REINFORCEMENT LEARNING-BASED ACTIVE FILTER FOR THE POWER MANAGEMENT OF OFF-ROAD HYBRID ELECTRIC VEHICLES

Changle Xiang, Xuzhao Hou*, Yue Ma

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Citation (Scopus)

Abstract

The high-frequency power management is an essential issue for off-road hybrid electric vehicles (HEVs). On rough terrains, the driving power of HEV changes rapidly. It's hard for the diesel generator set to supply all of the driving power. Therefore, the total power consumption is divided into high-frequency power and low-frequency power by a first-order filter. And then most of the high-frequency power is provided by ultra-capacitor (UC) and battery. By combining reinforcement learning (RL) algorithm, the active power filter is adaptive and attempts to save fuel and prolong battery life. Finally, analysis and simulation results show that the ultra-capacitor pack absorbs lots of high-frequency power and plays an important role in accelerating the response of power supply.

Original languageEnglish
Title of host publicationIET Conference Proceedings
PublisherInstitution of Engineering and Technology
Pages53-58
Number of pages6
Volume2020
Edition3
ISBN (Electronic)9781839534195
DOIs
Publication statusPublished - 2020
Event2020 CSAA/IET International Conference on Aircraft Utility Systems, AUS 2020 - Virtual, Online
Duration: 18 Sept 202021 Sept 2020

Conference

Conference2020 CSAA/IET International Conference on Aircraft Utility Systems, AUS 2020
CityVirtual, Online
Period18/09/2021/09/20

Keywords

  • HYBRID ELECTRIC VEHICLE
  • HYBRID ENERGY STORAGE SYSTEM
  • POWER MANAGEMENT
  • REINFORCEMENT LEARNING

Fingerprint

Dive into the research topics of 'REINFORCEMENT LEARNING-BASED ACTIVE FILTER FOR THE POWER MANAGEMENT OF OFF-ROAD HYBRID ELECTRIC VEHICLES'. Together they form a unique fingerprint.

Cite this