Detection of On-Load Tap-Changer Contact Wear Using Vibroacoustic Measurements

The On-Load Tap-Changer (OLTC), as the sole moving component within a transformer, necessitates routine maintenance to ensure dependable operation. Among the primary maintenance challenges is contact wear in oil-breaking type tap changers, typically assessed through labour-intensive and invasive inspections that require the transformer to be de-energized and the OLTC's active part to be disassembled. Vibro-acoustic measurement (VAM) is a non-invasive method that has been used successfully by several investigators to produce high quality fingerprints of the OLTC operation on energized transformers in a high-voltage substation environment. VAM has proven effective in identifying mechanical faults or degradation, historically relying on visual assessments that demand expert interpretation and are somewhat subjective. To fully leverage the diagnostic capabilities of Vibro-acoustic Measurement (VAM), the development of intelligent algorithms is essential for its practical application in the field. Recent research efforts, including those detailed in this paper, have primarily focused on advancing this aspect of the technique. Most important task for maintenance decisions is the quantitative determination of the contact wear and of the timing sequence of electrical contacts. Using laboratory measurements on a diverter type OLTC, in which different levels of contact wear were simulated in different combinations, a data space for the training of a model and its validation was created. It was shown that the accuracy of the model for the prediction of contact wear is well within the tolerance of OEM specifications for condition-based maintenance decisions without the need for costly outages and internal inspections. Additionally, this study presents a temperature compensation scheme integrated into the linear model, further improving the reliability of the diagnostic.