Year: 2026
Type: article
Abstract: In the following paper, a high-performance control approach to three-phase Permanent Magnet Synchronous Motors (PMSMs) is introduced, relying on the enhanced Field-Oriented Control (FOC) scheme with the application of artificial neural networks. The traditional FOC approach that uses proportional-integral (PI) controllers is characterized by poor performance under parameter changes and nonlinear operating conditions. A neural network-based supervisor is proposed to overcome these deficiencies and enable active reaction, diminish turn ripple, and improve robustness. Both conventional and intelligent control applications are based on a detailed mathematical model of the PMSM and Voltage Source Inverter (VSI) developed in the dq reference frame. The suggested neural network is built into the FOC framework to substitute or augment the traditional PI controllers and be able to provide adaptive and nonlinear control. Simulation tests are conducted at a step reference speed of 0-1000 rpm and different load torque conditions of 0, 5, and 10 Nm. The judgments signify that the NN-based FOC has a reduced rise time, reduce omit, and a taller constant-state veracity than the established technique. In particular, the offered approach decreases the speed error to about 50 rpm at high load conditions and reduces the torque ripple to about 0.3-0.5 Nm, as compared to about 100 rpm and 1 pm-1.5 Nm. Moreover, the system has better rejection of disturbances and resistance to parameter changes. The results obtained prove the fact that the neural network-based FOC gives a considerable performance improvement, thus becoming a viable solution for high-performance PMSM drive applications in electric vehicles, robotics, and industrial automation.
DOI: 10.66261/3q80bb15
Authors: Saif Talal Bahar, Saeed Raheem Salih
Institution: Middle Technical University
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7205006773