This presentation will cover the topic of AI based aging aware control of electric drives for all wheel drive vehicles. The degradation of Permanent Magnet Synchronous Motors (PMSM) over time significantly impacts electric vehicle performance, manifesting in reduced range, diminished power output, and potentially leading to drive failures. Addressing this challenge, this talk focuses on enhancing the longevity of power switches used in traction inverters. The solution relies on integrating Artificial Intelligence and Machine Learning (AIML) alongside predictive optimal control on automotive-grade embedded controllers.
The primary objective is to implement AI/ML techniques, particularly Recurrent Neural Networks (RNN), for anomaly detection in motor performance. By continuously monitoring the Inverters, aging-related anomalies can be promptly identified. When deterioration in any of the four motors is detected, including the components used to control the motors, particularly in traction inverters, the system responds by dynamically adjusting power distribution and torque loads across the motors. This adaptive control mechanism aims to optimize the efficiency and reliability of all-wheel drive EVs in real-time.
The proposed system architecture demonstrates a topology designed to manage multiple motors (up to four) concurrently. Utilizing AIML algorithms embedded within automotive-grade controllers enables seamless integration with existing vehicle systems, ensuring robust and reliable operation under varying conditions.