Fault Diagnosis Methods for Wind Turbines Health Monitoring: a Review

Recently, the rapid expansion of wind energy activity has led to an increasing number of publications that deal with wind turbine health monitoring. In real practice, implementing a prognostics and health management (PHM) strategy for wind turbines is challenging. Indeed, wind turbines are complex electro-mechanical systems that often work under rapidly changing environment and operating load conditions. Although several review papers that address wind turbines fault diagnosis were published, they are mostly focused on a specific component or on a specific category of methods. Therefore, a larger snapshot on recent advances in wind turbine fault diagnosis is presented in this paper. Fault diagnosis approaches could be grouped in three major categories according to the available a priori knowledge about the system behavior: quantitative/qualitative model, signal analysis and artificial intelligence based approaches. Each of the proposed methods in the literature has its advantages and drawbacks. Therefore, a comparison between these methods according to some meaningful evaluation criteria is conducted.