An Improved Distributed Maximum Power Point Tracking Technique in Photovoltaic Systems Based on Reinforcement Learning Algorithm

Zhihong Ge; Xingshuo Li; Fei Xu; Haimeng Wu; Ruichi Wang; Shuye Ding

doi:10.1109/jestie.2023.3332572

An Improved Distributed Maximum Power Point Tracking Technique in Photovoltaic Systems Based on Reinforcement Learning Algorithm

Zhihong Ge, Xingshuo Li, Fei Xu, Haimeng Wu, Ruichi Wang, Shuye Ding

Mathematics, Physics and Electrical Engineering

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

Abstract

The mismatch problem is commonly happened in photovoltaic (PV) systems due to partial shading conditions(PSCs). Distributed maximum power point tracking
(DMPPT) architectures can be used to solve such problem. Reinforcement learning (RL) method which is one of the advanced artificial intelligence(AI) methods is proposed to improve the tracking speed. However, the drawbacks such as lack of limited adaptability and exploration-exploitation trade-off theory make the RL method low in efficiency. Therefore, this article combines the Beta method and ε - greedy algorithm with the RL method to address this problem. The simulation and experimental tests have been carried out. The result shows the efficiency of the proposed RL method is up to 96.85%, which verifies the superiority of the proposed scheme.

Original language	English
Pages (from-to)	1-12
Number of pages	12
Journal	IEEE Journal of Emerging and Selected Topics in Power Electronics
Early online date	13 Nov 2023
DOIs	https://doi.org/10.1109/jestie.2023.3332572
Publication status	E-pub ahead of print - 13 Nov 2023

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1109/jestie.2023.3332572

An Improved Distributed Maximum Power Point Tracking Technique - full paperAccepted author manuscript, 2.83 MB

Cite this

@article{eb3cc9615cb7453fa3080739860f7fcd,

title = "An Improved Distributed Maximum Power Point Tracking Technique in Photovoltaic Systems Based on Reinforcement Learning Algorithm",

abstract = "The mismatch problem is commonly happened in photovoltaic (PV) systems due to partial shading conditions(PSCs). Distributed maximum power point tracking(DMPPT) architectures can be used to solve such problem. Reinforcement learning (RL) method which is one of the advanced artificial intelligence(AI) methods is proposed to improve the tracking speed. However, the drawbacks such as lack of limited adaptability and exploration-exploitation trade-off theory make the RL method low in efficiency. Therefore, this article combines the Beta method and ε - greedy algorithm with the RL method to address this problem. The simulation and experimental tests have been carried out. The result shows the efficiency of the proposed RL method is up to 96.85%, which verifies the superiority of the proposed scheme.",

keywords = "Distributed maximum power point tracking (DMPPT, reinforcement learning(RL), partial shading conditions (PSCs), PV systems",

author = "Zhihong Ge and Xingshuo Li and Fei Xu and Haimeng Wu and Ruichi Wang and Shuye Ding",

year = "2023",

month = nov,

day = "13",

doi = "10.1109/jestie.2023.3332572",

language = "English",

pages = "1--12",

journal = "IEEE Journal of Emerging and Selected Topics in Power Electronics",

issn = "2168-6777",

publisher = "IEEE",

}

TY - JOUR

T1 - An Improved Distributed Maximum Power Point Tracking Technique in Photovoltaic Systems Based on Reinforcement Learning Algorithm

AU - Ge, Zhihong

AU - Li, Xingshuo

AU - Xu, Fei

AU - Wu, Haimeng

AU - Wang, Ruichi

AU - Ding, Shuye

PY - 2023/11/13

Y1 - 2023/11/13

N2 - The mismatch problem is commonly happened in photovoltaic (PV) systems due to partial shading conditions(PSCs). Distributed maximum power point tracking(DMPPT) architectures can be used to solve such problem. Reinforcement learning (RL) method which is one of the advanced artificial intelligence(AI) methods is proposed to improve the tracking speed. However, the drawbacks such as lack of limited adaptability and exploration-exploitation trade-off theory make the RL method low in efficiency. Therefore, this article combines the Beta method and ε - greedy algorithm with the RL method to address this problem. The simulation and experimental tests have been carried out. The result shows the efficiency of the proposed RL method is up to 96.85%, which verifies the superiority of the proposed scheme.

AB - The mismatch problem is commonly happened in photovoltaic (PV) systems due to partial shading conditions(PSCs). Distributed maximum power point tracking(DMPPT) architectures can be used to solve such problem. Reinforcement learning (RL) method which is one of the advanced artificial intelligence(AI) methods is proposed to improve the tracking speed. However, the drawbacks such as lack of limited adaptability and exploration-exploitation trade-off theory make the RL method low in efficiency. Therefore, this article combines the Beta method and ε - greedy algorithm with the RL method to address this problem. The simulation and experimental tests have been carried out. The result shows the efficiency of the proposed RL method is up to 96.85%, which verifies the superiority of the proposed scheme.

KW - Distributed maximum power point tracking (DMPPT

KW - reinforcement learning(RL)

KW - partial shading conditions (PSCs)

KW - PV systems

U2 - 10.1109/jestie.2023.3332572

DO - 10.1109/jestie.2023.3332572

M3 - Article

SN - 2168-6777

SP - 1

EP - 12

JO - IEEE Journal of Emerging and Selected Topics in Power Electronics

JF - IEEE Journal of Emerging and Selected Topics in Power Electronics

ER -