Crosswell Logging for Acoustic Impedance

Crosswell seismic data are recorded by placing the seismic source in one well and receivers in a second well to measure physical properties between the two wells. When reflected seismic energy is extracted from th recorded seismogranis. the resultant picture is a function of acoustic impedance (density times velocity) between the wells. Previous crosswell seismic studies have predominantly used direct waves and measure only seismic velocities by tomographic techniques. Synthetic crosswell seismic data are generated by a computer to illustrate the concepts and procedures used for crosswell impedance logging in this study. Such synthetics are useful for investigating how the experiment should work, but real data reveal how the experiment does work. Crosswell seismic data recorded at the Multiwell Experiment (MWX) site in western Colorado revealed a few surprises when compared with synthetic data. Real data consist of a lot more than simple direct arrivals and reflections. There are numerous converted waves, borehole (tube) waves, and general noise. Source directivity proves to be a very limiting problem, because energy is concentrated in a horizontal plane perpendicular to the borehole. The final product of crosswell logging for acoustic impedance is a seismogram similar to the classic CDP seismogram recorded on the surface for exploration. The crosswell data in this paper are in the kilohertz range as opposed to surface data, which are generally below 100 cycles/sec [100 Hz]. Seismic data in the kilohertz range can theoretically resolve layers as thin as 2 ft 10.6 mi vertically. Actual resolution indicated is about 5 to 10 ft [1.5 to 3 m].

Well logs are normally recorded in individual wells and then correlated from well to well to deduce lithologic continuity. One problem with this qualitative log-correlation procedure is the interpretation of discontinuities. When a geologic horizon is identified in two adjacent wells, there is a fair degree of certainty that the horizon is continuous between the two wells. But there is always a chance that the horizon is discontinuous between wells and still observed at the two wellbores. Another problem exists when a geologic horizon is observed only in one well and the precise location of the discontinuity is unknown. Such problems are theoretically solved by a high-resolution. three-dimensional surface seismic survey. Such surveys, however, have extreme limitations in resolution. Crosswell seismology offers the possibility of obtaining extremely high resolution of both lateral discontinuities and vertical variations. Seismic surveys have vertical resolution that is predominantly controlled by the seismic wavelength. Typical wavelengths are 200 to 400 ft 161 to 122 mi, and resolution of distinct horizons vertically is about a one-fourth wavelength or 50 to 100 ft 115 to 30 m]. Geologic layers of this thickness are often continuous over a moderate-size field. Thinner layers are often discontinuous and very important for hydrocarbon-reservoir description and development. Horizontal resolution obtained from surface seismic surveys is becoming the biggest limitation in resolution as the reservoirs of interest become deeper. Resolution becomes progressively worse as targets become deeper. The Fresnel zone approximates the limit of horizontal resolution. In a simplified model, the Fresnel zone is in the shape of a disk-the smallest one that can be resolved laterally. A Fresnel zone of 2,200 ft 1671 m] in diameter is obtained from typical 40-cycle/sec 140-Hz seismic data when the target is about 10,000 ft [3048 m] deep. If the target is only 5,000 ft 11524 mi below the surface, the Fresnel zone decreases to only 1,600 ft 1488 mi in diameter. When a reservoir is 5,000 ft 11524 ml deep or more, surface seismic surveys are nearly impossible to use for mapping on the scale of 40-acre 116-ha] spacing, where 1,320 ft [402 m] usually separates wells. The limit of vertical resolution is often thicker than the pay zone and horizontal resolution wider than the well spacing. Surface seismic profiling is great as an exploration tool but poor as reservoir development tool unless the reservoir is very shallow. The two basic problems with seismic resolution are the distance to the target and the frequency of the data. These two problems are interrelated in that high frequencies are lost at greater distances. If a portion of the distance contains weathered rocks on the surface, then additional high frequencies are lost and resolution is again decreased. These problems can be eliminated most directly by placing the entire seismic experiment in holes beneath the weathered layer and closer to the targets to be studied. This experiment resembles a crosswell sonic log. Sonic logs provide velocity information about the medium surrounding the source and receiver, which are normally both in the same well. When the source is in one well and the receiver in a second well, the velocity structure between wells can be measured. This approach uses transmitted energy through the zone of interest and is often referred to as tomography. If the direct transmissions are neglected, then the remainder of the seismic energy is reflected and scattered from discontinuities in the geology. This reflected wave field is useful and the amplitude of reflected energy is a function of acoustic impedance contrast. Acoustic impedance is the product (multiplication) of density and velocity. When only the reflected portion of crosswell sonic logs is used. the experiment can be described as crosswell impedance logging.

Crosswell logging is a new science that has experienced relatively slow development in its brief 9 years of existence. Most crosswell logging studies in the past have been concentrated on measuring, the direct transmission of energy between wells and inverting the matrix of measured quantities (time and/or amplitude) for physical properties between the observation points. The theory has been applied in the past to sonic data and electromagnetic tomography. Electrical crosswell experiments have also been proposed for measuring interwell hydrocarbon saturation. Crosswell seismic reflection profiling for acoustic impedance has received very little attention. The procedure used in this paper is similar to that used by Baker and Harris. Such crosswell impedance logging is not necessarily advocated as the best high-resolution crosswell technique. It is presented as a new crosswell logging theory that should be considered in the ongoing development of many reservoir mapping, techniques. Crosswell sonic data were recorded at the MWX site near Rifle. CO1 by Los Alamos Natl. Laboratory with very encouraging results. A tomographic inversion for physical properties between wells was obtained on the basis of picking arrival times of P-waves and also S-waves.

P. 75^