I am currently writing my final thesis which consists of 2D ldc and I am wondering how does PYFR handle 2D problems. In the PYFR paper, does it reduce the dimension of the matrices or just set to 0 all the variables in z dimension?
Yes, PyFR can handle 2D as well as 3D problems. The dimensionality is set from the mesh you use. For example, if you are generating a mesh in pointwise you’ll need to set the CAE dimension to 2D.
The operator matrices are defined in a very abstract way, and only really care about the points sets and basis that you use. So, no the 3D terms aren’t zeroed, but rather the 2D version of the matrices are constructed. If you haven’t already found the file, the matrices are defined in shapes.py. It is fairly straightforward to modify the functions there to print out the matrices as they are generated. The matrices are only generated when needed so, you might need to play with the ini file if you are after a specific matrix.