Performance evaluation of instrumented buildings

Performance-based seismic design requires reliable methods to estimate earthquake demands to be met by structural capacities or to be limited through appropriate active or passive control systems. Several methods have been proposed to estimate seismic demand and capacity of existing structures through numerical models. The accuracy of these models in reproducing the real structural behavior depends on the correct calibration of the parameters used in the analyses. A satisfactory calibration can be achieved through identification and model updating methods using responses recorded on the structure during a moderate seismic event. Generally, due to economic reasons concerning the cost of instrumentation and of data analysis, responses are recorded in a limited number of locations; hence a reduced number of dynamic parameters may be estimated and used for the calibration of the numerical models. In order to obtain a more detailed description of the structure, the number of available responses can be increased by interpolating recorded ones. This paper describes a new technique to estimate the seismic demand on multistory buildings in terms of appropriate performance parameters such as story absolute accelerations, velocity and displacements, interstory drifts, story and base shear, from a limited number of recorded signals. A reduced number of signals is assumed to be recorded by sensors placed according to an optimal distribution defined by local minima of a function of the effective participation factors of the dominant modes of the structure. Unknown responses, in locations where sensors are not available, are calculated using a spline shape function to interpolate recorded responses along the height of the considered building. Recorded and interpolated signals are then used to estimate the seismic performance parameters. The method is applied to a real instrumented multistory building using records from recent earthquakes characterized by different intensities and frequency contents. Calculated performance parameters are compared to those obtained from recorded responses showing that the method leads to an excellent estimate of the seismic performance parameters even with a very limited number of recording sensors. The level of accuracy for a given number of recording sensors is linked to the order of the modes contributing to the response: the stronger the influence of the higher modes, the higher the number of sensors needed to attain a given level of accuracy in the estimate of seismic performance parameter