CFD Online Logo CFD Online URL
Home > Forums > FLUENT

If you need any help

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

LinkBack Thread Tools Display Modes
Old   January 13, 2003, 08:17
Default If you need any help
Posts: n/a
Hi CFD researchers,

Finally its over !!

I just submitted a thesis for a PhD.

I have 5 years experience in Fluent and Gambit.

Multiphase flows (VOF), particulate flows, turbulence dispersion, coating, pipe flow are my expertise.

I will be starting a job in six months time, and I don't have nothing to do !!

If you have any difficult problems in these area, please do not hesitate to contact. I would like to gain more experience.

You can also add me if you have yahoo and/or hotmail messenger:

Then we can chat online.

Regards, phd_y2k2001
  Reply With Quote

Old   January 13, 2003, 13:20
Default Re: PhD research
Matthew Roberts
Posts: n/a
Dear y2k2001(?)

I am currently in my third year PhD and have run into a few problems regarding the cfd side to my research studying open-channel meandering rivers. Unfortunately there are very few 3D FLUENT users at the university and the help from fluent support has been ineffective. I would greatly appreciate some help/advice on the task at hand. I will attach my mesh journal file (on another e-mail) and provide a project description below.

I am trying to simulate the 3D flow field and water surface elevation within a meandering reach of a river. So far I have built a scaled model in the laboratory, during my second year, and collected 3D velocity data and water surface elevation data at several cross-sections for validation purposes. I have also generated several meshes of varying grid size resolution in GAMBIT representing the channel geometry, however I am unable to accurately predict the free surface of the flow using the VOF multiphase model.

Project Description: The river channel is rectangular in cross-section, 30cm wide and 25cm deep with smooth channel bed and side walls. The channel reach consists of three meander bends of constant radius of curvature (0.75m to the channel centreline). The average flow depth is 10cm and the flow is turbulent.

In GAMBIT I produced a mesh that is 12cm in height (approx 10cm for the water flow and 2cm for the air above), and split the inlet face along the free surface between the water and the air (as measured in the experiments) to create two VELOCITY INLET's, inlet1_air and inlet1_water. I did this in GAMBIT because the free surface isn't uniform across the channel and I thought it simpler to specify a volume fraction of 1 for the water inlet and 0 for the air inlet(phase1 air, phase 2 water), rather than try to split the mesh and specify vol fractions in FLUENT? The outlet face was not split into separate phases as I want the model to determine this and I assigned the boundary condition OUTFLOW. The top surface (at 12cm) of the mesh is given the boundary condition SYMMETRY, I was advised to do this to represent an "open" channel???, however I think that maybe some sort of pressure inlet to specify atmospheric pressure is needed?

FLUENT setup

Read Case file Check grid Scale grid: grid created in cm, Scale Define Model: Multiphase, VOF (default) Define Model: Viscous, k-epsilon (default) Define Phases: Phase 1 air, Phase 2 water Define Operating Conditions: Gravity ON, z component -9.81. Specified Operating Density 1.225 (air) and Reference Pressure Location (0.3441, 0.4915,0.12) at the inlet surface right bank so that the closest cell centre is air.

(As an initial test I just applied constant velocities normal to the inlet boundary, rather than using the experimental data)

Boundary Conditions: Inlet1_wat

Phase Mixture: Magnitude, normal to the boundary Velocity magnitude 0.4m/s, Constant

Phase Water: Volume Fraction 1


Phase Mixture: Magnitude, normal to the boundary Velocity magnitude 0.002m/s, constant

Phase Water: Volume fraction 0 All other BC's default

Solution Control: Default (if I set the pressure to body force weighted, the pressure-velocity coupling to PISO and change all the under relaxation factors to 1, as the user manual recommends, I get divergence!!)

Solution Initialization: Default. (I would like to initialize the entire flow field with a more realistic flow of say 0.4m/s in the downstream direction parallel to the side channel walls, however the options only include initial guesses for the x,y,z velocities, which vary dramatically as you proceed around the bends!!!)

Monitor residuals: plot and print

Iterate: Time step (0.5s)

No. of time steps 500

Max Iterations per time step 20

After the solution had converged (approx 450 time steps) just based on the residuals, I then looked at the contours (levels=2) of volume fractions of water. The plot shows that virtually the entire mesh is filled with water?

If you can spot any errors in my procedure please let me know. Also would I have more success in modelling this type of free surface flow using FIDAP, or should FLUENT suffice?

Any comments will be gratefully received, Best Regards

Matthew Roberts Rm 9.10a Schools of Mathematics/Earth Sciences University of Leeds LS2 9JT (0113) 2335203

P.S Sorry about the long e-mail, I just wanted to make sure that I explained the situation clearly. Cheers M.

  Reply With Quote

Old   January 14, 2003, 02:31
Default Re: PhD research
Posts: n/a
Every thing is ok, but look on the volume fraction,both said water fraction 1 or 0, this could be problem..... another problem could be gravity... u set z direction, it could be y direction -9.81 depending on your geometry...


  Reply With Quote

Old   January 14, 2003, 12:40
Default Re: If you need any help
Adriana Catanase
Posts: n/a
I'm prepearing my doctoral thesis about the dynamics of Pelton turbines. I understand that with Fluent you could simulate the flow in the injector of this type of turbines.
  Reply With Quote

Old   January 21, 2003, 20:20
Default Re: PhD research
Posts: n/a
Hi Matthew,

I think you miss one step: Define a register for the bottom part of the domain by Adapt -> Region... -> Mark. It patchs the initial distribution of water in the bottom part of the channel.

Hope it helps,

-- Jean
  Reply With Quote

Old   January 22, 2003, 07:21
Default Re: If you need any help
Posts: n/a

Thank you very much for your support, I have a problem, hope you can help me in this regard. The technical support from fluent has been ineffective.

Lets take an example of a Pipe in Pipe structure in which Hot Water at nearly 80c flows through the Inner pipe and Air at room temperature flows through the Outer pipe. Assume that both the fluids flow in opposite directions.

My interest is to model the conjugated heat transfer method inorder to analyse the heat transfer from hot water ------> cold air, assuming the pipe material offers virtually no resistance to the heat transfer.(therefore no need to resolve the heat transfer via conduction through the pipe wall).

For this problem, I have created 3d mesh for a Pipe in Pipe structure in Gambit with two different fluid zones seperated by a thin circular wall. Assuming the Inner pipe consists of a special cylindrical wall surface structure which repeats many number of times along the length of the pipe, hence the question to use translational periodicity boundary conditions arises, as the length of pipe is assumed to be very long. But, in FLUENT, once after I make my boundaries translational periodic, I cannot be able to input the flow conditions (my flow conditions are assumed to be fully developed turbulent) for each fluidzone.

So, now my question, Is it really possible to use translational periodic boundary conditions in combination with conjugated heat transfer model in FLUENT.

or else, Is there any better approach to model the kind of problem I have explained above.

Please let me know, if you need any further description of the problem. Thank you very much for your co-operation.

  Reply With Quote

Old   January 22, 2003, 08:47
Default Re: If you need any help
Posts: n/a
Please use an email address and send me a doc file with a picture of what are you trying to say.

I don't want to mislead you. Based on my understanding:

Regarding your dilema, use 3 zones first zone as normal input, the output of the zone at the contact surface would be used as an input of the 2nd zone where you will use perdioc BC, in the 3rd zone there be (like the first zone) you may have an exit. It will take time though to make it work. I have used this apraoch, it works.
  Reply With Quote

Old   January 24, 2003, 07:02
Default Re: If you need any help
Posts: n/a
I am modelling the laminar flow in the curved duct of rectangular cross section (a x b). How can I change the hydraulic diameter in the system of equations by the channel width (a). Many thanks

  Reply With Quote


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

All times are GMT -4. The time now is 22:26.