# Compared MRFSimpleFoam and Fluent in a centrifugal pump!

 Register Blogs Members List Search Today's Posts Mark Forums Read

 December 11, 2009, 02:59 Compared MRFSimpleFoam and Fluent in a centrifugal pump! #1 Member   任芸 Join Date: Jun 2009 Posts: 75 Rep Power: 9 Hi all, Recently,I am calculating the hydraulic performance and internal field of a centrifugal pump by using OpenFoam-1.5-dev and Fluent.now,I want to recommend my steps.firstly,I generated the mesh in Gambit,save as ***.msh for Fluent and MRFSimpleFoam,and the numerical method i dopted both in Fluent and MRFSimpleFoam are: standard k-epsilon model, simple algorithm and first order upwind. Additionally,after setting the initial conditions and boundary conditions and the MRF in OF,I modified the discrete format and the under-relaxation factors as followed: divSchemes { default none; div(phi,U) Gauss upwind; div(phi,k) Gauss upwind; div(phi,epsilon) Gauss upwind; div((nuEff*dev(grad(U).T()))) Gauss linear; } fvsolution: SIMPLE { nNonOrthogonalCorrectors 0; pRefCell 0; pRefValue 0; } relaxationFactors { p 0.5; U 0.5; k 0.25; epsilon 0.25; } Compared the result in Fluent and MRFSimpleFoam: ① the residual: Fluent: http://www.cfd-online.com/Forums/mem..._residuals.jpg OF: http://www.cfd-online.com/Forums/mem...3-residual.jpg ② P and U: Fluent: http://www.cfd-online.com/Forums/mem...nt_p_total.jpg http://www.cfd-online.com/Forums/mem...v_absolute.jpg OF: http://www.cfd-online.com/Forums/mem...cture105-p.jpg http://www.cfd-online.com/Forums/mem...cture101-u.jpg http://www.cfd-online.com/Forums/mem...2-vector-u.jpg http://www.cfd-online.com/Forums/mem...r-velocity.jpg ③ the hesad and torque: fluent: head=28.6m, the torque=161.52 OF: head=26.1m,the torque are as followed: norm of(blade) : (3.35282e-09 3.39945e-09 0.000766278) pressure torque(blade) : (0 0 0)[Nm]; power: 0[W] Evaluation of GGI weighting factors: Largest slave weighting factor correction : 0.000544509 average: 0.000108346 Largest master weighting factor correction: 0.00247013 average: 2.03543e-05 viscous torque (blade) : (0 0 0)[Nm];power 0[W] norm of(wallqgb) : (-2.0884e-09 -7.9866e-10 -0.0519531) pressure torque(wallqgb) : (0 0 0)[Nm]; power: 0[W] viscous torque (wallqgb) : (0 0 0)[Nm];power 0[W] norm of(wallhgb) : (-1.26442e-09 -2.60051e-09 0.0712859) pressure torque(wallhgb) : (0 0 0)[Nm]; power: 0[W] viscous torque (wallhgb) : (0 0 0)[Nm];power 0[W] at last, my problems are: 1.strangely,when I calculated the impeller without the volute,and the head of impeller is only 20m,why it went up after adding the volute? 2.the p in OF is static pressure,dynamic or total pressure? While which is total pressure in Fluent. 3.Is the residual in OF above convergenced?and how to determine whether the result is convergenced? 4. or is there any things will cause the differents?whether is wrong or right in my setup? Regards, Yours jennifer Last edited by renyun0511; December 11, 2009 at 04:19.

 December 11, 2009, 06:22 #2 Member   Etienne Lorriaux Join Date: Mar 2009 Location: Compiegne, France Posts: 43 Rep Power: 9 Hello Jennifer. For steady-state incompressible flows, OF computes (static pressure)/rho. There is a 'ptot' tool in OF to compute the total pressure, I'm not sure but I think it's computing (total pressure)/rho, so you will surely have to rebuild rho * ptot in paraview. There are several points missing in your analysis, such as continuity residuals, turbulent inlet/outlet conditions, it has a great influence on convergence and results. It seems you are using tets (prisms?) in your mesh, it requires a particular attention concerning the convergence of p and continuity. I've already performed quite a lot of comparisons between Fluent and OF on steady-state incompressible flows, with excellent results in really close agreements (OF is even a little less diffusive than Fluent). Regards, Etienne.

 December 12, 2009, 08:35 #3 Member   任芸 Join Date: Jun 2009 Posts: 75 Rep Power: 9 hi Etienne, Thank you for your reply,it’s my negligence to make it clearlier. I used tetrahedron in my mesh. under the 0/U:I give a velocity for inlet ,and give shloud, hub and blades for fixedValue (0 0 0),interface patch for ggi, outlet for zeroGradient;the 0/p: outlet for fixedValue uniform 0;and interface type for ggi,and the other patch type for zeroGradient. Κ=0.07.ε=0.29for inlet,and I calculate them by the equantion: My problems are here: 1. how to give a particular attention concerning the convergence of p and continuity? 2. I use ‘calcPressureDifference’ utility to calculalte the Hydraulic head of pump.the differece of Pressure=InletPressure-OutletPressure =rho*inletPressure-rho*outletPressure so before I use the ‘ptot’ utility,the head I calculate is wrong because of the pressure I used is static pressure,is it right? regards yours jenifer

 July 1, 2010, 07:24 #4 New Member   Sunny William Join Date: Jun 2010 Location: London Posts: 6 Rep Power: 8 Hello, I am currently working on a 3d impeller/diffuser stage case and have one question. How did you compute the pump head and efficiency in OpenFOAM? Many thanks San

 July 1, 2010, 21:59 #5 Member   任芸 Join Date: Jun 2009 Posts: 75 Rep Power: 9 there are two adds-on software which are used for computing rotating machines.here you are: Attachment 3972 another is torque computation software,but i'm afraid it is too large enough to upload,because it displays" 708.7 KB bytes exceeds the forum's limit of 97.7 KB for this filetype".So,would you like to give me your e-mail ,please? i will give you as soon sa possible! good luck! your jennifer blake likes this. Last edited by renyun0511; May 11, 2013 at 11:08.

July 2, 2010, 02:47
#6
Senior Member

Alberto Passalacqua
Join Date: Mar 2009
Location: Ames, Iowa, United States
Posts: 1,910
Rep Power: 27
Quote:
 Originally Posted by renyun0511 3.Is the residual in OF above convergenced?and how to determine whether the result is convergenced?
Your residuals plots show the slope of the residuals is not zero in both FLUENT and OpenFOAM, so the solution is still changing. Strictly speaking, for a steady state simulation, you might want to compare the results you obtain when residuals stop lowering (the residual plots become flat).
__________________
Alberto Passalacqua

GeekoCFD - A free distribution based on openSUSE 64 bit with CFD tools, including OpenFOAM. Available as in both physical and virtual formats.
OpenQBMM - An open-source implementation of quadrature-based moment methods.

To obtain more accurate answers, please specify the version of OpenFOAM you are using.

 July 2, 2010, 09:51 #7 New Member   Sunny William Join Date: Jun 2010 Location: London Posts: 6 Rep Power: 8 Hi Jennifer, Thanks for your quick reply, I will send my email address now. Also I have been trying to get my case to convergence but without any luck. Could you kindly take a quick look at my case settings also? I'd really appreciate it. Regards, San

July 3, 2010, 23:30
#8
Member

Join Date: Jun 2009
Posts: 75
Rep Power: 9
Quote:
 Originally Posted by Santana Hi Jennifer, Thanks for your quick reply, I will send my email address now. Also I have been trying to get my case to convergence but without any luck. Could you kindly take a quick look at my case settings also? I'd really appreciate it. Regards, San
I'd like to!

July 6, 2010, 06:24
#9
New Member

Sunny William
Join Date: Jun 2010
Location: London
Posts: 6
Rep Power: 8
Thanks for kindly accepting.

Here is my case:

- I use MRFSimpleFoam with ggi interface
- The periodicity of the impeller/diffuser is different (6 and 7 passages, repectively) and I understand the limitations of running a frozen rotor case. However my aim here is to just confirm that convergence can be reached.
- I have not impletemented any turbulence right now.

Problem:

I cannot seem to reach convergence further than shown in the plot below. I have played around with different schemes, solver algorithms and BCs with no luck.

I have also tried a case with very slow impeller rotation (designed is 1000 rpm but tried 10 rpm) and then the solution converges with physically feasible solution - Thus I am also suspecting the possibility of it to be a more fundamental problem with the rotational reference frame, ggi interface, etc. Though have not yet found any solution.

Here are some of the visualization and the residual plot:

c1.JPG

c5.JPG

res.JPG

c7.jpg

Schemes:
Quote:
 gradSchemes { default Gauss linear; grad(p) Gauss linear; grad(U) Gauss linear; } divSchemes { default none; div(phi,U) Gauss upwind; div(phi,k) Gauss upwind; div(phi,omega) Gauss upwind; div(phi,R) Gauss upwind; div((nuEff*dev(grad(U).T()))) Gauss linear; } laplacianSchemes { default Gauss linear corrected; laplacian(nuEff,U) Gauss linear corrected; }
Solvers:
Quote:
 p GAMG { tolerance 1.0e-6; relTol 1.e-3; smoother DIC;//GaussSeidel; nPreSweeps 0; nPostSweeps 2; cacheAgglomeration true; nCellsInCoarsestLevel 800; agglomerator faceAreaPair; mergeLevels 1; maxIter 30; }; U PBiCG { preconditioner { type DILU;} smoother { type DILU;} minIter 1; maxIter 4; tolerance 1e-07; relTol 0; }; k PBiCG { preconditioner { type DILU;} smoother { type DILU;} minIter 1; maxIter 3; tolerance 1e-07; relTol 0; }; omega PBiCG { preconditioner { type DILU;} smoother { type DILU;} minIter 1; maxIter 3; tolerance 1e-07; relTol 0; };
BC Pressure:
Quote:
 R1-INLET { type rotatingTotalPressure; U U; phi phi; rho none; psi none; gamma 1.4; p0 uniform 101.3; value uniform 101.3; omega (0 0 104.72); } ggi-imp { type ggi; value uniform 110; } ggi-dif { type ggi; value uniform 110; } S1-OUTLET { type fixedMeanValue; meanValue 230; value uniform 230; }
BC velocity:
Quote:
 R1-INLET { type zeroGradient; } S1-OUTLET { type inletOutlet; inletValue uniform (0 0 0); value uniform (0 0 0); }
Any suggestions and help are welcome and appreciated.

Many thanks,

San

 Thread Tools Display Modes Linear Mode

 Posting Rules You may not post new threads You may not post replies You may not post attachments You may not edit your posts BB code is On Smilies are On [IMG] code is On HTML code is OffTrackbacks are On Pingbacks are On Refbacks are On Forum Rules

All times are GMT -4. The time now is 12:58.