A Wind Turbine Fault Diagnosis Method with Self-updating Model based on SCADA Data Mining

This paper presents a fault diagnosis method with self-updating model (FDSU) based on supervisory control and data acquisition (SCADA) data mining. First, a two-step feature selection, including a correlation clustering method and a feature selection method, is proposed to extract the key features. Then based on the features, the expert-based and the learning-based models are combined together, so that the diagnosis can be more accurate and more widely applicable. Moreover, a self-updating model mechanism is proposed, which automatically updates the existed expert-based models to the learning-based models. The SCADA data collected from a wind farm in northern China is used in experiments. The results show that FDSU is more effective in WT fault diagnosis. What's more, the efficiency of FDSU can be continuously improved with the accumulation of data.