Enhanced direct torque control based on intelligent approach for doubly-fed induction machine fed by three-level inverter

This paper presents an adaptive neuro-fuzzy inference system (ANFIS) based on 24 sectors direct torque command (DTC) for a doubly-fed induction machine (DFIM) by using a 3-level neutral point clamped inverter. The DTC approach is used in this paper with 24 sectors based on the ANFIS regulator to minimize the torque fluctuations, flux fluctuations, and stator stream THD (Total Harmonic Distortion) of the DFIM drive. The composed technique is accomplished by replacing the hysteresis controllers of the flux and torque with the ANFIS controller. On the other hand, the results of the designed approach based on ANFIS controls were obtained compared to the traditional approach that uses usual controls, as MATLAB was used to realize these approaches in different working conditions. In all tests, the designed method shows improved performance in minimizing torque and flux undulations while reducing the THD of the stator stream. These results highlight the extent of efficiency, high competence, and effectiveness in improving machine features compared to the conventional approach, where the designed approach minimized the THD of current by ratios of approximately 77.50 %, 48.34 %, 75 %, and 81.43 % in all tests. Also, the torque undulation value was reduced by 30 %, 39.24 %, 31.94 %, and 59.31 % in all tests compared to the DTC. The designed approach minimized the speed overshoot compared to the DTC by percentages estimated at 98.83 %, 97.11 %, 96.75 %, and 95.56 % in all tests. These percentages demonstrate the high efficiency and effectiveness of the 24 sectors DTC-ANFIS compared to the conventional approach in improving the features of the control system, making it a promising solution in all electrical fields in the future.