Summary of results
======================

The results are very close to the analytical solution (error less than 1.0E-05) except for two values (N to :math:`\theta \mathrm{=}90°` and Vy to :math:`\theta \mathrm{=}0°`) for which the error is 2.0E-03. We notice that this error decreases slowly when we refine the mesh, but the precision of 1.0E-05 is still not achieved with 1000 elements per quarter circle.