proved to be capable to simulate the Kelvin Helmholtz instabilities. Visual
aspect and most numerical results are matching experimental and theoretical
results. It can be used as a powerful tool to understand the behavior of
this complex phenomenon. Yet the bad stability prediction is annoying (although
it may be improved) as it will amplify numerical and discretization errors.
This problem also demonstrates the difficulty of representing numerically
an infinitesimal perturbation.
Moreover this study has enabled
us to deepen our knowledge of Fluent by using new features like the UDFs.
Particularly interesting was to to develop two viewpoints of the same problem
(each model having both advantages and downsides). This showed us that
there may not always be one solution, but an array of possibilities to
be used depending on the studied parameter.