混合不确定性下的多学科协同可靠性分析方法  

Multidisciplinary collaborative reliability analysis method under mixed uncertainties

作　　者：张立香 孟欣佳 敬石开[2] 张贺 ZHANG Lixiang;MENG Xinjia;JING Shikai;ZHANG He(College of Mechanical and Equipment Engineering,Hebei University of Engineering,Handan 056038,China;School of Mechanical Engineering,Beijing Institute of Technology,Beijing 100081,China;Beijing Shenzhou Aerospace Software Technology Co.,Ltd.,Beijing 100094,China)

机构地区：[1]河北工程大学机械与装备工程学院,河北邯郸056038 [2]北京理工大学机械与车辆学院,北京100081 [3]北京神舟航天软件技术有限公司,北京100094

出　　处：《计算机集成制造系统》2019年第9期2198-2207,共10页Computer Integrated Manufacturing Systems

基　　金：国家重点研发计划资助项目（2017YFB1102804）;河北省青年科学基金资助项目（E2018402229）;邯郸市科技研发计划资助项目（1721203048-3）;河北工程大学博士基金资助项目（17129033124）~~

摘　　要：多学科可靠性设计优化被认为是解决不确定性因素影响下复杂产品设计优化问题的有效方法,然而高额的计算花费严重影响了其在实际中的运用。多学科可靠性分析作为多学科可靠性设计优化的重要组成部分,对其计算效率起着直接的主导作用。目前,混合不确定性下的多学科可靠性分析仍是嵌套或顺序执行模式,易导致复杂产品可靠性分析时学科自治性差、效率不高的问题。为此,针对混合不确定性下的多学科可靠性分析问题,提出一种混合不确定性下的多学科协同可靠性分析方法。该方法首先将多学科极限状态函数向标准正态空间转换,利用区间可靠性分析的KKT条件进行模型转化,构建多学科可靠性分析整体模型,然后将整体模型分解成几个并行的单学科优化问题,并利用一致性约束对各学科进行协调,实现多学科可靠性分析的并行求解。通过数值算例和复合缸设计实例,验证了所提方法的可行性和有效性。Reliability-based Multidisciplinary Design Optimization(RBMDO)has been considered as an effective method to solve the problem of complex product design optimization under the uncertainties.However,due to its excessive computational cost,there have some difficulties in the application of RBMDO.As an important part of RBMDO,multidisciplinary reliability analysis plays a direct leading role in the computational efficiency of RBMDO.The current multidisciplinary reliability analysis under mixed uncertainties has still been a nested or sequential execution mode,which is easy to cause the problems of poor discipline autonomy and expensively computational cost in the reliability analysis of complex products.Aiming at the multidisciplinary reliability analysis problem under mixed uncertainties,a multidisciplinary collaborative reliability analysis method was proposed to solve the computational problems.The method transformed the multidisciplinary limit state function into the standard normal space,and then the KKT conditions of interval reliability analysis were used to conduct model transformation for constructing the integral model of reliability analysis.The integral model was decomposed into several parallel single-subject optimization problems,and the condition of consistency constraints was used to coordinate all disciplines,which could realize the parallel solution of multidisciplinary reliability analysis.A numerical example and a composite cylinder design example were used to verify the feasibility and effectiveness of the proposed method.

关 键 词：多学科可靠性分析 混合不确定性 KKT条件 并行求解 协同 

分 类 号：TB114.3[理学—概率论与数理统计]