In the input power of single-phase Pulse Width Modulation (PWM) rectifiers there is a pulsating portion with twice the grid frequency. To prevent the transferring of pulsating power to the DC side, it should be filtered properly. A convenient method is to install a high-capacitance capacitor at the DC side which cannot be an appropriate solution due to its undesirable characteristics. This paper proposes an active method for eliminating the pulsating power in which the control system is based on Adaptive Network-based Fuzzy Inference System (ANFIS) architecture. To improve the controller performance, an online supervisory learning algorithm based on the Error Back Propagation (EBP) learning method is employed. This approach leads to a control system with minimal structure, enhanced accuracy and improved dynamics. Also, a Proportional Resonant (PR) controller is proposed and the results of two controllers are compared with each other and with one of the most outstanding previous works. Moreover, a superseded control method based on ANFIS architecture for the conventional control system of single-phase PWM rectifiers is suggested. The proposed methods’ efficiency are confirmed by extensive simulations in MATLAB/Simulink.
Read full abstract