Orthotropic and temperature dependent CTC?

DaveStupple

I was hoping to simulate a material with temperature dependent thermal conductivity that is also lower in the z-direction. I don't think this is currently achievable. Would a proportional increase in the z-direction dimensions achieve the same thing? How about adding thin layers of lower CTC material (sounds very clumsy)? Any other tricks?

Victor

It's not available with the internal solver.

You can do it using CalculiX by manually entering the cards for a material, including *CONDUCTIVITY which allows orthotropic temperature dependent conductivities.

Your idea of scaling the model in the Z direction sounds quite ingenious but it doesn't work in a test I just did. See attachment. Temperatures aren't the same on the stretched and non-stretched copies of the ball. Maybe it would be OK with simpler more 2D like geometry.

DaveStupple

I have attempted to use the CCX solver on my model, with the following errors:

Error: pyr13 element in Thermal Steady State cannot be converted.

Error: convection could not be converted.

My model has many pyramid elements (used in grading from smaller to larger elements). As far as I can tell, Calculix does not do pyramids at all. Is this right? 

Is convection just one of those things that Mecway does not transfer to the inp. file? 

The second point above is probably fixable by inserting *FILM functions to the inp. file, but I think the pyramid thing might be a show stopper.

Victor

Pyramids are currently a problem and I don't know a practical solution sorry. Version 6 will export them as collapsed hex elements which are OK for CCX but it's hard to do that by hand.

Correct that convection isn't currently exported but you could add it as you suggested.

DaveStupple

Hi all, 

I'm trying to get CCX to run my thermal steady state model, using Mecway 6 Beta 2, hoping that CCX will now cope with pyramid elements in the mesh.

1. I see *CFLUX entries in the CCX inp file. I guess these are representing my internal heat generation in the Mecway model....?

2. I notice also that the inp file seems to recognize my named selections from the Mecway model, e.g. the face selection called 'Cooled' appears as:

*SURFACE,NAME=Cooled

65214,S3

65215,S3

65216,S3

65217,S3... etc.

I tried to use this selection to apply a convection to this selection, so I added a *FILM entry to the custom step facility in Mecway. This has appeared at the end of the inp file as follows:

*NODE FILE,GLOBAL=YES

NT

*EL FILE

HFL

*FILM

Cooled, 283.15, 3000

(NB: this is the end of the inp file, there is no *END STEP)

When I tried to solve with CCX, Mecway reports: solver failed to produce an output file

and the CCX messages output contains:

 *ERROR reading *FILM. Card image:

        COOLED,283.15,3000

I suspect that I have used the named selection incorrectly, or I need to include some more detail in my custom step (e.g. adding *END STEP).

I am trying to make sense of this, but if one of you guys could save me give me a pointer it could save me a lot of time/pain.

Thanks,

Dave

Victor

Unfortunately *FILM doesn't allow you to use a predefined surface like that. It needs the individual faces to be defined directly in the *FILM block.

If all the faces are the same face number of each element, then you can use an element set instead, otherwise you might need to copy and paste the *SURFACE data and replace the S with F on each line. See the CCX manual on *FILM for more details.

Please start a new thread for a new topic so that it's easier for other people to identify by the subject line when searching.

DaveStupple

Thanks Victor, I've tried to progress this but hit a brick wall. I'll start a new post.