A novel real-time energy management strategy based on Monte Carlo Tree Search for coupled powertrain platform via vehicle-to-cloud connectivity. (1st October 2022)

• Record Type: Journal Article
• Authors: Yu, Xiao; Lin, Cheng; Xie, Peng; Liang, Sheng
• Is Part Of: Energy. Volume 256(2022)
• Publication Date: 2022-10-01
• Subjects: Energy management -- Monte Carlo tree search -- Vehicle-to-cloud connectivity -- Electric vehicle -- Coupled powertrain platform; Power resources -- Periodicals; Power (Mechanics) -- Periodicals; Energy consumption -- Periodicals; 333.7905
• DOI: 10.1016/j.energy.2022.124619
• Languages: English
• ISSNs: 0360-5442

Abstract

Abstract: To improve the performance and efficiency of the energy management strategy used in electric vehicles equipped with a dual-motor coupled powertrain platform, this study proposes a systematic real-time search approach via vehicle-to-cloud (V2C) connectivity to reduce the battery degradation and electrical consumption by control working mode and split torque. To be specific, the Monte Carlo Tree Search (MCTS) is employed to search for optimal control sequence in the velocity feasible range in the cloud platform, considering battery loss and electric cost. The logic of time and velocity range updating is proposed as the solution for abrupt traffic changes. To evaluate the effectiveness of the proposed method, a rule-based and an online DP (Dynamic Programming) -based strategy is developed as the baseline approach. Meanwhile, the assessment conditions include standard cycles following power noise and real-world driving cycles. Finally, actual vehicle and hardware-in-the-loop (HIL) experimental results demonstrate that the proposed method significantly outperforms other strategies, the average total cost is 0.36 USD/km, and the improvements are 12.9 % and 11.4 % compared to the rule-based and online DP-based approaches, respectively. Highlights: A novel real-time energy strategy is designed for electric vehicles. Monte Carlo Tress Search is employed to search for optimal control sequence. The traffic information is acquired via vehicle-to-cloud connectivity. HIL bench and vehicle experiments are performed to verify the potential of the proposed strategy.