An efficient Bayesian method with intrusive homotopy surrogate model for stochastic model updating

• DOI: 10.1111/mice.13206
• Publication Date: 2024-04-16T16:44:55Z
• Authors: Hui Chen; Bin Huang; Heng Zhang; Kaiyi Xue; Ming Sun; Zhifeng Wu
• Keywords: Surrogate model; Statistic; Metropolis–Hastings algorithm

Abstract This paper proposes a new stochastic model updating method based on the homotopy surrogate (HSM) and Bayesian sampling. As novel intrusive model, HSM is established by finite element (FE) method. Then combining advanced delayed‐rejection adaptive Metropolis–Hastings sampling technology with HSM, structural FE can be updated uncertain measurement modal data. The numerical results show that effectiveness of proposed better than methods non‐intrusive models, such as response surface Kriging model. Compared to second‐order perturbation are more accurate, especially when fluctuation measured data large stiffness structure significantly changes. cable‐stayed bridge statistic properties have very good agreement