A New Decision Making Approach to Wind Energy Site Assessment

摘要

This paper describes an objective decision making approach for wind energy site assessment. It is often assumed that a full year of measured data is necessary to make a decision of whether or not to build a wind farm at a site, due to larger uncertainty in estimating the long-term wind resource associated with measurement periods less than one year. The approach in this paper considers if it is possible to make a decision well before a year of measurement is complete, despite the larger uncertainty. The basic premise is that at any point in the site assessment process, one can choose between stopping measurement and building a wind farm, stopping measurement and not building a wind farm, and continuing measurement. Portable ground-based device, such as SODAR and LIDAR, allow measurements to be stopped at a site and recommenced at another site. A decision making framework is developed, in which the three options are evaluated and compared after a period of data acquisition. This approach begins by utilizing Measure-Correlate-Predict to estimate the long-term annual mean wind speed. Next, cash flow analysis, with realistic constraints on the debt service coverage ratio, is used to analyze the economics of building a project. A recursive dynamic program is used to evaluate the option to continue measurement. Finally, the net present value of the cash flows is used to evaluate the three possible choices, and the "best" of the three choices is determined. When the optimal choice is to continue measurement, another measurement period is considered. This approach is applied to fifteen pairs of data sets, and the results indicate that it is an extremely effective method for reducing the average measurement time at a site. Furthermore, the accuracy of the decisions made using this method is equivalent to the accuracy when an entire year of measured on-site data is used. Lastly, this approach achieves significant savings on average in the site assessment process, for all 15 sites, due to both reduced measurement time and more rapid development.