Anomaly detection for condition monitoring data using auxiliary feature vector and density‐based clustering

High‐quality condition monitoring data can provide vital information on power equipment condition assessment and fault diagnosis. However, data quality is difficult to guarantee because valid and invalid anomalies, and different normal data patterns inevitably occur in datasets. This study presents a method for anomaly detection based on auxiliary feature vector and density‐based spatial clustering of applications with noise (DBSCAN). The auxiliary feature vectors of each condition variable are constructed for clustering to recognise normal data patterns and different types of anomalies. Furthermore, a heuristic method based on the ‘number of clusters– Eps ’ curve is proposed for the parameter selection of DBSCAN in an unsupervised setting. Different application examples are implemented on data of dissolved gas content in transformer oil. Compared with state‐of‐the‐art anomaly detection techniques, the proposed method shows an ability to identify and distinguish normal data patterns and valid and invalid anomalies accurately, provided that the condition monitoring data satisfy the assumption of stationarity.