Context Navigation

Posts for the month of May 2009

Spatial convergence with a smaller drop

The radius of the drop in the following simulations is 4.3 $\times 10^{-7}$ m and $\Delta x=1 \times 10^{-8}$ m. The previous simulations had drop with radius 1.1 $\times 10^{-6}$ m and $\Delta x=1.1 \times 10^{-8}$ m.

ID	Machine		scheme
104	alice (4) stopped	150	central difference
105	renfield (4) 10589	300	central difference
106	luggage (32) 40986	600	central difference
107	benson (4) 15976	150	power law
108	poole (4) 15934	300	power law
109		600	power law
110	sancho (4) stopped	150	van Leer
111	alice (4) 11750	300	van Leer
112	benson (4) 19216	150	QUICK

Posted: 2009-05-12 15:58 (Updated: 2009-05-18 17:45)
Author: wd15
Categories: (none)
Comments (0)

Spatial Convergence

The Figure below shows triple junction position against time for thesimulations 93, 99, 100, 101, 102, 103. What is clear is that they don't all agree. In and off itself this is not a huge problem. As long as we can show that for some grid resolution their is convergence. If under resolved calculations spread in a similar manner and have the same trends we can get away with running parameter search-type simulations at under resolved conditions. Simulation 102 is running with a 900 x 900 grid. I think the drop size can be reduced for these convergence calculations. It is fairly clear that the weird spurious velocities at the triple junction go away when the van Leer or power law schemes are used.