基于响应面和ASA的抑爆球注塑质量多目标优化  

Multi-Objective Optimization of Explosion Suppression Ball Injection Molding Quality Based on Response Surface Model and ASA Algorithm

作　　者：邬斌扬[1] 周天意 于洋洋 季宁[2] 张俊红[1,2] 孟祥德[2,3] WU Binyang;ZHOU Tianyi;YU Yangyang;JI Ning;ZHANG Junhong;MENG Xiangde(State Key Laboratory of Engine,Tianjin University,Tianjin 300072,China;Tianjin University Renai College,Tianjin 301636,China;Tianjin Jieqiang Power Equipment Co.,Ltd.,Tianjin 300410,China)

机构地区：[1]天津大学,内燃机燃烧学国家重点实验室,天津300072 [2]天津大学仁爱学院,天津301636 [3]天津捷强动力装备股份有限公司,天津300410

出　　处：《塑料》2021年第5期147-155,共9页Plastics

基　　金：天津市教委科研计划(2019KJ152);天津市技术创新引导专项基金(20YDTPJC02020)。

摘　　要：建立了抑爆球一模八腔仿真分析的有限元模型,基于注塑试验对模型进行验证,采用仿真分析可知,最大锁模力和隔片收缩率对成型质量的影响较大,但两者呈负相关,将其作为优化目标,以5种注塑参数作为影响因素,应用最优拉丁超立方抽样法和不同阶响应面模型建立两者的数学关系。通过计算复相关系数可知,四阶响应面模型的精度最高,并且将其作为代理模型。利用自适应模拟退火算法(ASA)对代理模型进行自主寻优,得到满足注塑质量要求的最优工艺参数组合,使提高抑爆球的生产效率显著提高。优化后最大锁模力为535.53 k N,与优化前相比,降低了12.15%;优化后隔片收缩率为4.449%,与优化前相比,降低了20.04%,这表明,提出的优化方法效果显著。The finite element model for simulation analysis of a mould containing eight cavities of explosion suppression ball was established,and the model was verified based on the injection experiment.The simulation analysis showed that the maximum clamping force and the shrinkage rate of the spacer had great influence on the molding quality,but they were mutually conflicting,so were taken as the optimization objects,and the five injection parameters were taken as the influencing factors.The optimal Latin hypercube sampling method and different order response surface model were used to establish the mathematical relationship between the optimization objective and influence factors,and the complex correlation coefficient was calculated.It showed that the fourth-order response surface model had the highest accuracy,which was used as the surrogate model.The adaptive simulated annealing algorithm(ASA)was used to optimize the surrogate model,and the optimal combination of processing parameters meeting the injection quality requirements was obtained,which improved the production efficiency of explosion suppression ball.After optimization:The maximum clamping force was 535.53 k N,which was 12.15%lower than that before optimization;the shrinkage rate of spacer was 4.449%,which was 20.04%lower than that before optimization,which could prove that the optimization method was effective.

关 键 词：抑爆球 多目标优化 最优拉丁超立方 响应面模型 ASA算法 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]