12.07.2015 Views

The Boundary Element Method for the Helmholtz Equation ... - FEI VÅ B

The Boundary Element Method for the Helmholtz Equation ... - FEI VÅ B

The Boundary Element Method for the Helmholtz Equation ... - FEI VÅ B

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

66 4 Discretization and Numerical Realization4.4.3 Single Layer Potential OperatorTo compute <strong>the</strong> matrices arising from <strong>the</strong> Galerkin equations we need to evaluate <strong>the</strong>corresponding double surface integrals. In <strong>the</strong> following three sections we provide analyticevaluations of <strong>the</strong> singular parts of <strong>the</strong> inner integrals related to <strong>the</strong> single layer potentialmatrix V κ,h , <strong>the</strong> double layer potential matrix K κ,h and <strong>the</strong> hypersingular operator matrixD κ,h . <strong>The</strong> outer integrals can be computed using <strong>the</strong> 7-point quadrature rule discussed inSection 4.4.1.<strong>The</strong> analytic <strong>for</strong>mulae corresponding to <strong>the</strong> single layer potential operator and <strong>the</strong>double layer potential operator are given in [17], Section C.2, and we only provide <strong>the</strong> finalresults with some discussion. For <strong>the</strong> hypersingular operator we provide a more detailedcomputation.<strong>The</strong> entries of <strong>the</strong> single layer operator matrix are given byV κ,h [k, l] := 1 e4πτ kτ iκ∥x−y∥l∥x − y∥ ds y ds x= 14πτ kFor <strong>the</strong> latter integrand we have1τ l∥x − y∥ ds y ds x + 14πτ ke iκ∥x−y∥ − 1τ l∥x − y∥e iκ∥x−y∥ − 1 e iκa − 1 llim= lim′ H = lim iκe iκa = iκx→y ∥x − y∥ a→0 a a→0ds y ds x .and thus <strong>the</strong> integral is not singular and can be evaluated numerically. For <strong>the</strong> remainingsingular part of <strong>the</strong> integral we use an analytic computation based on <strong>the</strong> local coordinatesystem introduced in Section 4.4.2 combined with a quadrature scheme. For <strong>the</strong> innerintegral we obtain (see [17], Section C.1.2.)S V (τ, x) := 1 14π τ ∥x − y∥ ds y = 1 sτ α2 s1 dt ds4π 0 α 1 s (s − sx ) 2 + (t − t x ) 2 + u 2 x= 1 sτ ln t − t x + α2(s − s x )4π2 + (t − t x ) 2 + u 2 sx ds0α 1 s= 1 F V (s τ , α 2 ) − F V (0, α 2 ) − F V (s τ , α 1 ) + F V (0, α 1 ) ,4πwhereF V (s, α) := (s − s x ) ln αs − t x + (s − s x ) 2 + (αs − t x ) 2 + u 2 x − s+ αs x − t√ x ln 1 + α 2 (s − p) + (s − s x ) 2 + (αs − t x ) 2 + u 21 + α 2 x (sq − αsx−tx1+α −+ 2u x arctan2 sx ) 2 + (αs − t x ) 2 + u 2 x + (αs − t x − q)q(s − p)u x(4.25)

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!