A Monte Carlo simulation based inverse propagation method for stochastic model updating

作者：Bao Nuo^*; Wang Chunjie

来源：Mechanical Systems and Signal Processing, 2015, 60-61: 928-944.

DOI：10.1016/j.ymssp.2015.01.011

• Model updating
• Uncertainty
• Analysis of variance
• Response surface model
• Monte Carlo simulation

This paper presents an efficient stochastic model updating method based on statistical theory. Significant parameters have been selected implementing the F-test evaluation and design of experiments, and then the incomplete fourth-order polynomial response surface model (RSM) has been developed. Exploiting of the RSM combined with Monte Carlo simulation (MCS), reduces the calculation amount and the rapid random sampling becomes possible. The inverse uncertainty propagation is given by the equally weighted sum of mean and covariance matrix objective functions. The mean and covariance of parameters are estimated synchronously by minimizing the weighted objective function through hybrid of particle-swarm and Nelder-Mead simplex optimization method, thus the better correlation between simulation and test is achieved. Numerical examples of a three degree-of-freedom mass-spring system under different conditions and GARTEUR assembly structure validated the feasibility and effectiveness of the proposed method.

• 出版日期2015-8
• 单位北京航空航天大学