工程力学 ›› 2018, Vol. 35 ›› Issue (9): 44-53.doi: 10.6052/j.issn.1000-4750.2017.08.0644

• 基本方法 • 上一篇    下一篇

稀疏偏最小二乘回归-多项式混沌展开代理模型方法

赵威1,2,3, 卜令泽1, 王伟1,2,3   

  1. 1. 哈尔滨工业大学土木工程学院, 哈尔滨 150090;
    2. 结构工程灾变与控制教育部重点实验室, 哈尔滨 150090;
    3. 土木工程智能防灾减灾工业与信息化部重点实验室, 哈尔滨 150090
  • 收稿日期:2017-08-23 修回日期:2017-12-22 出版日期:2018-09-29 发布日期:2018-09-15
  • 通讯作者: 卜令泽(1993-),男,黑龙江人,博士生,主要从事结构可靠度与全局灵敏度方法的研究(E-mail:17b933010@stu.hit.edu.cn). E-mail:17b933010@stu.hit.edu.cn
  • 作者简介:赵威(1982-),男,黑龙江人,讲师,博士,主要从事结构可靠度方法研究(E-mail:spritewei@163.com);王伟(1957-),男,黑龙江人,教授,博士,博导,主要从事结构可靠度方法研究(E-mail:wwang@hit.edu.cn).
  • 基金资助:
    国家自然科学基金面上项目(11572106)

SPARSE PARTIAL LEAST SQUARES REGRESSION-POLYNOMIAL CHAOS EXPANSION METAMODELING METHOD

ZHAO Wei1,2,3, BU Ling-ze1, WANG Wei1,2,3   

  1. 1. School of Civil Engineering, Harbin Institute of Technology, Harbin, 150090, China;
    2. Key Lab of Structures Dynamic Behaviour and Control of the Ministry of Education, Harbin Institute of Technology, Harbin, 150090, China;
    3. Key Lab of Smart Prevention and Mitigation of Civil Engineering Disasters of the Ministry of Industry and Information Technology, Harbin Institute of Technology, Harbin, 150090, China
  • Received:2017-08-23 Revised:2017-12-22 Online:2018-09-29 Published:2018-09-15

摘要: 为解决传统多项式混沌展开方法在高维全局灵敏度和结构可靠度分析当中存在的维数灾难与多重共线性问题,该文提出一种稀疏偏最小二乘回归-多项式混沌展开代理模型方法。该方法首先采用偏最小二乘回归技术得到多项式混沌展开系数的初步估计,然后根据回归误差阈值允许下的最大稀疏度原则,采用带有惩罚的矩阵分解技术自适应地保留与结构响应相关性强的多项式,并采用偏最小二乘回归得到多项式混沌展开系数的更新估计。通过对展开系数进行简单后处理即可得到Sobol灵敏度指数。在此基础上保留重要输入变量并按新方法重新进行回归可实现对代理模型的精简,从而在不增加计算代价的情况下实现高精度结构可靠度分析。算例结果表明在保证精度的情况下,采用新方法进行全局灵敏度和结构可靠度分析比传统方法在计算效率方面有显著优势。

关键词: 高维模型, 稀疏偏最小二乘回归, 多项式混沌展开, 全局灵敏度, 结构可靠度

Abstract: To circumvent the curse of dimensionality and multicollinearity problems of traditional polynomial chaos expansion approach when analyzing global sensitivity and structural reliability of high-dimensional models, this paper proposes a sparse partial least squares regression-polynomial chaos expansion metamodeling method. Firstly, an initial estimation of polynomial chaos expansion coefficients is obtained with the partial least squares regression. Secondly, according to the principle of maximum sparsity under the allowance of regression error threshold, polynomials which have strong correlation with the structural response are adaptively retained with the penalized matrix decomposition scheme. Next, an updated estimation of the polynomial chaos expansion coefficients is obtained with the partial least squares regression. Sobol sensitivity indices are obtained with a simple post-processing of the expansion coefficients. Finally, the metamodel is greatly simplified by regressing with important inputs, leading to accurate estimations of the failure probability without additional computational cost. The results show that with acceptable accuracies, the new method overperforms the traditional counterpart in terms of computational efficiency when solving high-dimensional global sensitivity and structural reliability analysis problems.

Key words: high-dimensional models, sparse partial least squares regression, polynomial chaos expansion, global sensitivity, structural reliability

中图分类号: 

  • TU311
[1] Hasofer A M, Lind N C. Exact and invariant second-moment code format[J]. Journal of the Engineering Mechanics Division, 1974, 100(1):111-121.
[2] Breitung K. 40 years FORM:Some new aspects?[J]. Probabilistic Engineering Mechanics, 2015, 42:71-77.
[3] Au S K, Beck J L. A new adaptive importance sampling scheme for reliability calculations[J]. Structural Safety, 1999, 21(2):135-158.
[4] Wang Z, Song J. Cross-entropy-based adaptive importance sampling using von Mises-Fisher mixture for high dimensional reliability analysis[J]. Structural Safety, 2016, 59:42-52.
[5] Au S, Beck J L. Estimation of small failure probabilities in high dimensions by subset simulation[J]. Probabilistic Engineering Mechanics, 2001, 16(4):263-277.
[6] Au S. On MCMC algorithm for subset simulation[J]. Probabilistic Engineering Mechanics, 2016, 43:117-120.
[7] Koutsourelakis P S. Reliability of structures in high dimensions. Part Ⅱ. Theoretical validation[J]. Probabilistic Engineering Mechanics, 2004, 19(4):419-423.
[8] Koutsourelakis P S, Pradlwarter H J, Schueller G I. Reliability of structures in high dimensions, part I:algorithms and applications[J]. Probabilistic Engineering Mechanics, 2004, 19(4):409-417.
[9] Chowdhury R, Rao B N, Prasad A M. High dimensional model representation for structural reliability analysis[J]. Communications in Numerical Methods in Engineering, 2009, 25(4):301-337.
[10] Chakraborty S, Chowdhury R. Assessment of polynomial correlated function expansion for high-fidelity structural reliability analysis[J]. Structural Safety, 2016, 59:9-19.
[11] Bichon B J, Eldred M S, Swiler L P, et al. Efficient global reliability analysis for nonlinear implicit performance functions[J]. AIAA journal, 2008, 46(10):2459-2468.
[12] Hu Z, Mahadevan S. Global sensitivity analysisenhanced surrogate (GSAS) modeling for reliability analysis[J]. Structural and Multidisciplinary Optimization, 2015, 3(53):1-21.
[13] Wiener N. The homogeneous chaos[J]. American Journal of Mathematics, 1938, 60(4):897-936.
[14] Ghanem R, Spanos P D. Polynomial chaos in stochastic finite elements[J]. Journal of Applied Mechanics, 1990, 57(1):197-202.
[15] Choi S K, Grandhi R V, Canfield R A, et al. Polynomial Chaos Expansion with Latin Hypercube Sampling for Estimating Response Variability[J]. Aiaa Journal, 2004, 42(6):1191-1198.
[16] Choi S, Canfield R A, Grandhi R V. Estimation of structural reliability for Gaussian random fields[J]. Structures and Infrastructure Engineering, 2006, 2(3/4):161-173.
[17] 李典庆, 蒋水华, 周创兵. 基于非侵入式随机有限元法的地下洞室可靠度分析[J]. 岩土工程学报, 2012, 34(1):123-129. Li Dianqing, Jiang Shuihua, Zhou Chuangbing. Reliability analysis of underground rock caverns using non-intrusive stochastic finite element method[J]. Chinese Journal of Geotechnical Engineering, 2012, 34(1):123-129. (in Chinese)
[18] 蒋水华, 李典庆, 周创兵. 基于拉丁超立方抽样的边坡可靠度分析非侵入式随机有限元法[J]. 岩土工程学报, 2013, 35(增刊2):70-76. Jiang Shuihua, Li Dianqing, Zhou Chuangbing. Non-intrusive stochastic finite element method for slope reliability analysis based on Latin hypercube sampling[J]. Chinese Journal of Geotechnical Engineering, 2013, 35(Suppl 2):70-76. (in Chinese)
[19] 蒋水华, 冯晓波, 李典庆, 等. 边坡可靠度分析的非侵入式随机有限元法[J]. 岩土力学, 2013, 34(8):2347-2354. Jiang Shuihua, Feng Xiaobo, Li Dianqing et al. Reliability analysis of slope using non-intrusive stochastic finite element method[J]. Rock and Soil Mechanics, 2013, 34(8):2347-2354. (in Chinese)
[20] Blatman G, Sudret B. An adaptive algorithm to build up sparse polynomial chaos expansions for stochastic finite element analysis[J]. Probabilistic Engineering Mechanics, 2010, 25(2):183-197.
[21] Blatman G, Sudret B. Adaptive sparse polynomial chaos expansion based on least angle regression[J]. Journal of Computational Physics, 2011, 230(6):2345-2367.
[22] 吕震宙, 李璐祎, 宋述芳, 等. 不确定性结构系统的重要性分析理论与求解方法[M]. 北京:科学出版社, 2015. Lü Zhenzhou, Li Luyi, Song Shufang, et al. Analysis theory and computational methods for uncertain structural systems[M]. Beijing:Science Press, 2015. (in Chinese)
[23] Sobol', Il'ya Meerovich. On sensitivity estimation for nonlinear mathematical models[J]. Matematicheskoe Modelirovanie, 1990, 2(1):112-118.
[24] Homma T, Saltelli A. Importance measures in global sensitivity analysis of nonlinear models[J]. Reliability Engineering & System Safety, 1996, 52(1):1-17.
[25] Stanfill B, Mielenz H, Clifford D, et al. Simple approach to emulating complex computer models for global sensitivity analysis[J]. Environmental Modelling & Software, 2015, 74(C):140-155.
[26] Oakley J E, O'Hagan A. Probabilistic sensitivity analysis of complex models:a Bayesian approach[J]. Journal of the Royal Statistical Society:Series B (Statistical Methodology), 2004, 66(3):751-769.
[27] Sudret B. Global sensitivity analysis using polynomial chaos expansions[J]. Reliability Engineering & System Safety, 2008, 93(7):964-979.
[28] Blatman G, Sudret B. Efficient computation of global sensitivity indices using sparse polynomial chaos expansions[J]. Reliability Engineering & System Safety, 2010, 95(11):1216-1229.
[29] 胡军, 张树道. 基于多项式混沌的全局敏感度分析[J]. 计算物理, 2016(1):1-14. Hu Jun, Zhang Shudao. Global sensitivity analysis based on polynomial chaos[J]. Chinese Journal of Computational Physics, 2016(1):1-14. (in Chinese)
[30] 邬晓敬, 张伟伟, 宋述芳, 等. 翼型跨声速气动特性的不确定性及全局灵敏度分析[J]. 力学学报, 2015, 47(4):587-595. Wu Xiaojing, Zhang Weiwei, Song Shufang et al. Uncertainty quantification and global sensitivity analysis of transonic aerodynamics about airfoil. Chinese Journal of Theoretical and Applied Mechanics, 2015, 47(4):587-595. (in Chinese)
[31] 宋彦. 基于随机响应面法的结构整体可靠度与全局灵敏度分析[D]. 哈尔滨:哈尔滨工业大学, 2016. Song Yan. Global reliability and sensitivity analysis of structures based on stochastic response surface methods[D]. Harbin:Harbin Institute of Technology, 2016. (in Chinese)
[32] 陈光宋, 钱林方, 吉磊. 身管固有频率高效全局灵敏度分析[J]. 振动与冲击, 2015, 34(21):31-36. Chen Guangsong, Qian Linfang, Ji Lei. An effective global sensitivity analysis method for natural frequencies of a barrel[J]. Journal of Vibration and Shock, 2015, 34(21):31-36. (in Chinese)
[33] 王惠文, 吴载斌, 孟洁. 偏最小二乘回归的线性与非线性方法[M]. 北京:国防工业出版社, 2006. Wang Huiwen, Wu Zaibin, Meng Jie. Linear and nonlinear methods of partial least squares regression[M]. Beijing:National Defend Industry Press, 2006. (in Chinese)
[34] Deng B, Yun Y, Liang Y, et al. A new strategy to prevent over-fitting in partial least squares models based on model population analysis[J]. Analytica Chimica Acta, 2015, 880:32-41.
[35] Witten D M, Hastie R T T. A penalized matrix decomposition, with applications to sparse principal components and canonical correlation analysis[J]. Biostatistics, 2009, 10(3):515-534.
[36] Pan Q, Dias D. Sliced inverse regression-based sparse polynomial chaos expansions for reliability analysis in high dimensions[J]. Reliability Engineering & System Safety, 2017, 167:484-493.
[37] 刘娇, 刘敬敏, 余波, 等. 工程结构体系可靠度分析的最新研究进展[J]. 工程力学, 2017, 34(增刊1):31-37. Liu Jiao, Liu Jingmin, Yu Bo, et al. Recent research progress on structural system reliability analysis[J]. Engineering Mechanics, 2017, 34(Suppl 1):31-37. (in Chinese)
[38] 孙威, 黄炎生, 杨惠贤. 改进响应面法计算钢筋混凝土框架体系可靠度[J]. 工程力学, 2016, 33(6):194-201, 208. Sun Wei, Huang Yansheng, Yang Huixian. System reliability calculation of RC frame structures by improved response surface method[J]. Engineering Mechanics, 2016, 33(6):194-201, 208. (in Chinese)
[1] 陈超, 吕震宙. 模糊分布参数的全局灵敏度分析新方法[J]. 工程力学, 2016, 33(2): 25-33.
[2] 孟增, 李刚. 基于修正混沌控制的一次二阶矩可靠度算法[J]. 工程力学, 2015, 32(12): 21-26.
[3] 杨绿峰,袁彦华,余波. 基于等概率近似变换的 向量型层递响应面法分析结构可靠度[J]. 工程力学, 2014, 31(7): 185-191.
[4] 范文亮,张春涛,李正良,韩枫. 考虑交叉项的自适应响应面法[J]. 工程力学, 2013, 30(4): 68-72.
[5] 赵 威,王 伟. 偏最小二乘回归在响应面法可靠度分析中的应用[J]. 工程力学, 2013, 30(2): 272-277.
[6] 白冰,张清华,李乔. 结构二次二阶矩可靠度指标的回归分析预测算法[J]. 工程力学, 2013, 30(10): 219-226,235.
[7] 程蕾,吕震宙,李璐祎. 基于失效概率偏导数的局部与全局混合灵敏度分析[J]. 工程力学, 2013, 30(10): 272-276,281.
[8] 吴小强;姚继涛;刘雅君. 住宅楼面活荷载的统计分析及楼板可靠度的分析[J]. 工程力学, 2012, 29(3): 90-94.
[9] 陈向前,董 聪,闫 阳. 自适应重要抽样方法的改进算法[J]. 工程力学, 2012, 29(11): 123-128.
[10] 吕大刚;贾明明;李 刚. 结构可靠度分析的均匀设计响应面法[J]. 工程力学, 2011, 28(7): 109-116.
[11] 胡 冉;李典庆;周创兵;陈益峰. 基于随机响应面法的结构可靠度分析[J]. 工程力学, 2010, 27(9): 192-200.
[12] 陈朝晖;郑小宇. 低层民居台风可靠度分析初探[J]. 工程力学, 2009, 26(增刊Ⅱ): 230-234,.
[13] 吕大刚. 基于线性化Nataf变换的一次可靠度方法[J]. 工程力学, 2007, 24(5): 0-086,.
[14] 刘成立;吕震宙. 非线性隐式极限状态方程失效概率计算的组合响应面法[J]. 工程力学, 2006, 23(2): 29-33.
[15] 程耿东;蔡文学. 结构可靠度计算的近似重要性抽样方法及其应用[J]. 工程力学, 1997, 14(2): 1-8.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
[1] 李忠献,黄 信. 行波效应对深水连续刚构桥地震响应的影响[J]. 工程力学, 2013, 30(3): 120 -125 .
[2] 王进廷;张楚汉;金峰. 有阻尼动力方程显式积分方法的精度研究[J]. 工程力学, 2006, 23(3): 1 -5 .
[3] 熊铁华;常晓林. 响应面法在结构体系可靠度分析中的应用[J]. 工程力学, 2006, 23(4): 58 -61 .
[4] 贾超;张楚汉;金峰;程卫帅. 可靠度对随机变量及失效模式相关系数的敏感度分析及其工程应用[J]. 工程力学, 2006, 23(4): 12 -16,1 .
[5] 史宝军;袁明武;宋世军. 流体力学问题基于核重构思想的最小二乘配点法[J]. 工程力学, 2006, 23(4): 17 -21,3 .
[6] 杨璞;刘应华;袁鸿雁;岑章志. 计算结构极限载荷的修正弹性补偿法[J]. 工程力学, 2006, 23(3): 21 -26 .
[7] 罗战友;夏建中;龚晓南. 不同拉压模量及软化特性材料的球形孔扩张问题的统一解[J]. 工程力学, 2006, 23(4): 22 -27 .
[8] 吴琛;周瑞忠. 小波基无单元法及其与有限元法的比较[J]. 工程力学, 2006, 23(4): 28 -32 .
[9] 曹晖;Michael I. Friswell. 基于模态柔度曲率的损伤检测方法[J]. 工程力学, 2006, 23(4): 33 -38 .
[10] 李华波;许金泉;杨震. 等厚双层涂层材料受切向集中力作用的三维理论解[J]. 工程力学, 2006, 23(4): 45 -51 .
X

近日,本刊多次接到来电,称有不法网站冒充《工程力学》杂志官网,并向投稿人收取高额费用。在此,我们郑重申明:

1.《工程力学》官方网站是本刊唯一的投稿渠道(原网站已停用),《工程力学》所有刊载论文必须经本刊官方网站的在线投稿审稿系统完成评审。我们不接受邮件投稿,也不通过任何中介或编辑收费组稿。

2.《工程力学》在稿件符合投稿条件并接收后会发出接收通知,请作者在接到版面费或审稿费通知时,仔细检查收款人是否为“《工程力学》杂志社”,千万不要汇款给任何的个人账号。请广大读者、作者相互转告,广为宣传!如有疑问,请来电咨询:010-62788648。

感谢大家多年来对《工程力学》的支持与厚爱,欢迎继续关注我们!

《工程力学》杂志社

2018年11月15日