Indirect Vector Control of Multilevel Inverter fed Induction motor Using ANN Estimator and ANFIS Controller

Abstract

Abstract-The indirect vector control of multilevel inverter fed three-phase induction motor using an adaptive neuro fuzzy controller and neural network estimator is proposed in this paper. The proposed scheme is realized by a three-level inverter, an adaptive neuro-fuzzy controller and two feed forward neural networks. The proposed three-level inverter is a neutral point clamped (NPC) inverter employing hysteresis current control technique for switching the IGBTs. A five layer artificial neural network (ANN) structure is used to tune the fuzzy logic algorithm in adaptive neuro-fuzzy controller. The adaptive neuro-fuzzy inference system (ANFIS) is used in place of conventional PI controller. To tune the fuzzy inference system a hybrid learning algorithm has been adopted. Two feed forward neural networks are used as estimator, learned by the Levenberg-Marquardit algorithm with data taken from PI control simulations. The performance of this scheme is investigated under various load and speed conditions. The simulation results show the suitability and robustness for high performance drive applications.