A STATISTICAL TECHNIQUE OF VELOCITY ANALYSIS IN SEISMIC PROSPECTING

Abstract

The modern geophysical exploration for hydrocarbons leans heavily on seismic prospecting techniques. The initial results in seismic surveys consists of time sections depecting the structural altitude of subsurface horizons. The results are more useful if they can be converted in terms of depths. The key to transformation from one domain to other is seismic velocity, Recent developements in interpretational techniques have used velocity and have gone a few steps further to stratigraphically reconstruct the subsurface on the basis of seismic velocities. This has become sn important stage in interpretation of seismic data. Moreover, the modern data acquisition techniques like the multi-channel C,D.P, method requires a precise knowledge of the velocity distribution. The success of modern seismic surveys depends critically on a accurate lnowlege of seismic velocities. The estimation of velocity has necessarily to be done from seismic data itself. most of the time, well information is not available in virgin areas. The usual. analysis consists of searching for the appropriate velocity to provide correct dynamic corrections for the various reflectors. In modern seismic practice the data is available in digital form. Hence, this determination is carried out on a digital computer. Since the accuracy of seismic sections and the value of seismic interpretation can be improved considerably by carrying out velocity analysis continuously along a profile, the volume of calculations in such nalysis Is necessarily large. Further, the conventional algorithms available to extract velocity information from reflection holographs are slow to implement on. a digital comps ter*. In the present Investigation we have developed new algorithms for ultimately calculating velocity with the help of Taylor series expansion of the a ,aression for reflection travel time* We have investigated the ranges of validity of this approximate but faster method of calculation and have found it to be useful over a wide range of conditions that are actually met in the case of real data. An optimum search technique has also been developed for estimating the stacking seismic velocity from the travel time data, It is based on Fibonacci search technique (P.5.T.). The relevant equations have been derived and a computer pro gra ime has been developed. The method is extremely fast In providing the optimum solutions. This technique is applicable to unimodal functions. However, this limitation can easily be overcome by dividing the initial interval of uncertainty of the velocity into subintervals each having its on maxima.