A Trainable Equivalent Consumption Minimization Strategy with Soft Actor-Critic for Fuel Cell Hybrid Electric Buses

Li Kang; Qicong Su; Ruchen Huang; Hongwen He

doi:10.1109/VPPC66000.2025.11392891

A Trainable Equivalent Consumption Minimization Strategy with Soft Actor-Critic for Fuel Cell Hybrid Electric Buses

Li Kang^*
, Qicong Su
, Ruchen Huang
, Hongwen He^*

^*Corresponding author for this work

School of Mechanical Engineering

Beijing Institute of Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Energy management strategies (EMSs) are critical for power distribution and fuel economy in hybrid electric vehicles (HEVs). Addressing limitations of deep reinforcement learning (DRL) in optimization and control reliability, this paper proposes a trainable equivalent consumption minimization strategy (ECMS) framework for fuel cell hybrid electric buses (FCHEBs). Firstly, the framework implements the soft actor-critic (SAC) algorithm for dynamic optimization of equivalent factors to enhance fuel economy. Subsequently, an improved ECMS control architecture is constructed with optimized parameters. Ultimately, multi-dimensional validation and comparative analysis using real-world driving cycle data quantitatively demonstrate the proposed EMS's performance. Simulation results demonstrate that the proposed strategy exhibits superior convergence characteristics and SOC stability, achieving 3.81% and 8.29% fuel economy improvements over conventional SAC and adaptive ECMS respectively.

Original language	English
Title of host publication	2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9798331598464
DOIs	https://doi.org/10.1109/VPPC66000.2025.11392891
Publication status	Published - 2025
Event	2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Hangzhou, China Duration: 22 Oct 2025 → 25 Oct 2025

Publication series

Name	2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings

Conference

Conference	2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025
Country/Territory	China
City	Hangzhou
Period	22/10/25 → 25/10/25

Keywords

energy management
fuel cell hybrid electric bus
soft actorcritic optimization
trainable equivalent consumption minimization strategy

Access to Document

10.1109/VPPC66000.2025.11392891

Cite this

Kang, L., Su, Q., Huang, R., & He, H. (2025). A Trainable Equivalent Consumption Minimization Strategy with Soft Actor-Critic for Fuel Cell Hybrid Electric Buses. In 2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings (2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/VPPC66000.2025.11392891

Kang, Li ; Su, Qicong ; Huang, Ruchen et al. / A Trainable Equivalent Consumption Minimization Strategy with Soft Actor-Critic for Fuel Cell Hybrid Electric Buses. 2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2025. (2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings).

@inproceedings{cb05a032206c43e0a8eb94a32f76c8fa,

title = "A Trainable Equivalent Consumption Minimization Strategy with Soft Actor-Critic for Fuel Cell Hybrid Electric Buses",

abstract = "Energy management strategies (EMSs) are critical for power distribution and fuel economy in hybrid electric vehicles (HEVs). Addressing limitations of deep reinforcement learning (DRL) in optimization and control reliability, this paper proposes a trainable equivalent consumption minimization strategy (ECMS) framework for fuel cell hybrid electric buses (FCHEBs). Firstly, the framework implements the soft actor-critic (SAC) algorithm for dynamic optimization of equivalent factors to enhance fuel economy. Subsequently, an improved ECMS control architecture is constructed with optimized parameters. Ultimately, multi-dimensional validation and comparative analysis using real-world driving cycle data quantitatively demonstrate the proposed EMS's performance. Simulation results demonstrate that the proposed strategy exhibits superior convergence characteristics and SOC stability, achieving 3.81\% and 8.29\% fuel economy improvements over conventional SAC and adaptive ECMS respectively.",

keywords = "energy management, fuel cell hybrid electric bus, soft actorcritic optimization, trainable equivalent consumption minimization strategy",

author = "Li Kang and Qicong Su and Ruchen Huang and Hongwen He",

note = "Publisher Copyright: {\textcopyright} 2025 IEEE.; 2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 ; Conference date: 22-10-2025 Through 25-10-2025",

year = "2025",

doi = "10.1109/VPPC66000.2025.11392891",

language = "English",

series = "2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings",

address = "United States",

}

Kang, L, Su, Q, Huang, R & He, H 2025, A Trainable Equivalent Consumption Minimization Strategy with Soft Actor-Critic for Fuel Cell Hybrid Electric Buses. in 2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings. 2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings, Institute of Electrical and Electronics Engineers Inc., 2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025, Hangzhou, China, 22/10/25. https://doi.org/10.1109/VPPC66000.2025.11392891

A Trainable Equivalent Consumption Minimization Strategy with Soft Actor-Critic for Fuel Cell Hybrid Electric Buses. / Kang, Li; Su, Qicong; Huang, Ruchen et al.
2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2025. (2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - A Trainable Equivalent Consumption Minimization Strategy with Soft Actor-Critic for Fuel Cell Hybrid Electric Buses

AU - Kang, Li

AU - Su, Qicong

AU - Huang, Ruchen

AU - He, Hongwen

PY - 2025

Y1 - 2025

N2 - Energy management strategies (EMSs) are critical for power distribution and fuel economy in hybrid electric vehicles (HEVs). Addressing limitations of deep reinforcement learning (DRL) in optimization and control reliability, this paper proposes a trainable equivalent consumption minimization strategy (ECMS) framework for fuel cell hybrid electric buses (FCHEBs). Firstly, the framework implements the soft actor-critic (SAC) algorithm for dynamic optimization of equivalent factors to enhance fuel economy. Subsequently, an improved ECMS control architecture is constructed with optimized parameters. Ultimately, multi-dimensional validation and comparative analysis using real-world driving cycle data quantitatively demonstrate the proposed EMS's performance. Simulation results demonstrate that the proposed strategy exhibits superior convergence characteristics and SOC stability, achieving 3.81% and 8.29% fuel economy improvements over conventional SAC and adaptive ECMS respectively.

AB - Energy management strategies (EMSs) are critical for power distribution and fuel economy in hybrid electric vehicles (HEVs). Addressing limitations of deep reinforcement learning (DRL) in optimization and control reliability, this paper proposes a trainable equivalent consumption minimization strategy (ECMS) framework for fuel cell hybrid electric buses (FCHEBs). Firstly, the framework implements the soft actor-critic (SAC) algorithm for dynamic optimization of equivalent factors to enhance fuel economy. Subsequently, an improved ECMS control architecture is constructed with optimized parameters. Ultimately, multi-dimensional validation and comparative analysis using real-world driving cycle data quantitatively demonstrate the proposed EMS's performance. Simulation results demonstrate that the proposed strategy exhibits superior convergence characteristics and SOC stability, achieving 3.81% and 8.29% fuel economy improvements over conventional SAC and adaptive ECMS respectively.

KW - energy management

KW - fuel cell hybrid electric bus

KW - soft actorcritic optimization

KW - trainable equivalent consumption minimization strategy

UR - https://www.scopus.com/pages/publications/105036123358

U2 - 10.1109/VPPC66000.2025.11392891

DO - 10.1109/VPPC66000.2025.11392891

M3 - Conference contribution

AN - SCOPUS:105036123358

T3 - 2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings

BT - 2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025

Y2 - 22 October 2025 through 25 October 2025

ER -

Kang L, Su Q, Huang R, He H. A Trainable Equivalent Consumption Minimization Strategy with Soft Actor-Critic for Fuel Cell Hybrid Electric Buses. In 2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2025. (2025 IEEE Vehicle Power and Propulsion Conference, VPPC 2025 - Proceedings). doi: 10.1109/VPPC66000.2025.11392891

A Trainable Equivalent Consumption Minimization Strategy with Soft Actor-Critic for Fuel Cell Hybrid Electric Buses

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this