Adaptive H∞ filter:: Its application to structural identification

adding the function of memory fading for past observation data to the H-infinity filter, the adaptive H-infinity filter was developed for identifying structural systems with nonstationary dynamic characteristics. Identification algorithms are proposed for time-varying structural systems in which the velocity and displacement of each floor are available for observation, as well as for the case when only the velocity and displacement of some floors are available. The Akaike-Bayes information criterion is used to determine the optimal forgetting factor. Identification algorithms that use the adaptive H-infinity and Kalman filters are applied to a five-degree of freedom (DOF) linear system with nonstationary dynamic characteristics and to a five-DOF nonlinear structural system. Digital simulation results show that the adaptive H-infinity filter efficiently traces the time-varying properties of structural systems. The behavior of the adaptive H-infinity filter is better than that of the adaptive Kalman filter for identifying a structural system with time-varying dynamic characteristics. The former is more efficient and robust for identifying structural systems with nonstationary dynamic characteristics.