Hi, I'm having a problem with comsol : I do not know if it come from a ill-posed physic or if it's a "numerical problem". Indeed, I use the creeping flow model on a rectangle with an inlet (p = 1) to the right, outlet (p=0) to the left and symmetry up and down. The result is given in the first picture below and it doesn't work. Then I consider the same geometry minus a circle (its center is on the lower border of the rectangle ; see the second attachment below) and with the same boundary condition : here the computation is done by comsol and the result seems consistent (cf. last attachment). I do not understand why it works in one case but not in the other. Do you have any idea where this may come from?
↧