|
[Sponsors] |
![]() |
![]() |
#1 |
New Member
Patrick Reis
Join Date: Mar 2018
Posts: 4
Rep Power: 8 ![]() |
Hello dear Community,
this is my first thread so first of all I want to thank everybody contributing to this forum! You guys helped me a lot! So this is going to be a longer one but I still hope someone is willing to sacrifice some time. Now to my problem: I am trying to simulate a pipe flow through a expanding and afterwards narrowing geometry, a diffuser-pipe-nozzle configuration so to say. Both times with a 45deg angle and a inlet Re-number of 50000 (calculated with bulk velocity) and developed inlet conditions. I used different turbulence models which all showed similar wrong results so I will refer to the kOmegaSST version below. It is an incompressible case (water-nu=1e-6) and I solved for a steady state solution with simpleFoam. The simulation was 2D as a wedge. CheckMesh was satisfied with my mesh quality and the average yPlus value has been approx. 0.5. In my first configuration I chose a rather high area Ratio of the connected pipes of 16 (diam. 0.025m to 0.1m). The solution converged but some results were questionable. First of all I expected a static pressure rise after the diffuser along the axis which was hardly achieved (k-epselon even showed a drop) but only a huge pressure drop in the nozzle (see plots attached). The second issue was the eddy viscosity ratio nut/nu which seems to be unreasonable high - around 1200 (see attached picture). I expected to see some value well below 500. To verify my case setup I simulated the ERCOFTAC Case 75 (sudden Geometry expansion). Which is basically a diffuser with a 90deg half angle. The Re number was with 15600 lower but the flow type seemed to be similar. The results were very good compared to experimental data. And the eddy viscosity was approx. 180 (as expected by the cfd-online turbulence tool). In that magnitude i expected nut/nu to be for my initial case After that I simulated my case again with a lower Re number but the results didn't get better. My last try was to reduce the area-ratio of my case and adapted it to the one from the ERCOFTAC-case (much lower with 2.56). Inlet Re number was still 50000. And I got very much better Results. There was a higher pressure rise along the Diffuser (although much smaller area ratio) and the linear pressure drop in the big pipe indicates developed pipe flow (pressure was nearly constant along the big pipe in the first configuration). The eddy viscosity has a maximum of about 400 with what I could live but the contour plot may also suggest that there is much more convection of k and omega involved whereas in the first place there seems to be a strange eddyViscosity bubble. I attached a pressure and eddyViscosity plot as well. So this is basically why I assume that the first case is wrong whereas the lower area ratio leads to reasonable results with the same setup. I would like to find out why?! I would appreciate any help and suggestions! I can hardly believe that the area ratio is the factor that leads to such wrong results. I also tried out a lot different solver configurations and schemes, even one simulation with an unstructured mesh. But nothing gave me better results. I will upload my setup files and the complete case as well. I used second order linearUpwind schemes. Best wishes Patrick ...ahh yes and OpenFOAM 3.0.1 is my version. Initial-files: k: Code:
dimensions [0 2 -2 0 0 0 0]; internalField uniform 0.015; boundaryField { inlet { type fixedValue; value nonuniform List<scalar> 45 ( 0.00933555 0.00983949 0.0107606 0.0119197 0.0131983 0.0145152 0.0158171 0.0170712 0.0182588 0.0193703 0.0204024 0.0213552 0.0222311 0.0230335 0.0237665 0.0244342 0.0250407 0.0255896 0.0260845 0.0265282 0.0269234 0.0272721 0.0275758 0.0278355 0.0280514 0.0282229 0.0283486 0.0284258 0.0284504 0.0284162 0.0283142 0.0281319 0.027851 0.0274454 0.0268765 0.026085 0.0249641 0.0233483 0.0209806 0.0175078 0.0125525 0.00641524 0.0018269 0.000262463 0.000213873 ); } outlet { type inletOutlet; value $internalField; inletValue uniform 0; } walls { type fixedValue; value uniform 1e-12; } "(front|back)" { type wedge; } axis { type empty; } } Code:
dimensions [0 1 -1 0 0 0 0]; internalField uniform (2 0 0); boundaryField { inlet { type fixedValue; value nonuniform List<vector> 45 ( (2.3977 -4.38574e-06 4.10496e-17) (2.38947 -3.19018e-05 9.86728e-17) (2.37359 -4.41901e-05 1.87704e-16) (2.35264 -5.38954e-05 2.69189e-16) (2.32815 -5.72397e-05 3.28321e-16) (2.30123 -5.73149e-05 3.65412e-16) (2.27266 -5.50426e-05 4.26e-16) (2.24301 -5.16693e-05 4.9447e-16) (2.21269 -4.76964e-05 5.23782e-16) (2.18198 -4.35582e-05 5.33742e-16) (2.15107 -3.94613e-05 5.85707e-16) (2.12009 -3.55532e-05 6.15477e-16) (2.08912 -3.19225e-05 6.27487e-16) (2.0582 -2.86069e-05 6.48472e-16) (2.02735 -2.56122e-05 6.59231e-16) (1.99655 -2.2926e-05 6.67904e-16) (1.9658 -2.05268e-05 6.7409e-16) (1.93506 -1.83888e-05 6.78534e-16) (1.90429 -1.64855e-05 6.77416e-16) (1.87344 -1.47913e-05 6.77412e-16) (1.84244 -1.32824e-05 6.76851e-16) (1.81123 -1.19369e-05 6.73991e-16) (1.77972 -1.07356e-05 6.69179e-16) (1.74782 -9.66107e-06 6.64312e-16) (1.71542 -8.69819e-06 6.58646e-16) (1.6824 -7.83365e-06 6.51936e-16) (1.64861 -7.05578e-06 6.43942e-16) (1.61387 -6.35445e-06 6.35235e-16) (1.57798 -5.72102e-06 6.24993e-16) (1.54068 -5.14755e-06 6.1312e-16) (1.50167 -4.6266e-06 6.00062e-16) (1.46055 -4.15171e-06 5.85673e-16) (1.41684 -3.72524e-06 5.7007e-16) (1.36988 -3.35115e-06 5.52832e-16) (1.31881 -2.91031e-06 5.33624e-16) (1.26245 -2.03916e-06 5.11128e-16) (1.19891 -2.16759e-06 4.8554e-16) (1.12542 -2.34127e-06 4.566e-16) (1.0384 -2.59479e-06 4.21496e-16) (0.932893 -1.90032e-06 3.78671e-16) (0.802023 2.63587e-07 3.25737e-16) (0.638514 -5.14399e-07 2.58909e-16) (0.449021 -3.81458e-07 1.81614e-16) (0.25914 -1.19122e-07 1.04628e-16) (0.0837018 -7.53384e-09 3.37192e-17) ); } outlet { type inletOutlet; value $internalField; inletValue uniform (0 0 0); } walls { type fixedValue; value uniform (0 0 0); } "(front|back)" { type wedge; } axis { type empty; } } Code:
dimensions [0 0 -1 0 0 0 0]; internalField uniform 500; boundaryField { inlet { type fixedValue; value nonuniform List<scalar> 45 ( 97.309 101.697 110.073 121.108 133.957 148.115 163.337 179.571 196.894 215.471 235.524 257.313 281.134 307.316 336.229 368.294 403.998 443.904 488.674 539.093 596.098 660.819 734.631 819.218 916.67 1029.61 1161.34 1316.12 1499.49 1718.76 1983.76 2307.95 2710.23 3217.77 3870.79 4731.42 5899.09 7532.41 9892.61 13455.8 19243.6 30173.7 58660.6 174787 1.06099e+06 ); } outlet { type zeroGradient; } walls { type omegaWallFunction; value uniform 500; } "(front|back)" { type wedge; } axis { type empty; } } Code:
dimensions [0 2 -2 0 0 0 0]; internalField uniform 0; boundaryField { inlet { type zeroGradient; } outlet { type fixedValue; value uniform 0; } walls { type zeroGradient; } "(front|back)" { type wedge; } axis { type empty; } } Code:
dimensions [0 2 -1 0 0 0 0]; internalField uniform 0; boundaryField { inlet { type calculated; } outlet { type calculated; } walls { type calculated; value uniform 0; } "(front|back)" { type wedge; } axis { type empty; } } Code:
ddtSchemes { default steadyState; } gradSchemes { default cellLimited Gauss linear 1; grad(U) cellLimited Gauss linear 1; } divSchemes { default none; /* second order div(phi,U) bounded Gauss linearUpwind grad(U); div(phi,k) bounded Gauss linearUpwind default; div(phi,omega) bounded Gauss linearUpwind default; div((nuEff*dev2(T(grad(U))))) Gauss linear; } laplacianSchemes { default Gauss linear corrected; } interpolationSchemes { default linear; } snGradSchemes { default corrected; } wallDist { method meshWave; } Code:
solvers { p { solver GAMG; tolerance 1e-06; relTol 1e-06; smoother GaussSeidel; nPreSweeps 0; nPostSweeps 2; cacheAgglomeration on; agglomerator faceAreaPair; nCellsInCoarsestLevel 10; mergeLevels 1; } U { solver PBiCG; preconditioner DILU; tolerance 1e-08; relTol 0.0; } "(epsilon|R|omega|k|nut)" { solver smoothSolver; smoother GaussSeidel; tolerance 1e-06; relTol 0.0; nSweeps 1; } } SIMPLE { nNonOrthogonalCorrectors 0; //pRefValue 0; //pRefCell 1001; residualControl { p 0.5e-7; U 0.5e-7; "(k|epsilon|omega)" 1e-7; } } relaxationFactors { fields { p 0.3; } equations { "(k|omega|epsilon|U)" 0.7; } } |
|
![]() |
![]() |
![]() |
![]() |
#2 |
New Member
Patrick Reis
Join Date: Mar 2018
Posts: 4
Rep Power: 8 ![]() |
And here are the case files for the simulation with the "wrong"results
|
|
![]() |
![]() |
![]() |
|
|
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
mass flow in is not equal to mass flow out | saii | CFX | 12 | March 19, 2018 05:21 |
Multiphase flow - incorrect velocity on inlet | Mike_Tom | CFX | 6 | September 29, 2016 01:27 |
reversed flow at pressure inlet and turbulent viscosity is limited.... | cfdiscool | FLUENT | 10 | June 10, 2015 06:15 |
Link between eddy viscosity and subgrid-scale turbulent viscosity in LES | Marvin | Main CFD Forum | 6 | December 4, 2009 11:08 |
length scales at inlet for internal flows | Anne-Marie Giroux | Main CFD Forum | 3 | July 5, 1999 21:28 |