Multiagent Deep Reinforcement Learning-Aided Output Current Sharing Control for Input-Series Output-Parallel Dual Active Bridge Converter

Yu Zeng; Josep Pou; Changjiang Sun; Ali I. Maswood; Jiaxin Dong; Suvajit Mukherjee; Amit Kumar Gupta

doi:10.1109/TPEL.2022.3181243

Multiagent Deep Reinforcement Learning-Aided Output Current Sharing Control for Input-Series Output-Parallel Dual Active Bridge Converter

Yu Zeng, Josep Pou, Changjiang Sun^*, Ali I. Maswood, Jiaxin Dong, Suvajit Mukherjee, Amit Kumar Gupta

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

21 Citations (Scopus)

Abstract

This letter proposes a multiagent soft actor-critic (MASAC) enabled multiagent deep reinforcement learning (MADRL) algorithm for output current sharing of the input-series output-parallel dual active bridge converter. The modular converter is partitioned into different submodules, which are treated as DRL agents of Markov games. Furthermore, all agents are executed decentralized to provide online control decisions after collaborative training. The proposed MASAC algorithm verified in a 150 V/50 V hardware prototype shows optimal dynamic performance.

Original language	English
Pages (from-to)	12955-12961
Number of pages	7
Journal	IEEE Transactions on Power Electronics
Volume	37
Issue number	11
DOIs	https://doi.org/10.1109/TPEL.2022.3181243
Publication status	Published - Nov 1 2022
Externally published	Yes

Bibliographical note

Publisher Copyright:
© 1986-2012 IEEE.

ASJC Scopus Subject Areas

Electrical and Electronic Engineering

Keywords

Input-series output-parallel dual active bridge (ISOP-DAB) converter
multiagent soft actor-critic (MASAC)
output current sharing (OCS)

Access to Document

10.1109/TPEL.2022.3181243

Cite this

@article{e8c0335a1a1e4afa821023a6196bf755,

title = "Multiagent Deep Reinforcement Learning-Aided Output Current Sharing Control for Input-Series Output-Parallel Dual Active Bridge Converter",

abstract = "This letter proposes a multiagent soft actor-critic (MASAC) enabled multiagent deep reinforcement learning (MADRL) algorithm for output current sharing of the input-series output-parallel dual active bridge converter. The modular converter is partitioned into different submodules, which are treated as DRL agents of Markov games. Furthermore, all agents are executed decentralized to provide online control decisions after collaborative training. The proposed MASAC algorithm verified in a 150 V/50 V hardware prototype shows optimal dynamic performance.",

keywords = "Input-series output-parallel dual active bridge (ISOP-DAB) converter, multiagent soft actor-critic (MASAC), output current sharing (OCS)",

author = "Yu Zeng and Josep Pou and Changjiang Sun and Maswood, {Ali I.} and Jiaxin Dong and Suvajit Mukherjee and Gupta, {Amit Kumar}",

note = "Publisher Copyright: {\textcopyright} 1986-2012 IEEE.",

year = "2022",

month = nov,

day = "1",

doi = "10.1109/TPEL.2022.3181243",

language = "English",

volume = "37",

pages = "12955--12961",

journal = "IEEE Transactions on Power Electronics",

issn = "0885-8993",

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

number = "11",

}

TY - JOUR

T1 - Multiagent Deep Reinforcement Learning-Aided Output Current Sharing Control for Input-Series Output-Parallel Dual Active Bridge Converter

AU - Zeng, Yu

AU - Pou, Josep

AU - Sun, Changjiang

AU - Maswood, Ali I.

AU - Dong, Jiaxin

AU - Mukherjee, Suvajit

AU - Gupta, Amit Kumar

PY - 2022/11/1

Y1 - 2022/11/1

N2 - This letter proposes a multiagent soft actor-critic (MASAC) enabled multiagent deep reinforcement learning (MADRL) algorithm for output current sharing of the input-series output-parallel dual active bridge converter. The modular converter is partitioned into different submodules, which are treated as DRL agents of Markov games. Furthermore, all agents are executed decentralized to provide online control decisions after collaborative training. The proposed MASAC algorithm verified in a 150 V/50 V hardware prototype shows optimal dynamic performance.

AB - This letter proposes a multiagent soft actor-critic (MASAC) enabled multiagent deep reinforcement learning (MADRL) algorithm for output current sharing of the input-series output-parallel dual active bridge converter. The modular converter is partitioned into different submodules, which are treated as DRL agents of Markov games. Furthermore, all agents are executed decentralized to provide online control decisions after collaborative training. The proposed MASAC algorithm verified in a 150 V/50 V hardware prototype shows optimal dynamic performance.

KW - Input-series output-parallel dual active bridge (ISOP-DAB) converter

KW - multiagent soft actor-critic (MASAC)

KW - output current sharing (OCS)

UR - http://www.scopus.com/inward/record.url?scp=85132784941&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85132784941&partnerID=8YFLogxK

U2 - 10.1109/TPEL.2022.3181243

DO - 10.1109/TPEL.2022.3181243

M3 - Article

AN - SCOPUS:85132784941

SN - 0885-8993

VL - 37

SP - 12955

EP - 12961

JO - IEEE Transactions on Power Electronics

JF - IEEE Transactions on Power Electronics

IS - 11

ER -

Multiagent Deep Reinforcement Learning-Aided Output Current Sharing Control for Input-Series Output-Parallel Dual Active Bridge Converter

Abstract

Bibliographical note

ASJC Scopus Subject Areas

Keywords

Access to Document

Other files and links

Researchers from Nanyang Technological University Detail New Studies and Findings in the Area of Technology (Multiagent Deep Reinforcement Learning-aided Output Current Sharing Control for Input-series Output-parallel Dual Active Bridge ...)

Cite this

Multiagent Deep Reinforcement Learning-Aided Output Current Sharing Control for Input-Series Output-Parallel Dual Active Bridge Converter

Abstract

Bibliographical note

ASJC Scopus Subject Areas

Keywords

Access to Document

Other files and links

Press/Media

Researchers from Nanyang Technological University Detail New Studies and Findings in the Area of Technology (Multiagent Deep Reinforcement Learning-aided Output Current Sharing Control for Input-series Output-parallel Dual Active Bridge ...)

Cite this