A Time-Domain Covariance-Based Parameter Estimation Method for Torsional Shear Buildings

A time-domain parameter estimation algorithm for direct estimation of the stiffness properties of a torsional shear building has been developed. The algorithm decouples the banded global stiffness matrix into story stiffness matrices, and therefore is capable of identifying the lateral and torsional stiffness parameters of each story independent of others. Covariance matrices of the structural response to unknown sources of excitation are used in order to eliminate the external forces from the equations of motion. The method can be applied to ambient vibration data gathered before and after a severe event for damage identification and localization up to the story level. The proposed method is verified using analytical data from Phase 1 of the IASC-ASCE benchmark problem.