CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   Quenching simulation - wall boiling model (

Michael.J March 11, 2010 08:35

Quenching simulation - wall boiling model

I am trying to simulate quenching of steel ball in water with phase change of quenching liquid in order to obtain heat transfer coefficient.

Geometry consists of steel ball (5mm in diamater) in cylindric tank, in which bottom surface is inlet and top is opening. Water enters with speed of 10 mm/s and temperature of 50 C. Initial temperature of steel ball is 950 C.

The simulation is transient. I use two phase model: one is water as continus phase and second is vapour as dispersed phase with initial volume fractions acordingly 1 and 0. I've setup SST turbulence model, Thermal energy model for each phase, bouyancy and phase change through saturation temp of 100 C. I've also setup drag model as Grace.

I'd like to be able to recreate film creation and also bubble detachment and their strong influence on convection when they detach (basicly normal boiling).

So far with this setup it goes quite well, however the calculation will take quite some time. What i am afraid of is the lack of bubble creation, and just simple phase change into one film around quenched ball. If I suspect correctly this setup wont give me the proper results and through whole simulation no bubbles wont be created but only constant vapour film which size will depend on temperature gradient.

Thus i wanted to use wall boiling model (RPI). The problem is that CFX crashes every time I turn it on. I've even went down with time step to 1e-6 s, but it still crashed with good RMSes and very low Courant numbers after around 10 terations.

If anyone has any sugestions, i'll appriciate them :) I've started also considering to change the vapour phase into continus or particle transport fluid. Could anyone with experince say his opinion on this matter?


ghorrocks March 12, 2010 00:51

This is a very challenging simulation. Don't expect any easy answers!

Are you directly modelling the steel balls or you using a eularian or langrangian approach for the balls? Are there lots of balls or just a few?

Michael.J March 12, 2010 08:25

Thanks for reply.

This is cylindric tank 20 mm in diamater, length 30 mm and there is one ball in middle of it with 5 mm in diameter.

EDIT (forgot to add this): ball is part of model and there is domain interface betwen tank and ball with conservative heat transfer boundary condition.

So far I used Eulerian model, with one continous phase and other dispersed. From what I understand Eulerian model doesnt give specific volume to dispersed phase and it concernes me actualy. I'd like to use Lagrangian approach but the question is will it be suitable to use for thermal phase change from liquid phase to gaseous? From what i've seen/read it's used mostly for simulation of already existing liquid droplets and their evaporation. But will it be suitable for reversed situation, when in the begining there is no gas phase and they come into existense due to phase transformation. Will it be also capable to reacreate thin vapour film around object ?

And the most important question is why Wall Boiling Model available for Eulerian approach crashes all the time? Is it only applicable for wall with steady T, H. Flux or heat transfer coefficient? Cause from what I've read this is exactly THE THING i want to use and so far after few iterations while using this model CFX crashes (without any specific error - I've encountered this situation before when i had very high RMSes, but while using this model I went down with timestep to 1e-6 s, and the RMSes were fine).

EDIT2: One more question: While using Eulerian, setting the water vapour phase as continus has any sense at all? So there are only continus phases: water and water vapour.


Felja August 18, 2013 13:24

Quenching simulation - wall boiling model
I have a very similar problem!
Can we share experiences?
If you agree, my e-mail is:

RicochetJ August 18, 2013 14:19

Why don't you just post and share your experiences here for all to read?

Felja August 19, 2013 09:00

RPI-Boiling model
I agree with your comment!

Problem description:
Vertically the cylinder speed of 25 mm / s, size (75x225) mm, the initial temperature of 850 C, the water temperature of 40 C, which bottom-up flow speed of 0.25 m / s!
During the immersion cylinder in water, coolant temperature kept constant (subsequent cooling).
My ------- *. Cas fille is:
axi - axisymmetrical
dp - double precision - Looking Eulerian model
pbns - pressure based
dynamesh - I got it moving (dynamic mesh layering only)
Eulerian multiphase (RPI - boiling model) - I do not have an example of the settings for this model (Help is not enough, because I used it)
And the most important question is why Wall Boiling Model available for Eulerian approach crashes all the time? Is it only applicable for wall with rising T, H. Flux or heat transfer coefficient?

ske - standard k-e model (including viscous heating and drift force and per phase)
Finally, I want to get Heat transfer coeficient as a function of time and temperature!
-------------------------------------------------- ------------------------------
The problem is complex!

Most suits me contact with Michael Jasiczek!
Thanks to everyone who is involved in this problem!


orlijda August 25, 2013 08:47

Hi, I'm new in this forum and I'm starting with the Ansys Fluent.

Could someone tell me how to set the initial temperature at the initial conditions of a domain in fluent?

I'm doing the cooling of a piece of steel with a wáter flow.

Thanks in advance.

Felja August 25, 2013 15:13

Is your problem is steady or unsteady?
Does water flow around or idle?

orlijda August 25, 2013 18:16

1 Attachment(s)
Hi, I'm trying to arround piece and also idle.

I donīt know how to set the initial steel temperatura. When the mesh interface is configured at fluent, four boundaries are created by software.

Which boundary should be set as hot steel and where?

Attached you can find a picture.

Thanks in advance.:)

Felja August 26, 2013 03:19

Dear Orlijda,
Best you do ..... ini.cas file. for example, at the initial time t = 0.01 seconds with a fixed value for steel, such as 850 C. Make a convergent solution! After that, turn off the fixed value and continue with that file still counting!

Contact surface, solid fluid, I created Fluent (face and shadow face)!
Initialization is the best at the inlet of water in the control volume (velocity or mass flow rate and temperature of the fluid)
If the problem axisymmetric, geometric see to that.
You must set the four boundary conditions on the outer surfaces of the fluid that surrounds the steel piece (If 3D then go six!)

orlijda August 27, 2013 17:02

Thanks a lot, Felja.

According to your instructions, calculation tests are running fine.
Now, I will try to simulate the steam.



All times are GMT -4. The time now is 12:52.