Improvements to the Eddy Viscosity Wind Turbine Wake Model

Abstract

This paper presents improvements to the eddy viscosity wake model to significantly improve its fidelity without compromising its low computational cost. The eddy viscosity wind turbine wake model, proposed by Ainslie, has been a staple component of the wind yield toolchain for decades. Its low computational cost and excellent predictive power are likely to keep it as such for years to come. However, the standard implementations of the model rely heavily on several heuristics which, although well validated, may not scale well for large turbines or large windfarms.In this paper, some non-physical heuristics are examined, and alternatives proposed which better reflect the underlying physics the model seeks to represent, which is shown to provide much better fidelity at no additional computational cost. The proposed model allows for better predictive capabilities of the performance of the next generation of offshore wind farms. The two key areas investigated are: first the impact of the ambient condition (TI, stability, …) on wake development; and second a re-examination of the wake combination method, leading to the inclusion of three additional processes in the model. It is shown that modifications to these areas provide a more verifiable wake model than the baseline implementations, and allow identification of the remaining physical processes to be further incorporated into future versions.