Learning Forum

[CFDaniel]

Hello all,

i'm doing the following tutorial which i found on youtube: FLUENT SIMULATION OF PCM IN REFRIGERATOR || SOLIDIFICATION AND MELTING ( 

At 03:00 when creating the calculation activities i get the following error:

Error: GENERAL-CAR-CDR: invalid argument [1]: improper list

 

Error Object: global

 

I choose CDAT for CFD-Post & EnSlight. When i run the simulation everything works fine, no divergence.

 

Does anyone know what i'm doing wrong?

 

Kind regards

CFDaniel

[Rob]

A CAR-CDR error usually means there's a setting missing somewhere. Given the model runs OK, check the execute commands and whether Fluent is spitting out data. 

[CFDaniel]

Thank you rwoolhou.

I think there is something wrong with the modeling of the PCM.

When creating a contour plot with the liquid fraction of the PCM there is only 0 and 1.

But when creating a contour plot with the temperature, it shows me a realistic temperature field.

 

But i don't know what i did wrong with the modeling of the PCM, i followed all the steps several times and it makes everything sense to me

[Rob]

Does the PCM occupy the whole volume? Check liquidus & solidus temperatures: are they the same?

[CFDaniel]

The simulation is a 2D axisymmetric simulation and the PCM occupys the three little rectangles on the left side which is also the midline.

Liquidus temperature is 274K

Solidus temperature is 271K

[Rob]

How much mesh is there in those regions? 

[CFDaniel]

In General:

 

Mesh Quality:

 

Minimum Orthogonal Quality = 9.99966e-01 cell 14254 on zone 22 (ID: 25804 on partition: 2) at location ( 2.45466e-01 3.09200e-01)

(To improve Orthogonal quality , use "Inverse Orthogonal Quality" in Fluent Meshing,

where Inverse Orthogonal Quality = 1 - Orthogonal Quality)

 

 

Maximum Aspect Ratio = 2.17979e+00 cell 1445 on zone 23 (ID: 42601 on partition: 0) at location ( 2.49484e-01 5.39000e-01)

 

There are 2700 faces in each of the small rectangle and 3600 faces in the largest PCM surface.

[Rob]

Plot temperature with node values off, and also see what happens if you reduce the time step: you could be melting the material too quickly. 

[CFDaniel]

There is no big difference by plotting the temperature wit node values off.

Reducing the time step doesn't change the result.

If i plot the volume integral of liquid fraction over flow time, the liquid fraction increases.

[DrAmine]

Can you please add screenshots of solidification panel, temperature contours and Liquidus fraction contours?

[DrAmine]

Please try again uploading the images (via insert Image). As ANSYS Employee I do not download any attachments or go to any external site to download them there.

 

The solidification and melting panel: The settings applied there.

[CFDaniel]

Everytime i upload an image via Browse i got an error: Sorry an error occured...

Do you know what could be wrong with that?

[DrAmine]

I will check that. Probably we are experiencing some issues.

[CFDaniel]

Chrome and Edge doesn't work...

[Rob]

It's us: IT are working on it..... 

 

[DrAmine]

Can you try now inserting the screenshots?

[CFDaniel]

Now it is working. Thanks

[DrAmine]

Is is axissymmetric? Is the x-axis the axis of symmetry?

[DrAmine]

Is it expected that the PCM solid regions will start melting? Have you run for longer time?

[CFDaniel]

The model is axissymmetric and the y-axis is the axis of symmetry.

It is expected that the PCM solid regions melts. I also run for longer time, but no difference.

Here you can see the temperature field at t=0s.

[DrAmine]

Axis of symmetry has to be the x-axis.

[CFDaniel]

Sorry for the late response, i was in vacation.

I looked through the simulation and made a mistake last time. The simulation isn't axissymmetric but the left side is the symmetry line. I got it mixed up.

Do you know what else can be a possible error?

[Rob]

If it's 2d all that will do is mirror on that line. 

[DrAmine]

For 2d axis-symmetric  cases  the x-axis has to be the axis of symmetry (otherwise fluxes and field might be wrong). So make the vertical axis the x-axis by flipping the whole by -90°. If your case is not axis-symmetric just use 2D planar.

[CFDaniel]

@rwoolhou

what do you exactly mean?

@abenhadj

my case is not axissymmetric. i use 2d planar and the whole left side of my geometry is a symmetry boundary.

[Rob]

Symmetry is effectively a mirror plane. As it's 2d (and no axisymmetric) it means you're effectively solving half of a cuboid shape. 

[CFDaniel]

Yes, that's the way i run the simulation. But i get the resuts shown above...

[DrAmine]

Something is completely wrong with your case either material properties or boundaries. As you are referring to that tutorial video, have you contacted the author of the video, perhaps he can provide you with this case.

[CFDaniel]

I have already contacted the author of the video but he didn't know this error and furthermore can't help me.

Is it possible that someone of you can follow the steps through the tutorial and see if it works?

[Rob]

I've used the model fairly recently and it was working fine.  Start at the beginning and work through the whole model set up, boundary conditions, materials etc.  Double check there's a wall-wall:shadow pair between the different fluids zones if they're non-mixing.