화학공학소재연구정보센터
Applied Energy, Vol.148, 220-233, 2015
Data mining techniques for performance analysis of onshore wind farms
Wind turbines are an energy conversion system having a low density on the territory, and therefore needing accurate condition monitoring in the operative phase. Supervisory Control And Data Acquisition (SCADA) control systems have become ubiquitous in wind energy technology and they pose the challenge of extracting from them simple and explanatory information on goodness of operation and performance. In the present work, post processing methods are applied on the SCADA measurements of two onshore wind farms sited in southern Italy. Innovative and meaningful indicators of goodness of performance are formulated. The philosophy is a climax in the granularity of the analysis: first, Malfunctioning Indexes are proposed, which quantify goodness of merely operational behavior of the machine, irrespective of the quality of output. Subsequently the focus is shifted to the analysis of the farms in the productive phase: dependency of farm efficiency on wind direction is investigated through the polar plot, which is revisited in a novel way in order to make it consistent for onshore wind farms. Finally, the inability of the nacelle to optimally follow meandering wind due to wakes is analysed through a Stationarity Index and a Misalignment Index, which are shown to capture the relation between mechanical behavior of the turbine and degradation of the power output. (C) 2015 Elsevier Ltd. All rights reserved.