Active Vibration Control of a Nonlinear System with Optimizing The Controller Coefficients Using Metaheuristic Algorithms

An active vibration absorber is utilized in this study for a nonlinear system with unknown multi-harmonic frequency disturbance. At first, a function for disturbance force and its first and second derivatives are estimated. Then the position of the main system is controlled by feedback linearization and sliding mode controllers. A magnetic actuator is designed, which is controlled by a sub-controller. Liunberger observer estimates disturbance function, and the feedback linearization and sliding mode controllers regulate the main system’s position. Metaheuristic algorithms obtain the controller’s coefficients to minimize settling time and errors. Four different techniques, namely, Genetic algorithm, Particle swarm optimization, Simulated annealing, and Teaching-learning-based optimization, are utilized for the optimization process. A magnetic actuator is designed using Faraday and Lorentz’s law for applying the controlling force to the system. Simulation results of the observer have been compared to real value, and the results show the excellent effect of active vibration absorbers on vibration suppression. Moreover, optimizing the controller coefficient shows an improvement in settling time and error. Comparing the algorithms, particle swarm optimization has the best cost function, where Teaching-learning-based optimization has the best-averaged results.