Optimized Speed Regulation of BLDC Motors: A Comparative Performance Study of Artificial Neural Networks and Super-Twisting Sliding Mode Controllers

Abstract

Brushless DC (BLDC) motors have seen substantial development across a wide range of applications, driven by their high performance and inherent robustness. As a result, the demand for accurate control of speed and torque has grown considerably. Despite numerous control strategies from traditional PID controllers to more recent artificial intelligence-based methods achieving precise speed control for BLDC motors remains a significant challenge. This article explores the enhancement of BLDC motor speed control through a comparative analysis of two advanced control strategies: an Artificial Neural Network (ANN) controller, representing intelligent control methods from the field of artificial intelligence, and a Super-Twisting Sliding Mode Controller (STSMC), recognized for its strong robustness against external disturbances. The originality of this work lies in studying systematically the trade-off between control quality and computational complexity for both ANN and STSMC. The simulation results obtained using Matlab/Simulink demonstrate and validate the performance of both controllers. The tests are generally conducted under two scenarios. In the first scenario, the motor is subjected to a constant load torque; the results show that the ANN controller has a response time of 0.25 seconds, while the STSMC is faster with a response time of 0.17 seconds. In the second scenario, under a load torque disturbance, the response times are 0.42 seconds for the ANN and 0.15 seconds for the STSMC.