Application of kurtosis in on line detection of transformer iron core looseness

Guo Jie,Chen Xiangxian,Huang Hai,Lin Rilei,Zhou Caikang
DOI: https://doi.org/10.19650/j.cnki.cjsi.2010.11.001
2010-01-01
Abstract:This paper proposes a method that combines wavelet decomposition and kurtosis to detect the looseness of transformer core. The loose conditions are simulated by decreasing the clamping pressure of the core gradually, and spectral kurtosis is used to analyze the impulses occurred in core vibration signals in high frequency band. The kurtosis of high frequency components derived by discrete wavelet analysis is applied to measure the clamping pressure trend. The analysis of multiple groups of data sampled from different measuring points shows that spectral kurtosis can be used to detect the impulses and determine if the core clamping pressure drops considerably, while the kurtosis of high frequency wavelet coefficients can reflect the clamping pressure trend effectively. This method can accurately describe the core loose fault related statistical characteristics of core vibration signals and provides a feasible way to detect the core looseness.
What problem does this paper attempt to address?