Reliability assessment for Chinese domestic wind turbines based on data mining techniques

In this work, based on the field operating data of a Chinese domestic wind farm, which came from the supervisory control and data acquisition system, data mining techniques are applied to analyze the reliability characteristics of wind turbines and their components. The reliability indexes including time among failures, failure rate, and downtime are analyzed. On that basis, the key components that influence the wind turbines' reliability most seriously are determined. The internal relation between the failure rate and the environmental temperature is identified with correlation function, and time series approach is used to analyze the seasonal feature of the wind turbines' failure rate. The results show that compared with the wind turbines mentioned in the literatures, the failure rate of the current sample is higher. Among the components, the failure rates of electrical and control systems are the highest, while the corresponding repair time is relatively short; on the contrary, the failure rates of main shaft, gearbox, and generator are relatively low, while the average time for maintenance is comparably long. Furthermore, there is an obvious dependency between the failure rate and environmental temperature, and the failure rate has a clear seasonal feature.