Metrics for partial discharge detection by the acoustic emission technique to support decision-making

The identification of metrics or ranges of parameters for the detection of partial discharges by the acoustic emission method for the diagnosis and/or classification of power transformers according to their state is essential to support decisions about equipment stoppages at the proper time, i.e., before a failure occurs. This paper discusses these parameters based on the analyses of several field trials, using an algorithm to generate decision trees, as well as a case study to validate the methodology, which was successfully applied in preventive maintenance.