The single neuron adaptive PI control of SRM based on IPSO

Abstract

Switched reluctance motor (SRM) has a strong nonlinear characteristic. This makes the traditional PI controller hard to get a good control effect. Single neuron has the simplest structure and fastest calculation speed. It is suitable to be applied on-line. Under certain condition, it can approach any nonlinear function with arbitrary precision and has strong study ability and adaptive ability. So, by combining it with traditional PI controller, the parameters of PI controller can be adaptively adjusted on-line. It is suitable to control nonlinear SRM. To get a super performance, the scale factor is set as a variable varied with dynamic response and improved PSO algorithm is proposed to optimum parameters of single neuron in this paper. This improved the convergence speed and precision of single neural controller. The scale factor K is also treated as a variable. Experimental results show that the system respond quickly, there is little overshot, the precision of stable state is high, also has a strong disturbance reject ability under the control of the proposed control scheme.