Computationally-efficient CFD solver with user interface written in Python.

• Very efficient solver for unsteady lid-Driven cavity flow for problems with grid-sizes of up to 256x256 nodes.

• The pressure (Poisson) equation is efficiently solved using a 2-dimensional cosine transform based on the solver written by Michio Inoue

• The unsteady solver automatically stores velocity and pressure fields for post-processing

• The solver can compute 100 convective times ($n_t=100$) @ $Re=10^3$, in a grid of 256x256 nodes, in less than 1 second! (if output results are not stored)

The software contains a comprehensive UI that allows for parameters input and to pre-visualization of domain and grid sizing.

• Firstly, the user needs to input the following flow parameters:

  • nx, ny: number of grid elements in x- and y-directions
  • lx, ly: domain sizes in x- and y-directions
  • nt: number of time-steps
  • Co: target simulation's Coarant number (see )
  • Re: target flow's Reynolds number. Reynolds number is based on cavity's x-dimension.

• Secondly, the user has to select a directory to save the files under the Run console

• Finally, the simulation can be started by clickin on "Run".

For pip:

• Mac or Linux: install with sudo easy_install pip (Mac or Linux) - or (Linux) find a package called 'python-pip' in your package manager.
• Windows: install Distribute then Pip using the linked .MSI installers.

For git:

• Mac: install Homebrew first, then brew install git.
• Windows or Linux: see Installing Git from the Pro Git book.

• Selection of time-steps to save the data at is missing (currently, the software stores data at every 10 time-steps only).
• Real-time plotting of velocity and pressure fields (avoided at the current stage of development for improved computational performance).
• The provided example of plotting function does not work for grids of $n_x \neq n_y$