A Monte Carlo simulation based inverse propagation method for stochastic model updating. (August 2015)

• Record Type: Journal Article
• Authors: Bao, Nuo; Wang, Chunjie
• Is Part Of: Mechanical systems and signal processing. Volume 60/61(2015)
• Page Start: 928
• Page End: 944
• Publication Date: 2015-08
• Subjects: Model updating -- Uncertainty -- Analysis of variance -- Response surface model -- Monte Carlo simulation; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2015.01.011
• Languages: English
• ISSNs: 0888-3270

Abstract

Abstract: 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. Highlights: The incomplete fourth-order polynomial RSM is constructed by using F -test and DOE. RSM and Monte Carlo method are integrated to achieve rapid random sampling. The weighted objective function is minimized by applying hybrid optimization. The statistical characteristics of parameter is synchronously estimated.