Connection between Vave and Ve50

[Petar Dundović]

Dear Wasp Team,

When we upload all required data in Weng to calculate GEWC, which method is used to calculate it?

When we prepare data for WAT, how is Vave calculated and what is the connection between Ve50 and Vave, because on one research we got similar Vave but Ve50 was drastically smaller?

Thank you!

Edited May 3, 2023 by Petar Dundović

[Morten]

WAsP engineering (WEng) has three methods for calculating extreme wind climates 

• Annual maximum method (AM)
• Peak over threshold method (POT)
• Spectral correction method (SCM)

The first two methods works with an observed extreme wind climate (OEWC) at a reference site prepared by the WAsP Climate Analyst (WACA) program. There are two variants of the OEWC since the data for the two methods are sampled differently, either supporting annual extremes or storms above a threshold with a suitable storm separation period. The OEWCs are converted to GEWC by WEng, using flow modelling and geostrophic drag law, similar to the wind atlas method in WAsP. Both OEWC and GEWC objects stores extreme-wind candidates for several directions, since the speed up from reference site or generalized conditions to individual turbines site will depend on direction, and we do not know, which candidate wind will be most extreme at individual turbine sites. Finally, the Gumbel distribution is fitted to local AM or POT extremes wind at the turbine site and used for extrapolation to a fifty-year return period. The Gumbel fitting method for POT is fairly conventional. For AM we use a fitting method based on probability-weighted moments.  

The SCM method is based on reanalysis data, which we store on a server. The data coverage is not yet global, but we are working on it. The length of the reanalysis data is 20 years, which is good for extremes, but reanalysis data are too smoothe in time and space, so direct application would underestimate the extremes. We corrected this error in spectral space, by a combination of the reanalysis spectrum for low frequencies and a high-frequency spectrum of local 10-min data from a mast. The extremes are calculated by moments of the combined spectrum. You can prepare the 10-min time series by WACA. 

I generally recommend the SCM method when supported by our data base, and AM as an alternative if you have local observations from 5+ years. I do not recommend the POT method (in WEng), because our sampling method applies a threshold for the reference mast data and not at the turbine site. WEng POT will generally not converge to a Gumbel distribution for turbines in complex terrain, unless all the extreme winds comes from the same direction so we avoid directional variation in speed up from mast to turbine site. 

WAT input is usually created by the Prepare data for WAT tool inside WEng. The mean wind climate for this is calculated by a WAsP DLL. Thus, WAT applies a mixture of data from WAsP and WEng. WAT recalculates turbine-specific Vave values by sector-wise Weibull distributions coming from WAsP. WAT applies the Ve50 values from WEng and sometimes make a the correction by Ve100 value, as specified in edition 4 of the IEC 61400-1 standard. 

Simple question - long explanation. 

[Petar Dundović]

Thank you for a quick reply, it is very helpfull.