Numerical enhancement of a mesoscale model for large-eddy simulation of the wind over steep terrain

Mesoscale modelling of the atmospheric boundary layer has advanced significantly over the past decades, although there are still different numerical aspects that must be enhanced to achieve accurate wind simulations over steep topography. This has become a necessity since many applications, such as wind resource assessment, now require high fidelity results for viability analysis and decision-making. With the advent of high performance computing and more sophisticated software, the wind energy industry is increasingly interested in multiscale models based on combined configurations capable of yielding higher resolution results. The size of the modern wind farms now requires a multiscale analysis that allows the evaluation of the joint meso- and microscale processes triggered over complex topography. For this reason, mesoscale models with imbedded large-eddy simulation capabilities are well suited to become the next mainstream family of simulation toolkits for wind engineering. The Mesoscale Compressible Community (MC2) model, subject of this work, is a good example since it is employed as the kernel of the Wind Energy Simulation Toolkit (WEST), introduced by the Recherche en Prevision Numerique (RPN) group of Environment Canada. MC2 performs well for wind simulations over flat, gentle and moderate terrain slopes, which led the wind energy community to be confident enough on employing it to generate the Canadian Wind Atlas. However, as with other similar models, several numerical issues such as wind overestimation and distorted circulation patterns have been identified in recent years from orographic flow simulations in presence of steep slopes. Hence, wind resource assessment over high impact topography, such as the Rocky Mountains or the Niagara Escarpment, cannot be entirely reliable and needs a revaluation with enhanced multiscale modelling. By applying an eigenmode analysis, we have recognized the numerical instability and precisely measured the spurious noise problem, inherent of MC2’s classical three time-level semi-implicit (SI) scheme. With the appropriate redefinition of the prognostic thermodynamic variables, the SI time discretization, coupled with the semi-Lagrangian (SL) scheme, is now consistently structured in a way that it enables MC2 to solve the compressible non-hydrostatic Euler equations (EE) in a more stable and accurate fashion. MC2 is now able to perform wind simulations over steep slopes in the absence of time decentering, frequency filtering and other numerical damping mechanisms. Additionally, the climate-state classification of the statistical-dynamical downscaling (SDD) method is upgraded by including the Brunt-Vaisala frequency that accounts for the atmospheric thermal stratification effect on wind flow over topography. The present study provides a real orographic flow validation of these numerical enhancements in MC2, assessing their individual and combined contribution for an improved initialization and calculation of the surface wind in presence of high-impact terrain. Lastly, the metric tensor adaptation of MC2’s imbedded large-eddy simulation (LES) method, necessary for wind modelling over mountainous terrain, has been achieved preserving the enhanced numerical stability and accuracy. Test results indicate that the enhanced MC2-LES model reproduces efficiently the expected flow patterns, separation and recirculation zone over steep terrain, and yields accurate results comparable to those reported from experimental data or by other researchers who use numerical models with similar or more sophisticated turbulence closure schemes.

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.