ĐÁNH GIÁ BỘ ĐIỀU KHIỂN DÙNG MẠNG NƠ-RON MỜ HỒI QUY TRÊN ĐỘNG CƠ KHÔNG ĐỒNG BỘ BA PHA

Abstract

This study aims to investigate the ability of using recurrent fuzzy neural network (RFNN) – based controller on 3-phase induction motors. Accordingly, the supervisory control mechanism is built by combining the traditional PI controller and the RFNN – based controller for self-adjusting parameters, to adapt to changed control conditions. Simulation results on MATLAB show that, when the PI controller runs independently, it gives a high overshoot response. However, when combined with a supervisory controller using a RFNN, the IM speed response has a negligible overshoot. The investigative results also show that online training algorithms of the RFNN - based identifier and controller have monitored and updated the control signal more appropriately, overcoming the limitation of fixed parameter PI controller.