Raymon L. Brown

Anomalous dispersion due to hydrocarbons The secret of reservoir geophysics

When P- and S-waves travel through porous sandstone saturated with hydrocarbons, a bit of magic happens to make the velocities of these waves more frequency-dependent (dispersive) than when the formation is saturated with brine. This article explores the utility of the anomalous dispersion in finding more oil and gas, as well as giving a possible explanation about the effect of hydrocarbons upon the capillary forces in the formation.

AVO for one- and two-fracture set models

A theoretical comparison is made of PP and PS angle–dependent reflection coefficients at the top of two fractured-reservoir models using exact, general, anisotropic reflection coefficients. The two vertical-fracture models are taken to have the same total crack density. The primary issue investigated is determination of the fracture orientation using azimuthal AVO analysis. The first model represents a single-fracture set and the second model has an additional fracture set oblique to the first set at an angle of 60°. As expected, the PP-wave anisotropy is reduced when multiple fracture sets are present, making the determination of orientation more difficult than for the case of a single-fracture set. Long offsets are required for identification of dominant fracture orientations using PP-wave AVO. PS-wave AVO, however, is quite sensitive to fracture orientations, even at short offsets. For multiple-fracture sets, PS signals can potentially be used to determine orientations of the individual sets.

Characterization of dispersion, attenuation, and anisotropy at the Buena Vista Hills field, California

We create a log of intrinsic dispersion and attenuation for the Antelope Shale formation of the Buena Vista Hills field, San Joaquin Valley, California. High dispersion (or low Q) values correlate with thin sand and carbonate beds within the Antelope Shale. These beds are at least ten times as permeable as the host shale formation, so this effect provides a possible avenue for seismic prediction of permeability. The dispersion log is formed through comparison of crosswell seismic velocities (measured at approximately 1 kHz) and sonic log velocities (measured at approximately 10 kHz). In order to provide a proper basis for comparison, the sonic log must first be adjusted for field anisotropy, scaling effects, and resolution of measurement. We estimate a local shale anisotropy of about 20% based on correlations generated from published measurements of other shale fields. We apply resolution enhancement to capture the thin sand and carbonate beds, and windowed Backus averaging to match the measurement scales. A modeling study verifies the technique, and shows that beds of thickness greater than 30 cm have a measurable signature. The actual resolution is on the order of the crosswell Fresnel length, or about 7 m for the model study.

Three-parameter AVO crossplotting in anisotropic media

Amplitude versus offset (AVO) interpretation can be facilitated by crossplotting AVO intercept (A), gradient (B), and curvature (C) terms. However, anisotropy, which exists in the real world, usually complicates AVO analysis. Recognizing anisotropic behavior on AVO crossplots can help avoid AVO interpretation errors. Using a modification to a three-term (A, B, and C) approximation to the exact anisotropic reflection coefficients for transversely isotropic media, we find that anisotropy has a nonlinear effect on an A versus C crossplot yet causes slope changes and differing intercepts on A versus B or C crossplots. Empirical corrections that result in more accurate crossplot interpretation are introduced for specific circumstances.

DEVELOPMENT OF RESERVOIR CHARACTERIZATION TECHNIQUES AND PRODUCTION MODELS FOR EXPLOITING NATURALLY FRACTURED RESERVOIRS

For many years, geoscientists and engineers have undertaken research to characterize naturally fractured reservoirs. Geoscientists have focused on understanding the process of fracturing and the subsequent measurement and description of fracture characteristics. Engineers have concentrated on the fluid flow behavior in the fracture-porous media system and the development of models to predict the hydrocarbon production from these complex systems. This research attempts to integrate these two complementary views to develop a quantitative reservoir characterization methodology and flow performance model for naturally fractured reservoirs. The research has focused on estimating naturally fractured reservoir properties from seismic data, predicting fracture characteristics from well logs, and developing a naturally fractured reservoir simulator. It is important to develop techniques that can be applied to estimate the important parameters in predicting the performance of naturally fractured reservoirs. This project proposes a method to relate seismic properties to the elastic compliance and permeability of the reservoir based upon a sugar cube model. In addition, methods are presented to use conventional well logs to estimate localized fracture information for reservoir characterization purposes. The ability to estimate fracture information from conventional well logs is very important in older wells where data are often limited. Finally, a desktop naturally fractured reservoir simulator has been developed for the purpose of predicting the performance of these complex reservoirs. The simulator incorporates vertical and horizontal wellbore models, methods to handle matrix to fracture fluid transfer, and fracture permeability tensors. This research project has developed methods to characterize and study the performance of naturally fractured reservoirs that integrate geoscience and engineering data. This is an important step in developing exploitation strategies for optimizing the recovery from naturally fractured reservoir systems. The next logical extension of this work is to apply the proposed methods to an actual field case study to provide information for verification and modification of the techniques and simulator. This report provides the details of the proposed techniques and summarizes the activities undertaken during the course of this project. Technology transfer activities were highlighted by a two-day technical conference held in Oklahoma City in June 2002. This conference attracted over 90 participants and included the presentation of seventeen technical papers from researchers throughout the United States.

Wavelet estimation: An interpretive approach

There appears to be no uniform agreement on how to determine the polarity and shape of the wavelet in seismic data. In addition, when the determination is performed, the interpreter is usually the last to be informed of the results. Although the wavelet estimation problem can become buried in mathematics, the interpreter is best suited to judge the quality of the estimated wavelet and the applicability of the wavelet to the final interpretation. An approach to getting the interpreter involved in the wavelet estimation process is described in this article, and we believe that it offers some conceptual advantages over many of the popular methods currently in use.

Estimation of components of elastic scattering and intrinsic attenuation via well control at the Buena Vista Hills Field

Attenuation of seismic energy and velocity anisotropy consists of an intrinsic component and an apparent component caused by strictly elastic scattering from velocity and density heterogeneities. The intrinsic component of attenuation is believed to be related to fluid properties in the porous portions of a reservoir. However, it may be difficult to distinguish between the apparent (layer-induced) effects caused by elastic scattering and the intrinsic effects when both mechanisms are operative. In this case, a detailed velocity model of the formation is required in order to predict the elastic scattering component of velocity. We have implemented such a scattering correction technique to determine the magnitude of the intrinsic attenuation from observed sonic velocity data. We constructed a high-resolution velocity model with layers one inch thick based on FMI logs and quality factors derived from sonic log data recorded at the Buena Vista Hills field. Quality factors for intervals were derived using the two station spectral ratio method applied to the full wave sonic log data at the Buena Vista Hills field. The wave response for intervals at different frequencies was obtained using a viscoelastic plane-wave modeling code to model a finely layered region. Intrinsic attenuation is determined in one way by measuring amplitudes of sonic full wave seismograms. The amplitudes at two adjacent stations are processed using the spectral ratio method to obtain an estimate of the attenuation. A second approach to computing the intrinsic attenuation of the formation is to make an elastic scattering correction, based upon the contribution of the reflection coefficients that is caused by thin layers, to the total velocity measured for the interval at sonic frequencies. The resulting difference between the observed velocity and the velocity predicted as a result of elastic scattering is assumed to be due to intrinsic attenuation. The applicability of this method is demonstrated by comparing the predicted intrinsic attenuation results, based on the layered model, with the intrinsic attenuation determined from sonic logs.

Seismic attenuation and flow properties in fractured reservoirs

Summary Seismic attenuation due to fluid movement at different scales within a fractured formation can be used to estimate the permeability tensor. Poroelastic models accounting for the fluid movement via a permeability tensor are used here to predict the intrinsic attenuation of a fractured formation. In addition to the permeability tensor, the length scales for the fluid movement and the stiffness matrix for the formations are required for the model. Using the type of modeling described in this paper, the intrinsic attenuation associated with seismic signals at different frequencies can be used to estimate matrix and fracture permeability. Another independent estimate of permeability can be obtained from well tests or production data. By combining the two estimates of permeability (seismic and well tests), constraints can be placed upon fracture aperture and connectivity (as defined in this paper) at the well. Both of these parameters are important for engineering purposes. If the connectivity is known as a function of the crack porosity, the effective aperture and crack porosity of the fractures can be mapped using seismic data. Separation of the effects of elastic scattering attenuation from the intrinsic attenuation will be a part of the problem facing the ideas presented here. An example is given for a fractured reservoir neglecting the effects of elastic scattering.

Fracture roughness: The key to relating seismic velocities, seismic attenuation and permeability to reservoir pressure and saturation

Summary: The roughness of a fracture surface plays a role in the fracture permeability, the fracture compliance and the pressure dependence of the fracture aperture. Ultimately these properties of fractures are related to the time-dependent properties of a fractured reservoir. Since all reservoirs are fractured to some extent, these properties of fractures may have a more general application than is generally believed. It is proposed here that the fracture roughness also controls the seismic attenuation associated with fluid motion within fractures during the passage of a seismic P- or S-wave through a fractured reservoir. In fact, a shear wave can become more sensitive to fluid content than P-waves as a result of fracture roughness. A fracture modeling procedure is presented that can be used to relate seismic velocities and attenuation to the permeability and the pressure dependence of the permeability tensor for fractured reservoirs. Issues of scale and frequency dependence are discussed.

Who will win the long-distance propagation race between boreholes; the tortoise (channel wave) or the hare (direct P-wave)?

This article is on the subject of estimating the relative distances over which different types of waves might be detected. While emphasis is on the transmission and detection of waves between boreholes, the basic philosophy can be applied to any problem in seismology. The comparison is expressed in terms of an imaginary race between the wave types which are disguised as characters from a famous fairy tale, “The Tortoise and the Hare.” The long‐distance race between the wave types is an important one because some of them can potentially be used to directly study reservoirs between wells. This paper attempts to lay the foundation for estimating the maximum transmission distance between wells for the different wave types considered. Let the race begin!