Application of Operating Deflection Shapes to the Vibration-Based Mechanical Condition Monitoring of Power Transformer Windings

• Published in: IEEE transactions on power delivery Vol. 36; no. 4; pp. 2164 - 2173
• Main Authors: Shi, Yuhang, Ji, Shengchang, Zhang, Fan, Dang, Yongliang, Zhu, Lingyu
• Format: Journal Article
• Language: English
• Published: New York IEEE 01-08-2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Calibration; Coils (windings); Condition monitoring; Conductors; Correlation coefficients; Deflection; Electric potential; Excitation; histogram of oriented gradient; Histograms; Load tests; Mathematical model; Mechanical condition monitoring; operating deflection shapes; Power transformers; Resonant frequency; Transformers; Vibration measurement; Vibration monitoring; Vibrations; Voltage; Winding; Windings
• ISSN: 0885-8977; 1937-4208
• DOI: 10.1109/TPWRD.2020.3021909

Operating deflection shapes (ODSs) are used for visualizing the vibration patterns of a structure. In this study, ODSs are applied to diagnose the mechanical condition of power transformer windings. First, a synchronous calibration measuring system is developed to simultaneously acquire tank vibration signals and excitation voltage signals. To perform the synchronous calibration, the phase of the excitation voltage signals is set as the reference. Next, the no-load test, load test, and variable load test are conducted on a 50-kVA transformer. The histogram of oriented gradient (HOG) and Pearson correlation coefficient (CC) are proposed to describe the differences between the different ODSs. Experimental results show that the ODSs are insensitive to the operating voltage and current, but sensitive to the clamping pressure of the windings. Field results further verify the significant change in the ODS after winding deformation. Finally, the partial ODS, which is suitable for practical transformers, is proposed to locate winding mechanical defects in power transformers. The proposed method can improve the efficiency and accuracy of the vibration-based mechanical condition monitoring of power transformers.