Fanwu Chu
Anomaly detection, hydropower generating set, multiple operationconditions, adaptive principal component analysis, statistics moni-toring charts
Anomaly detection of hydropower generating set (HGS) has great significance to the safe operation of power system and to prevent itself from accidents. The conventional principal component analysis (PCA)-based method is unsuitable for anomaly detection of HGS whose operating process is time varying and includes multiple operation conditions (OCs). This paper presents the use of two adaptive PCA algorithms combined with OC of HGS for adaptive process monitoring and anomaly detection of HGS. The proposed method consists of (i) recognition of OCs of HGS, (ii) window size of sample update, and (iii) PCA model update with recursive PCA or moving window PCA (MWPCA). A simulated fault based on filed monitoring data is tested to show the superior performance of the proposed method compared with conventional MWPCA with a fixed window size. Moreover, a real field fault case is analysed to further demonstrate the effectiveness of this method on anomaly detection of HGS.
Important Links:
Go Back
