No abstract

Abstract.To effectively detect the hidden fault of transformer winding, the combination of grey correlation analysis and entropy weight theory is proposed to detect the winding condition of power transformer based on the inherent relations of vibration frequency response (VFR) curves of several measured points. The vibration frequency response experiment is made on a real 220kV transformer winding both in normal and loosened conditions. The grey correlation degree and grey area correlation degree of VFR curves in each measured point. Then the entropy weight correlation degree is defined to determine the weight value of all the obtained features. The results have shown that the detection results based on the proposed method are agreed well with the preset condition of transformer winding, which is better illustrated the loosened degree of transformer winding. Compared with the existed FRA method, the vibration frequency response method is more sensitive to the variations of winding condition and has high application value.

Transformer power system is one of the most important electrical equipment, winding transformer failure is an important cause of damage and threaten the safe operation of the power grid. Through theoretical analysis, we propose a Lissajous analysis of transformer voltage / current characteristic analysis method to detect winding faults, while the design of the experimental method, by specific experiments further confirmed Lissajous characteristic analysis method based on effectiveness of transformer winding faults provides a new detection method.

No abstract

No abstract