Nastran Punch Reader (pch2csv)

Description

A simple tool to convert Nastran Punch format into CSV.

It is written in C++ and uses the C++ standard library.

History

The Nastran Punch Format inherits from the punched cards.

• A punch card:

https://en.wikipedia.org/wiki/Punched_card_input/output https://en.wikipedia.org/wiki/Punched_cards

Features

The tool reads a Nastran Punch file (*.pch) and converts its scary format into Comma-Separated Values (*.csv) readable by Calc or Excel.

Basically, it converts this:

$TITLE   = MY FEA MODEL                                                        1
$SUBTITLE=MY FIRST LOAD CASE                                                   2
$LABEL   =MY FIRST LOAD CASE                                                   3
$ELEMENT STRESSES                                                              4
$REAL OUTPUT                                                                   5
$SUBCASE ID =         666                                                      6
     12345          80004230        BAR                                        7
-CONT-                  2.288704E+04     -3.404367E+03      1.639255E+03       8
-CONT-                 -7.163730E+04      9.975631E+05      3.060709E+06       9
     12345          80004231        BAR                                       10
-CONT-                 -2.301775E+04     -3.107557E+03      4.195733E+02      11
-CONT-                  7.232352E+04     -9.979151E+05     -3.062225E+06      12

into this:

"TITLE";"SUBTITLE";"LABEL";"SUBCASE ID";unknown;unknown;unknown;unknown;unknown;unknown;unknown;unknown;unknown;unknown;
"MY FEA MODEL";"MY FIRST LOAD CASE";"MY FIRST LOAD CASE";"666";12345;80004230;BAR;;2.288704E+04;-3.404367E+03;1.639255E+03;-7.163730E+04;9.975631E+05;3.060709E+06;
"MY FEA MODEL";"MY FIRST LOAD CASE";"MY FIRST LOAD CASE";"666";12345;80004231;BAR;;-2.301775E+04;-3.107557E+03;4.195733E+02;7.232352E+04;-9.979151E+05;-3.062225E+06;

Remark:

The Nastran solver produces a PUNCH file as result of a FEA when the PUNCH keyword replaces the default PRINT or PLOT output option, which is defined in the Finite Element model CASE CONTROL section.

For instance, Nastran creates a PUNCH file containing all the GRIDs displacement if the FE model file contains the line DISPLACEMENT(SORT1,PUNCH,REAL)=ALL in its CASE CONTROL section.

Punch are sometimes more 'readable' than the F06 (see PRINT default option).

Remark 2:

Punch formats aren't that bad, however several types of elements (with different formats) can be stored inside (and also totals...), so it can sometimes be tricky to convert them easily into a single 2-dimensional table.

The work-around found here is to output as many CSV as element / totals types there are. It's not ideal but this allows a better reorganisation of the data.

In such case, the command:

$ ./pch2csv input.pch -o output.csv
Warning: pch2csv detected 4 different formats.
Then, 4 files are produced.

will produce these files:

output_format_0.csv
output_format_1.csv
output_format_2.csv
output_format_3.csv

Note that the tool emits warnings when such case appears. However this 'auto-split' feature can be disabled with -u.

Under the hood

Simple case

The simplest case is given below. One punch file containing one format of data, is converted into an unique CSV.

Concatenate several punch files into an unique output

If several files are proceeded (i.e. with the command pch2csv file1 file2 file3), the tool concatenates them into an unique CSV if they only contain one format of data. If not (i.e. when several formats are detected), pch2scv outputs one CSV file per format, except if the option -u is used.

Split formats to separated files

If the -u flag isn't specified, the tool writes each format from Punch files to a separated CSV.

Build

1. Download pch2csv.tar.gz.

2. Unzip:

    $ tar zxvf pch2csv.tar.gz
    $ cd pch2csv/
   

3. Build with CMake

   • On Unix / Mac OS X

      $ mkdir -p build
      $ cd build/
      $ cmake ..
      $ make
      $ make install
     

   • On Windows (MinGW)

      > mkdir build
      > cd build/
      > cmake .. -G "MinGW Makefiles"
      > make
     

   • On Windows (MSVC)

      > mkdir build
      > cd build/
      > cmake .. -G "Visual Studio 14 2015"
      > start .
     

     Then, double click the Visual Studio project (vcxproj).

Usage

On Unix / Mac OS X

$ ./pch2csv input.pch -o output.csv

On Windows

> pch2csv.exe input.pch -o output.csv

Remarks:

• By default, CSV aren't overwritten. Instead, pch2csv increments the former CSV (backup).
• Additionally, pch2csv writes user-defined column titles in CSV with -c option.

Options:

• -h, --help
  Show the help message and exit

• -v, --version
  Show the version

• -o OUTPUT, --output=OUTPUT
  Specify the name of the output file. By default, the output takes the same name as the input, with .csv extension

• -c %HEADER%, --column-header=%HEADER%
  Specify the csv header content. By default, the column headers are unknown. %HEADER% must be a sequence of words separated by commas, surrounded by a double-quote, for instance: -c "CBUSH ID;;CS;X in mm;Y [mm];CSout"

• -s, --skip-header
  Do not print the csv header. Data begins at the first line.

• -u, --unique
  Force the tool to produce an unique csv, even if several formats are detected.

Similar work from Github's Community

At the time of writing these lines, some interesting things about Nastran and PUNCH files can be found on Github:

• The nastran_pch_reader by anick107, a Python Parser for PUNCH files: https://github.com/anick107/nastran_pch_reader

• The pyNastran by SteveDoyle2, a Python-based interface tool for Nastran's file formats: https://github.com/SteveDoyle2/pyNastran (in particular, see pyNastran/bdf/test/unit/test_read_write.py)

Punch files found in those projects are used here, mainly for demo and unit testing.

License

The code is released under the GNU LGPLv3 open source license.

https://www.gnu.org/licenses/lgpl-3.0.en.html