Modeling the Effect of Environmental and Operating Conditions on Power Converter Reliability in Wind Turbines with Time-Dependent Covariates

In wind turbines, the power-electronic converters used for connection to the power grid are among the most frequently failing subsystems. For root-cause analysis and a subsequent development of effective countermeasures, it is an important task to identify influences having a significant effect on the converter reliability. Based on data from a large fleet of wind turbines spread across five continents, the failure behavior of power converters is modeled by means of a nonhomogeneous Poisson process (NHPP) regression model in the present work. Besides constant covariates characterizing design-related aspects, site-specific environmental and operating conditions of the wind turbines are included in the analysis as time-dependent covariates. As suggested by seasonal and load-dependent patterns observed in the failure behavior, the analysis results confirm that both humidity and the electrical load of the turbines have a significant effect on the reliability of the converter core components and allow for a quantification of these effects.

Keywords: Reliability, Field data, Regression models, Time-dependent covariates, Humidity, Power electronics, Converter, Wind turbines.

Download PDF