Data repair and anomaly detection of power meter based on deep reinforcement learning

Jieqiong Han; Wenjing Fan; Chengxin Huo; Zixiang Zhou; Yang Gao

doi:10.1117/12.3015736

6 February 2024 Data repair and anomaly detection of power meter based on deep reinforcement learning

Jieqiong Han, Wenjing Fan, Chengxin Huo, Zixiang Zhou, Yang Gao

Proceedings Volume 12979, Ninth International Conference on Energy Materials and Electrical Engineering (ICEMEE 2023); 129795E (2024) https://doi.org/10.1117/12.3015736
Event: 9th International Conference on Energy Materials and Electrical Engineering (ICEMEE 2023), 2023, Guilin, China

Abstract

The data quality of power metering device is very important for the operation and management of power system. However, these data are often faced with missing, abnormal and noise problems, which may lead to the difficulty of accurate evaluation of power system state and performance. Therefore, the purpose of this study is to propose a method based on deep reinforcement learning to restore missing values in power metering device data, and to detect anomalies. Firstly, the missing data are interpolated by the antagonistic network reconstruction, and the complete data of the power metering device is obtained to repair the power metering data. By constructing a model of abnormity detection based on deep reinforcement learning, a random matrix is established, and a feature index matrix is constructed to decompose the singular value of feature vector. Experimental results show that the proposed method has advantages in accuracy and time efficiency. This method can achieve more than 97% accuracy of anomaly detection and has shorter error detection time. These results verify the feasibility of the proposed method in data processing of power metering devices.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Jieqiong Han, Wenjing Fan, Chengxin Huo, Zixiang Zhou, and Yang Gao "Data repair and anomaly detection of power meter based on deep reinforcement learning", Proc. SPIE 12979, Ninth International Conference on Energy Materials and Electrical Engineering (ICEMEE 2023), 129795E (6 February 2024); https://doi.org/10.1117/12.3015736

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available