Optimization of control forces in a three-dimensional frame with magnetorheological dampers using a hybrid algorithm

This research presents a new methodology for managing control forces in magnetorheological dampers (MR dampers), applicable to three-dimensional structures. It is based on the programming of a non-dominated classification genetic algorithm NSGA-II, combined with fuzzy logic. The proposed control system, called FLC-1, was implemented on a three-dimensional frame that is part of a real building. The structure was subjected to 8 ground accelerations with different frequency ranges. The FLC-1 simultaneously optimized 50 different fuzzy logic controllers for 100 generations. In each of these controllers, the displacements and velocities of the first floor of the building were used as input parameters and the voltage function was used as the only output parameter to generate the control forces through the MR device. At the same time, a second controller called FLC-2 which operates with classical fuzzy logic was developed. This controller was implemented to verify the efficiency of the FLC-1 to generate optimized control forces in the structure. The results obtained indicate that the FLC-1 significantly improves the structural response of the building with respect to the uncontrolled system and the FLC-2, reducing the maximum floor drifts up to 45%, the maximum RMS response of accelerations up to 29% and the maximum RMS response of displacements up to 50%.