A new low-correlation MCP (measure-correlate-predict) method for wind energy forecasting

In the absence of long-term data, measure-correlate-predict methods are of great importance in the assessment of regional wind energy potential. In this study, a new MCP method is introduced for wind energy applications and tested using hourly wind data from four different regions. The method, named as multiple principal least squares (MPLS) method, has an advantage of applicability in the presence of low correlation between the target and reference site wind data. Therefore, this new method might particularly be advantageous when concurrent measurements are not available or when they contain major defects. The results show conclusively that the MPLS method is a strong competitor to the variance ratio method in the existence of concurrency. And without concurrency, the results indicate that it has potential of providing accurate predictions associated with more than 40% improvement using one year or six months long data.