Posters

Abstract

Prior to repowering a wind farm it is necessary to assess the wind climate and evaluate the energy resource it represents in the same way one would when developing any wind farm. Production data is available from the operational wind farm that is being repowered that is a valuable source of information, potentially providing a longer dataset than pre-construction measurement campaigns. In addition, it is possible that the pre-construction measurement campaign may have been unsatisfactory, the nacelle anemometry on the wind turbines may be unreliable, and there may have been no independent source of wind measurements on site. Under these circumstances the production data is our primary source of wind climate information. A number of challenges arise when deriving a wind climate from production data, even after the necessary filtering to balance or remove a variety of influences that may introduce bias has been performed. Assuming satisfactory power performance, wind speeds can be derived from power production only between cut-in wind speed and rated wind speed. However, the wind climate assessment for the repowered configuration requires information about the wind climate across all wind speeds. A technique is presented in which the wind speeds predicted from observed production are plotted on a quantile-quantile (Q-Q) plot. The Weibull parameters of the underlying wind speed distribution can then be derived from a straight line fit to this plot which excludes ranges of wind speed where the plot is distorted by power performance below cut-in or above rated wind speeds. This allows the recovery of the complete Weibull distribution in the absence of wind data in these ranges as a result of the derivation of the wind data from power production records. The technique will be demonstrated with a real-world case study in which multiple “single stick” wind turbines were repowered. As a consequence of the installation of single wind turbines, no pre-construction wind assessments were undertaken, as these were not considered economically viable. In addition, no independent wind monitoring was undertaken, and the calibration of nacelle anemometry and the application of a suitable nacelle transfer function could not be assumed. Therefore, power production was the primary source of wind data to inform energy yield assessment of the repowered configuration. The application of a truncated Q-Q plot methodology made resource assessment for the repowered wind farms possible.