工程设计学报2017，Vol.24Issue(5)：536-544,9.DOI:10.3785/j.issn.1006-754X.2017.05.008
基于自适应SVR-ELM混合近似模型的镁合金差温成形本构参数反求
Constitutive parameter inverse for nonisothermal stamping of magnesium alloy based on adaptive SVR-ELM mixture surrogate model
摘要
Abstract
To improve the accuracy of finite element model for nonisothermal stamping, a new method of parameter inverse based on adaptive SVR-ELM mixture surrogate model and quantum genetic algorithm was presented. Based on the finite element model for nonisothermal stamping of magnesium alloy, the SVR-ELM ensemble surrogate model was established between parameters of Johnson-Cook constitutive model and temperature of parts. The weight coefficient of ensemble surrogate was calculated by heuristic algorithm. Based on adaptive method, sample spaces and surrogate model could be updated by local optimal solution obtained during optimization process. The optimum constitutive parameters for nonisothermal stamping of magnesium alloy AZ31B would be obtained by using quantum genetic algorithm. Compared with single surrogate model, adaptive SVR-ELM mixture surrogate model had higher accuracy and inverse efficiency. Taking NUMISHEET2011 cross-shaped cup part as an example, optimal material constitutive parameters were obtained by inverse model. Compared with the inverse results of test data, the error of temperature was 0.39%, which showed that this method was effective. The results indicate that the finite element model established based on constitutive parameters obtained by SVR-ELM can predict forming parts more effectively in the actual production.关键词
差温成形/自适应SVR-ELM混合模型/量子遗传算法/参数反求Key words
nonisothermal stamping/adaptive SVR-ELM mixture surrogate model/quantum genetic algorithm/parameter inverse分类
矿业与冶金引用本文复制引用
唐维,谢延敏,黄仁勇,张飞,潘贝贝..基于自适应SVR-ELM混合近似模型的镁合金差温成形本构参数反求[J].工程设计学报,2017,24(5):536-544,9.基金项目
国家自然科学基金资助项目(51005193) （51005193）
国家大学生创新创业训练计划项目(201710613033) （201710613033）
