Uncertainty Analysis for In-Plane Capacity of Unreinforced Masonry Walls Assisted by Control Variate Multi-Fidelity Approach: Statistics Estimation
摘要
Unreinforced masonry walls are particularly vulnerable to horizontal in-plane loading due to their low tensile strength. The composite nature of masonry leads to significant challenges in predicting in-plane resistance of unreinforced masonry wall. Moreover, the uncertainties in the masonry material may lead to a large scatter. However, quantifying the uncertainty (e.g., estimating the mean value of in-plane capacity) is particularly challenging due to the computational cost. In this study, high-fidelity model and low-fidelity model are integrated into the multi-fidelity framework, leveraging the information from both models. The HF model is developed based on the simplified micromodel approach for masonry structures, whereas the design code-based model is adopted as the LF model. Variance reduction and accuracy improvement are obtained in the multi-fidelity estimation compared to classical Monte Carlo estimation, providing a huge potential for the uncertainty behavior quantification for masonry walls.