An Optimally Convergent Coupling Approach for Interface Problems Approximated with Higher-Order Finite Elements
For computational scientists solving interface problems, this method overcomes geometric variational crimes and achieves optimal accuracy with non-matching meshes.
This paper introduces a virtual interface coupling method for interface problems that achieves optimal convergence rates in H^1 and L^2 norms using higher-order finite elements on polytopial meshes, without requiring matching interfaces.
In this paper, we present a new numerical method for determining the numerical solution of interface problems to optimal accuracy with respect to the polynomial order of the Lagrangian finite element space on polytopial meshes. We introduce the notion of a virtual interface, and on this virtual interface we enforce that "extended" interface conditions are satisfied in the sense of a Dirichlet--Neumann coupling. The virtual interface framework serves to bypass geometric variational crimes incurred by the classical finite element method. Further, this approach does not require that the geometric interfaces are spatially matching. Our analysis indicates that this approach is well--posed and optimally convergent in $H^1$. Numerical experiments indicate that optimal $H^1$ and $L^2$ convergence is achieved.