Paul Van Lieshout
The installed capacity of wind farms is getting similar to those of conventional power stations. The financing necessary to build these wind farms is getting larger. The financial analysis of such wind farms should thus be robust and should particularly address the quantification of energy produced by the wind farm based on the varying wind resource. It is now time to move away from ‘average’ wind speeds and thus average annual energy productions. The larger investment decisions should be based on confidence levels associated with energy production. This article describes the statistical methodology used to calculate generation output of a wind farm as a function of required confidence levels, for each of the future years in which the wind farm is intended to be operational.
PB Power’s Wind Power and Developing Technologies Group has recently been awarded a further major New Zealand governmental research project aimed at increasing the confidence level of investors and developers in regards to the calculation of future energy production of proposed wind farms. The research undertaken by PB Power, tackles the difficult task of forecasting what the wind resource will be during the operational life of the wind farm, i.e. the wind resource of the future, 20 to 30 years into the future.
The project has been granted based upon the already large amount of knowledge PB Power has accumulated and verified over the years. This document discusses the general methodology and shows the statistical data analysis that PB Power presently undertakes.
As part of the research program, the Wind Power Group is evaluating and developing several different forecasting methodologies including Persistence models, Fast Fourier Transforms, Auto Correlation and Neural Networks. The research has found immediate practical usage in working with financial institutions and developers of wind farms in Europe, China, Australia, the US and New Zealand.