Detection of Abnormal Sound of Power Plant Equipment Fault based on Self-supervised Learning

Based on the detection of abnormal sound of equipment based on self-supervision, it can help to solve the detection method of abnormal sound, extract artificially constructed algorithms, identify the characteristics of sound, and detect abnormal sound. In this process, the human factor is relatively...

Full description

Saved in:
Bibliographic Details
Published in2022 IEEE 4th International Conference on Power, Intelligent Computing and Systems (ICPICS) pp. 174 - 178
Main Authors Peng, Yumin, Zhong, Xuehui, Yang, Xipeng, Hu, Liehao
Format Conference Proceeding
LanguageEnglish
Published IEEE 29.07.2022
Subjects
Behavioral sciences
Computational modeling
Data processing
equipment failure
Human factors
Production
Self-supervised learning
Semisupervised learning
sound detection
Online AccessGet full text

Cover

More Information
Summary:Based on the detection of abnormal sound of equipment based on self-supervision, it can help to solve the detection method of abnormal sound, extract artificially constructed algorithms, identify the characteristics of sound, and detect abnormal sound. In this process, the human factor is relatively large. Due to human factors, the generality of the method is limited, which affects the accuracy of abnormal sound detection of power plant equipment to a certain extent. The steps include the acquisition of sound from power plant equipment, data processing, production of sound samples, identification of abnormal sound and normal sound samples, and application of the trained sound in the system through an encoder. This makes it possible to detect whether the sound is abnormal. This research is mainly based on a self-supervised learning for abnormal sound detection of power plant equipment faults. The specific research is as follows.
DOI:10.1109/ICPICS55264.2022.9873544