Development of an adaptive chaotic fuzzy neural network controller for mitigating seismic response in a structure equipped with an active tuned mass damper

Nowadays, the issue of reducing a structure’s seismic responses has received much attention. Accordingly, numerous advancements have been made in developing the most effective controllers for reducing the seismic response of structural systems during both near- and far-earthquakes. Hence, we propose a robust adaptive chaotic fuzzy neural network controller system built upon type-2 fuzzy systems. The estimation is conducted by utilizing the multilayer perceptron neural network , from which the system’s Jacobian is subsequently derived. The adaptive training of the controller parameters is performed by incorporating the error backpropagation approach and the Extended Kalman filter . Then, the control force on the active tuned mass damper is applied to achieve the desired control objectives. An adaptive interval type-2 fuzzy neural network controller , which uses an Extended Kalman filter and feedback error learning technique, is added as a compensator to the proposed controller to enhance the stability and flexibility of the proposed controller strategy against estimation error, seismic vibrations, and unknown nonlinear functions . Furthermore, this study introduces an additional innovation in the form of an optimized simple adaptive controller , which utilizes an improved grey wolf optimization algorithm for optimization purposes. To evaluate the performance and efficiency, a comparison was made between the results of the adaptive chaotic fuzzy neural network controller and the improved simple adaptive controller and counterparts presented in related works. The results show that significant advantages exist between the adaptive chaotic fuzzy neural network controller and the improved simple adaptive controller in minimizing the maximum displacement, drift, base shear, and acceleration yet preserving a high performance in seismic conditions and parametric uncertainties . The adaptive chaotic fuzzy neural network controller significantly outperforms the optimized simple adaptive control method and other contemporary control strategies to reduce maximum displacement and improve robustness against structural uncertainties and seismic excitations in an 11-story structure equipped with an active tuned mass damper.