Semi-Active Control of a Benchmark Building using Neuro-Inverse Dynamics of MR Damper

A six-storey benchmark problem with semi-active controller based on artificial neural network (ANN) is studied. Linear quadratic regulator (LQR) is used to generate optimal control forces. The MR damper is modeled based on parallel plate Bouc-Wen model. The combination of LQR and ANN controller is utilized to study the benchmark problem for various historical ground motions records available. It is shown that the performance of LQR and ANN is superior over the other control law for controlling the building using MR dampers.