Discharge fault diagnosis model for transformers based on free characteristic gas

Abstract

Transformers are the critical equipment in power systems. During the operation of transformers, there is a risk of discharging faults such as arc discharge. High-energy discharge fault occurs with a large amount of free gas, which causes transformer fire, explosion, and other accidents. Based on this, this paper designs a discharge fault simulation device in transformer oil. It conducts breakdown discharge, spark discharge, and arc discharge tests to collect free gas generated when a discharge fault occurs. This paper selects four typical characteristic gases to form characteristic gas histograms. It uses a support vector machine (SVM) and ant colony optimization support vector machine (ASVM) to establish a fault recognition model to compare and diagnose the characteristic parameters of the map. The results show that the ant colony optimization support vector function can effectively identify the free distinct gas map of the transformer. This model is suitable for diagnosing and analyzing discharging faults of transformers based on free characteristic gases.