Research of marine voltage inverter based on neural networks PID control

WANG Ze-kai, HE Tong-neng, WANG Chong-fang, XV Dong

(School of Information Engineering, Zhejiang University of Technology, Hangzhou 310023, China)

Abstract: Aiming at the problems of the poor dynamic performance and the difficulty caused by tuning parameters in traditional PID control, which was unable to meet the new requirement of modern marine three-phase inverter power supply, a three-phase half bridge topology was selected for the voltage inverter. The coordinate transformation was used to help analyzing the problems in two-phase rotatingcoordinate axis and also the SVPWM algorithm was chosen to control it. A neural network method combined with PID control was presented because of the characteristics of the BP neural networks which make the control flexible, suitable for time-varying and nonlinear systems and can take good advantages of both. At the same time, the initial weights of BP neural network were optimized through the improved PSO algorithm and it was going to be verified through experiments. The results indicate that three-phase inverter based on the BP-PID control has a very good effect, the total harmonic distortion is controlled in a range and has better dynamic performance smaller overshoot and faster response in sudden-loading/unloading.

Key words: marine inverter; SVPWM algorithms; PID control; BP neural networks; improved PSO