A Novel Integral Reinforcement Learning based H∞ Control Strategy for Proton Exchange Membrane Fuel Cell in DC Microgrids

Yulin Liu; Tianhao Qie; Yang Yu; Yuxuan Wang; Tat Kei Chau; Xinan Zhang; Ujjal Manandhar; Herbert Iu; Tyrone Fernando

doi:10.1109/TSG.2022.3206281

A Novel Integral Reinforcement Learning based H∞ Control Strategy for Proton Exchange Membrane Fuel Cell in DC Microgrids

Yulin Liu, Tianhao Qie, Yang Yu, Yuxuan Wang, Tat Kei Chau, Xinan Zhang, Ujjal Manandhar, Herbert Iu, Tyrone Fernando

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

Abstract

In this paper, a novel integral reinforcement learning (IRL) based H∞ control algorithm is proposed for proton exchange membrane fuel cells (PEMFCs) to enhance its performance in DC microgrids. The proposed control algorithm produces superior dynamic and steady-state responses without requiring the system model information. The stability of the proposed control algorithm is rigorously proven. Furthermore, its superiority over the PI control, model predictive control (MPC), fuzzy PI control and adaptive MPC is analyzed. Hardware-in-the-loop experimental results are given to verify the effectiveness of the proposed control algorithm.

Original language	English
Number of pages	14
Journal	IEEE Transactions on Smart Grid
DOIs	https://doi.org/10.1109/TSG.2022.3206281
Publication status	E-pub ahead of print - 14 Sep 2022

Access to Document

10.1109/TSG.2022.3206281

Cite this

@article{3ffc4f5bf6b647fa8ae2daa96b7805d4,

title = "A Novel Integral Reinforcement Learning based H∞ Control Strategy for Proton Exchange Membrane Fuel Cell in DC Microgrids",

abstract = "In this paper, a novel integral reinforcement learning (IRL) based H∞ control algorithm is proposed for proton exchange membrane fuel cells (PEMFCs) to enhance its performance in DC microgrids. The proposed control algorithm produces superior dynamic and steady-state responses without requiring the system model information. The stability of the proposed control algorithm is rigorously proven. Furthermore, its superiority over the PI control, model predictive control (MPC), fuzzy PI control and adaptive MPC is analyzed. Hardware-in-the-loop experimental results are given to verify the effectiveness of the proposed control algorithm.",

keywords = "Adaptation models, Batteries, DC Microgrid, Fuel cells, Heuristic algorithms, H∞, Integral Reinforcement Learning, Mathematical models, Microgrids, Observers, Proton Exchange Membrane Fuel Cell",

author = "Yulin Liu and Tianhao Qie and Yang Yu and Yuxuan Wang and Chau, {Tat Kei} and Xinan Zhang and Ujjal Manandhar and Herbert Iu and Tyrone Fernando",

year = "2022",

month = sep,

day = "14",

doi = "10.1109/TSG.2022.3206281",

language = "English",

journal = "IEEE Transactions on Smart Grid",

issn = "1949-3053",

publisher = "IEEE, Institute of Electrical and Electronics Engineers",

}

TY - JOUR

T1 - A Novel Integral Reinforcement Learning based H∞ Control Strategy for Proton Exchange Membrane Fuel Cell in DC Microgrids

AU - Liu, Yulin

AU - Qie, Tianhao

AU - Yu, Yang

AU - Wang, Yuxuan

AU - Chau, Tat Kei

AU - Zhang, Xinan

AU - Manandhar, Ujjal

AU - Iu, Herbert

AU - Fernando, Tyrone

PY - 2022/9/14

Y1 - 2022/9/14

N2 - In this paper, a novel integral reinforcement learning (IRL) based H∞ control algorithm is proposed for proton exchange membrane fuel cells (PEMFCs) to enhance its performance in DC microgrids. The proposed control algorithm produces superior dynamic and steady-state responses without requiring the system model information. The stability of the proposed control algorithm is rigorously proven. Furthermore, its superiority over the PI control, model predictive control (MPC), fuzzy PI control and adaptive MPC is analyzed. Hardware-in-the-loop experimental results are given to verify the effectiveness of the proposed control algorithm.

AB - In this paper, a novel integral reinforcement learning (IRL) based H∞ control algorithm is proposed for proton exchange membrane fuel cells (PEMFCs) to enhance its performance in DC microgrids. The proposed control algorithm produces superior dynamic and steady-state responses without requiring the system model information. The stability of the proposed control algorithm is rigorously proven. Furthermore, its superiority over the PI control, model predictive control (MPC), fuzzy PI control and adaptive MPC is analyzed. Hardware-in-the-loop experimental results are given to verify the effectiveness of the proposed control algorithm.

KW - Adaptation models

KW - Batteries

KW - DC Microgrid

KW - Fuel cells

KW - Heuristic algorithms

KW - H∞

KW - Integral Reinforcement Learning

KW - Mathematical models

KW - Microgrids

KW - Observers

KW - Proton Exchange Membrane Fuel Cell

UR - http://www.scopus.com/inward/record.url?scp=85139411690&partnerID=8YFLogxK

U2 - 10.1109/TSG.2022.3206281

DO - 10.1109/TSG.2022.3206281

M3 - Article

AN - SCOPUS:85139411690

JO - IEEE Transactions on Smart Grid

JF - IEEE Transactions on Smart Grid

SN - 1949-3053

ER -

A Novel Integral Reinforcement Learning based H∞ Control Strategy for Proton Exchange Membrane Fuel Cell in DC Microgrids

Abstract

Access to Document

Other files and links

Fingerprint

Cite this