Seismic hazard models for typical urban masonry structures considering optimized regression algorithms.

Seismic risk estimation for urban building clusters is essential when developing regional seismic resilience models. A masonry structure is a type of building with a large stock, broad applicability, and simplified construction technology. The traditional seismic intensity measures and fuzzy quantification of structural vulnerability levels result in low accuracy in estimating and predicting seismic damage to masonry structure clusters. This paper proposes a method for evaluating bivariate seismic intensity measure. Nonlinear dynamic and spectral analyses were conducted on real acceleration records monitored by ten stations during the Wenchuan earthquake in Sichuan, China, on May 12, 2008. The outdated seismic vulnerability level of masonry structures is updated, and an optimized structural seismic risk model based on the seismic damage database of 2407 masonry structures in Dujiangyan city under the influence of the Wenchuan earthquake is established. An innovative seismic risk membership index algorithm was developed, and a vulnerability curve was generated for masonry structures considering multiple risk membership indices. An improved seismic risk index calculation function is proposed using the proposed updated vulnerability level, and the seismic risk index curve and stripe model of typical urban masonry building clusters are established. According to the field investigation of the actual structures in Dujiangyan city affected by the Wenchuan earthquake, the typical failure features and disaster mechanism of masonry structures are reported and analysed, and measures and methods to improve and enhance the seismic capacity of such buildings are proposed. The developed seismic risk model for urban masonry structures can contribute positively to the development of relatively accurate seismic resilience models for urban buildings and can improve the accuracy of seismic risk estimation for masonry structures. [ABSTRACT FROM AUTHOR]