gkowal/amun-code
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

EQUATIONS: Improve the guess in SRHD primitive variables solver.

Browse Source 
Signed-off-by: Grzegorz Kowal <grzegorz@amuncode.org>
This commit is contained in:
Grzegorz Kowal 2015-02-19 11:35:34 -02:00
parent f46981fb03
commit 56da64d935
1 changed files with 16 additions and 27 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

43
src/equations.F90
View File

@ -3246,13 +3246,10 @@ module equations
en = u(ien,i) + u(idn,i)
dn = u(idn,i)
! prepare the initial guess using pressure from the previous step
! prepare the initial guess
!
wm = en + q(ipr,i)
! set the initial W to the minimum value
!
w = wm
w = en + pmin
vv = mm / (w * w)
! find the exact W using an Newton-Ralphson interative method
!
@ -3501,7 +3498,7 @@ module equations
!
logical :: keep
integer :: it, cn
real(kind=8) :: vm, gm, tm
real(kind=8) :: vm, vs
real(kind=8) :: pr, dpw
real(kind=8) :: f, df, dw
real(kind=8) :: err
@ -3527,36 +3524,38 @@ module equations
!
do while(keep)
! calculate |V|², (1 - |V|²), and the Lorentz factor
! calculate (1 - |V|²) the square root of it
!
vv = mm / (w * w)
vm = 1.0d+00 - vv
gm = 1.0d+00 / sqrt(vm)
vs = sqrt(vm)
! calculate the thermal pressure and its derivative
!
! P(W) = (γ - 1)/γ (W - DΓ) (1 - |V|²)
! dP/dW = (γ - 1)/γ [(1 - D dΓ/dW) (1 - |V|²) - (W - DΓ) d|V|²/dW]
!
tm = w - dn * gm
pr = gammaxi * tm * vm
dpw = gammaxi * (1.0d+00 + tm * vv / w)
pr = gammaxi * (w * vm - dn * vs)
dpw = gammaxi * (1.0d+00 + (w - dn / vs) * vv / w)
! calculate F(W) and dF(W)/dW
!
! F(W) = W - P - E
! dF(W)/dW = 1 - dP/dW
!
f = w - pr - en
df = 1.0d+00 - dpw
f = w - en - pr
df = 1.0d+00 - dpw
! calculate the increment dW
!
dw = f / df
dw = f / df
! correct W
!
w = w - dw
w = w - dw
! calculate |V|² from W
!
vv = mm / (w * w)
! calculate the normalized error
!
@ -3579,10 +3578,6 @@ module equations
end do ! continue interations
! calculate |V|² from W
!
vv = mm / (w * w)
! print information about failed convergence
!
if (err >= tol) then
@ -3668,12 +3663,6 @@ module equations
it = nmax
cn = next
! calculate the initial |V|² from the guess of W
!
! |V|²(W) = |M|² / W²
!
vv = mm / (w * w)
! iterate using the Newton-Raphson method in order to find the roots W and |V|²
! of functions
!