BoundaryElementMethode

BoundaryElementMethode, known in English as the Boundary Element Method (BEM), is a numerical technique for solving linear boundary value problems. It reformulates domain PDEs as boundary integral equations using fundamental solutions, so the unknowns reside only on the boundary. This leads to a reduction in dimensionality: a volume problem becomes a surface problem in three dimensions, or a area problem becomes a line problem in two dimensions.

The method rests on Green’s representation formula or similar integral representations for the governing operator (for example Laplace, Helmholtz, or linear elasticity). By expressing the solution in terms of boundary values and their normal derivatives, BEM converts the original PDE into equations defined solely on the boundary. This makes BEM particularly well suited to exterior problems and to domains with complex or infinite extents, since radiation and Sommerfeld-type conditions are naturally handled.

Discretization involves partitioning the boundary into elements and approximating the boundary fields with shape functions. The resulting system is a dense linear matrix equation, in contrast to the sparse matrices typical of domain methods. Special techniques are required to handle singular and nearly singular integrals that arise from boundary integrals. Practical implementations include collocation and Galerkin formulations, and sometimes hybrid BEM-FEM schemes.

Applications span acoustics, elastodynamics, potential problems, electrostatics, and fluid mechanics, especially where unbounded domains or sharp boundary geometries are important. Advantages include reduced dimensionality and accurate boundary representation, while limitations involve the need for linear governing equations with known fundamental solutions, dense system matrices, and challenges for highly inhomogeneous or nonlinear media.