[dan86]

I'm having divergence issues with eulerian wall film model. I'm using a solid cone injection of water into air (DPM model), spraying a flat plate (30 mm thick) positioned normal to the axis of the spray. The spray axis horizontal and flat plate vertical, gravity enabled to act on the liquid film in z (downward direction).

The injection is water set at 353 K, while the continuous phase is initialised at 303 K. The point of the simulation is to get an idea of the heating of the plate due to the impinging spray/liquid film.

I'm running as transient solver, with a very small time step 1e-04 s, which gives convergence in around 25 iterations / time step.

Everything is fine up until the liquid film just begins to build up. At first the temperature of the film is equal to the temperature of the incoming droplets (as expected) but maybe 10 time steps later the temperature of the liquid film suddenly increases by around 80K to way above the spray temperature and the solution eventually diverges.

If anyone has any experience with this model or any suggestions, it would be much appreciated. Thank you.

[MDinc]

Hi,

I am using the same models EWF-DPM for spray cooling purposes in Ansys Fluent 14.5. I will first give some information about this model.

My computational domain is 3D rectangular 90 degree (quarter domain).
I am using full quad computational cells with 0.5 mm size. My time step size is 1e-05 s for DPM and gas phase. I am using coupled DPM and gas phase with both unsteady. I am using solid cone water spray at 300 K which is the same temperature with gas phase (air). I use k-epsilon turbulence model for air flow calculations. I test constant surface temperature and constant surface heat flux boundary conditions at the impact surface. The impact surface temperature is 393 K (20 degree saturated) for the constant surface temperature boundary condition case. The initial film thickness at the impact surface is 1 micron and initial film temperature is 372 K.

My simulations run and converge without any issues. Film thickness and temperature variation looks feasible but mesh optimization will be performed at the impact surface region so that more accurate results could be obtained. Also, since the EWF model does not include any boiling, I will try to add it using UDFs so that boiling and phase change calculations can be performed.

Based on your information, it is hard to understand what could be the reason for the divergence. Which version of ANSYS Fluent are you using? Can you check the EWF model setup? Can you also check the DPM model setup including the injection? It would be also helpful if you can show some different snapshots of your model setup (EWF, DPM, Turbulence, initialization, solution controls, etc.).

Thanks,
Murat

Originally Posted by dan86

I'm having divergence issues with eulerian wall film model. I'm using a solid cone injection of water into air (DPM model), spraying a flat plate (30 mm thick) positioned normal to the axis of the spray. The spray axis horizontal and flat plate vertical, gravity enabled to act on the liquid film in z (downward direction).

The injection is water set at 353 K, while the continuous phase is initialised at 303 K. The point of the simulation is to get an idea of the heating of the plate due to the impinging spray/liquid film.

I'm running as transient solver, with a very small time step 1e-04 s, which gives convergence in around 25 iterations / time step.

Everything is fine up until the liquid film just begins to build up. At first the temperature of the film is equal to the temperature of the incoming droplets (as expected) but maybe 10 time steps later the temperature of the liquid film suddenly increases by around 80K to way above the spray temperature and the solution eventually diverges.

If anyone has any experience with this model or any suggestions, it would be much appreciated. Thank you.

[MDinc]

Hi,

Also, what is the nozzle-to-impact surface distance? In my case, it is 38 mm but we neglect the 10 mm section which is spray atomization region so that spray is injected 28 mm above the surface with approximate spray information (average drops velocities and diameters etc.). By doing this we could reduce the total number of computational cells in high amount since spray atomization part needs very small cells.

What is the minimum film thickness that is given in the EWF model setup? It should be zero otherwise it could calculate some wrong values during calculations.

Murat

[surojittaurus]

Hi Murat,

For the simulation that you described in this thread, did you create interface between the solid and liquid domain for your simulations?

Thanks,
Sujit

[sutiut]

Hello,
I'm working on wall film formation using droplet. I am using Eulerian wall film and DPM in fluent. I start with DPM of droplets and inject droplets to the domain using plain orifice atomizer, till all the things is ok. then I activate EWF for droplet collection, but after that, the droplets dissapear. I changed every thing and ever the film material and chose a different material than the droplets, but yet, the droplets disappear.... I'm so confused, please help and tell if you need any extra information...
I'm using steady flow and unsteady particle tracking and EWF.

[Tensian]

Dear all,

I am working in spray cooling using DPM. My model is not working (I have a target solid piece that is not changing temperature due to the particle heat tranfer with the solid).

I have tried several thing (I use to ran steady state, but if I switch to transient, it seems that simulation exploedes when particles contact the surface of the solid body)

Could someone of you share your model with me to check some FLUENT settings?

Thanks in advance.

Regards,

Tensi