H5 output file does not show any subcase names when I load into Patran.

Right now I must use op2 or xdb format because those formats include the subcase names. For some reason I get no subcase names with h5. I just get the subcase number but no names in Patran. Please let me know what I am doing wrong. Surely someone at MSC knows the value of subcase names instead of just numbers. I would like to use h5 format since it is the latest and greatest. The name is supposed to come from the SUBTITLE command in the nastran input file.

Attached Files (2)

0 Don.Graff over 6 years ago

Hi Toby, This should work as you expect... I just made a sample and it seems to work... here is the Nastran input portion:

And here is the Patran Results form after hd5 import:

I used Nastran/Patran 2019... what version are you using and I'll try that out.

Don
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 Don.Graff over 6 years ago
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 Don.Graff over 6 years ago
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 toby.t.norris over 6 years ago

Thanks Don. Very interesting. This gives me hope. I will have to investigate further. Probably has something to do with the fact that I may have old leftover subcase definitions in my patran db file from previous solution file attachments. I will check.
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 toby.t.norris over 6 years ago

I double checked my files. I am using same format you have. I was using mdlprm,hdf5,3 to get the h5 file. Even after I delete all the old scrap subcases in Patran, I still never get any of the subcase names when I attach the new h5 file.

The names do show up just fine if I attach the op2 file or xdb file. But no names with the h5 file.

I wonder what command you are using to save your h5 file. Next I will try mdlprm,hdf5,1 to see if that makes any difference.
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 Don.Graff over 6 years ago

yes, I replicated your issue using format "3"- try using:

MDLPRM HDF5 0
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 toby.t.norris over 6 years ago

I am glad you noticed the problem.   Also, I did have success when I used mdlprm,hdf5,1.

Maybe someone at MSC should build a PCL script that automatically tests all of your command options prior to releasing to the public.   That would have caught this issue (bug or undocumented feature).

I have very large database. So I am trying to do all I can to minimize the size. That is why I was using the 3 option. The manual says that option is supposed to give minimum size without model input data and compressed. However the manual does not say anything about there being no subcase names with the results when you use the 3 option. Thankfully, I was able to figure this out on my own by experimenting with the various mdlprm,hdf5 options.

Thanks very much for your help on this issue. It is definitely resolved and I am very pleased with the attach speed for the h5 file format as compared to the op2.   So far the h5 is working great with minimum file size and fast attach time.
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 Don.Graff over 6 years ago

I will open a support case for you as well as a Nastran enhancement request and cross reference them together so you are informed when this is resolved.

FYI- I tracked down exactly why this happens. With the hdf5 creation using "2" or "3" the user is explicitly asking for the result data only, and no model data. It turns out that the SUBTITLE is stored with the model data portion of the hdf5 file. I've requested this be investigated, and possibly store the case control data always, even when using the "2" or "3" setting. The possible drawback is when a user has a large SET defined, or MANY subcases, which might increase the file size significantly.
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 Martin MSC Nastran over 6 years ago

Options 2 and 3 do not include input data for HDF5. This is documented in the system cell area and in the MDLPRM description with the Quick Reference Guide. The CASE CONTROL section with the TITLE, SUBTITLE and LABEL commands are part of the input data held in the CASECC datablock. The DMAP module CRDB_IN writes the CASECC datablock, along with other geometry datablocks. HDF5 options 0 and 1 include a call to CRDB_IN. With the POST option you control geometry in your .op2 file with the parameter OGEOM. OGEOM is set to yes by default. Looking at the output datablocks such as OUG1, which are table format, the subcase ID is present but I see no ASCII text records that would be associated with TITLE, SUBTITLE and LABEL. Further interest on the .op2 or .h5 content I will leave as an exercise to others using DMAP and PARAM settings.

Attached Files (1)
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 Martin MSC Nastran over 6 years ago

QRG entries for HDF5

Attached Files (1)
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel