Identification of shear buildings using an instrumental variable method and linear integral filters

Abstract

This paper develops a method for estimating the parameters of a seismically excited building. Acceleration measurements of the ground and of the building floors, containing offsets and noise, are used for identification purposes. The proposed scheme estimates the complete model of the building if all the floors are equipped with accelerometers. Moreover, it also estimates a reduced model of the structure if only some floors are instrumented. The methodology is based on the combined use of the Instrumental Variable method and Linear Integral Filters. The Instrumental Variable method employs as instrument an auxiliary model of the structure, and it is able to directly identify the continuous-time structure model using discrete-time data without resorting on model transformations from continuous-time to discrete-time, and vice-versa. Using Linear Integral Filters allows obtaining a linear in the parameters expression that depends only on acceleration measurements suitable for parameter identification purposes. These filters eliminate measurement offsets and attenuate high-frequency measurement noise. The above features together with the use of the Instrumental variable method reduce the likelihood of biased parameter estimates. Experiments on a testbed employing a reduced-scale five-story structure allow comparing the results obtained using the Instrumental Variable method and those produced by the standard Least Squares method.