Hierarchical operation switch schedule algorithm for energy management strategy of hybrid electric vehicle using adaptive dynamic programming

• Published in: Sustainable Energy, Grids and Networks Vol. 35; p. 101107
• Main Authors: Li, Fangyuan, Gao, Lefei, Zhang, Yubo, Liu, Yanhong
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.09.2023
• Subjects: Adaptive dynamic programming (ADP); Energy management; Hybrid electric vehicle (HEV); Neural networks; Optimal switch schedule; Hybrid electric vehicle (HEV); Optimal switch schedule; Neural networks; Energy management; Adaptive dynamic programming (ADP)
• ISSN: 2352-4677; 2352-4677
• DOI: 10.1016/j.segan.2023.101107

The cost-benefit of the hybrid electric vehicle (HEV) is strongly dependent on the operation of the energy management system (EMS). In this paper, we propose a novel algorithm named hierarchical operation switch schedule (HOSS) algorithm to solve the optimal operation problem in energy management of the HEV. The original energy management problem of EMS is reconstructed as an optimal switch schedule problem of the engine and motor. In the reconstruction, the time-dependent switch cost of the engine and motor is integrated, which is normally considered difficult and challenging for dynamic programming (DP). The proposed HOSS algorithm can get the optimal switch schedule of the engine and motor of the HEV with a closed-loop feature. In this algorithm, once the neural network is trained, it can cope with different initial conditions and disturbances with no need to retrain. In comparison, re-optimization is usually required for other optimization-based strategies when initial conditions vary. We also compare the HOSS algorithm with other algorithms by conducting simulations, and the results show the unique feature and effectiveness of the HOSS algorithm. •The original energy management problem of EMS is reconstructed as an optimal switch schedule problem of engine and motor of HEV. In the reconstruction, the time-dependent switch cost between the engine and motor is integrated, which is normally considered difficult and challenging for DP.•The proposed HOSS algorithm can get the optimal switch schedule of the engine and motor of the HEV with a closed-loop feature. Facing different initial conditions and disturbance, the HOSS algorithm can still attain optimal results with no need to retrain the neural network. The power schedule of engine and motor can also change accordingly to attain the minimum total cost of the HEV. In comparison, other algorithm such as DP+A-ECMS algorithm cannot deal with different initial conditions and disturbances unless the neural networks are retrained.•By conducting comparison simulations, the algorithm put forward in this paper can attain the optimal switch schedule even when initial conditions are changed and disturbances are existent, which perhaps contribute to the real-time optimal operation of HEV.