Apply evolution to the moving boundaries. To use this method (which will be available only for evolution problems), select Enable evolution on moving boundaries in the Numerical parameters menu or in the Global remeshing menu.

This technique progressively introduces the kinematic condition into the system of equations being solved over the course of multiple evolution steps. When the evolution parameter is equal to 0, the kinematic condition is ignored. When reaches a maximum value of 1, the full kinematic condition is taken into account.

This method has been shown to work well for most flow problems (generalized Newtonian, viscoelastic, variations of shape, etc.), except those involving a recirculation on the free surface, as in coating applications. For such cases, you should switch to a time-dependent calculation.

When guiding devices are invoked, you should simultaneously apply evolution on moving boundaries (see above), as well as on gravity. Also, it is a good idea to replace the zero velocity along the die wall by a Navier’s slipping law and to apply an evolution on the slipping coefficient.

Replace all free-surface conditions with a free-slip condition (that is, , ). The normal force will then not correspond to the desired value, but all problems linked to position nonlinearity will be eliminated and the solution will be a good starting point for the desired free-surface-problem solution. Use this solution as an initial guess for your free-surface problem. Alternatively, you can use evolution on moving boundaries, as described in the previous section.

For steady 2D cases, it can be a good idea to enable upwinding on the free surfaces and moving interfaces.

Variations of shape can occur due to important nonuniformities of the velocity at the die outlet. If convergence difficulties arise in solving a steady-state problem without surface tension, it is recommended that you replace the zero-velocity boundary condition along walls adjacent to free surfaces by a slip condition. Free slip occurs when the slip coefficient is equal to zero, and the velocity is equal to zero when is very large. If the problem converges for a low value of , use an evolution scheme to increment . In order to estimate the value of , you can perform a 2D channel flow simulation with the appropriate material properties, and then define so that the value of the wall velocity is 10 to 30% of the average fluid velocity. In many practical situations, nonzero slip along the die wall predicts more realistic extrudates.

See Slip Condition for information about defining slip conditions, and Evolution for information about evolution.

Replace all moving-interface conditions with , conditions on both sides of the interface. If Ansys Polyflow converges, then start the moving-interface calculation from the converged fixed-interface solution. If Ansys Polyflow does not converge, the problem is not caused by the moving boundaries; you should look elsewhere to determine the cause of the convergence trouble. If the fixed-interface problem requires an evolution scheme to converge, start the moving-interface calculation from the converged fixed-interface solution obtained upon completion of the evolution.

For small capillary numbers (that is, when capillary forces are large compared to viscous forces), use quadratic coordinates and quadratic velocities for 2D, or linear coordinates and the mini-element for 3D. The capillary number is defined by Equation 15–23.

An infinite value of corresponds to a straight free surface (or moving interface). If this is acceptable from the point of view of mesh deformation, start the calculation with a straight mesh and a large value of , and decrease it using an evolution scheme.

For integral viscoelastic problems, decouple the free-surface updates from the velocity and pressure calculations. See Calculations Involving Moving Boundaries for details.