A Fast Q-learning Energy Management Strategy for Battery/Supercapacitor Electric Vehicles Considering Energy Saving and Battery Aging

Abstract

Electrified powertrain system brings advantages of improved energy efficiency and reduced fossil fuel consumption, which leads to the electrification of powertrain systems as a key target of automotive industry. Advanced battery technology has been exploited in electric vehicle application and has made considerable progress. However, the degradation of batteries in the vehicle operation could cause adverse effects on the performance and lifespan of electric vehicles. Research on battery/supercapacitor hybrid energy storage systems of electric vehicles that considers energy saving and battery degradation is also lacking. This paper presents a fast Q-learning based energy management strategy to maximize energy saving and minimize battery degradation. Besides, battery only electric vehicle is also studied, and acts as a baseline vehicle. An electrified powertrain system model considering the battery degradation effect is established to form the environment for the Q-learning strategy. Under the training and validation driving cycles, the comparison indicates that the fast Q-learning strategy improves the energy efficiency by 3.83%, and the battery degradation is relieved by 26.36%.