General

How can I improve the convergence of calculations involving the use of porous media with a large pressure drop?

• FAQ
Participant

In calculations involving the use of porous media, the pressure tends to change rapidly in and around the porous regions of your model. Attempting to resolve these large pressure gradients through higher order discretization schemes such as second order may cause over and undershoots of the solution variables leading to poor convergence or even divergence of the calculation. We can take the following approach to improve convergence â€¢Patch the zones upstream and downstream of the porous zone with appropriate pressure values (better guess) for good convergence. â€¢Solve first with Porous zone turned OFF. Once the solution converges, turn ON porous zone. â€¢Do not set the resistance values (inertial and viscous) in the direction of no flow, to be more than 1000 times the resistance values in the direction of primary flow. â€¢Convergence is very sensitive to the grid resolution near the boundary b/w porous and non-porous zones. Ensure that there are no cell jumps near these boundaries. â€¢Start with first order discretization of pressure and momentum. Once first order solution converges, move to second order discretization. Use low URFs of pressure and momentum. â€¢Use PRESTO! or Body-Force-Weighted scheme instead of Standard for Pressure. PRESTO! is specifically formulated to handle cases with large body forces such as huge pressure drops in porous media. PRESTO! is recommended for Hex meshes. For Tet/Tri mesh, use Linear Pressure interpolation scheme. â€¢If the resistance coefficients are too high so as to cause huge pressure drop in the porous region, start with lower values of resistance coefficients and slowly increase them up. â€¢Try to use put hex-mesh in the porous medium as much as possible