ORIGINAL ARTICLEThermalstructural analysis of bi-metallic conformal coolingfor injection mouldsA. B. M. Saifullah & S. H. Masood & I. SbarskiReceived: 28 July 2011 /Accepted: 21 November 2011 /Published online: 9 December 2011#Springer-Verlag London Limited 2011Abstract In injection moulding process, cooling timegreatly affects the total cycle time. As thermal conductivityis one of the main factors for conductive heat transfer incooling phase of IMP, a cooling channel made by higherthermal conductive material will allow faster extraction ofheat from the molten plastic materials, thus resultingin shorter cycle time and higher productivity. The mainobjective of this paper is to investigate bi-metallic con-formal cooling channel design with high thermal con-ductive copper tube insert for injection moulds.Thermalstructural finite element analysis has been carriedout with ANSYS workbench simulation software for amould with bi-metallic conformal cooling channels andthe performance is compared with a mould with conven-tional straight cooling channels for an industrial plasticpart. Experimental verification has been carried out forthe two moulds using two different types of plastics,polypropylene (PP) and acrylonitrile butadiene styrene,in a mini injection moulding machine. Simulation andexperimental results show that bi-metallic conformalcooling channel design gives better cycle time, whichultimately increases production rate as well as fatiguelife of the mould.Keywords Bi-metallic conformal cooling.Injectionmoulding.Copper tube.Cooling time.Finite elementanalysis1 IntroductionInjection moulding is one of the most important manu-facturing processes available in plastic manufacturingindustries. With the broader use of plastics parts forconsumer products, the injection moulding process(IMP) has been renowned as the most widely used massmanufacturing process. The moulding cycle of IMPconsists of mould closing, injection/holding, cooling,mould opening and product removal stages. In anyinjection moulding cycle, cooling stage is the dominantcomponent, and generally accounts for 50% to 70% ofthe moulding cycle time 1. Thus the production rate isgenerally affected by the cooling channel design in themould. In a typical mould, cooling channel design is themost complex as well as vital factor for economical perfor-mance of injection moulding. The cooling system of aninjection mould can be divided mainly into three categories,cavity cooling, core cooling and sometime cooling of thestripper plate. There are many types of cooling channels thathave been invented by mould designers and researcherssince the invention of injection moulding. Among all typesof cooling channels, straight-drilled cooling channel (SDCC)system is the cheapest and the most popular cooling methodthe injection moulding companies are using. Figure 1 showsthe cross-sectional view of an injection mould with thelocation of SDCC around a plastic part cavity.SDCCs are also used as baffles and bubblers for coolingof cores in the injection moulds. These cooling channels arealso machined using drilling or boring 2, 3. They have astraight flow path and a circular cross-section. Baffles areused for cooling of cores with highly convex geometry 4.In baffle design, holes are first bored to the rear face of thecore, and the lower end of each hole is plugged. The boringsare then interconnected by another hole drilled from a sideA. B. M. Saifullah:S. H. Masood (*):I. SbarskiIndustrial Research Institute Swinburne,Faculty of Engineering and Industrial Sciences,Swinburne University of Technology,Melbourne, Australia 3122e-mail: .auInt J Adv Manuf Technol (2012) 62:123133DOI 10.1007/s00170-011-3805-5face. To ensure that the coolant flows into each hole,blades are fitted in each hole and act as baffles to divertwater to the top of each hole. To achieve more efficientcooling, twisted baffle are applied in practice. Bubbler isanotherpopularmethod used in the core cooling. Bubblersare similar to baffles, but instead of flowing from oneside to the other, the incoming coolant passes through atube fitted in the centre of the hole and returns to theoutlet through the passage in between the tube and thehole. But, it is more expensive 1.There has been significant improvement of coolingchannel design since the innovation of CAD/CAM/CAEand Rapid Tooling (RT) technology. Since early 1980s,mould cooling simulation provided substantial attentionwith different methods to predict the temperature distributionof the mould and part in injection moulding process 59. In1999, Jacobs 10 described the use of conformal coolingchannels in an injection mould insert. In that study, channelswere built by electroformed nickel shells, and finite elementsimulation shows that the conformal cooling channelformed by copper duct bending can increase the uniformityof mould temperature distribution. It can also decrease thecycle time and part distortion. As common injectionmoulding materials, such as steel, have not been includedin his research, the application is only restricted to copper ornickel duct bending. Xu et al.11 in 2001 applied the 3Dprinting process to fabricate injection moulds with conformalcoolingchannelsinside.ChaandPark12 in 2007 describedsome conformal cooling methods, with direct metal lasersintering and spray-formed tooling process. However, theincrease in complexity of part geometries hinders therealization of conformal cooling layout fabrication insome RT processes. Therefore, it is worthwhile to investigatefurther other effective approaches in order to obtain bettercooling performances. One such approach is to use materialswith higher thermal conductivity in injection moulds. Copperis a well-known material with higher thermal conductivitythan steel while its strength is generally lower than steel.Attempts have been made by some researchers to use copperorhighstrengthcopperalloysindiestoallowfasterextractionof heat to reduce cycle time. Kelly et al. 13 in 2011 haveinvestigated the performance of high strength copperalloy mould tool materials in injection moulding withregard to cycle time, part quality and energy consumption incomparison with tool steel. They concluded that copper alloytoolingcanachievesignificantreductionincycletimewithoutaffecting process or part quality. Beal et al. 14 in 2007 haveused the concept of functionally graded materials to developtooling inserts made of copper and steel using selective lasermelting technique. They observed that as copper was addedto tool steel, it provided more efficient heat transfer but ithad less capacity to absorb steel.This research work presents a novel cooling system designinvolving bi-metallic cooling channels in injection moulds,with high thermal conductive copper tube insert (CTI)used to replace conventional straight cooling channel(CSCC). It is expected that a copper tube insert in thechannel will further enhance heat transfer process duringcooling time of injection moulding. Bi-metallic coolingwith CTI can also be used for baffles and bubblers, but itmay be more suitable for plastic parts that do not havecurved surface other than round or fillet, and as a resultsuch cooling channel will maintain equidistance from thecavity surface asshown bydistance x,inFig.1. According toFouriers law of conduction of heat transfer, the distance thatheat is conducting through is inversely proportional to thetotal conduction of heat transfer energy. As a result, uniformheat transfer will take place in the moulding process. Inthis paper, the performance of bi-metallic straight coolingchannel (BSCC) and bi-metallic conformal cooling channel(BCCC), with two different thicknesses of CTI, have beeninvestigatedfor acavitymouldandcoremouldwithbubblers,through thermalstructural finite element analysis, supportedby experimental verification.2 Design of bi-metallic cooling channelsThe part chosen for this study is an injection mouldedplastic canister (0.5 L) made of polypropylene (PP)thermoplastic. Actual mould for this part is of six cavitiesmould, but only single cavity type has been consideredfor this investigation. Figure 2 shows the CAD model ofthe plastic part, which has outer dimensions of 160, 120 and48 mm with wall thickness of 2 mm and the weight of thepart is 69.5 g. Note that the part has curved surfaces at thecorners.Fig. 1 Sectional view of SDCC layout in a mould cavity124 Int J Adv Manuf Technol (2012) 62:123133Figure 3 shows the cavity and core moulds for the plasticpart with CSCC including bubbler cooling in the core.Figure 4a shows the design of BSCC with CTI fitted.Figure 5a shows the design of BSCC with CTI fitted. CTIhas also been used for bubbler system of core in both cases.Two different thicknesses, 2 and 3 mm of CTI, have beenused for BSCC and BCCC. The difference between BSCCand BCCC design is that in case of BSCC, the channels arestraight with no curved corners as shown in Fig. 4b, while incase of BCCC, the channels have curved shape corners,which are conformal with the plastic part corners, and as aresult, these cooling channels maintain same distance fromsurfaces of the plastic part as shown in Fig. 5b. Table 1 givesthe names and abbreviations of five types of coolingchannels that will be used in this study. The outer dimensionsof the single cavity mould are height of 232 mm, diameter of300 mm and the inner diameter of cavity and core coolingchannels are 12 and 15 mm, respectively.3 Thermalstructural finite element analysisThermalstructural FEA of the proposed bi-metallic coolingchannel moulds has been carried out with ANSYS work-bench simulation software to demonstrate that such mouldcan extract faster heat from molten plastic material in injec-tion moulding process, as well to check the robustness andlongevity of the mould with such bi-metallic channels. Inthe analysis, the mould material was taken Stavax Supreme(SS), a stainless steel tool alloy, as recommended by a localmould manufacturer, and the cooling channel insert materialwas high thermal conductive berylliumcopper (BC) alloy,which is capable of transferring heat at a higher rate thansteel. Table 2 shows a comparison of the physical propertiesof SS and CA.ANSYS workbench simulation software is capable ofsimulating both the steady state and transient behaviourwhen subjected to different structural and heat loads. Inthis simulation, transient analysis has been used becausein injection moulding, mould experiences variable tempera-ture, pressure and forces. Automatic meshing (elementsthat are automatically created depending on the physicalstructure) with tetrahedral elements have been used. Finerelevance centre and medium smoothing has been appliedin the meshing.In the simulation process first, the transient thermalanalysis has been carried out in the mould assembly andthen thermal analysis results have been coupled withtransient structural analysis, to calculate equivalent stress(von Mises) in thermal loading conditions. In the mouldingprocess of thermoplastics, three types of heat transfertake place:conduction through mould, convection in thecooling medium and outer surface of the mould, andfinally, radiation heat transfer, which is of very negligibleamount. In this analysis, radiation heat transfer has beenneglected. For thermal analysis, conduction and convec-tion heat flux (heat energy per unit area) have been usedas a boundary condition.Conduction heat transfer energy, which is of vitalimportance in IMP, has been calculated by Eq. 1,asdescribed in 15. This equation which is suitable forsteady-state one-dimensional heat transfer process has beenderived from Fourier conduction heat transfer equation forcomposite material. Though Eq. 1 has been used to calculateconduction heat flux values as an input boundary condition,tabular values of heat flux for different timing of themoulding cycle have been used rather than constantFig. 2 CAD model of plastic canisterFig. 3 Cavity and core with conventional straight cooling channel(CSCC)Int J Adv Manuf Technol (2012) 62:123133 125values to get the transient heat transfer effect in thesimulation process.Conduction heat energy;QCTiC0TwlsksAlckcA1where,ksThermal conductivity of SSkcThermal conductivity of CTIA Cross-sectional area through heat is transferringTWInside surface temperature of CTI.TiTemperature of cavity or core surface interface withplasticlsDistance from cavity or core surface to correspondingCTI outer surfacelcThickness of CTISimilarly, Eq. 2 13 has been used to calculate convectiveheat transfer energy inside the cooling channels surface.Convective heat energy;Qh hcATWC0TC 2where,A Surface area of the cooling channels in contact withflowing fluidTwAverage temperature of the inside surface of CTITCAverage temperature of the coolanthcConvection heat transfer coefficientThe convection heat transfer coefficient hc,has beencalculatedusingEq.3, basedonDittus-Boetler16 correctionequation for forced convective heat transfer by turbulentflow in a circular pipe. These coefficients were calculatedto be 5,397 and 5,709 watt/m2C for core and cavitycooling channels respectively.hc 0:023kDRe0:8Pr0:4310,000Re120,000 and 0.7Pr 120where,hcHeat transfer coefficientk Thermal conductivity of coolant (water)D Diameter of the cooling channelsRe Reynolds NumberPr Prandtl NumberOther thermal boundary conditions are the natural con-vection on the side surface of the cavity mould which isexposed to the air and the channel around the sprue bush, inwhich, air has been passed for additional cooling of spruebush. This additional cooling of sprue bush is necessary as itcarries the hot molten plastic material for injection into themould cavity. Convection co efficient has been used asboundary conditions in these cases, and the values for theseare 5106Watt/mm2C and 6.083103Watt/mm2C asrecorded by local mould manufacturer. So altogether, eleveninput boundary conditions have been used for thermalFig. 4 a Bi-metallic straightcooling channel (BSCC)with copper tube insert (CTI)in core and cavity; b sectionaltop view of cavity mould,showing the orientation ofBSCC in the mouldFig. 5 a Bi-metallic conformalcooling channel (BCCC) withcopper tube insert (CTI) in coreand cavity; b sectional top viewof cavity mould, showing theorientation of BCCC in themould126 Int J Adv Manuf Technol (2012) 62:123133analysis, seven for conduction heat flux (heat flux for SS, RSand BS surfaces for both core and cavity side, and SHSsurface shown in Fig. 6), two for convection heat flux (heatflux for CoCs and CaCS surfaces shown in Fig. 6) and twofor convection coefficient.Figure 6 shows the cross section of the entire mouldassembly, showing different interface surfaces. In order tocalculate different heat fluxes using Eqs. 1 and 2,itisnecessary to know the variable temperature, Tw, Tiand TCat these different surfaces (the interface of plastic andmould cavity, at the interface of cooling channel innersurface and cooling medium), as shown in Fig. 6. To getthese temperature values, a complete injection mouldingflow simulation (cool+flow+pack+warp analysis) has beencarried out separately with Autodesk Moldflow Insight(AMI) software. Flow simulation with AMI, also gives thevalues of the variable injection pressure at different surfaces(surfaces that plastic materials are in contact during injectionmoulding process as shown in Fig. 6) and clamping forces,which will be used as boundary conditions for thermalstructural analysis. For plastic flow analysis with AMI, dualdomain mesh has been used with 9,228 elements, mould andmelt temperature were 50C and 250C respectively, totalcycle time was 20 s (9, 8 and 3 s for injection/hold on,cooling and ejection, respectively), plastic and mouldmaterials were PP and SS stainless tool steel correspond-ingly. Pure water with a temperature of 10C has beenused as coolant. Table 3 shows the temperature values atsix different times (from 0 to 17 s) of the moulding cycle atdifferent interface surfaces of the assembly mould (asshowninFig.6) for the case of CSCC, from AMI flowsimulation. Similar values of temperature have beenobtained for other four cases of bi-metallic cooling channelcase. Values indicate that all four main interfaces cool downgradually during the moulding cycle. Average values ofthe temperature for each surface were taken for heat fluxcalculation using Eqs. 1 and 2. For a particular surface,same temperature values have been taken for core and cavityside because there is not much difference between them.Table 4 shows the heat flux values calculated at differenttimes (0 to 17 s) and used as boundary conditions forvarious interface surfaces for CSCC case. Similar valuesof heat flux have been calculated for other four cases ofbi-metallic cooling channel.Result of transient thermal analysis, which includes thetemperature response over the mould assembly for entirecycle, has been imported in the interface of transientstructural analysis to perform thermalstructural FEAanalysis. For structural analysis, four types of boundaryconditionshavebeenused,whicharefixedsupport(bottomofthe mould), injection pressure, clamping force (top of themould)andthetemperaturefromthethermalanalysis.Table5gives the values of variable clamping forces and injectionpressure for different surfaces for entire cycle, recordedfrom AMI flow simulation.4 Results and DiscussionFrom transient thermal analysis, temperature distributionhas been found for entire mould. Figures 7a, 8 and 9 showthe comparative temperature distribution for all coolingchannel moulds. In case of CSCC, after 1 cycle (20 s),temperature of the mould ranges from a minimum 13C tomaximum 74C (Fig.7a), whereas, for BSCC and BCCCCTI, average minimum to maximum temperature rangesTable 1 Different coolingchannels and their abbreviationsType of cooling channel DescriptionCSCC Conventional straight cooling channelBSCC 2-mm CTI Bi-metall