Probabilistic blind identification of soil-structure systems using extended kalman filter

This paper puts forward a probabilistic approach for simultaneous identification of the dynamic stiffnesses of the soil-foundation and the foundation input motion of soil-structure systems from sparsely-measured structural responses. The soil-structure system in this approach is represented by a shear building resting on swayand rocking-representative springs and dashpots. The proposed approach employs extended Kalman filtering in time domain. This method accounts for two major sources of uncertainty in the problem: the system noise and the measurement noise. In the proposed method, the state vector of the system is augmented by unknown parameters of the soil-structure system, including the soil-foundation representative springs and dashpots. In each time step, a threefold procedure is carried out. First, the current states of the system are predicted by solving the equation of motion in which the posterior estimates of the dynamic stiffnesses and the input of the previous time step are employed. Second, the input of the current time step is estimated using an unbiased, minimum-variance estimator. Third, the current states and the dynamic stiffnesses of the system are updated to obtain a posterior estimate of the dynamic stiffness at the current time step. This iterative process of prediction and updating is repeated for the full duration of the excitation. The end results are the mean vector and the covariance matrix of the unknowns, including the soil-structure parameters and the foundation input motion. Such results have direct applications in risk and reliability analysis of flexible-base structures and rapid damage detection of a building portfolio following an earthquake. 1 MSc student, Department of Civil and Environmental Engineering, Sharif University of Technology, Tehran, Iran 2 Assistant Professor, Department of Civil and Environmental Engineering, Sharif University of Technology, Tehran, Iran (mahsuli@sharif.edu) 3 Postdoctoral Researcher, Department of Civil and Environmental Engineering, University of California, Los Angeles, CA 90095 Jabini A, Mahsuli M, Ghahari S.F. Probabilistic blind identification of soil-structure systems using extended Kalman filter. Proceedings of the 11th National Conference in Earthquake Engineering, Earthquake Engineering Research Institute, Los Angeles, CA. 2018 Probabilistic Blind Identification of Soil-Structure Systems Using Extended Kalman Filter A. Jabini1 M. Mahsuli2 and S.F Ghahari3