Development of Anomaly Detection Method for Gas Circuit Breaker with 1D-CAE

Abstract

Gas circuit breakers (GCBs) are commonly used equipment in transmission and distribution systems. GCBs are designed to protect the system from over voltage and over current. When GCBs are damaged, conducting operations can cause expansion of failure parts and delay their restoration, therefore diagnostic methods for GCBs are needed to ensure a stable power supply. In this study, simulation model that can predict behavior of GCBs in the opening operation was constructed with one-dimensional computer aided engineering (1D-CAE). Then, data were generated by the model alternatively measured by sensors for diagnostics. Machine learning method which uses the data for training was employed for anomaly detection and the results showed that it could predict the status of GCBs with good accuracy.