Alternating current system fault arc diagnosis method and system based on feature weight enhancement

By using a diagnostic method with enhanced feature weights in the AC system, using VMD and CatBoost models to identify arc faults, the problem of difficulty in identifying low-voltage AC arcs in traditional methods is solved, and the diagnostic accuracy and economicality of fault detection are improved.

CN119720076APending Publication Date: 2025-03-28CHINA ELECTRIC POWER RESEARCH INSTITUTE CO LTD
0 Cites 0 Cited by

Patent Information

Application Number
CN202411715999.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-27
Publication Date
2025-03-28

AI Technical Summary

Technical Problem

In complex household electricity environments, traditional electrical protection devices find it difficult to effectively identify low-voltage AC arcs, especially series fault arcs, resulting in low diagnostic accuracy and making it difficult to effectively detect and prevent fault arcs.

Method used

Using a diagnostic method based on feature weight enhancement, a current signal is sampled by a high-frequency current transformer, and a variety of modal decomposition (VMD) is performed to obtain multi-order IMF components, its Pearson correlation coefficient, variance contribution rate and kurtitude are calculated, the optimal IMF component is determined and weighted and reconstructed, the time and frequency domain characteristic data of the reconstructed signal are extracted, and arc fault detection is detected using CatBoost fusion multi-classifier model.

Benefits of technology

It improves the diagnostic accuracy of AC system fault arcs, and can effectively identify series fault arcs in complex household electricity environments, achieving economical and reasonable fault arc detection and prevention.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides an AC system fault arc diagnosis method and system based on feature weight enhancement, and the method comprises the steps: carrying out the point-by-point sampling of a current in an AC system circuit, and obtaining a point-by-point sampling current signal; performing variational mode decomposition on the current signal to obtain multi-order IMF components, and calculating a Pearson correlation coefficient, a variance contribution rate and kurtosis of each IMF component, and a screening threshold value of the optimal IMF component; determining an optimal component in the plurality of IMF components according to a screening threshold value, and performing weighted reconstruction; extracting time domain and frequency domain feature data of the reconstructed signal; training and testing a pre-constructed CatBoost fusion multi-classifier model used for detecting whether an arc occurs or not through the time domain and frequency domain feature data; and detecting whether an arc occurs in each time window by using a model passing training and testing, and if the number of arc faults occurring in a plurality of continuous time windows is greater than a preset threshold value, determining that the arc fault occurs in the alternating current system. And the accuracy of fault arc diagnosis of the trunk is improved.
Need to check novelty before this filing date? Find Prior Art