Self-prediction (2008-11-11)

[Old Forum Archivist]

I have also a question about self prediction: is it possible to correct this effect when the difference between measured wind speed in .tab file (OWC wizard) and the self prediction exist? If yes, how? In some cases the difference is as large as 0.2, 0.3 m/s and this will affect the entire .wrg file or .rsf estimations because wind speed at mast position will be higher then measured wind speed.

--------

[legrandbleu]

I also encountered this problem in one project. WAsP self prediction of a mast fails by 0,1 in average wind speed (all sectors), and sectorwise results are terrible (deviations of up to 1m/s) particularly in sectors with few data. The widrose is distorted. I wonder if this is an issue of WAsP limitation of capabilities or there is something wrong with the project model.

Any opinions, experiences ? Terrain is rough.

Thanks,

Nick

[Pedrovozone]

I have also experiencing similar issues.
WAsP is overestimating AWS and Power Density by a factor of 1.4 which is a bit appalling.
I have tried to sort this out using two methods:
1 - Applying user corrections to the Reference site. Speedup works ok, while turn is a world of trouble. I could not find a logical way to set the sectors to the right sides. I would then apply these user corrections to all WTG's in the region, which is a lengthy and probably inaccurate enterprise.
2 - Applying a roughness rose on the original Met mast used to do the self-prediction. By iteratively varying the z0 and matching the predicted met mast speed to the measured met mast speed, results look ok in terms of speed, but wind rose is still wrong, so this isn't a complete solution either.
This is the best I have so far.
Anyone else?

[Niels Gylling Mortensen]

I find, that even in quite complex terrain, the overall selfprediction is usually quite ok, even though sector-wise results may vary quite a bit, because of the large speed-ups and turnings. In such cases, the terrain is also often close to (or past) the limitations of the WAsP flow model and one would expect a higher than usual uncertainty. The amount of data available may also influence the selfprediction.

A rescaling procedure based on the selfprediction at the anemometer site will be determined for one point only and the corrections found in this way may not apply to other (turbine) sites (and heights) where the topographical effects are different. So, in most cases, I would not consider or recommend to use corrections based on the selfprediction.

If you work in complex (steep) terrain, I would strongly recommend a RIX analysis of your sites, see e.g. http://www.wasp.dk/Support/Literature.html