When? 27-30 September 2021
Porting energy yield calculation to the time domain is an emerging development area that has attracted the attention of different specialists in wind resource analysis practices. One key question is how to integrate into our standard methods the wind resource virtualisation technology that now can offer a time series of modeling wind flow conditions at the turbine scale and which enables generating sub-hourly streams of power production (with the embedded losses) for each wind turbine location and spanning multiple years.
One of the key questions is to evaluate whether the effort to assimilate such amounts of information has a clear return on the investment in terms of resource accuracy improvements and uncertainty reduction.
In this work, we explore these and other questions around the benefits and barriers of upscaling wind resource analysis by adding the time dimension while keeping the required high resolution spatial and temporal granularity.
For that, we present the results from a flagship benchmark exercise for a real wind farm in Texas, exceptionally monitored with 40 met mast. We used as a virtual time series dataset a block of WRF time series downscaling with 100 m and 10’ resolution.
The work presents the results of the multi-tower validation exercise and the comparison against standard distribution based wind resource estimation derived from aggregated values from the WRF data.
We also have analysed the resource analysis exercise from a pure end-user practice by evaluating options to facilitate the use of multidimensional wind time series data.