sfrtmva2d (4.0)
index
user/pyang/Mrtmva2d.c
RTM with checkpointing in 2D acoustic media

 
Synopsis
        sfrtmva2d < Fv.rsf rho=Frho.rsf tau=Ftau.rsf tauo=Ftauo.rsf > Fw.rsf p1=Fp1.rsf p2=Fp2.rsf verb=n nb=20 nt= dt= fm=20.0 ns= ng= kt= nob=(int)log2f(nt) jsx= jsz=0 jgx=1 jgz=0 sxbeg= szbeg= gxbeg= gzbeg= csdgather=y vmute=1500 tdmute=2./(fm*dt)
The real value of checkpointing technology resides in the backpropagation with
viscoacoustic and viscoelastic wave equation, where the wavefield
reconstruction method using saved boundaries fails. Here, we only
demonstrate how to implement it in acoustic media without dissipation.
Note the backpropagation operator should be the adjoint of forward modeling!
Here we just use forward modeling operator for the time being!

 
Parameters
       
 
bool csdgather=y [y/n]
default, common shot-gather; if n, record at every point
 
float dt=
time sampling interval
 
float fm=20.0
dominant freq of Ricker wavelet
 
int gxbeg=
x-begining index of receivers, starting from 0
 
int gzbeg=
z-begining index of receivers, starting from 0
 
int jgx=1
receiver x-axis jump interval
 
int jgz=0
receiver z-axis jump interval
 
int jsx=
source x-axis jump interval
 
int jsz=0
source z-axis jump interval
 
int kt=
output px and pz component at kt
 
int nb=20
thickness of PML ABC
 
int ng=
number of geophones/receivers per shot
 
int nob=(int)log2f(nt)
number of buffers, default=optimal value
 
int ns=
number of shots
 
int nt=
number of time steps
 
file p1=
auxiliary output file name
 
file p2=
auxiliary output file name
 
file rho=
auxiliary input file name
 
int sxbeg=
x-begining index of sources, starting from 0
 
int szbeg=
z-begining index of sources, starting from 0
 
file tau=
auxiliary input file name
 
file tauo=
auxiliary input file name
 
int tdmute=2./(fm*dt)
number of deleyed time samples to mute
 
bool verb=n [y/n]
verbosity
 
float vmute=1500
muting velocity to remove the low-freq noise, unit=m/s

 
Used In
       

 
XJTU
        test/rtmva2d