Electromagnetic Waveform Identification of Electric Spark Based on BP Neural Network

Electromagnetic wave of electrical spark is a potential cause to eletrical equipment failure. This research focused on identificating and comparative analyzing the different types of electromagnetic waveform generated by eletrical equipment failure based on BP neural networks. After analyzing and ex...

Full description

Saved in:
Bibliographic Details
Published in:Journal of physics. Conference series 2018-08, Vol.1069 (1), p.12108
Main Authors: Li, Tongtong, Tong, Ziyuan, Tang, Shoufeng, Tong, Minming, Ji, Mingyu
Format: Article
Language:English
Subjects:
Back propagation networks
Electric sparks
Electromagnetic radiation
Failure analysis
Feature extraction
Model accuracy
Neural networks
Parameter identification
Physics
Waveforms
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Electromagnetic wave of electrical spark is a potential cause to eletrical equipment failure. This research focused on identificating and comparative analyzing the different types of electromagnetic waveform generated by eletrical equipment failure based on BP neural networks. After analyzing and extracting the features the electromagnetic waveform, BP neural network was chosen to build and train model because of fewer actual samples. The collected standard electromagnetic waveforms were used as the input of the train model, so that the nonlinear mapping between the input electromagnetic wave characteristics and the output electromagnetic wave types of the network can be maximally simulated. The model accuracy was improved by adjusting training parameters after analyzing the results. Then adjusted models were used to identificate the types of electromagnetic waveforms. The result shows that the identification of electrical spark electromagnetic waveform based on BP neural network is effective and feasible.
ISSN:1742-6588
1742-6596
DOI:10.1088/1742-6596/1069/1/012108