一种台灯罩注塑模具设计【1模2腔含11张CAD图】
一种台灯罩注塑模具设计【1模2腔含11张CAD图】【需要咨询购买全套设计请加QQ1459919609】图纸预览详情如下:
附 录 BInverse thermal mold design for injection moldsAddressing the local cooling demand as quality function for an inverse heat transfer problemCh. Hopmann1 P. Nikoleizig1Received: 26 September 2016 /Accepted: 7 December 2016 Springer-Verlag France 2016Abstract: The thermal mold design and the identification of a proper cooling channel design for injection molds becomes more and more complex. To find a suitable cooling channel system with objective rules based on the local cooling demand of the part a new methodology for the thermal mold design based on an inverse heat transfer problem was introduced. Based on a quality function regarding production efficiency as well as part quality, additional aspects to model the injection molding process are dis- cussed. Aim of those extensions is the improvement of the inverse optimization of the problem.Keywords: Injection molding, Thermal mold design, Inverse heat transfer problem, Heat transferIntroductionWith injection molding, increasingly complex components can be produced, but at the same time the requirements of the necessary injection mold rise. Concurrently due to the economic pressure, e.g. by global competition, strives for high efficiency and short production cycles are essential. Since the injection molding cycle is primarily characterized through the cooling of the melt into a dimensionally stable state, it is contiguous to focus on the cooling channel system of the injection mold for additional improvement in efficiency (Fig.1b). Usually, the cooling channels are realized through bores in the injection mold, which are connected by fittings to a complete channel system. Innovative technologies such as the selective laser melting (SLM) now enable the layered structured buildup of molds from metal pow- der. With this approach, the cooling channel system can be generated almost in any desired shape and course. The creation of a proper cooling channel system is a challenging task, also hindered by these opportunities and at the same time more complex parts.28Additionally thermal mold design phase is impeded due to particular thermoplastic materials, which are often used in technical parts and tend to a comparatively large shrinkage (Dependent on temperature and pressure) as a result of the crystallization process (as illustrated in Fig. 1a between points 3 to 5). This shrink- age causes stresses inside the part, if local differences in the shrinkage potential occur. Furthermore, the stresses can only be compensated through a deformation of the part. This so-called warpage may prevent the correct usage of the part and therefore must be avoided 1, 2.State of the artBesides the wish for a fast and efficient injection molding cycle, the aforementioned challenges lead to investigations to describe and simplify the thermal mold design phase. The efforts reach from a transfer of analytical approaches into the computer aided design to full mathematical and computational descriptions of the solidification process. Those efforts have a forward looking character and need an intense interpretation after the solution is calculated. A fully auto- mated thermal mold design phase is still not available.Fig.1Visualisation of the injection molding cycle with process variables (a) and a pie chart (b)Nowadays, this issue is progressively addressed through different research activities into a user independent optimization strategy for a proper cooling channel design (e. g. 35).Mehnen et al. rely on the use of evolutionary algorithms and model a mold29system based on light exchanging surfaces 3. The heat exchange is then calculated by a ray tracing method, which is faster than solving all governing equations. Spheres are used only in the first step of the sys- tem as parts to be analyzed. Maag and Kufer, in contrast, study a cluster algorithm which is combined with a branch and bound search algorithm to find the ideal cooling channel position 4. Contrarily, Fanacht et al. approach an auto- mated tempering system positioning by an artificial neural network, which covers numerous problems concerning temperature control 5. This also means that solutions may only come from the space of simulated training problems respectively from the possible interpolations in between. Common to all of these approaches is the forward headed nature which emphasizes the cooling channel system, but the evaluation is only possible after the simulation 35.Though also for those computer-aided optimizations a precise definition of the tempering system design is necessary in advance and essential for the quality of the result. Without knowledge regarding the local cooling demand for minimal part warpage and control of the polymer a targeted use of an optimization is not possible. Hassan et al. focus mainly on the criteria of part quality and study possibilities to realize a dynamic cavity tempering as well as a description of the influence of the cooling channel system on shrinkage and cooling of plastics 6. Thereby, an auto- mated generation of cooling channel systems is not the main focus of their work. Finally Agazzi et al. show a promising approach which is based on an inverse heat conduction problem 7, 8. Thus, in this case a part is defined as polymerwith homogeneous temperature. Along a given cooling area, surrounding the part, an optimized temperature distribution is calculated with a conjugate gradient algorithm in respect to a given objective function, which is based on fast heat removal as well as a homogeneous part temperature. Indeed, an inverse design is performed, but also along the analytical approach of thermal homogeneity.Setup of the proposed methodologyIn the light of the aforementioned technical development, the work of Agazzi et al. seems to be a promising starting point for further investigation.Besides a significant improvement for part warpage, their approach also shows some simplifications. For example the phases of the injection molding cycle are not modelled and implemented in the optimization. This refers especially to the injection and the holding pressure phase. Also the objective function refers to a rapid cooling and a homogeneous part temperature as the two aims 8. This approach seems30reasonable, but regarding to the phases of the injection molding cycle and the temperature and pressure dependent pvT-behavior, a different design of the objective function can be considered. Also the derived cooling channels require further investigations.Within the framework of the here proposed expanded methodology a model is favored, which also considers the following aspects. First the methodology should be able to include more phases of the injection molding process. So it is based on a conventional injection molding simulation. Results for pressure, temperature and inner properties can be exported at different stages of the process as boundary conditions for the optimization. So the expanded methodology is based on a hybrid simulation approach, which connects injection molding simulation with optimizing an inverse heat transfer problem. Also the objective function should be redesigned carefully. On one hand, the design should address minimal cycle times to fulfil the request for an efficient process like the one used by Agazzi et al.On the other hand, also part quality should be addressed, which refers to mechanical, visual and geometric requirements. Whereas mechanical and visual properties can be met by appropriate slow cooling rates, especially geometric proper- ties, paraphrased the dimensional accuracy of the part, turn out to be a severe element of the thermal mold design phase. By using a proper tempering system, a locally homogeneous shrinkage should be targeted to minimize the tendency of the part to warp. The objective of the analysis is to bring local heat and cooling demand of the part in equilibrium to local heat and cooling supply of the molds tempering system.The postulate of an even shrinkage potential can be modelled via homogeneous local densities as an objective function, so that the problem is still addressable as an inverse heat conduction problem 9.A modified exemplary extended objective function to be introduced to the31methodology is given in Eq. 1.This objective function Q(TC ) addresses a quick cooling through the first term, where a desired ejection temperatureTEjec for the surface 1 of the part is given and compared to the actual local temperatures Tloc (xi,t,TC) of the part. The second term addresses density homogeneity, with the differences of local density loc (xi,t,TC) compared to a mean density Ejec, which should be reached on a surface 2 within the part. Both terms are integrated over their respective areas 1/2 and can be weighted with the variable wm/k. The temperatures Tc on the outer mold contour according to Fig. 2b are then varied to minimize the quality function.For the proposed approach the exact modeling of the designed methodology will be carried out as a hybrid simulation approach, containing an injection molding simulation as input for a heat conduction simulation calculated with a multiphysics simulation. With this hybrid approach, all plastics related properties and more phases of the injection molding cycle can be modelled and made available for a thermal optimization at the same time.With the presented objective function an exemplary cooling channel system for a plate shaped specimen with ribs is analyzed 10,13,14. Measurements of the specimen are shown on Fig. 2a. Simultaneously, the specimen contains typical elements of injection molded parts with three ribs in different heights. The thickness of the specimen is 1.5 mm, which is typical for injection molded parts. Based on the specimens geometry a cooling area is generated, with a constant distance to the part and an area inside the part, for which the objective function is solved (see Fig. 2b). The specimen is optimized using a 2D calculation approach, in order to save computing time.After solving the optimization for density and cooling time, as stated in Eq.1 also with a conjugate gradient algorithm, the cooling channels can be derived from isothermal lines of the desired mold temperature of 80C 10. The gradient algorithm follows the steepest ascent of the objective function and calculates the necessary temperature distribution along the outer mold contour defined as cooling area to minimize the objective function. The input datafields for the optimization are shown in Fig.3a. Based on the result of the optimization, which is shown in Fig. 3b, the cooling channels were identified. Although, temperature distribution of the optimization lead to very low temperatures of 100 C, this distribution can be used to32derive the contour of a cooling channel by using an isothermal line. Those derived 2D channel contours are then modelled as extruded 3D geometries. With the injection molding simulation software Sigmasoft, Sigma engineering GmbH, Aachen, Germany, an entire 3D injection molding simulation is set up using the boundary conditions presented in Table 1. The implemented material is a widely used polyamide 6 (unfilled B 30 S) of Lanxess AG, Cologne, Germany (see Table 2 for properties). In Fig.4 two different setups of the specimen are presented and compared. One is modelled without cooling channels as a neutral reference and a second one with the derived cooling channels from the optimization. Figure 4a illustrates the resulting temperature distribution inside the mold. Figure 4b shows the resulting warpage of the specimen. Comparing these two cases, a significant reduction of the specimens warpage can be achieved, with the second case resulting in a relatively low warpage of the specimen. Merely, the end of the ribs show bigger deviations from the original geometry in both cases. Here, it has to be noted, that in the scope of the optimization, heat has to be brought into the system as shown in Fig. 3b. This demand is not yet considered, as the standard process uses just cooling. In addition the natural thermal shrinkage of the part has to be taken into account, which is already included in the results, due to the simulation software. This natural shrinkage is not part of the optimization yet.Fig. 2 Specimen with ribs and measurements (a) and outer contour around part(b)Fig. 3 Mold contours, initial temperatures (a), optimization,results and derived33cooling channel system (b)a) Initial data of temperature and pressure distributionb)Temperature distribution at optimization pointFurther extension of methodologyWith the principle functionality of the methodology shown, it will be investigated to what extent a more accurate.Table 1 Settings for the optimization calculation and the injection molding simulation of cooling and shrinkage and warpage.ParameterMelt temperature 270 CEjection temperature 110 CCooling fluid temperature 80 CInjection pressure 1000 barHolding pressure 800 barCycle time 5.6 sHandling time 1.5 sInjection time 0.248 sHolding pressure time 3.1 sTable 2 Material properties of the implemented mold and plastic 11, 1234ParameterDensity steel 7830 kg/m3 Heat conductivity steel 46.5 W/m K Heat capacity steel 440 J/kg KDensity modeling PA 6 Tait-approachacc. to material supplierViscosity modeling PA 6Cross-WLF-approach acc. to material supplierHeat conductivity PA 60.2 W/m KHeat capacity PA 6 2390 J/kg KHeat transfer coefficients 2000W/m2Fig.4 Resulting temperatures (a) and warpage (b) of the specimen for a 3D simulation without (upper part) and with automatically derived cooling channels for minimal part warpage (lower part)a) Temperature distribution at the end of the cooling phase after 15 cycles Without cooling channels:With derived cooling channels:b) Deformation of the part: Visual amplification factor: 2035Modeling of the injection molding process can increase the quality of the results, in comparison to the work of Agazzi et al.This will be presented with the following three additional aspects. The implementation of a multi- cycle approach should be discussed first. The influence of a modeling of the handling times will follow. And finally the modeling of the injection phase will be investigated. Unless specifically stated, besides heat conductivity of the mold, which is changed to more suitable mold making steel of 25.3 W/m K, all other material properties and boundary conditions remain the same 14.Implementation of multicycle analysisUsually the injection molding process is used to produce a large quantity of molded parts with the same geometry.This is done through the periodic repetition of the molding. Assuming a uniform temperature distribution in the mold at the beginning of the first cycle, a temperature distribution adjusts itself after some cycles 1. Local cooling demand for warpage minimized cooling of the part should be deter- mined in the stable state, because it is otherwise influenced by the heat change in the mold. For this reason, the model used for the generation of heating/cooling systems should use the initial values of a stable cycle for the optimization. Since these temperatures are but strongly dependent on the cooling channel system to be determined, an estimation or determination before optimization is not useful. Therefore, the method is expanded so that it determines the initial temperature field in the mold itself. The methodology so far uses, as well as the approach of Agazzi et al,a multicycle approach, but is now compared to a single cycle setup.Fig.5 Influence of multicycle analysis on minimal part temperatures (a) and quality function (b)Z01: Optimization covering one cycle Z15: Optimization covering 15 cycles Minimal temperature within part CQuality function value -Time s36TimesFig. 6 Results of optimization(a) and derived cooling channel contours (b)a) Temperature distribution at optimization point Z01: Optimization covering onecycleZ15: Optimization covering 15 cyclesb) Contours of isothermal lines of T= 80 CFor the quality function, still only the last cycle should be considered to save computation time.First, it is examined how many cycles are necessary to achieve a stable cycle at the time of optimization. To check this, the boundary conditions determined in the optimization are used to perform simulations with additional cycles. If the temperature field in these additional cycles does not change anymore, it can be assumed, that steady state is achieved. Figure 5a shows the curves of minimum temperature in the cavity for 25 cycles. It has to be recognized that the mini- mum37temperatures using the boundary conditions, which are determined by optimizing one cycle (Z01, black line), first take a nearly periodic course after about ten cycles. At the time at which the boundary conditions are optimized, the process is therefore still not in steady state. The significant changes of the temperature in the later cycles arise because the cooling demand is determined not only for the temperature of the part, but also for the temperature of the mold. The blue curve in the graph a) represents the minimum temperature in the part, which is calculated with the help of the boundary conditions obtained from the optimization over 15 cycles (Z15, blue line). The curves of the quality function value of the two calculations are shown in Fig. 5b. It can be seen that the quality function value with the boundary conditions determined in the optimization of Z15 remains low in the long run. In the calculation, which uses the boundary conditions determined in the optimization Z01, the value of the quality function increases from 0.0083 at the end of the first cycle to 0.6477 at the end of the 25th. According to the modeling of the methodology, which is used to derive the quality function, this should correspond to a significant increase of the part warpage.Fig. 7 Resulting temperature distribution (a) and warpage (b) without and with multi cycle considerationa) Temperature distribution at the end of the cooling phaseZ01:Optimization covering one cycleZ15:Optimization covering 15 cyclesb) Deformation of the part38Table 3 Warpage of the cooling channel systems Z01 and Z15Total deformation Measuring Measuring Measuringpoint1mm point2mm point 3 mmZ01 1.890 1.247 2.848Z15 1.616 1.007 1.807Since the change of the quality function is less than 1 % when viewing 15 cycles against 25 cycles, thus 15 cycles can be seen as a useful optimization range. Figure 6a shows the temperature distribution at the point of optimization of the setups Z01 and Z15. The local cooling demand is estimated on the basis of the reference temperatures. In both cases temperature distribution alternates between high and low values compared to the calculated reference temperatures. The reference temperatures, which are determined in optimizing Z01, vary much more than those which are determined in the optimization of Z15. Because in the inner side of the corners the cooling demand is greater than in the plate-shaped sections, very low reference temperatures are intended here. Figure 6b also shows the 80C isothermal lines of the temperature for three different cycle numbers (1 cycle, 15 cycles and 50 cycles). The isothermal lines optained by Z15 do not differ significantly from those obtained by the optimization at 50 cycles. However, the difference to the isothermal lines of Z01 is considerable. This is evident, for example, in the lower areas of the ribs, where more isothermal lines lines emerge for the multicycle optimization Z15.The cooling channel systems are analysed in conventional injection molding39simulations to investigate the effects of the differences between the systems on the value of the quality function and the part warpage. For the cooling channel system a heat transfer c
收藏