Fault Interpretation Algorithm Using Frequency-ResponseAnalysis of Power Transformers

GONZÁLEZ ARISPE, J. C.; MOMBELLO E.E.

IEEE TRANSACTIONS ON POWER DELIVERY

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC

Lugar: New York; Año: 2015 vol. PP p. 1 - 1

0885-8977

Sweep Frequency Response Analysis (SFRA) is apowerful technique to detect mechanical damage in transformerwindings, such as deformations and displacements, as well asother failures, helping to prevent severe damage. Since SFRA is asensitive diagnostic technique, there are nowadays severalcommercially available measurement instruments to perform thetests, but the failure interpretation is usually based on expertopinion. One of the most advisable ways to face the task ofinterpretation of SFRA test results is the use of mathematicalindexes to detect abnormalities along with an interpretationscheme. The failure interpretation methodology of SFRA testsmust establish the possible relationships between frequencyranges and specific failures, because different failure types can bedetected in different frequency ranges of the SFRA trace. Thisarticle proposes a new methodology for power transformerfailure diagnosis using the relationships between the differentdecomposition levels of the original trace obtained by the use ofthe Discrete Wavelet Transform and the definition of frequencyregions with variable limits, according the shape of the trace.Moreover, an interpretation map and a vector fitting analysis arealso proposed in order to establish a specific failure