Adaptive neuro-fuzzy and simple adaptive control methods for full three-dimensional coupled buildings subjected to bi-directional seismic excitations

Abstract

Attenuation of three-dimensional coupled buildings under bi-directional seismic excitations using semi-active control devices is pursued in this paper. Multiple magneto-rheological (MR) dampers are employed to connect two adjacent buildings at multiple levels for real-time control of the structural responses. The MR dampers are managed by adaptive neuro-fuzzy inference systems (ANFIS) and simple adaptive control (SAC) methods so a comparison between the performance of both controllers can be made. The displacement feedback type is used for both control methods to design the closed-loop action. The structural system modeled as two three-dimensional buildings connected by frame elements in which the MR dampers are implemented. The equations of motions of the three-dimensional model are formulated by assuming that each floor diaphragm is rigid in its own plane but flexible in the vertical coordinate. Each link is assumed to have three degrees-of-freedom two translational and one rotational at its ends. The adaptive neuro-fuzzy inference system is designed based on Sugeno-type model. Seven triangular membership functions are chosen to fuzzify the input crisp data in the fuzzy logic controller. The training data for ANFIS are generated by the Linear Quadratic Regulator (LQR) under white-noise disturbance. For the simple adaptive controller, LQR is also used to generate the desired trajectories of the reference-model. Numerical simulations are carried out for both symmetrical and asymmetrical coupled buildings under eleven pairs of major earthquakes. The results show that both ANFIS and SAC can deal very successfully with modeling complexities associated with full three-dimensional models subjected to multi-excitations. In terms of seismic responses reduction, both methods have shown a great potential in enhancing the structural performance under seismic activities.