The use of renewable energy is essential for global development and wind power is one of the most attractive among them. Wind turbines occupy a central role in this context and they have had a significant technological development in the last years. Usually the maintenance applied to them is based on preventive models. In the case of critical components is desirable to have a more accurately and frequently enough health forecasting to determine its condition and make decisions instantly before the fault develops. This paper describes a method for estimating health condition indicators based on the evolution of the system life (normal and overload behavior). The idea is that the modeling of these indicators associated with the wind turbine components failure modes help to reschedule dynamically the maintenance plan according to its reallife developments.
Keywords: Wind turbines, normal behavior models, anomaly detection, failure mode risk indicator, maintenance.
Anales de Mecánica y Electricidad. Volume: XC Issue: I Pages: 17-26
DOI reference: ANALES
Published on paper: February 2013.