AbstractPurpose – The purpose of this paper is to present the results of an investigation into the effect of injection molding process parameters on theperformance of direct metal laser sintered (DMLS) mold in producing quality Zytel nylon 66 plastic parts with consistency in part shrinkage and shot/partweight.Design/methodology/approach – The injection mold for an industrial component (hub gear) was fabricated in EOS M-250 machine using bronze-based material. The effect of four injection molding parameters (injection pressure, melt temperature, speed, and injection time) on part shrinkage andweight were studied experimentally using L9 orthogonal array. The weight of the part just after ejecting from the cavity, and the average shrinkagemeasured after cooling, were used in grey relational analysis technique to assess the effect of each molding parameter. Further, surface properties suchas surface finish, wear, scratch and corrosion resistance tests were conducted on DMLS mold material samples, in order to evaluate its use in rapidmanufacturing applications.Findings – The study found that injection speed and melt temperature have significant influence on part weight and shrinkage. The optimized moldingprocess variables were slightly more in the case of DMLS molds as compared with the parameters suggested in the plastic datasheet. Scanning electromicroscope (SEM) analysis of the mold surface after producing 5,000 glass filled Nylon 66 (Zytel) moldings did not indicate any surface degradation,confirming the use of DMLS mold in rapid manufacturing of few thousands of moldings.Research limitations/implications –
The grey relational analysis does not compute the effect of any two or more variables together unlike ANNOVA.Second, this study alone is not enough to estimate life of DMLS mold, although 5,000 glass filled nylon 66 moldings are successfully produced withoutany damage on mold surface.Practical implications – This investigation demonstrates a generic approach of using grey relational analysis to quantify the effect of differentmolding process variables on selected quality parameters. This method can be easily extended for new processes and materials. The preliminary tests onsurface finish, scratch, wear and corrosion resistance performed on DMLS mold samples have highlighted the need for improving surface properties toenhance their life. The authors are currently working on hard coating of DMLS molds as one of the solutions.Originality/value – Use of grey relational analysis is new to the problem of injection molding process optimization. Moreover, effect of injectionmolding parameters on part weight and shrinkage in DMLS mold has not been studied previously. This study helps while considering DMLS molds formanufacturing few thousands of parts.Keywords Rapid prototypes, Lasers, Sintering, Formed materialsPaper type Research paper1. IntroductionConventional mold making requires detailed process planningand involves a series of machining operations to generate thedesired mold cavity. This consumes significant amount of timeand cost. This has led to the exploration and growth of rapidprototyping (RP) based tooling, referred to as rapid tooling(RT), which has been shown to reducemold development timeby 50 percent or more (Levy et al., 2003). These processes canbe classified as direct RT, in which molds are fabricated in a RPsystem, or indirect RT, in which a RPmaster is converted into amold using a secondary process. Recently, more than 25 RTprocesses are available. The most important RT processesuseful for developing injection molds are listed in Table I,indicating the their parent RP process, and classification direct,indirect, soft, bridge and hard tooling. The soft and bridgetooling methods are suitable for producing molds that can be used to produce a limited number of parts.
Thesemolds usuallycannot withstand the pressures and temperatures involved inconventional injection molding. The hard tooling methodsenable fabricating metal tooling that can be used in injectionmolding machines, and result in better quality and largerquantity of parts (compared to soft and bridge tooling). Themetalmolds produced by the following RT processes includingDTM’s selective laser sintering (SLS), direct metal lasersintering (DMLS), RP-investment casting (RP-IC) and 3DKeltool could withstand conventional molding pressure andtemperature (Levy et al., 2003).Injection molds are expected to produce parts very close tospecifications. The accuracy, thermal conductivity andmechanical properties of the mold have significant influenceon injection molding cycle, part quality, geometric complexityand competitiveness.The effortsmade over the last few years toimprove the RT capabilities compliment significantdevelopments in materials and processes. This includedadaptive slicing (Pandey et al., 2003), build processoptimization and hybrid RP (layer machining) (Karunakaranet al., 2004). Production of die steelmolds using RP patterns ordirectly from layered manufacturing processes is still maturingand they need further improvements in accuracy, longer-lifemold material, and surface quality. Recently, Nagahanumaiahet al. (2006, 2007) have developed computer-aidedmethodology for RT process-selection for injection moldingbased on mold manufacturability evaluation and cost analysis.However, the database built into this system is limited to RTprocess capabilities, and database is not rich enough to predictthe quality of the moldings produced by the selected RT mold.This is because the processing knowledge using these new RTmold material is not very well established, leading to potentialincompatibility among process capabilities, mold design andproduct specifications. It is therefore, essential to investigate theeffect of injection molding process variables on quality of partproduced using these new and non-conventional RT moldmaterials. This paper is focused on understanding these effectswhile producing glass-filled Nylon 66 (Zytel) product usingDMLS bronze-based mold, which has not been studied indetail. A grey relational analysis technique has been used toquantify these effects of injection-molding process variables onselected quality parameters.The grey relational analysis developed by Deng enablesoptimizing process variables for multiple quality parameters.The response of these quality factors can be considered forany of the following three criteria: higher the better, lower isthe better, and the response factor is discrete value (Deng,1989). Grey relational analysis and its variants have been usedfor a variety of optimization problems. Recently, thistechnique has been used in EDM process to study the effectof discharge current, spar-on-time and flushing pressure onmaterial removal rate, electrode wear rate and surface finish,while using electrodes made by DMLS process (Vijay andNagahanumaiah, 2006). Therefore, considering stochasticnature of injection molding process, grey relational analysistechnique was selected for this study, to understand the effectof four important molding process variables on two qualityparameters.2. Previous work on DMLS moldDMLS process is similar to SLS process in that it too canproduce molds directly from 3D computer-aided design(CAD) model of core and cavity inserts layer by layer basedon liquid phase sintering using laser. It differs from SLS inthat it does not use polymer binder, and has been pioneeredby EOS Germany. Figure 1 shows a schematic representationof the EOSINT M-250 machine used for producing moldinserts. It consists of a laser unit, a control computer, a buildchamber, a powder dispenser, a wiper blade and a buildcylinder. A 200W CO2 laser with spot size 0.3mm is used forliquid phase sintering of metal powder. The mold is built on asteel plate, which facilitates mold assembly. Two differentpowder systems: bronze-, and steel-based (with a nitrogenatmosphere) are employed (Simchi et al., 2003). The bronzemolds can be used for more than 10,000 injection-moldingcycles, and steel molds made by this process could withstandas high as 100,000 shots (Dolinsek, 2005).Several attempts were made to understand and improve theDMLS process, especially towards reducing the porosities inthe mold. A study of microstructure of DMLS models usingscanning electro microscope (SEM), revealed them to haveover 20 percent porosity (Khaing et al., 2001). Simchi studiedthe issues related to residual porosity, bending strength, artifact hardness, and density, and presented a beamcompensation technique for improvement of dimensionalaccuracy of the part/mold being produced in DMLS (Simchiet al., 2003). 注射成型模具英文文献和中文翻译:http://www.youerw.com/fanyi/lunwen_43984.html