Robust speed control of induction motor using neuro-fuzzy controller

In the vector controlled induction motor drives, a good torque control can be obtained, since the coupling between torque and flux of the motor is eliminated. However, adaptive and robust speed control is required for the induction motor drives, due to the nonlinear and unknown motor dynamics and, variations of the motor parameters and loads. Neural-Fuzzy Controllers (NFC) which is an adaptive and nonlinear system is widely used in the control of nonlinear and time varying systems. In the studies with related to electrical drives, NFC is designed using the speed error and the change of error and, another controller is used for eliminating the steady state error seen on the motor speed which is operated under loads. In this thesis, a NFC which is designed using the integral of the speed error is proposed for the vector controlled induction motors. Speed control of the induction motor is implemented using the proposed NFC and, simulation and experimental results are given and compared with results obtained from the other NFCs. Robustness of the NFCs is tested for different speed references and motor loads. The results show that the transient and steady state speed control performance of the proposed NFC is sufficiently good. MATLAB/Simulink is used for simulations and, experimental results are obtained using the DSPACE 1 104 controller. Keywords: Induction motors, Neuro-fuzzy controllers, Vector control.