Publicaties

Skip Navigation Links.
Recent verschenen
Expand per documenttypeper documenttype
Expand per Unitper Unit
Expand per Clusterper Cluster

Zoeken naar publicaties:
Beperk het zoeken tot de velden:

ECN publicatie:
Titel:
Evaluation of iSpin anemometry uncertainty analysis
 
Auteur(s):
 
Gepubliceerd door: Publicatie datum:
ECN Windenergie 9-2-2017
 
ECN publicatienummer: Publicatie type:
ECN-E--16-065 ECN rapport
 
Aantal pagina's: Volledige tekst:
28 Download PDF  (625kB)

Samenvatting:
ECN has evaluated the uncertainty assessment of ROMO Wind’s iSpin anemometry comprising the iSpin uncertainty contributions, the iSpin uncertainty in the measured horizontal wind speed and the uncertainty in iSpin applications as for instance in nacelle power curves, etc. This work is based on the documentation provided by ROMO Wind and mainly conducted by DTU.

Regarding the uncertainty contributions ECN believes that the provided set-up is a comprehensive inventory of uncertainty components and a thorough assessment of their contributions to the overall uncertainty budget. ECN believes that the only signi?cant ?aw is the currently unexplained wind tunnel results regarding the angular response. We recommend that more sensors be subjected to angular response tests in the wind tunnel and ideally the source of the deviations be explained.

In addition, some of the theory for uncertainty estimation is based on potential ?ow theory for ideal ?uids on a sphere. ROMO Wind speci?es in their requirements that the “turbine should have an aerodynamic, symmetrical spinner”. ECN recommends that the spinner shape requirements be de?ned in terms of the limitations of the ?ow model. Moreover, it is recommended to specify how the uncertainty assessment is a?ected if the spinner does not have a spherical shape.

In DTU’s report, the uncertainty contributions lead to an equation that expresses the horizontal wind speed as a function of nine key variables following the widely used and ISO 17025 related GUM method. In order to come to concrete values for speci?c cases DTU has developed a complex numerical model to estimate the overall uncertainty of the horizontal wind speed. This model is relatively hard to be applied by pares other than DTU or ROMO Wind.

A more practical approach is desirable that does not rely on complex models and ideally is a simple set of equations that can be applied to any turbine that meets the spinner geometry requirements. To this end, ECN has evaluated a simpli?ed method presented by Demurtas. ECN believes that the method provides a good basis for a best practice uncertainty assessment method. However, the method currently does not take into account important uncertainty contributions, such as the lateral angular response of the iSpin sensor. Given that the source of the lateral angular response of the iSpin sensor is currently unclear, this is not a trivial task and may make the evaluation become more complex.

This is a crossroads in our evaluation. On the one hand ECN has seen a complete and thorough uncertainty contribution assessment leading to a complex numerical model that can hardly be applied outside DTU and ROMO Wind. On the other hand ECN has seen a simpli?ed model, which currently does not incorporate all contributions, and which might become complex upon completion.

In communication by telephone professor Pedersen provided an alternave option in which a detailed, complex uncertainty assessment is provided to the user community as a documented computer model. ECN believes this is an interesting proposal and could even imagine a mixed solution in which ROMO Wind ulizes a complex model to provide parameters for a simpli?ed assessment to the end user, i.e. an uncertainty parameter database for each turbine type as a service for their iSpin users.

Both the research of unclear uncertainty contributions, such as the iSpin angular response, and the investigation of a suitable implementation path of the uncertainty assessment are subject of research at DTU within the framework of the Performance Transparency Project led by ROMO Wind. ECN has also evaluated the iSpin uncertainty assessments for nacelle power curves according to the IEC 61400-12-2 standard. Arguably, this standard has mainly been developed considering nacelle anemometry, so one might question to what extend the uncertainty contributions also apply to spinner anemometry in general and to iSpin technology in particular. Of course this depends on the intended application and three application cases were evaluated.

For the use case of generic application of IEC 61400-12-2 for the estimation of the annual energy production, DTU suggests to neglect the e?ect that the terrain has on the ?ow over the nacelle. Although the iSpin anemometer is mounted on the spinner, in front of the nacelle, the local ?ow conditions may still be a?ected by the terrain. DTU’s claim that “at the spinner, there has not been any ?ow distortion” lacks a reference to the related work.

The use case on relative power curves deals with uncertainty estimation of power curves obtained from the same turbine. In DTU’s estimation of the di?erence of the computed annual energy production, only uncertainty contributions that are uncorrelated between two power curve measurements are included. ECN believes this approach will hold, as long as the set-up remains unaltered and the calibration of the instruments remains valid. ECN believes that for a quantitative comparison, seasonal uncertainty should be included in the uncertainty as well. The in?uence of di?erences in environmental conditions, most notably the turbulence intensity, should be reduced by ?ltering or normalization.

This is also re?ected in the iSpin Guardian approach, where the iSpin nacelle power curve on mulple turbines is determined simultaneously. Here, DTU proposes to exclude the uncertainty of seasonal variability from the nacelle transfer function, because the power curves are measured in the same period. ECN believes that for this application it is valid to exclude seasonal uncertainty from the nacelle power curves, but excluding the in?uence from the nacelle transfer function should only be considered if it is regenerated each time a nacelle power curve is made.


Terug naar overzicht.