Implement target system

[0mar]

Target system as described in results/ell-ell.tex has the following elements:

• [x] multiscale
  • [x] domain mapped
• [x] mixed boundary conditions
• [x] system of equations

After implementation of these, we can benchmark and model some toy problems

[0mar]

In 2e32fa3 we have convergence, but not 2nd order yet. Further inspection into the problem gives the following observations:

• Isolating microscale and halving the grid size give subsequent error reductions: 3.39, 3.69 (4 would be second order convergence)

• Isolating macroscale and halving grid size gives: 4.04, 4.02. One might presume from this that there is some slight error in the microscale.

Even further examining the microscale error shows that the error is localized in one point. Perhaps some boundary issues?

[0mar]

Closed in 8801852