Article: Detection method of abnormal vibration state of electrical equipment based on random forest Journal: International Journal of Modelling, Identification and Control (IJMIC) 2025 Vol.46 No.2 pp.91 - 99 Abstract: To overcome the problems of low signal-to-noise ratio, low accuracy, and long task completion time in traditional electrical equipment abnormal vibration state detection methods, a new detection method based on random forest is proposed. A signal acquisition architecture is built using fibre Bragg grating sensors to obtain vibration signals of electrical equipment. The joint approximation diagonalisation algorithm performs blind separation on the collected signals, with the high-quality signals obtained through blind separation being input into a random forest to obtain detection results of abnormal vibration states. Experimental results indicate the maximum signal-to-noise ratio of electrical equipment vibration signals reaches 43.67 dB under this method, with abnormal vibration state identification accuracy consistently exceeding 95.6%, and the minimum task completion time being 3.58 s, demonstrating high accuracy and efficiency characteristics. Inderscience Publishers - linking academia, business and industry through research

Title: Detection method of abnormal vibration state of electrical equipment based on random forest

Authors: Xiaoli Xing; Jin Huang

Addresses: Intelligent Manufacturing Institute, Xinxiang Vocational and Technical College, Xinxiang, 453006, China ' Zhengzhou Commercial Technician College, Zhengzhou, 450000, China

Abstract: To overcome the problems of low signal-to-noise ratio, low accuracy, and long task completion time in traditional electrical equipment abnormal vibration state detection methods, a new detection method based on random forest is proposed. A signal acquisition architecture is built using fibre Bragg grating sensors to obtain vibration signals of electrical equipment. The joint approximation diagonalisation algorithm performs blind separation on the collected signals, with the high-quality signals obtained through blind separation being input into a random forest to obtain detection results of abnormal vibration states. Experimental results indicate the maximum signal-to-noise ratio of electrical equipment vibration signals reaches 43.67 dB under this method, with abnormal vibration state identification accuracy consistently exceeding 95.6%, and the minimum task completion time being 3.58 s, demonstrating high accuracy and efficiency characteristics.

Keywords: random forest; electrical equipment; abnormal vibration state; state detection; fibre Bragg grating sensors; blind separation.

DOI: 10.1504/IJMIC.2025.150861

International Journal of Modelling, Identification and Control, 2025 Vol.46 No.2, pp.91 - 99

Received: 20 Dec 2024
Accepted: 29 Jul 2025
Published online: 24 Dec 2025 *

Title: Detection method of abnormal vibration state of electrical equipment based on random forest

Keep up-to-date