Bayesian inference for seismic damage assessment in a reinforced concrete structure

This article presents the application of a Bayesian updating method toassess seismic damage in Reinforced Concrete (RC) frame structures. The struc-ture is represented through a fishbone model that can account for the stiffness ofcolumns and beam-column connections. The fishbone model is updated using aBayesian approach using values of natural frequencies and modal shapes esti-mated through ambient vibration tests. Seismic damage is quantified by compar-ing the stiffness of the fishbone model before and after the earthquake. TheBayesian updating method is validated on a half-scale RC frame structure builtat Tongji University and artificially damaged during shaking table tests. Specifi-cally, several levels of damage were inflicted to the RC frame structure by apply-ing seismic excitations of increasing intensity. Natural frequencies and modalshapes of the RC frame structure, needed for Bayesian updating, were retrievedfrom responses recorded during ambient vibration tests carried out on a shakingtable in the undamaged and several damaged conditions. Results showed that theBayesian updating method is able to assess seismic damage on the analyzed struc-ture. In addition, the stiffness of columns was identified with high accuracy.