Changed projection point in time to before sampling. Good results similar to original paper are achieved.

Added distFromSol as member variable and public functions for accessing the final iteration, change, residual and number of points in the NB.

Updating DCPSE

See merge request openfpm/openfpm_numerics!13

Deleted options for higher order upwind gradient from the Sussman redistancing because I cannot show numerical convergence for these.