In this paper, optimization of the design of a fuzzy-genetic controller was done through the IDA and ET analysis methods. Combination controllers versus classical control methods have the ability to handle nonlinear and complex problem, learnability, adaptability and robustness to errors. However, one of the main shortcomings of fuzzy Type-1 systems is ignorance of the uncertainties in the fuzzy rule base. Studies which have been conducted on the benchmark structures show that the training of زontrollers is generally based on the control of the index of damage in a particular earthquake. Controller performance in case of different loading conditions depends on the characteristics of the earthquake used in the training phase, and the controller performance, in other conditions, is not optimal. In this study, the results of the IDA and ET methods for training and designing controllers, as a solution for reducing the uncertainty in structural behavior and loading, were used. Also, the IDA method was used to evaluate the performance of controllers. Examining the strengths and weaknesses of controllers in a wider range of dynamic loading conditions through using the IDA method is applicable. Moreover, evaluation of the results based on a wider range of earthquakes with different maximum accelerations can be done through the IDA method. A benchmark building which was equipped with a MR damper was used in order to evaluate the proposed control method. The results of structural analysis in the proposed controller with uncontrolled structure response, and controlled structures under fuzzy controller that was trained under a special earthquake, were compared and evaluated. Numerical results show that controller training under ET-generated accelerogram in the proposed controller is more effective than other controllers and structural responses under different loading conditions will be optimized. The use of the proposed controller reduces damage in the structure by 15–20% in comparison with other controllers.

Keywords