Comparison of hydrocodes in Wengen test4

A study of influence of some numerical parameters on "F" run

modifier	description
-B2	computational domain doubled in size to [-16,16]³
-PC	predictor-corrector evaluation of gravity forces switched on
-Cx	CFL number reduced from 0.8 to x/10.
-AV	artificial viscosity enabled with default value of 0.1
-Xx	resolution limited to dx = 1/x
-Z1	dx = dy = dz
-IC256	initial condition file was test4-256

Any diffusivity weakens clumps.
Without predictor-corrector there is some braking of the clump, which forces the clump to lose its angular momentum and helps to gather more matter.
A slope change in "F-Z1" when refining to new level indicates somewhat insufficient resolution. All other models have doubled z-resolution and do not show such effect.
Influence of periodic boundary conditions compares in strength with influence of other numerical factors.
Limiting resolution to initial conditions file resolution (model "F-Z1-X64") severely suppresses clump formation.
Resolution criterium of all above models was J ≥ 8, and finest lrefinement level has cells no smaller than dx = 1/256. Each block on the following picture has 16×16 cells. Picture below shows midplane density slice (covering [-18.75,18.75]×[-18.75,18.75] AU) taken at t = 4.8 ≃ 38 yr, the most prominent clump is at 6 o'clock

Back to main page of Wengen test4 project.