Interface kernels are meant to assist in coupling different physics across sub-domains. The most straightforward example is the case in which one wants to set the flux of a specie A in subdomain 0 equal to the flux of a specie B in subdomain 1 at the boundary between subdomains 0 and 1. In mathematical terms, we might be interested in establishing the condition:

$$$$-D_0 \frac{\partial c_0}{\partial x} = -D_1 \frac{\partial c_1}{\partial x}$$$$

where $$$D_i$$$ is the diffusion coefficient of specie $$$i$$$ in subdomain $$$i$$$, and $$$c_i$$$ is the concentration of specie $$$i$$$ in subdomain $$$i$$$. An example of this condition is shown in the Moose test directory; see files below:




Interface kernels can be used to provide any general flux condition at an interface, and even more generally can be used to impose any interfacial condition that requires access to values of different variables and gradients of different variables on either side of an interface. In an input file, the user will specify at a minimum the following parameters:

type: The type of interface kernel to be used
variable: This is the "master" variable. Note that the master variable must exist on the same subdomain as the sideset specified in the boundary parameter. The existence of a "master" and "slave" or "neighbor" variable ensures that the interface kernel residual and jacobian functions get called the correct number of times. variable could be $$$c_0$$$ from our example above.
neighbor_var: The "slave" variable. This could be $$$c_1$$$ from our example above.
boundary: The interfacial boundary between the subdomains. Note that this must be a sideset and again must exist on the same subdomain as the master variable. The fact that this boundary is a sideset allows access to variable gradients.

For additional information about the interface kernel system, don't hesitate to email the moose list at