Presentations

Abstract

This presentation describes the RES method for measuring AEP uplift of wind farm enhancements such as aerodynamic upgrades or wake steering. Accurate uplift measurement is challenging, however this is an important capability to ensure the development of technologies which increase output of existing wind farms. The presentation will detail a variety of methods and approaches and identify their merits and shortfalls to progress the industry ability to accurately measure uplift with uncertainty quantification. The RES method can be used for side-by-side turbine analysis where the before and after (or toggling) performance of upgraded turbines is analysed using unchanged reference turbines to measure the energy uplift. The measured uplift can then be modelled and extrapolated to long-term conditions to estimate AEP uplift. The RES method leverages applicable techniques from the power performance standard IEC 61400-12-1 including: * Selection and filtering of 10-minute data * Identifying which neighbouring turbines have wake influence in each wind direction * Site calibration, adapted to calculate the relationship of wind speed as a function of wind direction between two turbines * Calculation of the power curve and associated Category A uncertainty Innovative aspects of the RES method include: * Ability to use multiple types of references including turbines, masts, LiDARs and even reanalysis data * Option to analyse the uplift using all wind directions combined (including waked sectors) by using directional detrending and accounting for the operational state of all upwind turbines for each data record. * Reanalysis data is used in several ways including northing of direction signals and checking for drift over long timescales. * A reverse assessment is performed where the roles of the test and references are swapped. This analysis should yield the opposite result to the original assessment but often does not due to low correlation between test and reference turbines. A correction and uncertainty term are derived from this process which protect the final result from bias. * The method verifies all reference turbines measure 0% uplift for each other. * Results from multiple turbines are combined with careful combination of the uncertainty to calculate a final uplift and uncertainty. * A sensitivity analysis, which leverages techniques from AI and machine learning, shows how the uplift varies against each input field available (temperature, time of day, etc…). This information is used in the long term extrapolation and can be used to inform the next steps of upgrade testing and model development. The benefits of the method include: * Accurate AEP uplift validation is possible in reasonable timescales without extra instrumentation such as LiDAR. * The sensitivity analysis gives more information about the upgrade, revealing the conditions where it is most effective. * A model of the upgrade (for example FLORIS in the case of wake steering) can be simultaneously analysed with the data to determine modelling uncertainty and enable robust long term extrapolation. The presentation will include three real-world case studies of wind farm enhancements on commercial wind farms: * Aerodynamic upgrade using winglet tips, vortex generators and gurney flaps (before vs after test) * Wind farm collective yaw control (toggle test) * Wake steering (toggle test)