Running older versions of .bdf files on Nastran 2017? 3

[BorisGeorgiev93]

Dear all,

1) I would like to ask whether anyone has ever had problems in running .bdf files, created in Patran 2016 or older versions, on Nastran 2017?
Basically, if one works with the most recent version of software, does that impose any limitations in terms of working with files that are created in older versions?

2) Finally, a question related to the database (.db) file. When I open Patran and set a new database, the default working directory is C:\Windows\temp ( or on other PCs it is C:\Users\Documents ). In short, is it a problem to save the .db file ( together with any other analysis files like .f06,.xdb etc. ) at any other place on the hard disc. Or Patran/Nastran really requires the user to work in C:\Windows\temp or in C:\Users\Documents.

Really hope that anyone can shed some light on that matter!

Many thanks in advance!

Boris

 

[FFB12]

1) do you have a specific example? it should work in any version

2) you can use any work directory that you want

 

[BorisGeorgiev93]

Filipe, hello and thank you very much for the response!

1) Yes, I tried to run a .bdf file (about a FEM of an aircraft wing) directly in Nastran, and surprisingly the .f06 file that was outputted is 1) very small in size and 2) contains messages for errors. I asked a friend of mine who has Nastran/Patran 2016 to run it on Nastran for me. The .f06 that my friend produced was much bigger in size and showed no error messages-all the needed information was contained in it. If you or anybody else would like to give it a try, I can attach the .bdf file in my next message.

So this is why I decided to ask about 1) and 2) in my first post - just in case that this may be the problem.

Both my and my friend run 64 bits Windows 7 on a 64 bits laptop architectures. I think that maybe it is good to mention that I do not have any problems when I build my own models in Patran and produce the .bdf files.

Thank you again!

Boris

 

[FFB12]

What is the error? can you post the error message from .f06 here?

I can run the model and check if any errors come up

 

[BorisGeorgiev93]

Filipe, hello!

Please, click here in order to get the f06 file that Nastran 2017 outputs for me. Inside you can see the error messages.

Unfortunately, after asking today I was told that I am not allowed to share with anyone the structural model of the wing. We do a small internal project here and these are the rules...

Thank you again!

Boris

 

[FSB1]

can you create a simple model, for example a beam, or a box and try the same, generating the same error, and attach here?

 

[nlgyro]

I eye-balled your f06 from your nastran v2017 run & the bulk data echo found on the f06 file is missing a MAT1 card of ID 200 which is referenced by the PBAR card. This error originates in the 'XB2_GRID.DAT' include file of your model. You need to include your MAT1 card for the model to run!

Having said this I'm surprised that this model would run in any version of nastran as a missing/defective MAT1 card is a fundamental problem that will always cause nastran to FATAL out.

Nastran also reports your BEGIN BULK delimiter is missing in your input file.

For your query on Patran, what you need to do is this:

1. Copy the MSC.Patran (what ever version is installed on your machine) desktop shortcut onto your working directory (say for example C:\WORKING\PROJECTXXX)

2. In C:\WORKING\PROJECTXXX, right click on the just copied patran shortcut.

3. Select the 'shortcut' tab

4. Set the 'start in:' directory as C:\WORKING\PROJECTXXX

5. Click on 'Apply' and 'OK' and close the window.

6. Now double click on the Patran shortcut icon present in C:\WORKING\PROJECTXXX

This would launch Patran in your current working directory of C:\WORKING\PROJECTXXX and would enable all nastran files generated using this session of Patran to fall into C:\WORKING\PROJECTXXX

You can launch nastran from the same working directory to enable the o/p files to fall into the same location.

Repeat the same procedure when you change your working directory.

 

[BorisGeorgiev93]

Nlgyro, thanks for the post!

1) To be honest, I really do not think that there is anything wrong/missing with the model. After running the .bdf file in Nastran 2016. No errors are present.

2) As far as I understand you provided the above 6 steps in the cases when I would like to change my working directory, so that it is possible to place the working directory anywhere on the hard drive, each time I want to change it.

However my question was not about changing the working directory at all. All I wanted to know is whether there is any problem with Nastran to do analyses, if I save the Patran database and run Nastran at a completely different place on the hard disc. Currently, each time when I open Patran the working directory, as I mentioned above, is in C:\Windows\temp. Is there a problem if I want to save and run everything in Local Discs D or E (my hard disc is split into 3 parts: C, D and E and the OS together with any software is in C). Since I prefer to use more discs D and E, I do not want to bother changing every time the working directory...hope it can stay in C:\Windows\temp and nothing bad happens with Patran/Nastran when i run files that are saved in D or E. From Filipe's answer from above I got the impression that I can do that. Hope I made myself clear enough.

 

[nlgyro]

Your f06 file output from the nastran v2016 run reads the MAT1 card as can be seen from the model summary echo. But for the nastran v2017 run and based on the f06 file output you posted earlier, does not read the MAT1 card. This can be verified by looking at the bulk data in the f06 file and the model summary echo. So its obvious to say that there is something amiss with the MAT1 card. This is the only logical conclusion that can be reached sitting here without having access to your model.

If you can just post just the MAT1 card as-is on your input file, we can take it from there.

And as far as your query on working folders... its advisable to do your analysis on drive other than C:\ so that the system and OS files remain un-touched.

 

[dmapguru]

Hi Boris,

The MAT1 entry with ID 200 has an integer 0 in field 7. Field 7 is the thermal expansion coefficient if I remember correctly. Either delete the trailing zero on the MAT1 line, or add a decimal point immediately after the zero.

The input file reader has strict checking on the data types in the fields of the bulk data section.

There is nothing wrong with the file XB2_GRID.dat; this just happens to be the next piece of data read after the offending line.

 

[FFB12]

and are you sure you have the nastran input file, as well as the files
XB2_MAT1.dat
XB2_GRID.dat
XB2_GRID_AERO.dat
XB2_CBAR.dat
XB2_PBAR.dat
XB2_CONM2.dat

all in the same folder?


anyway, it looks like you have an extra digit in
"1520.0000"
try this one instead
MAT1 200 5.17e+12 0.3000 1520.000
4.77e+084.77e+082.75e+08

 

[nlgyro]

Hi Boris,

I think you're right when you claim that the model runs successfully in v2016 as-is (meaning with this defective MAT1) entry but FATAL's out with v2017. I was thinking about this and one plausible explanation for this can be could be the use of IFPSTAR.

If my memory serves me right, since v2010 MSC introduced the IFPSTAR which is an enhancement to the nastran bulk data processor. One of the things this did is precisely what you have faced in v2016, which is automatically convert an integer to a real number if the integer was found in a field which accepted only real numbers (as in the case of the CTE field of your MAT1 card). The user would be oblivious to this as the model would still go through. Now is this a good thing? I personally don't think so as this would escape the check of any accidental entries of integers’ in real field. The versions of Nastran prior to v2010 would give a fatal error for an integer in a real field.

I eye-balled the nastran v2017 QRG and the remarks for ifpstar=yes (a default for all nastran runs since v2010) which now says and I quote:

"Selects the MSC Nastran 2010 release Common Data Model for data processing, do not allow an integer in a field that the only possible format is real."

Which means that starting v2017, if IFPSTAR detects a bulk data entry where there is an integer in a field reserved for real numbers, nastran will FATAL out. This is precisely what you have observed in your v2017 run. Now this is a big improvement over what was there earlier.

Personally if you are a newbie to nastran I would recommend you run models using ifpstar=no so that such kind of defective bulk data entries would not slip through the cracks and resurface later and more importantly it puts the onus on the user to ensure that the format of the inputs are correct.

Hope this helps!

 

[BorisGeorgiev93]

Hello again!

Many thanks to all of you for the help!

Felipe, yes all the files are in the same folder, when I run the BDF file.

Indeed when I type, in the XB2_MAT1.dat file, 1520.000 instead of the original 1520.0000 the analysis was completed successfully.

nlgyro, please let me know how IFPSTAR can be controlled in Patran, since I do not see it anywhere in the BDF code or the DAT code? Or i add/type it somewhere in the files I have?

Thank you all again!

 

[nlgyro]

ifpstar=no should be typed in the nastran command line when you submit the job and type things like (for example) scr=yes etc..

 

[VN1981]

This is related to Patran 2017. I think there is a bug with Laminate PCOMP creator in Patran 2017. It incorrectly writes ply thickness values in bdf files. For ex: I created a 5 ply SYM laminate with each ply of thickness 0.1 inches. In the bdf file it writes ply thickness value of just 10.

Can any else confirm the presence of this bug so that we can pass the message to MSC?

 

[FFB12]

now I'm curious how your friend didn't have any error using exactly the same same files

 

[VN1981]

I guess not many Nastran users here are on Patran 2017 yet!

Filipe, I guess Nastran won't throw any errors while using bdf file from 2017 in 2016...will just get erroneous output/incorrect results.

 

[nlgyro]

@VN1981

Does your observation happen when generating PCOMP's using Create--->Material--->Composites or are you using the Patran Lam Modeler?

 

[nlgyro]

...and also what OS are you running your Patran 2017 on?

 

[VN1981]

I rarely use Patran Lam Modeler...PCOMP bug is caused using Create--->Material--->Composites

We are running Patran 2017 on Win-7 x64 OS.