Header Integers Error Message

[mesheb82]

from a user.....
---------------------------------------------------------------------------------

I have successfully installed pynastran, and got it to work on the
test files, but when I tried running for some of our large op2 files,
I was presented with an error. Attached is the message.

After confirming that it is an 'assertion error', though I didn't
understand in full, I commented this line out after reading that this
is something to do with debugging, and found that the code worked.

I confirmed that the output files are identical whether this is
commented out or not for the test files, but I wondered if I could get
the author to verify this. Please could you confirm the following
statement?

Inside the file, /pyNastran_v0.4.0_r674/pyNastran/op2/fortranFile.py
line 77, in readHeader
the following line of code
assert ints[0]==ints[2]==4, "header ints=(%s) expected=%s\n" %
(str(ints[0:5]),expected)
does not affect the output of the program.


jiho@fairlead09:~/Desktop/Work/elemselection$ sudo python extract.py
Traceback (most recent call last):
File "extract.py", line 4, in <module>
op2Obj.readOP2()
File "/home/jiho/pyNastran_v0.4.0_r674/pyNastran/op2/op2.py", line
517, in readOP2
(isAnotherTable) = self.skipNextTable()
File "/home/jiho/pyNastran_v0.4.0_r674/pyNastran/op2/fortranFile.py",
line 563, in skipNextTable
marker = self.getMarker()
File "/home/jiho/pyNastran_v0.4.0_r674/pyNastran/op2/fortranFile.py",
line 292, in getMarker
tableCode = self.readHeader(expected,debug)
File "/home/jiho/pyNastran_v0.4.0_r674/pyNastran/op2/fortranFile.py",
line 77, in readHeader
assert ints[0]==ints[2]==4, "header ints=(%s) expected=%s\n" %
(str(ints[0:5]),expected)
AssertionError: header ints=([1239127744, 12900952, 1243166646,
1201398220, 0]) expected=None

[mesheb82]

my reply....

It's hard to say exactly why it happened without the BDF and OP2 (if
you can send me them I can take a look), but my guess would be you
either are NOT using PARAM, POST, -1, or you have a table in your OP2
that is not skipped properly. There are a lot tables structures in
the OP2 and many of them are not documented. If you get remove the
broken tables from the file by rerunning nastran without the offending
table, it should be fine. You should be able to get a good idea from
the debug messages.

------------------------------
after receiving the BDF...the case control deck is as follows...

CEND
TITLE = TEST2.NAS.BDF
ECHO = NONE
FORMAT = OUTPUT2
FORMAT OUTPUT2
SUBCASE 2940
LOAD = 2940
SPC = 1
DISP(plot,print) = ALL
STRESS(plot,print) = ALL
BEGIN BULK
-----------------------------------------------------------
removing the following lines fixes the problem...
FORMAT = OUTPUT2
FORMAT OUTPUT2

there was a superelement table (GEOM1S) that is not handled properly

[mesheb82]

since i have no idea what other consequences that had :), the reply:

I have solved the problem. We are actually trialing RADIOSS which has
a NASTRAN deck, and it turns out that by default, it outputs the
model, and something called 'interior points of external
superelements', which sounds somewhat similar to 'superelement grid
point table' (GEOM1S) as you have said.

In our case, it turns out we needed to actively disable this by adding

OUTPUT OUTPUT2 NOMODEL NODMIG

I am guessing I don't need NOMODEL, but I added it anyway.