An Anomaly Detection Method for Protection Relay System in Distribution Networks based on KPCA-IF Model

Abstract

Due to the complex topology, multiple line branches, and dense spatial distributions of the distribution networks, the disturbances and failures cannot be eliminated it is difficult to completely avoid potential operation disturbances and faults. Thus, a reliable and stable protection system is necessary the protection system is bound to ensure a high level of reliability and stability. In that case, the monitoring and identification of the potential abnormal operation status of the protection system must be ensured the monitoring and identification of potential abnormal operation status of the protection system are facing new challenges. To this end, a data-driven-based real-time anomaly detection model is proposed in this paper. The kernel principal components analysis (KPCA) process is deployed to compress the dimensionality of raw data, which can then reduce the computational complexity within a high-dimensional data environment. Next, the isolated forest (IF) model is applied to excavate potential outliers according to the numeric range of the normal operating state of each feature. It can maintain high detection performance in the biased or sparse distributions and present a high reaction speed. Finally, the operation data of a relay system in a regional distribution network are utilized in the case study. The results verify the better performance of the proposed model in practical application, which can assist in the automatic identification and response of the risks of the distribution network.