插座底板注塑模具設(shè)計(jì)與CAE分析（桂理工）

插座底板注塑模具設(shè)計(jì)與CAE分析（桂理工）,插座,底板,注塑,模具設(shè)計(jì),cae,分析,理工 本科畢業(yè)設(shè)計(jì)(論文) 外文翻譯（附外文原文） 學(xué) 院： 機(jī)械與控制工程學(xué)院 課題名稱： 插座底板注塑模具設(shè)計(jì)與CAE分析 專業(yè)(方向)： 機(jī)械設(shè)計(jì)制造及其自動(dòng)化(模具) 班 級(jí)： 機(jī)械11-2班 學(xué) 生： 蘇勇孫 指導(dǎo)教師： 楊 輝 日 期： 2015/3/11  外文翻譯（原文） Simulation of injection molding into rapid-prototyped molds Rajitha Aluru Michael Keefe and Suresh Advani The authors Rajitha Aluru is a Graduate Student, Michael Keefe is Associate Professor and Suresh Advani is Professor, all at the Mechanical Engineering Department, University of Delaware, Newark, Delaware 19716, USA. Key words Rapid prototyping, Rapid tooling, Stereolithography, Injection molding Abstract Injection molding is a very mature technology, but the growth of layer-build, additive, manufacturing technology (rapid prototyping) has the potential of expanding injection molding into areas not commercially feasible with traditional molds and molding techniques. This integration of injection molding with rapid prototyping has undergone many demonstrations of potential. What is missing is the fundamental understanding of how the modifications to the mold material and mold manufacturing process impact both the mold design and the injection molding process. This work expanded on an approach to utilize current numerical simulation programs and use of non-metal molds for injection molding. Verification and validation work is presented. The model was exercised by studying the effect of varying the thermal conductivity on final-part distortions. This work clearly showed that one could not obtain reasonable results by simply changing a few input parameters in the current simulations. Although the approach did produce more realistic results, more work will be required for a tool capable of accurate, quantitative predictions. Introduction Design is one of the more time-consuming steps in product development. Functional parts are needed for design verification testing, field trials, customer evaluation, and production planning. Tens to hundreds of parts are frequently needed long before final production. Low-volume prototyping becomes highly desirable but requires a small number of parts to be produced quickly and economically. Rapid tooling technologies offer viable methods for producing durable tooling for industry and could produce economical, yet functional molds and dies in as little as one week (MMS Online, 1997). In addition, if one can reduce the tooling costs, rapid tooling has the potential to enable traditional high-volume processes, such as injection molding, to be competitive at lower production volumes (Hilton, 1995). Injection molding is of particular interest since, for plastic part production, it clearly dominates the field (Hallum, 1997). Several low-volume tooling options are currently used in the rapid tooling industry. Two broad categories are soft tooling and hard tooling. Soft tooling can provide limited quantities of true prototypes in the desired end use materials. It is normally associated with low cost and use of materials with low hardness levels (e.g. silicone, rubber, epoxies, zinc alloys, etc.). Hard tooling is a process that allows a tool for injection molding and die casting to be manufactured quickly and efficiently so the resultant parts are representative of final production parts as well as the production material. Hard tooling is also often referred to as the use of hard materials for tooling and the prototype material properties are comparable with the final part material properties. Hard tooling includes castable steel alloy tools, milling aluminum tool inserts using a numerically-controlled machine, use of reinforced composite tools, sprayed metal tools, investment casting and tooling using rapid prototyping (RP) technologies. The RP approaches all use a layer-build, additive manufacturing process. Details regarding these processes are available elsewhere (Jacobs, 1996; Jacobs, 1992). Background In injection molding of thermoplastics (Isayev, 1987; Kennedy, 1995), solid plastic is melted, and the melt is injected into the mold under high pressure (usually between 0.689 and 34.5MPa). The mold is cooled to release the part. Thus, the mold material needs to have thermal and mechanical properties capable of withstanding the temperatures and pressures of the mold cycle. Figure 1 shows a typical pressure profile in the mold cavity during injection molding; there would be a similar temperature profile. Figure 1 Pressure profile during an injection-molding cycle The focus of many studies has been to create the mold directly by an RP process. By eliminating multiple steps, this method of tooling holds the best promise of reducing the time and cost needed to create low-volume quantities of parts in a production material. One approach has been to try and produce RP molds that have near-metal mechanical and thermal properties. In this manner, one can immediately utilize the large knowledge base that has developed around traditional injection molding. However, the wide variety of available RP materials and the relative ease of creating complex geometries with RP techniques, suggests that perhaps one could optimize the entire mold-making process. This integration of injection molding with RP technologies has undergone many demonstrations of potential. However, there is not a fundamental understanding of how the modifications to the mold material impact both the mold design and the injection molding process parameters. One cannot obtain reasonable results by simply changing a few material properties in current models. Also, using traditional approaches when making actual parts may be generating suboptimal results. Although many materials are available for use in RP technologies, this work concentrated on using stereolithography (SL), the original RP technology, to create polymer molds. The SL process uses photopolymer and laser energy to build a part layer by layer (Jacobs, 1996). Using SL takes advantage of both the commercial dominance of SL in the RP industry and the subsequent expertise base that has been developed for creating accurate, high-quality parts. More importantly, the use of SL photopolymers gives a mold material with properties that are very different from those of traditional metal molds. Until recently, SL was primarily used to create physical models for visual inspection and form-fit studies with very limited functional applications. However, the newer generation stereolithographic photopolymers have improved dimensional, mechanical and thermal properties making it possible to use them for actual functional molds. The photopolymers used in the study that generated the experimental data in this paper (Li, 1997) had improved glass transition temperatures and heat deflection temperatures. The glass transition temperature, Tg is the temperature at which amorphous polymers soften or harden. The heat deflection temperature is the temperature at which the polymer starts to distort under slight external load. Work has also been done to improve the elastic modulus of the photopolymer. Table I shows the range of properties for some of the photopolymers currently available. Table I Properties of cured SL photopolymers One key property difference between traditional metal molds and polymer molds is the thermal conductivity; the polymers used have a thermal conductivity that is less than one thousandth that of an aluminum tool. Previous work (Li, 1997) studied injection molding into a polymer SL shell backed with aluminum and also into a polymer SL mold that included cooling channels. Theoretical calculations predicted that these modifications increased the thermal conductivity by about ten and one hundred (respectively) when compared to the polymer alone. However, one is limited by the wall thickness and subsequent strength reductions caused by these modifications, as well as concerns over creating local hot spots. In using SL to create molds, the entire mold design and injection-molding process parameters need to be Modified and optimized from traditional methodologies due to the completely different tool material. To correct dimensions of parts that fall outside specifications of the desired part in injection molding, technologists routinely adjust molding machine settings, making use of known dependence of shrinkage on mold temperature and holding pressure (Hallum, 1997). This distortion during injection molding is due to shrinkage and warpage (uneven shrinkage) of the mold as well as the plastic part. Due to non-linearity of these relationships, and lack of access to accurate models and database, the influence of the molding conditions is typically not anticipated in the mold design (Hallum, 1997). One needs to quantitatively study the effect of various parameters on shrinkage and distortion in order to be able to make these adjustments. One also must consider incomplete solidification of the part when it is removed from the mold by being physically pushed out by ejectors or other means. Ejector pins acting on a still soft molding will stretch the material, or a very soft portion of a molding prematurely taken from the mold may flow slightly (Isayev, 1987). Thus there is a need to optimize the creation of SL molds and the process of injection molding into these molds. One of the problems is insufficient information availability regarding the physics of direct tooling. A complete set of the most important components, the dynamic relationships among them and their contribution to the part quality is yet to be clearly identified and documented. An effective and inexpensive alternative to obtaining information experimentally is via computer simulations. Reliable and targeted optimizations using systematic theoretical experiments can thus be obtained, which help to get a better understanding of the process. Methodology The goal of this research was to develop a simulation tool to predict the results of the injection process into an RP fabricated SL mold (Aluru and Keefe, 1999). Rather than develop an approach from scratch, the initial idea was to utilize existing simulation tools. However, one cannot simply run the existing codes. Using existing flow-analysis simulation software with aluminum and SL molds and comparing with experimental results (Li, 1997), though the simulation values of part distortion were reasonable for the aluminum mold, the results obtained for an SL mold were unacceptable. To illustrate this, the results for a dipole rectangular solid final part are reproduced in Table II. Table II Error in the slab geometry using existing software In order to simulate the use of an SL mold in the injection molding process, this work proposed a three-step process (Aluru and Keefe, 1999). Different software packages were used to accomplish this task. The main assumption was that temperature and load boundary conditions cause significant distortions in the SL mold, and current simulation packages fail to capture this. Step I The first step in the simulation involved flow analysis of the melt into a photopolymer mold. The part geometry was modeled as a solid model (the package IDEASTM was used), which was translated to a file readable by a flow analysis package (the package Moldflow was used). This analysis solved the mold-filling problem for the flow front, and gave the resulting temperature and pressure profiles. The inputs required to simulate the process were the molding conditions and the mold material properties. The important molding conditions, for the filling stage, were the melt temperature, mold temperature, injection time/flow rate and the packing pressures. The necessary mold material properties for the filling analysis were the thermal conductivity, heat capacity and the density. The material properties of the injected polymer included user inputs such as melt temperature as well as the physical properties of the material required for filling analysis such as thermal conductivity, specific heat, no-flow temperature, viscosity model, shrinkage data, etc. The mesh elements used in the flow analysis software were triangular shell elements. After the injection cycle had been completed, the outputs (temperature and pressure (load) profiles at the nodes) were obtained and used as input in the next step. Step II Structural analysis was then performed on the photopolymer mold model using the thermal and load boundary conditions obtained from Step I. Finite element analysis (the package IDEASTM was used) arrived at the distortions in the photopolymer mold due to the application of the two boundary conditions separately, as well as simultaneously. This was done in order to develop a sense of the relative importance of each of the boundary conditions and also to determine whether the resulting distortions could simply be superimposed to generate the combined case. The new dimensions of the mold after distortion were then used as input in the next step. This simulated the distortion that the mold underwent during the injection process. In reality, these distortions would occur simultaneously with the injection of the melt. This analysis solved the structural problem of mold distortion under thermal and load boundary conditions. The analysis inputs required for this step were the temperature and pressure boundary conditions (from Step I), the mold material properties and the geometry information (from a solid model CAD file) for the mold. As this was a structural analysis package, the additional mold material properties needed were Young’s modulus and the coefficient of thermal expansion. During the structural analysis, three-dimensional solid tetrahedral elements were used to mesh the mold halves. The outputs from this step were the distortions in the mold. Step III The third step involved modeling the distorted mold cavity (changes in the dimensions of the cavity were obtained from the changes in the mold from Step II), and injecting the melt into the distorted mold following the injection procedure outlined in Step I. Thus the distortions in the mold induced further distortions in the part. Analysis was now carried out on the final part, which gave the final distortion in the part. Since this step was again a flow analysis simulation, the inputs and the mold material properties were the same as those in Step I. The difference, however, was in the cavity model that now had the changes in the mold from Step II. The outputs from this step were the final distortions in the injection-molded part. Different part geometries were used for the simulations. This allowed comparison of the simulation results to previously generated experimental results (Li, 1997). The two geometries investigated were: (1)a thin slab (82.5mm long by 25.4mm wide by 2.54mm thick) that kept the model and tool design simple, shown in Figure 2; and (2)a complex generic part, shown in Figure 3, with a flat surface and thin walls: features common to many injection molded parts. For both the slab geometry and the complex geometry, in the experimental test (Li, 1997), the injected polymer was polypropylene with a melt temperature of 250℃ at 20.7MPa. The injection time was 2 sec. Figure 2 Slab geometry Figure 3 Complex part geometry For the slab geometry, the maximum linear shrinkages in the length and width directions were the significant shrinkage dimensions. No angular distortion of this part was considered and the warpage was measured as the maximum out-of-plane distance in the thickness direction when the part was placed on a flat length-width plane. The slab dimensions are shown in Figure 2 along with the x, y, and z directions as considered in the study. For the more complex part, the shrinkage in the length and width directions was important. Distortion also included the bow (displacement in the height direction) and the twist (the difference in angle between two initially parallel edges)-as shown in Figure 4 and Figure 5. Figure 4 Distortion in the complex geometry Figure 5 Twist distortion in the complex geometry Verification To assess the usefulness of the three-step model to simulate different cases, verification and validation were important. The issue in verification involved the simulation tools used whereas validation involved comparing model input-output transformations to real-world input-output transformations. Verification of the flow simulation software was accomplished by running the case of a simple model, to which the closed form/analytical solution was available. The model used here was one-dimensional isothermal heat transfer between the mold walls and the completely filled mold cavity (Greenberg, 1998). The boundary condition imposed was a constant mold wall temperature. The problem is shown schematically in Figure 6. Using the same material properties, boundary and initial conditions, the results obtained from the software were compared with the analytical solution. These did not match initially since the analytical problem assumed ideal and static conditions at the mold wall boundary. The software used, on the other hand, assumed a finite thermal conductivity of the mold, which resulted in changes in the mold wall temperature. The input properties of the mold material were modified in the flow analysis software to simulate a hypothetical material with very high thermal conductivity (thermal conductivity over 20 times normal). Also, the number of laminate layers was reduced in the mold wall. This helped to create the constant boundary wall temperature. The results obtained using these hypothetical material properties in the analytical solution compared well, as seen in Figure 7. Figure 6 On-dimensional heat transfer (verification problem) Figure 7 Verification of flow analysis software The exact solution to the problem involved the following equation: where: T: temperature across the mold : mold wall temperature : diffusivity b: thickness of cavity t: time x: distance across the mold For the finite-element structural analysis, mesh refinement was done to improve the accuracy of the solution (Bathe, 1992). As the number of elements was increased, the values of the dependent variables (distortions) converged to the experimental values. However, as expected, beyond a certain number of elements, the changes in the distortion values were not very significant. This gave the final mesh size. For the slab model, convergence was from a coarse mesh of 40 elements to 220 elements – after running simulations and comparing solutions for the different number of elements. Further refinement to 440 elements did not give a significant improvement in the results. Similarly, for the complex model, convergence was from a coarse mesh of 28 elements to a final mesh size of 260 elements. Validation results The validation was done by comparison to previous experiments that measured distortions of parts made by injecting polypropylene into aluminum and SL molds (Li, 1997). The distortions obtained from injection molding into aluminum and SL molds were compared with the ones from the simulation model. The validation results for the slab geometry are shown in Figures 8, 9, 10; and Figures 11, 12, 13 and 14 give the results for the complex geometry. These figures also include the distortion values of the direct simulations, using current injection-molding software in a one-step approach compared to the developed three-step model. Although these results are very encouraging for the slab geometry, more work needs to be done before the approach could be used to generate accurate absolute shrinkage and distortion values for a more realistic part geometry. However, the trends are reasonable and therefore those figures also include a