Neural aided discrete PID active controller for non-linear hysteretic base-isolation building

Abstract

The combination of isolation system and active control devices has been increasingly considered in the structural control community to design an efficient smart hybrid baseisolation system for seismic protection. In this paper, a control scheme based on a combination of discrete PID controller and discrete direct adaptive neural controller is proposed for the active control of a nonlinear base isolated building to reduce superstructure responses and base drifts under near-fault earthquake excitations. Even though the PID controller is a traditional and widely used in many control applications, the performance of PID controller is not satisfactory in time varying and nonlinear systems. But the efficiency of its performance can be enhanced by combining the PID controller along with neural controller. The neural controller is constructed based on a single hidden layer feed forward network and the parameters of the network are modified using extreme learning machine (ELM) — like algorithm. To ensure the stability of the system, unlike original ELM algorithm, Lyapunov update law is used to update the output parameters of the network. This approach is validated by simulating a non-linear three dimensional benchmark base-isolated structure with time history records of three near-fault earthquakes. The performance of the proposed control scheme is measured in terms of a comprehensive set of performance indices. The results show that the proposed neural aided discrete PID active controller is more effective in reducing the superstructure acceleration, inter-storey drifts and base displacement by giving an active feedback control force to the base-isolated structure.