Automated Multicriterion Building Damage Assessment from Seismic Data

Abstract

An automated damage assessment method for buildings equipped with PC-based strong-motion recording systems is proposed. The major components of the method include a postearthquake data analysis package and a multicriterion damage assessment package. The data analysis package performs the examination of recorded data and the execution of a system identification method. The damage assessment package includes a set of user supplied rules on damage indices identifiable from the recorded data and a damage inference method based on fuzzy logic. Four damage states, ranging from no damage to severe damage, and four probabilistic statements, ranging from very likely to unlikely, are combined to make a statement on the damage state. The damage statements obtained from each damage index are then combined and defuzzified to arrive at a single set of damage statement. Three examples are given to illustrate the process by which an assessment statement on the damage state is reached. This proposed method can be used as a supplement to on-site inspection by experienced engineers.