Enhancing Dual-Mode Power-Split HEV Performance: An Intelligent Energy Management Considering Powertrain Dynamic

TY - JOUR

T1 - Enhancing Dual-Mode Power-Split HEV Performance

T2 - An Intelligent Energy Management Considering Powertrain Dynamic

AU - Ren, Xiaolei

AU - Liu, Hui

AU - Nie, Shida

AU - Qin, Yechen

AU - Guo, Lingxiong

N1 - Publisher Copyright: © China Society of Automotive Engineers (China SAE) 2026.

PY - 2026/2

Y1 - 2026/2

N2 - For dual-mode power-split HEVs, the complex transmission configuration of multi-energy sources and dynamic reaction delay makes it difficult to enhance an optimal power-coordinated control performance. Specifically, the power components of a dual-mode HEV face difficulties in quickly aligning with the optimal control directives generated by the energy management strategy (EMS) due to the dynamic reaction delay, negatively affecting EMS’s effectiveness. Motivated by this issue, this article proposes an intelligent EMS of dual HEV, considering the dynamic response of the powertrain mechanism. This strategy uses the improved soft actor-critic (ISTA) approach to generate optimal power allocation commands based on the steady-state assumption. It effectively solves the sensitivity and poor convergence issues commonly observed in current reinforcement learning (RL) investigations. Simultaneously, considering the actual response process of the engine and motors, a model reference adaptive controller (MRAC) is proposed to coordinate the torques among different power components and thus guarantee a balanced mode shift dynamic process and adequate power output. Ultimately, the simulation results show that the fuel consumption of the proposed strategy is reduced by 11.01%, 12.20%, and 12.11% over the benchmark method under three different driving cycles, respectively. Meanwhile, the setting time of output power is decreased by 61.07% to ensure the vehicle’s dynamic performance.

AB - For dual-mode power-split HEVs, the complex transmission configuration of multi-energy sources and dynamic reaction delay makes it difficult to enhance an optimal power-coordinated control performance. Specifically, the power components of a dual-mode HEV face difficulties in quickly aligning with the optimal control directives generated by the energy management strategy (EMS) due to the dynamic reaction delay, negatively affecting EMS’s effectiveness. Motivated by this issue, this article proposes an intelligent EMS of dual HEV, considering the dynamic response of the powertrain mechanism. This strategy uses the improved soft actor-critic (ISTA) approach to generate optimal power allocation commands based on the steady-state assumption. It effectively solves the sensitivity and poor convergence issues commonly observed in current reinforcement learning (RL) investigations. Simultaneously, considering the actual response process of the engine and motors, a model reference adaptive controller (MRAC) is proposed to coordinate the torques among different power components and thus guarantee a balanced mode shift dynamic process and adequate power output. Ultimately, the simulation results show that the fuel consumption of the proposed strategy is reduced by 11.01%, 12.20%, and 12.11% over the benchmark method under three different driving cycles, respectively. Meanwhile, the setting time of output power is decreased by 61.07% to ensure the vehicle’s dynamic performance.

KW - Dual-mode power-split HEV

KW - Energy management

KW - Improved soft actor-critic

KW - MRAC

KW - Torque coordination control

UR - https://www.scopus.com/pages/publications/105033523133

U2 - 10.1007/s42154-024-00351-7

DO - 10.1007/s42154-024-00351-7

M3 - Article

AN - SCOPUS:105033523133

SN - 2096-4250

VL - 9

SP - 242

EP - 262

JO - Automotive Innovation

JF - Automotive Innovation

IS - 1

ER -