Current Research 09.11.2006  --  2nd Run Iteration Sequence

This is a visual representation of the sequence of iterations performed using feature aligned embedded mesh surfaces to perform solution adaptive meshing (discussed in my current research section). The page is titled "2nd Run" because this is the second time this process has been performed. The "1st Run" was performed on a different domain which hindered the flow solution at various iterations to not converge fully. The pictures below show the NACA 0012 airfoil with each iterations respective maximum tangential velocity extent surface. This is the surface that was used to then refine around to produce the volume grid for the next iteration. As you can see, the max tangential velocity extent surface is quite crude for iteration 1. This is because the initial volume grid spacing is very coarse aft of the wing, therefore giving poor solution resolution. The only a priori refinement was focused on the surface of the wing and the wing's boundary layer only. As the iterations progress, the max tangential velocity surface gets quite tight and focuses in on where the vortex should be. At the bottom of the page is a graph of maximum tangential velocity vs. radius for each iteration compared to experimental data. From this, one can determine that iteration 3 yielded the best result and the subsequent iteration got worse while the one after that, iteration 5, only improved from iteration 4 slightly. This answer why this anomaly occurs is still unknown. Intuitively, as the iterations progress the solutions should improve and then level off/converge. In this instance, this is not the case and time is being put in to determine why this is happening.

Iteration 1

Iteration 2

Iteration 3

Iteration 4

Iteration 5

Tangential Velocity vs. Radius Comparison