Development Of Alternative Rapid Screening Method To Determine Regional Risk Distribution Of Masonry Structures

Abstract

The seismic risk of buildings are of great concern of societies as the loss of buildings interests both the public administration and the insurance companies. More importantly, the collapse of the buildings unfortunately gives rise to loss of inhabitants' lives. Therefore, the determination of the seismic risk of buildings has been an important topic among the structural engineers. The seismic risk assessment of building inventories could only be accomplished by changing the strategy from seeking safety to filtering the vulnerable buildings from the large building stock (i.e. low-pass filtering). To this end, decisions should be made on inexpensively acquired building data and the evaluation process should quickly be implemented. Therefore, a new rapid screening method is developed to estimate the seismic risk of unreinforced masonry (URM) buildings in this study. The method is based on a novel use of binary logistic regression of a large database. The database is composed of 543 URM buildings with detailed seismic assessment results. The method considers number of stories (N), type of slab system (SS), vertical irregularities (VI), visual damage (D), type of masonry material (M), typical story height (H) and typical plan area (A) as the basic estimation variables. These estimation variables have assigned some penalty scores depending on the coefficients derived from the binary logistic regression analysis of the database. 443 buildings from the database are used to generate the penalty scores and 100 buildings are reserved for the test of the proposed method. The correct overall estimation rates of the proposed method for the database (i.e. 443 buildings) and the test database (i.e. 100 buildings) are determined as approximately 95% and 86%, respectively. An evaluation form for the rapid assessment of masonry buildings is also presented in this study.