【作者】 褚宝磊；

【导师】 于同敏；

【作者基本信息】 大连理工大学 ， 机械制造及其自动化， 2009， 硕士

【摘要】 注塑成型是塑料成型加工领域的重要成型方法之一,各工业领域及人们日常生活中使用的各种塑料制品,大多都是通过注塑成型方法生产的。但随着现代工业产品的快速发展,人们对注塑成型制品的质量及性能要求越来越高。由于注塑制品的成型是一个十分复杂的多因素耦合作用下的非线性动态加工过程,这一过程中的任何因素都可能对制品的成型质量产生重要影响,进而导致制品的质量缺陷。因此,如何能在多因素作用下的非线性动态加工过程中,保证注塑成型制品质量稳定可靠,并对其实施有效的预测与控制,一直是注塑成型加工领域始终追求的目标之一。本文在深入分析了注塑成型工艺过程和制品缺陷类型及产生原因的基础上,针对制品成型过程中的非线性及多因素耦合作用对塑件质量的影响,提出了一种基于偏最小二乘法和人工神经网络相结合的质量预测方法。并以塑件重量和熔接痕处的断裂强度为评价指标,通过对实际塑件成型时的模具温度、熔体温度、注塑压力、注塑速度、保压压力、保压时间等工艺参数对塑件重量和熔接痕处强度的影响进行分析,建立了塑件质量预测模型。通过单因素和随机选取的工艺参数试验对模型进行验证,结果显示,塑件熔接痕处断裂力的预测值和试验测量值的最大误差为1.55%,塑件重量预测值与测量值的最大误差仅为0.28%,表明建立的质量预测模型能够较为准确地反映注塑工艺参数和制品质量指标之间的变化关系。基于所建质量预测模型的计算结果,再利用遗传算法对工艺参数进行寻优,便可得到优化的工艺参数。应用优化的工艺参数进行的注塑成型试验表明,建立的预测模型和工艺参数寻优模型能够有效地保证塑件质量。更多还原

【Abstract】 Injection molding is one of the important methods in plastic molding fields. Most of plastic parts used in industrial fields and people’s daily life are produced through injection molding. With the rapid development of modern industrial products, people have higher and higher requirements on the quality and properties of parts. As the injection molding is a very complex multi-factor coupling, nonlinear, dynamic processes, any factor in the process has an important influence on the part’s quality, which may lead to product quality defects. Therefore, how can ensure stable and reliable product quality, effective prediction and control, in the multiple factors, nonlinear, dynamic process, has always been one of pursuing goals in the field of injection molding.Based on the deep analysis of injection molding process and the type and cause of parts defects, considered the non-linear and multi-factor coupling effect on the quality of plastic parts, a quality prediction method is proposed using PLS combined with ANN. Moreover, a forecasting model is set up, taking the part weight and fracture force of weld line as evaluation indexes, through effect analysis of mold temperature, melt temperature, injection pressure, injection velocity, holding pressure and holding time. Test experiments with single factor and process parameters randomly selected have been used to check out the prediction model. The results show that the maximum error of fracture force of weld line and measured value is 1.55%, while the part weight is only 0.28%. This indicates the forecasting model can comparatively accurately reflect the influence relation of the injection process parameters on part quality indexes.Based on the results of quality prediction model, genetic algorithm has been used to optimize the process. Optimal process parameters are obtained in the end. According to the optimal process parameters, injection molding experiments are done. The results show that the plastic parts’ quality is ensured through prediction model and process parameter optimization model.更多还原