A novel hysteresis current control strategy based on neural network

The relationship among principle of variable-band hysteresis current control, compensation capacity of hysteresis and switching frequency is dissertated in this paper. Aimed at changing the disadvantages of variable-band current control strategy resulted from the fuzzy controller, a control strategy based on command current and current error is proposed and realized by neural network. Power electronics model is built by MATLAB and PSIM and logical control circuit is built according to the fuzzy rules and neural network respectively. Co-simulation results indicate that the hysteresis based on neural network controller has better compensation capacity than the hysteresis based on fuzzy controller and reduces switching frequency.