A finite element method for beam-like structures.
Author: S. Andrew Ning
C++ compiler, Boost C++ Libraries, LAPACK, NumPy
These instructions assume you are using MinGW and have already installed gcc and g++. Also you should already have successfully installed Python (for a single user only), NumPy, and setuptools. The example directories may need to be modified depending on where you installed things. See this Windows guideline set for additional support on installing python.
-
Edit (or create) a distutils config 'distutils.cfg' file in your Python directory.
C:\Python27\Lib\distutils\distutils.cfg
and put the following in it:
[build]
compiler=mingw32
- Download Boost (v 1.55 as of this writing) and setup bjam
At the command prompt:
> cd boost_1_55_0\tools\build\v2\engine
> build.bat mingw
This should create a folder called: bin.ntx86. For convenience in the next step you can add this folder to your PATH so bjam is accessible. Otherwise, use the whole path when calling bjam.
C:\boost_1_55_0\tools\build\v2\engine\bin.ntx86
- Compile Boost.Python
In the boost root directory (must be in the root directory) type the following at the command prompt:
> bjam toolset=gcc --with-python link=shared
the libraries should be built in stage/lib and will be needed in steps 5 and 6.
-
Install LAPACK and BLAS. I just used prebuilt libraries. Make sure to grab all three libraries - BLAS, LAPACK and LAPACKE and make sure they are the 32-bit versions. Remember the location for steps 5 and 6.
-
Make sure the following are on your system PATH. The dynamic libraries are needed in order to actually run pBEAM.
C:\Python27 (for Python) C:\Python27\Scripts (for easy_install) C:\MinGW\bin (for g++, gcc, etc.) C:\lapack (LAPACK dynamic libraries) C:\boost_1_55_0\stage\lib (Boost Python dynamic libraries)
-
Modify the 'setup.py' script in pBEAM's main directory. Unlike GCC on *nix systems, Windows does not have typical locations to store headers and libraries (e.g., /usr/local/include) and so you will need manually specify them. Add the header locations for Boost in the include_dirs. Add the library locations for Boost and LAPACK. You may also need to rename the boost_python library. Use the example below, modifying as needed based on where you installed things. Note that setup.py expects unix style slashes (forward), and that you do not need to include 'lib' at the front of the library names (i.e., 'lapack' corresponds to 'liblapack.dll' or 'liblapack.a'). Note: make sure your boost version matches the boost version installed (i.e. mgw48, mgw46, etc).
include_dirs=[join(path, 'pBEAM'), 'C:/boost_1_55_0'], library_dirs=['C:/boost_1_55_0/stage/lib', 'C:/lapack'], libraries=['boost_python-mgw48-mt-1_55', 'lapack']
For the remainder of the setup, use the below directions for *nix systems.
Install pBEAM with the following command.
$ python setup.py install
To check if installation was successful run Python from the command line
$ python
and import the module. If no errors are issued, then the installation was successful.
>>> import _pBEAM
pBEAM has a large range of unit tests, but they are only accessible through C++. They are intended to test the integrity of the underying code for development purposes, rather than the python interface. However, if you want to run the tests then change directory to src/twister/rotorstruc/pBEAM and run
$ make test CXX=g++
where the name of your C++ compiler should be inserted in the place of g++. The script will build the test executable and run all tests. The phrase "No errors detected" signifies that all the tests passed. You can remove the remove the test executable and all object files by running
$ make clean
For software issues please use https://github.com/WISDEM/pBEAM/issues. For functionality and theory related questions and comments please use the NWTC forum for Systems Engineering Software Questions.
Access the online version at http://wisdem.github.io/pBEAM/