Wind resource forecasting using enhanced measure correlate predict (MCP)

The enhancement of Measure Correlate Predict (MCP) using Principal Component Analysis (PCA) is a new wind prediction method based on studying the patterns of historical wind data. The method is trained based on past wind data to predict the wind speed using an ensemble of similar past events. The method is tested based on Meteorological Office (MET-Office) wind speed from a reference site that spans from 2000 to 2010. The last two years (2009 to 2010) were used as training years where the MCP - PCA algorithm learns the wind patterns between the reference(s) and target(s) site. The prediction result is then compared to the actual wind speed distribution at the target site of the training years. The method is further tested with an increase in number of reference sites for predictions. The new prediction results show that the prediction error improves to 23.1 % in average in comparison to a standard linear regression method.