SimScale CAE Forum

Thermomechanical memory allocation error


#1

Hello everyone,

i am trying to do a thermomechanical simulation of an reactor cell with an ZnSe window for IR analysis. Unfortunately i get an memory allocation error every time even when i use 16/1 cores (I only have a free account so i cant use more than 16 cores). Are those just not enough or what could be the reason for this? The Project in question is “DRIFT cell complete”.

Regards,
Phillip Timmer


#2

How many mesh elements do you have?


#3

Sorry for the late response. I thought i solved the problem and wanted to check first. After removing two tiny surfaces that caused the tet dominant mesh to fail it now only shows “an error occurred” without further explanation :thinking: The mesh now has about 800 k cells and 160 k nodes. Before (i used “new mesh preview” since it worked fine for thermal analysis) i tried between 60 and 293 k cells. The Projekt with the corrected CAD is now called “Phills cell 2” if you want to have a look.


#4

I see. 800k is fine with 16 cores.perhaps send a link to your project so I can take a look


#5

https://www.simscale.com/projects/ptimmer/phills_cell_2/

Thank you.


#6

I have had a look at your project. My first question is, why a transient simulation? This looks like a simulation where you are only interested in the end result and not a transient response in which case a steady state simulation can be used (it is less resource intensive and more stable). If you feel you need a transient case, then your time step length (under simulation controls) must be much much shorter than what it is now.

Lastly, the boundary condition of elastic support, why do you choose this on the face of the Viton seal? It will already deform easily due to its low young’s modulus. The fixed or elastic support gives the simulation a reference point to calculate the thermally induced displacement.I think (i’m not certain) it may be incorrect to define an elastic boundary in an interface between two bodies as to me it seems like a physically impossible scenario. Try making this boundary fixed, or if that does not work then try deleting this boundary condition and replacing it with a fixed support on an external surface.

Let me know how it goes!


#7

I noticed that i didnt run the simulation with 16 cores wiht the last parameter set and it actually went through with 16 cores. I think it was the elastic support that made the difference in the end. I’m not sure tho. Thanks for your help.