The Study Of Dual-PWM Inverter Based On Artificial Neural Networks

Because of uncontrolled rectifier based on diode, general inverter has some shortcomings, such as un-revisable power flow, low power factor and DC voltage uncontrollable and so on. On the basis of general inverter, Dual-PWM inverter use PWM rectifier replacing uncontrolled rectifier. This kind of device can feedback power, improve device's ac-side power factor and achieve DC voltage controllable.In this paper, the Dual-PWM Inverter consists of two parts: PWM rectifier and PWM inverter. First of all, The paper analysis the working principle of VSR(Voltage Source Rectifier),deduce the mathematical model of VSR based on switch function, and deduce the mathematical model under reference frame dq by introducing the idea of rotating reference frame. Then employ feed-forward control technology, the paper designs PI regulator control of current close-loop and voltage closed-loop. Designing PI regulator based on ANN's (Artificial Neural Network) self-learning function to realize parameters self-tuning. Constructing PWM rectifier hysteresis control system and vector control system by Power System Toolbox of MATLAB to carry out simulation. Research of PWM inverter, the paper analysis principle of AC induction motor, and deduce mathematical model under reference frame ABC, dq. Then focus on VOC(vector control system) based on field orientation and SVPWM technology, and carry out simulation. Based on the above-all study, experimental device has been designed, including power circuit, control circuit, signal detection circuit and so on. Realizing V/F control of PWM inverter based on SPMC75LF2313,get good results, and design software of VOC. The design adopts TMS320F2812 to realize numerical control of PWM rectifier double-loop, get parts of results.This paper will offer practical value and to research and develop Dual-PWM Inverter.