Page 212 - Origin and Prediction of Abnormal Formation Pressures

SEISMIC METHODS OF PRESSURE PREDICTION 187

(as the pressure is decreased) results in a decrease in seismic velocity which, in
turn, decreases the acoustic impedance. The decrease in acoustic impedance alters the
reflection coefficient and the seismic amplitude at the reservoir interface. As water is
injected into the reservoir, pressure builds up and free gas is pushed back into solution
in the oil. A large impedance contrast is observed between the area where there is a
free gas with the oil and where there is oil with redissolved gas. This will generate a
profound seismic amplitude variation between the two areas.

Estimation of sonic velocity from resistivity logs

The sonic velocity of formations obtained by acoustic logs is necessary in solving
many problems arising during exploration. This information is often lacking, however,
because sonic logs are sometimes obtained only in the productive portions of the
formation. Thus, it is necessary to estimate the sonic velocity on the basis of other
logging data. This is especially critical for shales, because even if the sonic logs are
available, they are very hard to interpret in badly washed-out beds.
One of the methods in solving this problem is using the correlations between the
sonic velocity and other logging data (electrical, gamma-ray, neutron, etc.). In some
regions, this approach gives positive results. For example, for various lithologies in
west Siberia, Bazylev (1987) obtained a system of equations relating longitudinal and
transverse wave velocities to the parameters of neutron gamma-ray or thermal neutron
logs. These equations are characterized by sufficiently high correlation coefficients
ranging from 0.74 to 0.97, with mean-squares error of sonic velocity estimation of
50-150 m/s. However, development of one- or multidimensional equations for accurate
estimation of sonic velocity is not always possible.
In normally compacted formations with hydrostatic pressure, sonic velocity can be
estimated from the velocity vs. depth relationships [V -- f(D)] obtained for certain
lithologies (and regions) (Kerimov, 1987; Averbukh, 1988). This method, however, is
not applicable to overpressured formations.
Kerimov et al. (1996) proposed a method for estimating (1) the sonic velocity, Va,
in abnormally pressured shales using resistivity logs and, thus, (2) shale bulk density.
Analysis of the well-log data for productive strata in Azerbaijan showed that there is a
poor correlation between the sonic velocities and other well-log data, such as resistivity,
SP, neutron and gamma-ray.
Introducing the normal trend for sonic velocity Va and resistivity Pn allowed these
authors to express sonic velocity (and, therefore, the bulk density of the shale) as a
nonlinear function of resistivity with good correlation between the normalized velocity
and normalized resistivity. The best-fit regression equation is of the following form:

where Pa and Pn are the resistivities of abnormally-pressured and normally-compacted
shales, respectively.