I'm trying to simulate a counter-current packed bed using the DPM, EWF and coupled with Species Transport with the transient (time-step for bulk= 0.001)pressure-based solver. Basically, moist air is injected at the inlet and on the other side droplets are being released. The droplet are gonna impinge the geometry and form a film. I'm solving both energy and momentum equation as I need to model the mass transfer from the film to the bulk. I have problem with the film convergence as soon as I activate the "pressure gradient" terms in the EWF dialog box. I'm using the first order implicit. When the pressure gradient terms are deactivated everything is going fine and 10 sub-iterations are enough to converged.Typical Residual line WITHOUT pressure gradient terms: Everything is going fine no divergence is detected.
Then as soon as the pressure gradient are being modeled even with a smaller time-step. At the begining everything is going just fine and every iteration converge then for some time step it starts alternating with one time step fully converged and the following one is not:
Finally after several time step the residuals are "converging to a diverging" value:
It looks like each residual as a fix value and increasing the number of sub-iteration would be pointless. Secondly the max_cfl value looks good because it should be lower than 0.01. I tried lowering the time steps even more but it doesn't help at all. Same with switching to second order on continuity, momentum and energy. Could it be the mesh ? I already have a mesh 300,000 cells (Quality is already good when checking orthogonality and skewness) in a 30*50 mm computational domain with inflation on the impinging surfaces. Anyone has an idea of what's going on and how to solve the problem ?