Seismic velocity model optimization method using simulated annearling to determine prestack travel-times

A method of operating a computer to determine a reference seismic velocity model that includes lateral and vertical variations in a stratified subterranean medium, the method comprising:

• storing seismic data acquired for selected seismic events, said data being recorded as seismograms depicting seismic traces representing the amplitude of seismic reflections for said events as a function of time (t) and common midpoint surface location (x) along a path of exploration of said medium;
• constructing a stacked section of common midpoint gathers of said seismic traces;
• displaying said stacked section;
• picking normal incidence traveltimes (T.sub.OL (x)) from said display and determining the associated stacking velocities (V.sub.stL (x)) as a function of said location (x) for selected seismic events (L) from said stacked section;
• determining true prestack traveltimes (T.sub.L (x, h)) for each said selected event (L) as a function of said location (x) and an offset distance (h), by summing a term representing the best fit hyperbola for said event with a perturbation term representing the non-hyperbolic component thereof, said hyperbolic term being determined from said picked normal incidence travel times and said associated stacking velocities, and said perturbation term being determined by simulated annealing;
• constructing an initial reference seismic velocity model for said medium;
• determining travel times (T.sub.cal (x, h)) from said velocity model to be compared to said true prestack traveltimes (T.sub.L (x, h) by map migrating said picked normal incidence travel times from time to depth and then ray tracing said seismic velocity model;
• comparing said travel times (T.sub.cal (x, h)) determined from said velocity model with said true prestack travel times (T.sub.L (x, h)), such that when they are substantially the same within a predetermined error norm, said velocity model provides an accurate representation of said subterranean medium strata;
• updating said velocity model when said predetermined error norm is not satisfied, said velocity model being updated by a constrained generalized least squares solution;
• and
• returning to said step of determining travel times from said velocity model until said error norm is satisfied.

A computer program stored on a computer-readable medium for execution by a computer for determining a reference seismic velocity model that includes lateral and vertical variations in a stratified subterranean medium, the computer program comprising:

• instructions for storing seismic data acquired for selected seismic events, said data being recorded as seismograms depicting seismic traces representing the amplitude of seismic reflections for said events as a function of time (t) and common midpoint surface location (x) along a path of exploration of said medium;
• instructions for constructing a stacked section of common midpoint gathers of said seismic traces;
• instructions for displaying said stacked section;
• instructions for picking normal incidence traveltimes (T.sub.OL (x)) from said display and determining the associated stacking velocities (V.sub.stL (x)) as a function of said location (x) for selected seismic events (L) from said stacked section;
• instructions for determining true prestack traveltimes (T.sub.L (x, h)) for each said selected event (L) as a function of said location (x) and an offset distance (h), by summing a term representing the best fit hyperbola for said event with a perturbation term representing the non-hyperbolic component thereof, said hyperbolic term being determined from said picked normal incidence travel times and said associated stacking velocities, and said perturbation term being determined by simulated annealing;
• instructions for constructing an initial reference seismic velocity model for said medium;
• instructions for determining travel times (T.sub.cal (x, h)) from said velocity model to be compared to said true prestack traveltimes (T.sub.L (x, h) by map migrating said picked normal incidence travel times from time to depth and then ray tracing said seismic velocity model;
• instructions for comparing said travel times (T.sub.cal (x, h)) determined from said velocity model with said true prestack travel times (T.sub.L (x, h)), such that when they are substantially the same within a predetermined error norm, said velocity model provides an accurate representation of said subterranean medium strata;
• instructions for updating said velocity model when said predetermined error norm is not satisfied, said velocity model being updated by a constrained generalized least squares solution;
• and
• instructions for returning to said instructions for determining travel times from said velocity model until said error norm is satisfied.