I recently started using Patran/Nastran 2018. I just noticed that my nastran input files now have the following command at the top: NASTRAN SYSTEM(216)=YES. I do not recall seeing that with previous patran/nastran versions. Unfortunately, I cannot find this particular system cell documented in the QRG or any other manual.
Please tell me why there is no mention of system cell 216 in the QRG or any other version 2018 manual. I would like to know what this command is doing and if it is necessary. I expect all commands to be documented in your manual. Am I expecting too much?
I found what this does... it relates to the iterative solver, so if Patran is writing it out, I suspect you have chosen the iterative solver instead of the default?
Here is a description:
system(216) keyword=ITER type=Integer
Controls execution of iterative solver in SOLs 1, 8, 11, 101, 108, 111 and 106.
1 or YES: Execute iterative solver.
0 or NO: Do not execute iterative solver (Default).
You can also control the iterative solver using the SMETHOD/ITER case control/bulk data command.
I played around a bit and couldn't get Patran to export a bdf with that system cell... did you happen to import an existing bdf into Patran? Or could there be a Direct Text input defining that line?
I will create a bug report asking for this to be added to the documentation if I can reproduce Patran writing it...
Thanks for that information Don. I normally use the SMETHOD=ELEMENT to turn on the iterative solver. I was not aware of the system(216) method since it is apparently undocumented. I do not have any direct text in Patran and it is just a standard analysis output from Patran. I have lots of subcases defined and there is a good chance that I did select the iterative solver for one or more of those cases from within patran awhile back. But then I found that the iterative solver was nearly 10 times slower for my huge tet model. So I went back to the sparse solver. I was selecting the solver by manually editing the bdf file and not from with Patran. I will double check my patran db to make sure that the iterative solver option is unchecked because I do not want to use the iterative solver.
While we are on the subject of the iterative solver, I wonder if you can point me to the documentation where the difference between the ELEMENT and MATRIX methods are clearly explained. There must be two methods for a reason, and I would like to know when one method should be chosen over the other. Thanks,
In looking for some information to answer your second question, I came across the only reference to System Cell 216 I have seen... it was in the Numerical Methods User Guide, in the Iterative Solver Solutions section:
"The NASTRAN keyword ITER (equivalent to SYSTEM(216)) is used to select the iterative solution by ITER = YES. The default is ITER = NO."
Of course, ITER is a keyword so it is not defined in the QRG... it is defined in the Installation and Operation Guide.
The Numerical Methods User Guide is where you can find some theory and details about most of these options:
CASI is generally recommended for large solid element dominated models... it does have some limitations related to large number of load cases... check out some of the information there.
In any event, using SMETHOD=ELEMENT is turning on the CASI iterative solver specifically, while SMETHOD=MATRIX it turning on the regular iterative solver using all the defaults. If you want to override the defaults for either form of iterative solver, you can use SMETHOD=n where "n" is the ID of the ITER bulk data entry... this entry gives you total control over the various iterative solver options. These are decently described in the QRG.
Note that System Cell ITER (216) was listed in 2005R2 QRG, but not in 2007R1 QRG. I don't see anything about it in the 2007R1 release notes, but apparently that was the time frame where it was removed from the QRG.