Abstract
The reliability and health of bushings in high-voltage (HV) power transformers is essential in the power supply industry, as any unexpected failure can cause power outage leading to heavy financial losses. The challenge is to identify the point at which insulation deterioration puts the bushing at an unacceptable risk of failure. By monitoring relevant measurements we can trace any change that occurs and may indicate an anomaly in the equipment’s condition. In this work we propose a machine-learning-based method for real-time anomaly detection in current magnitude and phase angle from three bushing taps. The proposed method is fast, self-supervised and flexible. It consists of a Long Short-Term Memory Auto-Encoder (LSTMAE) network which learns the normal current and phase measurements of the bushing and detects any point when these measurements change based on the Mean Absolute Error (MAE) metric evaluation. This approach was successfully evaluated using real-world data measured from HV transformer bushings where anomalous events have been identified.
Highlights
A description of Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) networks is first provided details on how LSTM is incorporated in an Auto-Encoder along with anomaly detection algorithm steps are discussed
Two analysis experiments using for Long Short-Term Memory Auto-Encoder (LSTMAE) model are conducted, and their respective results are presented and discussed
Anomalous events were successfully detected on leakage current and phase angle measured from three bushing taps that are part of a field operating transformer
Summary
40% cited in [3], were found to be related to bushing breakdown in power transformers. Such failures often come with heavy financial consequences, continuous monitoring of bushing condition is justified. The latter can be achieved by monitoring the leakage current measured at the individual bushing test tap [4,5]. There exist other measurement types to assess bushing condition such as power factor [6], capacitance, dissipation factor and partial discharges [7]. Few bushing monitoring methods that analyse leakage current using Machine Learning (ML) methods have been proposed in the literature. In [8], an artificial neural network was trained in a supervised learning manner where the measured amplitude and phase were used as inputs and the bushing condition as an output or label
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
