Optimal Fractional Order PID Controller Design for AVR System Based on Improved Genetic Algorithm

Fractional order PID controllers have more flexibility in form and have better control effect when compared with conventional integral order PID controllers. The extra two parameters in fractional order PID controllers not only bring flexibility in controller designing but also challenge in parameter tuning. In this paper we propose an improved genetic algorithm for the tuning of fractional order PID controllers and apply the controllers to the automatic voltage stabilization (AVR) system. A lion group selection operator, which simulates the lion group reproduction mode, is proposed to relieve the contradiction between the convergence speed and premature in the genetic algorithm. In order to eliminate the divergence of parameters in the process of tuning, the penalty function with the maximum value of the controller output is added into the objective function. Numerical experiments show that the improved genetic algorithm-lion group genetic algorithm (LGA) has advantages in convergence speed and avoidance of local optimal solutions. Meanwhile, the parameters of fractional order controllers which are given by LGA under the proposed objective function in this paper have excellent convergence.