Posts for the month of January 2009

Working 2D simulations with whole matrix method (unsegregated solver)

Finally have working 2D simulations with the whole matrix method. Simulation source:trunk/reactiveWetting/65 and source:trunk/reactiveWetting/66 both have $\mu=2\times10^{-3}$, but $\bar{M}=1\times10^{-6}$ and $1\times10^{-7}$, respectively. Movies are in the source directories for velocity, density, concentration and phase. The sharp interface simple analytic solution is plotted in the images. It is questionable whether the results have reached a final equilibrium, there is a sift from the analytic curve, especially for simulation 65. The diffusion is an order of magnitude slower in this case. It is probable that this simulation has to run for longer in order to see convergence to the equilibrium result. The final equilibrium shapes for simulation 65 and 66 are in source:trunk/reactiveWetting/65/conc0005400.png and source:trunk/reactiveWetting/66/conc0005400.png, respectively.

Convergence

Figures source:trunk/reactiveWetting/65/velocity.png and source:trunk/reactiveWetting/66/velocity.png show the maximum velocity against time. Neither simulation has completely converged, it may be that the drops still need to adjust to equilibrium or the velocities could be parasitic in nature.

Next steps

  • Run a simulation with $\bar{M}=1\times10^{-4}$. This may converge a little better. (67)
  • Run a simulation with $\bar{M}=1\times10^{-5}$ but for a grid of 200 * 200 (68)
  • Profile for speed and adjust.
  • axi-symmetric drops.

  • Posted: 2009-01-16 17:17 (Updated: 2009-04-29 16:04)
  • Author: wd15
  • Categories: (none)
  • Comments (0)