SteveDoyle2/pyNastran

grid type 10 error

Opened this issue · 4 comments

Vybornak2 commented

Hello,
I was trying to use pyNastran to read my OptiStruct results. Everything went well until I tried it on model with contact. This error has been caused by nodes that have been in contact as I found out. I actually do not understand this issue much, since I was not really aware of different grid types.

Error:

def recast_gridtype_as_string(self, grid_type):
    """
    converts a grid_type integer to a string

    Point type (per NX 10; OUG table; p.5-663):
    -1, Harmonic Point (my choice) -> '    ' = 538976288 (as an integer)
    =1, GRID Point
    =2, Scalar Point
    =3, Extra Point
    =4, Modal
    =5, p-elements, 0-DOF
    -6, p-elements, number of DOF
    """
    try:
        grid_type_str = GRID_TYPE_INT_TO_STR[grid_type]
    except KeyError:
        if grid_type in NULL_GRIDTYPE: # 32/64 bit error...
            warnings.warn(''.join(self.get_stats()))
        raise RuntimeError(f'grid_type={grid_type!r}')
    return grid_type_str

RuntimeError: grid_type=10

I actually do not need contact forces here. Only stress values at shell elements are of interest to me.

  1. Are there any ways of avoiding this error?
  2. As I would specify for bdf file to read certain keywords, can I specify this for op2? How? (I couldn't find it in documentation)

If there is not workaround I wouldn't mind helping with this if you would provide some clarification. But it also depends on the scale of such task :).

If implementation/solution would be not possible at the moment, would it be possible to at least implement some custom error that could be caught?

SteveDoyle2 commented
Vybornak2 commented

Hello, I am sorry. Here is a full trace:

File c:\Users\vybjan\py_projects\fatigue_evaluation2\fatigue_evaluation2\model\read_op2.py:37, in ResultsReader._read_op2_file(self, debug)
     35 self.logger.info("Reading op2 file: %s", self._file_path.name)
     36 logger = logging.getLogger("pyNastranOP2")
---> 37 self._op2 = read_op2(self._file_path, log=logger, skip_undefined_matrices=True, mode="optistruct")  # type: ignore
     39 self.logger.debug("Getting subcases")
     40 self._subcases = {key: " ".join(value[3].split()) for key, value in self._op2.isubcase_name_map.items()}

File c:\Users\vybjan\py_projects\fatigue_evaluation2\.venv\Lib\site-packages\pyNastran\op2\op2.py:1424, in read_op2(op2_filename, load_geometry, combine, subcases, exclude_results, include_results, log, debug, build_dataframe, skip_undefined_matrices, mode, encoding)
   1420     model.set_subcases(subcases)
   1421     model.include_exclude_results(exclude_results=exclude_results,
   1422                                   include_results=include_results)
-> 1424     model.read_op2(op2_filename=op2_filename, build_dataframe=build_dataframe,
   1425                    skip_undefined_matrices=skip_undefined_matrices, combine=combine,
   1426                    encoding=encoding)
   1428 ## TODO: this will go away when OP2 is refactored
   1429 ## TODO: many methods will be missing, but it's a start...
   1430 ## doesn't support F06 writer
ref='c:\Users\vybjan\py_projects\fatigue_evaluation2\.venv\Lib\site-packages\pyNastran\op2\op2.py:1'>1</a>;32m   (...)
   1435     #setattr(obj, attr_name, attr)
   1436 #obj.get_op2_stats()
   1437 return model

File c:\Users\vybjan\py_projects\fatigue_evaluation2\.venv\Lib\site-packages\pyNastran\op2\op2.py:618, in OP2.read_op2(self, op2_filename, combine, build_dataframe, skip_undefined_matrices, encoding)
    616     OP2_Scalar.close_op2(self, force=True)
    617     raise
--> 618 self._finalize()
    619 op2_reader._create_objects_from_matrices()
    620 if build_dataframe:

File c:\Users\vybjan\py_projects\fatigue_evaluation2\.venv\Lib\site-packages\pyNastran\op2\op2.py:651, in OP2._finalize(self)
    649 for obj in values:
    650     if hasattr(obj, 'finalize'):
--> 651         obj.finalize()
    652     elif hasattr(obj, 'tCode') and not obj.is_sort1:
    653         raise RuntimeError('object has not implemented finalize\n%s' % (
    654             ''.join(obj.get_stats())))

File c:\Users\vybjan\py_projects\fatigue_evaluation2\.venv\Lib\site-packages\pyNastran\op2\result_objects\table_object.py:657, in TableArray.finalize(self)
    655 for ugridtype in ugridtypes:
    656     i = np.where(gridtypes == ugridtype)
--> 657     self.gridtype_str[i] = self.recast_gridtype_as_string(ugridtype)

File c:\Users\vybjan\py_projects\fatigue_evaluation2\.venv\Lib\site-packages\pyNastran\op2\result_objects\op2_objects.py:591, in ScalarObject.recast_gridtype_as_string(self, grid_type)
    589     if grid_type in NULL_GRIDTYPE: # 32/64 bit error...
    590         warnings.warn(''.join(self.get_stats()))
--> 591     raise RuntimeError(f'grid_type={grid_type!r}')
    592 return grid_type_str

RuntimeError: grid_type=10
SteveDoyle2 commented

Try turning off pandas (build_dataframe=False). I don't know why Optistruct uses that gridtype, but contact has never been tested with it.

Vybornak2 commented

Hello, I did run read_op2() function with its default for argument build_dataframe. So this did not solve anything.

Actually I found out, that building dataframe using pyNastran turned out to be very slow for me, so I am not utilizing this option. I am actually not sure why. DataFrame itself looks very nice though. I am not sure if that is just nature of things with pandas when multiindexing, or if there is something else slowing up the process. So I am building a bit simper Df myself from np.ndarrays.