Recently, artificial intelligence technique is increasingly receiving attention in solving complex and practical problem and they are widely applying in electrical machine domain. The authors consider also the direct ...Recently, artificial intelligence technique is increasingly receiving attention in solving complex and practical problem and they are widely applying in electrical machine domain. The authors consider also the direct torque control （DTC） as an alternative to conventional methods of control by pulse width modulation （PWM） and by Field oriented control （FOC）, so the application of the DTC based on artificial intelligence can show more advantages and simplified control algorithms with high performance. For this reason, the objectives of this paper can be divided into two parts, the first part is to apply neural networks and fuzzy logic techniques to the DTC control in the presence of a loop speed control comparing to the conventional regulators （as PI） to show the feasibility of these approaches, the second part is to further improve the performance of the neural network by using a neural-fuzzy regulator which combine the advantages of two techniques. Simulation results confirm the validity of the proposed techniques.展开更多
文摘Recently, artificial intelligence technique is increasingly receiving attention in solving complex and practical problem and they are widely applying in electrical machine domain. The authors consider also the direct torque control （DTC） as an alternative to conventional methods of control by pulse width modulation （PWM） and by Field oriented control （FOC）, so the application of the DTC based on artificial intelligence can show more advantages and simplified control algorithms with high performance. For this reason, the objectives of this paper can be divided into two parts, the first part is to apply neural networks and fuzzy logic techniques to the DTC control in the presence of a loop speed control comparing to the conventional regulators （as PI） to show the feasibility of these approaches, the second part is to further improve the performance of the neural network by using a neural-fuzzy regulator which combine the advantages of two techniques. Simulation results confirm the validity of the proposed techniques.
