CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > OpenFOAM Running, Solving & CFD

Turbulent compressible and subsonic gas flow

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

Reply
 
LinkBack Thread Tools Display Modes
Old   April 21, 2005, 07:32
Default >What OpenFOAM version are you
  #21
Senior Member
 
Daniele Panara
Join Date: Mar 2009
Posts: 101
Rep Power: 8
panara is on a distinguished road
>What OpenFOAM version are you running?

1.1

>What solver?

rhoTurbFoam

>What thermodynamics package?

// Thermophysical model
thermoType hThermo<puremixture<consttransport<speciethermo<hc onstthermo<perfectgas>>>>>;

mixture air 1 28.9 1300 2.544e+06 1.84e-05 0.7;

>Have you tried changing the gradient?
>Does it run if the gradient is small enough?
>Have you tried changing the sign of the >gradient?

No I didn't try.
I will, and I'll let you know,

Thanks,

Daniele
panara is offline   Reply With Quote

Old   April 21, 2005, 07:35
Default I have checked the thermo libr
  #22
Senior Member
 
Join Date: Mar 2009
Posts: 854
Rep Power: 13
henry is on a distinguished road
I have checked the thermo library and the error in the thermodynamics only related to certain combustion thermo packages not the basic ones used by rhoTurbFoam.

Another useful check would be if you were to run with an isothermal wall and increase it's temperature and see what happens.
henry is offline   Reply With Quote

Old   May 19, 2005, 05:40
Default Dear all, I am using rhoTur
  #23
Senior Member
 
Daniele Panara
Join Date: Mar 2009
Posts: 101
Rep Power: 8
panara is on a distinguished road
Dear all,

I am using rhoTurbFoam (URANS) for simulating a heated air pipe flow with high amplitude pulsation of the flowrate.

Im using the following BC

Inlet: prescribed pulsating velocity, gradient of p = 0, temperature inlet constant.

Outlet: gradient of velocity = 0, gradiet p = 0, gradient T = 0

Wall: Twall = constant

I am comparing Low Reynolds and High Reynolds Turbulence models. I would like also to change the wall functions.

I have some questions:

1) rhoTurbFoam uses the turbulence models in src/turbulencemodels/compressible or incompressible ?
2) what is G in (compressible) wallDissipationI.H? is it the integrated contribution of epsilon in the near wall cell in the UMIST approach?
3) why the boundarySource of epsilon is multiplied by 1.0e+10?

Daniele
panara is offline   Reply With Quote

Old   May 19, 2005, 05:46
Default 1) src/turbulencemodels/compre
  #24
Senior Member
 
Join Date: Mar 2009
Posts: 854
Rep Power: 13
henry is on a distinguished road
1) src/turbulencemodels/compressible

2) G is the turbulent kinetic energy generation rate
(take a look at the sources in the k-equation)

3) In wall function approaches epsilon is usually fixed in the near-wall cell and multiplying the source by a large number and setting the central-coefficient to that number is a clunky but effective way of doing this.
henry is offline   Reply With Quote

Old   May 20, 2005, 03:32
Default Thanks for the usefull explana
  #25
Senior Member
 
Daniele Panara
Join Date: Mar 2009
Posts: 101
Rep Power: 8
panara is on a distinguished road
Thanks for the usefull explanation,

I have another question, I am experiencing problem in convergence when the amplitude of the pulsating flow in the channel is high, and I have reverse flow...

the Co goes up and then I get the error message as Jarrod.

Can be due to the BCs I am using?
Schall I use inletOutlet and OutletInlet BC?
If yes can somebody explain me how this BC works?
In the manual it is written that using those BCs two value should be given, inletValue and value, what are they?

Daniele
panara is offline   Reply With Quote

Old   May 20, 2005, 04:01
Default Do you have outflow at the inl
  #26
Senior Member
 
Join Date: Mar 2009
Posts: 854
Rep Power: 13
henry is on a distinguished road
Do you have outflow at the inlet or inflow at the outlet or both?

The inletOutlet BC simply changes the boundary condition for stability based on the direction of the flux, i.e. uses zero-gradient for outflow and the value specified to inflow.

outletInlet is the opposite which I created for pressure but it is unlikely to be appropriate in your case.

I do not know what inflow value you should specify, do you know what the inflow should be? If you want to limit the reverse flow try specifying a zero inlet value.
henry is offline   Reply With Quote

Old   May 20, 2005, 05:18
Default It is a pulsating channel, it
  #27
Senior Member
 
Daniele Panara
Join Date: Mar 2009
Posts: 101
Rep Power: 8
panara is on a distinguished road
It is a pulsating channel, it means that I have both conditions, the inlet at some point becomes an outlet and the outlet an inlet.

I prescribe an oscillating value of velocity at the inlet and a fix value of pressure at the outlet.. I am not sure that this BCs are consistent..

Do you have any suggestions?

Daniele
panara is offline   Reply With Quote

Old   May 20, 2005, 05:32
Default You can use inletOutlet for th
  #28
Senior Member
 
Join Date: Mar 2009
Posts: 854
Rep Power: 13
henry is on a distinguished road
You can use inletOutlet for the velocity and outletInlet for the pressure but you will have to adjust the inlet value of velocity according to the way in which you are driving the pulsations.
henry is offline   Reply With Quote

Old   May 20, 2005, 05:58
Default Do you think that the BCs I wa
  #29
Senior Member
 
Daniele Panara
Join Date: Mar 2009
Posts: 101
Rep Power: 8
panara is on a distinguished road
Do you think that the BCs I was giving are badposed or refining the grid and using a smaller time step can help?

Since I have already an oscillatingFixedValue BC, can I derive an inletOutletOscillatingFixedValue BC?

Can you give me some hint on how to do that?
where is the src of the inletOutlet BC ?

Thanks for your time,

Daniele
panara is offline   Reply With Quote

Old   May 20, 2005, 06:07
Default Having boundary conditions tha
  #30
Senior Member
 
Join Date: Mar 2009
Posts: 854
Rep Power: 13
henry is on a distinguished road
Having boundary conditions that change the bulk direction of the flow is likely to give problems and if you can think of ways to setup the case to avoid such complex and critical BCs it would be good idea.

You will find the implementation of the complex physical BCs in

OpenFOAM-1.1/src/OpenFOAM/fields/fvPatchFields/derivedFvPatchFields
henry is offline   Reply With Quote

Old   May 22, 2005, 23:37
Default Hi Guys I'm looking for the
  #31
Member
 
Wladimyr Mattos da Costa Dourado
Join Date: Mar 2009
Location: Sao Jose dos Campos, SP, Brazil
Posts: 36
Rep Power: 8
mattos is on a distinguished road
Hi Guys

I'm looking for the case root directory entry of rhoTurbFoam tutorials into foamX without success? Should I do something more to see the case root directory rhoTurbFoam tutorial dirs into the foamX trees? I guest that only if I can see this directory I should to edit and clone the case of rhoturbfoam class using foamX, is it?

Many tanks in advance

Wladimyr
mattos is offline   Reply With Quote

Old   May 23, 2005, 16:09
Default Hello Guys I have got the s
  #32
Member
 
Wladimyr Mattos da Costa Dourado
Join Date: Mar 2009
Location: Sao Jose dos Campos, SP, Brazil
Posts: 36
Rep Power: 8
mattos is on a distinguished road
Hello Guys

I have got the same problem reported by Jarrod at MArch/17 - item 1. It means, the solver crashes with the following message:--> FOAM FATAL ERROR : Maximum number of iterations exceeded
Function: specieThermo<thermo>:...

How can I solve this problem?

The test case is a NACA profile discretized in 3D. The spanwise is 3 mts and the cord is 1 m. The BC in the root and tips plane is simmetry (empty). The Mach is 0.7 and the farfield is divide in two parts: I defined the front farfield faces as inlet and the back farfield faces as outlet. Which are the best BC for farfield boundaries in such type of test case?

And what means R in the fields? Is the Reynolds Stresses? Using the K-epsilon model, have I impose some value or it is calculated from the flowfield?

Could someone help me?


Many tanks in advance!

Wladimyr Mattos Dourado
mattos is offline   Reply With Quote

Old   May 23, 2005, 16:12
Default Have you tried running incompr
  #33
Senior Member
 
Join Date: Mar 2009
Posts: 854
Rep Power: 13
henry is on a distinguished road
Have you tried running incompressible?

Yes R is the Reynolds stress and is not used unless you choose a Reynolds stress model.
henry is offline   Reply With Quote

Old   May 24, 2005, 12:26
Default Hi Weller I initialy tried
  #34
Member
 
Wladimyr Mattos da Costa Dourado
Join Date: Mar 2009
Location: Sao Jose dos Campos, SP, Brazil
Posts: 36
Rep Power: 8
mattos is on a distinguished road
Hi Weller

I initialy tried to run the compressible solver rhoturbfoam. The last message is results of this temptative. Just now I run the incompressible solver (turbfoam) and its work. Why your question about to try the incompressible? How solve that problem (crashing due thermophysical function)?

Second, I have the results of the incompresible solve. How can I transfer this field for the compressible case to use as initial conditions? The only way that I see now is copying the field files to the new comrpessible case first time directory. Have it another way to make this transfer using some foam tool ?

Really opem Foam is a nice CFD code. COngrutulations for you initiative and soo I hope to colaborate more actively to develop new features for foam.

Tank in advance for your help

Wladimyr
mattos is offline   Reply With Quote

Old   May 26, 2005, 13:36
Default Running compressible is alway
  #35
Senior Member
 
Join Date: Mar 2009
Posts: 854
Rep Power: 13
henry is on a distinguished road
Running compressible is alway more tricky than running incompressible: there is more to go wrong. You may find that starting from the incompressible solution helps your start-up problem and yes you can do this by copying the fields.

Are you running with the transonic option for the pressure? Given that your Ma is 0.7 you might find it helpful for stability.

You might also find it useful to start with a very small time step and crank it up as the solution progresses.
henry is offline   Reply With Quote

Old   May 27, 2005, 15:07
Default Hi Weller Tank very much fo
  #36
Member
 
Wladimyr Mattos da Costa Dourado
Join Date: Mar 2009
Location: Sao Jose dos Campos, SP, Brazil
Posts: 36
Rep Power: 8
mattos is on a distinguished road
Hi Weller

Tank very much for your help and advises. Your answer has grown some doubts in my mind such as: "How running with the transonic option for the pressure?". How is it turned on (or setup)?

I already try to begin the calculation with small time step. In fact, I use variable time set option and I impose Courant number between 0.15 and 0.3.

The copy of the incompressible solution field to compressible one is donne only using "cp" OS command directly, i.e., there are not a foam utility that do it, is it?

I already had success in the calculations of incompressible flow over this geometry (Airfoil). The air speed was set 10 m/s and the cord is 1 m long. I calculated using k-ep model and the boundary layer does not appear. The mesh is fine enough near the wall boundaries ( I hope!). Now I'm trying to guest what does not appear the boundary layer in such case!

Many tanks for your help!

Wladimyr
mattos is offline   Reply With Quote

Old   May 28, 2005, 05:41
Default transonic is an option in the
  #37
Senior Member
 
Join Date: Mar 2009
Posts: 854
Rep Power: 13
henry is on a distinguished road
transonic is an option in the PISO dictionary of fvSolution.

Why do you need an OpenFOAM utility reproducing the functionality of cp? If you like you could create an alias to change it's name or wrap it with a more complex alias or script if you would like to make it do more than the raw cp, it all depends on what you want from it.

What patch type have you given the airfoil?
What boundary conditions have you specified for p, U, k and epsilon?
henry is offline   Reply With Quote

Old   May 28, 2005, 16:23
Default Hi Weller Tank vary much fo
  #38
Member
 
Wladimyr Mattos da Costa Dourado
Join Date: Mar 2009
Location: Sao Jose dos Campos, SP, Brazil
Posts: 36
Rep Power: 8
mattos is on a distinguished road
Hi Weller

Tank vary much for your prompt answer.

Excuse for my stupid questions!

I have looked for some explicity transonic option in the PISO dictionary without success. And I nothing have found in the User's and Programmer's manual. Thus, is the fluxGradp switch that one for transonic flow? I haven't deep knowledge in the pressure corection methods such as PISO and SIMPLE. I come from the compressible density coupled methods schools!

About patch in the airfoil, it is defined as wall and the velocity is imposed equal to zero in the U field dictionary. All anothers are left with their default values.

Tank you very much again!
mattos is offline   Reply With Quote

Old   May 28, 2005, 17:02
Default The option is transSonic, set
  #39
Senior Member
 
Join Date: Mar 2009
Posts: 854
Rep Power: 13
henry is on a distinguished road
The option is transSonic, set it to yes:

transSonic yes;

You will also have to set the pressure solver type to BICCG rather than ICCG because the pressure matrix will become assymetric. You may also have to make some changed in fvSchemes but the code will tell you what entries are required.
henry is offline   Reply With Quote

Old   May 29, 2005, 21:57
Default Hi Weller Tank very much fo
  #40
Member
 
Wladimyr Mattos da Costa Dourado
Join Date: Mar 2009
Location: Sao Jose dos Campos, SP, Brazil
Posts: 36
Rep Power: 8
mattos is on a distinguished road
Hi Weller

Tank very much for your prompt answer and right advise. I had success in the airfoil simulations using your advise and correcting some mistakes in the boundary conditions definitions. I have two simmetry planes which I have defined in not right way as empty. Additionally, I have acquired experience in the incompressible test case about wall BC. The better way is initialise the calculation up some converge using wall after swap it for wall function BC. For transonic case, using rhoTurbFoam, I did sucess using adiabatic wall BC and late swaping to wall function BC after some convergence is achieved. In transonic case, I did use your sugesttion: switch on the transonic flag; However, the FoamX does not acept this variable and erases it in the solution dict. when we save the case in this GUI. The same problem happens in the solution dict. when we replaces BICCG in the rho and p solvers. How and where can us modify the sources such that this problem can be solved? I started the calculations of transonic flow with uniform distribution of the properties and Courant Num = 0.3 without problem! I did not need to begin with calculations from incompressible solution. Of coarse, I spent more time to find the final solution ;-|

You did commented something about fvSchemes needs be changed. Can you guide me where can I find the informations to do it?

Tank again and in advance for your help and advises!
Wladimyr
mattos is offline   Reply With Quote

Reply

Thread Tools
Display Modes

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 Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
Compressible Turbulent Flow CFDtoy Main CFD Forum 5 January 19, 2005 05:41
bench mark for subsonic compressible turbulent fl javadi Main CFD Forum 0 June 14, 2004 08:40
need tubulence compressible subsonic benchmark javadi Main CFD Forum 1 June 14, 2004 08:36
compressible subsonic flow Joel CD-adapco 2 April 24, 2003 08:18
Compressible turbulent flow FVS Main CFD Forum 0 April 13, 2002 17:07


All times are GMT -4. The time now is 20:17.