Open pscrozi opened 7 months ago
Can we do "validation" vs an LES model? Can we use Yulia's test case?
We want to build confidence in the key algorithms we've coded into ExaWind. Look at the core algorithm and combined cases. Build up to more realistic conditions like the Pazy wing from Kevin.
Phil had dropped a pitching and plunging plate (JFM paper) in the slack channel. This was an experiment that had been done. Phil doesn't have a lot of comments other than sharing it with us. Kevin voiced concerns about scope explosion, but Phil clarified that this was just sharing, and not necessarily wanting us to simulate this. Does the Pazy cover our bases, or do we need more? Phil is advocating for 3 tiers for verification testing: (1) low level fundamentals (Julia Pete's work) (cheap to run with lots of simulations); (2) turbulence cases --- Pazy wing --- grid refinement, etc.; (3) full turbine (e.g. IEA 15 MW, ~6 simulations). This is still a lot of scope. Aiming to prove that the full FSI simulation capability is working, and what would be expected in a publication.
We're done with this. We will move forward with the 3 cases outlined above.
Kevin is working to redistribute some of the work with Neil, and in coordination with the ML work.