Comparison of the new wind flow model FUROW with WAsP and OpenFoam for different topographic features

Introduction

SOLUTE has developed a new software for wind data analysis, wind resource assessment, energy production estimation and class verification called FUROW. This software allows calculating wind speed and turbulence maps at multiple heights based on wind measurements and taking into account topographic effects and thermal stability of the site.

Approach

The wind flow model used by FUROW is based on the old UPMORO code, first developed by the Research Group “Mecánica de Fluidos aplicada a la Ingeniería Industrial” who belongs to Universidad Politécnica de Madrid - Escuela Técnica Superior de Ingenieros Industriales (UPM-ETSII). During FUROW development, the code has been optimized and improved from its original version, including significant modifications on the physical formulation. This model estimates the effects of orography and roughness on wind speed and wind direction at any particular point and height and its horizontal extrapolation is perform by means of the geostrophic wind and assuming its invariance for the whole area. Wind profile used for simulations takes into account thermal stability through the Monin-Obukhov length.

Main body of abstract

FUROW wind flow model has been compared with commercial software such as WAsP (linear model) and OpenFoam (CFD model) in order to characterize its behavior under different orographic and roughness conditions. For this purpose wind speed has been simulated over Gaussian shaped hills with average slopes from 10% up to 50%, that is, from smooth orography with attached flow up to large slopes typical of very complex terrain where flow is detached and important recirculations appear. Moreover, roughness length has been modified from 0.03 to 0.4 in order to simulate different land cover types.

Wind speed vertical profile has been initially calibrated on flat terrain in such way that Monin-Obukhov length can be adjusted in FUROW to represent the same default conditions as in WAsP (typically neutral or slightly stable conditions with ΔHoff=-40W/m2 and ΔHrms=100W/m2 for onshore conditions). Additionally, three wind speed values have been run (5m/s, 10m/s and 15m/s at 80m high) to cover the typical range of observed wind speed distributions. The calculated vertical profile for each simulation has been introduced numerically in OpenFoam to carry out the simulation with the same initial conditions.


Conclusion

Simulations results have been compared on different reference points along a horizontal transect evaluating the difference on the vertical profile obtained by FUROW, WAsP and OpenFoam. Simulations with OpenFoam have been run using a structured mesh with the same spatial resolution as on the linear models (25m), and several turbulent models based on k-ε model with different coefficients widely proposed on state of the art literature.

Results derived from different simulations indicate that FUROW reproduces the same vertical wind profile as that obtained with WAsP along the whole horizontal transect, being the differences between 0% and 8% depending on the reference point and orographic complexity. FUROW tends to slightly underestimate WAsP, especially at the hilltop. However, when results are compared with OpenFoam, it can be observed how FUROW yields more similar results than those obtained with WAsP, in particular for moderate slopes (up to 20%), where it seems that WAsP tends to overestimate OpenFoam. For more complex orography, the differences between FUROW and WAsP with OpenFoam are similar, and none of them capture recirculations downwind of the hill.


Learning objectives
As a conclusion of this work, it can be stated that FUROW model simulates correctly the wind flow at sites with different topographic features, with results comparable to those of CFD type such as OpenFoam, but with the advantage of needing less computation time. Likewise, results are comparable with those of WAsP, allowing in FUROW the possibility of calculating multiple heights simultaneously. It is also remarkable that thermal stability (and implicitly wind shear) could be easily modified by using a well-known variable such as Monin-Obukhov length.