3+1 formalism and bases of numerical relativity - LUTh ...
3+1 formalism and bases of numerical relativity - LUTh ...
3+1 formalism and bases of numerical relativity - LUTh ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
9.2 Choice <strong>of</strong> foliation 159<br />
type ∂N/∂λ in the right-h<strong>and</strong> side <strong>and</strong> to compute the “λ-evolution” for some range <strong>of</strong> the<br />
parameter λ. This amounts to resolve a heat like equation. Generically the solution converges<br />
towards a stationary one, so that ∂N/∂λ → 0 <strong>and</strong> the original elliptic equation (9.15) is solved.<br />
The approximate maximal slicing has been used by Shibata, Uryu <strong>and</strong> Taniguchi to compute<br />
the merger <strong>of</strong> binary neutron stars [226, 239, 240, 237, 238, 236], as well as by Shibata <strong>and</strong><br />
Sekiguchi for 2D (axisymmetric) gravitational collapses [227, 228, 221] or 3D ones [234].<br />
9.2.3 Harmonic slicing<br />
Another important category <strong>of</strong> time slicing is deduced from the st<strong>and</strong>ard harmonic or De<br />
Donder condition for the spacetime coordinates (x α ):<br />
gx α = 0, (9.23)<br />
where g := ∇µ∇ µ is the d’Alembertian associated with the metric g <strong>and</strong> each coordinate<br />
x α is considered as a scalar field on M. Harmonic coordinates have been introduced by De<br />
Donder in 1921 [106] <strong>and</strong> have played an important role in theoretical developments, notably in<br />
Choquet-Bruhat’s demonstration (1952, [127]) <strong>of</strong> the well-posedness <strong>of</strong> the Cauchy problem for<br />
<strong>3+1</strong> Einstein equations (cf. Sec. 4.4.4).<br />
The harmonic slicing is defined by requiring that the harmonic condition holds for the<br />
x 0 = t coordinate, but not necessarily for the other coordinates, leaving the freedom to choose<br />
any coordinate (x i ) in each hypersurface Σt:<br />
gt = 0 . (9.24)<br />
Using the st<strong>and</strong>ard expression for the d’Alembertian, we get<br />
1 ∂<br />
√<br />
−g ∂x µ<br />
<br />
√−gg µν ∂t<br />
∂xν <br />
=δ0 <br />
= 0, (9.25)<br />
ν<br />
i.e.<br />
∂<br />
∂x µ<br />
√ µ0<br />
−gg = 0. (9.26)<br />
Thanks to the relation √ −g = N √ γ [Eq. (4.55)], this equation becomes<br />
∂ √ 00<br />
N γg<br />
∂t<br />
+ ∂<br />
∂xi √ i0<br />
N γg = 0. (9.27)<br />
From the expression <strong>of</strong> g αβ given by Eq. (4.49), g 00 = −1/N 2 <strong>and</strong> g i0 = β i /N 2 . Thus<br />
Exp<strong>and</strong>ing <strong>and</strong> reordering gives<br />
∂N<br />
∂t<br />
− ∂<br />
∂t<br />
√ γ<br />
− βi∂N − N<br />
∂xi N<br />
<br />
+ ∂<br />
∂x i<br />
√ γ<br />
N βi<br />
<br />
1 ∂<br />
√γ<br />
√ γ 1<br />
−<br />
∂t<br />
<br />
= 0. (9.28)<br />
∂<br />
√<br />
γ ∂xi √ i<br />
γβ <br />
<br />
=Diβi <br />
= 0. (9.29)