24.02.2013 Views

Read Back Signals in Magnetic Recording - Research Group Fidler

Read Back Signals in Magnetic Recording - Research Group Fidler

Read Back Signals in Magnetic Recording - Research Group Fidler

SHOW MORE
SHOW LESS

Create successful ePaper yourself

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

Numerical Methods<br />

current flow<strong>in</strong>g out of the volume. Normally the Neumann boundary conditions are given<br />

either for each boundary triangle<br />

j() r = j , ∀r ∈Δ<br />

(4.18)<br />

ijk ijk<br />

or for each boundary node of the mesh<br />

j( r) = j , ∀r ∈∂V<br />

. (4.19)<br />

i i i<br />

For the first case we get for the boundary <strong>in</strong>tegral term <strong>in</strong> (4.15)<br />

1<br />

σ() r ∇u() r ϕ () r dA=− j() r ϕ () r dA=−∑ j ϕ () r dA=− ∑A<br />

j<br />

3<br />

�∫ � ∫ ∫<br />

. (4.20)<br />

i i ijk i ijk ijk<br />

∂V ∂V Δijk Δijk<br />

j< k<br />

Δijk<br />

j< k<br />

For the second case we get for the current j by l<strong>in</strong>ear <strong>in</strong>terpolation between the sampl<strong>in</strong>g<br />

po<strong>in</strong>ts (po<strong>in</strong>ts at the boundary) us<strong>in</strong>g the test functions i ϕ<br />

∑<br />

j() r = j ⋅ϕ () r . (4.21)<br />

i<br />

i i<br />

Then the result of the boundary <strong>in</strong>tegral is<br />

�∫ �∫<br />

σ() r ∇u() r ϕ () r dA=− j() r ϕ () r dA=<br />

i i<br />

∂V ∂V<br />

1<br />

−∑ ∫ ( jiϕ i() r + jjϕ j() r + jkϕk() r ) ϕ i() r dA=− ∑Aijk<br />

( 2 ji + jj + jk)<br />

.<br />

12<br />

Δijk Δ<br />

Δ<br />

ijk<br />

ijk<br />

j< k j< k<br />

4.1.3 L<strong>in</strong>ear equation solver<br />

The FEM method leads to a system of l<strong>in</strong>ear equations (4.16)<br />

(4.22)<br />

Au ij j = ri<br />

⇔ Au ⋅ = r (4.23)<br />

with a sparse symmetric and positive def<strong>in</strong>ite matrix A. To solve this system of l<strong>in</strong>ear<br />

equations a Cholesky factorization method for sparse matrices seemed to be practicable [23].<br />

The advantage of this method over iterative methods is the <strong>in</strong>sensibility of “bad” meshes. The<br />

convergence fails with iterative methods if the tetrahedrons show much difference from<br />

regular tetrahedrons, so that they have small angles. This is problematic for meshes of th<strong>in</strong><br />

layers, as used <strong>in</strong> GMR elements.<br />

45

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

Saved successfully!

Ooh no, something went wrong!