International Symposium on Recent Advances in Exploration Geophysics (RAEG 1995)

Since the introduction of digital data processing in reflection seismology, new algorithms have been developed, in step with the progress of computer power. Massively parallel computers have appeared in the oil industry, and it has been recognized that parallel computation is very attractive in cost performance. In this paper we investigate algorithms in parallel computation for seismic modeling and pre-stack depth migration. The pseudospectral method which we adopted for the seismic modeling requires a 2-D FFT. Each processor can contains only the part of data, therefore 2-D FFT cannot be achieved in one pass. Since the data swapping is necessary, we developed here a parallel swapping algorithm and a parallel disk I/O method to reduce the computation time. There are many theoretical algorithms to apply pre-stack depth migration. For parallel computation we adopted the explicit extrapolation finite difference method in the frequency and space domain. Implementation was done using “ parallel frequency slice" computations. All algorithms developed here are easy to implement in workstation clusters since these algorithms are based on the message passing method.

The paper reviews the seismic data processing techniques based on a version of the Radon transform, developed for characterizing non-planar and spatially limited structures of arbitrary orientations in crystalline rock. The examples are taken from the VSP surveys carried out in the borehole KR6, at the Olkiluoto site in SW Finland. Polarization analysis applied in the transform space allows the positions of the reflectors to be determined by estimating the direction of the arrival of the reflected waves. The approach permits the determination of both the 3D-position and local orientation of the observed reflectors. Anisotropic effects are included and accounted for in the analysis.

This paper presents the model experiment on the enhancement of seismic signals for fracture detection. To enhance seismic signals from fractures, we propose a repeated measurement and a differential analysis by changing the seismic properties of fractures actively. Air bubbles are injected in the fluid-filled fracture to change its seismic properties. We performed the bar model experiment and the block model experiment to investigate the effect of the injection of air bubbles on P-wave propagation. As the result, the amplitude of the transmitted P-wave is attenuated and that of the reflected P-wave is emphasized by injecting air bubbles in the fracture. The effect of air bubbles can be represented as the variation of the density, the velocity and the attenuation characteristics. We can estimate the area where air bubbles are distributed by monitoring the amplitude attenuation of the transmitted P-wave.

A new two-dimensional (2D) magnetotelluric (MT) inversion scheme is proposed in this paper. This scheme is based on a locally two-dimensional analysis in order to minimize computational time and computer memory. The MT governing equation is linearized in terms of the magnetic field and electrical conductivity for the perturbation analysis. The perturbed equation is then multiplied by a test function and integrated over the cross section. Integrating by parts and then substituting this test function with local magnetic fields, a new equation is obtained that is a 2-D variational integral for the electrical conductivity. The new equation is general in the sense that it can explicitly include the horizontal derivative of the magnetic field. If the horizontal derivative term is eliminated, the new equation becomes identical to the Rapid Relaxation Inversion (RRI) scheme proposed by Smith and Booker (1991).

The new EM tomography system was planned for the detailed survey of mineral exploration. EM tomography system consists of surface; transmitter (3-component loop sources) and borehole; receiver (vertical component), and uses frequencies in the 8 - 8,192 Hz range. As the first step of the system development, we conducted model studies using SLDM (3D) developed by Druskin (1994). The phase data of EM tomography is very sensitive to the anomaly and 3-component sources are useful especially for multi-dimensions structure case. Also, these simulation results were used for the system design such as necessary sensitivity of receiver, transmitting moment and so on.

The present paper proposes a new algorithm for the three-dimensional depth transform of the surface reflection, the cross-well reflection and VSP data by use of equi-traveltime planes. The equation of equi-traveltime planes of waves which travel from source to receiver through reflection point on reflector is derived. The derived equation shows the equation of ellipsoid of which foci are source and receiver. This means that a reflection point on reflector is located on ellipsoid. Ellipsoids are drawn for each pair of source and receiver. Consequently, reflector is determined by drawing a common tangent plane to these ellipsoids. The procedure of depth transform is as follows. 1) The velocity distribution within survey area is assumed. 2) The ellipsoids for each sample time in seismograms for each pair of source and receiver are drawn. 3) Reflector is determined as the common tangent plane to these ellipsoids. Here, reflector is automatical1y drawn as follows. 1) Survey area is divided into cubic cells of appropriate size. 2) An ellipsoid is drawn from seismogram for each pair of source and receiver by changing with one sampling interval from the designated time after the travel time of first break. 3) The intersecting points of the ellipsoid with the vertical lines of cells are shifted to the nearest cell corners. 4) Amplitudes for the corresponding sample times on seismograms are stacked on these cell corners. The amplitudes on the cell corners where the common tangent plane of ellipsoids passes are amplified by summation with in-phase. However, the amplitude on the other grid cell corners would be canceled because of the summation with out-phase. Judging from the model studies, it was clarified that this algorithm reconstructed the structure with good accuracy, the short calculation time and the small requirement for core memory.