Discussion Forum

I'm attempting to model glacial flow, which has been performed by a few others. I've been able to set up a 2D glacier that deforms how it should using a nonlinear viscosity (eta_nl), but I'm having a hard time figuring out how to put in sliding of the glacier. Ideally, I would like to specify slip at the bed, where u_sliding = C*tau_xz^n, where C and n are constants and tau_xz=eta_nl*(uz+wx) and this is evaluated only at the basal boundary. This is all based on accepted glacial flow theories, though I would like to have the ability to change the sliding law around to test some other ideas.
For now, is there some accepted way to force a velocity at the base that is parallel to that boundary? I've tried using an inlet velocity boundary condition where I define uo and wo as the portions of the sliding that would parallel to the flow (I have functions that define the slope at the base and slope at the surface). I've also attempted to use a sliding wall boundary condition that has the the velocity defined above except with the opposite sign. I have not been able to achieve convergence with either of these for realistic values (the deforming flow does what I expect, however). Are there any suggestions or anything obvious that I might be missing?