Temperature Prediction of Permanent Magnet Synchronous Motor Using AI Techniques for Effective Traction Control

In the realm of traction drive applications, there exists a compelling need for harnessing Artificial Intelligence (AI) techniques to predict electric motor behaviour. AI-driven predictions enable proactive energy management, facilitating the implementation of optimal control strategies that minimize energy consumption and maximize overall system efficiency. In this research, various machine learning (ML), deep learning (DL), and long short-term memory (LSTM) approaches are used to analyze sensor data obtained from a Permanent Magnet Synchronous Motor (PMSM) mounted on a experimental test bench. Different regression models, including Linear Regressor, AdaBoost Regressor, Gradient Boosting Regressor, ElasticNet, K-Neighbor Regressor, and Decision Tree Regressor were utilized as a part of the research to develop efficient control strategies and enhance the efficiency and reliability of PMSMs. Performance of these techniques are evaluated using different metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and the R-squared (RS) coefficients. This research also cross-evaluated different DL and LSTM architectures temperature using various parameters like convergence speed, average training time, optimal epochs for achieving peak accuracy and accuracy improvement. The knowledge gained through this in-depth exploration hold the potential to steer the formulation of advanced control strategies, thereby enhancing the efficiency and dependability of PMSMs for traction drive applications.