Abstract

The condition-based maintenance strategy has received significant attention to reduce unscheduled downtime of wind turbines and increase the technical availability. This strategy relies on the analysis of data collected by SCADA (supervisory control and data acquisition system) and/or CMS (condition monitoring system). In practice, the size of collected data can be extremely large if an efficient scheme is not used for data classification and management. This challenge becomes a serious concern in the design of an optimal condition monitoring system, as the numbers of sensors mounted in a wind turbine and wind turbines installed in a farm are expected to increase gradually. This paper proposes a method to classify and manage data in wind turbines. Inverse cumulative frequency function of operational data from a wind turbine and their derivative are used to define criterion for optimal classifications. Four non-trivial classes and one trivial class are defined and proper diagnostics plans for each class are suggested. 7-month operational data from a onshore wind turbine is employed for the research in this paper.