move easily relative to each other. This behaviour distinguishes fluids in general from solid materials. At the same time, these molecules are packed sufficiently close to each other to experience strong forces of mutual attraction that make it hard to pull a piece of liquid apart. For the same reason, it is also hard to compress a liquid to a smaller volume. Therefore, liquids, unlike gases, can be treated as essentially incompressible materials, a property that greatly simplifies the governing equations.2 Fluid flow behaviour is characterized by density, pressure, temperature and velocity. Density, U, is the amount of mass, represented by molecules, in a unit volume of fluid. The incompressibility property implies that density stays constant during flow. In other words, no matter how a fluid is stretched, sheared or pressed, the number of molecules in a fixed volume stays more or less constant, even though some molecules may have moved out of it and others have entered it in their place. 2.1.1 Pressure The resistance of fluid to compression is characterized by pressure. Huge pressures must be applied to compress a fluid by as little as 1% of its original volume. In

2 In this chapter, when referring to metals we will use the terms liquid and fluid. Although the term fluid generally includes both incompressible liquids and compressible gases, we will primarily mean the former unless specifically clarified otherwise.

14

Casting: An Analytical Approach

most situations, even in high-pressure die casting, the pressures are not sufficient to change the fluid volume noticeably. Pressure is one of the main parameters that control the flow of fluid. It is related to the rate at which molecules transfer the momentum of their random microscopic motion to their neighbours through collisions. Since this random motion occurs in all directions, pressure at a point in the fluid also acts in all directions. But when molecules in one part of the fluid transfer more momentum to the molecules in an adjacent region than they receive in return, a macroscopic force arises between these fluid regions. This force can be described as the pressure gradient, which is the difference in pressures at two locations in the fluid, divided by the distance between those points. When pressure P is a function of the threedimensional coordinates x, y and z, then at every point in the fluid, the pressure gradient is a vector, defined as

’P

§ wP wP wP · ¸¸ . ¨¨ , , © wx wy wz ¹

(2.1)

In the absence of other forces, flow is initiated in the direction of the pressure variation from high values to low, that is, in the direction opposite to the direction of the pressure gradient, as shown in Figure 2.1.

Figure 2.1. Iso-lines of pressure (isobars) showing the distribution of pressure and the direction of the pressure gradient

2.1.2 Viscosity

As with any moving objects, the motion of most fluids experiences additional forces due to friction. Frictional forces also arise from the collisions of molecules in a moving fluid with molecules in the slower adjacent fluid regions. These collisions result in a net transfer of momentum from the faster flow regions to the

Introduction to Fluid Dynamics

15

slower ones, giving rise to frictional force. This force acts in the direction opposite the faster flow, dissipating its energy and generating heat, similar to the effect of taxes on the flow of capital. For example, if two streams of fluid are moving at different speeds parallel and next to each another, the faster moving stream will gradually slow down and the slower one will accelerate. As a result, the boundary between the two layers will widen with time leading to the development of the viscous boundary layer, the region where flow transitions from the velocity in one fluid layer to the velocity in the other. It is noteworthy that, as the two fluid layers exchange momentum, the overall kinetic energy in the flow decreases.

Figure 2.2. Viscous frictional force acting between two streams of fluid

The frictional properties of a fluid are conveniently described with a single variable called the dynamic viscosity coefficient P, and these forces are called viscous forces or stresses. Fluids with larger dynamic viscosity coefficients generate higher viscous forces than less viscous fluids in the same flow conditions. Additionally, larger differences in velocities result in a higher rate of transfer of the momentum from the faster moving fluid to the slower and hence in more viscous friction. Finally, the distance between two fluid regions also plays a role: the closer they are, the faster the transfer of momentum occurs. According to these observations, the viscous frictional force Ffr acting between the two streams of fluid in Figure 2.2 can be estimated as Ffr

AP

U1  U 2 . d

(2.2)

where A is the contact area between the streams, U1 and U2 the average velocities in the two streams and d the distance between them. In this form, Ffr is the force acting on the fluid moving with the velocity U2. For the fluid stream with the velocity U1 the sign of the force is the opposite. In a differential form, the righthand side of Equation 2.2 can be expressed for a unit contact area as

16

Casting: An Analytical Approach

Ffr A

P

wU . wz

(2.3)

where z is the coordinate axis normal to the direction of flow. The expression on the right-hand side of Equation 2.3 is called shear stress. Note that viscous stresses are reduced to zero in a uniform flow since the righthand side of Equation 2.3 vanishes when there is no velocity variation. However, it is hard to achieve uniform flow in practical situations where a fluid is typically confined by the walls of the channel or the container. Fluid molecules collide with these walls and bounce back. The surface of a typical material has roughness that far exceeds the size of a fluid molecule. Even a super finished metal surface at best has a roughness in excess of tens of nanometres. This is still more than a hundred times bigger than a fluid molecule. Other cutting and finishing techniques produce roughness in the range from 100 to 50,000 nm (0.1 to 50 Pm). So for a fluid molecule hitting a wall, its surface looks like the Black Forest to a football. After several collisions, it is very likely to lose all information about where it was coming from before it hit the surface. The usually irregular shape of the molecules and atoms only accelerates the “loss of memory.” When a fluid molecule returns into the flow after interacting with the wall, its original momentum component normal to the wall may be retained, but the direction of the tangential component is completely random. In macroscopic terms this behaviour is expressed in the form of the no-slip boundary condition. It means that the fluid velocity component tangential to the surface of a wall boundary is equal to zero. The no-slip boundary condition means that friction, or viscous shear stress, is always present in a flow near walls. In addition to the pressure gradient, it is one of the main factors controlling flow. It leads to the development of viscous boundary layers, in which flow transitions from zero velocity at the surface to the flow in the bulk. Moreover, the relatively large size of the surface roughness may produce more flow loss than can be suggested just by its interaction with the individual fluid molecules. Large clusters of these molecules can be deflected, redirected and trapped by the small bumps and pits on the surface that make up the surface roughness. This may contribute to the development of turbulence in the flow. Turbulence can be described as a form of flow instability, when random oscillatory motion develops in the otherwise ordered mean fluid flow. This random motion occurs on much larger time and length scales than the molecular motion, but its effect is similar. It accelerates the transfer of momentum between different parts of the fluid and, therefore, results in more friction. 2.1.3 Temperature and Enthalpy

The thermal state of a fluid is usually represented by temperature, T, which is a measure of and proportional to the kinetic energy of the chaotic motion of its molecules. Fluid specific thermal energy¸ I, is proportional to the temperature I

CVT ,

(2.4)

Introduction to Fluid Dynamics

17

with the coefficient of proportionality CV, called the specific heat at constant volume. It is equal to the amount of heat that is needed to raise the temperature of a unit mass of fluid by 1o. The subscript ‘V’ means that the volume of fluid would be kept constant during such a procedure. This clarification is necessary for a compressible gas, which, if allowed to expand upon heating, would require more energy to raise its temperature. For incompressible fluids, this distinction is not very important. As a result, the value of the specific heat at constant volume is very close to that of the specific heat at constant pressure, Cp. For obvious reasons, it is easier the measure Cp for metals by simply keeping the specimen at atmospheric pressure during measurement, whereas for a gas placed in a fixed container, it is easier to measure CV. CP is used to calculate another useful quantity called enthalpy, E, C P T  L(1  f S ) .

E

(2.5)

The second term on the right-hand side accounts for the release of thermal energy during solidification. Fluid molecules in the liquid phase have more freedom to move than in the solid state where they are locked in a crystalloid structure. As the metal cools and passes from the liquid state to the solid, the excess energy is released in the form of latent heat. The solid fraction, fS, is the mass fraction of the solidified phase in a given amount of metal. Upon cooling, its value changes from 0.0 in the pure liquid to 1.0 in the pure solid phase allowing for the latent heat release in Equation 2.5. One of the mechanisms for the exchange of thermal energy within fluids is thermal conduction. As molecules collide with each other, they transfer momentum, which is responsible for pressure and viscous forces, and also the kinetic energy of their chaotic motion. Consequently, any temperature variations in a thermally insulated volume of fluid would disappear over time, resulting in a uniform temperature distribution. The rate of heat exchange by conduction is described by the thermal conduction coefficient, k. The heat flux q between two regions of fluid at temperatures T1 and T2 separated by the distance d is then calculated as q

k

T2  T1 , d

(2.6)

wT . wx

(2.7)

or in differential form,

q

k

Equation 2.7 is the Fourier law stating that the heat flux by thermal conduction is linearly proportional to the temperature gradient [Holman, 1976]. Note that the form of Equation 2.7 is similar to that of Equation 2.3 for the viscous dissipation of momentum.

18

Casting: An Analytical Approach

2.2 Equations of Motion Pressure gradients and viscous stresses are the main internal forces present in fluids. External forces can include gravity and electro-magnetic forces. According to Newton’s second law, the sum of all these forces results in a net acceleration of the fluid, which is inversely proportional to its mass, or density. This can be expressed in the form of the Navier-Stokes equations, which for an incompressible viscous fluid can be written in the following form [Batchelor, 1967] wu wu wu wu  u  v  w wt wx wy wz



wv wv wv wv  w  v  u wz wy wx wt



ww ww ww ww  u  v  w wt wx wy wz

1 wP P § w 2u w 2u w 2u · ¸  Gx  ˜ ¨¨ 2   U wx U © wx wy 2 wz 2 ¸¹

w 2v · w 2v 1 wP P § w 2v (2.8) ¸  Gy   ˜ ¨¨ 2  wz 2 ¸¹ wy 2 U wy U © wx 1 wP P § w2w w2w w2w · ¸  Gz .   ˜¨   wy 2 wz 2 ¸¹ U wz U ¨© wx 2

Here u, v and w are the three components of the fluid velocity vector U at any point in the flow, and G = (Gx, Gy, Gz) is the external force, which we will assume here consists only of gravity. The left-hand side of Equation 2.8 represents the components of fluid acceleration, the components of the pressure gradient, viscous stresses and gravity are summed up on the right-hand side. These forces are divided by fluid density U, therefore, the same forces would produce a higher acceleration for a lighter fluid. The ratio of the dynamic viscosity coefficient and density is often called the kinematic viscosity coefficient Q PU. Mass conservation is another important law governing the motion of fluids. It states that mass cannot be created or lost and is expressed through the continuity equation. For incompressible fluids, this equation reduces to the condition of zero divergence of the velocity vector divU

wu wv ww   wx wy wz

0.

(2.9)

and simply means that for any amount of fluid entering a given volume from one side, exactly the same amount must leave on the other side. When heat transfer and solidification are of interest, then additional equations are needed to track the evolution of temperature and the solid fraction. This is done in the energy conservation equation, which, similar to the mass conservation one, says that energy is not lost or created. As for the equation of motion, the energy transport equation is simplified by the assumption of incompressibility. Written in terms of enthalpy, defined in Equation 2.5, it has the following form:

Introduction to Fluid Dynamics

wE wE wE wE u v  w wt wx wy wz

w 2T w 2T k § w 2T ¨ 2   U ¨© wx wy 2 wz 2

· ¸. ¸ ¹

19

(2.10)

The left-hand side of Equation 2.10 constitutes the rate of change of enthalpy, and the right-hand side describes thermal conduction. An appropriate relationship between solid fraction and temperature must also be devised to complete the model. Equations 2.8 – 2.10 constitute the basic set of equations describing the evolution of an incompressible fluid such as metal. It can be applied to a wide range of flow problems, from ocean currents to MEMS, from external to internal flows, steady-state or transient. Metal casting, of course, is one of the areas where the rules of fluid dynamics can be used. When turbulence is present, conventional turbulence models seek to enhance viscous mixing and dissipation in the flow by evaluating the turbulent dynamic viscosity coefficient and using it in Equation 2.8 in place of the molecular value [Batchelor, 1967]. The left-hand sides of Equations 2.8 and 2.10 have similar forms and describe the transport of the quantities shown in the partial derivatives (u, v, w in Equation 2.8, and E in Equation 2.10). The leading term is called the temporal derivative. It is the rate of change of a quantity at a given point in the flow. For instance, wE / wt could be evaluated by inserting a thermocouple into the flow and then using its readings and Equation 2.5. The rest of the terms on the left-hand sides of these equations are convective terms. They are responsible for carrying fluid quantities with the flow and are characteristic of continuum mechanics when a particle of fluid moves, another particle comes in its place bringing with it its unique properties such as temperature and velocity. Diffusion is another means of transport in fluids. The diffusion of thermal energy is described by the thermal conduction terms on the right-hand side of Equation 2.10. The diffusion of momentum is represented by the terms in parentheses on the right-hand side of Equation 2.8. In incompressible fluids, as well as in solids, pressure can actually be negative because the intermolecular forces in these materials include the forces of attraction that are responsible for keeping the molecules close together. Pressure in Equation 2.8 can be relative, or gauge pressure. For example, it can be set relative to one atmosphere, in which case the normal pressure will be equal to zero. This is possible for incompressible materials because pressure in the equations of motion is present only in the gradient operand, therefore, adding or subtracting a constant does not change the flow dynamics.

2.3 Boundary Conditions Equations 2.8 – 2.10 are usually solved in a finite domain that has external and internal boundaries. Therefore, proper descriptions of these boundaries, or boundary conditions, are needed to find the flow solution. In addition to material properties, boundary conditions distinguish low-pressure from high-pressure die

20

Casting: An Analytical Approach

casting or lost foam casting from gravity pour. Boundary conditions, therefore, play an important role in determining the solution, and it is worth saying a few words about them here. 2.3.1 Velocity Boundary Conditions at Walls

There are two flow boundary conditions at the walls bounding the flow. Since fluid cannot penetrate solid obstacles, the component of the velocity normal to the wall must be equal to zero:

U˜n

u ˜ nx  v ˜ n y  w ˜ nz

0,

(2.11)

where n(nx,ny,nz) is the unit length vector normal to the wall surface. The second boundary condition enforces the no-slip condition, that is, the velocity component tangential to the wall must also be equal to zero: UIJ

0.

(2.12)

Combined together, Equations 2.11 and 2.12 simply state that flow velocity at the wall is equal to zero. It is useful, however, to define the two conditions separately since Equation 2.12 is not necessary when an inviscid flow approximation is used (i.e., when viscous stresses are small and can be neglected in Equation 2.8). 2.3.2 Thermal Boundary Conditions at Walls A boundary condition at walls is also needed for Equation 2.10 for enthalpy. This is typically done by defining a heat flux, q, at the interface between fluid (metal) and wall (mould) as follows: q

h ˜ (Tfluid  Twall )

(2.13)

with the heat transfer coefficient, h, representing the thermal properties of the interface itself. Factors like surface roughness, coating or lubrication affect the value of h. The wall boundary condition given by Equation 2.13 can be replaced by the one that directly specifies the heat flux, possibly as a function of time, q

q0 (t ) .

(2.14)

Equation 2.14 is useful when modeling exothermic sleeves or water-cooled mould surfaces.

Introduction to Fluid Dynamics

21

2.3.3 Free Surface Boundary Conditions Free surface is a special type of boundary; it moves with the liquid. The influence of air on flow can usually be ignored because air is much lighter than most liquids, especially metals. The fact that free surface is a boundary between a liquid and the ambient air is expressed in the so-called kinematic boundary condition, stating that the velocity of the free surface, Ub, is equal to the velocity of the liquid: Ub

U(t , x, y, z ) .

(2.15)

This obvious condition is nevertheless necessary to include free surface properly in the flow model. Equation 2.15 ensures that liquid and free surface do not get separated. The lightness of the ambient air in comparison with liquid gives rise to the dynamic boundary conditions at a free surface. The first one states that fluid pressure at a free surface, P0, is equal to the air pressure, Pa. P0

Pa .

(2.16)

Moreover, if we ignore the variation of pressure in the air due to gravity, then Pa is constant along a contiguous section of the free surface. This does not necessarily mean that it is constant in time, however. For example, during filling, the air may not be able to escape quickly enough, causing the air pressure in the cavity to increase, thus making Pa a function of time. Moreover, multiple air pockets will generally have as many different pressures, each serving as the boundary condition for the segment of metal surface bounding the respective air pocket. Surface tension forces at a free surface can also be taken into account. A liquid molecule located at the free surface interacts with the liquid molecules on one side of the interface and with the adjacent air molecules on the other side. The asymmetry of the inter molecular forces gives rise to a macroscopic force, which is proportional to the curvature of the interface. This force is typically expressed in terms of the surface tension pressure, Ps, which is a product of the surface tension coefficient, V, and the interface curvature, N, Ps

VN

V (’n) ,

(2.17)

where ’ ˜ n wnx wx  wn y wy  wnz wz is the divergence of the unit outward normal vector of the surface ( Figure 2.3). Liquid metals have the highest surface tension coefficients among liquids, with mercury leading the pack. Additionally, the buildup of a surface film due to the oxidation of metal in contact with air adds to the molecular forces at a free surface [Campbell, 1991].

3

The surface tension coefficient is not so much a property of the fluid as of the interface between two media, such as aluminium and air.

22

Casting: An Analytical Approach

Figure 2.3. Surface tension pressure acting on an element of free surface

Surface tension is an important force when the free surface curvature is large as, for example, in small droplets in an atomized flow common in high-pressure die casting. Equation 2.16 then needs to be modified to include the surface tension force. Pa  Ps .

P0

(2.18)

The second dynamic boundary condition at a free surface is derived from the assertion that viscous friction between fluid and air is negligibly small or, using Equation 2.3, wU wn

0.

(2.19)

where the derivative of the fluid velocity near a free surface is taken in the direction normal to the surface. Thermal boundary conditions at the free surface during casting are often assumed to be adiabatic, i.e., for simplicity heat losses to the air are neglected in comparison with the heat fluxes inside metal and at mould walls. However, more realistic relationships, similar to that given by Equation 2.13, can also be used. For example, radiative heat losses, qR, which maybe important for high temperature alloys, can be computed as qR

4  Tair4 ) , H ˜ ] ˜ (Tfluid

(2.20)

where H is the emissivity of the surface (H < 1), ]=5.5604˜10-8 kg s-3 K-4 the StefanBoltzmann constant, and temperature is expressed in the absolute units of degrees

Introduction to Fluid Dynamics

23

Kelvin, K. Due to the power of four on the right-hand side of Equation 2.20, the radiative heat flux grows quickly with an increase in surface temperature. For example, the pouring temperature of steels, 1700 – 1800 K, is around twice that of a die-cast aluminium alloy and, therefore, with similar emissivity coefficients, the radiative heat loss from the surface of the steel is about sixteen times larger.

2.4 Useful Dimensionless Numbers Equations 2.8 – 2.10, together with the appropriate boundary conditions, describe a very wide range of flows. It is often useful to estimate the relative importance of various terms in these equations and thus determine the most significant aspects of the physical behaviour of the fluid in a given situation. This, in turn, may enable simplification of the equations before one proceeds with the solution. As a minimum, it would be useful to understand what type of flow to expect. A set of dimensionless numbers, each representing an estimate of the ratio of a pair of forces, can be conveniently employed for that purpose. These numbers are derived from the dimensionless form of the equations of motion. This form, in turn, is obtained by scaling the equations by the characteristic values of length and velocity. As their name suggests, for a given flow each dimensionless number has the same value, irrespective of the units system employed to evaluate it. 2.4.1 Definitions The commonly used dimensionless numbers are Reynolds number: Re

UUd P

Weber number:

UU 2l V

Bond number:

Froude number:

We

Bo

Fr

UGl 2 V

U Gh

fluid inertia . viscous forces

(2.21)

fluid inertia

.

(2.22)

surface tenstion

gravity surface

.

(2.23)

tension

fluid inertia . gravity

(2.24)

24

Casting: An Analytical Approach

Here U is the characteristic velocity, d, l and h denote the appropriate characteristic lengths and G is gravity. 2.4.2 The Reynolds Number For the Reynolds number U is the average variation of the velocity in the flow between its minimum and maximum values, and d is the distance over which this variation occurs. According to Equation 2.12, in a typical filling, U can be defined as the difference between the velocity at the walls, which is zero, and in the bulk of the flow, or as the average metal velocity. Then d becomes half of the minimum wall thickness or half of the channel width. The Reynolds number is one of the most important parameters characterizing fluid flow. When its value is small, Re < 1, then flow is dominated by viscous forces. For very small values of Re, the convective terms in Equation 2.8 can be neglected in comparison with viscous dissipation of the momentum, reducing it to the so-called creeping, or Stokes, flow approximation. As is shown in the next section, the Reynolds number in metal flow in most castings is much greater than one, indicating that, at least during filling, viscosity plays a secondary role to fluid inertia. With the increase in the speed of the flow, it transitions from laminar to turbulent due to the development of flow instabilities initiated by spatial variations in fluid velocity. The transition begins at Re | 2000 and turns into a fully turbulent flow when Re exceeds 10,000. Only in extremely carefully controlled flow experiments can the laminar regime be extended to Re up to 20,000. Fully developed turbulence enhances the dissipation of fluid momentum, in addition and significantly beyond the dissipation due to the molecular viscosity, even though a large value of the Reynolds number may suggest that viscous forces are not important in the flow. 2.4.3 The Weber Number In Equation 2.22 for the Weber number, U characterizes the average variation in fluid velocity near a free surface. To be more precise, it is the velocity component normal to the free surface that is of the interest here. Due to the no-slip boundary condition at walls and Equation 2.15, we can say the U is the average velocity of the free surface. As with the Reynolds number, the distance l then becomes the minimum width of the flow channel. During filling, internal fluid forces can cause distortion of the metal surface, sometimes called surface turbulence [Campbell, 1991], that would lead to folding of the surface, additional oxidation and other undesirable effects. The process can be visualized by imagining a submerged jet of metal directed at an area of the free surface. Its energy will create a bulge on the initially undisturbed surface. The Weber number can be used to determine if the surface tension forces can prevent the rupture of the surface film and restore its shape. The velocity U and distance l in Equation 2.22 in this case relate to the jet velocity and size of the bulge, respectively. If We < 1, then we can hope that the energy of the flow will be contained within the confines of the existing free surface. If We > 1, as is the case

Introduction to Fluid Dynamics

25

in most filling scenarios, then the folding and entrainment of the surface oxide film and possibly air cannot be avoided. It has also been observed experimentally that a free surface breaks up into small droplets when the Weber number exceeds the critical value of around 60 [Manzello and Yang 2003]. 2.4.4 The Bond Number The Bond number is another measure of the relative importance of surface tension. This time it is compared to gravity, which is useful to determine if a free surface will stay flat or bulge. The natural tendency of the surface tension forces is to bend the initially horizontal free surface to reach a constant curvature at its every point, and in the absence of other forces it will do just that. Gravity in this case acts in the opposite direction trying to flatten it. When gravity is strong and the surface’s horizontal extent l is large, that is Bo > 1, a free surface is likely to stay flat and undisturbed by the surface tension forces as in a glass of water or a metal pouring cup. If the size of the container is gradually reduced, then at some point the value of the Bond number will drop below unity and the shape of the free surface will be determined more by the surface tension than by gravity. This can be observed inside a half-filled (transparent) drinking straw or when placing a small droplet of water on a dry surface. 2.4.5 The Froude Number The Froude number is often employed to estimate the importance of such as surface waves in open-channel flows, like rivers and canals. It is also useful to look at the waves in the horizontal runners in gravity pour castings and shot sleeves in high-pressure die casting. In all these cases, the waves are driven by gravity. The variable h in Equation 2.24 is the average depth of the fluid. When Fr is much smaller than one, Fr 1, the flow is faster than the surface waves, or super critical. Any such waves are quickly swept away by the flow toward the boundaries of the flow domain. The fact that these waves can move in only one direction may result in their accumulation at the downstream walls. This, in turn, produces a buildup of fluid near the walls and eventually develops into a hydraulic jump, a narrow area in the flow in which the fluid transitions from the high velocity upstream to the low velocity downstream of the jump. The transition of the flow from one side of a hydraulic jump to the other is also characterized by an abrupt change in pressure, fluid depth and, of course, turbulence. The latter often results in excessive entrainment of air into the bulk of the fluid at the transition point, which is highly undesirable during mould filling.

26

Casting: An Analytical Approach

2.5 The Bernoulli Equation Once of the most commonly used solutions of the general fluid motion equations is the Bernoulli equation. It can be derived from Equations 2.8 and 2.9 when the flow is steady and inviscid, and can be expressed in the following form P 

1 UU 2  Ugh 2

C,

(2.25)

where g is the magnitude of the gravity vector and h is the height above a reference point. C is an abitrary parameter that is constant along any streamline. It can be evaluated by using pressure and velocity at a single point along the streamline P 

1 UU 2  Ugh 2

P1 

1 UU12  Ugh1 . 2

(2.26)

Stagnation, Dynamic and Total Pressure If the variation in fluid elevation h is small or gravity forces are negligible compared to pressure and inertia, like in high pressure die casting or in air, then Equation 2.26 can be reduced to P 

1 UU 2 2

P1 

1 UU12 . 2

(2.27)

As fluid accelerates along a streamline, pressure drops so that the sum on the lefthand side of Equation 2.27 stays constant. The maximum value of pressure occurs at the point where velocity is zero, or at the stagnation point. This pressure is called stagnation pressure. The term 1/2UU2 is the dynamic pressure, as opposed to the static pressure represented by P. The sum of static and dynamic pressures in Equation 2.27 is termed the total pressure. The Bernoulli equation in the form of Equation 2.27 led to the development of the theory of the airfoil [Abbott, 1959]. The difference between the static pressures on the lower and upper surfaces of an airplane wing creates the lift necessary to keep the plane in the air.

2.6 Compressible Flow Strictly speaking, all fluids are somewhat compressible. In other words, if external pressure is applied to a fluid volume, the latter will decrease in size. Among other things, compressibility of materials enables the propagation of acoustic waves. For most liquids, however, this change is negligible, even if the pressure is large. Fluids for which the compressibility effect is significant are called gases. The average distance between molecules in a gas is large, much larger than the size of

Introduction to Fluid Dynamics

27

the molecules themselves. This allows them to move freely in space, interacting with other molecules mostly through collisions. Unlike liquids, gases occupy all available space bound by solid or liquid surfaces as, for example, propane in a steel tank or an air bubble inside liquid metal. 2.6.1 Equation of State If a gas is not too dense and sufficiently hot, then two things can be said about its molecules. First, they interact with each other mostly through collisions, with only two molecules participating in any collision. Second, the kinetic energy of the molecules comes primarily from their translational motion. That is, molecules of a gas can be closely approximated by small, elastic, identical spherical balls moving around and colliding with each other in a chaotic manner. Such a fluid is called an ideal or perfect gas [Sedov, 1972]. The variables that define a thermodynamic state of a gas are pressure, density and temperature. For an ideal gas, they are related to each other through the equation of state: P

RUT ,

(2.28)

where R=8.3144 J mol– 1 K–-1 is the universal gas constant. One important result of this equation is that the thermodynamic state of an ideal gas is defined by just two parameters: density and temperature, pressure and temperature or pressure and density. Equation 2.28 is a very common equation of state that has been successfully applied to many real gases. In general, molecules in a real gas are far from spherical, or elastic, or even of the same size. Therefore, their rotational and oscillatory motions contribute to the total kinetic energy and are also exchanged during collisions. Moreover, if the gas is dense and cold, interactions between a pair of molecules cannot be described as simple collisions. In this case, the exchange of energy and momentum between molecules occurs over longer distances and times and with multiple molecules interacting at the same time. All these factors result in the behaviour that deviates from Equation 2.28. However, it is only significant at near cryogenic temperatures or very high pressures. For most gases, they are negligible in a wide range of temperatures and pressures. Air is an example of a compressible multi component real gas that can be described by Equation 2.28 with good accuracy. When modelling gas flow, the absolute values of pressure and temperature must be used. Degrees Kelvin or Rankine should be used for temperature. Unlike incompressible fluids, gauge pressure is not used for gases because pressure is present in the equation of state. The use of the absolute scale for these parameters is important for Equation 2.28 to be valid. A pressure of one atmosphere is 1.013 106 dyne/cm2 in CGS units or 1.013 105 N/m2 in SI units. Pressure, temperature and density for gases are always positive.

28

Casting: An Analytical Approach

2.6.2 Equations of Motion In general, the density of a gas can vary in time and space. The continuity equation that we wrote for incompressible fluids, Equation 2.9, is not valid in this case. It must be replaced by the full transport equation for density wU wUu wUv wUw    wt wx wy wz

0.

(2.29)

The full form of the specific thermal energy transport equation for gases. w ( UI ) w ( UI ) w ( UI ) w ( UI ) u  v  w wt wx wy wz § w 2T w 2T w 2T k ¨¨ 2   2 wz 2 wy © wx

· § wu wv ww · ¸  P¨¨ ¸.   ¸ x y w w wz ¸¹ © ¹

(2.30)

The last term on the right-hand side is the work term associated with the compression and expansion of the gas. It is equal to zero for incompressible fluids. Equation 2.30 manifests the first law of thermodynamics described in Section 2.6.3 below. Solution of the flow equations for liquids, Equations 2.8 and 2.9, is not coupled to the energy equation since neither density nor pressure depend directly on temperature, so that, generally, the solution of the energy transport equation, Equation 2.10, for liquids is optional. This is no longer true for gases. Both pressure and density depend on temperature through the equation of state. Therefore, the thermal energy transport equation above must always be included in the solution for gas flow. Compared to the momentum equations for incompressible fluids, Equation 2.8, the viscous terms in the momentum equations for gases include extra terms associated with compression:

Introduction to Fluid Dynamics

29

wu wu wu wu  u  v  w wt wx wy wz w 2u w 2u ·¸ 1 wP P §¨ w 2u P w  ˜    ˜ ¸ ¨ 2 2 2 U wx U U wx wy wz ¹ © wx wv wv wv wv  u  v  w wt wx wy wz



§ wu wv ww · ¨¨ ¸  Gx   wy wz ¸¹ © wx

w 2v ·¸ w 2v wv ww · 1 wP P §¨ w 2v P w § wu ¨¨ ¸  Gy     ˜  ˜   ¸ 2 2 w w w wz ¸¹ y x y U wy U ¨ wx 2 U © w w y z © ¹ ww ww ww ww  w  v  u wz wy wx wt 

1 wP P §¨ w 2 w w 2 w w 2w ·¸ P w  ˜    ˜ ¸ U wz U ¨ wx 2 U wz wy 2 wz 2 ¹ ©

(2.31)

§ wu wv ww · ¨¨ ¸  Gz .   wy wz ¸¹ © wx

The second term in parentheses on the right-hand side contains velocity divergence and represents the viscous force associated with the compression and expansion of the gas. According to Equation 2.8, it is equal to zero for incompressible fluids. 2.6.3 Specific Heats As mentioned in Section 2.1.3, specific heats at constant volume, CV, and at constant pressure, CP, differ significantly from each other for a gas. Because, when held at constant pressure, the gas expands upon heating. A part of the thermal energy goes into the work against the external pressure, leaving less energy for the actual heating of the gas. Consequently, more thermal energy is required to raise the gas temperature by 1o than when the gas volume is kept constant, and, therefore, CP is larger than CV. The difference between CP and CV is constant and identical for all ideal gases. It can be calculated from Equation 2.28 and the first law of thermodynamics. The latter states that the change in the total thermal energy of a gas, MdI, is equal to the amount of heat added to it, q, minus the amount of work done by the gas, W, MdI

q W,

(2.32)

where M is the total mass of the gas (see Figure 2.4). The work done by an expanding or contracting gas is the product of the gas pressure and the change in its volume

W

PdV .

(2.33)

30

Casting: An Analytical Approach

Figure 2.4. Heating of a gas volume at constant pressure P, during which its temperature and density change and the gas expands by volume dV, producing work PdV

Since V

M

U

,

(2.34)

then for the change in volume, dV V



dU

U

.

(2.35)

From Equation 2.28, it follows that when a gas is heated at a constant pressure, the corresponding changes in its density and temperature are related to each other: TdU  UdT

0,

(2.36)

and the heat flux in Equation 2.32 is by definition, q

MC P dT .

(2.37)

C v dT .

(2.38)

while from Equation 2.4, dI

Now combining Equations 2.32 – 2.38 yields

Introduction to Fluid Dynamics

CP  CV

31

(2.39)

R

Equation 2.39 states that the difference between the specific heats is the same for all gases that fit the ideal gas model. 2.6.4 Adiabatic Processes An adiabatic process is a process during which no heat is added or subtracted from the system, i.e., q = 0 in Equation 2.32. Using Equations 2.32, 2.35 – 2.39, it can be derived that in such a process the change in pressure is related to the change in density according to C P dU . CV U

dP P

(2.40)

After combining Equations 2.35 and 2.40, the equation of state, exrepssed in terms of volume and pressure, can be written in the following form J

P

§V · P0 ¨¨ ¸¸ , © V0 ¹

J

CP , CV

(2.41)

with pressure P0 and volume V0 expressing the state of the gas at a certain point in time. Note that, according to Equation 2.39, J> 1 for ideal gases. According to Equation 2.41, the thermodynamic state of gas during an adiabatic process is defined by only one parameter. This parameter can be either volume (or density), or pressure, or temperature. Given the simplicity of the equation of state, the adiabatic gas model is a useful approximation to gas flows where heat fluxes are small compared to other factors affecting the energy. This is often true when the process takes a relatively short time, for example, in supersonic flows. A Bernoulli-type solution can also be derived for processes governed by Equation 2.41 (see Section 2.6.7 below). 2.6.5 Speed of Sound In an acoustic, or sound wave, the material undergoes small, localized compressions and expansions. These changes in density result in corresponding changes in pressure and temperature. In turn, the variations in pressure create a force that causes these perturbations to propagate through the medium, that is, to actually behave like a wave. The rate at which acoustic waves propagate is called the speed of sound and is a property of the material. Note that, even though the sound speed in most solid and fluid materials is relatively large (of the order of several hundred metres per second), the actual displacement of the medium in an acoustic wave is small because of the small amplitude of the fluctuations in it. In other words, there is no transport of mass, energy and momentum associated with acoustic waves.

32

Casting: An Analytical Approach

Heat transfer due to conduction in a sound wave is also negligible because temperature gradients are small and also because the rate of heat transfer due to conduction is usually small compared to the speed of sound. This means that the propagation of sound waves is an adiabatic process. For small velocities in a sound wave, the viscous effects on its propagation are also negligible. The speed of sound, a, can be easily derived from Equations of motion 2.29 and 2.31 in the following way. Let’s assume for simplicity that the gas is initially at rest, in with a uniform pressure, temperature and density and that a single acoustic wave propagates in the x direction. As the wave moves, it introduces perturbations in the gas. Since all flow perturbations in such a wave are small, we will ignore all terms in these equations that are second order and higher with respect to these parameters. A change of density dU produced by the passing wave results in a corresponding change in pressure dP, which in turn causes a change in the velocity du. The latter two are related to each other through the momentum Equation 2.31 

du

dt dP , U dx

(2.42)

where dx is the distance travelled by the wave in the time dt, that is, dx . dt

a

(2.43)

According to the mass conservation Equation 2.29, du and dU are also interrelated

dU

dtU

du . dx

(2.44)

Now substituting Equation 2.44 in Equation 2.42 yields a2

dP dU

,

(2.45)

where the derivative on the right-hand side is taken with the condition of adiabaticity. With the help of Equations 2.28, 2.34 and 2.41, this derivative can be evaluated as a

JP U

JRT .

(2.46)

Introduction to Fluid Dynamics

33

Equation 2.46 shows that in an ideal gas the speed of sound is a function solely of temperature. At higher temperatures, the molecules are more energetic and, therefore, are more capable of transmitting local changes in pressure and density to the adjacent gas volume in the end, resulting in an increase in the speed of sound. 2.6.6 Mach Number Sound waves provide the means of transmitting information about flow conditions in different parts of the gas. When the speed of the flow exceeds the speed of sound, that is, the flow travels faster than information about it, then shock waves can arise. Typically, a shock wave separates two regions of the gas with principally different flow conditions. On the upstream side, the flow is supersonic, and on the downstream side, it becomes subsonic. The transition occurs on a very small length scale which can be estimated as the distance travelled by a gas molecule between two successive collisions with other molecules. That is to say that the thickness of a shock wave can be as small as 0.03 micron. The flow of gas is characterized by the Mach number which is the ratio between the flow speed and the speed of sound M

U . a

(2.47)

Obviously, M > 1 in a supersonic flow and is less then one in the subsonic. When the Mach number if less than about 0.1, the compressibility effects can usually be neglected. It is possible then to model the gas as an incompressible liquid. Shock waves may occur when a supersonic flow meets with geometric obstructions, such as a flying airplane, or near sudden changes in the flow path, such as at an orifice or at the entrance to a vent. 2.6.7 The Bernoulli Equation for Gases The equations of motion 2.29 and 2.31 can be integrated for an adiabtic process to yields a solution similar to the Bernoulli equation for the incompressible fluids, Equation 2.25. Gravity is usually omitted for gases because its effect is small compared to pressure forces. The result is § J ·§ P · 1 2 ¸¸¨¨ ¸¸  U ¨¨ © J  1 ¹© U ¹ 2

C.

(2.48)

With the help of Equations 2.28 and 2.41, Equation 2.48 can be rendered in several other useful forms, for example,

34

Casting: An Analytical Approach

§ J ·§ P0 ¨¨ ¸¸¨¨ © J  1 ¹© U 0

·§ P ¸¨ ¸¨ P ¹© 0

· ¸ ¸ ¹

J 1 J



1 2 U 2

C,

(2.49)

and CPT 

1 2 U 2

C.

(2.50)

where the constant C is the same for all three equations. It can be determined by calculating the left-hand side at some point along the streamline. For example, it can be defined by the flow parameters at the stagnation point, that is, where U=0:

C

§ J · P0 ¨¨ ¸¸ © J  1¹ U0

C P T0 ,

(2.51)

where the subscript ‘0’ indicates the values of the respective parameters at the stagnation point. P0 and T0 are also called the total pressure and temperature, respectively.

Maximum Speed The Bernoulli equation can be used to calculate the maximum speed, Umax, that can be achieved in a steady-state adiabatic gas flow. Since J!1¸ then, according to Equation 2.49, the maximum speed is obtained at a point where pressure is equal to zero, or U max

2J P0 J  1 U0

2CPT0 .

(2.51)

Note that the maximum velocity is only a function of the gas stagnation temperature. Flow with close to adiabatic and steady-state conditions can be obtained by letting a gas escape from a large container (or cavity) through a small hole. If the container is sufficiently large compared to the hole, then we can assume that the conditions inside are close to stagnation. The maximum velocity can then be achieved by placing the container in vacuum. For air at room temperature of 300 Kelvin and CP = 1000 J/Kg K, the maximum velocity comes to about 775 m/sec.

Introduction to Fluid Dynamics

35

2.7 Computational Fluid Dynamics In this and other sections, we have presented full systems of equations of motion for fluids, together with an array of approximate approaches to solving them. These approximations have the benefit of offering analytical dependencies for various flow parameters in simple situations and are useful tools in a design process. At the same time, they provide only limited information and often cannot be successfully applied to general, transient three-dimensional flows with turbulence, heat transfer and phase change. In this case, numerical methods must be employed to solve the fluid flow equations. The science (and often art) of developing numerical approximations to the differential and integral equations of fluid motion is called computational fluid dynamics [Roache, 1985]. Fluid motion is described with non linear, transient, coupled, second-order differential equations. A numerical solution of these equations involves approximating its various terms with algebraic expressions. The resulting equations are then solved to yield an approximate solution to the original problem. The process is called simulation. 2.7.1 Computational Mesh Typically, a numerical model starts with a computational mesh, or grid. It consists of a number of interconnected elements of various shapes, e.g., tetrahedrals or bricks. These elements subdivide the physical space into small volumes where at least one node is associated with each such volume. The nodes are used to store values of the unknowns, such as pressure, temperature and velocity. The mesh is effectively the numerical space that replaces the original physical one. It provides the means for defining the flow parameters at discrete locations, setting boundary conditions and, of course, for developing numerical approximations of the fluid motion equations. The mesh discretizes the physical space. Each fluid parameter is represented in a mesh by an array of values at discrete points. Since the actual physical parameters vary continuously in space, a mesh with fine spacing between nodes provides a better representation of reality than a coarse one. We arrive then at a fundamental property of a numerical approximation: any valid numerical approximation approaches the original equations as the grid spacing is reduced to zero. If an approximation does not satisfy this condition, then it is incorrect. Reducing the grid spacing, or refining the mesh, for the same physical space results in more elements and nodes and, therefore, increases the size of the numerical model. But apart from the physical reality of fluid flow and heat transfer, there is also the reality of design cycles, computer hardware and deadlines, which combine in forcing the simulation engineers to choose a reasonable size for the mesh. Reaching a compromise between satisfying these constraints and obtaining accurate solutions is a balancing act that is no lesser an art than the CFD model development itself.

36

Casting: An Analytical Approach

Figure 2.5. Regular two-dimensional staggered computational mesh. The cell indexing i and j are shown for a two-dimensional case

The regular two-dimensional grid shown in Figure 2.5 is called staggered. All scalar quantities, such as pressure and temperature, are calculated and stored at the centre of each rectangular cell, and velocity vectors are assigned to the respective faces of the cell for better stability of the numerical algorithm. The full velocity vector can be reconstructed at the centre of the cell by simply averaging its components from the four faces. Each element of such grid has five nodes one for the scalar quantities and four for the respective velocity components. Each cell shares velocity nodes with its immediate neighbours. Grids like that one shown in Figure 2.5 are very easy to generate and store because of their regular, or structured, nature. A non uniform grid spacing adds flexibility when meshing complex flow domains. The computational cells are numbered in a consecutive manner using three indices: i in the x direction, j in the y direction and k in the z direction. This way each cell in a three-dimensional mesh can be identified by a unique address (i, j, k), similar to coordinates of a point in the physical space. Structured rectangular grids carry additional benefits of the relative ease of the development of numerical methods, transparency of the latter with respect to their relationship to the original physical problem and, finally, accuracy and stability of the numerical solutions. The oldest numerical algorithms based on the finite difference and finite volume approaches have been developed on such meshes and are still widely in use. 2.7.2 Numerical Approximations The finite difference (FD) method is based on the properties of the Taylor expansion and on the straighforward application of the definition of derivatives. It

Introduction to Fluid Dynamics

37

is the oldest of the methods applied to obtain numerical solutions to differential equations and the first application is considered to have been developed by Euler in 1768. The idea of the finite difference method is quite simple. For a function u(x), the derivative at a point x can be approximated as wu u ( x  'x)  u ( x) | . wx 'x

(2.51)

The spatial increment 'x can be now selected as the distance between two adjacent mesh nodes, u(x+'x) and u(x) are the values of the function at these nodes. Then Equation 2.51 can be rewritten in the form that is more common in CFD: u( xi )  u( xi  1 ) wu | wx xi  xi  1

ui  ui 1 . xi  xi  1

(2.52)

Here we use a subscript i to indicate the location on the x axis where the values of u and x are taken. Thus we obtain an approximation of the derivative of the function u(x) on a structured computational grid. From the definition of derivative, the accuracy of the approximation given by Equation 2.52 improves with smaller grid spacing. A similar approach is taken to approximate all first and second-order spatial derivatives in the fluid equations of motion, Equations 2.8 – 2.10 for incompressible fluids and Equations 2.29 – 2.31 for gases. The approximation of the second-order derivatives, which are present in the diffusive terms, typically requires at least three nodes. A time step, 't, is used to divide time into discrete increments so that the temporal derivatives can also be approximated: wu u (t  't )  u (t ) | wt 't

uin  1  uin , tn 1  tn

(2.53)

where superscript n refers to the order of the time points at which the values are taken. Once all derivatives are replaced with their respective finite difference approximations, the differential equations are replaced by a set of algebraic equations. As with the original equations of motion, proper boundary conditions must also be defined for the finite difference equations. These, for example, include the no-slip and heat transfer conditions at the interface between the fluid and the surrounding walls. The unknowns in that system of algebraic equations are the values of the flow parameters velocity, pressure and temperature at the mesh nodes. For a unique solution of this system to exist, the number of algebraic equations must match the number of nodes. Therefore, the size of a numerical model increases with the increase in mesh resolution. For example, for a three-dimensional mesh with a

38

Casting: An Analytical Approach

hundred nodes in each coordinate direction, a million equations must be solved. Moreover, the solution must be found at each incremental time point tn, starting at the beginning, t = 0, and finishing at a predefined moment tN = T, multiplying accordingly the number of operations required to find the complete solution. The actual procedure for the solution of each equation in the numerical model is usually quite simple. It is the huge number of the elementary operations that is so daunting for a human brain and where the electronic brain power is really needed. CFD simulations in this book use a combination of the finite difference and finite volume approaches to the solution of the fluid flow equations on computational grids similar to that shown in Figure 2.5. Unlike the finite difference method, the finite volume (FV) method uses the integral form of conservation laws. For every elementary control volume, fluxes of the conserved quantity (e.g., mass) are calculated at its boundaries. The net flux then translates into the change in that quantity inside the control volume. One of the strengths of the finite volume method is its conservation property adherence to the physical conservations laws that naturally arises from the method’s definition. For structured rectangular meshes, the look of the formal expressions for the control volume approximations is often indistinguishable from those of the finite difference method. 2.7.3 Representation of Geometry As mentioned above, finite difference methodology has been in use for many decades. In modern times, it is difficult to accurately describe the complex geometry within the framework of the traditional finite difference method. A rectangular cells is either fully open or fully blocked by the mould resulting in a familiar stair-step representation of the curved surfaces. Such a limitation makes the definition of accurate boundary conditions problematic. For example, the zigzag shape of the interface introduces unphysical additional flow losses due to friction. Also the surface area of a zigzag surface is larger than that of the original smooth surface, and the difference does not improve with mesh refinement. An innovative technique called Fractional Area Volume Obstacle Representation (FAVORTM) has been developed to remedy this problem [Hirt, 1985]. In this method, a geometric surface can cut through a rectangular mesh cell dividing it into blocked and open portions, as shown in Figure 2.6. The ratio of the open volume in a cell to its total volume is called the fractional volume. The intersections of the surface with the faces of the cell (six in three dimensions) are computed and stored as fractional areas, which are the ratios of the open area to the total area at respective cell faces. Similarly to the velocity vectors, the fractional areas are computed and stored at the staggered, cell-face locations. The complete geometry in every computational cell is thus converted into fractional volumes and areas. With adequate mesh resolution, the reconstruction of the original geometry from those fractional quantities is possible with a high degree of reliability.

Introduction to Fluid Dynamics

39

Figure 2.6. Geometric representation (shaded area) in the finite difference mesh using fractional area and volumes

The area and volume fraction are subsequently incorporated into the numerical equations and are also used to define boundary conditions. The resulting model provides significantly better accuracy in the numerical solution than the traditional stair-step approach. 2.7.4 Free Surface Tracking Free surface exists in most metals flows and, certainly, during filling. It is challenging to model a free surface in any computational environment because flow parameters and materials properties, such as density, velocity and pressure, experience a discontinuity at it. Moreover, the motion of the free surface is the result of a combination of dynamic and kinematic flow conditions. Proper account of these conditions is critical to accurate modelling of free surfaces. One of the commonly used methods to model free surfaces is the Volume-offluid (VOF) method [Hirt, 1981; Rider, 1998]. It consists of three main components: the definition of the volume-of-fluid function F, a method to solve the transport equation for F and the proper boundary conditions at the free surface.

Volume-of-fluid Function The VOF function F is defined equal to one in the fluid and to zero outside. Averaged over a cell volume, it becomes the fractional volume of fluid, that is, the amount of fluid in the cell divided by the cell’s total volume. Thus F=1.0 in a cell that is full of fluid and F=0.0 in an empty cell. A cell with free surface would have a value of F in the range between zero and one as shown in Figure 2.7.

40

Casting: An Analytical Approach

Figure 2.7. Illustration of the calculation of the Volume-of-fluid function in selected cells. The shaded area represents fluid

The VOF Equation The VOF function can be interpreted as a kind of a tracer ink added to a fluid. Then it it should be carried through space by the fluid. This consideration leads to a transport equation for the volume-of-fluid function: wFw wFv wFu wF    wz wy wx wt

0.

(2.54)

Equation 2.54 is similar to Equation 2.29 for gas density. It is often called the kinematic equation because there are no forces present. It just states that the values of F move around according to the velocity field, like smoke in the air. This applies to the free surface itself making Equation 2.54 automatically include the kinematic free surface boundary condition given by Equation 2.15.

Tracking Free Surface If conventional computational methods described earlier in this section are applied to the VOF equation the result will most likely be unsatisfactory because the free surface represents a boundary between the values of F in the fluid and outside. The derivative of F across the free surface is effectively infinite since its value changes from one to zero on an infinitely small length scale, that is, across the “thickness” of the free surface. Therefore, any attempt at approximating this variation with the finite difference method, e.g., using Equation 2.52, is bound to be very inaccurate. The most viable way to solve Equation 2.54 numerically is to use the geometric method, where the shape and location of the free surface within a computational cell are reconstructed using the values of the VOF function in its vicinity before computing the fluxes of the fluid volume at the cell faces. For example, if we know that fluid is filling the left portion of a cell, as shown in Figure 2.8, it will be some time before it crosses into the neighbour on the right.

Introduction to Fluid Dynamics

41

Figure 2.8. The calculation of fluid volume fluxes using the geometric method. The shaded area represents fluid, while vectors represent the direction of fluid motion. The cell in the middle must fill completely before fluid can cross into its neighbour on the right

Typically, the segment of the free surface within a computational cell is represented in three dimensions in a piecewise linear fashion, that is, with a section of a plane. Then its slope can be computed from the gradient of the VOF function, and its location is pinned by the value of F inside the cell. Then the respective amounts of fluid moving in or out of the cell at each of its six faces can be evaluated. This particular approach is sometimes called piecewise linear interface calculation or the PLIC method.

Free Surface Boundary Conditions The components of the VOF method described so far can be applied to both twophase (metal and air) and one-phase (metal and void) flows. In the former case, the space outside the F = 1.0 region is filled with air, and flow equations are solved for both fluids. In the one-phase case, the inertia of the air is neglected and the F=0.0 space is effectively empty, void of mass, represented only by uniform pressure and temperature. This approach has the advantage of not wasting CPU time on modelling air since in most cases the details of its motion are unimportant for the motion of the much heavier metal. However, in this case, free surface becomes one of the fluid external boundaries and requires the definition of dynamic boundary conditions. These boundary conditions are contained in Equations 2.16 - 2.20. A proper definition of the boundary conditions is important for accurate capture of free surface dynamics. 2.7.5. Summary The finite difference/finite volume approach to numerically solving the equations of fluid motion, combined with the FAVORTM and VOF methods provide the basis of the CFD calculations for incompressible and compressible flows in this book. The numerical solutions also contain turbulence, heat transfer and solidification. These and other numerical models are part of a general-purpose CFD code FLOW3D® [FLOW-3D, 2006]. This code has been used through this book to validate analytical models and to provide transient, three-dimensional solutions that are beyond those simplified approaches.

3 Part Design

3.1 Design of Light Metal Castings The advantages of light metal parts have secured them a place in modern machinebuilding industries. The ability of casting processes to produce near net-shape parts explains their popularity as a leading manufacturing method. All aspects of the engineering process, including the product, tools, process design, and engineering must be considered as a whole to achieve the best possible product. The product engineer must keep in mind that an abrupt change in the cross-sectional area of the part, thick walls, or small fillets will reduce the mechanical properties of the part. The tool designer must also understand how the die design will affect the performance of the part. An incorrectly chosen location for the ejector pins, parting line, or the water line’s layout can cause premature failure of the part. Metal temperature, process variation, and the amount of moisture in the die will also influence part performance. In this chapter, product design will be the main topic of discussion. It will help die casters understand the thought processes of the product designer. The development of Visual Basic applications and fundamentals of the static analysis of the mechanical system will help the die-cast engineer to understand product functionality, suggest modifications, and find the process that best complements the final product.

3.2 Static Analysis Static analysis is the subject of engineering mechanics that studies stationary solid bodies. In most real-life problems, bodies are subjected to complex loading. Finding stresses using analytical equations is extremely difficult. Complex cases are solved by breaking down complex loading cases into the sum of simple linear states. To simplify analytical solutions, the following assumptions must be implemented:

44

Casting: An Analytical Approach

1. The beam has a uniform cross section. 2. The beam satisfies Hooke’s law, which states that stress in the material is linearly proportional to strain. 3. The beam is of a homogeneous material with the same modulus of elasticity in compression and tension. 4. Beam deformation is significantly smaller than its size. Loads and reactions remain normal in relation to the axis of the beam. 5. The maximum stress does not exceed proportional limits. 6. The cross section of the beam after deformation remains flat and perpendicular to the longitudinal plane of symmetry. 7. No unbalanced forces are acting on the beam. Let us define the major geometric characteristics of a beam’s cross section. 3.2.1 Moment of an Area The moment of an area of an element can be determined by multiplying each element of the area dA by its distance y from the axis 0x of axis 0y. Therefore, ydA ; dS y

dS x

xdA .

(3.1)

The sum of the entire moment of an area of an element will give us the moment of the area:

Sx

³ ydA ; S

y

A

³ xdA .

(3.2)

A

3.2.2 Moment of Inertia of an Area The moment of inertia of an area is the integral of the products, obtained by multiplying each element of an area by the square of its distance from the axis.

Jx

³ y dA ; 2

A

Jy

³ x dA . 2

A

Moments of inertia of common shapes are given in Table 3.1.

(3.3)

Part Design 45

Table 3.1. Moments of inertia

Moment of inertia 1

2

Ix

Iy

h4 12

Ix

Iy

2 3 h t 3

3

Ix Iy

4

Ix Iy

ah 3 12 ha 3 12 th 3 § a · ¨ 3  1¸ 6 © h ¹ 3 ta § h · ¨ 3  1¸ 6 © a ¹

Ix

1 4

§ ah 3 Sd 4 · ¸ ¨ ¨ 3  16 ¸ ¹ ©

Iy

1 4

§ ha 3 Sd 4 · ¸ ¨  ¨ 3 16 ¸¹ ©

5

6

Ix Iy



th 3 a  H 3  h3 12 12 ht 3 a3 H  h  12 12



46

Casting: An Analytical Approach

Table 3.1. Moments of inertia (continued)

7

t8

3 ª 4 a· º § 4 «2b  2b  a  a¨ h  2b  ¸ » 2 ¹ »¼ © «¬

Ix

1 3

Iy

h 4  h  a 12

Ix Iy

4

> >

@ @

1 3 3 t h  a  ba 3  b  t a  t 3 1 3 3 t b  c  hc 3  h  t c  t 3

9

Ix

Iy

Ix

Iy

10

Sd 4 64

S D 4  d 4 64

3.3 Loading Loading is the result of the rigid body’s contact with external components. Loading can be a force, torque, or a shearing traction. In a rigid homogeneous body, stress is always proportional to a load. There are two types of forces that are studied in engineering mechanics 1.

2.

External force Distributed force Concentrated force Dynamic force Impulse force Internal force

An external force is the result of interaction with other bodies. The most important question that engineering analysis can answer is how bodies behave under an external load. Three of Newton’s laws describe the behaviour of solid bodies under an external load:

Part Design 47

x x x

Newton’s first law: If external equally balanced forces are applied to a body, it will remain at rest or continue in motion at a constant velocity. Newton’s second law: The acceleration of a body is proportional to the resultant force in the direction of the resultant force. Newton’s third law: For every external force acting on a body, there is an equal reaction opposite in direction.

3.4 Stress Components An external load applied to a body causes stress inside the body. Stress is the result of the interactions between the elements of the body caused by the external loading applied to the body.

Figure 3.1. Solid body

Imagine that every object consists of an infinite number of small elements “A” (Figure 3.1). Under stress, these elements attempt to move closer together (compression stress) or away from each other (tensile stress). These types of stresses are called normal stresses. Stresses that cause elements to slide one against another are called shear stresses. If stresses on the side of any given element are pure compression or tension, the sides of this element are called principal planes. Stresses that act on the principal plane in a direction perpendicular to them are called principal stresses. Stresses acting on an element can be 1. Linear. When stresses in two principal directions are zero, the stress is considered linear or uniaxial. 2. Biaxial.

48

Casting: An Analytical Approach

When stress in one principal direction is zero, the stress is considered biaxial or plane. 3. Triaxial. When none of the three principal stresses equals zero, the stress is considered triaxial.

Figure 3.2. Single element

3.4.1 Linear Stress Linear stresses are often found in straight bars of uniform cross section. Let us assume that the bar (Figure 3.3) loaded on both ends is under uniformly distributed stresses. Uniformly distributed stress can be calculated as

V

P , A

(3.4)

where P is the force applied at the end of the bar and A the cross-sectional area of the bar. Stresses will be the same at any cross section parallel to the end of a bar. Shear stresses are equal to zero. To calculate stress at any cross-sectional area of a bar, let’s consider the section at angle Į to the direction of stress ı (Figure 3.4).

Part Design 49

Figure 3.3. Bar under uniformly distributed stresses

Figure 3.4. Bar section at angle Į to the direction of stress ı

The cross-sectional area of the bar is defined as AD

cos D . A

(3.5)

The total stress in the bar can be calculated as

V total

P AD

P cos D A

V cos D .

(3.6)

50

Casting: An Analytical Approach

Normal and shear stresses in any cross-sectional area can be calculated as

VD

V Total cos D

WD

V Total sin D

V cos 2 D .

V 2

sin 2D .

(3.7)

(3.8)

Normal stresses are considered positive when directed outwardly. Shear stresses are positive when they are clockwise. 3.4.2 Plane Stress An oblique plane at a random location is cut by a plane element as shown on Figure 3.5.

Figure 3.5. Stresses on an oblique plane

Two mutually perpendicular stress components ıx and ıy are acting on the side of the element. Normal and shear stresses on the edge of AĮ can be found by summing all of the forces caused by the stresses in the element.

V DX

Vx  Vy 2

V DY



Vx  Vy

Vx  Vy 2

2



cos 2D  W xy sin 2D

Vx  Vy 2

cos 2D  W xy sin 2D .

(3.9)

(3.10)

Part Design 51

Vx  Vy

W DXY

2

sin 2D  W xy cos 2D .

(3.11)

Summing both sides of Equation 3.9 yields

V DX  V DY

V X  VY .

(3.12)

Equation 3.11 shows that the sum of normal stresses in perpendicular directions are the same in any cross section of a body and are equal to the sum of principal stresses in the same direction. There are two ways to determine stresses in a plane element. Analytically, stresses can be determined using Equations 3.9 and 3.11. Differentiating Equations 3.9 and 3.10 with respect to Į, setting a derivative equal to zero yields the following results: 2W XY . V X  VY

tan 2D

(3.13)

In a similar manner, by differentiating Equation 3.11, we obtain tan 2D



V X  VY . 2W XY

(3.14)

Differentiating Equations 3.10 and 3.11 with respect to Į yields

V 1, 2

V X  VY 2

§ V X  VY · 2 ¨ ¸  W XY , 2 © ¹ 2

r

§ V  VY · 2 r ¨ X ¸  W XY . 2 ¹ ©

(3.15)

2

W 1, 2

(3.16)

Graphically, stresses can be determined using Mohr’s circle.

3.4.3 Mohr’s Circle Otto Mohr introduced the graphical method for determining principal and shear stresses in 1882. Mohr’s circle is an effective way to visualize a stress state at any given point. Mohr’s circle shown in Figure 3.6, completely describes the stresses in the plane element in Figure 3.7. Normal stresses are plotted on the abscissa, and shear stresses are plotted on the ordinate. By convention, tensile stresses are positive and are plotted on the right side of the origin; compressive stresses are negative and are plotted on the left side of the origin. Shear stresses acting in a

52

Casting: An Analytical Approach

clockwise direction are positive, and stresses acting in a counter clockwise direction are negative. As shown in Figure 3.6, normal stresses V 1 and V 2 are plotted on the abscissa.

Figure 3.6. Mohr’s circle

Part Design 53

Figure 3.7. Stresses in a plane element

Equation 3.10 can be rewritten as V  Vy Vx  Vy (3.17) V DX  x cos 2D  W xy sin 2D . 2 2 1 into Equation Substituting the basic geometric equation cos 2 2D  sin 2 2D 3.17 yields

V  VY · § 2 ¸  W XY ¨ V Dx  X 2 ¹ © The radius of the Mohr’s circle is 2

§ V X  VY · 2 ¸  W XY . ¨ 2 ¹ © 2

(3.18)

§ V X  VY · 2 (3.19) ¸  W XY ¨ 2 ¹ © The distance of the Mohr’s circle from the origin can be determined as V 2  V1 V Arg . (3.20) 2 From the centre “O” of the circle, we can draw a line at the 2Į angle. The intersection of this line with the circle will be at point “A.” Coordinates of point A will be V X and W XY . Coordinates of point B will be V Y and W YX Two problems can be solved using the graphical method: 2

R

1. Stresses acting parallel to the main coordinate system become known. Stresses in the direction inclined at angle Į to the main coordinate system must be determined.

54

Casting: An Analytical Approach

2. Stresses acting at an angle to a main coordinate system become known. Stresses parallel to the main coordinate system must be calculated.

Example 3.1

An aluminium bar of a uniform cross section is suspended vertically. A uniform symmetrical load is applied at the bottom of the bar. A crosssectional element is subjected to stress plane conditions.

V x = – 60 mPa

V y = 115 mPa

W xy = 25 mPa Determine the principal stress and shear stress. We will solve this example using MATLAB® function Mohrdirect. ............................. Calculated values: .............................

Sigma1 = 118.501 mPa with angle 8.1301o Sigam2 = í 63.5014 mPa with angle 81.8699o Taymax = 91.0014 mPa with angle 37.0273o and 127.027o. Average normal stress SigmaAver = 27.5 mPa.

Part Design 55

Example 3.1 (continued)

3.5 Hooke’s Law In 1660, Robert Hooke discovered the law of elasticity, which describes the mathematical relations between stress and strain. When a straight bar is subjected to an axial tensile load, it elongates. Changes in the length of bars under a tensile load are called stretch. The elongation of an elastic body per unit of length is called strain. Bernoulli’s hypothesis states that, if the cross section of a bar was flat before deformation, it will stay flat after deformation, and will move along the axis of the bar in the direction of the applied force.

'l l

H

const ,

(3.21)

where 'l = total elongation of the bar and l = length of the bar. According to Hooke’s law, the relation between stress and strain can be described a

H

V E

,

(3.22)

where E = modulus of elasticity (Young's modulus). The modulus of elasticity is a constant characteristic of a material. Force applied to the end of the bar can be defined as P

³ VdA , A

where A is the cross-sectional area of the bar.

(3.23)

56

Casting: An Analytical Approach

Using Equation 3.22, we can obtain

V

HE = const.

(3.24)

P . A

(3.25)

Then Equation 3.23 will yield

V

3.6 Saint-Venant’s Principle To simplify the definition of the stressed-boundary conditions of a body, SaintVenant’s principle must be applied. Saint-Venant’s principle states that if forces acting on an elastic body are replaced by a statically equivalent system of forces, acting on the same relatively small area of the body, stresses in the parts of the body at a significant distance from the area of loading do not undergo any noticeable changes. This principle is shown on Figure 3.8.

Figure 3.8. Saint-Venant’s principle

Elongation can be expressed as

H

P . EA

(3.26)

Part Design 57

Lateral strain can be found as

H Lateral

'$ . $

(3.27)

Now we can define Poisson’s ratio, which relates the lateral to the axial strain for an axially loaded body:

P

H Lateral . H

(3.28)

3.7 Criteria of Failure The most important outcome of engineering calculations is determining of whether or not the structure will fail under known stress conditions. Engineering materials can be divided into two categories: 1. Ductile materials 2. Brittle materials An ultimate elongation of 5% is the dividing point between this two material types. Yield or permanent deformation is the criterion used to predict the failure of ductile materials. Rupture is the criterion used to predict the failure of brittle materials. The major characteristics of both types of materials are described in Table 3.2. Table 3.2. Characteristics of ductile and brittle materials

Ductile materials

Brittle materials

Exhibit considerable elongation

Exhibit little ultimate elongation

Yield point, yield strength clearly defined

Yield point is not clearly defined

Ultimate tensile and compressive strength are approximately the same

Stronger in compression than in tension

58

Casting: An Analytical Approach

3.7.1 Ductile Material Failure Theory There are three commonly used failure theories for ductile material:

1. 2.

3.

The maximum stress theory, which states that failure occurs whenever maximum normal stress equals or exceeds maximum tensile strength. The maximum shear stress theory, which states that failure occurs when the maximum shear strength reaches one-half of the yield strength. The maximum distortion energy theory, which states that failure occurs when von Mises stress reaches the yield strength of the material.

3.7.2 Brittle Material Failure Theory There are three commonly used failure theories for brittle material: 1. 2.

The maximum normal stress theory, which states that failure occurs when the ultimate strength is reached. The Coulomb-Mohr theory, which states that failure occurs when the stress condition satisfies the following equation:

V V1  3 t 1. V UT V UC 3.

(3.29)

A modified Mohr’s theory, which states that failure occurs when Equation 3.29 is satisfied, except in the fourth quadrant, where V 1 is in tension and

V 2 is in compression. At this point, the material is stronger than the Coulomb-Mohr theory would have predicted. 3.8 Beam Analysis Recent studies have shown that Leonardo da Vinci (1452–1519) and Galileo Galilei (1564–1642) made the first attempts to develop the beam theory; however, it was not until the late 19th century that Leonard Euler and Daniel Bernoulli developed the elementary beam theory, which for the first time gave engineers a way to calculate a beam’s deflection and loading capacity in a relatively quick and easy way. This theory was first tested during the construction of the Eiffel Tower (1887–1889) and the Ferris Wheel (1893). Euler-Bernoulli’s equation for displacement is f ( x)

d2 dx 2

§ d2y· ¸, ¨ EI ¨ dx 2 ¸¹ ©

(3.30)

Part Design 59

where E = Young’s modulus I = area moment of inertia of the beam’s cross section y = displacement The beam under consideration is of uniform cross section and isentropic material. Based on these assumptions, E and I are constant. Then the Euler-Bernoulli equation can be written as EI

d4 w( x) dx 4

f ( x) , 0 d x d L .

(3.31)

The right side of the Equation 3.31 is a forcing function:

f ( x)

q( x)  f 0G ( x  x f )  W

d G ( x  xW ) , dx

where q(x) = Distributed external load f0 = Pointwise force applied at xf IJ = Torque applied at xIJ į = Dirac4 delta function. Table 3.3. Static beam boundary conditions

1

4

Dirac delta function was introduced by British physicist Paul Diarc.

(3.32)

60

Casting: An Analytical Approach

Table 3.3. Static beam boundary conditions (continued)

Moment 0