A novel anomaly indicator and early warning system for hydropower units
Xuhui Yue, Yulong Li, Xiao Hu, Dong Liu, Jiaying Liu · 2024
Within the new type of power system characterized by renewable energy, the role of hydropower units becomes increasingly crucial, while their operational conditions grow more challenging. To prevent the occurrence of serious accidents, it is imperative to initiate research into the intelligent fault early warning system for hydropower units. This study proposes a new model to calculate the anomaly indicator for hydropower units, i.e., time-frequency anomaly indicators of vibration signals based on ensemble empirical mode decomposition and sparse auto-encoder. Further, combined with the comprehensive assignment theory, an early warning system is designed for real-time fault warning of hydropower units. A real fault case of a hydropower unit is used to verify the effectiveness of the proposed method. The results show that the proposed time-frequency anomaly indicator is very sensitive to the abnormal change of the vibration status of the hydropower unit, and the proposed warning method has higher reliability compared with other methods.