I'm modelling a CFRP structure with 32 plies and using heat transfer shell elements. When trying to run the model I keep getting errors relating to the degree of freedom for the temperature outputs.
I have been reading up on the abaqus manual about this issue. I know that "A maximum of 20 temperature degrees of freedom can exist at a node” when employing shell heat transfer elements.
I also noted that "the number of integration points must be an odd number for Simpson's rule, and it determines the number of temperature degrees of freedom at a node of the element if this section is associated with heat transfer or coupled temperature-displacement elements. The maximum number of points for Simpson's rule is 99, and in the case of heat transfer or coupled temperature-displacement elements it is 19."
I want to know if its possible to only request outputs from selected points through the shell thickness or if this is fixed to the number of integration points and thus the number of material layers that I can employ?
Hi John Lyons
You can request outputs from selected points using *EL PRINT keyword in .inp file: https://abaqus-docs.mit.edu/2017/English/SIMACAEKEYRefMap/simakey-r-elprint.htm
Hi John Lyons ,
I encountered this error previously. I believe the maximum temperature DOFs relates directly to number of integration points and alas you are effectively limited on number of possible layers, as you have alluded to. This applies to Heat transfer and Coupled temperature-displacement steps. In my case, I was applying a uniform temperature profile that could be applied using a predefined field in a General step, thereby circumnavigating the need for temperature boundary conditions - I simulated 30 layers under uniform temperature.
In your case of heat transfer I suspect you are probably limited. My recommendation is to consider ply orientation, as I did, if you have flexibility to vary this in your problem. I imagine your layup is something classic like [+45/-45]16? If your problem allows, you can consider something like [(+45)8/(-45)8]2. Since we assume perfect bonding in laminate analysis we can model this sequence as four layers, +45/-45/+45/-45, with appropriate thickness in the layup editor (or composite section, whichever approach you use). Obviously this is a quick example; play around with different sequences if you can.
Hope this helps.
JCH
- Badges
- Science topic