I test the proficiency of SPH method dealing with a benchmark numerical example solving Poisson's equation; SPH shows the good efficiency in modeling the contours, but the value of results change for a different nodal distributions (different nodal distance) , in fact; although, the ratio of the results is nice, the value of the results has a noticeable different comparing to exact solutions. In addition, SPH results are severely affected by the radial of support domain.
Please guide me, how determine the logical and optimal radial for support domain.
