Wavelet Transformation for Recognizing Various Defects in Diagnosis of Power Transformers

The article discusses the use of discrete wavelet transform in diagnosing power oil-filled transformers using automated acoustic monitoring tools. The article is based on the results of experiments using a simulator with an expanded set of defects. Comparison of acoustic signals on a power transformer with signals from a defect simulator made it possible to formulate a criterion for the presence of an insulation defect. An experimental test of the proposed criterion was carried out on signals of various powers and using electrodes of various shapes. The paper presents expressions that implement the criterion for recognizing the type and degree of development of an insulation defect in power transformers. Examples of processing acoustic signals using wavelet transformation in the MatLab program are given. For each type of signals considered, numerical values of the defect recognition criterion are given. At the end of the article, conclusions and prospects for further development of the proposed approach are formulated.