Cascaded Soft Computing Technique for Enhanced BLDC Motor Performance

Abstract

Electric vehicles, industrial automation, and robots are using brushless DC (BLDC) motors more and more because they are more efficient, reliable, and have a better torque-to-weight ratio. However, traditional Proportional-Integral (PI) and PID controllers lack functionality as well when parameters change, loads change, or DC-link voltage changes, leading to diminished dynamic performance. This research presents a cascaded BLDC motor control system that incorporates two Adaptive Neuro-Fuzzy Inference System (ANFIS) controllers—one for speed regulation and the other for DC-link voltage stabilization—aiming for optimum parameter tuning using the Shuffled Frog Leaping Algorithm (SFLA). The optimization reduces a composite performance index that includes the Integral of Time-weighted Absolute Error (ITAE), overshoot, and DC-link ripple. Simulations in MATLAB/Simulink show that the SFLA-optimized ANFIS works better than regular ANFIS and PI controllers. It has less voltage ripple, smoother torque, more balanced phase currents, and a quicker transient response, all while keeping the steady-state error very low. The findings show that hybrid intelligent optimization works well to make BLDC drives more resilient for real-time application.