摘要：A new data-mining approach based on power curve profiles is put forward to monitor the power generation performance of wind turbines in this paper. Through assessing the wind-speed power datasets, the weakened power generation performance of turbines could be identified effectively by this approach. Shapes of power curve profiles over consecutive time intervals are constructed by fitting power curve models into wind-speed power datasets. In this research, we designed the Auto-adapt Optimal Interclass Variance algorithm, optimal constraint in each wind-speed power sub-dataset is explored for governing the data-driven method based on distance-based outlier detection and variance analysis model. The AOIV algorithm achieves the self-optimization of the threshold parameter and reaches a high degree of robustness to variations in wind-power generation performance monitoring. The blind industrial researches are conducted to validate the effectiveness of this approach, also indicates the decrease of error rates while detecting weakened power generation performance and the improvement of turbines' power output.
关键词：Wind turbine; Power curve; Data-mining; Performance monitoring
