Evaluating the accuracy of RANS wind flow modeling and its impact on capacity factor for moderately complex forested terrain

The growth of onshore wind energy into the second largest renewable energy source has depended on overcoming many technological challenges. One of the main research priorities has been minimizing the uncertainty associated with wind energy yield calculations. The success of these calculations strongly depends on accurate wind resource assessment mainly done with wind speed measurements. Notwithstanding, a lack of measurement data justifies the use of computational modeling with promising results. But significant modeling challenges remain when analyzing turbulence over forested complex sites. These challenges are considered in this work with the main objective of evaluating the uncertainty in the wind flow predictions over moderately complex forested terrain and its impact on capacity factor, using the Reynolds-Averaged Navier-Stokes (RANS) equations coupled with a modified k-ε turbulence closure in the open-source software OpenFOAM v.2.4.0. \n \nWith the effects of complex topography implicitly captured in the RANS equations, the effects of the forest are explicitly calculated with two models: a displacement height model, and a canopy model that estimates the pressure loss due to the forest through analogy with porous media. To properly simulate the atmospheric boundary layer (ABL), the specific boundary conditions that rely on the law of the wall are implemented based on the recommendations of Richards and Hoxey, and Hargreaves and Wright. To validate the canopy model, the case of a fully-developed wind flow within and above a horizontally homogeneous black spruce forest is reproduced. Furthermore, two practical limitations are considered: 1) the physical foliage parameters may not be accessible for all type of forests; therefore, a generic leaf area density (α) distribution that is in agreement with the published results is tested; and 2) the published case limits its use to cyclic boundary conditions which are not practical for real site cases. Therefore, for cases without cyclic boundary conditions, two sensitivity analyses on the friction velocity u∗ and roughness length at the inlet z0inlet are tested. Different values of either of them give no significant difference in the vicinity of the forest, but they do at higher altitudes approaching the top boundary. This highlights the importance of imposing a proper fully-developed flow at the inlet condition. \n \nFour model cases are calculated for a site located in Quebec, Canada: A) terrain only, B) displacement height, C) canopy model with a uniform forest, and D) canopy model with the real forest distribution. The results are compared in terms of speed-up factors S (normalized velocities) with two years of measurement data from EDF-EN. Overall, the canopy model provides a better agreement with the mean statistical results than the other models. And where the terrain is densely forested, the assumption of a constant forest height delivers promising results. Finally, it is shown that the uncertainty in the energy calculation in terms of capacity factor CF is a non-linear function of the uncertainty in S. In this case, the 2.76% uncertainty in speed-up factor associated with the real forest distribution model leads to an uncertainty in the energy calculation of just 5.76%.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.