Abstract
This portfolio documents an investigation into improving the ability to predict mean wind speed values in complex terrain from the perspective of the wind power industry. In complex terrain, the uncertainty in wind flow modelling techniques currently represents a significant contribution to the overall uncertainty when producing wind farm energy yield predictions. Energy yield predictions provide essential information to developers looking to valuate and finance wind farm projects. Reducing the uncertainty in these predictions can remove obstacles to development and so contributes to the viability of wind power in general. Two approaches have been taken. The first is to investigate the ability of numerical CFD-based flow modelling methods. The other looks at improving the current industry standard linear flow modelling technique using physically informed post-processing adjustments. Available datasets containing measurements which are suitable for validation of the considered and proposed modelling techniques have been collated. The two CFD-based packages applied here did not exhibit clear improvement in agreement with measurement when compared to current industry standard techniques. This conclusion must be considered in the context of the computational resource and codes available at the time the work was carried out. A pressure gradient parameter and a methodology to derive it from linear flow modelling output is proposed with the aim of predicting the location of flow separation and re-attachment points. This knowledge could then inform a terrain modification enabling a linear flow model to take some account of the region of separated flow so producing improved wind speed predictions in complex terrain. Although this work has not yielded a complete and useable technique it is considered that the results do demonstrate potential for future improvements. Further wind measurements are required to support this field of investigation and allow more generally applicable conclusions to be confidently drawn.