AWS Truewind Logo

SiteWind

SiteWind® provides the finest in micrositing services. In demanding situations where developers require more accurate wind resource and plant output estimates, SiteWind provides the best solution.

An enhanced version of AWS Truewind’s MesoMap technology, SiteWind outperforms conventional micrositing models such as WASP, and even relatively sophisticated CFD models, in complex wind climates. Case studies have demonstrated improvements of up to 40-70% in accuracy compared to WASP. SiteWind is compatible with leading wind plant design programs such as WindFarmer and WindPro and is accepted for project financing.

Value

All developers face the same challenge: how to obtain an accurate estimate of the potential energy production of their project quickly and at the lowest cost.

The traditional approach has been to take measurements and then feed the results into a simplified micrositing model such as WASP or, more recently, a computational fluid dynamics (CFD) model such as WindSim. Such models, whatever their type, extrapolate the wind resource from a single point, and thus their accuracy can decline dramatically with distance away from the mast, especially in complex wind climates.

In response to this problem, some developers have taken to erecting numerous masts, often as many as one every kilometer or two, to constrain the models. But this solution, though effective, is time-consuming and expensive.

SiteWind improves on other models because it first develops a comprehensive picture of the wind climate in the surrounding region using a sophisticated mesoscale atmospheric model before zooming into the project area. Technical details aside, the result is a dramatic reduction in error compared to conventional models driven from a single reference mast. SiteWind offers these specific benefits:

  • With data from one mast, SiteWind produces a more accurate estimate of wind plant output than other models do. This result can lead directly to lower risks and increased returns for investors.
  • Conversely, SiteWind requires fewer masts to achieve the same level of accuracy as other models. This feature can result in substantial savings for each unneeded mast.
  • SiteWind can be used as an effective site selection tool. Often developers have data for one site but want to know if other sites in the region have a better resource. For regions as large as 2500 to 5000 km2, SiteWind can provide a rapid, affordable, and reliable answer to this question. (For larger regions, MesoMap® can be used.
  • SiteWind is backed by AWS Truewind’s extensive experience in all aspects of wind project development. Furthermore, SiteWind has been field-tested in numerous areas, from Altamont Pass to the Gaspe Peninsula, and has been accepted for project financing.

Technology

SiteWind’s key advantage over other micrositing methods is that it combines the strengths of three tools: a mesoscale weather model (presently MASS), a microscale wind flow model, and on-site measurements. As a model developer—not just a user—AWS Truewind is continually improving and updating its models and techniques.

The mesoscale model simulates important features of the wind climate that are beyond the capacity of microscale models alone (even relatively advanced CFD models), such as katabatic (downslope) mountain winds, channeling through mountain passes, lake and sea breezes, low-level jets, and temperature inversions. Experience has shown that such effects can have a substantial impact on the wind resource even within the area of a single wind project.

Because of their huge computational demands, however, mesoscale models cannot achieve the high resolution required for micrositing. SiteWind overcomes this limitation by using a microscale model to sharpen the mesoscale picture and produce a highly detailed wind resource map (typically 100m resolution).

To complete the picture, on-site data from one or more meteorological (met) masts are used to correctfor any errors in the predicted speed and direction. The end result achieves the highest possible modeling accuracy while remaining closely tied to measurement.

SiteWind’s mesoscale model, MASS (Mesoscale Atmospheric Simulation System), is a non-hydrostatic weather model developed by AWS Truewind partner MESO, Inc. It is similar in most respects to well-known mesoscale models such as MM5 and WRF, but has been specifically adapted to wind resource modeling. MASS is run in a series of nested grids, with the innermost grid having a resolution of 1 km, an exceedingly high resolution for this type of model.

The microscale model, MSFD (Mixed Spectral Finite Difference), is a numerical model for atmospheric boundary-layer flow over complex terrain developed at York University in Canada. MSFD is a linear model with turbulence closure for neutrally stratified atmospheric surface flows. (The name derives from the fact that it uses a spectral technique to solve the momentum and energy equations in the horizontal dimension and a finite-difference technique in the vertical dimension.) MSFD has been validated against both wind tunnel measurements and field observations.

Accuracy

The accuracy of SiteWind has been verified against data and other models in several regions, including California, Saskatchewan, Quebec, Scotland, and New York. Two published case studies conducted by AWS Truewind illustrate the SiteWind advantage. (Papers can be found here.)

We compared SiteWind against WASP and measured wind speeds for 7 masts in the Cypress Hills area of Saskatchewan. In this study, the standard error in mean speed for SiteWind was 3%, whereas for WASP it was 12%. The masts with the smallest WASP errors were all located along the same bluff as the reference mast; the wind resource at stations away from this bluff was consistently overestimated.

We also compared SiteWind with WASP and measurements in a case study in Altamont Pass involving 132 met masts. Here the discrepancies between model and data are displayed as a function of distance from any reference mast. The discrepancy associated with WASP increases substantially with distance from a reference mast, whereas the discrepancy associated with SiteWind is virtually constant with distance. The estimated root mean square (rms) error for SiteWind was 5.7%, whereas that of WASP was 19%.

As these two examples illustrate, SiteWind is highly advantageous where turbines and met masts are sited in widely varying terrain (Saskatchewan), or where "mesoscale effects," induced for example by temperature differences, are significant (Altamont Pass).

Even where such conditions do not apply, SiteWind can greatly improve accuracy over micrositing models alone. For example, in a study we conducted in the Gaspe Peninsula of Quebec, with 8 high-quality met masts sited along ridgelines within a 10-km radius, the rms error for SiteWind was 3.7%, whereas the errors for WASP and WindSim were 6.8% and 5.7%, respectively. Compared to the other two models, SiteWind was 33 to 46% more accurate.