readnas (+FEA)

Purpose

Import data from NASTRAN Output2, Output4, bulk data, and punch files.

Syntax

f=readnas('filename')
f=readnas
f=readnas('directoryonly')
f=readnas('supported')
[f,fname,opdir,opt]=readnas('*.op2','blocks',{'OUGV1','GEOM1'})
f=readnas(filename,'blocks',BLOCK_CELL,OPT,'silent','noprogbar','logfile')
f=readnas(filename,'modes',1:5)

Description

READNAS reads data from a NASTRAN file. READNAS supports Output2 (OP2), bulk data (BLK or DAT), PCH (XYPUNCH, SORT1, and SORT2), and Output4 (OP4) files. OP2 files must be created with the PARAM,POST,-2 or PARAM,POST,-1 option to read geometry and results. READNAS supports geometry, property, matrix, and nodal result data blocks. Geometry data is output in the IMAT geometry structure format. Physical and material property data is returned as structures. Results are output as imat_shp for nodal results, and as result objects for all other result types. Matrices are output as structures. The OGPWG table(s) are output as a structure, which is described in detail below. Unrecognized datablocks are read in as a structure of INT32 vectors.

Several optional input arguments are supported. FILENAME is a string containing the filename to be read. It must include the extension of the file to read. If FILENAME is not specified, or FILENAME contains wildcards (i.e. '*.op2'), a standard file dialog will appear. If FILENAME contains wildcards, it will be used as a filter for the file dialog.

Several string input arguments are also supported.

For PCH files, BLOCK_CELL can be a cell array containing either a single vector of numeric indices of records to read (an empty matrix specifies that all of the records should be read), or a cell array of block names (i.e. 'OUGV1', 'PCHBULK', etc.). The block names are those returned in the .name field of the PCH file directory structure (described here).

OPT is a structure containing special processing options for READNAS. You may also save this structure to a MAT file called readnas_options.mat. If READNAS finds this file in the current working directory, and the MAT file contains a structure named 'OPT', and you have not passed in an OPT structure, READNAS will use the OPT structure found in the MAT file. Supported options are shown in the table below. They are also more fully described below.

Four output arguments are supported. F is the returned data. It is a structure with fields that describe the output.

The following fields are possible for BLK, OP2, or PCH files.

Matrices read from the OP4 file will be returned as a structure array.

FILENAME is a string containing the full path and filename of the file read.

The OP2 directory structure is as follows:

The PCH directory structure is as follows:

Supported sort types are listed in the table below.

For OP4 files, the directory output is a cell array of strings containing the matrix names.

The physical and material property output structure is defined below. Both sets of property data are returned in the .properties field of the output. Physical property data is returned in the .ept field, and material property data is returned in the .mpt field. Each of these fields contains a structure, the format of which is described below. The following physical property cards are supported: PBAR, PBARL, PBEAM, PBEAML, PBEND, PBUSH, PDAMP, PELAS, PRBAR, PROD, PTUBE, PCOMP, PSHEAR, PSHELL, and PSOLID. The following material property cards are supported: MAT1, MAT2, MAT3, MAT4, MAT5, MAT8, MAT9, and MAT10.

The OGPWG (grid point weight generator) output structure is described below. If there are multiple superelements in the OP2 file, this structure will be an array, where each index contains the table for each superelement.

The LAMA (eigenvalue) output structure is described below. Each field is an NMODESx1 numeric vector, where NMODES is the number of modes extracted during the eigenvalue analysis.

Bulk data file TABLED1 cards are imported into the .tables field. This field contains another structure containing the .tabled1 field (in the future the other TAB* cards will be returned in appropriate fields). Each table is contained in a structure with the following fields.

OPT is a structure containing the processing options used during the import.

Important notes

The notes below refer to the default behavior. Specifics for each program mask are described further below.

Generally speaking, when you import results from an OP2 file, you need the geometry datablocks present so READNAS can access information relevant to the result processing. If you opt not to transform results to the basic coordinate system or return actual node numbers, you can import nodal and energy results without the geometry.

Nodes will be imported in the definition coordinate system when read from both bulk data files and Output2 files. SPOINTs will import as nodes with the color yellow and coordinates 0,0,0.

Matrices will be contained in a structure, where each element of the structure contains the contents of one of the matrices read. The structure fields are described in the following table. If the matrices being read from the Output2 file were created by the substructure DMAP alter provided by MTS for creating a Guyan-reduced substructure (KXX, MXX, and GO), READNAS will also fill in the DOF field. Otherwise this field will contain empty DOF lists, since NASTRAN does not store matrix DOF lists directly in the Output2 file.

If you have a DMI matrix, and wish to obtain a matrix where the row and column numbers match the row and column numbers in the DOF fields, you can use the expandDMI function.

CONM1 and CONM2 mass and inertia properties will be imported as physical property tables.

To read a subset of SORT1 results from the OP2 file, use the 'subcases' and/or 'modes' specifiers. These act as filters on the results. For modes runs, 'modes' specifies the mode(s) to read. 'subcases' allows you to specify results from specific subcase IDs to read. The subcase numbers match the number specified on the SUBCASE case control cards.

Element Type-Specific Extra Data

Some NASTRAN elements have element-based data that is specific to the element type. This data can optionally be returned by readnas, and is located in the extra property of imat_elem. readnas only supports the extra type-specific data for certain element types. These are documented in the table below.

BLK File Output

READNAS reads most information from a bulk data file, including parts of the file management section and executive and case control. It handles INCLUDE statements as well. Relative paths on an INCLUDE statement are treated as relative from the top-level bulk data specified to READNAS.

If bulk data includes a comment line that has the form "$TITLE       =      #", where column 79 contains a digit, it will assume that this is the start of a punch data record. If this is not intended, then please modify the format of the comment line so that READNAS interprets it correctly as a comment.

PCH File Output

READNAS considers Punch file data to be stored as a series of records. A record consists of a header line or lines, followed by numeric data. SORT1 records are written as a matrix, where the rows are the entity (i.e. node or element), and the columns are the item codes. One record is written for each time or frequency step, and for each step, one record is written for each entity type. SORT2 records are written as a matrix, where the rows are the time or frequency, and the columns are the item codes. One record is written for each entity type. XYPUNCH records are written as a 2-column matrix, where the first column is the time or frequency, and the second column is the data for a specific entity and item code. The XYPUNCH header contains very little information, so for example it is not possible to determine whether the first column contains time or frequency values. READNAS thus allows you to specify what this axis contains.

BULK data found in PCH files will be returned in the fields that are returned when reading bulk data directly. Each section of bulk data found in the PCH file will appear as a separate record. If the same bulk record types are found in multiple bulk sections, READNAS will attempt to merge them. All matrices will be returned, even if they are duplicated. Only the first set of FEM geometry will be returned. Only the first set of each set type will be returned.

Nonlinear CBUSH results are written to the PCH file as a stress record for element type 226. The data has 18 values. The first 9 are translational forces, stresses, and strains, in that order. The following 9 are the rotational forces, stresses, and strains. Note that the Output2 file output is ordered differently. Please see the notes below for datablock OESNLXR.

Punch file results are reordered in the same way, such that the first result values are force, the next 6 are stress, and the final 6 are strain.

It is possible to have READNAS return the SORT1 records in SORT2 format. To do this, pass in the OPT structure with the .pch.sort1tosort2 field set to true.

SORT2 and XYPUNCH records (and SORT1 records converted to functions) are returned in the .functions field of the output as imat_fn.

Matrix records are returned in the .matrix field as an array of structures using the format specified above.

SORT1 records are returned in the .records field as a cell array. Each cell contains the data for a specific record. If the record contained nodal data (item codes 3-8), it will be stored as an imat_shp. Otherwise, it will be stored as a structure with the following fields:

Output2 Import Notes

This section contains general information about how Output2 stores its data. The following subsections define how the import options modify the results from the way they are stored in the Output2 file. If no transformation is requested, results will be returned the way they are stored in the Output2 file. Please see below for details of what transformations occur if they are requested.

Nodal displacement results (OUG) are stored in the displacement coordinate system unless their datablock name begins with 'B', in which case the results are stored in the basic coordinate system.

Nodal force results (OQG) are stored in the displacement coordinate system.

Element forces read from the OEF datablocks are always imported with some coordinate system transformation. 1D elements such as CBAR and CBEAM are always imported in the element coordinate system. Forces are transformed according to the cross section rotation angle. Eccentricities are also accounted for. CELAS and CBUSH forces are always imported as they are stored in the Output2 file. Please note that for bar/beam elements, NASTRAN defines moments IN a plane, where IMAT defines moments ABOUT an axis perpendicular to the plane. readnas returns the moment in plane 1 as Mz and the moment in plane 2 as My, which means that the moment values will appear flipped from what NASTRAN reports. There may also be a sign change on the moments, depending on the program mask. Please read the sections on I-deas and Femap masks for more important information.

SCN Nastran 2019.1 through 2401 has a bug where the sign on the shell stress resultant moment output for CQUAD8 is negated from what it should be. This bug was fixed in SCN Nastran 2406. readnas returns the results as they are stored in the Output2 file.

NASTRAN does not write out stress tensors for 1D elements. IMAT returns the stresses as components of a tensor, but the values returned are not tensor components.

The OESNLXD datablock contains nonlinear output for different element types. NASTRAN mixes the datatypes stored in this datablock. Since results in IMAT can only have a single datatype, READNAS returns OESNLXD results as a force dataset. The implications are that if you export these results to another format (e.g. Universal file), and then change units and read the results back in, non-force results will be scaled incorrectly. The table below describes how the results for the supported element types are returned.

Nonlinear CBUSH results are written to the OESNLXR datablock. READNAS returns a stress dataset. Layer 0 contains the force results. Layer 1 contains the stresses, and layer 2 contains the strains. Please note that the resulting output is tagged by READNAS as a force, so changing the units on export/import may result in incorrect values for the stresses and strains. Also note that the PCH file import retains the NASTRAN ordering. See the notes above.

For 2D elements, element stress resultants are stored in the element coordinate system. See the CQUAD and CTRIA documentation in the NASTRAN Quick Reference Guide for more details on the definition of the element coordinate system, as well as the material coordinate system. Midside node results for parabolic elements are averaged from the two adjacent corner nodes, since the Output2 file does not store midside node results.

Stress and strain data is stored in the Output2 file in the element coordinate system for all element types except PCOMP. Stresses and strains for composite shell elements (PCOMP) are written to the Output2 file in the ply coordinate system.

If requested, READNAS returns stress and strain invariants. It returns Maximum Principal, Minimum Principal, Maximum Shear, Octahedral Stress/Strain, and von Mises. Nastran does not store all of these invariants for each element type. When it is available from OP2, that value is returned. When it is not available from the OP2 and it can be calculated, READNAS computes the value and returns it. When this is not possible, READNAS returns 0. Note that READNAS only computes invariants for SORT1 data.

* Strain invariants are not stored or returned for this data type

Grid point force balances are written to the Output2 file in the displacement coordinate system. See the notes for GPFORCE in the NASTRAN documentation. These results are returned as Data on Elements at Nodes results. Special quantities are identified with a non-positive element ID. Please refer to the table below for a list of these quantities and their associated "element" identifier.

The layer number for solid composite elements (CPENTA and CHEXA that reference PCOMPS) is 10*ply layer + BMT layer. This element type supports output at up to three locations on each ply (Bottom, Middle, Top). This is the BMT layer, where 1 = Bottom, 2 = Middle, and 3 = Top. The ply layer is 1 for the bottom ply and increments by 1. To determine the output location from the layer number, divide the layer by 10. The integer portion is the ply layer number and the remainder is the BMT layer. For example, a layer number of 42 corresponds to ply layer 4 middle surface, and 103 corresponds to ply layer 10 top surface.

Output on each element is ordered in the nodal order expected by I-deas. The node order is the same between I-deas and NASTRAN for linear elements. However, parabolic elements have a different numbering sequence. I-deas interleaves corner and midside nodes, and NASTRAN orders the corner nodes first, followed by the midside nodes. The node numbering listed in the component listing follow the NASTRAN convention, but are ordered in the I-deas convention. This is so exporting these results via writeunv will result in a correctly ordered Universal file result dataset. To clarify the ordering, consider a CTRIA6 with corner nodes 1, 2, and 3, and midside nodes 4, 5, and 6. The component listing for the results at this element will be in node order 1 4 2 5 3 6.

Matrices are written to the OP2 file several different ways. readnas supports all matrices written as matrix datablocks. These datablocks are named with the matrix name, and they contain only the matrices with no row and column DOF information. External superelements are written to the OP2 in the XSOP2DIR datablock which contains the directory of matrices, and multiple EXTDB datablocks that contain these matrices. readnas uses the information in XSOP2DIR to rename the EXTDB datablocks to their corresponding matrix name, so that the matrix name is what is displayed in the readnas GUI. The MATPOOL datablock contains matrices of several different formats. readnas imports the DMIG-formatted matrices from MATPOOL. Matrices from all of these sources are returned as a structural array in the .matrix field of the readnas output.

Option: Femap

This option sets the transformtobasic option to false, atnodes to true, atcentroids to true, and actualnodenumbers to false. This combination has the effect of returning the data much more closely to the way it is stored in the OP2 file.

Only one spring force is returned for each CELAS element, and is returned at node 0, the "centroid" and node 1.

CBAR and CBEAM element Y and Z moment values are switched as compared to the way they are stored and printed in the NASTRAN output (for example, F06), since IMAT defines beam moments about an axis and NASTRAN defines moments in a plane. The signs on the moments will match what is displayed in the F06 file. Forces and moments are returned for nodes 1 and 2 (end A and end B of each element), as well as the "centroid". The centroidal values are the average of the node 1 and node 2 values. The axial and shear forces are the same for both nodes, since NASTRAN only stores one value per element for the axial, shear, and torque. However, the moments may be different.

CROD and CONROD only returns components 1 and 4 for element forces. All 6 components are returned for stresses as the stress tensor components are used to store various other quantities (see notes above).

CBUSH returns forces at the centroid and at node 1. The values returned are identical for both locations.

Shell stress resultants are returned with a layer number of 100. This layer number is used to identify stress resultants when writing a Femap Neutral file with writeFemap.

Only results at the corners of parabolic elements are returned. No midsize results are returned, since NASTRAN does not store midside node values in its output.

Option: I-deas

This option sets the transformtobasic option to true, atnodes to true, atcentroids to false, renumbercsys to true, and actualnodenumbers to false.

Nodal displacement data read from the OUG datablocks (displacements and pressures) is imported from the Output2 file in the global (basic) coordinate system if geometry information is present in the Output2 file.

Nodal (SPC) forces read from the OQG datablocks are transformed to the global (basic) coordinate system if geometry information is present in the Output2 file.

Midsize node results will be returned for parabolic elements, but since NASTRAN does not store midside results, these are simply the average of the two corner nodes on that edge.

For bar/beam/bend element forces, the end A Mz and end B My moments have their signs negated from what is stored in the Output2 file and printed to the F06 file. In addition, the end A axial and shear forces and end A torsion are negated from what is stored in NASTRAN.

Output on each element is ordered in the nodal order expected by I-deas. The node order is the same between I-deas and NASTRAN for linear elements. However, parabolic elements have a different numbering sequence. I-deas interleaves corner and midside nodes, and NASTRAN orders the corner nodes first, followed by the midside nodes. The node numbering listed in the component listing follow the NASTRAN convention, but are ordered in the I-deas convention. This is so exporting these results via writeunv will result in a correctly ordered Universal file result dataset. To clarify the ordering, consider a CTRIA6 with corner nodes 1, 2, and 3, and midside nodes 4, 5, and 6. The component listing for the results at this element is in node order 1 4 2 5 3 6.

Option: transformtobasic

Coordinate transformations require that geometry information is present in the Output2 file. If this information is not present, results will be returned as they are stored in the Output2 file.

Nodal displacement data read from the OUG datablocks (displacements and pressures) will be transformed to the basic coordinate system. If these results were written to the Output2 file in the basic coordinate system (datablock name begins with 'B'), this option has no effect, since the results are already in the basic coordinate system.

Nodal (SPC) forces read from the OQG datablocks are transformed to the basic coordinate system.

Elemental forces for 1D elements such as CELAS, CBUSH, CBAR and CBEAM are transformed to the basic coordinate system.

Shell stress resultant data is not transformed and is returned as it is stored in the Output2 file.

Stress and strain data is always returned in the element coordinate system for 1D elements. For 2D and 3D elements, it is returned in the basic coordinate system for most elements. For CHEXA1, CHEXA2, CPYRAM, CQUAD, CQUADX, CTRAPRG, CTRIARG, and CWEDGE elements, the data is always returned as it is stored in the Output2 file.

Stresses and strains for composite shell elements (PCOMP) are transformed from the element coordinate system (defined by the element first edge) to the basic. This means that there is still a transformation from the ply angle to basic that has not been performed.

Heat flux and heat gradient data for 1D elements will be returned exactly as it is stored in the Output2 file. Data for 2D and 3D elements is transformed to the basic coordinate system.

OP2: Generic Tables

Tables not explicitly recognized by READNAS can still be imported. They will be stored in the .table field of the output. This field contains one field for each unrecognized table read. The field names are the names of the tables read. Each of these fields is a structure with three fields, .header, .record and .trailer. The .record field is a cell array, where each cell contains an INT32 vector corresponding to the records in the table. The .trailer field is a 1x6 int32 array containing the 6 trailer words for the table. The .header field is a structure with four fields. This structure contains the actual contents of the 16-word header found in OP2 tables. The .name field is an 8-character array corresponding to the table name. This corresponds to the first 2 words in the header. The .fulltrailer field contains the 7-word trailer that immediately follows the table name. The last 6 words of this trailer correspond to the contents of the .trailer field. The .name2 field contains the contents of the next two words in the header. In most cases this should match the .name field, but it does not always do so. The .date field contains the last 5 words of the header, which is the date string. In many cases this is empty since it is not set in the OP2 file.

To properly interpret the table records and trailer, please visit the NASTRAN DMAP Programmer's Guide. The record data is read in as INT32. Since it can be real (float32), double, or character data, it will likely be necessary to recast the data into the appropriate datatype. READNAS can read in Output2 files created on platforms with different endian-ness from the platform on which it is running. READNAS will perform the necessary byte-swapping based on 4-byte words. The .endianswap field of the DIR structure specifies whether READNAS performed byte-swapping during the read. IMAT provides the function TYPECAST_OP2 to typecast data from INT32 to another data type, correctly handling any byte-swapping necessary.

IMAT provides several helper routines for processing some of tables in raw format. Please refer to the PARSEOP2TABLE_* functions for useful examples of how to process raw OP2 tables.

Examples

>> f=readnas('/users/testdata/output.op2','blocks',{'BOUGV1','OEF1'});
>> s=readnas('*.dat','silent');

>> opt = [];
>> opt.global.atnodes = true;
>> f=readnas('filename.op2',opt);

See Also

expandDMI, readodb, readadf, readunv, setunits, getunits, parseop4, parseop2table_dmig, parseop2table_sets, typecast_op2