Daniel A. Ebrom

Hyperbolic traveltime analysis of first arrivals in an azimuthally anisotropic medium; a physical modeling study

Wave propagation in a fractured medium is modeled physically using layers of Plexiglas with thin films of water, held under moderate uniaxial confining pressure. The system exhibits anisotropy comparable to that of measured earth materials; i.e., shear‐wave splitting to waves with 3 percent velocity differences and P-wave directional anisotropy of at least 20 percent. SV polarizations demonstrate the concept of the shear‐wave window with the conversion of an SV body wave to an internal head wave with P-wave velocity, a head wave which is present in both the fractured medium and the control solid (unfractured) medium. For an azimuthally anisotropic medium, moveout curves are hyperbolic for a surface line oriented parallel to the fractures but are nonhyperbolic for a line oriented perpendicular to the fractures. Q anisotropy is observed in the system, with strongest attenuation on propagation paths perpendicular to the fractures.

Comparison of experimental velocity measurements with theoretical results in a solid-solid composite material

Two-phase composite media were constructed of aluminum powder inclusions embedded randomly in a clear resin matrix. The shapes of the inclusions are elongate, with aspect ratios ranging from 0.1 to 1.0, averaging about 0.25. The effective velocities of ultrasonic elastic waves in the bulk media were measured for these composite materials, and the observed data were compared to the theoretical values calculated using five different numerical models. The observed velocities agree with the theoretical values calculated using a dynamic model under the assumption that the inclusions have an oblate spheroidal shape.

Thin layers and shear-wave splitting

The near‐surface weathering layer is considered by many to be strongly anisotropic. Any shear‐wave signal passing through this low‐velocity layer will inherit, to some degree, the anisotropic response of this layer. For thin weathering layers, information about previous anisotropic events may be distorted; when the thickness of this layer approaches some physically defined limit, however, a previous layer’s anisotropic signature is completely overwritten. Hodograms and Alford rotations are typically used to analyze shear‐wave splitting in the presence of azimuthal anisotropy. When the time‐delay generated by an azimuthally anisotropic layer is ⩾τ/8, where τ = one period of the wavelet’s dominant frequency, distortion of a shear‐wave signal is great enough to degrade the accuracy of the interpretation in hodogram analysis. We found that Alford rotations are superior to visual hodogram analysis when the time delay between the fast and slow shear‐waves is less than τ/8. When two azimuthally anisotropic layer...

Study of Fresnel zone and diffraction amplitudes in a physical modeling experiment

Fresnel zone and diffraction amplitude responses have been calculated using numerical modeling techniques. (Hilterman, 1970, 1975, 1980, 1982, and Knapp, 1991) Our research involves analyzing P-wave and shear-horizontal wave Fresnel zone amplitudes with the use of physical models. These models consist of Plexiglas cylinders mounted on a plate and placed in a water tank for P-wave data, and a block of Plexiglas with cylindrical holes drilled into the bottom for the SHwave case. These collected data allow us to examine the effects of diffractions on various sized Fresnel zone reflectors. Amplitude analysis of the data confirm theoretical predictions. In addition, physical modeling of both acoustic and elastic waves allows a one to one comparison between P and SH-wave reflection and diffraction amplitudes.

Interpretability and resolution in post‐migration time‐slices

A surprising result was that data which were acquired on an adequate spacing (as predicted by plane-wave theory) were nevertheless still improved in interpretability by interpolation onto a finer grid. In our experiments, we found that Kirchhoff migration performed a better job of interpolation than a cheaper two-pass post-migration algorithm. In this regard, we feel that the Neidell hypothesis may contain a valuable intuitive insight: specifically, that interpolation aids visual interpretability (the most important criterion for seismic data).

Nine‐component data collection over a reflection dome: A physical modeling study

All components of plastic waves in an isotropic medium (Plexiglas) have hvc>n recorded in a reflection seismics geometry. This was accomplished by using a source polarizable in any of threr directions (in-line horizontal or SV, cross-line horizontal or SH, and vertical or P) and a receiver similarly polarizal+. Nine acquisition polarization pairs were thus recordc\d. The struct,ure in the model was a simljle dome embetldrd in a flat plain, and lines of acquisil,ion passed over bol.h the crest of the dome and its flanks. The lines were acquired at a constant offset, and no summing of traces has been performed. S-wave, P-wave, and mode conversion ( S-wa.ve to P-wave and P-wave to S-wave) events are all present on the sections. The P/P, P/SV, SV/P, and SV/SV sections enhance those events whose rays lie in the vertical plane below the line of acquisition. The sections incorporating an SH transducer as either source or receiver (or both) enhance thos_e events whose rays reflect from out. of the plane. Although the sources and receivers are (#errned SV, SH, and P, in fact any of the polarizations can receive any of the wave types given an appropriate ray angle.

The Effect of Fold On Horizontal Resolution In a Physical Model Experiment

S u m m a r y inal fold of 30. The CMP binning of this dataset was Using two datasets with different acquisition geometries equivalent to 30 by 30 foot squares, corresponding to the collected over the same physical model, we attempted to spacing between the closest pegs. That dataset was used test Neidell’s conjecture. We found that fold contributes by Ebrom et al. in 1995 to demonstrate the power of a significantly to spatial resolution. However, given an Kirchhoff migration as an interpolation operator. equal number of traces, finer bin spacing provided a much Subsequently, greater increase in resolution than fold. a new single fold dataset has been collected over model 94. The new survey is collected on 8 by 8 foot bin centers. The new dataset has coincident source and

A Comparison of 3-D Seismic Time Imaging Results From Physical Model Strike And Dip Acquisitions

This study compares seismic time imaging results of laboratory derived three dimensional seismic data acquired with survey orientations in the strike and dip of the target feature. The physical model is a two layered model with the interface constructed similar to the “French Model” (French, 1974). The interface creates a velocity perturbation to the underlying imaging target which is a series of four metal rods. The strike acquisition is oriented parallel to the strike of the rods, and the dip survey is perpendicular to the rods. Similar experiments include the actual field acquisition on the Bullwinkle Field, Gulf of Mexico in 1988 (O’Connell, et. al, 1993), the salt swell model study (Postma and Jeannot, 1988), and the studies done by CGG (Manin and Hun, 1992).

The dependence of lateral resolution on the orientation of symmetry axis and elastic parameter in transversely isotropic media

Similar models having positive and negative values and hence, contrasting wavefront curvatures were also used in conducting the experiments. Cylindrical holes of various sizes were drilled into the bottoms of the Phenolite models. The hole sizes ranged from 0.83 to 1.834 and from 0.86 to 2.054 times the expected Fresnel-zone diameters in the horizontally layered and vertically fractured Phenolites respectively.

A study of subsalt imaging using physical models

Physical modeling techniques were used to investigate some of the problems associated with subsalt imaging. The physical model was constructed of Plexiglas and Sylgard 3110. Two 3D seismic data sets were collected in the water tank system. A simulated closely spaced survey was designed. A conventional imaging technique (3D NMO followed by 3D poststack time migration) was tested first. This method was found to be of insufficient quality to adequately image the base of the salt body and underlying target structures. Long-period multiples were a serious problem in subsalt imaging. Subsalt imaging was weakened by the strong seismic reflection energy attenuation and complex ray path distortion.