STRUCTURAL DAMAGE IDENTIFICATION BY UNSCENTED KALMAN FILTER WITH l1 REGULARIZATION

Due to the ill-posedness of inverse problems, the damage identification results of the traditional Kalman filter algorithm are easily influenced by measurement noise, and the identification algorithm may diverge. An Unscented Kalman Filter (UKF) combined with l1 regularization is proposed to identify structural damage. Because local structural damage leads to sparse distribution of damage parameters, the l1 regularization is combined with UKF through pseudo-measurement method to improve the ill-posedness of damage identification problems and to obtain more precise identification results. The numerical analysis and experimental study on beams and trusses show that the proposed algorithm has excellent robustness and can identify the damage location and extents accurately.