Year: 2025
Type: article
Abstract: Traditional direct torque control (TDTC) is a widely used technique for controlling Permanent Magnet Synchronous Motors (PMSMs).However, it has significant limitations, including large torque ripple and excessive overspeed.These issues can lead to suboptimal system performance and instability, which requires a more efficient control strategy.This study aims to develop a new control method that addresses the limitations of TDTC, improving both speed and torque controllability while minimizing oscillations and noise in PMSM drives.This paper proposes a new control method that combines the high-order sliding mode (HOSM) technique with the radial basis function neural network (RBFNN).The proposed method, called RBFNNH-DTC, leverages the robustness of HOSM for noise rejection and the adaptability of RBFNN for precise control, integrating the two methods to enhance the noise rejection and improve the accuracy of speed and torque control in PMSM systems.Simulation results in MATLAB/Simulink show that RBFNNH-DTC significantly outperforms TDTC and sliding mode control-based DTC (SMC-DTC).Specifically, RBFNNH-DTC achieves a rise time of 0.2 sec without settling time, while TDTC and SMC-DTC experience large speed fluctuations and high torque overshoots of 4.273% and 9.441%, respectively.Regarding torque, RBFNNH-DTC maintains stable operation with a maximum torque of 5.443 Nm while minimizing adverse torque fluctuations.In addition, RBFNNH-DTC generates a near-ideal stator current waveform with significantly lower total harmonic distortion (THD) than other methods.The results demonstrate that RBFNNH-DTC provides superior torque and speed control performance, minimizing oscillation and disturbance.This method significantly improves conventional DTC techniques, making it a promising solution for highperformance PMSM drives.
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4408494898