A Hybrid Pi-ANN Framework for Non-Linearity And Performance Analysis of the Torque and Speed of an Induction Motor

Abstract

Induction motors are crucial in industrial settings due to their reliability and efficiency in converting electrical energy into mechanical energy. Their extensive application across various industries underscores the necessity for ongoing research and development. However, the performance of induction motors are notably affected by nonlinearities caused by factors such as magnetic saturation, thermal variations, and fluctuating load conditions. When handling dynamic variations in control systems, traditional proportional-integral (PI) controllers struggle to adapt quickly and effectively due to their inherent limitations. In contrast, standalone Artificial Neural Network (ANN) models possess a greater capacity for handling these variations but often demand substantial training data and computational resources for optimal performance. Hybrid control systems that combine Proportional-Integral (PI) controllers with artificial neural networks (ANNs) have been designed to effectively manage nonlinear disturbances encountered during motor operations. To assess the efficiency of this hybrid control system, simulations were executed using MATLAB/Simulink, focusing on a model of a 3-phase, 5-horsepower induction motor. The results showed that the hybrid control system outperformed traditional PI controllers in terms of torque ripple, speed overshoot, settling time, and steady-state error, demonstrating the potential for improved motor performance and stability in real-world applications. In conclusion, the integration of artificial neural networks (ANNs) with traditional control methods has proven to be a promising approach for enhancing motor control systems.