Nonlinear Time Series Analysis of Transformer´s Core Vibration

The vibration signal of transformer´s core is correlative with the conditions of core´s compression and insulation, therefore, vibration monitor of transformer´s core is an effective method for on-line monitoring the power transformer delitescent faults. A new method of obtaining the energy distribution with time and frequency of transformer´s core vibration based on Hilbert-Huang transformation (HHT) was developed. Firstly, empirical mode decomposition (EMD) separated the vibration series into a finite and small number of intrinsic mode functions (IMF). IMF was directly decomposed from vibration series and could reflect the intrinsic physical property of vibration series more clearly. Secondly, Hilbert transform was applied to every IMF. Then, three dimensional amplitude(energy)-frequency-time spectrum designated as Hilbert spectrum was got, where both amplitude and frequency were functions of time. The Hilbert spectrum eliminated the spurious harmonics without physical mean and provided high-resolution time-frequency characteristics of core´s different work states, the test results show core vibration characteristics. It is demonstrated that the new detection method is one effective way for fault diagnosis of transformer´s core