Initializing the Phase Field Function
If the study type Transient with Phase Initialization is used in the model, the phase field variable is first initialized so that it varies smoothly between zero and one over the interface. For this study, two study steps are created, Phase Initialization and Time Dependent. The Phase Initialization step solves for the distance to the initial interface, Dwi. The Time Dependent step then uses the initial condition for the phase field function according to the following expression:
in Fluid 1 and
in Fluid 2. These expressions are based on a steady, analytic solution to Equation 11-10 and Equation 11-11 for a straight, non-moving interface.
If the Transient with Phase Initialization () study is being used, for the initialization to work it is crucial that there are two Initial Values nodes and one Initial Interface node. One of the Initial Values nodes is set to Specify phase: Fluid 1 and the other to Specify phase: Fluid 2. The Initial Interface node should have all interior boundaries where the interface is initially present as selection. If the selection of the Initial Interface node is empty, the initialization fails.
Studies and Solvers in the COMSOL Multiphysics Reference Manual