Dwight Sukup

Facilitating technologies for permanently instrumented oil fields

It is a commonplace of technological advances that each improvement in technology relies on a host of enabling, or facilitating, subtechnologies. In pursuing a lofty goal (such as putting a man on the moon or monitoring fluid flow in a petroleum reservoir), a large number of intermediate problems must be overcome. In the case of the space race, enormous numbers of supporting technologies were created to solve problems ranging from keeping pumps operating (while passing liquid oxygen!) to keeping small objects from floating free in the capsules. Unfortunately, the oil business cannot afford the NASA strategy—funding multiple technology paths and picking the most appealing result. Instead, the best “fit for purpose” technology must be prudently chosen from the field of possible contenders.

Anisotropic Prestack Depth Migration of Converted-Wave Data From the Gulf of Mexico

Summary We apply prestack depth imaging to converted-wave data from the Gulf of Mexico. Migrating these data with isotropic S-wave velocities images reflectors at depths greater than those in the corresponding P-wave sections. From these depth differences and the isotropic S-wave velocity model we estimate an anisotropic S-wave velocity model. Anisotropic migration with this model yields images more consistent with the P-wave images.

Analysis of multicomponent seismic data from offshore Gulf of Mexico

SUMMARY 3D 4C surveys are a natural extension of 3D streamer surveys for time lapse reservoir monitoring. Until the use of seafloorrecorded converted waves reaches a level of maturity comparable to that of P waves, extra care will need to be taken to assure that events interpreted as converted-wave arrivals add value to the interpretation process. Careful analysis of the velocities associated with interpreted converted waves can indicate whether the interpretation is plausible. Using horizontal-component seismic data collected at Teal South, we generated velocity profiles for P-wave and S-wave velocities that agree with earlier published work (Hamilton, 1979). Modeling based on velocity fields generated by focusing analysis (with extrapolations for the shallow S-wave velocities from Hamilton) yields a synthetic common-receiver gather that closely resembles the field data in terms of changes in arrival time and amplitude with offset. This agreement between modeled data and field data supports our interpretation of the high amplitude events as waves that convert from P to S at the reflector.

Migrating sparse-receiver data for AVO analysis at Teal South Field

Summary In processing a 3-D ocean-bottom-cable (OBC) survey from Teal South Fi eld in the Gulf of Mexico, we had to address unusual challenges to imaging and AVO analysis. Sources were located on a dense grid, but the receiver grid consisted of only 24 widely spaced locations. In processing the survey, we obtained the best results by applying conventional AVO preprocessing followed by prestack migration (of common-receiver gathers). In comparing prestack and poststack migrations of the survey, we found that for bright-spot analysis and structural interpretation, migration after stack is sufficient. However, comparison of unmigrated and migrated CMP stacks makes it clear that AVO analysis should be preceded by migration. The common practice of comparing limited-range stacks (with migration either before or after stack) conveys some useful information, but is also complicated by acquisition footprint effects, especially in the near-range data. The common-receiver migration provided imaged anomalies that are stronger, more sharply defined, and more consistent spatially with fault-bounded reservoir compartments. It also provided 24-fold image gathers that we input to linearized AVO inversion. A resulting fluid-factor volume highlights AVO anomalies that are consistent with well control and with structure and bright spots known from a previous 3-D streamer survey.

Combining Dual-Sensor Data with Pre-stack Depth Migration - Imaging the Ghost and Primary Reflection at Teal South

Summary In a dual-sensor Ocean Bottom Cable (OBC) survey, both hydrophones and geophones are used to record the seismic signal. Primary reflection and water-column reverberations exhibit different behavior on hydrophone and geophone data. Rather than focusing on eliminating the reverberation by careful calibration of the hydrophone and geophone data, we present here a method of using pre-stack depth migration to use the reverberations as well as the primary reflection as part of the imaging. Combining the subsurface image from both not only significantly reduced the multiple effect as noise but also has extended the coverage. We demonstrate by applying the method to the Teal South 4D_4C survey.

Anisotropic 3D Prestack Depth Imaging of the Donald Field With Converted Waves

We apply prestack depth migration to a four-component (4C) 3D ocean-bottom-cable (OBC) data set obtained in 1999 over the Donald Field in the Gulf of Mexico. This is an area where a shallow gas cloud obscures a large portion of the field in conventional images. We produce both a Pwave and a converted-wave image. In order to make these two images consistent with each other, we need to include anisotropy (with TIV symmetry) in the migration technique. We obtain a 3D converted-wave image of excellent quality which clearly shows reflectors and faults that have never been imaged before.

Fast 3-D Prestack Depth Migration With a Parallel PSPI Algorithm

Summary A parallel computing model is presented that produces scaleable code for performing 3D prestack depth migration using the phase-shift-plus-interpolation (PSPI) algorithm. The parallel code is constructed by building standard PVM messagepassing calls into an existing serial code. Very few modifications are required to the original algorithm because the parallel model simply farms out individual shot records to a number of machine processing elements (PE’s) that are each running an independently spawned copy of the migration code. This simple approach to parallelism allows large 3D seismic surveys to be migrated in a fraction of the time normally required by the equivalent serial code on a single-PE computer. The PVM model can be used to port serial codes to different types of parallel machines with minimal development effort. A description of the parallel model is shown, along with benchmark performance results for code versions written for a workstation cluster and a Cray T3D, using a subset of a full 3D seismic survey as test data. A factor of 100 increase in effective processing speed was observed for a 128-PE Cray T3D run, relative to a serial single-PE run on a Sun SPARC