Learning Forum

[antonio.ammendola]

Hi all,

I'm trying to do a convection model with moisture involved. When I setted all the parameters and tried to make a Hybrid Initialization the program gave me this error:

Error at host: malloc_shared_storage: unequal sv_lengths, thread 82, SV_STATUS.

Can you help me, please?

[Rob]

How much RAM have you got, and how many cells in the model? A malloc error is memory allocation so tends to mean you need a bigger computer. There are other causes, but that's the most common.

[antonio.ammendola]

Hi have a RAM of 32GB, With a mesh of around 3000000 cells.
Cosider that I've made already a model only for temperature distribution and it works very well. Like half an hour and I've the results

[Rob]

That should fit OK. What's different in this model if you've already modelled heat transfer? Have you got any UDFs in the case?

[antonio.ammendola]

No no i haven't fitted any UDFs. I run the model only with the heat transfer, without water vapor, only with air inside. Now I want to add water vapor and see if there are condensation fenomenon on the surface of the dominium. In order to do that I tried to add at the previous model the multiphase model (volume).

[Rob]

You had a single phase model then switched on multiphase? Which model? Are you wanting to model free stream or surface condensation?

[antonio.ammendola]

volume of fluid. I wanted to model surface condensation

[antonio.ammendola]

and yes i swiched from single to multiphase

[Rob]

Multiphase uses a slightly different data structure so the results you already have may not be valid with the current set up. You can get around this with an interpolation file. For surface condensation I'd also read up on the wall film models: VOF will be a very mesh intensive way to model condensation.

[antonio.ammendola]

So If i start another simulation from the beginning, using wall film model, I have to do an interpolation file anyway? Is possible to use Radiation model and multiphase model together? There are some tutorials that I can follow in ansys guide or something like that?

[Rob]

Wall film will run on from single phase as it's not a multiphase model. VOF won't, but you can interpolate the old data back onto the new set up assuming there isn't much/any of the other phase to consider.
Multiphase will generally work with radiation, but check the model compatibility and limitations.

[antonio.ammendola]

I'm sorry for bothering you again, i'm trying to set the Wall film, Do you know how can I set these values? Consider that my domain is closed, with no inlet no outlet. The condensation can be due only to the temperature variation caused by solar irray

[Rob]

Set the film initial condition, leave the Film boundary condition alone.

[antonio.ammendola]

Ok thanks. I've tryed but wen i initialize the case appear this message
===============Message from the Cortex Process================================

Compute processes interrupted. Processing can be resumed.

==============================================================================
Do you know why?

[Rob]

Did you initialise the film from the EWF panel? It's an odd set up and is separate from the rest of the solver initialisation.

[antonio.ammendola]

Yes, i enabled the wall film from the boundary condition and then i changed fron boundaty to initial condition, setting all to 0. then i initialised the case from the EWF setup. But always appear this error message.
===============Message from the Cortex Process================================

Compute processes interrupted. Processing can be resumed.

===========================================================================

[antonio.ammendola]

This is how i setted the EWF model

[antonio.ammendola]

[Rob]

Turn off the scalar and re-try.

[antonio.ammendola]

I'sorry for bothering you again but i've done what you suggest to me but i obtain the same error message

[Rob]

It's no bother. Shouldn't the film be water-liquid? Otherwise where does the latent heat go? Otherwise what you've shown looks fine. Try initialising from a flow boundary in case there's something odd being set in the hybrid scheme for the initial data.

[antonio.ammendola]

Good Morning I managed to avoid the error passing from water vapor to water liquid. But at the end of the simulation didn't result any presence of h2o as you can see from the residuals here attached.
Maybe I've to change something from the followings tabs?

[Rob]

Not sure, turning off surface tension shouldn't do anything to the film other than stop it beading up.

[antonio.ammendola]

Ok I did the set up like you said but i've this message
with this residuals trend