AN INVESTIGATION ON CONFORMAL

Transcription

1 AN INVESTIGATION ON CONFORMAL COOLING IN PLASTIC INJECTION MOULDING By Abul B M Saifullah A Thesis submitted to the Faculty of Engineering & Industrial Sciences, in fulfilment of the requirements for the Degree of Doctor of Philosophy Swinburne University of Technology Hawthorn, Victoria, Australia May, 2011

2 Declaration This thesis contains no material which has been accepted for the award of any other degree or diploma at any university and to the best of my knowledge and belief contains no materials previously published or written by another person or persons except where reference is made. Abul Bashar Mohammad Saifullah May, 2011

3 Acknowledgements All praises to Almighty Allah (God) who gives me the knowledge. First of all, I would like to give thanks and gratitude to my PhD supervisor Professor S.H. Masood for his constant supervision and support throughout my PhD study. His valuable suggestions and ideas at different stages of my PhD study not only helped me to get in to the correct steps to carry out this research but also motivated me to continue the research. His guidance and scrutinising helping hand assisted me to complete this thesis writing, which at stage looked very thorny tusk for me to finish. I would like to take the privilege to thank my co supervisor Dr. Igor Sbarski for his important support and suggestion during my research, especially his important ideas to carry out the experimental verification of injection moulding and crystallisation test of the plastic part. I want to give thanks to Diecraft Pty. Ltd. Australia for their technical support and providing the plastic parts and their mould design and guidelines for conventional cooling system. I want to thanks my colleague Mostafa Nikzad for his supports and encouragements during my PhD study. I want to give a big thank to my colleague Muhammad Rakib Mansur for helping me to conduct optical image testing for my research. Finally, I would like to mention few names of my colleagues, Kadhim Alamara, Mohammad Tarikul Islam Mazumder, Mohammad Khalid Imran and Rizwanul Haque who were always helpful giving me support and ideas during my research. i

4 This thesis is dedicated to my parents as well as my wife and my new born baby ii

5 Abstract One of the most important features of mould cooling in injection moulding is the provision of suitable and adequate cooling arrangements. For many years, mould designers and researcher have been trying to improve the performance of cooling system, despite the fact that cooling system complexity is physically limited by fabrication capability of conventional straight drilling methods. This research presents a comprehensive investigation on the application of conformal cooling channels in injection moulding for different cross sectional shapes, positions and diameter by endeavoring which conformal cooling configuration gives the best results in comparison with conventional cooling channels. Mould design with efficient cooling channels, comprehensive finite element analysis and flow simulation have been performed for different plastic parts with Pro/Engineer, ANSYS and Autodesk Moldflow Insight simulation software. Structural-thermal finite element analysis has been performed for two different types of industrial plastic part moulds with conventional, conformal and bimetallic conformal cooling channels to study the robustness and longevity of the conformal cooling channel moulds in comparison with conventional cooling channel moulds. Fatigue life of the mould with different cooling channel configuration and different materials has also been predicted using high cycle fatigue formulas and stress analysis results. Laser direct metal deposition (DMD) rapid manufacturing technology has been used to fabricate conformal cooling channel in injection mould for experimental validation of the results obtained by numerical modeling. Experimental verification has also been done for a test plastic part using conventional, conformal and bi-metallic cooling channel moulds, which have been manufactured by conventional computer numerical control machining process for mini injection moulding machine. iii

6 Finally, hardness test has been done with Shore D hardness testing machine for test plastic parts which have been produced with conformal and conventional cooling channel mould. In conclusion, simulation and experimental verification show that with optimal cross section, diameter and position, the conformal cooling channel moulds reduce significant amount of cycle time as well as improve the quality of plastic parts produced by injection moulding process. iv

7 Table of contents ACKNOWLEDGEMENS...I ABSTRACT...III TABLE OF CONTENTS...V LIST OF FIGURES... XI LIST OF TABLES... XVII 1 CHAPTER INTRODUCTION OVERVIEW MOULD CYCLE TIME AND IMPORTANCE OF COOLING SYSTEM RESEARCH OBJECTIVE AVAILABLE COOLING METHOD AND IMPORTANCE OF RESEARCH CAD/CAM/CAE AND RAPID TOOLING IN INJECTION MOULDING ORGANIZATION OF THE THESIS CHAPTER LITERATURE REVIEW BACKGROUND WORKS ON NUMERICAL AND ANALYTICAL SOLUTIONS FOR INJECTION MOULDING COOLING WITH DIFFERENT COOLING SYSTEMS Modeling and Boundary Conditions for Cooling Analysis Cooling Channel Design and Optimization Other Solution Methods WORKS ON THERMAL STRESS IN INJECTION MOULDED PLASTIC PART v

8 2.4 WORKS ON MOULD FABRICATION USING RAPID TOOLING WORK ON CONFORMAL COOLING CHANNEL BY RAPID TOOLING CONCLUSION ON LITERATURE REVIEW CHAPTER CONFORMAL AND CONVENTIONAL COOLING CHANNEL DESIGN AND ANALYSIS INTRODUCTION RECOMMENDED DEPTH AND PITCH OF COOLING CHANNELS FACTORS AFFECTING COOLING IN INJECTION MOULDING Temperature Differences Thermal Properties of Materials Cooling Medium and Flow of Coolant Cooling Channel Layout Guidelines for Straight Cooling Channel Conformal Cooling Channel A CASE STUDY OF COOLING CHANNEL DESIGN Design for Cross Section Design for Diameter, Depth and Mould Material HEAT FLOW IN THE MOULD AND GOVERNING EQUATIONS Heat Exchange System in Injection Moulding Conduction and Convection Heat Transfer in Injection Moulding COOLING TIME CALCULATION FINITE ELEMENT ANALYSIS OF COOLING CHANNELS FOR OPTIMUM CROSS SECTION Transient Thermal FEA with ANSYS Workbench Boundary Condition for Thermal Analysis Other Input Parameters for Thermal Analysis RESULT AND DISCUSSION vi

9 4 CHAPTER FINITE ELEMENT THERMAL STRUCTURAL ANALYSIS OF CONFORMAL COOLING INTRODUCTION FINITE ELEMENT METHOD FOR HEAT TRANSFER ANALYSIS The Method of Weighted Residuals Bubnov-Galerkin Method Integration by Parts FINITE ELEMENT METHOD SOLVING PROCEDURE IN HEAT TRANSFER ANALYSIS Discretising the Domain Selecting the Interpolation Functions Deriving the Element Equations with Boundary Conditions Solving the Elementary Equation for Transient Analysis Assembling the System Equations Solving the Matrix Equation THERMAL STRUCTURAL FINITE ELEMENT ANALYSIS A CASE STUDY OF THERMAL-STRUCTURAL ANALYSIS Cooling Channel Design for Thermal-structural Analysis Boundary Conditions for Thermal-structural Analysis RESULTS OF FINITE ELEMENT ANALYSIS FATIGUE FAILURE AND LIFE CYCLE CALCULATION OF THE MOULD SUMMARY CHAPTER BI-METALLIC CONFORMAL COOLING INTRODUCTION CONVENTIONAL COOLING METHODS vii

10 5.3 COOLING WITH HIGH THERMAL CONDUCTIVE MATERIAL HEAT TRANSFER THROUGH COMPOSITE MATERIAL A CASE STUDY OF BI-METALLIC COOLING CHANNEL Design of Part, Mould and Cooling Channels Thermal-structural Finite Element Analysis RESULTS OF FINITE ELEMENT ANALYSIS CALCULATION OF FATIGUE LIFE OF MOULDS CALCULATION OF COOLING TIME OF MOULD EXPERIMENTAL VERIFICATION SUMMARY CHAPTER FABRICATION OF CONFORMAL COOLING CHANNELS BY RAPID MANUFACTURING INTRODUCTION RAPID MANUFACTURING OF INJECTION MOULDS DIRECT METAL DEPOSITION (DMD) DESIGN OF MOULD WITH CONFORMAL COOLING CHANNEL FABRICATION OF MOULD WITH CONFORMAL COOLING CHANNEL BY DIRECT METAL DEPOSITION (DMD) OPTICAL MICROSCOPIC STUDY OF DMD FABRICATED MOULD EXPERIMENTAL VERIFICATION OF CONFORMAL COOLING RESULT AND DISCUSSION CHAPTER CONFORMAL COOLING EFFECT ON QUALITY OF PLASTIC PART INTRODUCTION viii

11 7.2 EFFECT OF COOLING CHANNEL ON PLASTIC PART QUALITY EFFECT OF COOLING ON CRYSTALLINE MATERIALS CONFORMAL COOLING CHANNELS AFFECTING THE PART QUALITY A CASE STUDY WITH CONFORMAL COOLING CHANNELS Conformal Cooling Channel Design and Plastic Flow Analysis with Autodesk Moldflow Insight Result and Discussion of Plastic Flow Analysis EXPERIMENTAL STUDY OF HARDNESS VARIATION DUE TO CONFORMAL COOLING CHAPTER CONCLUSION AND RECOMMENDATION INTRODUCTION MAJOR FINDINGS AND CONTRIBUTION RECOMMENDATION FOR FUTURE WORK REFERENCES Appendix A.210 Mathematical Calculations 210 A1 Sample Calculation for Heat Flux 211 A2 Sample Calculation for Reynolds Number and Convective Co-efficient A3 Sample Calculation for Theoretical Cooling Time.213 A4 Sample Calculation for Fatigue Life.214 ix

12 Appendix B.215 Temperature and Equivalent Stress Distribution Images B1 Images for Temperature Distribution from FEA Simulation.215 B2 Images for Equivalent Stress Distribution from FEA Simulation.219 Appendix C 225 List of Publications for the research x

13 LIST OF FIGURES FIGURE 1.1: TYPICAL MOULDING CYCLE OF INJECTION MOULDING PROCESS (IN SECOND) BASED ON [3]... 2 FIGURE 1.2: DIFFERENCE IN TIME AND STEPS BETWEEN TRADITIONAL AND CONCURRENT RAPID TOOLING MANUFACTURING, BASED ON [19] FIGURE 1.3: THE WORKING ELEMENTS OF THE DIRECT METAL DEPOSITION (DMD) SYSTEM, BASED ON [22] FIGURE 2.1: CONFORMAL COOLING DESIGN BASED ON A FEATURE-RECOGNITION [45] FIGURE 2.2: BASIC STEPS OF RAPID TOOLING AND MANUFACTURING [7] FIGURE 2.3 : STEPS OF METAL SPRAYING RAPID TOOLING TECHNIQUE ADOPTED FROM [7] FIGURE 2.4: (A) CONFORMAL COOLING CHANNEL MADE BY COOPER DUCT, (B) BENDING OF COPPER DUCT EVENLY AROUND THE CAVITY WALL [111] FIGURE 2.5: GREEN PARTS OF AN INJECTION MOULD WITH CONFORMAL COOLING CHANNEL DESIGN MADE BY MIT S 3D PRINTING [11,114] FIGURE 2.6: COMPARISON BETWEEN CONVENTIONAL AND CONFORMAL COOLING DESIGN FOR COOLING SIMULATION [118] FIGURE 2.7: SOFT RT MOULD WITH CONFORMAL COOLING CHANNEL [21] FIGURE 2.8: WORKFLOW OF DTM RAPIDTOOL PROCESS [119] FIGURE 2.9: THE HOT PLATENING PROCESS FOR LST PRODUCTION [120] FIGURE 2.10: ELECTRIC COVER (A) CONVENTIONALLY DRILLED COOLING CHANNELS FIGURE 2.11: CROSS-SECTIONAL VIEW OF AN INJECTION MOULD ASSEMBLY BY SL TECHNIQUE [124] FIGURE 2.12: CAD DESIGN AND PROTOTYPE OF A RAPID MOULD BY PROMETAL [128] FIGURE 3.1: DIFFERENT PARAMETERS OF COOLING CHANNELS, ARE SHOWN IN A CAVITY CROSS SECTIONAL MOULD FIGURE 3.2: PLASTIC TEMPERATURE VS TIME FOR AMORPHOUS PLASTIC (LEFT) AND CRYSTALLINE PLASTIC (RIGHT) [130] FIGURE 3.3: SCHEMATIC CROSS SECTIONAL VIEW OF A MOULD SHOWING INTERFACE LAYER BETWEEN PLASTIC PART AND CORE/CAVITY xi

14 FIGURE 3.4: SCHEMATIC DIAGRAM OF COOLANT FLOW INSIDE A COOLING CHANNEL FIGURE 3.5: SCHEMATIC OF STRAIGHT COOLING CHANNELS LAYOUT, SHOWING IN A CROSS SECTIONAL TOP VIEW OF A MOULD FIGURE 3.6: BUBBLER TYPE COOLING CHANNELS FOR A SLENDER MOULD, IN A SCHEMATIC CROSS SECTIONAL FRONT VIEW OF THE MOULD FIGURE 3.7: COMPARATIVE COOLING CHANNEL LAYOUTS FOR (A) CONVENTIONAL AND FIGURE 3.8: CAD MODEL OF (A) PLASTIC BOWL, AND (B) CORE AND CAVITY IN MOULD ASSEMBLY FIGURE 3.9: DIFFERENT CROSS SECTIONAL SHAPES OF COOLING CHANNELS (A) RECTANGULAR (L/W=1.4) (B) ELLIPTICAL (C) HEXAGONAL (D) SQUARE (F) CIRCULAR FIGURE 3.10: 3-D CAD MODELS OF THE CAVITY MOULD, AS DESIGNED BY PRO/E SYSTEM, WITH DIFFERENT TYPES OF COOLING SYSTEMS; (A) CONVENTIONAL, (B) CONFORMAL WITH CIRCULAR, (C) CONFORMAL WITH HEXAGONAL, (D) CONFORMAL WITH ELLIPTICAL, (E) CONFORMAL WITH RECTANGULAR, AND (F) CONFORMAL WITH SQUARE CROSS SECTIONAL COOLING CHANNELS FIGURE 3.11: SCHEMATIC DIAGRAM OF INJECTION MOULDING MACHINE WITH BASIC COMPONENTS FIGURE 3.12: THE SCHEMATIC OF TYPES OF HEAT FLOW IN AN INJECTION MOULD, BASED ON [7] FIGURE 3.13: TYPICAL INJECTION MOULDING HEAT TRANSFER PROCESS 3-D DOMAIN FIGURE 3.14: UNIT VOLUME (Ω) OF 3-D DIMENSIONAL CONDUCTION HEAT TRANSFER PROCESS IN CARTESIAN CO-ORDINATES FIGURE 3.15: CONVECTIVE HEAT TRANSFER PROCESS IN OPEN CHANNEL FIGURE 3.16: SCHEMATIC DIAGRAM OF CONVECTIVE HEAT TRANSFER PROCESS IN INJECTION MOULDING FIGURE 3.17: CROSS SECTIONAL CAVITY MODEL, SHOWING THE SURFACES THOROUGH HEAT TRANSFER HAPPENS DURING INJECTION MOULDING FIGURE 3.18: DIFFERENT TEMPERATURES POINTS FOR DIFFERENT SURFACES AND COOLING CHANNELS FOR CONVENTIONAL COOLING CHANNELS FIGURE 3.19:: (A) DIFFERENT INPUT AND OUT BOUNDARY CONDITIONS AND PROCESS PARAMETERS SETTING IN AMI SIMULATION FOR CONVENTIONAL COOLING. (B) TEMPERATURE VS TIME PLOT FOR DIFFERENT SURFACES OF PLASTIC PART/CAVITY WALL FOR A TYPICAL NODE POINT IN EACH SURFACES xii

15 FIGURE 3.20: TEMPERATURE RECORDING OF COOLING CHANNEL AFTER ONE CYCLE IN AMI SIMULATION (A) COOLING WATER TEMPERATURE (B) COOLING CHANNEL WALL TEMPERATURE FIGURE 3.21: PROCESS PARAMETERS AND BOUNDARY CONDITIONS OF TRANSIENT THERMAL FEA ANALYSIS FOR CONVENTIONAL COOLING CHANNEL IN ANSYS WORKBENCH SIMULATION SOFTWARE FIGURE 3.22: HALF OF CAVITY MOULD, SHOWING TETRAHEDRAL MESHING WITH RECTANGULAR CONFORMAL COOLING CHANNELS, NUMBER OF NODES AND ELEMENTS ARE ALSO SHOWN IN LEFT SIDE FIGURE 3.23: TEMPERATURE DISTRIBUTION IN THE CAVITY MOULD SECTIONAL VIEW, AFTER ONE CYCLE FOR (A) CONVENTIONAL, AND (B) CIRCULAR, (C) SQUARE, (D) RECTANGULAR, (E) HEXAGONAL AND (F) ELLIPTICAL, CROSS SECTIONAL CONFORMAL COOLING CHANNELS FIGURE 3.24: TEMPERATURE VS TIME PLOT OR COOLING CURVE OF THE CAVITY MOULD WITH DIFFERENT COOLING CHANNELS FIGURE 4.1: A 3-D RECTANGULAR DOMAIN IN INJECTION MOULD FIGURE 4.2: A TETRAHEDRAL ELEMENTS SHOWING 4 NODAL COORDINATES FIGURE 4.3: CROSS SECTIONAL MOULD, SHOWING DIFFERENT SURFACES USED FOR APPLYING FIGURE 4.4: VARIABLE CLAMPING FORCES ON THE TOP SURFACE OF MOULD ASSEMBLY FOR ENTIRE CYCLE FROM AMI ANALYSIS FIGURE 4.5: VARIABLE INJECTION PRESSURE ON DIFFERENT SURFACES OF PLASTIC PART THAT IS IN CONTACT WITH CAVITY SURFACES, FOR ENTIRE CYCLE FROM AMI ANALYSIS FIGURE 4.6: TEMPERATURE DISTRIBUTION AFTER ONE CYCLE (35 SECOND) IN THE CAVITY MOULD SECTIONAL VIEW WITH CONFORMAL COOLING CHANNELS (CIRCULAR DIAMETER, PITCH 2.5D H=30 MM) MADE OF (A) STAVAX SUPREME AND (B) ALUMINIUM MATERIALS FIGURE 5.1: SCHEMATIC CROSS SECTIONAL DIAGRAM OF THE MOST FREQUENTLY USED COOLING CHANNELS [1] FIGURE 5.2: SCHEMATIC OF ONE DIMENSIONAL HEAT FLOW THROUGH A COMPOSITE WALL OF STEEL AND COPPER PLATES FIGURE 5.3: CAD MODEL OF (A) PLASTIC CANISTER, AND (B) CORE AND CAVITY IN MOULD FIGURE 5.4: ASSEMBLY MOULD WITH CONVECTIONAL STRAIGHT COOLING CHANNELS xiii

16 FIGURE 5.5: BI-METALLIC STRAIGHT COOLING CHANNEL (BSCC) WITH COPPER TUBE INSERT (CTI) IN CAVITY, AND CTI IN CORE BUBBLER CHANNELS FIGURE 5.6: SECTIONAL TOP VIEW OF CAVITY MOULD, SHOWING THE ORIENTATION OF BSCC IN THE MOULD FIGURE 5.7:BI-METALLIC CONFORMAL COOLING CHANNEL (BCCC) WITH COPPER TUBE INSERT (CTI) IN CAVITY AND CTI IN CORE BUBBLER CHANNELS FIGURE 5.8: SECTIONAL TOP VIEW OF CAVITY MOULD, SHOWING THE ORIENTATION OF BCCC IN THE MOULD FIGURE 5.9: CROSS SECTIONAL ASSEMBLY MOULD, SHOWING DIFFERENT INTERFACE SURFACES FIGURE 5.10: DIFFERENT TEMPERATURES NOTATION FOR DIFFERENT SURFACES AND COOLING CHANNELS IN A TYPICAL MOULD FIGURE 5.11: DIFFERENT INPUT AND OUT BOUNDARY CONDITIONS AND PROCESS PARAMETERS SETTING AMI SIMULATION FIGURE 5.12:TEMPERATURE VS TIME PLOT FOR DIFFERENT SURFACES OF PLASTIC PART WHICH IS IN CONTACT WITH CAVITY AND CORE WALL FOR A TYPICAL NODE POINT IN EACH SURFACES, OBTAINED FROM AMI SIMULATION FIGURE 5.13: TEMPERATURE RECORDING OF COOLING CHANNEL IN AMI SIMULATION (A) COOLING CHANNEL WALL TEMPERATURE (B) COOLING WATER TEMPERATURE FIGURE 5.14:VARIABLE CLAMPING FORCES ON MOULD ASSEMBLY FOR ENTIRE CYCLE FROM AMI ANALYSIS FIGURE 5.15: VARIABLE INJECTION PRESSURE ON MOULD DIFFERENT SURFACE OF PLASTIC FIGURE 5.16::(A) TEMPERATURE DISTRIBUTION AFTER ONE CYCLE IN THE MOULD, AND (B) EQUIVALENT STRESS DISTRIBUTION AT 4.45TH SECOND OF CYCLE TIME FOR CSCC MOULD FIGURE 5.17: FIGURE 5.17: TEMPERATURE DISTRIBUTION ON MOULD ASSEMBLY AFTER ONE CYCLE (20 SECOND) FIGURE 5.18: TEMPERATURE DISTRIBUTION ON MOULD ASSEMBLY AFTER ONE CYCLE (20 SECOND) (A) BCCC 2MM CTI, AND (B) BSCC 2MM CTI FIGURE 5.19: EQUIVALENT STRESS DISTRIBUTION AT 4.45 SECOND OF CYCLE FOR (A) BSCC 3MM CTI AND (B) BCCC 3MM CTI MOULDS FIGURE 5.20: EQUIVALENT STRESS DISTRIBUTION AT 4.45 SECOND OF CYCLE FOR (A) BSCC 2 MM CTI AND (B) BCCC 2 MM CTI MOULDS FIGURE 5.21: MAXIMUM EQUIVALENT STRESS FOR DIFFERENT COOLING CHANNELS FOR ENTIRE CYCLE xiv

17 FIGURE 5.22: COMPARATIVE COOLING CURVE FOR CSCC AND BSCC 2MM CTI FIGURE 5.23: (A) CORE AND CAVITY MOULD, (B) COPPER TUBE IS BEING INSERTED, AND (C) SHOWS THE INJECTION MOULDED PARTS IN PP AND ABS PRODUCED DURING THE EXPERIMENT FIGURE 5.24: EXPERIMENTAL SET UP WITH MINI MOULDER FIGURE 5.25: COMPARATIVE COOLING CURVE FOR CSCC AND BCCC CTI MOULDS, SHOWING TEMPERATURE AT THE BOTTOM AND TOP SURFACE OF THE POLYPROPYLENE (PP) PLASTIC PART FIGURE 5.26: COMPARATIVE COOLING CURVE FOR CSCC AND BCCC CTI MOULDS, SHOWING TEMPERATURE AT THE BOTTOM AND TOP SURFACE OF THE ACRYLONITRILE BUTADIENE STYRENE (ABS) PLASTIC PART FIGURE 6.1: SCHEMATIC OF COMPONENTS IN DMD SYSTEM, BASED ON [86] FIGURE 6.2: SCHEMATIC OF FEEDBACK CONTROL SYSTEM IN DMD PROCESS, BASED ON [86] FIGURE 6.3: CAD MODEL OF CORE AND CAVITY MOULDS DESIGNED BY PRO/ENGINEER FIGURE 6.4: CAD MODEL OF (A) CAVITY MOULD WITH CSCC, AND (B) CAVITY MOULD WITH SSCCC, DESIGNED BY PRO/ENGINEER SOFTWARE FIGURE 6.5: DIMENSIONS OF DMD LAYERS FIGURE 6.6: BUILDING PATTERN OF DMD LAYERS FIGURE 6.7: POM DMD 505 MACHINE AT SWINBURNE UNIVERSITY OF TECHNOLOGY FIGURE 6.8:(A) FABRICATION OF CAVITY MOULD WITH SSCCC BY DMD PROCESS, AND (B) COMPLETE CAVITY MOULD WITH SSCCC BY DMD PROCESS FIGURE 6.9: (A) CROSS SECTION OF CAVITY MOULD SHOWING SSCCC AND INTERFACE AREA BETWEEN H13 DMD FABRICATION AND MILD STEEL (B) MOUNTING OF INTERFACE AREA FOR OPTICAL MICROSCOPIC IMAGE TESTING FIGURE 6.10: OPTICAL MICROSCOPIC IMAGE OF INTERFACE OF H13 DEPOSITION FIGURE 6.11:(A) EXPERIMENTAL MOULDING SETUP WITH MINI-INJECTION MOLDER, (B) TOP OF CORE SHOWING THE POSITION OF THERMOCOUPLE WITH RED CIRCLE, (C) BOTTOM OF CAVITY SHOWING POSITION OF THERMOCOUPLE WITH RED CIRCLE FIGURE 6.12: TEMPERATURE READING FOR TOP AND BOTTOM INTERFACE OF PP PLASTIC FIGURE 6.13: TEMPERATURE READING FOR TOP AND BOTTOM INTERFACE OF ABS PLASTIC xv

18 FIGURE 6.14: PROCESS PARAMETER FOR AMI SIMULATION FOR SSCCC MOULD FIGURE 6.15: TEMPERATURE DISTRIBUTION IN THE CAVITY SECTIONED MOULD WITH SSCCC FROM THERMAL FEA SIMULATION FIGURE 6.16: TEMPERATURE DISTRIBUTION IN THE CAVITY SECTIONED MOULD WITH CSCC FROM THERMAL FEA SIMULATION FIGURE 7.1: MOULD TEMPERATURE RESULTS OF THE ORIGINAL (LEFT) AND MODIFIED (RIGHT) COOLING SYSTEM [159] FIGURE 7.2: COMPARATIVE WARPAGE DEFLECTION RESULTS OF THE ORIGINAL (LEFT) AND MODIFIED (RIGHT) COOLING SYSTEM DESIGN [159] FIGURE 7.3: VARIATIONS IN CRYSTALLIZATION THROUGH CROSS SECTION OF THE PLASTIC WALL AND MOULD IN INJECTION MOULDING PROCESS [129] FIGURE 7.4: CONFORMAL COOLING CHANNELS DESIGN IN AMI SIMULATION (A) CIRCULAR CROSS SECTION, (B) SQUARE CROSS SECTION FIGURE 7.5: FLOW CHART OF FLOW SIMULATION WITH AMI FIGURE 7.6: TIME REQUIRED TO REACH EJECTION TEMPERATURE OR COOLING TIME OF THE PLASTIC PART FOR (A) CIRCULAR, AND (B) SQUARE, CROSS SECTIONAL CONFORMAL COOLING CHANNELS FIGURE 7.7: TIME REQUIRED TO REACH EJECTION TEMPERATURE OR COOLING TIME OF THE PLASTIC PART USING CONVENTIONAL COOLING FIGURE 7.8: WARPAGE DEFLECTION OF THE PLASTIC PART IN ALL DIRECTION FOR (A) CIRCULAR AND (B) SQUARE CROSS SECTIONAL CONFORMAL COOLING CHANNELS FIGURE 7.9: (A) WARPAGE DEFLECTION OF THE PLASTIC PART IN ALL DIRECTION, AND (B) PERCENTAGE OF VOLUMETRIC SHRINKAGE WITH THE ORIGINAL VOLUME OF THE PLASTIC PART, USING CONVENTIONAL COOLING FIGURE 7.10: PERCENTAGE OF VOLUMETRIC SHRINKAGE WITH THE ORIGINAL VOLUME OF THE PLASTIC PART, USING (A) CIRCULAR AND (B) SQUARE CROSS SECTIONAL CONFORMAL COOLING CHANNELS FIGURE 7.11: (A) MEASURING THE HARDNESS WITH SHORE D HARDNESS TESTING MACHINE, (B) POINTS AT WHICH HARDNESS HAS BEEN MEASURED, TOP SURFACE OF ABS (TOP), BOTTOM SURFACE OF PP (BOTTOM) FIGURE 7.12: COMPARATIVE HARDNESS PLOT IN DIFFERENT POINTS OF PARTS THAT HAVE BEEN PRODUCED BY SCC AND CCC FOR ABS FIGURE 7.13: COMPARATIVE HARDNESS PLOT IN DIFFERENT POINTS OF PARTS THAT HAVE BEEN PRODUCED BY SCC AND CCC FOR PP xvi

19 LIST OF TABLES TABLE 2.1: COMMERCIAL RAPID PROTOTYPING (RP) TECHNIQUES TABLE 2.2: DIRECT METAL RAPID MANUFACTURING (RM) PROCESSES TABLE 2.3: MECHANICAL PROPERTIES OF VARIOUS METAL FILLED EPOXIES [112] TABLE 3.1: DATA FOR CALCULATION OF HYDRAULIC DIAMETER D H TABLE 3.2: TEMPERATURE VALUES OF DIFFERENT SURFACES AND COOLING CHANNELS RECORDED TABLE 3.3: CONDUCTION HEAT FLUX VALUES USED AS BOUNDARY CONDITION FOR DIFFERENT TABLE 3.4: CONVECTIVE HEAT FLUX VALUES USED AS BOUNDARY CONDITION FOR DIFFERENT TABLE 3.5: STAVAX SUPREME PROPERTIES TABLE 3.6: NUMBER OF NODES AND ELEMENTS FOR TETRAHEDRAL MESHING FOR ALL COOLING TABLE 3.7: TEMPERATURE RECODING IN DIFFERENT COOLING CHANNEL MOULD FOR ENTIRE CYCLE TABLE 3.8: COMPARISON OF SURFACE AREAS FOR DIFFERENT CROSS SECTIONAL COOLING CHANNELS TABLE 3.9: COMPARATIVE COOLING TIME DATA FOR THEORETICAL AND SIMULATION RESULT TABLE 4.1:PROPERTIES OF STAVAX SUPREME AND ALUMINIUM [157] TABLE 4.2: DIMENSIONS OF CONFORMAL COOLING CHANNELS TABLE 4.3: VALUES OF REYNOLDS NUMBER AND CONVECTION COEFFICIENT FOR DIFFERENT TABLE 4.4: VALUES OF CLAMPING FORCES AND INJECTION PRESSURE USED AS BOUNDARY TABLE 4.5: VALUES OF MAXIMUM TEMPERATURE IN THE CAVITY MOULDS AFTER ONE CYCLE FOR TABLE 4.6: VALUES OF MAXIMUM EQUIVALENT STRESS OR VON-MISES STRESS IN THE CAVITY MOULDS DURING THE MOULDING CYCLE OF ALL TYPES OF COOLING CHANNELS TABLE 4.7: VALUES OF CONSTANT F FOR DIFFERENT S UT TABLE 4.8: PREDICTED LIFE CYCLE OF THE STAVAX SUPREME MOULDS WITH DIFFERENT COOLING xvii

20 TABLE 5.1: DIMENSIONS AND WEIGHT OF THE TEST PLASTIC PART TABLE 5.2: DIFFERENT COOLING CHANNELS AND THEIR ABBREVIATIONS TABLE 5.3: PROPERTIES OF STAVAX SUPREME AND COPPER ALLOY [157] TABLE 5.4: NUMBER OF NODES AND ELEMENTS FOR DIFFERENT COOLING CHANNEL MOULD TABLE 5.5: TEMPERATURE VALUES OF DIFFERENT SURFACES AND COOLING CHANNELS RECORDED FROM AMI FLOW SIMULATION FOR ALL COOLING CHANNELS FOR DIFFERENT TIMES TABLE 5.6: HEAT FLUX VALUES USED AS BOUNDARY CONDITION FOR DIFFERENT COOLING CHANNELS TABLE 5.7: VALUES OF CLAMPING FORCES AND INJECTION PRESSURE USED AS BOUNDARY CONDITIONS TABLE 5.8: VALUES OF MAXIMUM EQUIVALENT STRESSES (S ES) AND LIFE CYCLE OF THE VALUES OF MOULDS WITH DIFFERENT COOLING CHANNELS TABLE 5.9: COMPARATIVE COOLING TIME DATA FOR CSCC AND BCCC 2MM CTI TABLE 6.1: COMPOSITION AND PROPERTIES OF H13 AND MILD STEEL [159] TABLE 6.2: COMPARISON OF EXPERIMENTAL AND THERMAL FEA RESULT OF TEMPERATURE DISTRIBUTION AT THE MOULD SURFACE FOR CSCC AN SSCCC COOLING CHANNEL FOR PP AND ABS PLASTICS TABLE 7.1: COMPARATIVE FLOW SIMULATION RESULTS FOR DIFFERENT COOLING CHANNELS xviii

21 Introduction 1 Chapter Overview Injection moulding is effectively one of the most widely used plastic manufacturing processes available in plastic manufacturing industries. With the broader and broader use of plastics parts in varieties of products from consumer products to machineries, the injection moulding process has been renowned as an important manufacturing process. Plastic materials are commonly used in large variety of products of our daily life. The most important reason for this is because of the properties of the plastic material. Some of these properties are lightness, resistance to corrosion, and ease to give shape. Plastic material s physical and chemical properties can be changed as desired. Plastic material can be used in packaging, aerospace and aviation, building and construction, automotive, electronic goods etc. Injection moulding is highly efficient means of producing plastic parts. The injection moulded parts certainly have excellent dimensional tolerance and require very little amount of finishing and assembly operations. Injection moulding is used for producing parts from both thermoplastic and thermosetting plastic materials. Despite its tremendous advantages, injection moulding process also experiences some problems in mould design and consequently problems in plastic parts. Some of the main problems of injection moulded plastic parts are brittleness, warpage, uneven shrinkage and cracking [1, 2]. The main reasons of these problems are non uniform and inappropriate cooling system design in the mould, which results in too much or too less cooling phases in injection moulding cycle. This research work is carried out to address these problems by developing conformal cooling channel design in the mould of injection moulding process. 1

22 1.2 Mould Cycle Time and Importance of Cooling System The mould process cycle of injection moulding consists of mould closing, injection/holding, cooling, mould opening and product removal stages. Figure 1.1 (adopted from [3]) shows cyclic components of a typical mould cycle of 36 seconds. The sequence is repeated for each process cycle. This is why an injection moulding process is suited for mass production. The mould cycle time depends mainly on mould cooling design, mould material selection, and the plastic material moulded while other factors include the machine speed setting and the method of ejection from the mould. In practical applications, the mould cooling stage takes a substantial part, generally from 50-80% of the moulding cycle time [3-6]. Figure 1.1: Typical moulding cycle of injection moulding process (in second) based on [3]. Therefore, the mould cooling system is the most important and promising section for mould designers to minimize the cycle time. Besides affecting productivity, the mould cooling process also plays a significant role in 2

23 determining the part quality. A good cooling system must not only be efficient in reducing the cooling time, but must also be able to achieve uniform temperature distribution in the mould area to minimize undesirable defects. The longest time required in the moulding cycle is the solidification of the molten component, which occurs during the cooling time. So the reduction of cooling time will lead to considerable amount of shortening of the moulding cycle, which will increase production volume. But this is not easy task to accomplish, considering the complexity of mould design and the requirements of the injection moulding process. The cooling system is an essential mould feature, which requires special attention in mould design. The challenges of the cooling problem are due to the complex geometry introduced by the cooling channel layout and the significant differences of material properties, between the part and the mould. Many factors [1, 7], which affect the cooling of a mould, are as follows: o Thermal properties and geometry of the plastic part: density, specific heat, thermal conductivity, thickness and surface area of the part; o Thermal properties and geometry of the mould: density, specific heat, thermal conductivity of the mould material as well as the size of the mould; o Thermal and rheological properties of the coolant: density, thermal conductivity, specific heat and viscosity; o Coolant flow: Reynolds (Rn) and Prandlt (Pn) number; o Types, dimensions, locations and arrangements of the cooling channels; o Cooling operation conditions: melt injection temperature, ejection temperature, coolant temperatures and its variation. The cost effectiveness of the injection moulding process is dependent on the time spent in the moulding cycle. Amongst the phases in moulding cycle that have been shown in Figure 1.1, cooling phase is the most significant as it largely determines the rate at which the parts are produced. As in most modern 3

24 industries, time and cost are strongly linked; the longer the time is to produce parts, the more are the costs. A reduction in the time spent on cooling the part before it is ejected would drastically increase the production rate and hence would reduce the costs [8]. Both quality and productivity are two of the most important issues in mould cooling design. The physical qualities and the appearance of the moulded part depend largely on the rate of cooling. A part becomes brittle and shows lack of glossy appearance when cooled too quickly or cooled too excessively at low temperature, whereas, it shows unwanted crystallization when cooled insufficiently or too slowly. Defects, including hot spots, uneven shrinkage and warpage, would often result in expensive trouble-shooting and modification of the existing tooling. Residual stresses are often the main cause of uneven part shrinkage and warpage [9]. In chapter seven, more details of the quality of plastic that are affected by cooling channel geometry in injection moulding, have been discussed. Thin-wall parts are often very sensitive to uneven shrinkage. Ignoring the flow-induced residual stress, a major solution to reduce the thermal-induced residual stress is to provide uniform cooling of all the surfaces of the part [10]. The filling time and moulding conditions such as the injection pressure are also affected by the mould temperature gradients which, in turn, depend on the cooling effect. A designer needs to select suitable mould material and to optimise the cooling system so that the polymer is cooled efficiently and evenly inside the mould. It can be difficult to provide cooling to some smaller areas or near large accumulations of the polymer. However, every second saved from the cooling time contributes to a proportional increase in productivity and is well worth the additional design effort and higher manufacturing costs. For example, if a mould runs in a 36-second cycle; where cooling time is 18 seconds, with better cooling design, even though at a higher mould-making cost, if it is possible to reduce cooling time only by 6 seconds, while other cycle remain same, total moulding cycle will reduced to 30 seconds. This would amount to a 20% improvement in output, and in a long-running job of several 4

25 hundred thousand pieces, the financial savings could be impressive. If the life cycle for a 36 second cycle time mould is 1 million cycle, then with a single cavity mould, 1 million parts have been produced. But with a 30 second cycletime, an additional 200,000 parts can be produced. Therefore, for high production moulds, it is imperative that the cycle time be reduced to a minimum even if initial moulding cost is bit higher. 1.3 Research Objective Injection moulding is a mass production cyclic process, which consists of mould clamping, injection/holding, cooling, mould opening and product removal stages. Among these, cooling stage is the most time consuming and significant phase. For many years, researchers are trying to improve the cooling system performance by developing different cooling techniques. However, very little research has been directed on further development of cooling system of injection moulding using conformal cooling design and rapid tooling manufacturing technology. This research work is specially directed to investigate the application of conformal cooling channels with different crosssectional shapes, diameters and position from the cavity surface to determine which conformal cooling configuration gives the best result aiming for better production rate and quality product. Conformal cooling channels with novel bimetallic cooling channels are also investigated for different types of industrial plastic parts with flat and curve surfaces. Flow simulation (cool+flow+pack+warp analysis) and finite element analysis have been performed for developed conformal cooling systems with Autodesk Moldflow Insight (AMI) and ANSYS workbench simulation software respectively. Moulds have been fabricated with developed conformal cooling channels by Rapid tooling technique of laser Direct Metal Deposition (DMD) as well as traditional Computer Numerical Control (CNC) machining process. Comparative experimental verifications have been done with conformal and conventional cooling for a test plastic part of two different types of plastics, 5

26 polypropylene (PP) and Acrylonitrile butadiene styrene (ABS) with miniinjection moulding machine. Experimental verification also has been done with Shore D hardness testing machine to test the hardness/crystallization effect of test plastic parts, which have been produced with conformal and conventional cooling channel moulds. According to the demand of the moulding industry, the following outcomes of the research have been fulfilled which will facilitate the injection moulding industries to achieve better and quality production of plastic parts. 1. Development of conformal cooling channels with optimum crosssectional area, diameter and distance from the cavity surface. 2. Development of novel bi-metallic conformal cooling channels with high thermal conductive cooper tube for injection moulds. 3. Feasibility of fabrication of conformal cooling channels with rapid manufacturing technique of laser Direct Metal Deposition process. 4. Demonstration of decreased cycle time, increased life cycle of moulds and better quality of parts with conformal cooling channels in comparison with conventional straight cooling channels using numerical and experimental techniques. 1.4 Available Cooling Method and Importance of Research In injection moulding process, it is very difficult to achieve efficient and balanced cooling. It is also difficult to generate an optimal cooling design for a plastic part automatically. Because of the intricacy, the mould cooling system design in the mould industry is still largely based on established principles, the designer's experience and trial-and-error techniques. The design may also face the limitations of the capabilities and standards of the markets as well as the potential of the end user. The cooling design unfortunately, has to be placed at the end of design cycle. Mould designers have other elementary issues to 6

27 consider such as gate location, ejector pin and runner arrangement for multicavity mould before they can design and place cooling channels. So ultimately, after placing other elements, little space and flexibility are available to design efficient cooling channels [11, 12]. Some designer places as many cooling channels as they could in the mould to get fast and sufficient cooling, and have the mould tested to decide which of them would work well. When no computational analysis and simulation tool is applied, it is normally difficult for available mould cooling methods to achieve uniform and balanced cooling. Cycle time, which is vital factor for mould design, is normally estimated based on a designer s experience and a combination of trial-and-error approaches. As a result, many moulds are operated at a much longer cycle time than necessary which leads to low productivity and quality. As the part geometry becomes more complex, an experience-based approach becomes less feasible. Because the experience-based methods cannot design conformal cooling channels for complex geometrical shapes of plastic parts, whereas, proposed conformal cooling channels with proper size, shapes and locations could overcome these limitations, which will be thoroughly discussed in this thesis. Available conventional cooling methods being used are straight drilled cooling channels (SDCC), baffles, bubblers, thermal pin, helical channels, heat rods, heat pipes and milled grooves [13]. With the substantial advances in Rapid Prototyping (RP) and Rapid Tooling (RT) techniques, some researchers have turned these techniques into mouldmaking activities [14-16]. The advantages of RP and RT techniques are the ability to shorten the lead-time and the ability to fabricate complex internal mould cooling systems. The conformal cooling channels, which can be manufactured using the RP and RT, were reported to have achieved better temperature control for shortening the cycle time and improve the product quality. However, proper cooling channel design and geometrical location in the mould with respect to different moulding material, mould robustness and mould-making costs are some of the issues to be overcome using these technologies. In this research work, a comprehensive study has been done to 7

28 overcome some of these shortcomings of developing conformal cooling channels. Different cross-sectional shapes of conformal cooling channel, their position from mould surfaces and bi-metallic conformal cooling channel have been thoroughly investigated with complete simulation and experimental verification. 1.5 CAD/CAM/CAE and Rapid Tooling in Injection Moulding Computer-aided design, computer-aided manufacturing and computer aided engineering (CAD/CAM/CAE) technologies, which emerged during the past decades, have helped to increase engineering productivity significantly. Along with rapid tooling technology which is still developing, CAD/CAM/CAE have provided the total integration of design, analysis and manufacturing functions and have had a large impact on the engineering practices. While CAD/CAM/CAE has found a wide range of engineering applications, its applications to injection moulding design and manufacturing have been limited. This is because of complexity of injection moulding system and lack of interest. Mould making used to be mostly based on experiences and depends mainly on skilled craftsman, typically trained under apprenticeship scheme and having acquired their expertise through years of practice. The supply of trained manpower, however, is rapidly thinning as the younger generation is unwilling to go through the long training periods and would prefer other jobs. But since the introduction of CAD/CAM/CAE, this situation has changed since early Eighties. Most of the injection moulding companies use 3-D CAD software to make mould to increase their productivity. However, the lack of optimized mould design and traditional manufacturing process still could not solve the critical mould designing steps, specially the ideal cooling channel design and cost effective manufacturing process [17]. Integration of CAD/CAM/CAE and rapid tooling technology could be one of the solutions of these difficulties. In recent years, to be competitive in the international market, where new products have to be presented on shorter time 8

29 scales, the use of CAD/CAM/CAE in the mold industry is already a basic requirement. To work profitably for customers in many areas, it is necessary to install a complete network that directly links computer-generated design data to the computer numerical controlled (CNC) production metal-cutting machinery. The CAD system should have two- and three-dimensional capabilities to advance all relevant design activities. A powerful computer system enables one to react quickly to clients requirements, especially in terms of changes in design and savings in time and money. Specific software automatically produces the data for CNC programming and forms the connection between design and production departments. The CAM portion of a computer-automated system converts design data into numerical control data used by CNC machines, which will mill or drill the material onto the desired parts by material removal process [18]. And finally, using the rapid tooling technologies like Direct Metal Deposition (DMD) and Direct Metal Laser Sintering (DMLS) [14, 19, 20], mould design with any complex shape especially with conformal cooling channels can be manufactured. In recent years, RT processes [20, 21] have found widespread use in speeding up tooling production. These processes not only greatly reduce the manufacturing costs and lead time required for tool production but also give the flexibility for complex design for mould shape and cooling channels. Figure 1.2 (adopted from [19]) illustrates the various steps and time saving between traditional tooling production and contemporary RT fabrication. This research work also describes (in chapter 6) the feasibility of fabricating conformal cooling channel using the rapid tooling technology of DMD (direct metal deposition). DMD technology [22] provides an innovative CAD-driven metal deposition technique that integrates CAD, CAM, industrial lasers, and powder metallurgy to create a unique near net shape deposition process. The high energy generated within a 5kW CO2 laser, when combined with the injection of complex alloy powders into the laser beam, as shown in figure 1.3 (adopted from [22]) creates the basis for the deposition of the molten 9

30 metal pool onto a free standing substrate forming a metallurgical bond, while the DMD closed loop sensors monitor the melt pool to maintain dimensional integrity. Figure 1.3: Difference in time and steps between traditional and concurrent rapid tooling manufacturing, based on [19]. Figure 1.2: The working elements of the Direct Metal Deposition (DMD) system, based on [22]. 10

31 The benefit of this technology is its ability to deposit successive layers of metal onto the surfaces of complex components with the minimum of heat transfer, as well as the ability to mix metals to create unique alloys that are not commercially available. The low heat input property of DMD is the most critical attribute of this technology that makes it attractive for production of conformal cooling channels and tooling surfaces. The direct-metal rapid tooling from high quality steel is not commonly known in the injection molding industry today. Different RT technologies enable the production of inserts for injection molding tools, and a distinction between direct and in-direct RT is made where indirect methods usually involve more sequences than direct methods. Typically, an indirect method uses a master pattern to produce a mold or die, while the direct methods build the actual tooling inserts directly in the prototyping machine. Using the DMD process, research work proposes a direct rapid tooling system to make mould with H-13 steel powder incorporating conformal cooling channels. With the help of this DMD system, square shape conformal cooling channels have been fabricated inside the mould by direct RT technique, as described in this thesis. 1.6 Organization of the Thesis This thesis consists of eight chapters. Chapter one is the introduction, it gives brief introduction of research objective, importance of cooling channel in injection moulding process, limitations of present cooling channel system, hence importance of research in relation to CAD/CAM/CAE and rapid tooling technology. Chapter two presents a comprehensive literature review of related present and past research on rapid direct and indirect tooling technologies, mathematical solution, heat transfer and thermal stress analyses, mould cooling design and optimization. Chapter three describes design and analysis of conformal and conformal cooling channel. Conformal cooling channels have been designed for optimum cross-section, diameter and distance from cavity surface. Design and analysis 11

32 have been carried out for different materials and plastic parts with round and flat surface geometry. Fundamentals of cooling channel design and heat transfer modeling also have been described in this chapter. Cooling channel design includes guideline for conventional and conformal cooling channels and factors affecting cooling of injection moulding. Heat transfer modeling section explains basic heat transfer process in injection moulding and their mathematical formulation, cooling time calculation and heat transfer process through composite materials. Chapter four studies the FEM theories and procedures in the applications of heat transfer analysis and the derivation of FEM equations in transient thermal stress analysis. Comparative thermal-structural FEA analysis has been done for conformal and conventional cooling channel aiming for better results for conformal cooling channels for cycle time and fatigue life of mould. Chapter five talks about bi-metallic conformal cooling channel with high thermal conductive copper tube insert in straight cooling channels. Design and FEA analysis have been done for bi-metallic conformal and semi conformal cooling channel. Experimental verification also has been described in this chapter for bimetallic conformal cooling channel. Chapter six describe about experimental verification of conformal cooling channel, which has been manufactured with rapid manufacturing DMD technology and conventional CNC machined straight cooling channel. Experimental verification has been done for a sample part with two different plastics PP and ABS in mini-injection moulding machine. Chapter seven explains the properties of plastics that have been produced by injection moulding. Comparative Fill-cool-warp analysis has been done with AMI flow simulation software to check plastic properties with conformal and conventional cooling channel. This chapter also describe about experimental verification with Shore D hardness testing machine to check crystallization effect on plastic that has been made with conformal and conventional cooling channels. Chapter eight describes conclusions and gives some recommendations for future works. 12

33 Literature Review 2 Chapter Background For mass scale production of plastic parts, the injection moulding process (IMP) has been extensively investigated and modified during the last century since the invention of this process in There have been many research works accomplished, especially in the last few decades to make this process more productive and economical. Applications of the state of art technologies in IMP, by incorporating different cooling channel configurations in the mould and it s manufacturability using the Rapid tooling (RT) and Rapid Manufacturing (RM) technologies have also attracted a great deal of interest. But little success seems to have been made in implementing conformal cooling channels (CCC) efficiently in injection moulds. The main obstacle has been the manufacturability of such channels in the mould due to high costs and technological limitations. It has been mentioned in chapter one that this research work has been dedicated to investigate the appropriate CCC design based on its crosssectional shape, distance from cavity surface and suitable material selection for such design. Validations for such cooling channel design have been achieved through thermal-structural Finite Element Analysis (FEA) of the moulds and experimental verification by fabricating a mould containing such CCC with Rapid Tooling (RT) technology of laser Direct Metal Deposition (DMD), and producing a sample plastic part with the mould. This chapter reviews the literatures in the following categories highlighting outstanding gaps appropriate for outcome of this research work. 13

34 Works on numerical and analytical solutions for injection moulding cooling with different cooling systems. Works on thermal stresses in injection moulded plastic part. Works on mould fabrication by rapid tooling and rapid manufacturing. Work on conformal cooling by rapid tooling. 2.2 Works on Numerical and Analytical Solutions for Injection Moulding Cooling with Different Cooling Systems. Presently there are two types of research on numerical simulation in mould cooling design and analysis. Firstly, work is directed to the application of various computational methods in cooling analysis and simulation to help optimize the cooling channel design based on routine trial and error method of mould-making and mould cooling. Secondly, the work is directed to find new methods of mould making or cooling system design to eradicate non-uniform cooling of conventional system using conformal cooling approach. Use of CAD/CAM/CAE for cooling analysis can provide critical information for cooling design and processing parameters setting in injection moulding to ensure successful moulding process. The CAE methods predict the transient temperature profile, so as to estimate the cooling efficiency and part quality before the actual mould fabrication. It is a cost-effective alternative to the conventional trial-and-error methods. Consequently, cooling optimisation can be performed based on the CAE analysis results. The analysis of the mould cooling system has been studied extensively over the past many years. In 1970s, the studies focused on the polymer melt temperature variations during solidification. In 1980s, the studies focused on the influence of the cooling system on the mould impression surfaces (MIS) temperature. Although a complete analysis for the transient temperature variations of the mould and the polymer melt simultaneously is possible in principle, the computation cost 14

35 was too high in earlier days. It was particularly difficult to implement it during the actual design process, especially when the optimisation was required and many design parameters were involved. In the 1990s, mould cooling analyses focused on the development of numerical methodology, which was computationally efficient and, at the same time, accurate enough for design purposes. Since then there have been significant number of research work published on cooling channels design and analysis with different FEA techniques and CAD/CAM/CAE software, either on thermal or structural aspects or both, to evaluate different cooling channel designs. In this section of the chapter, different computational and simulation methods and related research works are systematically reviewed with advantages and drawbacks Modeling and Boundary Conditions for Cooling Analysis Normally, the cooling analyses are carried out to calculate both temperature distributions in the mould and the cooling time after the design parameters relating to coolant operation conditions as well as the arrangement and dimensions of the cooling channels have been decided. Due to the complexity in the coupling of many design parameters, assumptions are necessary to simplify the problem. In the preliminary cooling system design, an overall heat balance between the plastic melt and the coolant is usually evaluated by the shape factor approach or empirical calculations. The temperatures of mould surfaces and polymer melt have been assumed to be constant during the filling stage in most of the earlier studies. The cooling effect resulting from the cooling channel configuration is usually evaluated manually in an averaging scheme. The polymer part is normally assumed very thin such that the heat transfer within the melt region can be treated as transient, 1-dimensional heat conduction in the predominant thickness direction of the part [23-30]. The transient variation in mould cavity temperature within a steady cycle is then evaluated, using 2-dimentional boundary element method (BEM) using the 15

36 instantaneous heat flux introduced by the polymer melt at the MIS during the entire injection stage. It is then imposed to the cavity temperature calculation based on a cycle-averaged analysis. It is further assumed that heat exchange between the cooling channel surfaces and the coolant is steady [23, 25]. The exterior surfaces of the mould were treated as adiabatic for simplicity because the heat loss through the exterior surfaces is less than 5% of overall loss in most injection moulding applications. Rezayat et al. [31] developed a special method employing a periodic approach and using special modifications for MIS for circular cooling channels in order to further improve computational efficiency. Chen et al. [32] reported a comparison of mould cooling analysis based on different methods, such as Shape Factor Approach (SFA), Finite Difference Method (FDM), FEM and BEM. Chen et al. [33] presented a hybrid FEM/SFA method for preliminary cooling system evaluation and a cycle-averaged BEM for calculating the temperature distribution on the melt and mould interface. The standard and modified BEM plays a dominant role in mould cooling analysis to avoid mould domain discretisation so as to reduce the computation cost. Chiang et al. [24] pointed out that the standard BEM leads to numerical difficulties and convergence problems. Modifications have to be made to accommodate the thin mould cavities and long but small diameter cooling channels. To apply these hybrid 1-D/2-D approaches, the geometry simplification in the MIS and in circular cooling channels is necessary. However, the geometry simplification introduces some numerical errors around the corners and edges of parts, as well as locations lying directly beneath cooling channels, which subsequently influence the temperature calculation. Although 1-D methods are efficient for the calculation of MIS temperatures, their limitations and accuracy prevent them from more rigorous applications. Some researchers have also carried out works in CAE for injection and co-injection moulds [34-36] including the commercialised CAE package known as C-Mold. Chen and Chung [37] decoupled the heat conduction within the mould from within the part. The 3-D transient mould temperature is first calculated using 16

37 the Dual Reciprocity BEM (DRBEM). MIS temperatures are then used to evaluate heat flux along the MIS from the polymer melt using the finite difference method (FDM). They also revealed that the cycle-average based approach overestimates the temperature variation rate in mould cavity temperature within a steady cycle. Hu et al. [29] used Chen and Chung s approach to investigate the effect of various factors, including cooling line configuration, mould thermal properties, coolant temperature, part thickness and cycle time, on mould temperature distributions. Park and Kwon [38] proposed a thermal analysis system which was also based on the modified BEM in terms of accuracy and aimed to develop a design sensitivity analysis tool for the cooling system of injection moulds. In recent years, there has been a steady trend towards thinner part thickness, more complex part geometry, and more stringent quality and productivity requirements. With the rapid enhancement of computer capacity, a reliable 3-D transient mould cooling analysis becomes necessary to predict the temperature distribution of the plastic part and mould accurately. FEM is one of the best tools for this kind of analysis. Although some injection moulded plastic parts are thin, Tang et al. [39] suggested that 1-D or 2-D transient simulation for a complicated part geometry may not be accurate at sharp corners or at regions with sudden change in thickness. They proposed an implicit Galerkin FEM with a matrix-free conjugate gradient iterative solver to simulate a full 3-D cyclic, transient heat transfer problem. They reported that the technique is time accurate and efficient in solving the energy balance equation Cooling Channel Design and Optimization Several studies have been made on optimisation of the cooling system in injection moulding since early 1990s. Zou [40] has carried out a study on cooling system optimisation using an initial cooling configuration design. First an objective function is formulated as a measure of the temperature uniformity and cooling efficiency. The objective function is expressed in terms of 17

38 parameters related to the configuration of the cooling system and process conditions. By integrating an optimisation algorithm with a cooling analysis algorithm, the initial design can be fine-tuned to optimise the cooling system design. Matsumoto et al. [41, 42] applied 2-D BEM to simulate steady-state heat conduction within the injection mould and to optimise the layout of cooling channels. More recently, FEM approaches have been applied in the optimisation of mould cooling systems. Based on the time efficient FEM, Tang et al. presented an optimisation approach in which the objective function is chosen to represent the uniformity of mould temperature distribution and is iteratively evaluated using FEA until the minimum objective function is reached. The constrained optimal design problem is solved using Powell s conjugate direction method. They also applied the same technique in the optimisation of cooling systems for multi-cavity injection moulds [43]. Lin [44] developed an abductive network model to obtain limited number of important parameters in the cooling system and applied Simulated Annealing (SA) to achieve the optimal cooling design. FEM was applied to verify the feasibility of this method. Li [45] proposed a feature-based synthesis approach for automatic cooling channel design. In the context of cooling system design, a part with a complex shape is decomposed into shape features with simpler shapes. For each simpler shape feature, a cooling system can be obtained by direct application of design guidelines or examples. The cooling system of the entire part is then obtained by synthesizing from the cooling systems of the shape features. This method relates to feature recognition and definition of shape features. Figure 2.1 [45] illustrates the proposed conformal cooling design based on feature recognition algorithm. Most of the above mentioned works have used FEM to simulate the cooling channel performance. In this research work, a new approach of bi-metallic conformal and conventional cooling channel has been presented for a plastic part with thin flat surfaces with single and multi cavity mould, which will be thoroughly discussed in chapter five. 18

39 Figure 2.1: Conformal cooling design based on a feature-recognition [45]. Sun et al. [13, 46] proposed the U -shape milled groove and milled groove insert methods for large and complex moulded parts, especially for parts with large free-formed surfaces. The grooves need to be machined previously using CNC machines, making them more expensive than the straight-drilled cooling channels. However the method can be attractive to the moulding industry for higher productivity and part quality Other Solution Methods In injection moulding process, different thermal and structural steady or transient loading, can be modelled with Partial Differential Equations (PDEs), like other problems in engineering. There are lots of methods of solution of these PDEs researcher suggested. The selection of these methods depends on the feasibility and robustness, the accuracy requirement, capacity limitation of computer hardware, and the computational efficiency, cost and time. Finite Element Method (FEM) and Boundary Element Method (BEM) have very strong status in the field of computational methods in engineering, mostly 19

40 because of its greater flexibility and wider range of applicability [47]. It is particularly effective in problems of transient or steady state in regions of complex geometry [48-50]. Other numerical solution methods are Finite Difference Method (FDM) and Finite Volume Method (FVM). In both methods it is required that the computational domain be discretised using orthogonal structured meshes. For regular domains, the FDM is the simplest and the best numerical approach, and is the most widely used method in thermal analysis before the FEM was applied in this area. The mathematical concept of the FDM, approximating a continuous domain with a network of discrete points, called node, is relatively simple. The derivatives in the PDE are approximated by finite differences using the Taylor series on nodes. In general, first and second derivatives are estimated using second-order difference approximations. While the FVM, or the control volume method, is based on integral of the governing PDE over the control volumes. Both FDM and FVM lead to similar discretisation equations, the difference lies in the strictly mathematical approach in the FDM and the more physical one in the FVM. For problems with irregular geometric domains, finer discretisation and stair-stepping techniques near the curvature require more computational efforts. Zhou et al. [51] proposed an acceleration method for the BEM-based cooling simulation, in which the dense BEM matrix is split into a sparse dominant matrix and a dense residual matrix. The residual matrix is transformed from the inner iteration to the outer iteration, and the dominant item can be stored in the RAM memory so that the resulting system of the equation will be solved much more quickly and cost-effectively, which has been proved by numerical experiment. Qiao [52] presented a fully transient mould cooling analysis formulation using th e BEM based on the time-dependent formulation. In his study, he applied this technique in a T-shape plastic part and proved that using this technique, domain discretisation or use of point internal to the domain can be avoided. 20

41 Hassan et al. [53] designed cooling system of a plastic injection mould to provide thermal regulation in the injection moulding process. A numerical model by finite volume method was used for the solution of the physical model. A validation of the numerical model was also presented. 2.3 Works on Thermal Stress in Injection Moulded Plastic Part The thermally induced residual stresses are very important in case of injection moulded parts because they are directly responsible for the amount of part distortion. The temperature distributions in parts are used for the thermal stress analysis. Due to the difficulty of obtaining the temperature distribution of a complex part during cooling stage, only simple parts were simulated for the thermal stress analysis in earlier days. Jacques [54] analysed the thermal stress and warpage in flat parts due to unbalanced cooling. Thompson and White [55] measured the effects of temperature gradients on thermal stresses and distortions of parts. Matsuoka et al. [56] developed an analysis program, which takes into account mould filling, packing and cooling stages for the prediction of part warpage. Kabanemi and Crochet [57, 58] excluded the effect of packing pressure and employed the thermal visco-elasticity model to predict the residual stress and dimensional changes of a part when cooled inside the mould. Hastenberg et al. [59] measured the residual stress distributions in flat plates using a modified layer-removal method. Wang et al. [35] and Chiang et al. [24] have developed unified simulation programs to model the filling and postfilling stages in injection moulding. In their models, they perform simultaneous analyses of compressible fluid flow, heat transfer, fibre orientation and residual stress build-up in the material during flow and cooling. Several numerical approaches, such as hybrid FEM/Control Volume Method, FDM and FEM, are used. Chang and Tsaur [26] developed an integrated theory and computer program to simulate the shrinkage, warpage and sink marks of a rectangular flat part. Chang and Chiou [27] studied both flow and thermally induced 21

42 residual stresses during injection moulding of a thin cubic box part. Bushko and Stokes [60] assumed the polymer to be thermo-viscoelastic and model a molten layer of amorphous thermoplastic between cooled parallel plates. The solidification of this layer was studied to calculate part shrinkage, warpage and the build-up of residual stresses during the injection moulding process. Jansen et al. [61-63] calculated residual stresses and shrinkage of thin products using an elastic model to study the effect of in-mould shrinkage on the final product dimensions and measured the shrinkage under various moulding conditions. They have systematically studied the effect of processing conditions such as holding pressure, injection velocity, and mould and melt temperatures on shrinkage. Zoetelief et al. [64] investigated the influence of the holding stage on the thermal residual stress using a linear viscoelastic constitutive law and compared this with experimental results. Liu [65] simulated and predicted the residual stress and warpage using a viscoelastic phase transformation model, which assumed the solidified polymer to be a linear solid and the polymer melt to be a viscous fluid. Kabanemi et al. [57, 58]simulated residual stresses and deformations using 3-D FEM for a thermo-viscoelastic model and applied it to a complex shape. Choi and Im [66] adopted the model proposed by Bushko [60] and Stokes for analysing residual stresses during the packing and cooling stages. 3-D FEM, which enables the analysis for thin parts with complex geometries, was applied subsequently to analyse the part shrinkage and warpage. Chen et al. [28] employed thermoviscoelasticity with appropriate assumptions to establish the mathematical model governing the development of residual stresses in an amorphous polymer during the cooling stage of the injection moulding process, with the initial strain at the beginning of the cooling stage taken as the packing bulk strain. The residual-stress model during the cooling stage is solved with the FDM. Kansal et al. [67, 68] applied a 2-D FDM in analysing injection-moulded ring and gear. Zhang et al. [69] reported a 1-D approximation model of thermal residual stress in an injection moulded thin plate. Finally, it can be concluded 22

43 that according to the reviews, 3-D FEM appears to be the best tool for the thermal stress analysis of parts with complex geometry. Dimla et al. [4] presented a FEA thermal heat transfer analysis to determine an optimum and efficient design for conformal cooling channels in the injection mould. Analysis of virtual models showed that those with conformal cooling channels predicted a significantly reduced cycle time as well as marked improvement in the general quality of the surface finish when compared to a conventionally cooled mould. But no experimental verification has been presented. 2.4 Works on Mould Fabrication Using Rapid Tooling In last two decades, rapid prototyping (RP), rapid tooling (RT) and rapid manufacturing (RM) processes have found extensive use in speeding up tooling production. These processes greatly reduce the manufacturing cost and the lead time required for tool production. Rapid tooling is the process by which the moulds can be made using rapid prototyping technologies for injection moulding or die casting process, whereas rapid prototyping can only serve for functional prototyping before final manufacturing of the moulds. Table 2.1 gives a summary of the conventional RP processes, which have contributed significantly in manufacturing industries for functional prototypes and for development of RT and RM processes. 23

44 Table 2.1: Commercial Rapid Prototyping (RP) techniques Commercial Process Name How it works Depth/ Layers (mm) Layers/hour or Deposition rate Surface roughness Ref. Stereo Point to point laser Variable, >5µm [70-72] Lithography cure 4 in 3 /hour Selective laser Sintering of powder ± to [73, 74] sintering of metal or other - ± 0.25 materials Fused Extruded deposition thermoplastic modeling Laminated Bonded profile object cut sheet manufacturing 3-D Printing Select binder To powder Variable - ± mm [75] Variable >0.125mm [75] mm [75] Among the rapid prototyping techniques described in table 2.1 some of them have also been developed for metal freeform fabrication process. Table 2.2 summarizes the rapid manufacturing (RM) techniques, which have been used as main rapid tooling and metal based direct manufacturing processes. 24

45 Table 2.2: Direct metal Rapid Manufacturing (RM) processes. Process Selective laser sintering (SLS) Laser Engineered Net Shaping (LENS) Directed Light Fabrication (DLF) Direct Metal Deposition (DMD) Selective Laser Melting (SLM) Electron Beam Melting (EBM) Deposition type Accuracy : Horizontal Layer Thickness & deposition rate Materials used laser High - Steel, Cobalt sintering brazing, cladding 0.02mm 0.13 to 0.38 mm SS, Alloys and Z: 0.4mm low Numerous cladding to to mm SS, P20 and mm 1 to 2gm/min Numerous cladding mm H13, Al 0.1to 4.1cm 3 /min and Numerous Sintering/ µm Titanium Melting diameter alloy, SS and other materials Sintering / µm mm Titanium Melting high alloy, H13 steel and other materials Ref. [76, 77] [78, 79] [80-83] [22, 84-87] [88-94] [95-100] 25

46 In chapter one, it has been mentioned with an illustrative figure ( Figure 1.2) that rapid tooling (RT) technique can save up to 50% time in comparison with traditional tooling production. RT techniques can be classified into two categories, first one is direct RT and the second one is indirect RT techniques. Figure 2.2 (adopted from [7] ) illustrates the basic concepts of these two categories of RT techniques. In both techniques first 3D CAD model has been designed and then CAD data has been converted to closed volume elements. These data are processed and sliced into 2-D layers. These 2-D layers are then transferred to the RP machine, where the metal mould is built up layer by layer directly, or plastic pattern is built, which is later converted into metal mould by secondary processes. Figure 2.2: Basic steps of Rapid tooling and manufacturing [7]. 26

47 Indirect RT techniques can be chosen based on the underlying principle of making the positive pattern or negative pattern of plastic part. For example, in indirect RT technique using spray metal, the part is fabricated by RP machine and then the mould is made by spraying metal as shown in figure 2.3 (adopted from [7]). Figure 2.3 : Steps of metal spraying rapid tooling technique adopted from [7]. In direct RT technique, mould can be made directly using RM machine in any metal of the choice using techniques such as Selective laser Melting (SLM), Direct metal deposition (DMD) etc as shown in Table 2.2 or can be semimetallic processes such as FDM technique developed by Song et al. and Nikzad et al. [ ]. Nikzad et al. [103] have developed a new metal/polymer for FDM technique, which is suitable for making injection moulding die fabrication. This new metal/polymers are more thermally stable and conductive than the previously developed metal/polymer by Song et al. [102]. 27

48 2.5 Work on Conformal Cooling Channel by Rapid Tooling Conformal cooling [12, ] is defined as the cooling channels that conform to the surface of the mould cavity, core and stripper for effectively transferring the heat from the mould cavity to the coolant in the channels. The term conformal means that the geometry of the cooling channel follows the mould surface geometry. The aim is to maintain a steady and uniform cooling performance for the moulding part. From experimental results by several researchers, the injection mould cooling performance after utilizing conformal cooling channels can offer more uniform temperature distribution within the mould than the traditional cooling method. Heat can be evenly transferred or dissipated through the conformal cooling channel. However, very little work has been done in published work to fully investigate the performance and viability of such channels for injection moulding. Different from the traditional method of mould making, Sachs et al [12]. applied the 3-D printing (3DP) process, one of the RP techniques, to fabricate moulding tools with conformal cooling channels inside. As it is quite easy to create internal geometry using the RP process, the internal conformal cooling channels can be designed to maintain constant distance from the impression of the core and cavity so as to obtain more accurate temperature control. Such temperature control has the potential to minimise cycle times and to produce parts with lower residual stresses. Conformal cooling was reported to have no transient behaviour at the start of moulding and is able to maintain a more uniform temperature during an individual moulding cycle. Dalgarno et al. [108] applied indirect Selective Laser Sintering (SLS) in studies designed to assess the extent to which RP processes were capable of providing full production specification tooling for injection moulding. It is claimed that significant productivity benefits are available through the use of conformal cooling and the RP processes are economically competitive with existing production tool manufacture methods for injection mould tools which do not require significant finishing. 28

49 Ahn [109] presented a study by manufacturing a thermal management mould with three different materials and conformal cooling channels using a laser-aided direct metal tooling process to obtain both rapid and uniform cooling characteristics. The results of the experiments show that the cooling time of the thermal management mould can be shortened to 3s. Through comparison of the product from the thermal management mould with that of a previously designed mould, it is shown that the designed mould can improve both the cycle time and the quality of the product. Au [110] studied a novel design of variable radius conformal cooling channel (VRCCC), which is reported to achieve better uniform cooling performance. Thermal-FEA and melt flow analysis are used to validate the method. But no significant experimental mould making has been reported. Figures 2.4 (a) and (b) illustrate the conformal cooling channel of direct ACES injection moulding (AIM) prototype tooling, designed by 3D Systems in 1997 [111]. However, the geometry of the copper duct can only partially follow the shape of the moulding part. It cannot provide a true uniform temperature distribution in the injection mould. The bending of the copper duct is limited by its diameter, mechanical strength and the size of the moulding part. Further bending of the copper duct will damage the cooling channel. (a) (b) Figure 2.4: (a) Conformal cooling channel made by cooper duct, (b) Bending of copper duct evenly around the cavity wall [111]. 29

50 It is worth to focus on the relationship between the geometry of the moulding surface and the cooling channel. The technique shown in figures 2.4 is proposed to realize the conformal cooling channel with better cooling performance. Besides, properties like thermal conductivity and coefficient of thermal expansion are also important in the rapid tooling process. Thermal conductivity is based on the quantity of heat transmitted through a distance in a direction normal to a surface with a certain area due to a temperature difference. An increase in thermal conductivity of the mould shortens the time required to cool down the moulding part. As epoxy is the material having low thermal conductivity, aluminium filler can be added or mixed with epoxy. On the contrary, the coefficient of thermal expansion is the fractional change in dimensions of a material for a unit change in temperature. The value decreases when aluminium filled compounds are added. Aluminium filled epoxy have a better dimensional stability than unfilled epoxy for injection moulding in RT. Table 2.3 [112] indicates the coefficient of linear thermal expansion and thermal conductivity of various metal filled epoxies. Table 2.3: Mechanical properties of various metal filled epoxies [112]. Types of epoxies Coefficient of linear thermal expansion (10-6 / C) Coefficient of thermal conductivity (W/(m.k)) Unfiled Silica filled Aluminium filled The advancement of rapid manufacturing and Solid Freeform Fabrication (SFF) techniques give rise to the production of injection mould with intricate cooling channel geometry. Rapid tooling based on SFF technology such as RapidTool, SLS or rapid casting [113], also provide significant advantages to 30

51 plastic injection mould manufacturing. Much research has focused on improving the geometric design of the cooling channel via RT technologies. In 2001, Xu [114] studied injection mould with complex cooling channels based on SFF processes. He described the conformal cooling layout that can be realized with substantial improvements in part quality and productivity. He presented a modular and systematic technique for the design of cooling layouts by using 3DP. He suggested the decomposition of the injection moulded surface into definite controllable parts, called cooling zones. Then the cooling zones with the system of cooling layouts are further divided into definite cooling cells for analysis with the assistance of various design rules or constraints. He demonstrated his methodology via application to complex core and cavity for injection moulding. Figure 2.5 [11, 114] shows the green part of an injection mould with conformal cooling system design made by MIT s 3D printing. Figure 2.5: Green parts of an injection mould with conformal cooling channel design made by MIT s 3D Printing [11,114]. In 1999, Jacobs [72, 115] described the use of conformal cooling channels in an injection mould insert. The channels are built by electroformed nickel shells. From finite element simulation, the conformal cooling channel formed by 31

52 copper duct bending can increase the uniformity of mould temperature distribution. It can also decrease the cycle time and part distortion. As common injection moulding materials such as steel have not been included in his research, the application is only restricted to copper or nickel duct bending. Schmidt [116, 117] investigated and generated a series of design of experiments in an attempt to evaluate and measure the benefits of conformal cooling for injection moulding. He presented an overview of the mould design methodology, cooling channel simulation and analysis, and tool production through MIT s 3D Printing process. The simulation results showed that conformal cooling can reduce both cycle and cooling times. However, the mechanical strength, thermal stress of mould material and other mould defects were not taken into consideration in this work. Figure 2.6 [117] illustrates the comparison between conventional and conformal cooling design for cooling simulation. Figure 2.6: Comparison between conventional and conformal cooling design for cooling simulation [117]. Ferreira [21] attempted to use rapid soft tooling technology for plastic injection moulding. His work integrates rapid tooling with a composite material of aluminium-filled epoxy. The mould is cooled by conformal cooling channels. With the assistance of a decision matrix algorithm, a proper choice of materials and processes can be selected. The cooling layouts of the soft tooling are inserted with a bending copper duct before the epoxy filling process. However, 32

53 in reality, the geometries of the cooling layouts are not fully conformed to the model. The cooling and moulding performance are affected directly with the rough metal mould surface finish. Mould defects such as flash, weld line, sink marks and low back pressure appeared and cannot be avoided. Figure 2.7 [21] shows the soft RT mould with conformal cooling channel. Figure 2.7: Soft RT mould with conformal cooling channel [21]. As mould cooling is one of the limiting factors in the injection-moulding cycle, cooling channel design in RT is important for controlling the production time and quality. RapidTool [118] is a proprietary process from 3D Systems (formerly from DTM) based on selective laser sintering of LaserForm powder i.e. thermoplastic coated steel powder, and subsequent bronze infiltration. Conformal cooling channels can be incorporated into the moulds, which last for hundreds of thousands of shots of common plastic. Figure 2.8 [119] shows the workflow of the RapidTool process for tooling fabrication. Like RapidTool, the Copper Polyamide process is now available from 3D Systems and uses a mixture of bronze and polyamide powders and conformal cooling channels can be incorporated into the moulds. 33

54 Figure 2.8: Workflow of DTM RapidTooL Process [119]. Laminated steel tooling (LST) [ ] is a process that is employed to produce a laminated tool made of sheets of steel from laser-based cutting technology. The process is based on sequentially combining sheets of steel layer by layer with high-strength brazed joints for the laminated injection-mould fabrication. The advantage of LST is the production of tools that have dimensional accuracy comparable to injection moulding. The technology can give rise to produce complex geometric configuration within the injection mould. However, LST moulds are used only for low melting thermoplastics and are not appropriate for the injection-moulding process with thermosetting plastics or high-temperature glass fibre. The layered manufacturing feature of LST is capable of fabricating injection moulds insertion of conformal cooling channels into any shape or position required. Figure 2.9 [120] shows the hot platening process for LST production. 34

55 Figure 2.9: The hot platening process for LST production [120]. EOS s Direct metal laser sintering (DMLS) process utilizes specially developed machines and multi-component metal powders (mixture of bronze or steel with nickel). The SLS process is used for sintering, but no bronze infiltration is needed. Figure 2.10 [123] shows the electric cover mould with conformal cooling channel designed by EOS GmbH Electro Optical Systems. (a) (b) Figure 2.10: Electric cover (a) Conventionally drilled cooling channels (b) Optimized conformal cooling [123]. 35

56 Some conventional rapid prototyping process provides the capability for the development of rapid tooling for injection moulding via 3D Systems stereo lithography (SL). In the SL process, a photo-curable epoxy formed resin is solidified by exposing to an UV laser beam. In order to further improve thermal conductivity, copper channels or aluminium shots can be added to the low-melt alloy mix. The proposed design of cooling channel limits the consistency of the mould surface for heat transfer. Figure 2.10 [124] shows cross-sectional view of an injection mould assembly by SL technique. Figure 2.11: Cross-sectional view of an injection mould assembly by SL technique [124].. ProMetal [ ] is an application of MIT s Three Dimensional Printing Process to the fabrication of injection moulds. The ProMetal system creates metal parts by selectively binding metal powders layer by layer. It uses a wide area inkjet head to deposit a liquid binder onto the metal powders. The final metal mould is obtained by sintering and bronze infiltration similar to RapidTool of 3D Systems. Figure 2.12 [128] shows the design of the cooling channel in the mould by ProMetal. 36

57 Figure 2.12: CAD design and prototype of a rapid mould by ProMetal [128]. From the above review, much research has attempted to apply RP technologies to the design of conformal cooling channel. However, the increase in complexity of part geometries hinders the realization of conformal cooling layout fabrication in some RT processes. It is worthwhile to investigate further a more effective approach in order to obtain better cooling performances. This research work has been carried to investigate how different types of conformal cooling channels with different diameters and cross sectional shapes perform with different positioning from the cavity surfaces. For the first time, experimental mould making with square shape conformal cooling is presented using Direct Metal Deposition (DMD) rapid manufacturing technique to check that this cooling channel can be a potential alternative for injection moulding die fabrication. 37

58 2.6 Conclusion on Literature Review Review of the published work show that most studies focus on the development of their general purpose tool making techniques rather than on the cooling analysis and design of injection moulds. It is sceptical whether their methods were practical for industrial parts. In practice, SDCC is the most popular cooling method. Therefore, almost all the methods reported are based on this method, except those fabricated using RP processes. The conformal cooling channels are totally different from the conventional cooling solution that they cannot be machined by traditional mould-making methods. RP processes are normally much more expensive and cannot be adopted in the manufacturing of many products that essentially require the mould impression surface finishing. Besides, the durability of about 100,000 shots and the accuracy of ±0.2 mm [108] are still the major difficulties to be overcome using RP techniques. It is also notable that there is no significant research work done to check how different conformal cooling channels perform with different diameter, cross section, mould materials and positions from the cavity surfaces in injection moulding process. Considering the significance of these outstanding gaps of research, this thesis attempts to address some of these major issues. 38

59 3 Chapter 3 Conformal and conventional cooling channel design and analysis 3.1 Introduction In injection moulding, cooling begins with the mould filling during the moulding cycle. The major amount of heat exchange therefore happens during cooling time, which is the time until the mould opens and the plastic part has been ejected. To be able to eject the plastic parts, the hot molten plastic materials must need to be cooled down up to that extent that the product must withstand the forces during ejection without being deformed. Design of cooling channels should be in such a way that they cool molten plastic part rapidly, which ultimately improves process economics. But rapid cooling only is not sufficient for the part because it also needs to be uniformly cooled down till the demoulding temperature to avoid differential heat transfer. If the molten plastic materials do not cool down uniformly, it will cause differential shrinkage and internal stresses, which will reduce the part quality. Therefore, it is necessary to design cooling channels which will not only increase production rate but also improve product quality. Rapid and uniform cooling is achieved by sufficient number of properly located cooling channels. The location of these cooling channels should be consistent with the shape of the moulding as well as plastic part and as close to the cavity wall as allowed by the strength and rigidity of the mould. Increasing depth of the cooling line from the moulding surface reduces the heat transfer efficiency and too wide a pitch gives a non-uniform mould surface temperature. This chapter presents a transient thermal analysis of conventional and conformal cooling channels of various cross sections in a typical injection mould. The initial sections of the chapter describe the design parameters of cooling channels and factors affecting the cooling process performances in 39

60 injection moulding system. Difference between conventional cooling and conformal cooling channel has also been described with diagrams. Then cooling time equations for the mould are derived mathematically using governing heat flow equations to compare with the cooling times obtained from simulation. Finally, optimization of the cross section of conformal cooling channels has been described for an industrial plastic part mould with transient thermal FEA, by ANSYS Workbench simulation software. Figure 3.1: Different parameters of cooling channels, are shown in a cavity cross sectional mould. 3.2 Recommended Depth and Pitch of Cooling Channels As shown in Figure 3.1, cooling channels diameter d, depth D and pitch P, mostly depend on the mould materials and size of the mould or plastic parts. Strong mould materials i.e. structural steel can have lower values of depth and pitch, and, on the other hand, weaker mould material i.e. aluminium must have higher values of depth and pitch of cooling channels to withstand different moulding pressure and force. Different dimensions and locations of cooling channel based on some recommendation [1, 129, 130] are as follows. Diameter of the cooling channels, d= 10 mm to 25 mm, for standard size, single cavity mould (350x350x300mm 3 ) to big size, single cavity mould 40

61 (1.78x1.32x0.79m 3 ). For small mould (100x100x50mm 3 ), diameter could go down to 4mm. Depth of water line from the surface of the of the mould, D= 0.8d to 2d, for strong mould (i.e. tool steel) to weak mould (i.e. Aluminium) Distance between two cooling channels (pitch), P= 2.5d to 5d, for strong mould (i.e. tool steel) to weak mould (i.e. Aluminium). These recommendations are based on theory, to produce optimal and uniform surface temperature. It should take into consideration the fact that if the cooling channels are moved closer to the surface, the cooling effect on the surface will be greater near the channels and smaller in the space between them. If channels are moved closer together, the cooling will be more even, but the added cooling channels will increase the cost of the mould and require more flow of the coolant. Also the closer and additional cooling channels could weaken the mould, which may ultimately create fatigue failure and lessen the mould life. So, cooling channels should be designed in such a way that it must withstand the forces created by the plastic as well as the clamping forces during moulding process. These forces are repetitive and typically changes from zero to 150 MPa, and in some cases up to 200 MPa for thin plastic parts [9, 10, 131]. Safe and recommended values of stresses should always be predicted before designing cooling channels. For example, under cyclic conditions, the permissible stresses for steel mould should be held below 10% of the yield stress of the mould materials [132]. 3.3 Factors Affecting Cooling in Injection Moulding Main factors that affect the cooling rate of injection moulding are the temperature difference among molten plastic material, mould material and coolant temperature; thermal conductivity of plastic material, mould material and coolant; type of coolant flow and finally, the most important is cooling channel layout. This section of the chapter will describe some of the important 41

62 factors, which affect the cooling performance. Considering all these effects, different types of cooling channels design will be discussed afterwards Temperature Differences Temperature is a property of a body which determines the direction of heat flow from it or into it when it is placed in contact with another body. In injection moulding, cooling process removes a quantity of heat energy from the plastic part by transferring it through the mould, coolant and also in the moulding room to reduce the temperature of the plastic. This amount of heat energy removed from the molten plastic material raises the temperature of the coolant and the air around the mould and machine. Temperature difference is expressed as T in C. The following three important temperature differences are crucial factor in injection moulding a. Between injected molten plastic material and ejected plastic material or de-moulding temperature. b. Between molten plastic material and coolant temperature, and c. Between coolant inlet and outlet temperature. Temperature difference between injected hot molten plastic and ejected part, depends on the plastic material and it decides the amount of heat input. Normally, ejecting as hot as possible reduces the amount of heat removal by the mould and saves cycle time. However, if the cycle time is too short to remove most of the heat, the mould temperature will gradually rise and ultimately will be too hot, resulting in undesirable parts defects. If the cycle time is too long, the mould wastes time and production cost will not be economical. Heat flow from the hot plastic to the coolant depends largely on the temperature difference between them, the cooling channel design, and the mould material. Even though the plastic will cool faster with larger temperature difference, using lower temperature coolant will only slightly decrease the time required to 42

63 cool the plastic. Most of the cases, the added cost of chiller equipment and the power of using very cold coolant may not be justified economically, so it is advised that mould designer should emphasize more to improve cooling channel design and configuration of the coolant system. In most case, coolant temperature is kept in range 5-10 C. Some special cases, i.e. for materials ABS or PC, the cooling water are actually heated to the required temperature up to 60 C, by portable heating units. The coolant temperature rises as it flows through the mould and takes the heat away from them. In most of the injection moulding, in general, temperature difference between coolant inlet and outlet must be kept below 5 C. A greater value can result in uneven mould cooling and longer moulding cycles. For mass production moulds, this difference must be held below 3 C. Large quantities of coolant, supported by higher capacity and more expensive pumps and piping systems, are required to achieve such small temperature difference. The required investment and achieved benefits must be significantly evaluated [133] Thermal Properties of Materials Injection moulding normally uses high-strength moulds made of metals, primarily steel. While the frames are almost always made of steel, the cavities are frequently made of other high-quality materials. Now-a-days, non-metallic materials have been growing in importance in mould construction for rapid prototyping and rapid tooling. This is because of trend to new technologies, and also some of which can fabricate mouldings as quickly and inexpensively as possible that have been produced in realistic series production, so that they can inspect them to rule out weaknesses in the product and problems during later mass production. 43

64 Several factors determine the selection of materials for mould, most importantly cost, strength, wears resistance, thermal conductivity and ease of machining. The factor which influences most for cooling is the thermal conductivity, K. The value of K might be viewed as a consideration for cavity material selection, but in most cases the choice is made for performance over a large number of pieces, integrity of shape and controllability of fabrication and resulting values of K must be accepted. However, one must keep the values of K, of the selected materials in mind while designing the heat-transfer system. The K value can be used most advantageously for small-size deep cores where a straight and uniform opening is needed. For this application, it can be calculated whether a beryllium copper pin or steel pin with a copper core will be more effective in conducting the heat away. In this research work, a new method of cooling system with high thermal conductive copper tube insert has also been proposed; details of which will be discussed in chapter five. The plastic materials, which are generally poor conductors, need special attention while designing injection mould and cooling system. The thermoplastics can be classified into two types: the amorphous type plastics such as ABS (Acrylonitrile butadiene styrene), PS (polystyrene), PVC (polyvinyl chloride), etc. and the crystalline type plastics such as PA(Polyamides), PE (Polyethylene), PP(Polypropylene), etc. Amorphous plastics are heated up and then cooled down proportionally to get rid of the heat energy added [134]. Figure 3.2 (adopted from [129]) shows the difference in thermal behaviors of the two types of thermoplastics. 44

65 Figure 3.2: Plastic temperature vs time for amorphous plastic (left) and crystalline plastic (right) [129]. On the other hand, crystalline plastics require a certain amount of heat, which is called latent heat of fusion, at a certain temperature level just to melt the crystals without increasing the temperature of the plastic. Before and after the crystals have melted, the crystalline plastics are also heated up and then cooled down proportionally to remove heat that has been added. Therefore, a crystalline plastic needs more heat to raise a certain mass to the desired melt temperature than an amorphous material and also more cooling to remove its heat content. So this phenomenon of plastic has to be taken into account in designing cooling channels. Also the interface layer of plastic on the cavity or core walls, as shown in Figure 3.3, has very low thermal conductivity compared with that of the mould material; it forms an insulating layer between the mass of the plastic and the cold mould. This influences the rate of cooling to a much greater degree than the temperature of the mould or the material of the cavities. 45

66 Figure 3.3: Schematic cross sectional view of a mould showing interface layer between plastic part and core/cavity. This is of particular significance when designing moulds for products with heavy walls, or when there are isolated areas in the product which are thick and cannot be cooled properly. As the steady trend towards thinner part thickness continued in recent years [ ], thin-walled products, to some extent, nullify the insulating effect of the plastic as it cools down. This makes it important to efficiently carry the heat away from the cavity walls. 46

67 3.3.3 Cooling Medium and Flow of Coolant Cooling medium, which is mostly pure water, is a significant factor influencing cooling in injection moulding. Water is normally conditioned to minimize contaminations of the cooling system with lime deposits or other substances present in the water that corrode the cooling channels. These contaminations also could create deposit on the walls of the cooling channels, which ultimately reduces the heat transfer and can also block the passages and hence reducing the flow. Cooling medium temperature is also important factor, and in most cases it is kept between 5-10 C. The amount of heat removed from the mould is proportional to the temperature difference of the molten plastic and the coolant. There will be no effect of cooling if heated cooling medium is not removed and replaced by new coolant. In this regards, flow of coolant has to be taken into account. The factors which affect the flow of coolant are, o Cross section, length and layout of cooling channels o Reynolds number o Pressure difference between coolant inlet and outlet Cross section and layout of the cooling channels are the most important factors for necessary coolant flow. Normally, larger cross section allows more fluid to flow, as does the longer cooling channel. Flow rate is proportional to the square of the cross sectional diameter of the cooling channels. So large cross sectional area will flow more coolant, but it could reduce strength of the mould, therefore a balance should be achieved while designing cooling channels. Except in very large mould, the total length of cooling channels does not greatly change the flow, but direction of the length changes will add resistance of the flow, thereby reducing cooling effect. So it is recommended to have fewer change of direction of flow, especially with small cross sections of cooling channels. 47

68 Dimensionless parameter, Reynolds Number, is a measure of the effect of velocity of coolant flow and cross section of the cooling channels on the flow of coolant. Reynolds Number (Re) also depends on the density and dynamic viscosity of the coolant. If Re is less (<3200), flow is laminar and if it is above 4000, the flow is turbulent [13, 142]. From Figure 3.4 ( adopted from [129]), it can be seen that in case of laminar flow, only the coolant close to the wall removes heat while remaining flow just passes through the channel. On the other hand for turbulent flow, the entire flow of coolant contributes in the cooling by creating eddy current. In injection moulding, cooling channels in core, cavity or stripper must have turbulent flow. Figure 3.4: Schematic diagram of coolant flow inside a cooling channel (a) Turbulent flow, and (b) Laminar flow [129]. The temperature difference ( T) of the cooling water in and out, happens as the coolant flows through the mould and takes heat and hence its temperature rises. Now the question is: what should be the value of T? For normal moulding, this value should be kept between 4-6 C. Greater than this value can cause uneven mould cooling and longer moulding cycles. In high-production moulds, such as mould to make plastic bottles, the T should be kept below 3 C, even better between 1-2 C. To achieve required temperature difference, flow rate of water should be maintained in such a way that it takes sufficient heat energy from the mould. And also it is very important to have turbulent flow of water to remove required heat energy in the convection heat transfer 48

69 process in the cooling medium. In this research work, a new square shape cooling channel has been proposed to attain these requirements. Details of this new cooling channel have been discussed in chapter four and experimental verification has been done in chapter seven Cooling Channel Layout The cooling of an injection mould can be divided mainly into three categories cavity cooling, core cooling and cooling of the stripper plate. In literature review section, different types and layout of cooling channels that have been used since the invention of injection moulding have been discussed. Among all types of cooling channels, the straight drilled cooling channel (SDCC) system is the cheapest and the most popular cooling method, the injection moulding companies are using Guidelines for Straight Cooling Channel Because of simplicity and ease of manufacturing, straight cooling channels are the most commonly used cooling channels. Experiences and research shows that these channels are suitable for plastic parts that do not have curved surfaces other than round or fillet [129]. Straight cooling channels can be manufactured with conventional machining process i.e. straight drilling. If the plastic part surfaces are flat, this type of cooling channels will maintain equidistance from the mould cavity surface, shown as x, in Figure 3.5, and as a result uniform heat transfer will take place in moulding process. According to Fourier s law of conduction of heat transfer, distance that heat is conducting through, is inversely proportional to the total conduction heat transfer energy. 49

70 Figure 3.5: Schematic of straight cooling channels layout, showing in a cross sectional top view of a mould. So, distance has significant effect of total heat transfer process and in case of injection moulding, it is necessary to maintain equidistance from the interface of moulding surface and plastic part to the cooling channels for uniform transfer. This will ultimately give quality product. Straight cooling channels sometime need to use bubblers as shown in Figure 3.6, especially for slender cores. An inlet tube conveys the coolant into the hole of cooling channel of core and coolant flow changes direction and flows opposite through the hole to take away heat from core. The diameters of both tube and cooling channel have to be adjusted in such a way that the resistance of flow in both cross sections is equal. The condition for this is d/d=0.5, where d is the inner diameter of tube and D is the channel diameter. 50

71 Figure 3.6: Bubbler type cooling channels for a slender mould, in a schematic cross sectional front view of the mould Conformal Cooling Channel Conformal cooling channels which conform to the shape of the part in the core or the cavity or both are suitable for plastic parts that have significant curved surfaces. These channels are designed and placed as close as possible to the surface of the mould to increase heat transfer away from mould, ensuring part is cooled uniformly and efficiently. Figure 3.7 (adopted from [123]) shows a comparative diagram of convention straight cooling channels and conformal cooling channels [123]. In case of conventional tooling mould, an uneven conduction heat transfer will happen because of differential distance of the cooling channels from the molten plastic part or cavity surface. Also in the dead areas, marked green in Figure 3.7, in front of the threaded plug that are not passed through, dirt deposits accumulate, which leads to a steady decline of the general flow or unnecessary pressure loss. 51

72 (a) (b) Figure 3.7: Comparative cooling channel layouts for (a) Conventional and (b) Conformal [123]. On the other hand, as conformal cooling channels have close and uniform distance from the surface of the plastic part, it takes away heat effectively and uniformly from the molten plastic part. In chapter five detailed discussion of conformal cooling channel with optimized diameter, cross section and distance from the cavity surface will be discussed. 3.4 A case study of Cooling Channel Design A plastic bowl made of polypropylene has been selected as a case study. Plastic part is of 246 mm diameter, 68mm height, and 2.2 mm thick and it weighs 163 gm. Figure 3.8(a), shows the part as designed with Pro/Engineer CAD system. Figure 3.8(b), shows the core and cavity for the part as designed automatically with Pro/Mold system of Pro/Engineer. Finally, cooling channel has been designed for core and cavity with sweep cut tool of Pro/Engineer system. After designing the cooling channel, transient thermal finite element analysis has been done for optimisation of cross section to determine which cross sectional cooling configuration gives the best cooling time. Analysis will be discussed in a later section of this chapter. 52

73 Core (a) (b) Cavity Figure 3.8: CAD model of (a) Plastic bowl, and (b) Core and cavity in mould assembly Design for Cross Section Five different types of cross sectional areas of cooling channels have been designed for conformal cooling channel. For conventional cooling channel, one type i.e. circular has been designed. For conformal cooling, the five cross sectional shapes are rectangular, circular, square, hexagonal and elliptical as shown in Figure 3.9. These cross sectional shapes for the conformal cooling channels have been selected for the same hydraulic diameter, Dh=10.10 mm, as a benchmark (which is also used as the cross sectional diameter of the conventional cooling channel by the local industry). The hydraulic diameter Dh, [13] is defined by the following equation (3.1), D h =4A/P...(3.1) where, A= cross sectional area of the channel, and, P= wetted perimeter of the section. 53

74 The distance of the channels from the cavity or core surface has been taken as D=2Dh=22.2 mm. For the rectangular section, the ratio of l/w, as shown in Figure 3.9, has been taken as 1.4. With these parameters, which are industry standard for this particular plastic part, cooling channels have been designed for all cavity moulds as shown in Figure With these cooling channels, transient thermal finite element analyses (FEA) have been performed for optimization of cross section, which will be discussed in the last sections of this chapter. While designing the cooling channels sharp corners have been removed using rounds, which may cause fatigue failure of mould before expected life cycle due to stress concentration. Using equation (3.1), Dh has been calculated for all cross sectional areas which are shown in Table 3.1, Figure 3.9: Different cross sectional shapes of cooling channels (a) Rectangular (l/w=1.4) (b) Elliptical (c) Hexagonal (d) Square (f) Circular. Table 3.1: Data for calculation of hydraulic diameter D h. Type Area(mm 2 ) Perimeter (mm) D h (mm) a b c d e

75 1 st side cooling channel, S 1 (a) 2nd side cooling channel, S 2 (b) (c) Bottom cooling channel, S 3 (d) (e) Natural convection areas, S 4 (f) Figure 3.10: 3-D CAD models of the cavity mould, as designed by Pro/Engineer system, with different types of cooling systems; (a) conventional, (b) conformal with circular, (c) conformal with hexagonal, (d) conformal with elliptical, (e) conformal with rectangular, and (f) conformal with square cross sectional cooling channels. 55

76 3.4.2 Design for Diameter, Depth and Mould Material In this research, one of the main aims is to investigate how different positioning of the cooling channel as well as different diameter cross-section affects the cooling performance. Three different types of positions, that means the distance of the cooling channel from the cavity surface, have been investigated. For each of the distance, four different diameters types are used, and two different types of mould materials are used. So, total 3x4x2=24 metrics solution has been designed and analysed for each cross sectional types, with FEA transient thermal-structural analysis. Details discussion of this part will be presented in next chapter. 3.5 Heat Flow in the Mould and Governing Equations Before doing the analysis for optimum cross sectional area with thermal FEA, it is necessary to understand the heat flow system in injection moulding. In the injection moulding process, as shown in Figure 3.11, pellets or granules of polymer materials are fed through a hopper into a heated barrel, where it is mixed and melted, and then forced into a mould cavity through the nozzle either by a hydraulic plunger or by rotating screw system of an extruder. Molten material is then cooled by the cooling medium of cooling channels and hardens to the configuration of the mould cavity to get desire shape of the product. Heat normally flows from a body of higher temperature to a lower temperature body. Temperature difference between two bodies will determine the flow of heat. The greater the temperature difference between two bodies the higher the rate of heat flow between them. Heat can be transferred from one medium to another in three different ways: radiation, conduction and convention. 56

77 Figure 3.11: Schematic diagram of Injection moulding Machine with basic components Heat Exchange System in Injection Moulding Heat exchange between the injected plastic and the mould is a key factor in the economical performance of an injection mould. Heat has to be taken away from the thermoplastic material until a stable state has been reached for the plastic material for ejection or demoulding. The amount of heat to be taken off depends on the following factors, Temperature of the melt (plastic material) The demoulding temperature The specific heat of the plastic materials. Thermal diffusivity of the material. The ejection temperature of the polymer which is also called demoulding temperature, is lower than the injection temperature, but not necessarily the same as the room temperature. In injection moulding process, conduction heat transfer takes place in the mould material and goes through the interface of plastic material and mould material. Convection heat transfer happens in the cooling medium where heat energy comes through the interface of the mould material and cooling channels. Some convection heat transfer occurs also in the 57

78 mould surface, which is normally very insignificant. And finally, radiation heat transfer happens through the outer surface of the mould into the air which is also of negligible amount. In this research work, radiation heat transfer has been neglected. Figure 3.12 (adopted from [7]) shows a scheme of heat flow during injection moulding. Figure 3.12: The schematic of types of heat flow in an injection mould, based on [7]. The heat input by hot polymer melt must be removed as much as possible inside the mould before the mould can be opened to eject the part. Heat is extracted from the mould by the cooling system throughout the processing cycle. The mould cooling stages are necessary to ensure that parts are stiff enough to withstand the forces during ejection without being deformed while the other systems are idle. 58

79 3.5.2 Conduction and Convection Heat Transfer in Injection Moulding Whenever a temperature difference exists in a material, there will be an energy transfer from the high temperature domain to low temperature domain. This type of energy transfer is called conduction heat transfer. In injection moulding this energy start transferring through the interface of plastic surface (shown in Figure 3.13 & 3.17) and mould cavity surface and it continue through mould material to the direction of lower temperature domain i.e. cooling channels surfaces. Figure 3.13: Typical injection moulding heat transfer process 3-D domain. According to the French mathematical physicist Joseph Fourier, this energy transfer can be expressed by mathematical equation (3.2) [142], as follows, Q C =-KA T/L (3.2) Where, QC = conduction heat transfer energy K = thermal conductivity of the medium or material A = area of the cavity or core in contact with the plastic material or interface surface T= temperature difference between two domain (hot plastic material and coolant) 59

80 L = distance between two domain. Minus sign is inserted in equation (3.2), so that the second principle of thermodynamic will be satisfied; i.e., heat must flow into lower level of temperature. Equation (3.2) is valid for steady state heat transfer, which means temperature does not change with time. However, in injection moulding process, temperature of the material or domain changes with time. Therefore, it is necessary to develop transient heat transfer equation for such cases. Let s consider a unit cubic shape domain, Ω, of size dx.dy.dz as shown in Figure We consider a general case, where the temperature is changing with time and heat energy is generating within the body, which is Qg. So, for the element of thickness dx, energy balance equation can be written as: Figure 3.14: Unit volume (Ω) of 3-D dimensional conduction heat transfer process in Cartesian co-ordinates. Energy conducted in left face + energy generated within the element = change in internal energy + energy conducted out right face 60

81 Now, Conduction energy in the left face, Q x =-KA Energy generated within element, Q g = Q v Adx Change in internal energy = ρca dx Conduction energy out right face, Q x+dx =-KA x+dx =-A K K dx Where, Qv=energy generated per unit volume A =area of the face c =specific heat of material ρ = density T =temperature t =time Therefore, energy balance equation becomes, -KA + Q vadx =ρca dx - A K K dx or, K + Q v =ρc... (3.3) Equation (3.3) is one-dimensional heat conduction equation. Considering heat conduction in and out of the unit domain in all directions, as shown in Figure 3.14, the energy balance equation becomes, Q x +Q y +Q z +Q g =Q x+dx + Q y+dy + Q z+dz +ρcdxdydz Or, Or, K = K + K +Q v = ρc 61

82 + + + = α...(3.4) Equation (3.4) is the general 3-dimentional heat conduction equation where, α= ρ is called thermal diffusivity of the material. If α value is higher for a material, heat transfer rate will be higher for that material [142]. Another heat transfer phenomenon that happens in injection moulding is the heat transfer by convection. To express the convection heat transfer process, consider the heat transfer in an open channel as shown in Figure 3.15, where a fluid is flowing with temperature T, on a steel plate of surface temperature Tw, Figure 3.15: Convective heat transfer process in open channel. Now, according to Newton s low of cooling [ ] overall heat transfer by convection can be expressed by the following equation Q h =ha(t W -T )... (3.5) Where, A= surface area of the wall h= convection heat-transfer co-efficient 62

83 Equation (3.5) is the general equation for convective heat transfer for evaluating the heat energy loss for flow over an external surface. In case of injection moulding, convective heat transfer phenomenon happens in the cooling medium flowing inside the cooling channels. Heat which is conducted through mould to the cooling channel surfaces must frequently be removed by convection process for thermal balance as well as to allow plastic to return to its rigid state for ejection. Now, to calculate convective heat flow in a pipe, let s consider a cooling channel inside a mould, as shown in Figure 3.16, where wall temperature of the channel is TW and the cooling medium or coolant temperature is TC. So using equation (3.5), convective heat transfer on the surface of the cooling channel, can be written as, Figure 3.16: Schematic diagram of convective heat transfer process in injection moulding. Q h =ha(t W T C )... (3.6) Where, A=surface area of the cooling channels in contact with the flowing fluid. If we consider the real situation in injection moulding, values of Tw and TC are not constant, rather, it changes slightly. Equation (3.6) can be written as the bulk temperature difference, where bulk temperature gives the energy average 63

84 temperature values. Thus, for the cooling medium, total energy added can be expressed in terms of a bulk-temperature difference as given by, Q h =m t c p (T bo T bi )... (3.7) Where, mt =mass flow rate of cooling medium cp = constant specific heat of the cooling medium Tbi = inlet temperature of coolant Tbo =outlet temperature of coolant In some differential length dx the heat added dqh can be expressed either in terms of a bulk-temperature difference or in terms of convective heat transfer co-efficient. dq h =m t c p dt b =h c (2πr)dx(T W -T b )... (3.8) Where, TW and Tb are wall and bulk temperatures at the point of dx. The total heat transfer can therefore be expressed as Q h =h c.2πrl.(t W -T b ) av...(3.9) Where, r and L is the radius and total length of the cooling channel. Both TW and Tb values vary along the length of the cooling channel, so a suitable average value should be used for equation (3.9). To determine the values of convective heat transfer co-efficient, hc, Dittus and Boelter corrected [145, 146] equation, (3.10) as given below is used for calculating heat transfer in fully developed turbulent flow in smooth tubes. h c = R e 0.8 P r (3.10) Where, 64

85 K = thermal conductivity of flow material Dh = hydraulic diameter of the channel Re = Reynolds Number Pr = Prandtl Number Equation (3.10) is valid for 10,000<Re<120,000 and 0.7<Pr<120 [13] with moderate temperature difference between wall and fluid condition [142]. In this equation, the dimensionless parameter Reynolds Number (Re), characterises the ratio of inertia forces to viscous forces in a viscous flow and Prandtl Number (Pr), is a measure of the ratio of viscous diffusion to thermal diffusion. Re is typically used to express the transition from laminar to turbulent flow in forced convection, and is defined as Re=ρD h v/µ...(3.11) Where, ρ=density of flow material v=velocity of flow µ=dynamic viscosity of flow material Dh=Hydraulic diameter of the flow channel (can be measured by equation (3.1)) Prandlt number Pr, can be calculate by the following equation, Pr=Cpµ/k...(3.12) Where, Cp=Specific heat of the flow material µ = dynamic viscosity of flow material K = thermal conductivity of flow material 65

86 3.6 Cooling Time Calculation The heat exchange between plastic and the coolant takes place through thermal conduction in the mould core, cavity or stripper. Thermal conduction is described in 3-D form by the equation (3.4). But mouldings are primarily of a two-dimensional nature and because of cooling channel placement in the mould, heat can be considered to be removed or transferred only in one direction - the direction of their thickness (direction of y-axis as shown in Figure 3.13). Therefore, a one-dimensional computation can be considered for cooling time calculation. So, in case of one dimensional heat flow and with no internal heat generation, Qg=0, equation (3.4) can be written in following form =α...(3.13) Now, to solve this differential equation some assumptions are necessary. Assuming that, immediately after injection, the melt temperature of the plastic in the cavity has a uniform constant value of Tm, the temperature of the cavity and core wall that is in contact with molten plastic jumps rapidly to the constant value Tw and remains constant. Then the analytical solution of the equation (3.13) can be written as follows, T T w = 4 (T m-t w ).. α.sin....(3.14) where, Td =demoulding temperature, which is the temperature of molten plastic to just become solid enough to be ejected from the mould. s = maximum thickness of plastic part. Now, if only first term of the rapidly converging series is considered then the above equation becomes 66

87 π =. sin. Finally, resolving for cooling time tc, the above equation can be written as t c =. ln( )...(3.15) where, s= thickness of plastic part α= thermal diffusivity of plastic material = k= thermal conductivity of plastic material ρ ρ= density of plastic material Cp= specific heat of plastic material Tm=moulding temperature of plastic Td =demoulding temperature of ejected plastic material Tw= maximum temperature of interface wall between molten plastic and cavity wall Equation (3.15) will be used to determine cooling time, and to validate the result obtained from the FEA of heat transfer process for various design of cooling channels used in our investigation. 3.7 Finite Element Analysis of Cooling Channels for Optimum Cross Section After designing five types of cooling channels with different cross sectional geometry, transient thermal FEA has been carried out with ANSYS Workbench simulation software, which is one of the most versatile and widely used software for numerical modelling, simulation and optimization. Different mould cavity models that have been designed with Pro/Engineer CAD system have been imported in ANSYS thermal simulation module and then analysis has been carried out to find out which cooling design configuration gives the best temperature distribution in the mould. 67

88 3.7.1 Transient Thermal FEA with ANSYS Workbench By using ANSYS Workbench, engineers can easily evaluate product performance by simulating the behaviour of parts and assembly product in thermal loading conditions. ANSYS simulation module can perform steady state and transient analysis of a thermal problem. The steady state thermal analysis is used to calculate thermal response to heat loads subjected to prescribed temperatures and/or convection conditions. Steady thermal analyses assume a steady state for all thermal loads and boundary conditions. This characteristic is used to test the temperature distribution on the mould surface. Transient thermal analysis is used to calculate thermal responses over the period of time and therefore it is used to estimate the cooling time. After designing cooling channel systems for the injection mould, they need to be evaluated for the efficiency in terms of temperature distribution and cooling time. In this investigation, transient thermal FE analysis has been performed since temperature response on the mould needs to be checked for entire moulding cycle. Only the cavity mould has been considered for this simulation purpose to simplify the analysis timing. Figure 3.17 shows the cross section of the cavity mould containing the conventional cooling configuration (of Fig 3.10 (a)). In this configuration, there are two side cooling channels (shown as S1 and S2), which cools the side wall (SS) of the cavity or plastic part. Then, there is a bottom cooling channel (shown as S3) affecting the cooling of the bottom of the cavity wall(bs) or plastic part as well as sprue hole surface(shs). In ANSYS simulation the boundary conditions will be used for heat transfer taking places at these surfaces/walls (side, bottom and sprue) of the cavity mould. 68

89 Figure 3.17: Cross sectional cavity model, showing the surfaces thorough heat transfer happens during injection moulding Boundary Condition for Thermal Analysis To simulate the real conditions with CAE (computer aided engineering) software it is very much necessary to select the proper boundary conditions before running the simulation. Appropriate input values can just correctly predict real environment results. For thermal analysis, conduction and convection heat flux (Q/A) has been used as a boundary condition, and necessary equations and heat transfer diagrams have been described in the earlier sections of this chapter. Equation (3.2) has been used for conduction heat transfer that happens from the cavity surface or interface between cavity wall and plastic part, to the surface of the cooling channels. These two surface temperature readings have been taken from the results of cool+fill+pack+warp (complete injection moulding analysis) flow simulation analysis, which was run using the Autodesk Moldflow Insight (AMI), injection moulding filling simulation software. 69

90 Figure 3.18 shows typical example of different temperatures points for different surfaces and cooling channels, for conventional cooling channels. Figure 3.19(a) shows AMI simulation process parameters for conventional cooling channels, where, cooling channels have been considered for entire Figure 3.18: Different temperatures points for different surfaces and cooling channels for conventional cooling channels. mould assembly(detail discussion of this flow simulation with AMI, will be discussed in chapter seven). Figure 3.19(b) shows bulk temperature result of cavity wall or plastic part side, bottom and sprue surfaces, for conventional cooling channels, where, only a particular node point for various surfaces (T12060,T537 and B12598 respectively) temperatures have been plotted for entire cycle. For conduction heat transfer calculation, average values of temperature for each surface have been taken, which are shown in Table 3.2 for all conformal cooling channels. 70

91 (a) Temperature ( c) (b) Time (second) Figure 3.19: (a) Different input and out boundary conditions and process parameters setting in AMI simulation for conventional cooling. (b) Temperature vs time plot for different surfaces of plastic part/cavity wall for a typical node point in each surfaces. 71

92 For convective heat transfer, which happens through the cooling channel surfaces (S1, S2 and S3 in figure 3.17) to the cooling medium (water), equation (3.8), has been used, and values for cooling channel s walls (Tw) and cooling water temperatures (tb) have been calculated from AMI flow simulation. For each case an average value has been taken, because there is not much difference in temperature values through the cycle. Figure 3.20 shows a typical example for conventional straight cooling channel, showing (a) the temperature reading for tb as the coolant temperature and (b) the temperature reading for Tw as the wall temperature. These temperature values are also given in Table 3.2 along with different conformal cooling channels temperature values. (a) (b) Figure 3.20: Temperature recording of cooling channel after one cycle in AMI simulation (a) Cooling water temperature (b) Cooling channel wall temperature. 72

93 Table 3.2: Temperature values of different surfaces and cooling channels recorded from AMI flow simulation for all cooling channels for different times. Cooling channel types Conventional Straight cooling channels Temperature recording for different time in C Time (Sec) T S T b T SP T W T WS= 12 and T WB=12.5 t b Conformal cooling channels with circular cross section Conformal cooling channels with square cross section Conformal cooling channels with rectangular cross section Conformal cooling channels with hexagonal cross section Different interface surfaces and cooling water temperatures T S T b T SP T W T WS= 12 and T WB=12.35 t b 10.5 T S T b T SP T W T WS= 12 and T WB=12.30 t b 10.5 T S T b T SP T W T WS= 12 and T WB=12.30 t b 10.5 T S T b T SP T W T WS= 12 and T WB=12.40 t b 10.5 Conformal cooling channels with elliptical cross section T S T b T SP T W T WS= 12 and T WB=12.35 t b

94 It should be noted that though equation (3.2) has been used for conduction heat transfer, which is suitable for steady state heat transfer, heat flux values for conduction heat transfer have been used for different timing (shown in Table 3.3) of the moulding cycle to get the transient heat transfer effect in the simulation process. Different conductive heat transfer boundary conditions have been used for side, bottom and sprue surfaces, because these surface areas have different temperature gradient (table 3.2). Figure 3.21 shows process parameters and boundary conditions of transient thermal FEA analysis for conventional cooling channel in ANSYS Workbench. Bottom surface Conduction heat flux Convection heat flux Side surface Convection co-efficient for natural convection Sprue surface Typical conduction heat flux values for sprue surface Figure 3.21: Process parameters and boundary conditions of transient thermal FEA analysis for conventional cooling channel in ANSYS Workbench simulation software. 74

95 Table 3.3: Conduction heat flux values used as boundary condition for different cooling channels. Cooling channel types 1 Time (Sec) Heat flux values (watt/mm 2 ) Surfaces 2 CSCC SS BS SHS CCCS SS BS SHS CSCS SS BS SHS CRCS SS BS SHS CHCS SS BS SHS CECS SS BS SHS CSCC= Conventional straight cooling channels, CCCS= Conformal circular cross section, CSCS= Conformal square cross section, CRCS= Conformal rectangular cross section, CHCS= Conformal hexagonal cross section CECS= Conformal elliptical cross section 2 SS= Side surface, BS=Bottom surface and SHS=Sprue hole surface 75

96 For example, surface of the sprue bush area is the hottest spot of the mould; as a result it has the highest heat flux rate as shown in Figure 3.21 and Table 3.3, compared to the side and bottom surface of the cavity. The next highest heat flux is the bottom surface and then the side surface of the cavity mould. Conduction heat flux boundary conditions are shown for the conventional cooling channel in Figure 3.21, the first three in the left hand side. Table 3.4: Convective heat flux values used as boundary condition for different cooling channels Cooling channel types Time (Sec) Heat flux values (watt/mm 2 ) Surfaces CSCC S 1 S e-002 S e-002 CCCS CSCS CRCS CHCS CECS S 1 S 2 S 3 S 1 S 2 S 3 S 1 S 2 S 3 S 1 S 2 S 3 S 1 S 2 S e e e e e e e e e e : S 1S 2= Heat flux through side cooling channels surfaces, S 3 = Heat flux through bottom cooling channels surfaces, 76

97 Two different heat flux values, one for bottom cooling channels and the other for side cooling channels, have been used for convective heat transfer. The heat flux values for this case for different cooling configuration are shown in table 3.4. For convective heat transfer co-efficient, hc, equation (3.9) has been used. Reynolds Number (Re), calculated for mm hydraulic diameter (using equation (3.10)) cooling channel, is Reynolds Number is within the range (10,000<Re<120,000) of turbulent flow, which is one of the necessary conditions for optimum heat transfer process for forced convection heat transfer inside a tube. Prandtl Number (Pr) has been calculated by using equation (3.11), and is 7.22 and it is also within the turbulent flow range (and 0.7<Pr<120). It should be point out that heat flux values for final second (35 second) of the cycle has been taken as zero, because at that time, plastic part has been ejected. In real situation of injection moulding process, even after ejection of plastic part, some heat transfer in the mould happens. This is because mould remains hot for the next cycle to start and cooling water continues to flow in the channels. But in our simulation, this small amount of heat transfer has been neglected. Other thermal boundary conditions are the natural convection on the side surface of the cavity mould which is exposed to the air and the channel around the sprue bush, in which, air has been passed for additional cooling of sprue bush. This additional cooling of sprue bush is necessary as it carries the hot molten plastic material for injection into the mould cavity. Convection coefficients have been used as boundary conditions in these cases, and the values for these are 5x10-6 Watt/mm 2 C and 6.083x10-3 Watt/mm 2 C for the two convections respectively, as recorded by local mould manufacturer. So all together seven input boundary conditions have been used, three for conduction heat flux, two for convection heat flux and two for convection co-efficient as shown in Figure 3.21, left side. Temperature has been used as an output boundary condition, which is the prime interest in this study. Sample calculations for conduction and convection heat flux have been given in Appendices section A1. 77

98 3.7.3 Other Input Parameters for Thermal Analysis The mould material used for this simulation study is Stavax Supreme, a stainless tool steel alloy, as recommended by a local mould manufacturer. Properties of this material are shown in Table 3.5. Tetrahedral automatic generated meshing has been used, which has been optimized by software itself. Figure 3.22 shows meshing of a cavity mould. Table 3.5: Stavax Supreme properties Composition Rockwell Thermal conductivity Thermal Expansion Tensile Strength Hardness (HRC) (W/m* C) Coefficient (10-6 / C) S ut (MPa) Cr-Ni-Mo-V Figure 3.22: Half of cavity mould, showing tetrahedral meshing with rectangular conformal cooling channels, with number of nodes and elements shown in left side. 78

99 Numbers of nodes and elements for all types of cavity moulds have been given in Table 3.6. Room temperature has been taken as 20 C. Simulation is done for 35 second, which is the cycle time for this plastic part, in which 13 second is for injection and holding time, 18 second is for cooling time, and 4 second is for part ejection time. Temperature has been given as an output parameter, as purpose of simulation is to calculate the temperature distribution of mould over entire cycle of the injection moulding process. Table 3.6: Number of nodes and elements for tetrahedral meshing for all cooling channels. Type of cooling channel Number of Nodes Number of Elements CSCC CCCS CECS CHCS CSCS CRCS

100 3.8 Result and Discussion Figure 3.23 shows the temperature distribution in cavity moulds after 35 (a) (b) (c) (d) (e) (f) Figure 3.23: Temperature distribution in the cavity mould sectional view, after one cycle for (a) Conventional, and (b) Circular, (c) Square, (d) Rectangular, (e) Hexagonal and (f) Elliptical, cross sectional conformal cooling channels. 80

101 seconds (1 cycle) for all cases of different cross sectional type cooling channels. Also Table 3.7 gives the data of maximum temperature values in different moulds for each step of entire cycle time (1 to 35 seconds). From figure 3.23 and Table 3.7, it can be shown that all cooling channel temperature is rising till 11 th second, shown as red in Table 3.7. The reason for this is that the injection and holding time is 13 second, and so, no matter what types of cooling is being used, the mould will get heated, and consequently temperature will rise upto certain level for all moulds. Table 3.7: Temperature recoding in different cooling channel mould for entire cycle. But at 14 th second (red colored), which is the starting time of cooling, no further hot molten material is being injected. It is noted that, in cases of all conformal cooling channels, the temperature is sharply going down in comparison with conventional cooling. 81

102 It is also shown that using the conventional straight cooling channel, the maximum temperature in the mould after 35 seconds is C, which is also the demoulding temperature of the plastic part. Time Vs Temperature (Cooling curve) Coventional_Cooling_Channel Conformal_Circular_Cross-section Conformal_Square_cross-section Conformal_Rectangular_cross-section Conformal_Hexagonal_Cross-section Conformal_Elliptical_cross-section 110 Temperature ( C) Conventional Circular Elliptical Hexagonal 60 Rectangular 50 Square Time (Second) Figure 3.24: Temperature Vs time plot or cooling curve of the cavity mould with different cooling channels. 82

103 On the other hand, results of conformal cooling channel show better temperature distribution in terms of cycle time. Since the same max temperature of C in the mould can be achieved in around 26.5 seconds by using square and rectangular cross sectional conformal cooling channels, so about 8 seconds of reduction of cycle time can be possibly achieved using these cooling channels. It is also noted that by using circular, elliptical and hexagonal conformal cooling channels, cycle time can be reduced by 5, 3 and 3 seconds respectively. Now the question is which conformal cooling channel gives best cycle time? We can find the answer from Table 3.7 and Figure 3.24, which shows the cooling curves for all types of cooling channels used in our study. From Table 3.7 and Figure 3.24, it can be concluded that conformal cooling channels with all types of cross section give better cooling than conventional cooling channel. For this particular plastic part, it can reduce cooling time by around 3 to 8 seconds in comparison with cooling time of conventional cooling channel of 18 second. This is a redution of 17 to 44% of the conventional cooling time. Main reason for this is that the conformal cooling channel has the uniform distance from the cavity surface than the conventional cooling. Among the conformal cooling channel types, the square, rectangular and circular cross section types give the best results. Similar results have been found using other simulation software, such as Pro-Mechanica/Thermal for the same work, which have been mentioned in author s one of the published papers [6]. The reason for this is that, with the same hydraulic diameter (Dh), the rectangular and square sections have higher surface areas, and as a result, according to equation (3.8), more heat energy can be extracted from moltem plastic materials. Table 3.8 gives a comparison of total surface areas of all cooling channels used in this study. From Table 3.8, it can be noted that the rectangular cross section type has higher surface area than the square section but as the square surface has better uniformity than rectangular, it gives the best cooling time and temperature uniformity throught the cavity mould as shown in Figure 3.23 (c). 83

104 Table 3.8: Comparison of surface areas for different cross sectional cooling channels. Type of cooling channel Surface Area mm 2 CSCC CCCC CECC CHCC CSCC CRCC On the other hand, the conventional cooling channel has the higher surface areas than conformal circular and elliptical cooling channels, and also it has good surface uniformity, but because cooling channel distance from cavity surface is not the same for all cases so it has the worst result for cooling time. Although the rectangular cross section types gives almost the best cooling time, we are not considering it for our structural thermal FEA analysis because it has very high surface area in comparison with conventional cooling, which could cause structural failure. So considering the cooling time, uniformity and values of surface areas of cooling channels, the square and circular cross sectional types can be recomended for conformal cooling. In the next chapter, structurtural integrity of these two types of cross sectional cooling channels wil be discussed with various position, diameter and mould materials. Equation (3.14) has been used to calculate the theoretical values of cooling time ( sample calculation has been given in Appendices section A3) for all cases to verify the results of our simulation study. Table 3.9 shows the comparative values of cooling time calculatin for all cooling channels. 84

105 Table 3.9: Comparative cooling time data for theoretical and simulation result. Type of cooling channel Theoretical Cooling time (Second) Average cooling time from simulation result (second) Conventional Conformal Circular ,, Elliptical ,, Hexagonal ,, Square ,, Rectangular From Table 3.9, we can see that there are some discrepancies between theoretical cooling time values and the cooling time values obtained from the simulation results (from figure 3.24 and Table 3.7). This is obvious because the theoretical model is for only ideal cooling effect and equation (3.14) has been derived with a numbers of assumptions. But one point that is very much clear is that in both cases, cooling times have been reduced in all cases of using conformal coiling channels in injection moulds. 85

106 4 Chapter 4 Finite element thermal structural analysis of conformal cooling 4.1 Introduction In the previous chapter, cross sectional shapes of conformal cooling channel have been optimized. This optimization has been performed based on the result of the best temperature distribution in the cavity mould from thermal Finite Element Analysis (FEA) by ANSYS workbench simulation software. Necessary mathematical modelling for cooling channel design and boundary conditions for thermal analysis, have been discussed also in the last chapter. In this chapter, conformal cooling channels will be designed for different diameters and positions from the cavity surfaces. These cooling channels will then be tested for thermal-structural viability in injection moulding process, through thermal structural FEA by ANSYS simulation software. The major function of the finite element thermal-structural analysis is to accurately simulate the cooling process that takes place within an injection mould during moulding cycles. This analysis will be useful to test different cooling channel designs before manufacturing the expensive moulds. In this chapter, finite element method (FEM) for heat transfer analysis and related fundamentals including the method of weighted residual (MWR), the Bubnov- Galerkin method, and the integration by parts will be discussed first. Then FEM for thermal-structural stress analysis will be presented with a case study. 86

107 4.2 Finite Element Method for Heat Transfer Analysis The finite element method (FEM), also often known as finite element analysis (FEA) in practical applications is a numerical technique for finding approximate solutions of partial differential equations (PDEs) as well as of integral equations. The solution approach is based either on eliminating the differential equation completely for steady state problem, or interpretation of the PDEs, into an approximating system of ordinary differential equations, which are then numerically integrated using standard techniques such as Euler s method, Runge-Kutta [147] etc. The sequence of FEM solving procedure in heat transfer analysis includes [ ]. Discretising the domain or region, Selecting the interpolation functions, Deriving the element equations, Imposing the boundary conditions, Solving the time-dependent equation for transient analysis, Assembling the system equations, and Solving the system equations. Before discussing the above mentioned steps, FEM related fundamentals including the Method of Weighted Residuals (MWR), the Bubnov-Galerkin method, and the integration by parts will be discussed first, then the sequence of FEM solving procedure in heat transfer analysis for this research will be presented step by step. 87

108 4.2.1 The Method of Weighted Residuals Prior to development of the Finite Element Method, there existed an approximation technique for solving differential equations, called the Method of Weighted Residuals (MWR). This method will be presented as an introduction before discussing a particular subclass of MWR, the Galerkin Method of Weighted Residuals, which has been widely used to derive the element equations for the finite element method. The MWR is a global technique for obtaining approximate solutions to linear and non-linear PDEs. The application of this method in FEM is to find an approximate functional representation for a field variable in a PDE. For example, the solution of temperature T in Equation (3.4), (general 3-D heat conduction equation from chapter 3) is one of the examples. Because Equation (3.4) holds for any point in the solution in 3-D domain (Figure 3.14), it also holds for any collection of points defining an arbitrary sub-domain or element of the whole domain (Ω). Let s rewrite equation (3.4), for no internal heat generation (QV=0), as 2 T= α (4.1) where, Laplacian operator, 2 =. = + + Now, for an individual element in this 3-D domain (Ω), firstly, T is approximated by Ť, and heat flow per unit area (heat flux), ŎC is derived by substituting Ť into Equation (3.2) (from chapter 3), which is as follows, Ŏ C =-K Ť/L or, Ŏ C =K Ť ( considering L is very small ) 88

109 In matrix notation, Ŏ C =-KT D (4.2) where, TD is the 3-dimensional vector of partial derivative Ť. Normally, Ť is chosen to satisfy the global boundary conditions (i.e. heat flux in injection moulding process). Due to substitution of this approximation function into the PDE (Equation (4.1)), some error will exist in the results, which is called the residual R, and can be written as follows, R=. Ŏ C +ρc Ť 0 The MWR seeks to determine the unknowns in such a way that the residual R, over the entire solution domain is, on average zero. This is accomplished by forming a weighted average of the error and specifying that this weighted average would vanish over the solution domain. Hence, weighting functions W, which consists of the same number of linear independent terms as the unknowns, are chosen and can be written as follows, if R=0, in some sense. Ω RW i dω= Ω (. Ŏ C +ρc Ť )W i dω =0 (4.3) i=1,2,3..n Equation (4.3) represents a set of equations to be solved for the same number of unknowns. Many linear and even some non-linear problems have shown that Ť tends to T, when the number of equations tends to infinity. There are five MWR sub-methods, according to the choices for the weighting functions, W. These five methods are [150]: 89

110 1. Collocation method. 2. Sub-domain method. 3. Least Squares method. 4. Bubnov-Galerkin method. 5. Method of moments. Amongst the above methods of MWR, Bubnov-Galerkin method has been used most of the FEA simulation and therefore in this research work, Bubnov- Galerkin method has been used and discussed in next section Bubnov-Galerkin Method The error distribution principle most often used in FEM is known as the Bubnov-Galerkin method, or simply called the Galerkin method, in which the weighting functions, W are chosen to be the same as the interpolation functions. The weighting functions in the Galerkin method are the derivatives of the assumed solution with respect to the unknown coefficients. The unknown coefficients in the finite element form are the nodal degrees of freedom. Interpolation functions also play the role of weighing functions when using the Galerkin method. Selection of interpolation function will be discussed later in this chapter as a step to solve FEM Integration by Parts In mathematical analysis, integration by parts is a rule that transforms the integral of products of functions into other simpler integrals thus reducing the order of derivative in the integral function. It is also known as Green s and Gauss s theorems [151] in 2-D and in 3-D respectively. This method is widely used to solve the equations in FEM. For example, if this method has been used to solve Equation (4.3), it will reduce the order of the derivative function or number of products of the integral function, and hence it reduces the continuity constraint and results in wider choices of the interpolation functions. It also 90

111 offers a convenient way to introduce the natural boundary conditions that must be satisfied on some portion of the boundary. To define usefulness of integration by parts, let s substitute weighting functions W(x)=φ(x), into Equation (4.3), and integrate it using Gauss s theorem. Then the Gauss s theorem will give the surface integrals of the heat flow across the element boundary Γ or more precisely heat flux can be written as, Ω ρc Ť φ dω - Ω (Ŏ C. ) φ dω = Γ (Ŏ C.n) φ dγ (4.4) where, n is the unit normal vector to the boundary. From Equation (4.4), we can see that the highest order of derivatives in the integrands is reduced from the second to the first, which can be easily used as linear form of the element. It should be noted that for heat transfer FEA simulation in this study, heat flux has been used as one of the boundary conditions. 4.3 Finite Element Method Solving Procedure in Heat Transfer Analysis This section of the chapter will discuss about the different steps of the FEM solving procedure briefly, which will be used in finite element heat transfer analysis of our simulation. In the earlier section of this chapter, we have learned that, FEM solving procedure follows five important steps, and first step is to discretise the domain or continuum of any thermal or structural quantity. In the next section, discretising of a domain will be discussed. 91

112 4.3.1 Discretising the Domain In this research, only the discretisation with single type of element (i.e. tetrahedral) is considered for simplicity. Consider the same rectangular domain, Ω (figure 3.13 from chapter 3), as shown in Figure 4.1. Figure 4.1: A 3-D rectangular domain in injection mould. In figure 4.1, Ω= Ωm Ωp, and Ωm is the mould volume with cooling system bounded by Γ, and Ωp is the part volume bounded by Γp. The Γp is the mould impression surface, that is the plastic part outer surface which is in contact with mould surfaces. Γ is union of Γm, Γmp and Γmc, i.e. Γ=Γm Γmp Γmc, where, Γm is the boundaries of the mould exterior, Γmp is the mould impression surfaces i.e. Γp=-Γmp, and Γmc is the cooling channel surfaces respectively. Let s discretise the domain Ω into a union of Ee elements of r nodes each, where Ω= Ω, and Ω is an element either in the mould or in the plastic part i.e. Ω Ωm or Ω Ωp. Volume of the element is V (e), and V (e) 0 and Ω (i) Ω (j) = when i j. In finite element analysis, the temperature and temperature gradients 92

113 within an element can be approximated by the following equations [152], Temperature, Ť (e) (x, t) = φi (e) (x)t (e) i (t) (4.5) Temperature gradient, Ť, = φ T (e) i (t) (4.6) Similarly, Ť, = φ T (e) i (t) (4.6.1) and Ť, = φ T (e) i (t) (4.6.2) where i=1,2,...r, and Ti (e) (t) is the unknown, discrete, nodal temperature of the element. The nodes have only one degree of freedom which is the variable temperature. In this research work, this variable temperature will be used as an input parameter to calculate initial boundary condition (heat flux), in thermal FEA simulation. The function φi (e) (x) is the interpolation function, which holds the requirements given by the equation: φi (e) (x)=1 (4.7) It also satisfies compatibility requirement, having at least C n 1 continuity at element interfaces, and completeness requirement, having at least C n continuity within an element. If the field variable is continuous at element interfaces, it is called C 0 continuity, if first derivatives are also continuous, it is called C 1 continuity; and so on [147]. For simplification let s omit the superscript (e) and the parameters, x and t, in equations (4.5) and (4.6), and rewrite with matrix notation given by: 93

114 Ť = φ.t (4.8) Ť D = GT (4.9) where, φ is the r-dimensional vector of temperature interpolation, T is the r- dimensional vector of nodal temperatures, and G is the 3 r matrix of temperature gradient interpolation. G can be written as, φ φ φ G= φ= φ φ φ (4.10) φ φ φ Substituting Equation (4.9) into Equation (4.2), the approximation of heat flux can be written as, Ŏ C =-KGT (4.11) Selecting the Interpolation Functions Polynomial functions are normally used as interpolation functions, although there are other functions. The reason is that polynomials are relatively easy to manipulate mathematically, i.e., they can be integrated or differentiated without difficulty. In a typical 3-Dimensional domain, a complete n th -order polynomial can be written as follows, φ (x, y, z) φ (4.12) a+b+c n i=1,2,3.r where, r is the number of nodes of the elements in FEA analysis, which can be written as [153], 94

115 r= Linear form of the interpolation function, equation (4.12), can also be written as, φ(x, y, z)=ψ i1 +ψ i2 x+ ψ i3 y+ ψ i4 z where ψ s can be either constant or functions. The interpolation functions are derived from a natural volume coordinates system, which relies on the element geometry for its definition and whose coordinates range between zero and unity within the element. In this research, the 3-D, 4-node, linear tetrahedral finite element is applied. Figure 4.2: A tetrahedral elements showing 4 nodal coordinates. Let s consider a tetrahedral element as shown in figure 4.2, which is defined by nodes 1 (x1, y1 z1), 2 (x2, y2 z2), 3 (x3, y3 z3) and 4 (x4, y4 z4), and the nodes are numbered in such a way that, nodes 1, 2, and 3 are ordered counter clockwise when viewed from node 4, and then the 4 4 matrices can be defined by, 95

116 1 1 D= (4.13) 1 1 If d is the algebraic co-factor matrix of D, then, the interpolation functions of the above tetrahedral element are given by [2], φ= (d 1+d 2 x) (4.14) where, V is the volume of the tetrahedral element, which can be calculated using 6V= D, d1 is a 4-D constant vector, and d2 is a 3x4 constant matrix, and d1 and d2 are also sub-matrices of d, and can be defined as, d= d. Now, the definition of the interpolation functions, satisfies the requirement of interpolation function (Equation (4.7)), and approximated temperature at the node points calculated by using Equation (4.8) is equal to the nodal temperature, where φi(xi) = 1 and φj(xj) = 0 ( when i j). Temperature gradient interpolation and the approximation for that can also be derived from above equation, which are as follows, G = φ= (4.15) Ŏ C = - (4.16) where and is the transpose matrix [148] of d1 and d2 respectively Deriving the Element Equations with Boundary Conditions Equation (4.4) expresses the surface integral of the heat flow across the element boundary. Now, if we divide this integral in different surfaces and impose the boundary conditions, then it can be written as, 96

117 Γ (Ŏ C.n) φ dγ= Γmp (Ŏ C.n) φ dγ + Γp (Ŏ C.n) φ dγ + Γmc (Ŏ C.n) φ dγ But Γp=-Γmp, and heat flux at the surfaces of cooling channels is, ŎC =hc (Ť-TC) ((using Equation (3.6) from chapter three)). So, above equation can be written as, Γ (Ŏ C.n) φ dγ= Γmp (Ŏ C.n) φ dγ - Γmp Ŏ Γ φ dγ + Γmc h c (Ť-T C ) φ dγ (4.17) Equation (4.17), after using some boundary conditions (i.e. initial temperature values), and simplification, can be written as a first-order ordinary differential equation, which is as follows [149], MT D +KT = F (4.18) where, T is the 4-D, time-dependent, unknown, nodal temperature vector of which the initial conditions T(x,0) are known, TD denotes the first derivative of T over time, and M is the mass matrix (4x4) defined by equation (4.18.1), M= Ω ρcφφ dω (4.18.1) and K is the stiffness matrix (4x4) defined by equation (4.18.2) K= Ω (G T KG)dΩ + Γmc h c φ φ dγ (4.18.2) 97

118 and the discretised forcing matrix F is given by equation (4.18.3) F =- Γmp (Ŏ C.n)φ dγ + Γmp Ŏ Γ φ dγ + Γmc h c (T C ) φ dγ (4.18.3) Solving the Elementary Equation for Transient Analysis This section of the chapter will describe the fundamental approaches to solve the elementary equation (Equation (4.18)) in FEA. There are two fundamental approaches for solving elementary equation, which are frequently used in FEA, one is the mode superposition technique, and the other is the recurrence method [154]. Recurrence method is normally used when total number of nodes is large. In our FEA simulation, this method has been used and will be discussed briefly. To describe recurrence method, let s define t (n) as a typical time in 4-D field, and if t is the time increment, then t (n+1) = t (n) + t. The recurrence method will derive the recursion formulas that relate the values of temperature T, at t (n) to the value of T, at t (n+1), as a result, the solution marches in time, starting from the initial conditions at t=0, and continuing step-by-step until it reaches the desired time period. There is a popular time-marching method, which is called θ-method [155], where 0 θ 1, is introduced such that an intermediate time t (θ ) is defined by t (θ ) = t (n) + θ t within each time step. The first-order matrix equations at time t (θ ) are represented by finite difference approximations. Using this method, equation (4.18) can be written as, MT θ +KT (θ) = F (θ) (4.19) where, T θ = (4.20.1) and T (θ) =θt (n+1) +(1-θ)T (n) (4.20.2) where T (n) is initial known condition to advance the solution to time level t (n+1). 98

119 Similarly, if F is a function of time, it also can be approximated over the interval by, F (θ) =θf (n+1) +(1-θ)F (n) (4.20.3) Substituting Equations (4.20.1) to (4.20.3) into Equation (4.19) gives a set of linear algebraic equations of the following form, AX= B (4.21) where, X= T (n+1), which is unknown, A = M + θ TK (4.21.1) B = [M-(1-θ) TK]T (n) + T[θF (n+1) +(1-θ)F (n) ] (4.21.2) For a given θ, Equation (4.21) is a recurrence relation for calculating the vector of nodal values T (n+1) at the end of the time step from known values of the T (n) at the beginning of the time step. Equation (4.21) represents a general family of recurrence relations, where a particular algorithm that depends on the values of θ selected. For example, the θ-method are referred as the Euler or the forward difference method when θ=0, the Crank-Nicolson method when θ=1/2, the Galerkin method when θ=2/3, and the backward difference method when θ=0 [155] Assembling the System Equations The assembly of element equations is to combine all the element equations to form a complete set governing the composite of elements. The system assembly procedure is based on the compatibility that the values of the unknown nodal variables are the same for all elements connecting at an element nodes. The general procedure of assembly is summarised in the following steps after the element matrices have already been determined: It is necessary to set up an M M null matrix and an M-Dimensional (M- D) null vector, where M is the total number of system nodal variables, 99

120 which is equal to the total number of nodal temperatures in the domain for thermal FEA analysis. Then starting with one element, transform the element equation from local to global coordinates if these two coordinates systems are not coincident. Using the established correspondence between the local and global numbering schemes, change the local subscript indices of terms in matrices and vectors to the global indices. Then insert these terms into the corresponding M M matrix and M-D vector in the locations designated by their indices. Each time a term is placed in a location, where another term has already been placed, it is added to whatever value is there. Repeat this procedure (from second step) for one element after another until all elements have been treated. The result will be an M M master matrix A of stiffness coefficients and an M-D vector B of resultant nodal loads. After applying the boundary conditions of the specified nodal variables, the number of nodal unknowns and the number of equations to be solved are effectively reduced. However, it is most convenient to introduce the known nodal variables in a way that leaves the original number of equations unchanged and avoid major restructuring of computer storage. This can be accomplished by two straightforward iterative means of modifying the matrix A and the vector B in Equation (4.21). These two methods are, (a) If the i th nodal variable is specified, all the elements in the i th column and row of matrix A are set equal to zero and Aii is reset to unity. The Bi is calculated by substituting the specified Ti into Equation (4.21.2), while each of the M 1 remaining terms of Bj is subtracted by the product of Bi and Aji. 100

121 (b) Aii is multiplied by a large number, for example 10 15, and then the calculated Bi is multiplied by the modified Aii. The second method is simple and accurate enough with a large factor. It is also easy to be applied by computer coding, and, therefore, is adopted by most FEM packages (i.e ANSYS Workbench FEA simulation package) Solving the Matrix Equation Regardless of type of problem for which FEA is applied to solve, the final equation will have the form AX=B (Equation (4.21)), which represents a set of linear equations consisting of matrices associated with the various terms of the original differential Equation (4.18). As we can see in Equations (4.21), A is a sparse and symmetric matrix [148]. It is also positive-definite, i.e., p T A p > 0 for any non-zero vector p, which guarantees a unique solution X to Equation (4.21). It is customary to use iterative methods to solve large sparse matrix derived from the discretisation of the PDEs. Iterative methods are, in particular, well suited for large systems of equations (i.e. large M). Jacobi Preconditioned Conjugate Gradient (PCG) method [153] is one of the most popular iterative methods. In any iterative method, convergent rate depends also on the approximation of the initial vector. In the heat transfer analysis of injection moulds, the initial temperatures within the domain, or temperature distribution results in the previous step, are known and close to the unknown, therefore, are the most appropriate vector for the initial guess. PCG involves matrix-vector products and vector dot products, in which the former dominates the CPU time and requires large memory storage. The matrix-vector product can be split by expressing the product as the sum of the product of element matrix with vector, the so-called Element-By-Element (EBE) approach. It can be further split by applying Node-By-Node (NBN) approach. The basic system matrix-vector product of EBE and NBN method is: hg = AGPG 101

122 where, hg = h1 + h2 + +hm where the subscript G denotes global and r is the number of nodes per element. By applying EBE or NBN, the requirement of the formation and storage of a global matrix is eliminated, and therefore the total storage and computational costs become low. Furthermore, the amount of storage is independent of the node numbering, and mesh topology and depends on the number and type of elements in the mesh. EBE is well suited for higher order elements, while NBN is more suited for simple elements. Although both EBE and NBN support parallel processing, the parallel processing with NBN approach is more straightforward. As the 4-nodes, linear, tetrahedral element is appropriate in the mould cooling analysis using FEM, programs based on the NBN approach can be developed for parallel processing to accelerate the computations. Most of the commercial FEA simulation software coding is written based on the above fundamental steps of FEM. In our thermal-structural Finite Element Analysis, ANSYS workbench FEA package will be used. For thermal FEA, initial approximation of temperature values will be used, and for structural FEA, pressure and force will be used as initial approximation. 4.4 Thermal Structural Finite Element Analysis Thermal-structural Finite Element Analysis (FEA) has been performed for different cooling channels, which have been designed for different positions from the cavity surfaces, and diameters. First, transient thermal FEA has been carried out with ANSYS simulation software to check which cooling configuration with different position and diameter gives the best temperature distribution in the cavity mould. Then results of thermal FEA simulation, which gives the temperature response on the cavity mould for entire cycle, have been imported into the interface of transient structural analysis to calculate equivalent stress (von-mises) in thermal loading condition. A complete transient thermal-structural FEA has been carried out to check the robustness 102

123 and longevity of the moulds, with different cooling channels configuration for different diameter and positions from the cavity surfaces. In chapter three, detailed discussion of heat transfer in injection moulding has been done with necessary equations and boundary conditions. Same boundary conditions will be applied here as well for thermal FEA. Earlier in this chapter, theoretical discussion of thermal FEA has been done with mathematical equations and matrices. Same procedures have been followed for structural FEA as well. 4.5 A Case Study of Thermal-structural Analysis In chapter three, optimization of the cross sectional shape of the conformal cooling channel has been performed. Optimization has been done for an industrial plastic part s cavity mould. Same model has been taken for the case study of thermal-structural FEA analysis. Two different types of mould materials have been used; one is Stavas Supreme (SS) stainless steel and the other is, Aluminium (Al). SS is very strong material in comparison with Al, but Al has very high thermal conductivity as compared to SS. Table 4.1 shows the properties of SS and Al [156]. The reason for choosing these two mould material is to check how different types of cooling configuration with different diameters and positions from cavity surfaces perform with stronger or weaker, and high or low thermal conductive mould materials. 103

124 Table 4.1:Properties of Stavax Supreme and Aluminium [156]. Material Composition (%) Hardness Thermal conductivity (W/m* C) Density (Kg/m 3 ) Thermal Expansion Coefficient (10-6 / C) Tensile Strength Sut (MPa) Stavax Supreme (HRC 45, AISI 420 modified) Cr-Ni-Mo-V ( ) 45 (HRC) Aluminium (QC-7 Aluminum (HB) Alloy) Cooling Channel Design for Thermal-structural Analysis Two types of cross sectional shapes of conformal cooling channels, square and circular, which have been optimized in the last chapter based on the best temperature distribution in the mould and the cycle time, have been considered for this analysis. These conformal cooling channels have been designed for three different depth or positions from the cavity surfaces, which are 1.5Dh, 2Dh and 2.5Dh, where Dh is the hydraulic diameter of the cooling channel. For each of the three positions, four different diameters (Dh) cooling channels, which are 9mm, 10.10mm,11mm and 12mm, have been designed and considered for analysis. Table 4.1 shows the dimensions of all cooling channels that will be used in this study. 104

125 Table 4.2: Dimensions of conformal cooling channels. Types of cooling Diameter of the cooling Distance from channels Conformal circular cross sectional cooling channel Conformal square cross sectional cooling channel channels D h (mm) the cavity surface (mm) D h D h D h D h D h D h Boundary Conditions for Thermal-structural Analysis In chapter three, heat flux has been used as a boundary condition for conduction and convection heat transfer, and convection co-efficient for natural convection. Same boundary conditions will be used in this study for thermal FEA, because we keep the same plastic part as well as same cavity mould. Only difference will be in the values of heat fluxes in moulds surfaces with different diameters and positions of the cooling channels from cavity surfaces. In this study, different diameters of cooling channels have been used, and as a result, channel s coolant Reynolds Number (RN) as well as the convection coefficient 105

126 (hc) will be different. Table 4.3 shows the Reynolds Number and convection coefficient for different diameter cooling channels that will be used for this study. Sample calculations of RN and hc are given in Appendices section A2. Table 4.3: Values of Reynolds Number and Convection coefficient for different diameter cooling channels Diameter of the cooling channel D H (mm) Reynolds Number R N Convection coefficient h c (watt/m 2 C) Results of transient thermal analysis, which is the temperature response over the mould assembly for entire cycle, have been imported in the interface of transient structural analysis, to perform thermal-structural FEA analysis. For structural analysis, four types of boundary conditions have been used, which are fixed support, injection pressure, clamping force on the shut-off area and the temperature from the thermal analysis. Figure 4.3 shows the different surfaces used to apply boundary conditions for structural FEA. 106

127 Figure 4.3: Cross sectional mould, showing different surfaces used for applying injection pressure and clamping forces. To get the values of variable injection pressure at different surfaces (surfaces that plastic materials are in contact with mould surfaces during injection moulding process as shown in Figure 4.3) and clamping forces, which will be used as a boundary conditions for thermal-structural analysis, a complete injection moulding flow simulation (cool+flow+pack+warp analysis) has been carried out separately with Autodesk Moldflow Insight (AMI) software. In chapter three, it has been already discussed about the process parameters of the flow simulation with AMI. Also in chapter seven detailed discussion will be done for AMI simulations with different cooling configurations. 107

128 Force (Tonne) Figure 4.4: Variable clamping forces on the top surface of mould assembly for entire cycle from AMI analysis. Figure 4.4 show variable clamping forces, on the top surface of the mould assembly and Figure 4.5 shows injection pressure on different surfaces (surfaces that plastic materials are in contact with the cavity surfaces as shown in figure 4.3) from AMI simulation. 108

129 Pressure (MPa) Figure 4.5: Variable injection pressure on different surfaces of plastic part that is in contact with cavity surfaces, for entire cycle from AMI analysis. To calculate the injection pressure at different interface surfaces, the same procedure has been applied as that for calculating temperature values (as discussed in chapter 3). In should be noted that, Figure 4.5 shows the injection/hold-on pressure for a particular node point of a surface. For example, node point N714, gives the injection pressure on the side surfaces of the plastic part which is in contact with cavity surface. Similarly node points N1224, N2584 and N6311 gives the injection pressure for a particular point on round surface, bottom surface and sprue hole surface respectively. Table 4.4 shows the values of the clamping forces on the shut-off surfaces (as shown in figure 4.3, shut-off surfaces are the surface areas on the cavity which are in contact with core during injection moulding process, it also called the parting surfaces). Clamping forces on the shut-off surfaces have been 109

130 calculated, clamping force per unit areas on the top surfaces of the assembly mould from AMI simulation (figure 4.4). Table 4.4 also shows the pressure for different surfaces that has been found from AMI simulation, average values of pressure for each surfaces have been used for thermal-structural FEA. Table 4.4: Values of clamping forces and injection pressure used as boundary condition for thermal-structural FEA simulation. Injection pressure Clamping force Bottom Surface (BS) Round Surface (RS) Side Surface (SS) Sprue hole Surface (SHS) Shut off areas (SoA) Time (s) Pressure (MPa) Time (s) Pressure (MPa) Time (s) Pressure (MPa) Time (s) Pressure (MPa) Time (s) Forces (N) e e e e e

131 4.6 Results of Finite Element Analysis From transient thermal analysis, temperature distribution has been found for cavity moulds made of Stavax Supreme (SS) and Aluminium (Al) materials, and each material type having cooling channels with three different diameters, two different cross section shapes and four different positions from the cavity surfaces. Table 4.5 shows the maximum temperature in the cavity moulds made of SS and Al with cooling channels of different diameters, cross section and positions, after one cycle (35 seconds). Figure 4.6 and 4.7 show the typical temperature distribution after one cycle (35 second) in the cavity moulds made of two different materials (SS & Al) with conformal cooling channels (circular cross section, diameter mm and distance from cavity surfaces 2.5Dh=30 mm) and equivalent stress distribution in the moulds at approximately 3.9 th second respectively. Images of the temperature distribution for these cooling channel moulds, have been shown in the Appendix B in section B1. Table 4.6 gives the maximum equivalent stress or Von-Mises stree in the cavity moulds during the moulding cycle of all types of cooling channels. Also images of the equivalent stress distribution in the cavity moulds of these cooling channels, at approximately 3.9 th second of the moulding cycle have been given in the Appendix B in section B2. From Table 4.5 and Appendix B1 we can see that if the diameter of the cooling channel increases, cavity mould gives less maximum temperature after one cycle, which concludes less time required to cool down the plastic part. Table 4.5 also shows that if the distance of the cooling channel from cavity surface decreases, same result can be concluded. Among the two cross sectional shapes of cooling channels, square cross sectional shape gives slightly better result, in all cases of cooling channels than circular shape. Among the two types of materials of cavity mould, Aluminium (Al) mould gives better result than Stavax Supreme (SS) mould. 111

132 Table 4.5: Values of maximum temperature in the cavity moulds after one cycle for different cooling configurations Types of cooling channels Distance from the cavity surface (mm) Diameter of the cooling channels (mm) Maximum temperature in the cavity moulds after one cycle ( C ) Stavax Supreme Aluminium (SS) (Al) Conformal circular cross sectional cooling channel Conformal square cross sectional cooling channel 1.5D h 2D h 2.5D h 1.5D h 2D h 2.5D h Conventional Straight

133 (a) (b) Figure 4.6: Temperature distribution after one cycle (35 second) in the cavity mould sectional view with conformal cooling channels (circular diameter, pitch 2.5Dh=30 mm) made of (a) Stavax Supreme and (b) Aluminium materials. (a) (b) Figure 4.7: Equivalent stress distribution at about 3.9 th second of cycle for cavity mould sectional view with conformal cooling channel (circular diameter, pitch 2.5Dh=30 mm) made of (a) Stavax Supreme and (b) Aluminium materials. 113

134 Table 4.6: Values of maximum equivalent stress or Von-Mises stress in the cavity moulds during the moulding cycle of all types of cooling channels. Types of Distance Diameter Maximum equivalent or Von- cooling from of the Mises stress in the cavity channels the cavity surface (mm) cooling channels (mm) ( MPa ) Stavax Supreme Aluminium (SS) (Al) Conformal circular cross sectional cooling channel Conformal square cross sectional cooling channel 1.5D h 2D h 2.5D h 1.5D h 2D h 2.5D h Conventional Straight

135 On the other hand, from Table 4.6 and Appendix B2, we can see that if we consider the maximum stress distribution in the cavity moulds with different cooling configuration, the result is almost opposite with the temperature distribution. From Table 4.6, we can see that increasing the diameter of the cooling channel as well as decreasing the distance of the cooling channel from cavity surfaces, increases the maximum equivalent stress in the mould. Among the two cross sectional shapes of cooling channels, circular cross sectional shape gives less maximum equivalent stress in the mould, in all cases of cooling channels than square shape. Among the two types of materials of cavity mould, Aluminium (Al) mould gives slightly less maximum equivalent stress than Stavax Supreme (SS) mould. Values of maximum equivalent stress (Ses) in the cavity mould is a crucial factor to identify how long moulds can run before fatigue failure. If the values of Ses is higher and it exceeds the values of endurance limit (Se) or fatigue limit of the material, mould could fail before running the required amount of cycle. Next section of the chapter will discuss about the fatigue failure of the mould, and calculate life cycle of it, using maximum equivalent stress and high cycle fatigue formula. 4.7 Fatigue Failure and Life Cycle Calculation of the Mould When a structure is loaded steadily, and it s components properly designed, they will last theoretically forever. However, if it subjected to repeated load, even an insignificant change of load, the structure may fail early. Metals subjected to fluctuating loads break after a finite number of cycles, even though the stress is well below the ultimate strength of the metal. This is called fatigue failure. However tests have shown that a steel part which, under a certain repetitive load, has not broken after 2 million cycles [129], will not fail at all under this load. Fatigue plays important role in the life cycle of injection moulds. Injection moulds come across different variable loads i.e. 115

136 injection/hold on pressure, clamping forces etc. in thermal loading conditions. Depending on the material properties and value of stresses gathered from thermal-structural loading conditions, fatigue life of the mould can be calculated using high cycle fatigue formulas. In injection moulding, moulds frequently operates several millions cycles, so fatigue life of the moulds can be calculated using high cycle (10 6 to 10 7 ) fatigue formula. Using typical S-N (stress Vs life cycle) curve for steel mould [129, 157], Equation (4.22) can be used to calculate fatigue life of the mould. Life cycle of the mould, N, is given by: N= (S f /a) 1/b (4.22) where, 10 3 <N<10 7 a = (f*sut) 2 /Se and b = -(1/3)*log(f*Sut/Se) where, Sf = Ses= Equivalent stress or Von-Mises stress, Sut = Ultimate tensile strength Se= Elastic strength or endurance limit [157] = 0.504Sut Sut 1460 = 740 MPa Sut >1460 f = Constant (Values are given in Table 4.7 [157]) Table 4.7: Values of constant f for different S ut S ut (Mpa) f

137 Using the above equation and data, fatigue life or life cycle of the moulds made of Stavax Supreme (SS) have been calculated (Sample calculation has been shown Appendix A in section A4), using the values: Sut = 1500 MPa Se = 740 MPa f = 0.75 From Table 4.6, we can see that maximum equivalent stress for Aluminium (Al) moulds is around 500 to 600 MPa for all cases of cooling channels. But these values is very close to their ultimate strength (Sut), 610 MPa, or even higher than the Yield strength (Sy), 410 MPa. So, safe design of mould, with Al mould material for this particular plastic part is not possible even if it shows very good temperature distribution from thermal analysis. Even though in some cases of cooling configuration (Table 4.6), the maximum stress is below the ultimate strength, mould could experience fatigue failure because of repetitive high injection moulding cycles. Table 4.8 shows the values of fatigue life or life cycle of the moulds for two different cooling configuration, which is made of Stavax Supreme (SS) mould material. 117

138 Table 4.8: Predicted life cycle of the Stavax Supreme moulds with different cooling channels. Types of cooling Distance from the cavity surface Diameter of the cooling channels Life cycles (N) of the mould made of SS channels (mm) (mm) (Million cycles) Conformal circular cross sectional cooling channel Conformal square cross sectional cooling channel 1.5D h 2D h 2.5D h 1.5D h 2D h 2.5D h Conventional Straight Cooling

139 4.8 Summary From table 4.8, we can see that if the diameter of the cooling channel increases, and the distance of the cooling channel from cavity surface decreases, life cycle of the mould decreases, even though these two types of cooling configurations give better cooling effect in the mould (as shown in Table 4.5). On the other hand, opposite result is observed if cooling channels diameter decreases and distance of the cooling channel increases. So, compromise should be done while designing the cooling channels between better cooling effect and structural integrity of the mould. To get better cooling effect as well as structural robustness of the mould, it is found that conformal cooling with diameter Dh= mm and distance from the cavity surfaces, 2Dh =20.20 mm, gives the best result for this particular plastic part mould design. Now, if we consider the cross sectional shape of the conformal cooling channels, which shape should we design? We can get the answer from Table 4.8, in which, it clearly shows that square shape conformal cooling channel gives less fatigue life cycles than the circular shape, though it gives better cooling effect than the circular shape (as shown in Table 4.5). From Table 4.8, we can also see that, maximum life cycle using square shape conformal cooling channels (SSCCC) for all diameters and distance from cavity surfaces is 1.9 million. This concludes that, SSCCC can be used when mould life cycle more than 2 million is not desired rather than better cooling effect is necessary. On the other hand, circular cross sectional shape conformal cooling channel can be used for very high life cycle (more than 2 million) injection moulding process. In the chapter six, experimental mould making with square shape conformal cooling channels (SSCCC), by Rapid Tooling (RT) based on laser Direct Metal Deposition (DMD) will be discussed with a sample plastic part moulding. 119

140 5 Chapter 5 Bi-metallic conformal cooling 5.1 Introduction For optimum cooling effect, bi-metallic cooling channel with high thermal conductive copper tube insert has been proposed in this research for flat surfaces plastic part. This method is, to some extent, similar to the conventional straight cooling channels (CSCC). A highly thermally conductive material tube made of copper is inserted into the CSCC, to get higher heat transfer rate in the mould. Cost effectiveness and robustness of the mould, containing copper tube insert (CTI) cooling channel, have been taken into account, for optimizing the thickness of the copper tube. Optimization has been done through thermalstructural Finite Element Analysis (FEA), for an industrial plastic part mould assembly, using CTI cooling channel with different thicknesses of CTI. Experimental verification also has been carried out by moulding a sample part with different plastic materials with these cooling channels. 5.2 Conventional Cooling Methods Several cooling methods have been used in the mould industry. Most of them relate to core cooling which is normally more difficult than cavity cooling. Figure 5.1 (adopted from [1]) shows the scheme of the most frequently used cooling methods, showing the conventional straight (drilled) cooling channels including baffle and bubbler types. CSCC is the most popular method because of its simple and inexpensive nature. The cooling channels are machined using drilling or boring. They have a straight flow path and a circular cross-section. The cooling arrangement is preferably parallel, that is, each channel inside the 120

141 mould part, is supplied from a common coolant supply channel and empties into a common return channel. Tube Plug Figure 5.1: Schematic cross sectional diagram of the most frequently used cooling Channels [1]. Serial cooling layout, where several cooling channels are connected and share one inlet and outlet, can be applied only for a small mould. The mould has a lot of holes or recesses to accommodate ejector pins, guide pillars, sprue bush, insert, etc. As a result, the layout of a circuit is often constrained by the fact that, channels must not be drilled too close to any other channels or holes, horizontally or vertically, in the same mould part. Conventional straight cooling channels are also used as baffles and bubblers for cooling of cores in the injection moulds. These cooling channels as shown in figure 5.1 are also machined using drilling or boring. Baffle is a popular method for the cooling of cores with highly convex geometry. In this method, holes are bored to the rear face of the core, and the lower end of each boring is plugged (shown in figure 5.1). The borings are interconnected by a hole drilled from a side face. To ensure 121

142 that the coolant flows into each hole, baffles are fitted in each hole. To achieve more efficient cooling, twisted baffle are applied in practice. Bubbler is another popular method used in the core cooling. Similar to the baffle method, holes are bored to the rear face of the core. Instead of flowing from one side to the other, the incoming coolant passes through a tube in the centre of the hole (shown in figure 5.1) and returns to the outlet through the passage in between the tube and the hole. To be more efficient, sometimes the tube is replaced by a cylindrical helix with a hole in the centre. The coolant is forced to return to the outlet along the helical path. However, it is more expensive. In this chapter of the thesis, the performance of bi-metallic straight cooling channel (BSCC) and bi-metallic conformal cooling channel (BCCC), with two different thicknesses of copper tube insert (CTI), have been investigated for a cavity mould and core mould with bubblers, through the thermal-structural FEA analysis. 5.3 Cooling with High Thermal Conductive Material Injection moulding generally uses high strength mould materials, primarily steel. While the frame is almost always made of steel, cavities or cores sometimes are made with other metal inserts, especially the complex cooling channel. Their functions within the mould depend on specific properties and therefore, appropriate selection of the right materials is indispensable. Several factors determine the selection of these materials, which include economical consideration, strength, ductility and most importantly the thermal conductivity. For better cooling in injection moulding process, high thermal conductive material is crucial. Copper alloy could be the best solution for this case. The importance of copper alloy, as an auxiliary mould material for cooling, is based on high thermal conductivity and ductility, which equalizes stresses from non-uniform heating quickly and safely. Beryllium-copper alloys, however, are the best alloy materials for cooling in injection moulding because 122

143 of their high strength and high thermal conductivity. In this research work, this alloy has been used as an insert in the cooling channels for better cooling in injection moulding, which will be discussed in this chapter. Beryllium-copper alloy is one of the best nonferrous metals used in mould making. The mechanical and thermal properties of this alloy depend on the chemical compositions of the alloy. Interestingly, the mechanical properties of this alloy improve with increasing the beryllium (Be) content, while the thermal properties deteriorate. This alloy, with more than 1.7% Be, is mostly used in mould making. They have the tensile strength up to 1200 MPa and can be hardened up to 45 HRC. But for the functional components of mould, like cooling channel, a Be content below 1.7% is used, because it has high thermal conductivity as well as good dimensional strength. In our study, alloy with 1.6% Be has been used as copper tube insert (CTI). Beryllium-copper alloys can be employed as materials for mould s functional components, wherever high heat conductivity is important either for whole mould or part of it. By using this alloy, the temperature difference between cavity or core wall can be greatly reduced. Consequently, it results in higher output with the same, or even better, quality with the same cooling capacity. These alloys also offer an advantage, if high demands on surface accuracy have to be met. 5.4 Heat Transfer Through Composite Material Heat energy can be transferred through multilayer of different materials if two end side of the material have got differential temperature. We have already learned from chapter three that, heat energy will transfer from higher temperature domain to that of lower by conduction heat transfer process through the material. Now, in injection moulding, sometimes additional material with high thermal conductivity or high strength has been inserted for better performance of the mould. In this particular case, conductive heat transfer will be different because of different thermal conductivity of material. 123

144 To calculate conduction heat transfer for such a case, let s consider two layers wall of steel and copper plates as shown in Figure 5.2. Figure 5.2: Schematic of one dimensional heat flow through a composite wall of steel and copper plates. Assume that heat flow through both sections is same, so using equation (3.2), which is the general Fourier heat conduction equation, described in chapter three, conduction heat transfer energy can be written as, where, Q C =-k s A =-k c A (5.1) K s= thermal conductivity of steel plate K c= thermal conductivity of copper plate A = cross sectional area through heat is transferring l s = thickness of the steel plate l c = thickness of the copper plate T si = Inlet temperature of steel plate 124

145 T so = T ci= Outlet of steel plate or inlet of copper plate temperature T co=outlet temperature of copper plate. Solving the above equations, conduction heat transfer can be written as Q C = / / (5.2) This equation is called steady state one dimensional conduction heat transfer equation for composite materials, which will be used to calculate conduction heat flux as a boundary condition for thermal-structural FEA simulation for different cooling channel moulds in this chapter. 5.5 A Case Study of Bi-metallic Cooling Channel In this section of the chapter, a case study of bi-metallic cooling channel will be presented step by step. Performance will be discussed with the comparative thermal-structural Finite Element Analysis for bi-metallic cooling channel and conventional straight cooling channel. These thermal-structural FEA will be carried out with ANSYS workbench simulation software to demonstrate that bimetallic cooling channel can extract faster heat from molten plastic material in injection moulding process (IMP), and also to check the robustness and longevity of the mould containing bi-metallic cooling channel with copper tube insert (CTI) Design of Part, Mould and Cooling Channels The part chosen for this study is an injection moulded plastic canister (0.5 liter) made of PP thermoplastic. Necessary data and dimensions have been provided by a local die manufacturing company. Actual mould is of six cavities mould, but only single cavity type has been considered for this investigation. Figure 5.3(a) shows CAD model of plastic part, as designed with Pro/Engineer CAD system, and Table 5.1 gives the dimensions of the plastic part. From Figure 125

146 5.3(a), it can be seen that plastic part has flat surfaces with round corners. These flat surfaces will maintain uniform distance with the straight copper tube insert (CTI) cooling channel. Figure 5.3(b) shows the core and cavity mould for the part as designed automatically with Pro/Mold system of Pro/Engineer. Table 5.1: Dimensions and weight of the test plastic part. Length (mm) Width (mm) Height(mm) Weight(gm) Thickness (mm) Core (a) (b) Cavity Figure 5.3: CAD model of (a) Plastic canister, and (b) Core and cavity in mould assembly of the plastic part. 126

147 Figure 5.4 shows the cavity and core moulds for the plastic part with conventional straight cooling channel (CSCC) including bubbler cooling in the core. Figure 5.5 shows the design of bi-metallic straight cooling channel (BSCC) with copper tube insert (CTI) fitted. Figure 5.7 shows the design of bi-metallic conformal cooling channel (BSCC) with CTI fitted. CTI has also been used for bubbler system of core in both cases. Two different thicknesses, 2mm and 3mm of CTI, have been used for BSCC and BCCC. The difference between BSCC and BCCC, design is that, in case of BSCC, the channels are straight with no curved corners as shown in Figure 5.6, while in case of BCCC the channels have curved shape corners, which are conformal with the plastic part corners, and as a result these cooling channel maintain equal distance from surfaces of the plastic part as shown in Figure 5.8. Table 5.2 gives the names and abbreviations of five types of cooling channels that will be used in this study. The outer dimensions of the single cavity mould are height 232 mm, diameter 300, and the inner diameter of cavity and core cooling channels are 12mm and 15 mm respectively. Core Core cooling channels (bubbler type) Cavity Cavity cooling channels (Conventional straight cooling channel) Figure 5.4: Assembly mould with convectional straight cooling channels. 127

148 CTI for bubbler Bi-metallic straight CTI for cavity Figure 5.5: Bi-metallic straight cooling channel (BSCC) with copper tube insert (CTI) in cavity, and CTI in core bubbler channels. Plastic part Bi-metallic straight CTI, showing no curved corners as like plastic part Figure 5.6: Sectional top view of cavity mould, showing the orientation of BSCC in the mould. 128

149 CTI for bubbler Bi-metallic conformal CTI for cavity Figure 5.7:Bi-metallic conformal cooling channel (BCCC) with copper tube insert (CTI) in cavity and CTI in core bubbler channels. Plastic part Bi-metallic conformal CTI, showing curved shape corners as like Figure 5.8: Sectional top view of cavity mould, showing the orientation of BCCC in the mould. 129

150 Table 5.2: Different cooling channels and their abbreviations. Type of cooling channel CSCC BSCC 2mm CTI BSCC 3mm CTI BCCC 2mm CTI BCCC 3mm CTI Description Conventional straight cooling channel Bimetallic straight cooling channels with 2mm copper tube insert Bimetallic straight cooling channels with 3mm copper tube insert Bimetallic conformal cooling channels with 2mm copper tube insert Bimetallic conformal cooling channels with 3mm copper tube insert It should be noted that, bubbler system has a central tube, which helps the incoming coolant passes through it, in the centre of the cooling channel hole (as shown in figure 5.1). In this study, this tube has not been counted for simplicity of design and analysis Thermal-structural Finite Element Analysis Thermal-structural FEA of the proposed bi-metallic cooling channel moulds have been carried out with ANSYS workbench simulation software to prove that such mould can take away heat faster from molten plastic material in injection moulding process. In the transient thermal-structural FEA process, first, the transient thermal analysis has been carried out in the mould assembly with different cooling configurations, and then thermal result has been coupled for transient structural analysis to calculate equivalent stress (von-mises) in thermal loading condition. In chapeter three and four, details of thermal and structural analysis have been discussed with mathmatical modelling and then its application for a case study. Same mathmatical modelling and boundary conditions will be applied for this analysis also. In this analysis, the mould material was taken Stavax 130

151 Supreme (SS), stainless tool steel alloy, as recommended by local mould manufacturer. The cooling channel insert material was high thermal conductive Copper Alloy (CA) which is capable of transferring heat at a higher rate than steel. Table 5.3 [156] shows a comparison of the physical properties of SS and CA. Table 5.3: Properties of Stavax Supreme and Copper alloy [156]. Material Composition (%) Rockwell Hardness (HRC) Thermal conductivity (W/m* C) Density (Kg/m 3 ) Thermal Expansion Coefficient (10-6 / C) Tensile Strength Sut (MPa) Stavax Supreme (HRC 50, AISI 420 modified) Cr-Ni-Mo-V ( ) Copper alloy (Beryllium- Copper) Cu-Br-Ni (balance ) ANSYS workbench simulation software has been used to carry out thermalstructural analysis, which is capable of simulating both the steady state and transient behaviour when subjected to different structural and heat loads. In this simulation, transient analysis has been used because, in injection moulding, mould experiences variable temperature, pressure and forces. Automatic meshing (elements that are automatically created depending on the physical structure) with tetrahedral elements have been used. Fine relevance centre and medium smoothing have been applied in the meshing. Table 5.4 shows the meshing data for different cooling channel mould. As we can see from the Table 5.4 that maximum number of elements are about The reason for this is that medium smoothing mesh sizing have been used for ease of iteration in numerical analysis. 131

152 Table 5.4: Number of nodes and elements for different cooling channel mould. Type of cooling channel Number of nodes Number of elements CSCC BSCC 2mm CTI BSCC 3mm CTI BCCC 2mm CTI BCCC 3mm CTI It has been discussed in earlier chapters that in the injection moulding process, three types of heat transfer take place; conduction through mould, convection in the cooling medium and outer surface of the mould and finally, radiation heat transfer, which is of very negligible amount. For thermal analysis, conduction and convection heat flux (heat energy per unit area) have been used as a boundary condition. Necessary equations and heat transfer diagrams have been described in chapter three. Conduction heat transfer, which is of vital importance in injection moulding process, has been calculated by equation (5.2) [142]. Using equation (5.2), conduction heat transfer energy, for this moulding process can be written as, Q C = / / (5.3) Where, k ss = thermal conductivity of SS, k c = thermal conductivity of CTI, A = cross sectional area through heat is transferring, 132

153 T w = inner surface temperature of CTI. T i = temperature of cavity or core surface interface with plastic l s = distance from cavity or core surface to nearest CTI, inner surface l c = thickness of CTI, Equation (5.4) has been used to calculate convective heat flux inside the cooling channels surface. Using equation (3.8) (as described in chapter three), convective heat transfer for this case study can be written as, Q h =h c A(T W T C )... (5.4) where, A = Inner surface area of the CTI in contact with coolant. T W = average temperature of the inside surface of CTI. T C = average temperature of the coolant, h c = convective heat transfer co-efficient The convection heat transfer coefficient, hc, has been calculated using equation (3.10) (Dittus-Boetler correction equation), as explained in chapter three. These coefficients were calculated to be 5397 and 5709 watt/m 2 C for core and cavity cooling channel respectively. Other thermal boundary conditions are the natural convection on the side surface of the cavity mould, which is exposed to the air and the channel around the sprue bush, in which, air has been passed for additional cooling of sprue bush. This additional cooling of sprue bush is necessary as it carries the hot molten plastic material for injection into the mould cavity. Convection coefficient has been used as boundary conditions in these cases, and the values for these are 5x10-6 Watt/mm 2 C and 6.083x10-3 Watt/mm 2 C respectively, as recorded by local mould manufacturer. So, all together eleven input boundary conditions have been used for thermal analysis, seven for conduction heat flux (heat flux for SS, RS and BS surfaces for both core and cavity side, and SHS 133

154 surface shown in Figure 5.9), two for convection heat flux (heat flux for CoCs and CaCS surfaces shown in Figure 5.9) and two for convection co-efficient. Figure 5.9 shows the cross section of the entire mould assembly, showing different interface surfaces used to apply heat flux boundary conditions. In order to calculate different heat fluxes using equation (5.3) and (5.4), it is necessary to know the variable temperature, Tw, Ti and TC at these different surfaces as shown in figure 5.10 (the interface of plastic and mould cavity, at the interface of cooling channel inner surface and cooling medium as shown in Figure 5.9). To get these temperature values, a complete injection moulding flow simulation (cool+flow+pack+warp analysis) has been carried out separately with Autodesk Moldflow Insight (AMI) software. Flow simulation with AMI, also gives the values of the variable injection pressure at different surfaces (surfaces that plastic materials are in contact with mould surfaces during injection moulding process as shown in Figure 5.9) and clamping forces, which will be used as a boundary conditions for thermal structural analysis. Figure 5.9: Cross sectional assembly mould, showing different interface surfaces. 134

155 Figure 5.10: Different temperatures notation for different surfaces and cooling channels in a typical mould. For plastic flow analysis with AMI, dual domain mesh has been used with 9228 elements, mould and melt temperature were 50 C and 250 C respectively, total cycle time was 20 seconds (9, 8 and 3 seconds for injection/hold on, cooling and ejection respectively), plastic and mould materials were PP and SS stainless tool steel correspondingly. Pure water with temperature 10 C has been used as coolant. Figure 5.11 shows the typical process parameters and boundary conditions for CSCC, in flow simulation with AMI. 135

156 Figure 5.11: Different input and out boundary conditions and process parameters setting AMI simulation. Figure 5.12 shows bulk temperature result of plastic part surfaces which is in contact with cavity and core wall, and sprue surfaces (as shown in figure 5.9), for conventional cooling channels, obtained from AMI simulation. In this figure only a particular node point for each surface has been plotted for entire cycle. For example, node point T90656 represents a point in the side surface of plastic part that is in contact with cavity mould. Similarly, node points T89967, T88073 and B492 are points, for round, bottom and sprue surfaces respectively. Average values of the temperature for each surface have been taken for heat flux calculation. Table 5.5 shows the temperature values at six different times (0 to 17 second) of the moulding cycle at different interface surface of the assembly mould (as shown in figure 5.9) for all five cases of cooling channels, from AMI simulation. For a particular surface, same temperature values have been taken 136

157 for core and cavity side, because there is not much difference between them, as shown in figure 5.9. Node point T90656 temperature values is almost same with node point T91037 (which is a point in the core side of same surface) temperature values. Temperature( C) Figure 5.12: Temperature vs time plot for different surfaces of plastic part which is in contact with cavity and core wall for a typical node point in each surfaces, obtained from AMI simulation. 137

158 Table 5.5: Temperature values of different surfaces and cooling channels recorded from AMI flow simulation for all cooling channels for different times. Types of cooling channels Interface Surfaces Temperature ( C ) Time (s) Side surface (SS) CSCC Round surface (RS) T i Bottom surface (BS) Sprue hole surface Core cooling surface T W & T C T W = T C = (coolant temp.) Cavity cooling T W = 12 T C =10.15 (coolant temp.) BSCC 2mm CTI Side surface (SS) Round surface (RS) T i Bottom surface (BS) Sprue hole surface Core cooling surface T W & T C T W = T C = (coolant temp.) Cavity cooling T W = 12 T C =10.15 (coolant temp.) Side surface (SS) BCCC 2mm CTI Round surface (RS) T i Bottom surface (BS) Sprue hole surface Core cooling surface T W & T C T W = T C = (coolant temp.) Cavity cooling T W = 12 T C =10.15 (coolant temp.) Side surface (SS) BSCC 3mm CTI Round surface (RS) T i Bottom surface (BS) Sprue hole surface Core cooling surface T W & T C T W = T C = (coolant temp.) Cavity cooling T W = 12 T C =10.15 (coolant temp.) BCCC 3mm CTI Side surface (SS) Round surface (RS) T i Bottom surface (BS) Sprue hole surface Core cooling surface T W & T C T W = T C = (coolant temp.) Cavity cooling T W = 12 T C =10.15 (coolant 138

159 Figure 5.13 shows a typical example for conventional straight cooling channel (CSCC), showing (a) the temperature reading for Tw, as the coolant metal or wall temperature and (b) the temperature reading for TC, as the coolant temperature. These temperature values are given in Table 5.5 along with different CTI cooling channel temperature. (a) (b) Figure 5.13: Temperature recording of cooling channel in AMI simulation (a) Cooling channel wall temperature (b) Cooling water temperature. Table 5.6 shows the heat flux values calculated at different times (0 to 17 second) and used as boundary conditions for various interface surfaces for different cooling channels. In this table we can see that different interface surfaces have different heat flux values. The reason for this has been explained in chapter three. 139

160 Table 5.6: Heat flux values used as boundary condition for different cooling channels. Cooling channel types Interface surfaces Heat flux values (watt/mm 2 ) for different time (s) of the cycle CSCC BSCC 2mm CTI BCCC 2mm CTI BSCC 3mm CTI BCCC 3mm CTI SS RS BS SHS CoCS CaCS SS RS BS SHS CoCS CaCS SS RS BS SHS CoCS CaCS SS RS BS SHS CoCS CaCS SS RS BS SHS CoCS CaCS

161 Result of transient thermal analysis, which is the temperature response over the mould assembly for entire cycle, has been imported in the interface of transient structural analysis, to perform thermal-structural FEA analysis. For structural analysis, four types of boundary conditions have been used, which are fixed support, injection pressure, clamping force and the temperature from the thermal analysis. Figure 5.14 and 5.15 show variable clamping forces, on the top surface of the mould assembly and injection pressure on different surfaces (surfaces that plastic materials are in contact during IPM) respectively for entire injection moulding cycle, from AMI analysis. Force (Tonne) Figure 5.14: Variable clamping forces on mould assembly for entire cycle from AMI analysis. 141

162 To calculate the injection pressure at different interface surfaces, the same procedure has been applied as that for calculating temperature values. Figure 5.15 shows the injection/hold-on pressure for a particular node point in that surface. For example, node point N2025, gives the injection pressure on outer side surface of the plastic part which is in contact with cavity surface. Similarly node points N16423, N1967 and N867 gives the injection pressure for a particular point on sprue hole surface, bottom surface and round surface respectively. Pressure (MPa) Figure 5.15: Variable injection pressure on mould different surface of plastic part/cavity/core for entire cycle from AMI analysis. 142

163 Table 5.7 shows all the data for average clamping force and injection pressure that have been obtained from the AMI simulation and used as boundary conditions for different surface in transient structural analysis. It should be noted that, injection pressure for core side and cavity side, for a particular surface have been taken as same, because there is not much difference in values, shown at nodes N2025 and N2024 in figure 5.15, which are the node points for cavity and core side surfaces respectively. Table 5.7: Values of clamping forces and injection pressure used as boundary conditions. Injection pressure Clamping force Bottom Surface (BS) Round Surface(RS) Side Surface (SS) Sprue hole Surface(SHS) Top Surface Time Pressure Time Pressure Time Pressure Time Pressure Time Forces (s) (MPa) (s) (MPa) (s) (MPa) (s) (MPa) (s) (N) e e e e e e

164 5.6 Results of Finite Element Analysis From transient thermal analysis, temperature distribution has been found for entire mould. Figure 5.16(a), 5.17(a) & (b) and 5.18 (a) & (b), show the comparative temperature distribution for all five cases of cooling channel moulds. In case of CSCC, after one cycle (20 second), temperature of the mould ranges from a minimum 13 C to maximum 74 C (Figure 5.16(a)), whereas, for BSCC and BCCC CTI, average minimum to maximum temperature ranges from 16 C to 64 C (Figure 5.17 & 5.18). So by using CTI, temperature reductions by 10 C could be possible, which ultimately reduces cycle time. Now, among the bi-metallic types, both the BSCC and BCCC with 2mm and 3mm thickness, show almost similar results. But as copper is weaker than tool steel in strength, it is necessary to check the robustness of the moulds with CTI. Therefore, transient thermal-structural analysis has been carried out to verify whether weaker metal insertion may cause fatigue failure of the mould, before the required number of cycles it can operate. (a) (b) Figure 5.16: (a) Temperature distribution after one cycle in the mould, and (b) Equivalent stress distribution at 4.45th second of cycle time for CSCC mould. 144

165 (a) (b) Figure 5.17: Temperature distribution on mould assembly after one cycle (20 second) (a) (b) Figure 5.18: Temperature distribution on mould assembly after one cycle (20 second) (a) BCCC 2mm CTI, and (b) BSCC 2mm CTI. In order to determine the fatigue life, it is necessary to determine the maximum equivalent stress distribution for each case. Figure 5.16(b), 5.19 (a) & (b), and 5.20 (a) & (b) give the equivalent stress or von-mises stress (Ses) distribution in the mould at 4.45 th second of the cycle time, from transient thermal-structural FEA analysis, for all five cases of different cooling 145

166 configuration moulds. Figure 21 shows the complete stress vs time distribution graph for entire cycle for these five cases of moulds. It is noted that the peak of the maximum equivalent stress for all cases occur at approximately 4 th second of the cycle. Table 5.8 also gives the peak values of maximum von-mises stress during injection cycle for different moulds. (a) (b) Figure 5.19: Equivalent stress distribution at 4.45 second of cycle for (a) BSCC 3mm CTI and (b) BCCC 3mm CTI moulds. (a) (b) Figure 5.20: Equivalent stress distribution at 4.45 second of cycle for (a) BSCC 2 mm CTI and (b) BCCC 2 mm CTI moulds. 146

167 From figure 21 and Table 5.8, it can be shown that in case of CSCC, maximum equivalent stress or von-mises stress (Ses) is MPa, that mould experiences about 4 th second of the cycle time. On the other hand, for 2mm CTI, this value is around 650 MPa, for both BSCC and BCCC. For 3mm CTI, this values goes a bit higher, which is around 685 MPa for both BSCC and BCCC. 700 Maximum Equivalent Stress Vs Time Maximum Equivalent Stress(MPa) CSCC BCCC_3mm_CTI BSCC_3mm_CTI BCCC_2mm_CTI BSCC_2mm_CTI Time (second) Figure 5.21: Maximum equivalent stress for different cooling channels for entire cycle. These peak equivalent stress or von-mises stress values will be used to calculate fatigue life of the moulds with different cooling configuration which will be discussed in the next section of this chapter. 147

168 5.7 Calculation of Fatigue Life of Moulds In Chapter four, detailed discussion of metal fatigue and life cycle of the mould has been done with mathematical equations. Using maximum equivalent stress from thermal-structural FEA analysis and high cycle fatigue formula, which is given by equation (5.5), the fatigue life of the mould or number of cycle mould can operate before failure can be written as [157], N=(S es /a) 1/b (5.5) 10 3 <N<10 7 Where, a = (f*sut) 2 /Se and b = -(1/3)*log(f*Sut/Se) Ses= Equivalent stress or Von-Mises stress Sut = Ultimate tensile strength Se= Elastic strength f = Constant From Table 5.8 it can be noted that all these peak values of maximum equivalent stress in the mould are much less than the ultimate tensile strength (1780 MPa) of the Stavax Supreme mould, therefore, indicating structurally safe design of moulds with bi-metallic cooling channels. We will use these peak values of stress, to calculate the fatigue life of the moulds in each case, to find out which type of bi-metallic cooling channel will be the best from fatigue failure point of view. 148

169 Table 5.8: Values of maximum equivalent stresses (S es) and life cycle of the Values of moulds with different cooling channels. Type of cooling channel CSCC BCCC 2mm CTI BSCC 2mm CTI BCCC 3mm CTI BSCC 3mm CTI Von-Mises stress S es (MPa) Fatigue life N (Million cycle) Using the equivalent stress and high cycle fatigue formula (equation (5.5)), the fatigue life of the mould or the number of cycle (N) mould can operate before failure can be calculated. For Stavax Supreme, Sut= 1780 MPa, Se= 750 MPa. The values of f is normally taken as 0.9 [105] but it varies with Sut, and for Sut =1780 MPa, f=0.7. Applying Equation (5.5) and these values, the fatigue life of the mould with different cooling channels have been calculated and these values are also shown in Table 5.7 (Sample calculation is given Appendix A in section A4). Comparison of fatigue life (N) values in Table 5.8 shows that the bi-metallic conformal cooling channel (BCCC) with 2mm CTI, gives the highest fatigue life of million cycles, compared to all other cases. So, we will investigate further the cooling efficiency of the BCCC with 2mm CTI by analytical and experimental investigation as described later. 149

170 5.8 Calculation of Cooling Time of Mould In Chapter three, detailed discussion of theoretical cooling time calculation has been done. The same formulas will be used to calculate the theoretical cooling time and then compare it with the cooling time from the simulation results for this case study as well. So, using equation 3.15 (from chapter three), the theoretical cooling time can be written for this case study as, t C =. ln( ) (5.6) where, S= Thickness of plastic part=2mm α = thermal diffusivity of plastic material = ρ =8.87x10-8 m 2 /sec k= thermal conductivity of plastic material=0.14 w/m. c ρ= density of plastic material=830 kg/m 3 Cp= specific heat of plastic material=1900 J/kg. c Tm= moulding temperature of plastic=250 C Td=demoulding temperature of ejected plastic material =78 C Tw=Maximum temperature of interface wall between molten plastic and cavity wall after one cycle =74.32 C for CSCC. Using the plastic part material as Polypropylene (PP), theoretical cooling time for CSCC and BCCC 2mm CTI, have been calculated and compared with the cooling time obtained from simulation results, and are shown in Table 5.9. From the results shown in Table 5.9, we can see that there is good agreement between theoretical cooling time calculations and the cooling time from simulation result for both types of cooling channels. Table 5.9 also shows that using 2mm CTI, cooling time has been reduced by almost 3 seconds, thus providing faster cooling. Moreover, this bimetallic cooling channel also gives the highest fatigue life (as shown in Table 5.8). 150

171 Figure 5.22 shows the cooling curve for the plastic part for CSCC and BCCC 2mm CTI, for the entire cycle of the moulding process using the numerical analysis. The part reaches a maximum temperature in the mould and then cools down to the ejection temperature. For the same ejection temperature, the cooling curve for the BCCC shows much lower time (5 seconds) for ejection than the CSCC (8 seconds). Thus simulation and analytical results both confirm that BCCC provides a better robust design option for injection moulds. Table 5.9: Comparative cooling time data for CSCC and BCCC 2mm CTI. Type of cooling channel Theoretical Cooling time (Second) Average cooling time from simulation result (second) CSCC BCCC 2mm CTI Temperature Vs Time Temperature ( C) CSCC BSCC_2mm_CTI Time (second) Figure 5.22: Comparative cooling curve for CSCC and BSCC 2mm CTI. 151

172 5.9 Experimental Verification Experimental verification of the numerical analysis result has been carried out by injection moulding of a plastic part using two different plastics, PP and ABS, using conventional CSCC as well as BCCC moulds. The part was a disk shape with of a diameter 40 mm and thickness 7 mm. The mould was a square shape with overall dimensions of 100x100x25 mm. The mould material was mild steel. Experiments weree carried out on a mini injection moulding machine, TECHSOFT mini moulder. Figure 5.23(a) shows the core and cavity moulds with cooling channels. Figure 5.23(b) shows the copper tube insert (CTI) being fitted. Figure 5.23(c) shows the injection moulded parts in PP and ABS produced during the experiment. (b) (a) (c) Figure 5.23: (a) Core and cavity mould, (b) Copper tube is being inserted, and (c) shows the injection moulded parts in PP and ABS produced during the experiment. 152

173 Figure 5.24: Experimental set up with mini moulder. Two thermocouples TC08 K type of PICO technology, have been used to measure temperature of top and bottom interface of the test part and cavity for every second of the cycle. Melting temperature used was 250 C for both ABS and PP. Normal water has been used as a cooling medium, and water temperature has been measured as 20 C. Cooling channel diameter used was 6 mm for CSCC mould and CTI inner diameter was 6mm and CTI thickness was 1mm used in bi-metallic mould. Copper tube has been inserted inside the cooling channel by heating the mould and cooling down after press fit. In this process, the air gap between two materials has been minimized. Figure 5.24 shows the experimental set up with mini moulding machine. Figure 5.25 and 5.26 shows the comparative cooling time curves obtained from experimental results for CSCC and BCCC moulds for PP and ABS parts respectively. Cooling curves were obtained from the maximum temperature of the top and bottom thermocouple readings up to next 20 seconds. Both figures clearly show that for both types of plastic parts, using bi-metallic cooling 153

174 channel, the plastic part can be cooled down much faster than using the CSCC. In average, 8-10 C temperature and 3-5 seconds of cooling time can be reduced using the bi-metallic cooling channel moulds. 105 Cooling curve for PP Temperature C BCCC_BOT_SURFACE BCCC_TOP_SURFACE SCC_BOT_SURFACE SCC_TOP_SURFACE Time (second) Figure 5.25: Comparative cooling curve for CSCC and BCCC CTI moulds, showing temperature at the bottom and top surface of the Polypropylene (PP) plastic part. 154

175 140 Cooling curve for ABS Temperature C BCCC_TOP_SURFACE BCCC_BOT_SURFACE SCCC_TOP_SURFACE SCCC_BOT_SURFACE Time (second) Figure 5.26: Comparative cooling curve for CSCC and BCCC CTI moulds, showing temperature at the bottom and top surface of the Acrylonitrile Butadiene Styrene (ABS) plastic part Summary From experimental and thermal-structural FEA, it can be concluded that high thermal conductive copper tube can be a potential alternative to replace conventional straight cooling channel in injection moulding, as it reduces significant amount of cooling time (35%), as well as it increases fatigue life time of the mould. But it is also notable that, increasing too much thickness of the copper tube will also reduce the structural strength and fatigue life of the mould. So, the use of bi-metallic cooling channel with copper tube, should be in such a way that, it does not reduce the overall strength of the mould or increase the total life cycle of the mould before its failure. 155

176 6 Chapter 6 Fabrication of conformal cooling channels by Rapid Manufacturing 6.1 Introduction While conformal cooling system provides great benefits in terms of cycle time reduction and part quality, it is also true that such complex cooling system can not be produced by conventional machining process such as drilling. In this chapter, a detailed discussion on the feasibility of fabricating conformal cooling channels using Direct Metal Deposition (DMD) based rapid manufacturing (RM) tooling is presented. Experimental study of making square shape conformal cooling channel (SSCCC) by DMD is specifically described. Mould with SSCCC will then be compared with conventional straight cooling channel (CSCC) mould by injection moulding of a sample plastic part using two different types of plastic materials. Also comparison will be done between results from experimental moulding and thermal Finite Element Analysis (FEA) based on temperature distribution on the cavity surface. 6.2 Rapid Manufacturing of Injection Moulds To develop a new product, it is very important that it does fulfil the requirement of time and cost to manufacture it and bring into the market. Likewise, it is extremely important for the injection moulding industry to produce prototype that can go into production as quickly as possible. Since late Nineties, rapid prototyping has been used as a very efficient way of making prototypes. But using such prototyping techniques to make moulds directly cannot fully meet the requirements of injection moulding process to go into production quickly. Rapid Tooling (RT) and Rapid Manufacturing (RM) techniques have been developed recently by which metal moulds can be made 156

177 directly using layer by layer additive manufacturing. Such systems, even though they are expensive and have limitations, will eventually be accepted by industry for making tools and moulds rapidly for customised or mass production applications. In chapters 1 & 2, benefits of using RM and related research works have been discussed. In our research work, we investigate making square shape conformal cooling channels using DMD rapid manufacturing technique. 6.3 Direct Metal Deposition (DMD) Various RP & RM techniques that have been published in different research papers and books have been compared and summarised briefly in Table 2.1 and 2.2 (chapter 2). Amongst these techniques, only a few of them have the capabilities of producing metal parts directly from the CAD model with near 100% density and functional properties. Most of the RM processes are powder based techniques, which require infiltration and densification. The DMD process is one of the few RM processes, which offers the possibility of fabricating a 100% dense-metal component with functional properties, which can be used directly for making injection moulds. Also discussed in chapter one, in DMD process, a high power laser beam generates a melt-pool on a substrate material while the powder material is delivered into the melt-pool and forms a metallurgical bond with the substrate as shown in figure 1.3 (In chapter 1). Figure 6.1 (adopted from [87]) shows schematic of laser aided DMD system which consists of four key components, which are: 1. Laser cladding 2. CAD/CAM 3. Numerical control (NC) of machine tool, and 4. Feedback control 157

178 Figure 6.1: Schematic of components in DMD system, based on [87]. To fabricate a part, a CAD model is generated using a CAD/CAM system and then model is sliced with uniform thickness and then a machine tool path is generated (similar with CNC machining process) based on the slice. The machine tool path and the auxiliary functions are then uploaded into the computer the machine. A feedback control unit controls the deposition layer as well as deposition integrity. Figure 6.2 (adopted from [87]) shows the feedback control system, which consists of height sensing unit and feedback signal processing unit. An optical photo-sensor is developed for the height. One can either use one sensor or multiple sensors for closed-loop feedback control of the deposited layers. Multiple sensors will overcome any problem related to the field of view with respect to the cladding direction for a single sensor. 158

179 Figure 6.2: Schematic of feedback control system in DMD process, based on [87]. Before using the optical sensor, it is necessary to calibrate it based on the desired image of melt-pool. The image of the melt-pool is sent onto an optical photo sensor, which reflects the height change of the melt-pool on the deposited surface. When the height of the current melt-pool is higher than desired position, the sensor catches the image of the melt-pool. Then, a proper signal is sent out to signal processing unit and laser power controller. The signal processing unit also sends signal to the powder delivery system. The powder delivery system is designed to mix multiple powders and deliver uniform mass flow. Multiple powders delivery system consists of three powder containers which are a powder container, a concentric nozzle assembly and delivery gas inlets. Each hopper in the powder container has a dimpled-shaft delivery system and a motor assembly, which is driven by a servo-controller. The motor speed is controlled by a signal from the signal processing unit, so that the powder mixing and the mass flow rate can be precisely controlled and stabilized. 159

180 6.4 Design of Mould with Conformal Cooling Channel The mould to be fabricated with cooling channels is a square shape with overall dimensions of 100x100x25 mm. Two moulds, one with a conventional straight cooling channel (CSCC) and the other with a square shape conformal cooling channel (SSCCC) are fabricated in this study. The core and cavity parts of the mould have been shown in figure 6.3. The core has been made with conventional CNC (computer numerical control) machining process, which has a sprue hole on the top through bottom as shown in figure 6.3. Cavity part has CSCC, for conventional cooling channel mould, as shown in figure 6.4 (a). Cavity of the SSCCC has two parts, as shown in figure 6.4 (b). The base part, which works as a substrate for DMD deposition, has been manufactured by CNC machining, and the top part is fabricated by laser direct metal deposition (DMD) technique, containing SSCCC and cavity hole. Diameter of the CSCC is 6 mm and the total surface areas is 5261 mm 2, and the sectional dimension of the SSCCC is 6x6mm, and total surface areas is 5217 mm 2. The cavity hole for the plastic part is a disk shape with of a diameter 40 mm and thickness 7 mm. Cavity Core Sprue hole Figure 6.3: CAD model of Core and Cavity moulds designed by Pro/Engineer software. 160

181 Conventional straight cooling channels (CSCC) DMD fabricated part of cavity mould (a) Square shape conformal cooling channels (SSCCC) CNC manufactured part (base part) of cavity mould (b) Figure 6.4: CAD model of (a) Cavity mould with CSCC, and (b) Cavity mould with SSCCC, designed by Pro/Engineer software. 161

182 6.5 Fabrication of Mould with Conformal Cooling Channel by Direct Metal Deposition (DMD) Top section of the SSCCC cavity mould has been fabricated by the DMD fabrication process. As fabrication of DMD is a costly process, only the top section of the cavity mould containing SSCCC (as shown in figure 6.4(b)) has been fabricated by DMD, rather than making the whole cavity mould by DMD. In this process, optimization of cost by fabricating mould in DMD process has been achieved. Dimensions of the DMD deposited area was 86x86x13mm. Thickness and width of each layer of the deposition were 0.5 mm and 1.5 mm respectively, and overlapping between two layers was 0.75 mm as shown in figure 6.5. Total of 26 layers have been deposited to build up the top part of the SSCCC mould. Figure 6.5: Dimensions of DMD layers. The DMD fabricated areas of the cavity contains cavity hole as well as SSCCC, two different building patterns have been used as shown in figure 6.6, to fabricate it. Circular building pattern was used for cooling channels and cavity hole areas fabrication, and straight pattern was used to fabricate rest of the areas. Figure 6.7 shows a DMD 505 machine installed at Swinburne University of Technology and supplied by Precision Optical Manufacturing (POM) Inc, USA [22]. The machine has been used to deposit H13 tool steel powder on mild steel substrate (base of the cavity mould as shown in figure 6.4 (b)). H13 steel 162

183 Substrate Circular building pattern Straight building pattern Figure 6.6: Building pattern of DMD layers. powder was supplied by Sulzer Metco (Australia) Pty Ltd and mild steel was supplied by local manufacturer. Properties and chemical compositions of mild steel and H13 tool steel are given in Table 6.1[158]. Table 6.1: Composition and properties of H13 and Mild steel [158]. Material Chemical Rockwell Thermal Thermal Expansion Tensile composition Hardness conductivity Coefficient Strength (%) (HRC) (W/m* C) (10-6 / C) (MPa) H13 Tool Steel Mild steel (AISI 1029) C-Mn-Si-Cr-Mo- V-Fe ( Balance) C-Mn-P-S

184 Figure 6.7: POM DMD 505 machine at Swinburne University of Technology. DMD process parameters remained the same during the whole fabrication process. The power of the laser beam was 1500 watts, which produced a bead of 1.5 mm width and 0.5 mm height and each bead overlapped 50% with the previous one at every shot (as shown in figure 6.5). The size of the steel powder particles was micrometer. Time required to make each layer was 30 minutes. Total time of 15 hours approximately was taken to complete the full deposition process including the cool down time of the machine at every 5 layers for 20 minutes. Figure 6.8 (a) & (b) shows the fabrication process of cavity mould with SSCCC by DMD and the complete fabricated mould respectively. 164

185 (a) (b) Figure 6.8:(a) Fabrication of cavity mould with SSCCC by DMD process, and (b) Complete cavity mould with SSCCC by DMD process. 6.6 Optical Microscopic Study of DMD Fabricated Mould To check how H13 powder has been deposited by the DMD, the cavity mould has been cut crosswise, as shown in figure 6.9 (a) to observe the microstructure and crack formation between the interfaces of two layers. The sample, as shown in figure 6.9 (b) has been mounted using cold mounting technique. Initially the sample has been gradually polished with 320, 600, 1200 number (grit size) silicon carbide paper, then finally sequentially polished with 15, 6 and 1 µm diamond paste using polishing clothe. Then the sample has been checked under the optical microscope (LEICA MEF4M), and image has been taken as shown in figure 6.10, which is 10 times magnified. From figure 6.10, we can see that H13 deposition on the mild steel substrate look structurally sound as there is no crack formation between the interfaces. 165

186 SSCCC H13 DMD fabricated area Mild steel area (a) (b) Figure 6.9: (a) Cross section of cavity mould showing SSCCC and interface area between H13 DMD fabrication and mild steel (b) Mounting of interface area for optical microscopic image testing. H13 DMD fabricated area Mild steel area Figure 6.10: Optical microscopic image of interface of H13 deposition on mild steel. 166

187 6.7 Experimental Verification of Conformal Cooling. Experimental verification has been carried out by injection moulding of a plastic part using two different plastics, PP and ABS, using the CSCC as well as SSCCC moulds. Test part volume was 8.8 cm3, and part weight for ABS and PP were 8.68 g and 8.13 g respectively. Experiments were carried out on a mini injection moulding machine, as shown in figure 6.11(a), TECHSOFT mini molder (also described chapter 5, experimental verification of bi-metallic conformal cooling channel). Two thermocouples, TC08 K type of PICO technology, have been used to measure temperature of top and bottom interface (as shown in figure 6.11 (b) & (c)) of the test part and cavity, for every second of the cycle. All together 10 parts have been produced for PP and ABS each, and (b) ( a) (c) Figure 6.11:(a) Experimental moulding setup with mini-injection molder, (b) Top of core showing the position of thermocouple with red circle, (c) Bottom of cavity showing position of thermocouple with red circle. 167

188 for each case, temperature readings of top and bottom interfaces have been recorded. Average of these temperature values have been plotted, which are discussed in the next section. 6.8 Result and Discussion Figure 6.12 and 6.13 shows the comparative cooling time curves obtained from experimental results with CSCC and SSCCC moulds for PP and ABS plastic parts respectively. Cooling curves were obtained from the maximum temperature of the top and bottom thermocouple readings up to next 20 seconds. Both figures clearly show that for both types of plastic parts, using the SSCCC mould, the plastic part can be cooled down much faster than using the CSCC mould. In average, 4-5 C temperature and approximately 3 seconds of cooling time can be reduced using the SSCCC. Cooling curve for PP Temperature( C) Conven_TS Conven_BS Conformal_TS Conformal_BS Time (second) Figure 6.12: Temperature reading for top and bottom interface of PP plastic part with conventional(cscc) and conformal (SSCCC) moulds. 168

189 Cooling curve for ABS Temperature ( C) Conven_BS Conven_TS Conformal_TS Conformal_BS Time (second) Figure 6.13: Temperature reading for top and bottom interface of ABS plastic part with conventional (CSCC) and conformal (SSCCC) moulds. Experimental results obtained above have been compared with the thermal finite element analysis (FEA) simulation result, which has been carried out by ANSYS workbench simulation software. In chapter 3, 4 & 5, detailed discussion of thermal FEA, has already been done. Separate plastic flow simulation (cool+flow+warp analysis) by Autodesk Moldflow Insight (AMI) software, has been carried out to get the interface temperature. Typical process parameter used for AMI simulation has been shown in figure 6.14 for the SSCCC mould. For this particular case of thermal FEA simulation, Reynolds Number(RN) of 6mm hydraulic diameter cooling channel has been calculated as 8090 and convective heat transfer coefficient (hc) as 6555 Watt/mm 2 C. Figures 6.15 and 6.16 show the result of temperature distribution in the moulds after 25 second form thermal FEA simulation for SSCCC and CSCC moulds respectively. Table 6.2 gives the comparative temperature values on the 169

190 Figure 6.14: Process parameter for AMI simulation for SSCCC mould. top and bottom surface of the plastic part (interface surface) for PP and ABS after 25 seconds, which has been recorded from the experimental moulding of the plastic parts and thermal FEA simulation. From Table 6.2, we can see that there is good agreement between temperature reading from experimental result and thermal FEA simulation results. Both experimental and simulation result shows that using the SSCCC, the temperatures can be reduced faster than using the CSCC, which ultimately will reduce the cycle time of the plastic part. Thus verification further shows that the conformal cooling channels with square cross sectional shape reduce cycle time of the moulding process, which eventually will result in increase production rate. In the next chapter, it will be described how SSCCC also improves the quality of plastic part. 170

191 Figure 6.15: Temperature distribution in the cavity sectioned mould with SSCCC from thermal FEA simulation. Figure 6.16: Temperature distribution in the cavity sectioned mould with CSCC from thermal FEA simulation. 171

192 Table 6.2: Comparison of experimental and thermal FEA result of temperature distribution at the mould surface for CSCC an SSCCC cooling channel for PP and ABS plastics Temperature values after 25 second ( C) Type of cooling channel Experimental result Thermal FEA simulation result PP ABS PP ABS Top Bottom Top Bottom Top Bottom Top Bottom surface surface surface surface surface surface surface surface CSCC SSCCC

193 7 Chapter 7 Conformal cooling effect on quality of plastic part 7.1 Introduction Nowadays, the need for plastic products with specific properties, low cost and fast production is more important than ever. This is especially true for injection moulding process which is used for mass production of large varieties of plastic products with varieties of polymers. But, these injection moulded plastic parts often have some undesired defects such as sink marks, differential shrinkage, thermal residual stress built-up, as well as part warpage [9]. One of the main reasons for this is poor cooling system design. These defects can be relieved or even eliminated through a proper arrangement of cooling channels and the best combination of process conditions. This chapter looks at the effect of conformal cooling system on plastic parts quality based on simulation and experimental studies. 7.2 Effect of Cooling Channel on Plastic Part Quality The cooling system of the injection mould is significant not only for cycle time but also warpage, shrinkage and residual stresses issues. The mould can be thought of as a heat exchanger, where heat from the plastic part is transferred mostly through the coolant medium. The mould should be designed in such a way that it can minimize the temperature distribution across the mould surface. This can be accomplished by proper design of cooling channels, and proper location inside the cavity and core mould surface. The more uniformly the heat energy can be extracted from the mould, the fewer cooling-related problems will occur. 173

194 Cooling problems in the injection moulded plastic part can be addressed in a variety of ways. One of main problems is volumetric shrinkage, caused by uneven cooling in the mould. When plastic material cools down from molten state to solid state, it shrinks from its original volume, and without any injection and hold on pressure during injection moulding, volumetric shrinkage can be up to 25 % of the part original volume [3]. However, studies shows have shown that injection/holding pressure adjustment can only reduce volumetric shrinkage by about half [3]. In injection moulding process (IMP), application of pressure and balanced cooling system can also reduce the volumetric shrinkage but no means can be fully eliminate it. Another problem in the plastic part during IMP is warpage, which mainly occurs due to uneven shrinkage of the plastic part during solidifying process of the plastic part. And one of the main reasons of uneven shrinkage in the plastic part is attributed to unbalanced cooling system in the mould. Figure 7.1 [159] shows a comparative diagram of mould temperature with two different types of cooling channels. It shows that the modified and more uniform cooling channels cooled the part better as well as produced a better heat balance in the mould. Figure 7.1: Mould temperature results of the original (left) and modified (right) cooling system [159]. 174

195 Figure 7.2 [159] shows how the warpage deflection can also be improved (reduced) by the same modified cooling channels than the original for the same part. Figure 7.2: Comparative warpage deflection results of the original (left) and modified (right) cooling system design [159]. Changing the coolant temperature can reduce the warpage but it often leaves the part with higher residual stresses. The residual stresses can significantly affect the mechanical performance of an injection moulded part by inducing warpage or initiating cracks. Normally, the thermally induced residual stresses are dominant in parts and can cause parts with tension in the central position and compression on the surface of the part. In practice, the complexity of part geometry and cooling channel design can result in unbalanced mould wall temperature distribution, which in turn leads to the non-uniform cooling and build-up thermal residual stress in parts. An accurate calculation of the thermal stresses will allow improvement of mould cooling design and evaluation of potential part warpage. A linear elastic theory modelling with initial and boundary conditions and evaluation of thermal residual stress based on the transient cooling analysis has been done by Sun et al. [13]. They described how 175

196 the milled groove insert cooling method can reduce the residual stress in the plastic part. 7.3 Effect of Cooling on Crystalline Materials Cooling channel design also affects the crystalline content of plastic material. In the molten stage, all plastics are amorphous. In the solidification process of IMP they are cooled down, and crystallize gradually back to their original structure. However, crystallization requires time for the crystal to grow. If cooling is very fast and efficient, the heat is removed so fast that the resin does not get enough time to crystallize, and the material, even though it is crystalline, will remain in an amorphous state even after being cooled. On the other hand, if cooling is uniform and efficient, plastic gets sufficient time to crystallize before cooling and will again become crystalline in structure. Normally, areas in between the part centre and mould surfaces may result in smaller size crystal than the areas farther from the centre of the plastic part, as shown in figure 7.3 (adopted from [129]). Figure 7.3: Variations in crystallization through cross section of the plastic wall and mould in injection moulding process [129]. 176

197 Amorphus structures with small crystallinity improve impact strength, whereas, crystalline structure with large crystal improve compressive strength. Normally, in an industrial plastic part, both of these characteristics are required, and these qualities of plastic part can be only be achieved by adjusting the moulding process and providing uniform and optimised cooling channel design. Next sections of the chapter will describe how conformal cooling channel can improve such properties of plastic part based on flow simulation studies. An experimental verification of injection moulding with conformal and conventional cooling channel is also described, which shows how the hardness values of plastic parts is affected by different cooling channels, due to effect of cooling on crystallisation of plastics. 7.4 Conformal Cooling Channels Affecting the Part Quality Conformal cooling channel has significant effects on quality of injection moulded plastic part. The conformal cooling method uses contour-like channels of different cross-section, constructed as close as possible to the surface of the mould to increase the heat absorption away from the molten plastic. This ensures that the part is cooled uniformly as well as more efficiently (as described in chapter three). It has been discussed in chapter two (literature review) that there have been some research works accomplished which indicated that conformal cooling channel can be one of the solutions to improve quality of plastic part. With appropriate cross sectional shape, position from the cavity surface and diameter size, it is expected that conformal cooling channels can significantly reduce the warpage deflection, shrinkage and crystal structure of the injection moulded plastic part. 7.5 A Case Study with Conformal Cooling Channels To check how conformal cooling channel can improve quality of plastic part, a complete flow simulation, cool+fill+pack+warp analysis is carried out with 177

198 Autotodesk Moldflow Insight (AMI), which is injection moulding flow simulation software. An industrial plastic part (circular bowl) made of polypropylene (PP) thermoplastic has been considered for this study. The dimensions and design of the plastic part have been discussed in chapter three. The part has been exported to IGES (Initial Graphics Exchange Specification) surface model after designing with Pro/Engineer CAD software, and then imported in AMI. Material volume of the plastic part is cm 3 and its weight is 163 gm Conformal Cooling Channel Design and Plastic Flow Analysis with Autodesk Moldflow Insight Two different cross sectional types of conformal cooling channels, which are circular and square, have been designed for these analyses. From chapter three and four we have learned that these two cross sectional shapes give best result for conformal cooling channel design in terms of life time of the moulds as well as cycle time reduction. Figure 7.4 (a) & (b) show the circular and square cross sectional cooling channels design in AMI interface. Dimensions of the cooling channel section was mm and 10.10x10.10 mm for circular and square channels respectively. Distance from the cavity surface of the cooling channel has been taken as 2Dh= These dimensions of the conformal cooling channel have been considered as a result of the best thermal structural analysis described in chapter four. Autodesk Moldflow Insight (AMI) software has been used to carry out complete injection moulding flow analysis, cool+fill+pack+warp. AMI software can reliably simulate plastics flow and packing, mould cooling, and part shrinkage and warpage analysis for thermoplastic injection moulding, as well as other injection moulding process i.e. gas-assisted injection moulding, coinjection moulding and injection-compression moulding processes [ ]. Analysis with AMI software can be used in the very early stages of Mold design to avoid costly mold rework and minimize the delays associated with taking 178

199 moulds out of production. The software also lets one easily optimize gates, runners, cavity layouts, and other mold features and test out new ones and to (a) (b) Figure 7.4: Conformal cooling channels design in AMI simulation (a) circular cross section, (b) Square cross section. see how changes to wall thickness, gate location, material, and geometry affect manufacturability. This comprehensive suite of injection molding simulation tool can be used to analyze and optimize all phases of the injection molding process. AMI even includes tools for simulating special molding processes [159, 163]. Flow chart of the processes, which have been used for this simulation with AMI, is shown in figure 7.5. For simulation with AMI, dual domain meshing have been used, which is recommended by Moldflow for plastic part with medium thickness (i.e. 2 mm). Total number of elements was with global edge length of 9.95 mm. Mould and melt temperature of plastic have been used 179

200 as 50 C and 250 C respectively. Pure water has been used as coolant with a temperature of 10 C. Reynolds Number (RN) for mm hydraulic diameter of the cooling channels was calculated as Figure 7.5: Flow chart of flow simulation with AMI Result and Discussion of Plastic Flow Analysis Results of the flow simulation (cool+fill+pack+warp analysis) for different cooling channels with AMI have been described with various flow simulation diagrams. From figure 7.6(a) & (b), it can be shown that in case of conformal cooling channels with circular and square cross section, time required to reach ejection temperature or cooling time, ranges from approximately 6 to 53 seconds and 5 to 79 second respectively. And most of the areas of the plastic part have been cooled down to the ejection temperature by approximately 14 and 16 seconds for circular and square cross sectional conformal cooling channels respectively, except the few areas at the top corners. 180

201 (a) (b) Figure 7.6: Time required to reach ejection temperature or cooling time of the plastic part for (a) Circular, and (b) square, cross sectional conformal cooling channels. On the other hand, in case of conventional cooling channels, time required to reach ejection temperature, as shown in figure 7.7, ranges from Figure 7.7: Time required to reach ejection temperature or cooling time of the plastic part using conventional cooling. 181

202 5 to 83 seconds. And most of the areas of plastic part is cooled down by 22 seconds. So, using conformal cooling channels, significant amount of cooling time reduction, which is up to 36% of cooling time, can be possible. From figure 7.8 (a) & (b), we can see that using conformal cooling channels with circular and square cross section, warpage deflection in all direction ranges approximately from 0.6 to 3.4 mm and 0.2 to 3.5 mm respectively. And in both cases of cooling channels, average warpage deflection is not more than 1.5 mm, whereas, using conventional cooling channels, warpage deflection in all direction, as shown in figure 7. 9 (a) ranges from 0.23 mm to 3.63 mm. So, using conformal cooling channels of both circular and square cross section, warpage deflection has been reduced. (a) (b) Figure 7.8: Warpage deflection of the plastic part in all direction for (a) Circular and (b) Square cross sectional conformal cooling channels. 182

203 (a) (b) Figure 7.9: (a) Warpage deflection of the plastic part in all direction, and (b) Percentage of volumetric shrinkage with the original volume of the plastic part, using conventional cooling. From figure 7.10 (a) & (b), it can be shown that using conformal cooling channels with circular and square cross section, percentage of volumetric shrinkage with the original volume of the plastic part ranges from approximately 2 to % and 0.7 to % respectively. Thus in case of square cross section, the average values is within 9.6 %, whereas for circular cross section, it is within 10.5 %. On the other hand, for conventional cooling channels, the values of percentage of volumetric shrinkage with the original volume of the plastic part ranges from approximately 0.73 to %, as shown in figure 7.9 (b). And the average value is almost same with the values of conformal circular cross sectional cooling channel, which is within 10.5 %. So, in case of volumetric shrinkage, only square cross sectional conformal cooling channel shows better result than conventional cooling channels. Table 7.1 shows the comparative results summary for all cooling channels from the flow simulation by AMI software. 183

204 (a) (b) Figure 7.10: Percentage of volumetric shrinkage with the original volume of the plastic part, using (a) Circular and (b) Square cross sectional conformal cooling channels. Table 7.1: Comparative flow simulation results for different cooling channels. Properties of plastic part Conventional Cooling channel Conformal cooling channels with circular cross-section Conformal cooling channels with square cross-section Time to freeze (second) Range 5-83 Average 22 Range 6 53 Average 14 Range 5-79 Average 16 Warpage Range Range Range deflection Average to 3.5 (mm) 1.8 Average Average Total Range Range Range Volumetric shrinkage Average Average Average (%)

205 7.6 Experimental Study of Hardness Variation Due to Conformal cooling Flow simulation studies using AMI does not provide how hardness of injection moulded part is affected by cooling system design. An experimental study is presented here to show how the hardness values can be affected by cooling channel cross section of various cooling system design. Comparative experimental study of hardness number of the plastic parts that have been made with conformal cooling channel as well as conventional straight cooling channel has been carried out. Injection moulding has been done with two types of plastic materials, PP and ABS. In chapter five and six, dimensions of the experimental sample part and details of experimental moulding procedure with mini injection moulding machine have been discussed. Hardness of the test plastic part has been measured with the Shore D (Durometer) hardness testing machine of Zwick/Roell. The Shore hardness is measured with an apparatus known as a Durometer and consequently is also known as Durometer hardness. The hardness value is determined by the penetration of the Durometer indenter foot into the sample, as shown in figure 7.11(a). Hardness has been measured in twenty different points of top and bottom surfaces of the sample plastic part as shown in figure 7.11(b). In this experiment, hardness has been measured for 5 different test parts of each ABS and PP material, and average values of these readings have been plotted. From figure 7.12, it can be shown that for ABS plastic, measured hardness number for all different points of the plastic part which is made by square cross sectional conformal cooling channels (SSCCC) mould, is around 70, and almost uniform throughout the part. The uniformity of the hardness number of the plastic part is due of the fact that SSCCC helps plastic material to cool down uniformly. Because this cooling channel maintains equidistance from the cavity surface, it helps plastic part to cool down uniformly during the solidification process of plastic material in injection moulding process. On the other hand, in case of conventional straight cooling channel (CSCC), measured hardness number is not uniform throughout the part, which varies from 45 to 72. The 185

206 reason for this is, CSCC do not maintain uniform distance from the cavity surface, and as a result plastic part does not uniformly solidify or cool down during its crystallization process from molten state to solid state. (a) (b) Figure 7.11: (a) Measuring the hardness with Shore D hardness testing machine, (b) Points at which hardness has been measured, top surface of ABS (top), bottom surface of PP (bottom). 186

207 Hardness no(shored) Hardness for ABS Conventional straight cooling channel(cscc) Square cross sectional conformal cooling channel(ssccc) Different points of test part Figure 7.12: Comparative hardness plot in different points of parts that have been produced by SCC and CCC for ABS. From figure 7.13, similar result can be concluded in case of PP. For SSCCC, hardness number has been measured around 50 and for CSCC it is in the range of

208 60 Hardness for PP Hardness no(shored) Conventional straight cooling channel(cscc) Square cross sectional conformal cooling channel(ssccc) Different points of plastic part Figure 7.13: Comparative hardness plot in different points of parts that have been produced by SCC and CCC for PP. Finally, from flow simulation result and experimental study, it can be concluded that conformal cooling channel with circular cross section gives better result of cooling time and warpage deflection than conventional cooling channel. But conformal cooling channel with square cross section gives better results in all three plastic properties of cooling time, warpage deflection and volumetric shrinkage than the conventional cooling channel though it shows slightly higher cycle time than circular conformal cooling channel. So, square shape conformal cooling channel gives the best result for flow simulation as well as for hardness test. 188

209 8 Chapter 8 Conclusion and Recommendation 8.1 Introduction Plastic products are becoming more and more dominant and useful for our daily life than ever. Injection moulding is one of the most influential manufacturing processes of making plastic products. Like other products the part quality and the productivity are two of the most important issues in the injection moulding industry. These two key issues are mostly dependent on the mould cooling phase of injection moulding cycle. Proper cooling system design is very important for balanced cooling of injection moulding process, but normally it is not highly thought of as it should be before designing a mould for injection moulding process. The reasons are various. Firstly, it is difficult for an experienced mould designer to design efficient and uniform cooling systems without the help of CAE tools. Secondly, existing CAE tools are not fully ready for an optimal cooling design. Finally, the requirements of lead-time and cost constrain the effort to design and manufacture optimal cooling systems. Therefore, the cooling systems are normally designed at the last stage using the conventional straight drilling method, which is simple, low-cost, and widely applicable though it is not efficient and does not offer uniform cooling. However, with the advent of CAD/CAM/CAE, Rapid Manufacturing (RM) and Rapid tooling (RT) techniques, significant improvement of cooling system design can be achieved, especially designing complex mould with conformal cooling and manufacturing it with RT techniques. But still a lot of research need to be done for perfectly designing and manufacturing of mould with optimum cooling system and mass production of plastic parts with it. 189

210 This research work has been carried out to fill up some of the outstanding gaps of designing and manufacturing conformal cooling system in injection moulding. 8.2 Major Findings and Contributions In this research work conformal cooling channel with five different cross sectional shapes have been investigated with transient thermal Finite Element Analysis (FEA) for a circular shape industrial plastic part mould. Due to the large size of problems and considering the cost of the experiment, the investigations were performed using commercial FEA package, ANSYS Workbench simulation software. From the thermal FEA, it has been found that circular and square cross sectional shapes of conformal cooling channels give best result in terms of temperature distribution in the mould. These two cross-sectional shaped cooling channels are then further investigated with the same plastic part mould for four different diameters and three different positions from the cavity surfaces to check the robustness and longevity of the mould with different conformal cooling channels. This investigation has been carried out with transient thermal structural FEA by the ANSYS Workbench simulation software. Using transient thermal structural FEA results and high cycle fatigue formula, life cycles of moulds with different conformal cooling channels have been calculated. Life cycle calculation shows that conformal cooling channel with circular cross sectional shapes gives the highest life cycles (in average 10 million cycles). Conformal cooling channels with square cross sectional shape gives a maximum 2 million cycles even though it gives better temperature distribution in the mould than the circular cross sectional conformal cooling channels. A complete flow simulation, cool+fill+pack+warp analysis has been carried out with injection moulding simulation software Autodesk Moldflow Insight (AMI) to check how these conformal cooling channels affect the plastic part quality. 190

211 Simulation results with AMI show that both square and circular cross sectional conformal cooling channels show better cycle time, warpage deflection and volumetric shrinkage than the conventional straight cooling channels. Amongst the conformal cooling channels, square cross sectional shape gives better result than the circular shape though it shows a little better result in case of cycle time of the plastic part. In conclusion from the simulation results, it has been proved that conformal cooling channels with square cross sectional shape gives best result than that of circular cross sectional shape other than structural integrity of the mould, when it is necessary to produce very high volume of plastic parts, which is more that 2 millions cycles. An experimental mould fabrication with square shape conformal cooling channel by Rapid Manufacturing (RM) technique of Direct Metal Deposition (DMD) has been performed to justify that this complex conformal cooling channel can also be fabricated, while it cannot be manufactured by conventional machining process. Comparative experimental injection moulding has been carried out by producing a plastic part with two different types of plastic materials by using conventional straight drilled cooling channel mould and square shape conformal cooling channel mould fabricated by DMD. Comparative experimental result shows that square shape conformal cooling channel gives better temperature distribution in the mould than the conventional straight drilled cooling channel. An experimental verification has also been performed to test the hardness number of plastic parts by Shore D hardness testing machine to observe the effect of cooling channel on the crystallinity of the plastic part and hence their quality. Experimental harness testing results shows that plastic part that has been produced by square shape conformal cooling channel mould, shows uniform hardness number throughout the plastic part than that of the plastic part made by conventional cooling channel mould. Thus, simulation and experimental results verify that conformal cooling channel with square cross sectional shape can be a potential alternative to the 191

212 conventional straight cooling channel for reducing cycle time and improving part quality in injection moulding industries. This research work also presents a bi-metallic cooling channel with high thermally conductive copper tube insert for flat surface plastic parts for which conformal cooling channel may not be necessary to improve cycle time. Both bimetallic straight and conformal cooling channel have been investigated with transient thermal structural FEA by ANSYS Workbench simulation software for an industrial plastic part of rectangular shape. Experimental verification has also been carried out for this by making mould with bi-metallic conformal cooling channel and then injection moulding a plastic part with two different types of plastic materials. Both simulation and experimental results verify that bi-metallic cooling channel can extract faster heat from the molten plastic part, which ultimately reduces the cycle time more than the conventional straight cooling channel. Fatigue life cycle calculation also justifies that weaker copper tube insertion as a cooling channel will not reduce the life cycle of the mould; rather it increases production rate because of faster heat reduction during injection moulding process. 8.3 Recommendation for Future Work In this research work, major portion of the work has been performed using CAD/CAM/CAE software. Pro/Engineer package has been used for design purpose, ANSYS Workbench FEA simulation package has been used for transient thermal structural FEA analysis and Autodesk Moldflow Insight (AMI) software has been used for flow simulation in injection moulding process. Using these softwares reduce lots of time and cost than making expensive mould or using experimental verification. Though these softwares cannot give 100% accurate results but these are very reliable softwares to use for injection moulding cooling simulation process. Arguably, it would have been better if experimental verification could be done for every case. But it would be very costly process. 192

213 In our work, experimental verification has been done by fabricating small moulds of sizes 100x100x25 mm and producing plastic part with these moulds. Also DMD deposition has been done only on mild steel substrate to justify that complex square cross sectional shape conformal cooling channel can be possible to fabricate. Other substrate material i.e. tool steel, Moldmax could be investigated. Optical microscopic images of DMD deposition on mild steel substrate has been investigate, where no crack has been found. Thus, it arguably confirms structurally sound mould fabrication with square shape conformal cooling channel. But, no mass production of injection moulding has been carried out with square cross sectional conformal cooling channel mould. The present work has focussed on development of conformal cooling channels for single cavity mould with relatively common industrial part. Future work should include study of conformal cooling channels in multi-cavity moulds and for more complex industrial parts. By investigating a large variety of plastic parts, a comprehensive guideline can be produced for optimum size, shape and location of conformal cooling channels for single cavity and multicavity moulds. Work on conformal cooling study can also be extended to other specialised injection moulding process such as gas-assisted injection moulding, powder injection moulding, micro-injection moulding, micro-cellular moulding process, and even to other metal deformation processes such as die-casting. Fabrication of conformal cooling channels by various modern rapid manufacturing (RM) processes has also not received much attention. Future research should include a more comprehensive investigation on viability and process steps of speedy and direct manufacturing of injection moulds involving conformal cooling channels by RM techniques of electron beam melting (EBM), selective laser melting (SLM), laser engineered net shaping (LENS) and direct metal laser sintering (DMLS) processes. 193

214 References 1. Rosato, D.V.,Rosato, M.G., Injection moulding handbook. 3rd ed. 2003, Boston: Kluwer Academic Publishers. 2. Iyer, N.,Ramani, K., A study of localized shrinkage in injection molding with high thermal conductivity molds. Journal of Injection Molding Technology, (2): p Shoemaker, J., Moldflow Design Guide. 1st ed. 2006, Munich Hanser 4. Dimla, D.E., Camilotto, M.,Miani, F., Design and optimisation of conformal cooling channels in injection moulding tools. Journal of Materials Processing Technology, : p Wylinis, E.M., Production of injection moulding tooling with conformal cooling channels using three dimensional printing process. MS Thesis. 1995, USA: MIT. 6. Saifullah, A.B.M.,Masood, S.H., Optimum cooling channels design and thermal analysis of an injection moulded plastic part mould. Materials Science Forum, : p Menges, G., Michaeli, W.,Mohren, P., How to make injection mould. 3rd ed. 2000, Munich: Hanser Publishers. 8. Ho, J.K.L., Chu, K.F.,Mok, C.K., Minimizing manufacturing costs for thin injection molded plastic components. International Journal of Advanced Manufacturing Technology, (5-6): p Huang, M.C.,Tai, C., The effective factors in the warpage problem of an injection-molded part with a thin shell feature. Journal of Materials Processing Technology, (1): p Ozcelik, B.,Sonat, I., Warpage and structural analysis of thin shell plastic in the plastic injection molding. Materials and Design, (2): p

215 11. Xu, X., Sachs, E.,Allen, S., The design of conformal cooling channels in injection molding tooling. Polymer Engineering & Science, (7): p Sachs, E., Wylonis, E., Allen, S., Cima, M.,Guo, H., Production of injection molding tooling with conformal cooling channels using the Three Dimensional Printing process. Polymer Engineering and Science, (5): p Sun, Y.F., Lee, K.S.,Nee, A.Y.C., Design and FEM analysis of the milled groove insert method for cooling of plastic injection moulds. International Journal of Advanced Manufacturing Technology, (9-10): p Cha, B.S., Park, H.P.,Rhee, B.O. Manufacturing of plastic lens mold conformal cooling channel using direct metal laser sintering and spray-formed tooling process. Proceedings of ANTEC Cincinnati, OH. p Saifullah, A.B.M., Masood, S.H.,Sbarski, I., Finite Element Analysis of Bimetallic Conformal Cooling for Injection Mould. Submitted to Materials and Design.Oct Choi, J.,Hua, Y., Dimensional and material characteristics of direct deposited H13 tool steel by CO2 laser. Journal of Laser Applications, (4): p Fuh, J.Y.H., Zhang, Y.F., Nee, A.Y.C.,Fu, M.W., Computer aided injection mold design and manufacturing. 2006: Marcel Dekker, NY, USA. 18. Groover, P.,Zimmers, E.W., CAD/CAM: Computer-Aided Design and Manufacturing. 1984: Prentice/Hall Int. Inc USA. 19. Au, K.M.,Yu, K.M., A scaffolding architecture for conformal cooling design in rapid plastic injection moulding. International Journal of Advanced Manufacturing Technology, (5-6): p Shambley, W., From RP to AM: A review of prototyping methods., Metal Casting Design and Purchasing p

216 21. Ferreira, J.C.,Mateus, A., Studies of rapid soft tooling with conformal cooling channels for plastic injection moulding. Journal of Materials Processing Technology, (2): p POM, Precision Optical Manufacturing, The POM Group Inc Pontiac Rd. Auburn Hills MI 48326, USA. 23. Himasekhar, K., Lottey, J.,Wang, K.K., CAE of mold cooling in injection molding using a three-dimensional numerical simulation. Journal of Engineering for Industry, (2): p Chiang, H.H., Himasekhar, K., Santhanam, N.,Wang, K.K., Integrated simulation of fluid flow and heat transfer in injection molding for the prediction of shrinkage and warpage. Journal of Engineering Materials and Technology, Transactions of the ASME, (1): p Himasekhar, K., Hieber, C.A.,Wang, K.K. Computer-aided design software for cooling system in injection molding. Proceedings of ANTEC NY, USA. p Chang, R.Y.,Tsaur, B.D., Experimental and theoretical studies of shrinkage, warpage, and sink marks of crystalline polymer injection molded parts. Polymer Engineering & Science, (15): p Chang, R.-Y.,Chiou, S.-Y., A unified K-BKZ model for residual stress analysis of injection molded three-dimensional thin shapes. Polymer Engineering & Science, (22): p Chen, X., Lam, Y.C.,Li, D.Q., Analysis of thermal residual stress in plastic injection molding. Journal of Materials Processing Technology, (1): p Hu, S.Y., Cheng, N.T.,Chen, S.C., Effect of cooling system design and process parameters on cyclic variation of mold temperatures - simulation by DRBEM. Plastics, Rubber and Composites Processing and Applications, (4): p Chen, S.-C.,Hu, S.-Y., Simulations of cycle-averaged mold surface temperature in mold-cooling process by boundary element method. International Communications in Heat and Mass Transfer, (6): p

217 31. Rezayat, M.,Burton, T.E., Boundary-integral formulation for complex threedimensional geometries. International Journal for Numerical Methods in Engineering, (2): p Chen, S.C., Wang, S.M., Chang, Y.L.,Wang, C.H. A study of computer-aided mold cooling simulations based on different methods. Proceedings of ANTEC p Chen, S.C., Hu, S.Y.,Davidoff, A. Hybrid methods for injection mold cooling process simulation and mold cooling system analysis. Proceedings of ANTEC Montreal, Que, Can. p Turng, L.S., Wang, V.W.,Wang, K.K., Numerical simulation of the coinjection molding process. Journal of Engineering Materials and Technology, Transactions of the ASME, (1): p Wang, K.K. Twenty years of CIMP research towards CAE for injection molding. Proceedings of American Society of Mechanical Engineers, Materials Division (Publication) MD Chicago, IL, USA: ASME. p Turng, L.S.,Wang, K.K. Computer-aided cooling-line system for injection molds. Proceedings of Chicago, IL, USA: Publ by American Soc of Mechanical Engineers (ASME). p Chen, S.C.,Chung, Y.C., Numerical simulations of the cyclic, transient mold heat transfer in injection mold cooling process. International Communications in Heat and Mass Transfer, (3): p Park, S.J.,Kwon, T.H., Thermal and Design Sensitivity Analyses for Cooling System of Injection Mold, Part 2: Design Sensitivity Analysis. Journal of Manufacturing Science and Engineering, (2): p Tang, L.Q., Pochiraju, K., Chassapis, C.,Manoochehri, S., Threedimensional transient mold cooling analysis based on galerkin finite element formulation with a matrix-free conjugate gradient technique. International Journal for Numerical Methods in Engineering, (18): p Zou, Q. Injection molding cooling optimization. Proceedings of ANTEC p

218 41. Matsumoto, T., Tanaka, M., Miyagawa, M.,Ishii, N., Optimum design of cooling lines in injection moulds by using boundary element design sensitivity analysis. Finite Elements in Analysis and Design, (2-3): p Tang, L.Q., Pochiraju, K., Chassapis, C.,Manoochehri, S. Optimal design of cooling system for injection molding. Proceedings of ANTEC Indianapolis, IN, USA. p Tang, L.Q., Chassapis, C.,Manoochehri, S., Optimal cooling system design for multi-cavity injection molding. Finite Elements in Analysis and Design, (3): p Lin, J.C., Optimum cooling system design of a free-form injection mold using an abductive network. Journal of Materials Processing Technology, (1-3): p Li, C.L., A feature-based approach to injection mould cooling system design. Computer Aided Design, (14): p Sun, Y.F., PhD thesis.milled groove insert method for cooling of plastic injection moulds. 2004, National University of Singapore. 47. Bonnet, M., Boundary integral equation methods for solids and fluids. 1st ed. 1995, NY: John Wiley & Sons. 48. Heinrich, J.C.,Pepper, D.W., Intermediate finite element method: fluid flow and heat transfer applications. 1st ed. 1999, NY: Taylor & Francis. 49. Beer, G., Programming the boundary element method: an introduction for engineers. 1st ed. 2001, NY: John Wiley & Sons. 50. Raamachandran, J., Boundary and finite elements theory and problems. 2000, New Delhi,India: CRC Press. 51. Zhou, H., Zhang, Y., Wen, J.,Li, D., An acceleration method for the BEMbased cooling simulation of injection molding. Engineering Analysis with Boundary Elements, (8-9): p Qiao, H., Transient mold cooling analysis using BEM with the time-dependent fundamental solution. International Communications in Heat and Mass Transfer, (3-4): p

219 53. Hassan, H., Regnier, N., Le, B.C.,Defaye, G., 3D study of cooling system effect on the heat transfer during polymer injection molding. International Journal of Thermal Sciences, (1): p St. Jacques, M., Analysis of thermal warpage in injection molded flat parts due to unbalanced cooling. Polymer Engineering Science, (4): p Thompson, M.,White, J.R., Effect of a temperature gradient on residual stresses and distortion in injection moldings. Polymer Engineering and Science, (4): p Matsuoka, T., Takabatake, J.-I., Koiwai, A., Inoue, Y., Yamamoto, S.,Takahashi, H., Integrated simulation to predict warpage of injection molded parts. Polymer Engineering & Science, (14): p Kabanemi, K.K., Vaillancourt, H., Wang, H.,Salloum, G., Residual stresses, shrinkage, and warpage of complex injection molded products: Numerical simulation and experimental validation. Polymer Engineering and Science, (1): p Kabanemi, K.K., Vaillancourt, H., Wang, H.,Salloum, G. Numerical simulation of warpage and residual stresses and experimental validation. Proceedings of ANTEC Toronto, Can. p Hastenberg, C.H.V., Wildervanck, P.C., Leenen, A.J.H.,Schennink, G.G.J., The measurement of thermal stress distributions along the flow path in injection-molded flat plates. Polymer Engineering & Science, (7): p Bushko, W.C., Stokes, V.K., Solidification of thermoviscoelastic melts. Part II: Effects of processing conditions on shrinkage and residual stresses. Polymer Engineering and Science, (4): p Jansen, K.M.B., Pantani, R.,Titomanlio, G., As-molded shrinkage measurements on polystyrene injection molded products. Polymer Engineering and Science, (2): p

220 62. Jansen, K.M.B.,Titomanlio, G., Effect of pressure history on shrinkage and residual stresses - Injection molding with constrained shrinkage. Polymer Engineering and Science, (15): p Jansen, K.M.B., Van Dijk, D.J.,Husselman, M.H., Effect of processing conditions on shrinkage in injection molding. Polymer Engineering and Science, (5): p Zoetelief, W.F., Douven, L.F.A.,Ingen Housz, A.J., Residual thermal stresses in injection molded products. Polymer Engineering and Science, (14): p Liu, S.J., Modeling and simulation of thermally induced stress and warpage in injection molded thermoplastics. Polymer Engineering and Science, (6): p Choi, D.S.,Im, Y.T., Prediction of shrinkage and warpage in consideration of residual stress in integrated simulation of injection molding. Composite Structures, (1-4): p Kansal, G., Atreya, S.K.,Rao, P.N., Thermal and stress analysis of an injection-molded polystyrene ring. Polymer - Plastics Technology and Engineering, (1): p Kansal, G., Rao, P.N.,Atreya, S.K., Study: Temperature and residual stress in an injection moulded gear. Journal of Materials Processing Technology, (3): p Zhang, X., Cheng, X., Stelson, K.A., Bhattacharya, M., Sen, A.,Voller, V.R., Approximate model of thermal residual stress in an injection molded part. Journal of Thermal Stresses, (6): p Beleforte, D.A., Rapid Prototyping Speeds Product Development.. Laser Focus World (6): p Ashley, S., Rapid prototyping is coming of age. Mechanical Engineering, (7): p Jacobs, P.F., Fundamentals of StereoLithography. 1992, USA: Society of Manufacturing Engineers,. 200

221 73. Dimitrov, D., Moammer, A.,Harms, T. Cooling channel configuration in injection moulds. Proceedings of 4th International Conference on Advanced Research in Virtual and Physical Prototyping, VRAP Leiria. p Vallabhajosyula, P.,Bourell, D.L. Production of fully ferrous components using indirect selective laser sintering. Proceedings of 4th International Conference on Advanced Research in Virtual and Physical Prototyping, VRAP Leiria. p Kochan, D., Solid freeform manufacturing: advanced rapid prototyping. 1st ed. 1993, Amsterdam: Elsevier. 76. Bunnell, D., Bourell, D.,Marcus, H., Solid freeform fabricationof powders using laser processing. Advances in powder metallurgy and particulate materials. 1996, Metal Power Institute Federation. p Choi, J.,Mazumder, J. Rapid manufacturing by laser aided direct metal deposition process: Issues and examples. Proceedings of ASME Design Engineering Technical Conference and Computers and Information in Engineering Conference Pittsburgh, PA. p Lu, Z.L., Li, D.C., Lu, B.H., Zhang, A.F., Zhu, G.X.,Pi, G., The prediction of the building precision in the Laser Engineered Net Shaping process using advanced networks. Optics and Lasers in Engineering, (5): p Yin, H.,Felicelli, S.D., Dendrite growth simulation during solidification in the LENS process. Acta Materialia, (4): p Gary, L.,Dick, R. Rapid Prototyping and the ASME B5.54 Standard for Evaluation of Machine Tool Performance. Proceedings of Measurements and Standards Issues in Rapid Prototyping National Institute of Technology (NIST),USA. p Thoma, D.J., Lewis, G.K., Milewski, J.O.,Nemec, R.B. Free-form processing of near-net shapes using directed light fabrication. Proceedings of International Thermal Spray Conference Nice,France. p Milewski, J.O., Lewis, G.K.,Fonseca, J.C. Near net shape processing for solar thermal propulsion hardware using Directed Light Fabrication. Proceedings 201

222 of International Solar Energy Conference Albuquerque, NM, USA: ASME. p Milewski, J.O., Thoma, D.J., Fonseca, J.C.,Lewis, G.K., Development of a near net shape processing method for rhenium using directed light fabrication. Materials and Manufacturing Processes, (5): p Dutta, B., Singh, V., Natu, H., Choi, J.,Mazumder, J., Direct metal deposition. Advanced Materials and Processes, (3): p Koch, J.L.,Mazumder, J. Rapid Prototyping by Laser Cladding. Proceedings of International Congress on Applications of Lasers & Electro Optics p Mazumder, J., Schifferer, A.,Choi, J., Direct materials deposition: Designed macro and microstructure. Materials Research Innovations, (3): p Choi, J.,Chang, Y., Characteristics of laser aided direct metal/material deposition process for tool steel. International Journal of Machine Tools and Manufacture, (4-5): p Facchini, L., Magalini, E., Robotti, P., Molinari, A., Höges, S.,Wissenbach, K., Ductility of a Ti-6Al-4V alloy produced by selective laser melting of prealloyed powders. Rapid Prototyping Journal, (6): p Boi, Q., Fei, H., Shi, Y.S.,Wei, Q.S. Comparison of two scan strategies applied to the selective laser melting. Proceedings of International Conference on Computer Application and System Modeling, ICCASM Shanxi, Taiyuan. p Jinhui, L., Ruidi, L., Wenxian, Z., Liding, F.,Huashan, Y., Study on formation of surface and microstructure of stainless steel part produced by selective laser melting. Materials Science and Technology, (10): p Khan, M.,Dickens, P., Selective Laser Melting (SLM) of pure gold. Gold Bulletin, (2): p

223 92. Louvis, E., Fox, P.,Sutcliffe, C.J., Selective laser melting of aluminium components. Journal of Materials Processing Technology, (2): p Li, R., Liu, J., Shi, Y., Du, M.,Xie, Z., 316L stainless steel with gradient porosity fabricated by selective laser melting. Journal of Materials Engineering and Performance, (5): p Yang, Y., He, X., Wang, D., Wu, W., Su, X., Ding, H., Kong, W.,Wang, H. Selective laser melting for rapid prototyping of medical devices. Proceedings of Pacific International Conference on Applications of Lasers and Optics, PICALO, Wuhan,China. 95. Cormier, D., Harrysson, O.,West, H., Characterization of H13 steel produced via electron beam melting. Rapid Prototyping Journal, (1): p Gaytan, S.M., Murr, L.E., Martinez, E., Martinez, J.L., MacHado, B.I., Ramirez, D.A., Medina, F., Collins, S.,Wicker, R.B., Comparison of microstructures and mechanical properties for solid and mesh cobalt-base alloy prototypes fabricated by electron beam melting. Metallurgical and Materials Transactions A: Physical Metallurgy and Materials Science, (12): p Jiang, D.C., Dong, W., Tan, Y., Wang, Q., Peng, X.,Li, G.B., Investigation on removal of aluminum impurity in metallurgical grade silicon by electron beam melting. Journal of Materials Engineering, 2010(8): p Rännar, L.E., Glad, A.,Gustafson, C.G., Efficient cooling with tool inserts manufactured by electron beam melting. Rapid Prototyping Journal, (3): p Sabbadini, S., Tassa, O., Gennaro, P.,Ackelid, U. Additive manufacturing of gamma titanium aluminide parts by electron beam melting. Proceedings of 139th Annual Meeting and Exhibition,The Minerals, Metals & Materials Society Seattle, WA. p Qi, H.,Li, Y. Powder blowing and preheating experiments in Electron Beam Selective Melting. Proceedings of 2nd International Conference on 203

224 Computer Engineering and Technology, ICCET Chengdu,China. p Mostafa, N., Syed, H.M., Igor, S.,Andrew, G., A Study of Melt Flow Analysis of an ABS-Iron Composite in Fused Deposition Modelling Process. Tsinghua Science and Technology, : p Masood, S.H.,Song, W.Q., Development of new metal/polymer materials for rapid tooling using fused deposition modelling. Materials and Design, (7): p Nikzad, M., New Metal/Polymer Composites for Fused Deposition Modelling Applications-PhD Thesis, Faculty of Engineering & Industrial Science,. 2011, Swinburne University of Technology: Melbourne Nikzad, M., Masood, S.H., Sbarski,I, Thermo-mechanical Properties of a Highly Filled Polymeric Composite for Fused Deposition Modelling. Materials & Design, : p Saifullah, A.B.M., Masood, S.H.,Sbarski, I., Thermal-structural finite element analysis of injection moulding dies with optimized cooling channels. Materials Science Forum, p Grimm, T., User s guide to rapid prototyping. 2004, Dearborn.: Society of Manufacturing Engineering Park, H.S.,Dang, X.P., Optimization of conformal cooling channels with array of baffles for plastic injection mold. International Journal of Precision Engineering and Manufacturing, (6): p Dalgarno, K.,Stewart, T., Production tooling for polymer moulding using the RapidSteel process. Rapid Prototyping Journal, (3): p Ahn, D.G.,Kim, H.W., Study on the manufacture of a thermal management mould with three different materials using a direct metal tooling rapid tooling process. Proceedings of the Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture, (3): p Au, K.M.,Yu, K.M., Variable radius conformal cooling channel for rapid tool. Materials Science Forum, : p

225 111. Decelles, P., Direct AIM prototype tooling procedural guide, 1st ed. 1999, CA,USA: 3D Systems Harper, C., Handbook of plastics, elastomers, and composites. 1st ed. 1992, New York: McGraw-Hill Bennett, G., Developments in rapid prototyping and tooling,. 1997, UK.: Mechanical Engineering Publications Limited Xu, X., Conformal cooling and rapid thermal cycling in injection moulding with 3-D printing,phd thesis, 1999, MIT: USA Jacobs, P.F., New frontiers in mold construction: High conductivity materials & conformal cooling channels. American Society of Mechanical Engineers, Manufacturing Engineering Division, : p Schmidt, F., Pirc, N., Mongeau, M.,Chinesta, F., Efficient mold cooling optimization by using model reduction. International Journal of Material Forming, 2010: p Schmidt, W.R., White, R.D., Bird, C.E.,Bak, J.V. Conformal cooling versus conventional cooling: an injection molding case study with P-20 and 3DPprocessed tooling. Proceedings of Solid Freeform Fabrication Symposium. Austin, TX p Pham, D.,Dimov, S., Rapid manufacturing: the technologies and applications of rapid prototyping and rapid tooling 1st ed. 2001, New York: Springer Lee, K., Principles of CAD/CAM/CAE systems. 1999, Massachusetts: Addison-Wesley Bryden, B.G.,Pashby, I.R., Hot platen brazing to produce laminated steel tooling. Journal of Materials Processing Technology, (2): p Bryden, B.G., Pashby, I.R., Wimpenny, D.I.,Adams, C., Laminated steel tooling in the aerospace industry. Materials and Design, (4): p Bryden, B.G., Wimpenny, D.I.,Pashby, I.R., Manufacturing production tooling using metal laminations. Rapid Prototyping Journal, (1): p

226 123. Siegfried, M., Optimised mould temperature control procedure using DMLS, Technical Paper. May, 2008, EOS GmbH, Electro Optical Systems Geiger, M.,Ozel, T. Development of rapid tooling for injection molding using metal-filled stereolithography (SLA) cavity inserts Proceedings of 2nd Int. Conf. and Exhibition on Design and Production of Dies and Molds Kusadasi, Turkey p Jianxin, L.,Kuhn, H.A., Materials development for prometal, MS&T '05 Technical Program. 2005: Robert Morris University,Pittsburgh, PA. p Alfred, T.S., RCT takes layer-by-layer approach to mold, core production. Modern Casting, (2): p Shellabear, M., Kotila, J., Danzig, A.,Syvänen, T. Latest developments in direct metal laser-sintering for series production tooling. Proceedings of Society of Plastics Engineers - Modern Moldmaking Techniques Conference Frankfurt. p Sirinterlikci, A., Czajkiewicz, Z., Behanna, N.,Doswell, J. Direct and indirect rapid tooling. Proceedings of Rapid Prototyping Conference and Exposition Schaumburg, IL. p Rees, H., Mold Engineering. 2nd ed. 2002, Munich: Hanser Publishers Domoff, D.J., Plastics in Automotive, Plastics Engineering, Monthly Magazine,September. 2010, Wiley: Society of Plastic Engineers,USA. p Kalpakjian, S.,Schmid, S.R., Manufacturing engineering and technology, 4th ed. 2000: Pearson Education Asia Cracknell, P.S.,Dyson, R.W., Handbook of thermoplastic injection noulding design, 1st ed. 1993, Hong Kong: Chapman & Hall Kamal, M.R., Isayev, A.I.,Liu, S.J., Injection Molding technology and fundamentals, 1st ed, ed. J.L. White. 2009, Munich: Hanser Dym, J.B., Product design with plastics: a practical manual,. 1983, NY,USA: Industrial Press Inc. 206

227 135. Chen, S.C., Liaw, W.L., Peng, H.S., Huang, L.T.,Chang, Y.F., Characteristics of mould separation and its effects on qualities of thin wall injection moulded parts. Plastics, Rubber and Composites, (1): p Šercer, M., Godec, D.,Bujanić, B. Application of Moldex3D for thin-wall injection moulding simulation,. Proceedings of 9th International Conference on Numerical Methods in Industrial Forming Processes Porto. p Gava, A.,Lucchetta, G. Three dimensional viscoelastic simulation of thin-wall injection moulding with microfeatures. Proceedings of 66th Annual Technical Conference of the Society of Plastics Engineers, ANTEC Milwaukee, WI. p Kita, H., Mizoguchi, M.,Kuriyama, T. Evaluation of fracture toughness for thin-wall injection moulded plastics. Proceedings of Society of Plastics Engineers Annual Technical Conference, ANTEC Cincinnati, OH. p Kudlik, N., New possibilities for the injection moulding of thin-wall parts. Kunststoffe Plast Europe, (6): p Kim, H.,Colton, J.S., Fabrication and analysis of plastic hypodermic needles. Journal of Medical Engineering and Technology, (4): p Kuepper, H., Thin-wall injection moulding. Kunststoffe Plast Europe, (10): p Holman, J.P., Heat Transfer, 8th ed. 1997: McGraw Hill Inc Kakac, Y.Y.a.S., Heat Conduction 4th Ed ed. 2008: Taylor & Francis Serth, R.W., Process Heat transfer :Principles and applications, 1st ed. 2007: Elsevier Ltd Jeppson, R.W., Analysis of flow in pipe networks,. 1976, UK,Butterworth Publications. 207

228 146. Dittus, F.W.,Boelter, L.M.K., Heat transfer for automobile radiators of the tubular type,. University of California Publications in Engineering, : p Cook, R.D., Malkus, D.S.,Plesha, M.E., Concepts and applications of finite element analysis. 1989, New York, : Wiley Logan, D.L., A first course in the Finite Element Method 4th ed. 2007: Thomson,Canada Bhatti, M.A., Fundamental Finite Element Analysis and Applications,. 2005: John Wiley & Sons Grandin, H., Fundamentals of the Finite Element Method,. 1991, NY: McGraw-Hill Jeffrey, A., Advanced Engineering Mathematics 1st ed. 2001, NY: Academic Press Lakshmi, J.N., Finite element analysis,. Vol. 1st. 2009, Hyderabad, India: BS Publications 153. Reddy, J.N., An introduction to the finite element method 3rd ed. 2006, New York: McGraw-Hill Higher Education 154. Pepper, D.W.,Heinrich, J.C., The finite element method : basic concepts and applications, 2nd ed. 2006, New York: Taylor & Francis Irons, B.M.,Ahmad, S., Techniques of Finite Elements,. 1979: Ellis Horwood William, D.C., Jr, Materials Science and Engineering: An Introduction,. 2007, USA: John Wiley & Sons, Inc Shigly, J.E., Mischke, C.R.,Budynas, R.G., Mechanical Engineering Design, 7th ed. 2004: McGraw-Hill Sulzer Metco Australia Pty Ltd: 2/89 Gow St, PADSTOW, NSW, 2211,Australia Derek, H., Moldflow Pty. Ltd, A subsidiary of Moldflow Corporation: ,Colchester Road,VIC-3137, Melbourne,Australia. 208

229 160. Zhai, M., Lam, Y.C.,Au, C.K., Runner sizing and weld line positioning for plastics injection moulding with multiple gates. Engineering with Computers, (3): p Haddad, H., Masood, S.H.,Saifullah, A.B.M., Gate location and its effects on product quality in injection moulding,. Advanced Materials Research : p Kim, H.S., Son, J.S.,Im, Y.T., Gate location design in injection molding of an automobile junction box with integral hinges. Journal of Materials Processing Technology, (1-3 SPEC.): p Manzione, L.T., Applications Of Computer Aided Engineering in Injection Molding. 2004, Hanser Publishers.: Munich. 209

230 Appendix A Mathematical Calculations A1 Sample Calculation for Heat Flux A1.1 Calculation of conduction heat flux: Values of heat flux at sprue surface at 5 second for conventional straight cooling channels, Using Equation (3.2) from chapter three, QC /A =K T/L Where, K= Thermal conductivity of Stavax Supreme mould = 20 W/m C (from table 3.5) T = difference of temperature of sprue surface and cooling wall at 5 second (from table 3.2) = L= Average distance between two surfaces = 40 mm Heat flux (QC /A) = W/m 2 = W/mm 2 210

231 A1.2 Calculation of convective heat flux: Heat flux through side cooling channels surfaces for conventional straight cooling channels, Using Equation (3.8) from chapter three, Qh/A=hC (TW-Tb)av Where, hc = Convective heat transfer co-efficient for mm diameter cooling channel was calculated using equation 3.10 (given in the next section). = 5906 watt/m 2 C TW-Tb = (difference of temperature between side cooling channel wall temperature and coolant temperature, from Table 3.2) = Heat flux (Qh/A) = W/m 2 = W/mm 2 A2 Sample Calculation for Reynolds Number and Convective Co-efficient: A2.1 Calculation of Reynolds Number (Re) for mm hydraulic diameter cooling channel: Using Equation (3.11) form chapter three, Re=ρDhv/µ Where, 211

232 ρ= density of coolant ( pure water) =965 kg/m 3 Dh=Hydraulic diameter of cooling channel =10.10 mm V=Velocity of coolant water =1.4 m/sce µ = Dynamic viscosity of coolant water = 1.002x10-3 kg m -1 s -1 Re = A2.2 Convective heat transfer co-efficient for mm diameter cooling channel: Using Equation (3.10) from chapter three, h c = R e 0.8 P r 0.4 where, Reynolds Number, Re = K= thermal conductivity of coolant water = 0.58 W/m C Hydraulic diameter, Dh= mm Prandlt Number, Pr= Cpµ/k (Cp= specific heat of water =4.18x10-3 J kg -1 K -1 ) = 7.22 Convective heat transfer co-efficient, hc =5906 watt/m 2 C 212

233 A3 Sample Calculation for Theoretical Cooling Time A3.1 Cooling time for PP plastic using convntional straight cooling channel: Using Eqation (3.14) from chapter three, Cooling time, Where, tc = ln ( ). Thickness of plastic part S=2.2mm Thermal diffusivity α = ρ =8.87x10-8 m²/s Thermal conductivity of PP plastic =0.14 w/m. c Density of PP plastic =830 kg/m 3 Specific heat of PP plastic =1900 J/kg. c Tm=moulding temperature of plastic =250 C Td =demoulding temperature of ejected plastic material = 80 C Tw= maximum temperature of interface wall between molten plastic and cavity wall = 65 C theoretical cooling time tc =18.94 s 213

234 A4 Sample Calculation for Fatigue life cycle of the mould. A4.1 Calculation of fatigue life for conventional cooling channel made of Stavax Supreme Using equation 4.22 from chapter four Fatigue life of mould, N= (Sf /a) 1/b where, a = (f*sut) 2 /Se = 2050 and b = -(1/3)*log(f*Sut/Se) =0.08 where, Sf = Ses= Equivalent stress or Von-Mises stress=470 Mpa Sut = Ultimate tensile strength =1500 MPa Se= Elastic strength or endurance limit =750 Mpa Fatigue life of mould, N = 9 million cycle 214

235 Appendix B Temperature and Equivalent Stress Distribution Images B1 Images for Temperature Distribution from FEA Simulation (a) (b) Figure B1.1: Temperature distribution after one cycle (35 second) in the cavity mould sectional view with conventional cooling channels made of (a) Stavax Supreme and (b) Aluminium materials. 215

236 B1.2: Temperature distribution after one cycle (35 second) in the cavity mould sectional view with square sectional conformal cooling channels (D h= 10.10mm and pitch 2D h=20.20 mm) made of Stavax Supreme materials. B1.3: Temperature distribution after one cycle (35 second) in the cavity mould sectional view with circular cross sectional conformal cooling channels (D h= 12.10mm and pitch 1.5D h=18.15 mm) made of Aluminium materials. 216

237 B1.4: Temperature distribution after one cycle (35 second) in the cavity mould sectional view with square cross sectional conformal cooling channels (D h= 9 mm and pitch 1.5D h=13.5 mm) made of Aluminium materials. B1.5: Temperature distribution after one cycle (35 second) in the cavity mould sectional view with circular cross sectional conformal cooling channels (D h= 12.10mm and pitch 2D h=24.2 mm) made of Stavax Supreme materials. 217

238 B1.6: Temperature distribution after one cycle (35 second) in the cavity mould sectional view with circular cross sectional conformal cooling channels (D h= 12.10mm and pitch 2D h=24.2 mm) made of Aluminium materials. B1.7 :Temperature distribution after one cycle (35 second) in the cavity mould sectional view with circular cross sectional conformal cooling channels (D h= 9 mm and pitch 2D h=18 mm) made of Stavax Supreme materials. 218

239 B2 Images for Equivalent Stress Distribution from FEA Simulation (a) (b) Figure B2.1: Equivalent stress distribution at about 3.9 th second of cycle for cavity mould sectional view with conventional cooling channel made of (a) Stavax Supreme and (b) Aluminium materials. 219

240 Figure B2.2: Equivalent stress distribution at about 3.9 th second of cycle for cavity mould sectional view with square sectional conformal cooling channels (D h =10.10mm and pitch 2D h=20.20 mm) made of Stavax Supreme materials. Figure B2.3: Equivalent stress distribution at about 3.9 th second of cycle for cavity mould sectional view with square sectional conformal cooling channels ( D h =12.10mm and pitch 1.5D h=18.15 mm) made of Aluminium materials. 220

241 (a) (b) Figure B2.4: Equivalent stress distribution at about 3.9 th second of cycle for cavity mould sectional view with circular cross sectional conformal cooling channels (D h= 11.10mm and pitch 2D h=22.20 mm) made of (a) Stavax Supreme and (b) Aluminium materials. 221

242 Figure B2.5: Equivalent stress distribution at about 3.9 th second of cycle for cavity mould sectional view with circular cross sectional conformal cooling channels (D h = 9mm and pitch 2.5D h=22.5 mm) made of Stavax Supreme materials. B2.6: Equivalent stress distribution at about 3.9 th second of cycle for cavity mould sectional view with circular cross sectional conformal cooling channels (D h= 12.10mm and pitch 1.5D h=18.15 mm) made of Aluminium materials. 222

243 B2.7: Equivalent stress distribution at about 3.9 th second of cycle for cavity mould sectional view with circular cross sectional conformal cooling channels (D h= 12.10mm and pitch 2D h=24.2 mm) made of Stavax Supreme materials. B2.8: Equivalent stress distribution at about 3.9 th second of cycle for cavity mould sectional view with circular cross sectional conformal cooling channels (D h= 10.10mm and pitch 2D h=20.20 mm) made of Stavax Supreme materials. 223

244 B2.9: Equivalent stress distribution at about 3.9 th second of cycle for cavity mould sectional view with circular cross sectional conformal cooling channels (D h= 9 mm and pitch 2D h=18 mm) made of Stavax Supreme materials. 224