观测不确定性下变分贝叶斯高效模型修正
收稿日期: 2023-12-12
修回日期: 2023-12-26
录用日期: 2024-01-17
网络出版日期: 2024-02-02
基金资助
国家自然科学基金(U23B20105)
Efficient variational Bayesian model updating under observation uncertainty
Received date: 2023-12-12
Revised date: 2023-12-26
Accepted date: 2024-01-17
Online published: 2024-02-02
Supported by
National Natural Science Foundation of China(U23B20105)
考虑试验观测数据不确定性,针对以随机响应为目标的复杂仿真模型不确定性修正问题,提出一种以自回归模型提取信号特征,马氏距离构造近似似然函数,并由变分贝叶斯-蒙特卡洛进行参数后验识别的模型修正方法。对经过平稳性检验的随机信号进行自回归分析,得到模型特征向量实现信息降维;考虑试验观测数据混合不确定性并用概率盒法表征,以数值模拟数据与试验观测数据的特征向量之间的马氏距离构造近似对数似然;基于变分贝叶斯-蒙特卡洛方法求解边际似然,经过很少的迭代次数即可收敛,最终识别出参数的后验分布。在螺栓连接结构仿真案例和某民用飞机机翼模型工程案例中,修正后的模型具有很高精度,且在一定的观测不确定性水平下依然具有良好的修正效果,验证了所提方法对工程结构不确定性模型修正问题的有效性。
陶言和 , 郭勤涛 , 周瑾 , 马嘉倩 , 李效法 . 观测不确定性下变分贝叶斯高效模型修正[J]. 航空学报, 2024 , 45(19) : 229969 -229969 . DOI: 10.7527/S1000-6893.2024.29969
A model updating approach is proposed to address uncertainty in complex numerical models with stochastic responses. The approach involves using auto-regressive models for signal feature extraction, utilizing Mahalanobis distance as uncertainty quantification metric, and performing parameter posterior identification through variational Bayesian Monte Carlo. Initially, auto-regressive analysis is conducted on stationary stochastic signals to obtain model feature vectors for dimension reduction. The hybrid uncertainty of experimental observation data is then characterized utilizing the probability-box method. An approximate logarithmic likelihood is constructed based on the Mahalanobis distance between feature vectors of simulated data and experimental observation data. Finally, the variational Bayesian Monte Carlo method is used to solve the marginal likelihood, resulting in the identification of the posterior of parameters after very few iterations. Effectiveness of the proposed method for uncertainty updating in engineering structural models is validated through a numerical case of bolted connection structures and a civil aircraft wing model. The updated model exhibits high accuracy and retains good updating performance under certain levels of observation uncertainty.
1 | AHMAD BURHANI A B, WOOK C D, HYEONGILL L. Finite element model updating of composite with adhesive jointed structure under built-up internal stress[J]. Journal of Vibration and Control, 2022, 28(11-12): 1390-1401. |
2 | BI S F, BEER M, COGAN S, et al. Stochastic model updating with uncertainty quantification: An overview and tutorial[J]. Mechanical Systems and Signal Processing, 2023, 204: 110784. |
3 | CRESPO L G, KENNY S P, GIESY D P. The NASA langley multidisciplinary uncertainty quantification challenge[C]∥16th AIAA Non-deterministic Approaches Conference. Reston: AIAA, 2014, doi: 10.2514/6.2014-1347 . |
4 | BECK J L, KATAFYGIOTIS L S. Updating models and their uncertainties. I: Bayesian statistical framework[J]. Journal of Engineering Mechanics, 1998, 124(4): 455-461. |
5 | FENG K, LU Z, CHEN Z, et al. An innovative Bayesian updating method for laminated composite structures under evidence uncertainty[J]. Composite structures, 2023, 304(1):116429. |
6 | LIAO B P, ZHAO R, YU K P, et al. A novel interval model updating framework based on correlation propagation and matrix-similarity method[J]. Mechanical Systems and Signal Processing, 2022, 162(2): 108039. |
7 | PANDA A K, MODAK S V. A two-stage approach to stochastic finite element model updating using FRF data[J]. Journal of Sound and Vibration, 2023, 553: 117670. |
8 | 陈喆, 何欢, 陈国平, 等. 考虑不确定性因素的有限元模型修正方法研究[J]. 振动工程学报, 2017, 30(6): 921-928. |
CHEN Z, HE H, CHEN G P, et al. The research of finite element model updating method considering the uncertainty[J]. Journal of Vibration Engineering, 2017, 30(6): 921-928 (in Chinese). | |
9 | 王震宇, 王计真, 杨婧艺, 等. 基于新型粒子群算法的结构动力学热振模型修正[J]. 航空学报, 2023, 44(7): 226559. |
WANG Z Y, WANG J Z, YANG J Y, et al. Dynamics model updating of structures at high temperature based on novel particle swarm optimization algorithm[J]. Acta Aeronautica et Astronautica Sinica, 2023, 44(7): 226559 (in Chinese). | |
10 | 杨庆. 基于多级贝叶斯模型修正的结构损伤识别方法研究[D]. 大连: 大连交通大学, 2018: 17-30. |
YANG Q. Study on structural damage identification method based on hierarchical Bayesian model updating[D].Dalian: Dalian Jiaotong University, 2018: 17-30. (in Chinese) | |
11 | SEDEHI O, KATAFYGIOTIS L, PAPADIMITRIOU C. A time-domain hierarchical Bayesian approach for model updating[C]∥The 16th European Conference on Earthquake Engineering (16ECEE). Thessaloniki: European Association for Earthquake Engineering, 2018: 1-11. |
12 | SEDEHI O, PAPADIMITRIOU C, KATAFYGIOTIS L S. Probabilistic hierarchical Bayesian framework for time-domain model updating and robust predictions[J]. Mechanical Systems and Signal Processing, 2018, 123: 648-673. |
13 | KITAHARA M, BI S F, BROGGI M, et al. Bayesian model updating in time domain with metamodel-based reliability method[J]. ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems Part A Civil Engineering, 2021, 7(3), doi: 10.1061/AJRUA6.0001149 . |
14 | 程马遥, 金银富, 尹振宇, 等. 改进DE-TMCMC法及其在高级模型参数识别上的应用[J]. 岩土工程学报, 2019, 41(12): 2281-2289. |
CHENG M Y, JIN Y F, YIN Z Y, et al. Enhanced DE-TMCMC and its application in identifying parameters of advanced soil model[J]. Chinese Journal of Geotechnical Engineering, 2019, 41(12): 2281-2289 (in Chinese). | |
15 | TONG W T, GE W, HAN X, et al. A low-complexity algorithm based on variational Bayesian inference for MIMO channel estimation[J]. Applied Acoustics, 2023, 211(2): 109512. |
16 | ZHANG B, HOU Y, YANG Y, et al. Variational Bayesian cardinalized probability hypothesis density filter for robust underwater multi-target direction-of-arrival tracking with uncertain measurement noise[J]. Frontiers in Physics, 2023, 11, doi: 10.3389/fphy.2023.1142400. |
17 | 蔡巍, 陈明剑, 邓垦, 等. 基于变分贝叶斯自适应鲁棒滤波的动对动相对定位算法[J]. 中国惯性技术学报, 2023, 31(8): 760-767, 776. |
CAI W, CHEN M J, DENG K, et al. Dynamic to dynamic relative positioning algorithm based on variational Bayesian adaptive robust filtering[J]. Journal of Chinese Inertial Technology, 2023, 31(8): 760-767, 776 (in Chinese). | |
18 | 李明, 柴洪洲, 靳凯迪, 等. 基于可变因子的变分贝叶斯SINS海上动态对准[J]. 海洋测绘, 2023, 43(4): 47-51. |
LI M, CHAI H Z, JIN K D, et al. Variational Bayesian SINS marine dynamic alignment based on variable factor[J]. Hydrographic Surveying and Charting, 2023, 43(4): 47-51 (in Chinese). | |
19 | 田健. 基于变分贝叶斯滤波的MEMS-SINS/GPS组合导航算法研究[D]. 重庆: 重庆邮电大学, 2021: 35-52. |
TIAN J. Research on MEMS-SINS/GPS integrated navigation based on varitional Bayesian filtering[D].Chongqing: Chongqing University of Posts and Telecommunications, 2021: 35-52. (in Chinese) | |
20 | 王彦钧. 基于变分贝叶斯的协同目标跟踪方法研究[D]. 镇江: 江苏大学, 2021: 41-50. |
WANG Y J. Research on cooperative target tracking method based on variational bayes[D].Zhenjiang: Jiangsu University, 2021: 41-50. (in Chinese) | |
21 | 巴丽伟, 童常青. 基于变分贝叶斯推断的因子分析法[J]. 杭州电子科技大学学报(自然科学版), 2022, 42(3): 95-102. |
BA L W, TONG C Q. Factor analysis based on variational Bayes inference[J]. Journal of Hangzhou Dianzi University (Natural Sciences), 2022, 42(3): 95-102 (in Chinese). | |
22 | 于汀, 李璐祎, 刘昱杉, 等. 观测不确定性下的高效贝叶斯更新方法及其在机翼结构中的应用[J]. 航空学报, 2023, 44(24): 117-134. |
YU T, LI L Y, LIU Y S, et al. Efficient Bayesian updating method under observation uncertainty and its application in wing structure[J]. Acta Aeronautica et Astronautica Sinica, 2023, 44(24): 117-134 (in Chinese). | |
23 | 熊芬芬, 李泽贤, 刘宇, 等. 基于数值模拟的工程设计中参数不确定性表征方法研究综述[J]. 航空学报, 2023, 44(22): 92-123. |
XIONG F F, LI Z X, LIU Y, et al. A review of characterization methods for parameter uncertainty in engineering design based on numerical simulation[J]. Acta Aeronautica et Astronautica Sinica, 2023, 44(22): 92-123 (in Chinese). | |
24 | RANGANATH R. Black Box Variational Inference: Scalable, Generic Bayesian Computation and its Applications[D]. Princeton: Princeton University, 2017: 20-27. |
25 | ZHAO Y H. Gaussian process mixture model for prediction based on maximum posterior distribution[J]. Journal of Physics: Conference Series, 2021, 2014: 012007. |
26 | 恽鹏. 不确定量测下的变分贝叶斯目标跟踪算法研究[D]. 南京: 南京理工大学, 2022: 14-16. |
YUN P. Research on variational Bayesian target tracking algorithm under uncertain measurement[D].Nanjing: Nanjing University of Science and Technology, 2022: 14-16. (in Chinese) | |
27 | ACERBI L. Variational Bayesian monte carlo[J]. Advances in Neural Information Processing Systems, 2018, 31: 8222-8232. |
28 | ACERBI L. Variational Bayesian Monte Carlo with noisy likelihoods[C]∥Proceedings of the 34th International Conference on Neural Information Processing Systems. New York: ACM, 2020: 8211-8222. |
29 | CHE Y F, WU X, PASTORE G, et al. Application of Kriging and Variational Bayesian Monte Carlo method for improved prediction of doped UO2 fission gas release[J]. Annals of Nuclear Energy, 2021, 153(3): 108046. |
30 | ZHANG Q, LI Y P, HUANG G H, et al. Copula function with Variational Bayesian Monte Carlo for unveiling uncertainty impacts on meteorological and agricultural drought propagation[J]. Journal of Hydrology, 2023, 622: 129669. |
31 | 展铭. 螺栓连接结构动力学多响应模型修正与确认方法研究[D]. 南京: 南京航空航天大学, 2020: 74-75. |
ZHAN M. Research on model updating and validation of bolt jointed structures based on multi dynamic responses[D].Nanjing: Nanjing University of Aeronautics and Astronautics, 2020: 74-75 (in Chinese). | |
32 | 杨乐昌, 韩东旭, 王丕东. 基于Wasserstein距离测度的非精确概率模型修正方法[J]. 机械工程学报, 2022, 58(24): 300-311. |
YANG L C, HAN D X, WANG P D. Imprecise probabilistic model updating using A Wasserstein distance-based uncertainty quantification metric[J]. Journal of Mechanical Engineering, 2022, 58(24): 300-311 (in Chinese). | |
33 | 姜东, 吴邵庆, 史勤丰, 等. 基于各向同性本构关系薄层单元的螺栓连接参数识别[J]. 振动与冲击, 2014, 33(22): 35-40. |
JIANG D, WU S Q, SHI Q F, et al. Parameter identification of bolted-joint based on the model with thin-layer elements with isotropic constitutive relationship[J]. Journal of Vibration and Shock, 2014, 33(22): 35-40 (in Chinese). | |
34 | DESAI C S, ZAMAN M M, LIGHTNER J G, et al. Thin-layer element for interfaces and joints[J]. International Journal for Numerical and Analytical Methods in Geomechanics, 1984, 8(1): 19-43. |
35 | LIAO T, FASANG A. Comparing groups of life-course sequences using the Bayesian information criterion and the likelihood-ratio test[J]. Sociological Methodology, 2020, 51: 44-85. |
36 | 邓振鸿, 张保强, 苏国强, 等. 基于近似似然的频响函数不确定性模型修正[J]. 振动 测试与诊断, 2020, 40(3): 548-554, 628. |
DENG Z H, ZHANG B Q, SU G Q, et al. Uncertainty model updating of frequency response function based on approximate likelihood function[J]. Journal of Vibration, Measurement & Diagnosis, 2020, 40(3): 548-554, 628 (in Chinese). |
/
〈 |
|
〉 |