Data-Driven Finite-Set Predictive Current Control via Deep Q-Learning for Permanent Magnet Synchronous Motor Drives

Zichun Tang; Chenwei Ma; Jose Rodriguez; Cristian Garcia; Wensheng Song

doi:10.1109/PRECEDE57319.2023.10174508

Data-Driven Finite-Set Predictive Current Control via Deep Q-Learning for Permanent Magnet Synchronous Motor Drives

Zichun Tang^*, Chenwei Ma, Jose Rodriguez, Cristian Garcia, Wensheng Song

^*Corresponding author for this work

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

2 Scopus citations

Abstract

This paper proposes a finite-set current control strategy based on deep Q-learning for permanent magnet synchronous machine (PMSM) drives. Here, the model-based current prediction of conventional model predictive control is abandoned. Instead, the proposed method selects an optimal switching action in each control period for PMSM drives by training a Deep Q-Network (DQN) to approximate the optimal Q function. Simulations are conducted to demonstrate the effectiveness of the proposed method, showing close performance compared to the conventional finite control set model predictive current control (FCS-MPCC) method.

Original language	English
Title of host publication	2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9798350396867
ISBN (Print)	9798350396867
DOIs	https://doi.org/10.1109/PRECEDE57319.2023.10174508
State	Published - 2023
Event	2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023 - Wuhan, China Duration: 2023 → 2023

Publication series

Name	2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023

Conference

Conference	2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023
Country/Territory	China
City	Wuhan
Period	16/06/23 → 19/06/23

Bibliographical note

Funding Information:
P110420G02016-National Natural Science Foundation Excellent Youth Science Fund Project. Q110422S01010-Science and Technology Innovation Research Team - Electric Traction and Control Sichuan Province Youth Science and Technology Innovation Research Team.

Publisher Copyright:
© 2023 IEEE.

ASJC Scopus subject areas

Control and Optimization
Modeling and Simulation
Energy Engineering and Power Technology
Electrical and Electronic Engineering
Mechanical Engineering
Safety, Risk, Reliability and Quality

Access to Document

10.1109/PRECEDE57319.2023.10174508

Cite this

Tang, Z., Ma, C., Rodriguez, J., Garcia, C., & Song, W. (2023). Data-Driven Finite-Set Predictive Current Control via Deep Q-Learning for Permanent Magnet Synchronous Motor Drives. In 2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023 (2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/PRECEDE57319.2023.10174508

Tang, Zichun ; Ma, Chenwei ; Rodriguez, Jose et al. / Data-Driven Finite-Set Predictive Current Control via Deep Q-Learning for Permanent Magnet Synchronous Motor Drives. 2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023. Institute of Electrical and Electronics Engineers Inc., 2023. (2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023).

@inproceedings{c11d15c1224b4844bd9d57c46d8f04f6,

title = "Data-Driven Finite-Set Predictive Current Control via Deep Q-Learning for Permanent Magnet Synchronous Motor Drives",

abstract = "This paper proposes a finite-set current control strategy based on deep Q-learning for permanent magnet synchronous machine (PMSM) drives. Here, the model-based current prediction of conventional model predictive control is abandoned. Instead, the proposed method selects an optimal switching action in each control period for PMSM drives by training a Deep Q-Network (DQN) to approximate the optimal Q function. Simulations are conducted to demonstrate the effectiveness of the proposed method, showing close performance compared to the conventional finite control set model predictive current control (FCS-MPCC) method.",

keywords = "DQN, Data-driven control, PMSM, deep reinforcement learning",

author = "Zichun Tang and Chenwei Ma and Jose Rodriguez and Cristian Garcia and Wensheng Song",

note = "Funding Information: P110420G02016-National Natural Science Foundation Excellent Youth Science Fund Project. Q110422S01010-Science and Technology Innovation Research Team - Electric Traction and Control Sichuan Province Youth Science and Technology Innovation Research Team. Publisher Copyright: {\textcopyright} 2023 IEEE.; 2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023 ; Conference date: 16-06-2023 Through 19-06-2023",

year = "2023",

doi = "10.1109/PRECEDE57319.2023.10174508",

language = "English",

isbn = "9798350396867",

series = "2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023",

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

booktitle = "2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023",

}

Tang, Z, Ma, C, Rodriguez, J, Garcia, C & Song, W 2023, Data-Driven Finite-Set Predictive Current Control via Deep Q-Learning for Permanent Magnet Synchronous Motor Drives. in 2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023. 2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023, Institute of Electrical and Electronics Engineers Inc., 2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023, Wuhan, China, 16/06/23. https://doi.org/10.1109/PRECEDE57319.2023.10174508

Data-Driven Finite-Set Predictive Current Control via Deep Q-Learning for Permanent Magnet Synchronous Motor Drives. / Tang, Zichun; Ma, Chenwei; Rodriguez, Jose et al.
2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023. Institute of Electrical and Electronics Engineers Inc., 2023. (2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023).

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

TY - GEN

T1 - Data-Driven Finite-Set Predictive Current Control via Deep Q-Learning for Permanent Magnet Synchronous Motor Drives

AU - Tang, Zichun

AU - Ma, Chenwei

AU - Rodriguez, Jose

AU - Garcia, Cristian

AU - Song, Wensheng

N1 - Funding Information: P110420G02016-National Natural Science Foundation Excellent Youth Science Fund Project. Q110422S01010-Science and Technology Innovation Research Team - Electric Traction and Control Sichuan Province Youth Science and Technology Innovation Research Team. Publisher Copyright: © 2023 IEEE.

PY - 2023

Y1 - 2023

N2 - This paper proposes a finite-set current control strategy based on deep Q-learning for permanent magnet synchronous machine (PMSM) drives. Here, the model-based current prediction of conventional model predictive control is abandoned. Instead, the proposed method selects an optimal switching action in each control period for PMSM drives by training a Deep Q-Network (DQN) to approximate the optimal Q function. Simulations are conducted to demonstrate the effectiveness of the proposed method, showing close performance compared to the conventional finite control set model predictive current control (FCS-MPCC) method.

AB - This paper proposes a finite-set current control strategy based on deep Q-learning for permanent magnet synchronous machine (PMSM) drives. Here, the model-based current prediction of conventional model predictive control is abandoned. Instead, the proposed method selects an optimal switching action in each control period for PMSM drives by training a Deep Q-Network (DQN) to approximate the optimal Q function. Simulations are conducted to demonstrate the effectiveness of the proposed method, showing close performance compared to the conventional finite control set model predictive current control (FCS-MPCC) method.

KW - DQN

KW - Data-driven control

KW - PMSM

KW - deep reinforcement learning

UR - http://www.scopus.com/inward/record.url?scp=85166189622&partnerID=8YFLogxK

U2 - 10.1109/PRECEDE57319.2023.10174508

DO - 10.1109/PRECEDE57319.2023.10174508

M3 - Conference contribution

AN - SCOPUS:85166189622

SN - 9798350396867

T3 - 2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023

BT - 2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023

Y2 - 16 June 2023 through 19 June 2023

ER -

Tang Z, Ma C, Rodriguez J, Garcia C, Song W. Data-Driven Finite-Set Predictive Current Control via Deep Q-Learning for Permanent Magnet Synchronous Motor Drives. In 2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023. Institute of Electrical and Electronics Engineers Inc. 2023. (2023 IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics, PRECEDE 2023). doi: 10.1109/PRECEDE57319.2023.10174508

Data-Driven Finite-Set Predictive Current Control via Deep Q-Learning for Permanent Magnet Synchronous Motor Drives

Abstract

Publication series

Conference

Bibliographical note

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this