Hybrid KANN-DBSCAN with KNN for Anomaly Detection of Power Plug-in Board

Power plug-in boards are critical components of power control systems, straight influencing the safety and performance of the latter. The diverse categories of plug-in boards in power control systems pose challenges for traditional anomaly detection methods, which suffer from issues such as complex procedures, limited applicability, and results dependent on the expertise of the detection personnel. To address these problems, this paper proposes an anomaly detection method that combines KANN-DBSCAN and improved KNN, offering features like no need for specifying hyperparameters, high accuracy, and fast detection speed. The proposed method is applied to the detection of the main control board of a specific power control system, and experimental tests are conducted. The results demonstrate that compared to the baseline algorithm, the proposed method achieves superior performance and improved detection speed.