Inverse analysis of deformation moduli for high arch dams using the displacement reconstruction technique and multi‐objective optimization
Zefa Li,Zhenyu Wu,Jiankang Chen,Yanling Li,Huibao Huang,Yu Lu,Xiang Lu,Junru Li
DOI: https://doi.org/10.1111/mice.13141
IF: 10.066
2023-12-20
Computer-Aided Civil and Infrastructure Engineering
Abstract:The inverse analysis of the deformation moduli of high arch dams based on displacement monitoring data is essential for structural safety assessment. In traditional inverse analysis methods, the deformation moduli are identified based on the single‐objective optimization and the hydrostatic component derived from the statistical model. This type of method has two main shortcomings: First, it treats the essential multi‐objective optimization problem as a single‐objective problem; second, the extracted hydrostatic component may be biased due to the multicollinearity of variables in the statistical model. This paper presents a methodology for the inverse analysis of the deformation moduli of high arch dams under a multi‐objective optimization strategy. The methodology employs empirical mode decomposition to extract the aging component from displacement monitoring data. Then, thermomechanical analysis is used to reconstruct the remaining hydrostatic and temperature components, thereby avoiding the biases encountered in solving the statistical model. The adaptive polynomial chaos expansion method is embedded in the NSGA‐III algorithm to establish and solve multi‐objective functions in the inverse analysis. Additionally, a composite decision index considering errors and test information is proposed to determine acceptable deformation moduli from the Pareto solution set. A high arch dam is selected to illustrate this methodology with static and dynamic monitoring data. The results show that the identified deformation moduli have errors of 3.8% and 7.2% in displacement and acceleration, respectively. The proposed methodology can yield deformation modulus values that are more consistent with the physical implications than those of the single‐objective optimization method.
transportation science & technology,construction & building technology,computer science, interdisciplinary applications,engineering, civil