基于RMQGS-APS-Kriging的主动学习结构可靠性分析方法

doi:10.3901/JME.2022.16.420

机械工程学报 ›› 2022, Vol. 58 ›› Issue (16): 420-429.doi: 10.3901/JME.2022.16.420

扫码分享

基于RMQGS-APS-Kriging的主动学习结构可靠性分析方法

智鹏鹏^1,2, 汪忠来¹, 李永华³, 田宗睿³

1. 电子科技大学机械与电气工程学院成都 611731;
2. 电子科技大学广东电子信息工程研究院东莞 523808;
3. 大连交通大学机车车辆工程学院大连 116028

收稿日期:2021-09-01 修回日期:2022-05-01 出版日期:2022-08-20 发布日期:2022-11-03
通讯作者: 汪忠来(通信作者)，男，1980年出生，博士，教授，博士研究生导师。主要研究方向为可靠性设计、稳健设计、模型验证。E-mail：wzhonglai@uestc.edu.cn
作者简介:智鹏鹏，男，1989年出生，博士，助理研究员。主要研究方向为韧性设计、机电装备结构/疲劳可靠性、不确定性分析与优化。E-mail：zhipeng17@yeah.net
基金资助:
广东省基础与应用基础研究基金(2021A1515110308)、四川省杰出青年科技人才(2020JDJQ0036)和四川省自然科学基金(2022NSFSC1941)资助项目

RMQGS-APS-Kriging-based Active Learning Structural Reliability Analysis Method

ZHI Pengpeng^1,2, WANG Zhonglai¹, LI Yonghua³, TIAN Zongrui³

1. School of Mechanical and Electrical Engineering, University of Electronic Science and Technology of China, Chengdu 611731;
2. Institute of Electronic and Information Engineering in Guangdong, University of Electronic Science and Technology of China, Dongguan 523808;
3. School of Locomotive and Rolling Stock Engineering, Dalian Jiaotong University, Dalian 116028

Received:2021-09-01 Revised:2022-05-01 Online:2022-08-20 Published:2022-11-03

摘要/Abstract

摘要： 为兼顾具有黑箱问题的机械产品结构可靠性分析精度和效率，提出一种基于随机移动四边形网格抽样(Random moving quadrilateral grid sampling, RMQGS)和交替加点策略(Alternate point strategy, APS)的Kriging(RMQGS-APS-Kriging)主动学习结构可靠性分析方法。采用RMQGS方法选择初始样本点并计算其真实性能函数值，结合差分进化算法(Differential evolution, DE)，获得高精度优化Kriging代理模型；通过欧式距离构造抽样限定区域，确定交替加点的样本选取范围，依据迭代次数采用主动学习U函数和改进EI(Improved EI, IEI)函数交替筛选最佳样本点，并加入到每次迭代的样本库以更新优化Kriging代理模型；利用子集模拟(Set simulation, SS)方法对迭代过程中优化Kriging代理模型拟合的性能函数进行可靠度计算，并通过收敛准则确定最终的结构可靠度。算例分析表明，与传统基于代理模型的可靠度计算方法相比，所提方法具有较强的局部和全局性能函数拟合能力，能够以较少的性能函数调用次数和可靠度计算时间精确估算失效概率。

关键词: 结构可靠性, Kriging代理模型, 主动学习, 子集模拟

Abstract: In order to take into account the accuracy and efficiency of structural reliability analysis of mechanical products with black box problems, a novel Kriging method named RMQGS-APS-Kriging for active learning structural reliability analysis based on random moving quadrilateral grid sampling (RMQGS) and alternate point strategy (APS) is proposed. The RMQGS method is used to select the initial sampling points and estimate the real performance function. Combined with the differential evolution (DE) method, the Kriging-based surrogate model with high accuracy is obtained. Through Euclidean distance, the sampling limited area is constructed to determine the sample selection range of alternate points. According to the number of iterations, the active learning U function and the improved EI (IEI) function are employed to alternately select the best sampling points, which are added to the sample database of each iteration to update the optimized Kriging surrogate model. The subset simulation (SS) method is then used to calculate the reliability of the performance function fitted by the optimized Kriging surrogate model in the iterative process, and the final structural reliability is determined by convergence criteria. The examples analysis show that compared with traditional surrogate model-based reliability calculation methods, the proposed method has stronger local and global performance function fitting ability, and can estimate the exact failure probability with less times of performance function calls and reliability calculation time.

Key words: structural reliability, Kriging surrogate model, active learning, subset simulation

中图分类号:

U270

智鹏鹏, 汪忠来, 李永华, 田宗睿. 基于RMQGS-APS-Kriging的主动学习结构可靠性分析方法[J]. 机械工程学报, 2022, 58(16): 420-429.

ZHI Pengpeng, WANG Zhonglai, LI Yonghua, TIAN Zongrui. RMQGS-APS-Kriging-based Active Learning Structural Reliability Analysis Method[J]. Journal of Mechanical Engineering, 2022, 58(16): 420-429.

参考文献

[1] ZHOU Tong, PENG Yongbo, LI Jie. Structural reliability analysis using probability density evolution method and adative model[J]. Journal of Vibration Engineering, 2020, 33(5): 1035-1043. 周通, 彭勇波, 李杰. 结构可靠度分析的概率密度演化理论——自适应代理模型方法[J]. 振动工程学报, 2020, 33(5): 1035-1043.
[2] YANG Zhijun, CHEN Chaoran, HUANG Guanxin. Hybrid global optimization method based on dynamic kriging metamodel and gradient projection method for optimal design of robot[J]. Journal of Mechanical Engineering, 2019, 55(11): 61-68. 杨志军, 陈超然, 黄观新. 面向机器人优化设计的GA-非均匀Kriging-梯度投影混合全局优化算法[J]. 机械工程学报, 2019, 55(11): 61-68.
[3] HAN Zhonghua. Kriging surrogate model and its application to design optimization: A review of recent progress[J]. Acta Aeronautica et Astronautica Sinica, 2016, 37(11): 3197-3225. 韩忠华. Kriging模型及代理优化算法研究进展[J]. 航空学报, 2016, 37(11): 3197-3225.
[4] WEI Juan, ZHANG Jianguo, QIU Tao. Structural reliability algorithm based on improved dynamic kriging model[J]. Journal of Beijing University of Aeronautics and Astronautics, 2019, 45(2): 373-380. 魏娟, 张建国, 邱涛. 基于改进的动态Kriging模型的结构可靠度算法[J]. 北京航空航天大学学报, 2019, 45(2): 373-380.
[5] HUANG D, ALLEN T T, NOTZ W I, et al. Global optimization of stochastic black-box systems via sequential kriging meta-models[J]. Journal of Global Optimization, 2012, 54(2): 431-431.
[6] LI Yonghua, LIANG Xiaojia, GONG Qi. Improved kriging response surface reliability calculation method based on two-point addition strategy[J]. China Mechanical Engineering, 2019, 30(17): 2051-2057. 李永华, 梁校嘉, 宫琦. 基于双点加点策略的改进Kriging响应面可靠度计算方法[J]. 中国机械工程, 2019, 30(17): 2051-2057.
[7] LI Z, RUAN S, GU J, et al. Investigation on parallel algorithms in efficient global optimization based on multiple points infill criterion and domain decomposition[J]. Structural & Multidisciplinary Optimization, 2016, 54(4): 747-773.
[8] LI Guofa, CHEN Zequan, HE Jialong. New adaptive sampling strategy for structural reliability analysis[J]. Journal of Jilin University, 2021, 51(6): 1575-1581. 李国发, 陈泽权, 何佳龙. 新型结构可靠性分析自适应加点策略[J]. 吉林大学学报, 2021, 51(6): 1575-1581.
[9] SHU Z, JIRUTITIJAROEN P. Latin hypercube sampling techniques for power systems reliability analysis with renewable energy sources[J]. IEEE Transactions on Power Systems, 2011, 26: 2066-2073.
[10] GAO Y H, Wang X C. An effective warpage optimization method in injection molding based on the Kriging model[J]. The International Journal of Advanced Manufacturing Technology, 2008, 37: 953-960.
[11] YUAN Xiukai, KONG Chongchong, GU Jian. Reliability analysis approach based on kriging and advanced first-order second moment method[J]. Journal of National University of Defense Technology, 2020, 42(6): 150-156. 袁修开, 孔冲冲, 顾健. Kriging与改进一次二阶矩融合的可靠性分析方法[J]. 国防科技大学学报, 2020, 42(6): 150-156.
[12] HUANG X X, CHEN J Q, ZHU H P. Assessing small failure probabilities by AK-SS: An active learning method combining kriging and subset simulation[J]. Structural Safety, 2016, 59: 86-95.
[13] SHI Zhaoyin, LÜ Zhenzhou, LI Luyi, et al. Cross-entropy importance sampling method based on adaptive kriging model[J]. Acata Aeronautica et Astronautica Sinica, 2020, 41(1): 223123. 史朝印, 吕震宙, 李璐祎, 等. 基于自适应Kriging代理模型的交叉熵重要抽样法[J]. 航空学报, 2020, 41(1): 223123.
[14] ECHARD B, GAYTON N, LEMAIRE M, et al. A combined importance sampling and Kriging reliability method for small failure probabilities with time-demanding numerical models[J]. Reliability Engineering & System Safety, 2013, 111: 232-240.
[15] CADINI F, SANTOS F, ZIO E. Passive systems failure probability estimation by the meta-AK-IS 2 algorithm[J]. Nuclear Engineering & Design, 2014, 277(1): 203-211.
[16] XIE Yanmin, ZHANG Fei, PAN Beibei, et al. Optimization of drawbeads based on parallel infilling strategy and kriging models[J]. Journal of Mechanical Engineering, 2019, 55(8): 73-79. 谢延敏, 张飞, 潘贝贝, 等. 基于并行加点kriging模型的拉延筋优化[J]. 机械工程学报, 2019, 55(8): 73-79.
[17] ECHARD B, GAYTON N, LEMAIRE M. AK-MCS: an active learning reliability method combining kriging and Monte Carlo simulation[J]. Structural Safety, 2011, 33(2): 145-154.
[18] HONG Linxiong, LI Huacong, PENG Kai, et al. Structural reliability algorithms of Kriging model based on improved learning strategy[J]. Journal of Northwestern Polytechnical University, 2020, 38(2): 412-419. 洪林雄, 李华聪, 彭凯, 等. 基于改进学习策略的Kriging模型结构可靠度算法[J]. 西北工业大学学报, 2020, 38(2): 412-419.
[19] ZHI P P, LI Y H, CHEN B Z, et al. Bounds-based structure reliability analysis of bogie frame under variable load cases[J]. Engineering Failure Analysis, 2020, 114: 104541.

基于RMQGS-APS-Kriging的主动学习结构可靠性分析方法

RMQGS-APS-Kriging-based Active Learning Structural Reliability Analysis Method

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 13

编辑推荐

Metrics

本文评价

[1]	程洪鑫, 李璐祎, 周长聪. 基于Kriging模型和代理抽样的分布参数不确定性灵敏度分析方法[J]. 机械工程学报, 2024, 60(8): 370-383.
[2]	韦新鹏, 姚中洋, 宝文礼, 张哲, 姜潮. 一种基于主动学习克里金模型的证据理论可靠性分析方法[J]. 机械工程学报, 2024, 60(2): 356-368.
[3]	杨旭锋, 程鑫, 刘泽清. 一种融合交叉熵自适应抽样与ALK模型的可靠性分析方法[J]. 机械工程学报, 2024, 60(16): 73-82.
[4]	范小宁, 王凯, 余畅. 基于约束边界抽样的起重机金属结构可靠性优化[J]. 机械工程学报, 2023, 59(8): 288-298.
[5]	周金宇, 王志凌, 程锦翔, 韩文钦. 面向离散化临界状态空间的重要抽样法[J]. 机械工程学报, 2023, 59(14): 352-360.
[6]	何庆, 利璐, 李晨钟, 王平, 谢斯. 基于Kriging模型的在役高速列车悬挂参数近似贝叶斯估计[J]. 机械工程学报, 2023, 59(12): 139-148.
[7]	余萌晨, 龙湘云. 基于两阶段主动学习Kriging模型的广义概率区间混合可靠性分析[J]. 机械工程学报, 2022, 58(6): 274-288.
[8]	张文鑫, 吕震宙. 一种新的自适应Kriging法停止准则及其在涡轮盘疲劳寿命可靠性中的应用[J]. 机械工程学报, 2022, 58(6): 263-273.
[9]	于梦阁, 潘振宽, 蒋荣超, 张继业. 基于近似模型的高速列车头型多目标优化设计[J]. 机械工程学报, 2019, 55(24): 178-186.
[10]	姜潮, 张旺, 韩旭. 基于Copula函数的证据理论相关性分析模型及结构可靠性计算方法[J]. 机械工程学报, 2017, 53(16): 199-209.
[11]	张干清;龚宪生. 变量相关情况下基于杂交GA-PSO算法的结构协同优化[J]. , 2012, 48(15): 113-125.
[12]	李昆锋;杨自春;孙文彩. 结构凸集—概率混合可靠性分析的新方法[J]. , 2012, 48(14): 192-198.
[13]	宋述芳;吕震宙. 基于马尔可夫蒙特卡罗子集模拟的可靠性灵敏度分析方法[J]. , 2009, 45(4): 33-38.