基于Kriging和NSGA-Ⅱ的辅以困气改善的熔接痕优化策略  

Optimization strategy of weld line assisted by air traps improvement based on Kriging and NSGA-Ⅱ

作　　者：王梦寒[1] 涂顺利 余春丽 WANG Menghan;TU Shunli;YU Chunli(College of Material Science and Engineering,Chongqing University,Chongqing 400030,China)

机构地区：[1]重庆大学材料科学与工程学院,重庆400030

出　　处：《化工学报》2018年第10期4449-4455,共7页CIESC Journal

基　　金：中央高校基本科研业务费专项资金项目(CDJZR14130006)~~

摘　　要：从熔接痕的成型机理出发,考虑了困气状况和熔合温度对熔接痕熔合质量的影响,提出一种以困气改善辅助提升熔接痕熔合质量的方法,建立了以"气穴面积"和"熔接痕处流动前沿温度"为熔接痕熔合质量指标的多目标评价体系,并构建了一种基于Kriging模型和非支配排序遗传算法(NSGA-Ⅱ)的多目标寻优策略,以实现工艺参数的择优选择。以多孔板熔接痕优化为例,选用拉丁超立方抽样(LHS)设计实验,建立并检验反映工艺参数和优化目标之间影响规律的Kriging模型,采用NSGA-Ⅱ确定多目标优化问题的Pareto前沿及其最优解。仿真及生产实验结果表明所建立的多目标优化策略能有效实现熔接痕的多目标优化,同时以困气改善辅助提升熔接痕熔合质量的方法效果显著。Based on forming mechanism of weld line,effects of air traps condition and melt temperature on bonding quality of weld line were taken into consideration.Thus,an optimization strategy of weld line assisted by improving air traps condition was proposed.And a multi-objective evaluation system was established to evaluate bonding quality of weld line from two aspects,namely area of air traps and temperature at flow front in weld line.A multiobjective optimization methodology based on Kriging surrogate model and nondominated sorting genetic algorithmⅡ(NSGA-Ⅱ)was proposed to optimize process parameters.Optimization of weld line of a perforated plate was given as an example.Latin hypercube sampling(LHS)was utilized to arrange experiments.A surrogate model based on Kriging was established and verified to map relationship between process parameters and optimization objectives.Then NSGA-Ⅱwas used to gain Pareto Front and optimal solution of multi-objective optimization problem.Simulated and practical confirmation experiments were carried out to verify validity of obtained optimal solution.Results show that established multi-objective optimization methodology is effective to multi-objective optimization problems of weld lines.And proposed optimization strategy of weld line assisted by improving air traps condition was proved to be highly effective.

关 键 词：模塑 熔接痕 困气 多目标优化 数值模拟 KRIGING模型 NSGA-Ⅱ 

分 类 号：TQ320.66[化学工程—合成树脂塑料工业]