| Abstract |
Speed control of electrical drives are often made difficult by nonlinearities, difficulties in proper system parameters identification, unpredictable parameter variations and external load disturbances. As a possibility a model reference adaptive speed control using on-line trained fuzzy neural network (FNN) is presented. In this control method fuzzy-logic controller is equipped with additional option for on-line tuning its chosen parameters. In the paper PI-type fuzzy logic controller is used as the direct speed controller, whose connective weights are trained on-line according to the error between the states of the plant and the reference model. The FNN speed controller is on-line tuned to preserve a favourable model-following characteristics under various operating conditions. The analysis, design, simulation and experimental implementation and comparison of the proposed control schemes are described. |
