Development of an Advanced Online Adaptive FOPID Controller Using the Interval Type 2 Fuzzy Neural Network Optimized With the Levenberg–Marquardt Algorithm for a 20-Story Benchmark Building

Abstract

This paper proposes an innovative control method to reduce the seismic responses of nonlinear structures under the uncertainties of near- and far-field earthquakes. This method is crucial for controlling the seismic response and ensuring structural stability. For this purpose, the robust adaptive FOPID controller is combined with the interval Type 2 fuzzy neural network, whose parameters are optimized through the Levenberg–Marquardt algorithm. An MLP neural network trained using an error backpropagation algorithm is considered for structural system identification and plant estimation. The Jacobian of the estimated model is applied online to the controller. Also, an adaptive compensator, interval Type 2 fuzzy neural networks, is considered to increase the stability and robustness of the proposed controller against estimation error, seismic disturbances, and some unknown nonlinear functions. The extended Kalman filter with feedback error learning strategy is used to maintain the acceptable performance level in the compensator. The performance effectiveness of the proposed controller equipped with a compensator in reducing seismic responses was investigated on a 20-story benchmark building equipped with an active cable damper, and the evaluation criteria were compared with previous works. The results indicate that the IT2FNN-FOPID controller performs better than other controllers in mitigating the seismic responses of the structure during an earthquake and achieving the control objectives. Thus, the J₁ criterion in the El Centro earthquake with an intensity of 1.5 times has improved by about 70% of the ratio of the LQG controller, which is about 60% in the case of the Kobe earthquake.