06.08.2013 Views

Sequential Methods for Coupled Geomechanics and Multiphase Flow

Sequential Methods for Coupled Geomechanics and Multiphase Flow

Sequential Methods for Coupled Geomechanics and Multiphase Flow

SHOW MORE
SHOW LESS

Create successful ePaper yourself

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

92 CHAPTER 4. CONVERGENCE OF THE DRAINED AND UNDRAINED SPLITS<br />

errors of the <strong>for</strong>m ek Uj = γk eei(j)θ eU, ˆ ek Vj = γk eei(j)θ eV ˆ , <strong>and</strong> ek Pj = γk eei(j)θ eP, ˆ we obtain<br />

⎡<br />

⎢<br />

⎣<br />

γe −∆tγe 0<br />

∆tKdr<br />

ρbh 2 2(1 − cos θ)γe<br />

γe<br />

∆tb θ<br />

ρbh2i sin 2<br />

0 ∆tMb θ<br />

h 2i sin 2γe γe + ∆tMkp<br />

µh2 <br />

B<br />

2(1 − cos θ)γe<br />

<br />

∗ dr<br />

⎤⎡<br />

⎥⎢<br />

⎥⎢<br />

⎥⎢<br />

⎦⎣<br />

eU ˆ<br />

eV ˆ<br />

eP ˆ<br />

⎤<br />

⎥<br />

⎦ =<br />

⎡ ⎤<br />

0<br />

⎢ ⎥<br />

⎢ ⎥<br />

⎢ 0 ⎥<br />

⎣ ⎦<br />

0<br />

.<br />

(4.37)<br />

From det(B∗ dr ) = 0, the error amplification factors of the coupled flow <strong>and</strong> dynamics <strong>for</strong><br />

the drained split are given by<br />

γe = 0, −<br />

Equation 4.38 indicates that<br />

∆t 2 Mb 2<br />

ρbh2 2(1 − cos θ)<br />

<br />

1 + ∆t2Kdr ρbh2 <br />

2(1 − cos θ) 1 + ∆tMkp<br />

µh2 . (4.38)<br />

2(1 − cos θ)<br />

lim<br />

∆t→0 max |γe| = 0. (4.39)<br />

When we follow a similar procedure <strong>for</strong> the drained split in coupled flow <strong>and</strong> stat-<br />

ics (Equations 4.5 – 4.7) <strong>and</strong> add vk to the unknowns, convergence is obtained <strong>for</strong> the<br />

drained split of the coupled flow <strong>and</strong> dynamics from Equations 4.1 <strong>and</strong> 4.39 because<br />

<br />

<br />

lim e n+1<br />

<br />

<br />

= 0. As explained by Armero <strong>and</strong> Simo (1992), when we use the staggered<br />

∆t→0<br />

fs<br />

method (i.e., one iteration), convergence <strong>and</strong> first-order accuracy can be directly estimated<br />

by Lie’s <strong>for</strong>mula (Chorin et al., 1978; Lapidus, 1981), written as<br />

lim<br />

n→∞ [exp(A1t/n)exp(A2t/n)] n = exp[(A1 + A2)t] , (4.40)<br />

because Equation 4.29 is split by the drained split as<br />

˙ξ = A1ξ + s1, ˙ ξ = A2ξ + s2, (4.41)

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

Saved successfully!

Ooh no, something went wrong!