Major improvement in 3D seismic imaging expanding North Sea opportunities

Workstation interpretive depth imaging
solves gas cloud problem

Thad Dunbar
Amoco (U.K.) Exploration

Mark Lance
Landmark Graphics

• Schematic cross section of the regional geology and the project area outlined in red. [40,563 bytes]
• The gas-charged zone (center of diagram) that developed over the salt structure is largely responsible for the poor 3D seismic imaging. [11,927 bytes]
• Seismic cross section, 3D horizons with velocity overlay derived from ray traced based methods, and wellbore paths to the producing horizon. [8,657 bytes]
• 3D prestack depth imaging has improved the resolution of events through the gas cloud and around the salt intrusion. [15,310 bytes]

The global industry today continues to push exploration and production activities into areas of increasing complexity. Not only must asset teams make more complex inferences about subsurface geology; they must also support more complex business processes.

The rising cost and demand for seismic vessels and drilling rigs has created an urgent need to compress cycle times at every stage of exploration and exploitation. With projections that equipment and personnel shortages will continue, oil and gas companies are more focused than ever on the economic value of making the right technical and business decisions, fast!

Inspiring new play concepts and fueling ongoing exploitation projects in areas of great geologic complexity demand that geoscientists employ the most advanced geophysical technology.

For years, resolving subsurface interpretations in subsalt, deepwater and other difficult areas has been possible using sophisticated 3D seismic velocity modeling and prestack depth imaging technologies. However, business constraints often prohibited more routine use of these tools. Geoscientists were faced with the formidable task of balancing the accuracy of the seismic image and resulting earth model with the actual time and cost required to create them.

3-D velocity

Today, advances in integrated exploration software, visualization, and computing technology are enabling oil companies to realize the practical value of 3D velocity modeling and depth imaging, when they are used in appropriate integrated workflows.

Some years ago, Amoco (UK) brought onstream a development project in a stratigraphically and structurally complex part of the UK sector of the North Sea. As reservoir pressure declines over time, additional infill drilling is required in the near term to maintain production rates, meet sales contract obligations, and capture incremental reserves. In addition to the planned drilling, other exploration targets were identified below the producing formation. These deeper targets could be drilled at the same time and if successful, might reduce the number of infill wells.

The current producing field consists of deep marine Paleocene turbidites. Mid-Miocene salt intrusion uplifted the sands, creating four-way closure and the field trap. The deeper targets have much greater structural and stratigraphic complexity, translating to greater risk than the infill wells. One target is the Upper Jurassic sand.

These shallow marine sands were deposited into structural lows (rifts) created by a period of Jurassic extension. The rifts were also zones of weakness resulting in preferential zones of salt uplift. Later mid-Miocene salt movement inverted the structure, creating the trap for this target. A second potential target is the fractured Cretaceous chalk. The same late salt movement also uplifted and flexed the relatively brittle overlying chalk, creating fracture porosity with four-way closure.

Gas-charged zones

A gas-charged zone that developed over parts of the salt structure is largely responsible for the 3D seismic imaging problem. The velocity contrast of the gas cloud hinders imaging of both the producing horizon and deeper potential. Imaging the 3D seismic data within and below the gas cloud presents a challenge that must be overcome before recommending the infill-drilling plan to meet production obligations.

To solve the imaging problem, Amoco considered two key technologies, prestack depth migration and four-component (4C) ocean bottom cable (OBC) seismic acquisition. Both technologies have been successfully used to improve the image of seismic reflectors within a gas cloud.

The time and cost of acquiring OBC data, plus additional processing and interpretation cycle time, posed a risk to meeting Amoco's business milestones. Due to the lead-time required for contracting OBC vessels, Amoco had to make a decision quickly.

Given the complexity of the geology and the relative newness of OBC technology, Amoco geoscientists believed prestack depth migration offered a better chance of imaging target reflectors within the available time. Amoco partnered with geoscientists from Landmark Graphics to design and implement the most effective interpretive depth imaging workflows for the project. The objective was to improve the seismic image within and beneath the gas cloud while meeting timelines dictated by Amoco's business situation.

Integrated workflow

For prestack depth imaging to be successful, the project team needed to develop an accurate velocity model for the key geologic intervals and particularly the troublesome gas-charged zone. Amoco had already launched a major 3D prestack depth migration project covering 500 sq km to support ongoing exploration in the license area. But velocity modeling consumes a large percentage of overall cycle time, so this depth migration project was scaled down. A target-oriented volume, focused only on the area of the gas cloud and two deeper targets, was selected.

The Amoco/Landmark team then designed a practical workflow that would optimize the attributes of integrated software and a networked workstation environment. Key attributes of the integrated solution included:

• Flexible workflows that matched appropriate technology to geologic complexity
• Integrated data access among interpretation, modeling, and processing domains
• 3D visualization and animated analysis graphics
• Symmetric multi-processor computing architecture and network communication.

The well database, interpretation system and the velocity modeling tools were run on a dual-headed SGI Octane with R10000 CPU and SI/SSI graphics. 3D prestack depth imaging was executed on an eight-node SGI Power Challenge with two gigabytes of shared memory. The two systems were connected by a switched 10BaseT network, allowing rapid data transfer between team members.

Building velocity model

To keep the velocity-modeling project on schedule, appropriate technologies were applied to each interval depending on geologic/velocity complexity. In a three-phase approach, less rigorous and more rapid modeling techniques were used in the shallow, simpler portion of the model. More sophisticated and time-consuming technologies were used only in intervals requiring greater accuracy.

• Phase One - Vertical Depth Stretch: The initial interval velocity field was created by converting stacking velocities to depth interval velocities using a horizon-constrained Dix conversion. The original poststack time-migrated data and interpreted horizons were vertically stretched to depth. The accuracy of the depth conversion was evaluated in two ways.

Forward ray tracing provided a qualitative validation. In 3D views of the depth-converted seismic data and velocity field, image rays were vertical from the surface of the model to the top of the current producing horizon. Below this horizon and around the gas cloud, image rays were noticeably bending. Vertical depth conversion was not valid due to geologic complexity.

The model was further validated quantitatively. A comparison of well tops with depth-converted seismic revealed depth errors averaging nine percent at the top of the producing formation. Both validation methods indicated the initial velocity model was unacceptable and more accurate modeling techniques would be required.

• Phase Two - Coherency Inversion and Map Migration: To improve the accuracy of both the interval velocity and the structural position of reflectors in the depth model, two additional technologies were brought into the workflow. 3D common-midpoint (CMP) coherency inversion analysis measured lateral velocity changes between the sparse well control. The result was a gridded velocity map for each interval. Then the interpreted time horizon for each interval was converted to depth using map migration, a more accurate technique than simple vertical stretching.

Validating the model at known well locations showed that the previous nine percent depth error had been reduced to approximately three percent. Below the chalk, however, coherency inversion produced velocities that were inconsistent with the interpreter's experience. The rapid velocity contrast in the gas-charged zone was apparently responsible. To derive velocities below the top of chalk, a more rigorous method was necessary.

• Phase Three - Global Tomographic Inversion: Properly imaging target horizons below the chalk required even more sophisticated technology. The team needed to analyze the scope of the velocity problem and determine if tomographic inversion and prestack depth migration could indeed image the key horizons before committing time and resources to 3D prestack depth migration on the full target volume. To save time, the team targeted a single line from the 3D survey, then performed global tomographic inversion and prestack depth migration on it. Global tomographic inversion was perfectly suited to resolve the combined adverse effects of the gas cloud, the chalk layer, and the salt.

The resulting depth image was a clear improvement over the original time migration. The team proceeded to complete the full 3D-velocity model. In a target-oriented fashion, the velocity field below the chalk was updated using tomographic inversion along key lines then interpolated along structure to create the 3D model. Next, phase two results for the upper section were merged with phase three updates for the deeper, more complex portion of the data.

Rapid improvements in the accuracy of the velocity model were due, in part, to the integrated interpretation and analysis environment. Direct access to shared data by the interpretation system, well database and modeling tools significantly reduced cycle time by eliminating data reformatting and intermediate tape storage or transfer. Already, collaboration between interpreter and processor was paying off. The interpretive approach ensured that results from each phase were aligned with current knowledge of the regional geology and play concept.

Prestack depth migration

Expanding the target-oriented approach, which had proven so effective, the team 3D prestack depth migrated every 20th line of the survey using the updated velocity model. They quickly verified that coherency inversion had produced an accurate velocity solution down to the chalk, and that tomographic inversion had accurately resolved velocities in the deeper section. Both within and beneath the gas cloud, structures were better imaged. These results gave the team confidence to proceed with 3D prestack depth migration of the full target volume.

Interpretation of the prestack depth migration is currently underway. With the improved imaging, the infill wells can be better located, reserves can be more effectively exploited and the need for the 4C OBC data will probably be eliminated. The deeper exploration targets can also be evaluated better with the new depth data. If successful, these targets will not only provide new reserves but may reduce the number of infill wells.

Conclusion

Historically interpretation, modeling, and seismic imaging have taken place in isolation, stretching out cycle times. In this complex North Sea setting, an interpretive, target-oriented approach resulted in more accurate depth imaging in far less time. The integrated environment supported a collective effort that crossed traditional disciplines. Integrated applications enabled flexible workflows in which velocity data, interpreted horizons, and seismic data could be shared between team members without data reformatting or intermediate data storage.

3D prestack depth migration is becoming a critical tool for evaluating exploration and exploitation opportunities. Its greatest value is in areas of complex structure and rapidly varying velocities. Advances in computing power and software integration are making depth imaging more accessible to asset teams than ever before. Oil and gas companies that take advantage of these sophisticated technologies will achieve better technical solutions and make better business decisions, even under today's challenging conditions.

Acknowledgements

The authors would like to thank Amoco (U.K.) Exploration Company, Amerada Hess, and BG for allowing us to publish the results of this project. Special thanks also to Buzz Davis and Gregg Hofland of Landmark for their modeling and depth imaging expertise, and their technical contributions to this article.

Copyright 1998 Oil & Gas Journal. All Rights Reserved.