Em 14 Set 2021, foi publicado online no períodico Journal of Verification, Validation and Uncertainty Quantification, da American Society of Mechanical Engineers (ASME) dos Estados Unidos, o artigo "Lid-driven square cavity flow: a benchmark solution with an 8192x8192 grid".
O artigo é de autoria de Carlos H. Marchi, Cosmo D. Santiago e Carlos A. R. de Carvalho Jr., todos membros do
Grupo de Pesquisa em CFD, propulsão e aerodinâmica de foguetes da UFPR.
Com este artigo
quebramos dois recordes mundiais para o problema clássico de CFD da cavidade quadrada com tampa móvel, que foi criado há 60 anos:
1) A maior malha (8192x8192 pontos) já usada até Re = 10.000.
2) A maior acurácia e precisão dos resultados. Em geral, o erro estimado dos resultados é de 100 a 1000 vezes menor que resultados anteriores da literatura.
AbstractThe incompressible steady-state fluid flow inside a lid-driven square cavity was simulated using the mass conservation and Navier–Stokes equations. This system of equations is solved for Reynolds numbers of up to 10,000 to the accuracy of the computational machine round-off error. The computational model used was the second-order accurate finite volume (FV) method. A stable solution is obtained using the iterative multigrid methodology with 8192 x 8192 volumes, while degree-10 interpolation and Richardson extrapolation were used to reduce the discretization error. The solution vector comprised five entries of velocities, pressure, and location. For comparison purposes, 65 different variables of interest were chosen, such as velocity profile, its extremum values and location,
and extremum values and location of the stream function. The discretization error for each variable of interest was estimated using two types of estimators and their apparent order of accuracy. The variations of the 11 selected variables are shown across 38 Reynolds number values between 0.0001 and 10,000. In this study, we provide a more accurate determination of the Reynolds number value at which the upper secondary vortex appears. The results of this study were compared with those of several other studies in the literature. The current solution methodology was observed to produce the most accurate solution till date for a wide range of Reynolds numbers.
Siglas:
CFD é a abreviação de Computational Fluid Dynamics ou dinâmica dos fluidos computacional.
Re é o número de Reynolds.
UFPR é Universidade Federal do Paraná.
Geometria do problema da cavidade com tampa móvel,
dimensões e condições de contorno.
Campo da função de corrente para Re = 10.000