Hybrid Sensorless Control Strategy for EV Applications Based on High Frequency Signal Injection and Machine Learning

Wided Zine (1, 2, 3) , Zaatar Makni (4) , Eric Monmasson (1, 3) , Keyu Chen (2) , Lahoucine Idkhajine (3, 1) , Bruno Condamin (2)

1 SATIE - Systèmes et Applications des Technologies de l'Information et de l'Energie
2 Valeo-Siemens eAutomotive France SAS
3 UCP - Université de Cergy Pontoise
4 Valeo systemes de controle moteur, Cergy, France

This paper presents a novel position estimation scheme for interior permanent magnet synchronous machines (IPMSMs) based on merging two self-sensing techniques according to the operating speed. High frequency (HF) signal injection algorithm is deployed for zero and low- speeds, while a Machine Learning (ML) method is solicited for medium and high speeds. The scheme is intended for sensorless torque-controlled drives used in electric vehicle (EV) propulsion. Typically, the proposed sensorless strategy can be applied to safety and limphome applications.

Keywords: Torque; Estimation; Hafnium; Voltage control; Machine vector control; Signal processing algorithms; Machine learning algorithms

Domains: Engineering Sciences [physics]

Wided Zine, Zaatar Makni, Eric Monmasson, Keyu Chen, Lahoucine Idkhajine, et al.. Hybrid Sensorless Control Strategy for EV Applications Based on High Frequency Signal Injection and Machine Learning. 2017 IEEE Vehicle Power and Propulsion Conference (VPPC), Dec 2017, Belfort, France. pp.1-5, ⟨10.1109/VPPC.2017.8330887⟩. ⟨hal-04032562⟩