Direct adaptive neural-control system for seismically excited non-linear base-isolated buildings

This paper presents a direct adaptive control design to suppress vibrations in nonlinear base-isolated buildings arising due to severe earthquakes. The control design is based on discrete direct adaptive neural control, where the neural controller parameters are adapted using Lyapunov based tuning laws. There is no explicit identification phase in this control scheme, and the resulting controller operates directly on measurements without a state estimator. Performance of the proposed control scheme is evaluated on a full-scale nonlinear three-dimensional base-isolated benchmark structure incorporating lateral-torsion superstructure behavior, and biaxial interaction of nonlinear bearings. The results show that the proposed controller scheme is capable of achieving good response reductions for a wide range of near-fault earth quakes, without corresponding increases in the superstructure responses.