A new way of performing on-demand seismic monitoring of an oilfield is now being implemented, using a permanent array of seismic receivers installed in Norwegian waters.

For several years, the oil industry has been buzzing with talk about the e-Field, the Instrumented Oil Field, the Intelligent Oil Field, the Field of the Future. These labels have all been applied to a common vision, varying in its details but driven by the realization that, if we could know what is going on inside a reservoir, in real time, with full spatial coverage, we could certainly manage it better.

The Field of the Future vision begins with the idea of better downhole measurements (of pressure; temperature; and the flow of gas, oil and water) and with downhole controls of valves, etc.

But these measurements tell us only about conditions and events in the various boreholes. To assess the rest of the reservoir away from the boreholes, 4-D seismic is the tool of choice. Although it provides only seismic signatures (e.g. amplitudes), not what we really want (e.g. saturations), it is the only measurement we can make which covers the entire reservoir.

The high cost of conventional 4-D seismic means that usually it can be performed only at intervals measured in years. If 4-D seismic could be done cheaply and frequently, it would have an enormous impact on the way that oil reservoirs are managed.

About 8 years ago, BP conceived the idea of permanent installation, on the seafloor, of Ocean Bottom Seismometers (OBS) (Figure 1). With such a permanent seismic array, 4-D reshoots could be done with only a source-boat working and hence would be relatively inexpensive. Seismic 4-D reshoots could be acquired "on-demand" rather than infrequently and could become an integral part of active reservoir management. Life-of-field seismic (LoFS) could be an integral part of the Field of the Future.

With LoFS receivers permanently emplaced and immobile, a major source of artifacts (receiver-position variation) in the 4-D seismic difference is eliminated so that small changes in the reservoir, occuring in short time spans, can be reliably detected. For example, a reservoir engineer should be able to commission a reshoot a few weeks after starting an injection program to see exactly where the injected fluids are going.

The first version of such a permanent seafloor seismic array was the Foinaven Active Reservoir Mangement (FARM) system in 1995 (see Cooper, et al, "Foinaven Active Reservoir Management: Towed streamer and buried sea-bed detectors in deep water for 4-D seismic," 1999). However, the OBS technology was less mature then, and the seismic receivers were hydrophones only, so the promise was only partly realized. But the project did show the concept could work and established that remarkably repeatable seismic data could be acquired with receivers trenched into the seabed and rapidly processed for time-lapse differences.

Figure 2 shows the time-lapse difference from the FARM survey from 1995 to 1999 and also from a conventional (towed-streamer) 4-D survey over the same area (Cooper, et al, "Foinaven Active Reservoir Management: Towed streamer and buried sea-bed detectors in deep water for 4-D seismic," 1999). It is clear that the ocean-bottom hydrophone (OBH) 4-D difference is free of a number of artifacts that are apparent in the towed-streamer 4-D difference.

Not obvious in the figure is the fact that significant cross-equalization procedures were required for the conventional difference, while only a bulk time-shift (7 ms, presumably accounting for a difference in seawater velocity) was required for the OBH survey. This encourages our hope for similar simplicity of 4-D differencing for a modern LoFS survey.

Since then, four-component Ocean Bottom Seismology (4C-OBS) has become feasible, enabling the recording of converted shear (S) waves as well as conventional compressional (P) waves. The world's first permanent 4C-OBS seabed array was installed this summer in the southern sector of the Norwegian North Sea.

The technology and economics

At the leading edge of the oil business, high technology and advanced economics are always closely connected. With a permanently installed seafloor array, the costs of the reservoir-monitoring program are front-loaded instead of delayed. Hence, this particular high-tech idea is counter-intutitive, and so the economics must be carefully assessed. In fact, many of the issues were recently addressed in these pages (Donoghue, "4-D Back to the Future," March 2003).

The essential idea is for the reservoir manager to look into the future and to design a 4-D seismic program for the entire life of the field (instead of just planning for the coming year). The key question to ask is, "Is there a better way to invest that program budget, rather than in a few conventional surveys sparsely spaced over time? "

The permanent array offers the following technical advantages:

• Reservoir engineers can obtain cost-effective surveys when they are needed.
• The time-lapse seismic differences should be of higher quality than is achieved with conventional techniques.
• Later re-shoots would not be hampered by the growing obstructions of platforms over the field.
• The reservoir would be illuminated from all azimuths.
• Surveys would acquire converted shear waves in addition to conventional P waves.

These features all add value, but do they justify the cost? Clearly, any cost/benefit analysis is just speculation, unless based upon actual numbers.

The learning curve

With the advent of any new technology there is a learning curve. Following the early stages, experience brings efficiency, and costs decline. Sometimes, these initial costs are so discouraging that the technology never gets the chance to deliver its potential value.

Even with large upfront costs, the economic argument may be successfully made for a field that is characterized by:

• large reserves in place;
• an extensive drilling program planned and subject to influence by the 4-D seismic data; and
• some feature (surface or subsurface) that renders conventional 4-D seismic difficult.
All of these characteristics are shared by the Valhall field in the Norwegian sector of the North Sea, operated by BP on behalf of its partners Shell, Total, and Amerada-Hess.

Valhall

Valhall is one of the classic North Sea chalk reservoirs. Discovered in 1975, with total production to date of 480 MMboe and current estimated remaining reserves of 570 MMboe, it continues to produce at near-peak rates through aggressive reservoir management, including the application of advanced seismic technology.

Thirty wells are to be drilled into the area of Valhall covered by the LoFS array. A waterflood program will begin in late 2003. These activities will be managed more effectively because of the LoFS seismic imaging.

Near the crest of the Valhall structure, the conventional seismic image is severely distorted by the presence in the overburden of gas in low concentrations. Such gas-obscured zones can be imaged regardless, using converted shear waves recorded on the horizontal components of 4C-OBS receivers. In fact, Valhall provided many of the early learnings that were required before such images could be fully realized (Thomsen, "Converted-wave reflection seismology over inhomogeneous, anisotropic media," 1999).

During 20 years of production, the extraction of hydrocarbon fluids from the reservoir has lowered the pore pressure substantially. At the crest of the Valhall structure, the chalk framework (with depleting pore pressure) cannot support the overburden, so it compresses, leading to a sagging of the entire overburden and a subsidence of the seafloor surrounding the original production platform complex.

An elaborate plan has been created to deal with these effects. It includes the waterflood mentioned earlier along with and two additional platforms at the north and south ends of the field. The LoFS images will be crucial for monitoring the progess of this subsidence mitigation plan.

According to classical analyses, time-lapse effects in seismic data are supposed to be minimal in carbonate reservoirs because the stiffness of the framework of grains minimizes the seismic effect of fluid substitution and of changing pore pressure. However, at Valhall the porosity is so high that these arguments do not apply, and in fact recent time-lapse processing of conventional data (Figure 3) (Barkved et al, "4-D seismic response of primary production and waste injection at the Valhall field," 2003) has empirically revealed the existence of substantial time-lapse seismic changes due to production over 10 years. This analysis also provides a basis for estimating the size of changes to be expected after much shorter intervals.

Based on these facts, the Valhall partners agreed on the economic and technical basis for first LoFS installation.

LoFS at Valhall

The permanent seismic array at Valhall consists of about 75 miles (120 km) of OBS cable trenched into the seafloor in July 2003. The pattern of receivers, shown in Figure 4, is designed to avoid significant seafloor infrastructure while supporting 4-D seismic imaging of the highest quality.

Following initial testing, the baseline LoFS survey was acquired in August 2003. The shooting is done with a standby vessel, modified to accomodate new airguns of compact design and allowing for rapid retrieval so that the safety function of the standby vessel will not be compromised. The airgun array is designed to radiate energy equally in all compass directions since receivers are live in all directions (not just aft, as with conventional surveys).

The geophone receivers are ungimballed and installed without control of the twist so that only the orientation of the inline component is known initially. An orientation algorithm that does not assume vector fidelity of the receivers (Dellinger, et al, "Horizontal vector infidelity correction by general linear transform," 2001) is used to transform the data to the conventional (x, y, z) coordinate system and to optimally establish vector fidelity within the constraints of linearity and causality.

Current plans call for reshooting five times in 18 months, with further shooting on demand. This requires that the data processing be done with extraordinary speed, with much faster turnaround than is conventionally expected in a time-lapse survey. To this end, the data will be brought ashore via a fiber-optic link, first to a BP operations center and then to the processing contractor, which will have a dedicated team located in BP's Stavanger office.

It is anticipated that the processing of the first survey will require 3 months for the P-wave image volume and an additional month for the converted-wave image volume. Although these are very aggressive schedules for a field of this size, by the time the first volumes have been delivered, it is probable that the first reshoot will have already been acquired. However, processing of subsequent volumes should be faster as workflows and basic parametrization will have been largely determined on the first cycle. We expect that not many cycles will occur before we can analyze time-lapse differences prior to the commencement of subsequent cycles.

This will also be the world's first time-lapse converted-wave survey. "C-waves" convert from P to S at the reflector; and the upcoming S waves travel undistorted through the gas in the Valhall overburden, enabling decent imaging of the crest of the reservoir. Further, both the travel times and the reflected amplitudes of the C-waves carry rock property information that is different from that carried by P-waves, enabling more sophisticated physical characterization of reservoir changes.

It has been previously established (e.g. Thomsen, "Converted-wave reflection seismology over inhomogeneous, anisotropic media," 1999) that seismic anisotropy (both polar and azimuthal) is present at Valhall and must be taken into account for optimal imaging. The azimuthal anisotropy is likely caused by preferentially oriented fractures, and these are expected to be sensitive to changes in reservoir pressure (e.g. Angerer, et al, "Processing, modeling, and predicting time-lapse effects of over-pressured fluid-injection in a fractured reservoir," 2002). These changes will be monitored through 4-D changes in the azimuthal seismic signatures:

• azimuthally variable P-wave Amplitude Variation with Offset ("4-d P-AVOAz");
• differential velocity of the fast and slow modes of the upcoming shear waves ("4-d C-wave splitting"); and
• differential reflected ampli-tudes of the fast and slow C-waves.
These studies will be conducted offline due to their unconventional nature.

The ongoing merger prize

Prior to the 1999 merger of BP and Amoco, BP was among the leaders in 4-D seismic technology, while Amoco was not. Conversely, Amoco was among the leaders in 4-C technology, while BP was not. At the time of the BP-Amoco merger, it was widely noted that the business portfolios of the two partners were exceptionally complementary. Less widely noted was that their technology strategies and competencies were also complementary and, in fact, enabled and supported the development of novel ways of actively managing reservoirs.

Life-of-field seismic is not a seismic project; it is a reservoir management project. All of the advanced seismic technology embodied in the LoFS data will be only ornamentation unless it drives reservoir management decisions taken on the same time scale as the repeated surveys. This does not necessarily mean an update (at this frequency) of the finite-difference reservoir model, although eventually that should be possible. But it probably does mean that some 4-D changes observed seismically will be unpredicted by current reservoir thinking and that prompt action in response to these observations will lead to more effective reservoir management.

It also may mean that fundamentally new ways of managing reservoirs will emerge. After all, a reservoir manager is most concerned with future prediction of reservoir performance; the matching of history is only a means to estimate parameters (e.g. distributions of porosity, permeability and saturation) that are needed to predict the future. Frequent full-volume 4-D seismic snapshots of the reservoir will become an invaluable aid for estimating these parameters.