Modal Parameters Identification of Power Transformer Winding Based on Improved Empirical Mode Decomposition Method

Chao Geng,Fenghua Wang,Jun Zhang,Zhijian Jin
DOI: https://doi.org/10.1016/j.epsr.2013.12.002
IF: 3.818
2014-01-01
Electric Power Systems Research
Abstract:Modal parameters of power transformer winding are closely related to transformer manufacturing and detection technology of winding deformation based on vibration analysis method. Aimed at identifying the modal parameters of transformer winding accurately, a modal experiment is designed and made on a real 10 kV power transformer. An improved Empirical Mode Decomposition (EMD) algorithm is proposed to identify the modal parameters of transformer winding with obtained experiment data. First, Ensemble Empirical Mode Decomposition (EEMD) is applied to decompose the vibration signals roughly. Then a masking signal is introduced to eliminate the mode mixing effect with the selected masking frequency. Finally, EMD is used again to get the final Intrinsic Mode Functions (IMFs). By applying the Hilbert transform to the obtained IMFs, modal parameters of power transformer winding are calculated. The PolyMax method is also used to extract modal parameters of transformer winding as a comparison. Calculated results have shown that natural frequency obtained by two methods matches well and the proposed method shows better accuracy in the identification of damping ratio, which verifies the effectiveness of proposed method to recognize the modal parameters of transformer winding with high accuracy. (C) 2013 Elsevier B.V. All rights reserved.
What problem does this paper attempt to address?