基于概率-凸集混合模型的汽车正面碰撞结构可靠性优化设计  

Reliability-based optimization design of structure subjected to vehicle frontal impact based on probability-convex hybrid model

作　　者：李方义[1,2] 荣见华[3] 胡林[3] 李凤玲[3] 易继军[3] 

机构地区：[1]长沙理工大学桥梁工程安全控制省部共建教育部重点实验室,长沙410114 [2]汽车运输安全保障技术交通行业重点实验室,西安710064 [3]长沙理工大学工程车辆安全性设计与可靠性技术湖南省重点实验室,长沙410114

出　　处：《振动与冲击》2016年第3期215-220,230,共7页Journal of Vibration and Shock

基　　金：国家自然科学基金(11302033;11372055;51408069);汽车运输安全保障技术交通行业重点实验室开放课题资助项目(2014G1502013);湖南省教育厅科研项目(13C1033);长沙理工大学桥梁工程安全控制省部共建教育部重点实验室开放基金资助项目(12KB04)资助

摘　　要：基于概率和凸集模型研究汽车正面碰撞可靠性优化设计问题。根据汽车吸能结构厚度、材料参数等不确定参数类型,分别采用概率和多椭球凸模型进行描述,以汽车安全性可靠性指标为约束,考虑汽车吸能结构质量为优化目标,建立混合模型可靠性优化设计模型。采用拉丁方试验设计构造目标函数和约束函数的Kriging近似模型,利用功能度量法评定概率约束,通过基于移动因子序列优化与可靠性评定将嵌套优化解耦为单层次优化。实际算例表明算法具有较高的计算效率及精度,对实际设计工作有一定参考价值。In order to describe the different types of uncertainty parameters for the thickness and material parameters of absorbing energy structure, probability and convex set models were adopted. The reliability indices for vehicle frontal impact safety were chosen as the reliability constraints and the weight of absorbing energy structure was selected as the objective of optimization. Then, an optimization design of structures subjected to vehicle frontal impact was formulated. The Kriging models for obiective and constraint functions were constructed by using the method of Latin hypercube experiment design. A performance measure method was adopted to evaluate probabilistic constraints. The nested optimization was transformed into single-level optimization problem by using the sequential optimization and reliability assessment （SORA） based on shifting factors. Some practical examples show the efficiency and accuracy of the presented algorithm, which could supply some beneficial references to designers.

关 键 词：概率 凸模型 混合模型 序列优化与可靠性评定 优化 

分 类 号：U463.82[机械工程—车辆工程] TH122[交通运输工程—载运工具运用工程]