To conclude this study several things should be pointed out :

  • First of all, to realize CFD with Navier-Stokes equations, you should dispose of a fast computer, or of parallel machines that are not overloaded.
  • You should be very careful to the equations used : it might seem nice to simplify your calculus by using Euler or perfect uncompressibl fluid equations, but are they fully usable ? These equations make a loss of data and might make your simulation crash.
  • Artificial viscosity is a great mean of having your simulation converge, but it does include a non-physic term in equations and therfore should be used carefully.
  • This project was a little bit to long for the time we had. We could not do many simulations on the airfoil A cases, and most simulations done with the naca12 profile had to be done by night.