| Gas Assisted Injection Molding (GAIM) is one of the most important and innovative molding process. In GAIM, the polymer melt is firstly filled or partially filled into mold cavity, and then the compressed gas is injected into the mold to drive the molten polymer further into the mold, leaving a hollow gas channel in the part. The GAIM can not only reduce warpage, clamping force and product weight, but also improve surface quality and save costs compared with the traditional injection molding. On the other hand, there are many unstable factors which can effects product quality of GAIM in actual production, it will result in the control of GAIM become harder compared with injection molding, because new parameters are introduced to GAIM. Fluctuation of design and parameters and the noise factor will lead to the increase of waste product rate too.With the development of computer technology and Finite Element Method (FEM) theory, rapid progress has been made in the numerical simulation of GAIM. The processing optimization of GAIM based on numerical simulation can reduce the numbers of“trial and error”, and help designers obtain best process parameters quickly and effectively. There are several issues in numerical simulation and processing optimization of GAIM:Firstly, the simulation based on Hele-Shaw hypothesis treats the gas imposed on polymer melt as the boundary conditions simply. It can't describe the flow mechanism of gas and melt. Secondly, the relationship between process parameters and design response is recessive and nonlinear, traditional optimization algorithm based on gradient is not suitable for this kind of problem, and traditional optimization algorithm is time consuming and costly especially for the complex parts. Last, there are several kinds of criteria for the quality of GAIM parts,different kinds of criteria is conflict usually, it can't be solved by weighted methods simply. Therefore, it is necessary to coordinate the relationship between different goals and reduce the sensitivity of processing fluctuations.These are some key technical problems which should be solved in GAIM now. The main topics of the thesis are as follows: Firstly, the melt and gas in mold should be described by unified N-S equations, the gas which is act on polymer melt is treated as fluid too, therefore the gas should be endow with governing equations, constant density and viscosity. The VOF method is obtained to trace the free-interface of gas and polymer melt. Finally, the gas penetration of GAIM is simplified to the computation of velocity and pressure fields of two phase flow.Based on the governing equations and common shape of GAIM parts, different typical gas channel were modeled and simulated. Compared with the Hele-Shaw hypothesis, the free-interface of gas and polymer melt can be simulated exactly, the variation of free-interface can be obtained, and the motion pattern of free-interface can also be captured when the gas pass through the corner of gas channel. The edge shape of gas channel and the fillet effect are interpreted by simulation, and the recommended values of gas channel are proposed too.Secondly, a hybrid optimization approach, which is integrates CAE, surrogate model and intelligent evolution algorithm, is proposed. The Adaptive particle swarm optimization (APSO) algorithm is developed by Matlab. In this approach, the CAE is adopted as experimental means, the Latin Hypercube Sampling method (LHS), which can fill the whole design space, is used to sampling in feasible design space, the Kriging surrogate model is adopted to approximate the nonlinear relationship between processing parameters and optimization objective, APSO algorithm is adopted to search optimal solutions. APSO can accelerate the convergence rate of optimization approach, and obtain optimal processing parameters using the least computational costs.The Multi-Objective optimization method is discussed. The Multi-Objective optimization approach which is based on surrogate model, CAE is proposed. The Multi-Objective PSO Based on Crowding Distance (MOPSOCD) is developed based on Matlab platform by introducing Crowding distance mechanism and mutation operator. The crowing distance mechanism and mutation operator can maintain the diversity of population.The external archive can constantly save the non-dominate solutions generated by algorithm, it can search the non-dominate solutions as many as possible and guarantee the global optimization. Integrating it with the optimization approach based on CAE and surrogate model, both the efficiency of approach and distribution of Pareto solutions are improved.According to the design failure caused by uncertain factors in real GAIM processing, the robust design is introduced to the optimization of GAIM. By introducing reliability design and 6-sigma quality theory, the design quality can be improved to 6-sigma level. The robust design criterion is analyzed. Robust design is generally focused on reducing response variation, balancing“mean on target”and“minimize variation”performance objectives. So, the 6-sigma robust design is treated as multi-objective problem in this thesis. The computation costs caused by CAE are solved by integrating Monte Carlo Simulation with surrogate model.Finally, by applying 6-sigma robust design to a car back mirror which is molding by GAIM, a series of Pareto robust solutions are obtained. Compared with the deterministic optimization, the reliability and robustness are improved significantly. The robust solutions obtained by this approach can provide more optional scheme for the processing design of GAIM. |
PDF Full Text Request