Wind Turbine Tower Abnormal Monitoring Based on Cloud Model
Yali Zhao, Qiankun Hu, Ketao Dong, Peng Guo · 2024
Real-time monitoring of wind turbine output can find the problems in time and ensure the economic benefits of wind farms to the greatest extent. The power main belt formed by the normal operation data in the two-dimensional coordinate system of wind speed and power can accurately reflect the the power generation performance of a wind turbine. Firstly, Mahalanobis distance is used to calculate the deterioration degree of data. Then, Mahalanobis distance of monitoring data is analyzed and transformed by sliding window, and the unit state is visually presented by cloud model by specifying fuzzy comment set, and the classification information of normal, early warning and alarm are given. The results show that the method can react the operating state of unit in a real,objective, quantitative and qualitative way, also can provide reasonable guidance for the follow-up maintenance work.