In recent years, the utilization of time-lapse techniques has become an essential tool for the geophysical monitoring of producing oil and gas fields. As a result, an increasing number of seismic projects are planned and executed over obstructed areas (Figure 1).

In these operational environments, health, safety and environment (HSE) rules are very restrictive. As a result, large regions remain uncovered during repeated acquisitions, and the final quality of the acquired data can severely be affected.

In the context of time-lapse studies, a possible solution to the above problem is the

Figure 1. (top) Sketch of an ultradeepwater development scheme and related hazards for seismic acquisition. In color, (bottom) a streamer coverage simulation boxing an FPSO unit. Nodes can be used to ‘fill the gaps’ that conventional streamers can’t illuminate. (All figures courtesy of Total SA)
utilization of ocean-bottom recording units (nodes). By strategically locating them beneath production facilities, nodes can be used as a complement to streamer data to fill in the gaps left behind following marine operations. By simulating a typical deepwater scenario — undershooting a floating production, storage and offloading (FPSO) installation — we assess the feasibility of such a “nodes/streamer” approach, starting from the design of a joint survey to the interpretation of a finally processed merged dataset.

Background

Nodes are four-component (4-C) autonomous systems that, once positioned on the sea bottom, are programmed to record the wave fields emitted at surface by a towed seismic source. Nodes have many virtues: as they are deployed by remotely operated vehicles, they can be located in areas otherwise inaccessible; acquisition operations at sea are light and HSE-friendly; good positioning and hence repeatability can be achieved; coupling and data quality benefit from the absence of a cable relaying the sensors; and full-azimuth 4-C data can be recorded in ultradeep waters.

An experimental test conducted in 2004 by CGG and Total in offshore Angola (Boelle et al., 2005) has provided extremely encouraging results, demonstrating that high-quality data can be acquired with a relatively sparse distribution of autonomous ocean-bottom nodes.
Joint streamer/ nodes survey design. We consider a typical deepwater FPSO installation linked to a hydrocarbon offloading buoy with 1.2 miles (2 km) separation between the two. In the first phase, a streamer acquisition is designed to box the obstructed area (Figure 1b). This is done in full compliance with HSE restrictions, considering the local currents, the turning radius of the streamer vessel and the safety distances from the production units. Due to the parameters used for this example, a large gap of around 15 sq miles (40 sq km) is isolated.

Once the gap is defined, we fill it with a regular grid of nodes. A dense grid of shots is then designed in such a way that each node is equally represented in offsets and azimuths.
The two approaches exhibit totally different properties, mainly due to the different recording datum (the sea surface for streamer data, the ocean bottom for the nodes). Moreover, while narrow-azimuth information is collected with the streamer approach, wide-azimuth data can easily be recorded with nodes. These differences pose fundamental questions: will a joint migration be possible? Which will be the impact of the footprint on the expected 4-D signal after a full processing workflow?

Joint streamer/nodes data processing

In order to verify that merging data of different natures can generate valid processed results, we run 3-D prestack migration in both depth and time domains.
Figure 2. Prestack time migration results. Left: High-resolution streamer data, nominal fold 52. Right: Merged approach high-resolution streamer data depopulated plus nodes in the central area, nominal fold 4. Light blue squares represent the simulated streamer fold build-up.
Specific modifications are required to make the migration codes work correctly. In particular, Green’s functions are computed both from the water surface and from the sea bottom. For streamer data, surface Green’s functions are used while, for node traces, we use surface Green’s functions for the sources and sea bottom functions for the receivers.
Prestack depth migration is run on synthetic data, generated by ray tracing in a simplified depth model. Results highlight that the kinematics of the wave propagation are correctly handled for the merged data (Ceragioli et al, 2006).

Prestack time migration is run on the real data acquired in 2004 during the Girassol experiment (Boelle et al., 2005). Streamer data are de-populated to mimic the presence of an obstruction. The coverage loss is compensated by node data. The merged data set is then processed through an optimized joint prestack time migration workflow. Results show that, despite the low fold — streamer and node data are grouped in four offset classes only — satisfactory images can be obtained from this merged approach (Figure 2).

Interpretation aspects

In order to evaluate the impact of the joint prestack time migration approach on the
Figure 3. Amplitude extractions for ‘streamer only’ and joint prestack time migration and their relative difference over a channel complex in the center of the study area. Nodes are shown as blue dots.
interpretation of the turbidite fairways, amplitudes are extracted at different reservoir levels of the field. Results of this amplitude-based approach demonstrate that:
• The substitution of streamer by nodes does not hinder a proper definition of the channel system architecture. The imaging quality is preserved despite the scarcity of the data (Figure 3).
• Amplitude differences between the joint approach and the original streamer results are relatively high. In the context of the present test, this impact would go beyond the expected amplitude of a 4-D effect. However, we have to take into account that the node density and their total fold would be much higher in a “full scale” operational case.

Conclusions and perspectives

Studies and operational tests conducted between 2004 and 2006 have confirmed that high-quality data can be acquired through a sparse distribution of autonomous nodes.
A joint survey can effectively be designed, taking into account operational and HSE constraints along with the requirements for the detection of desired time-lapse signals.
The problem of the datum difference between ocean bottom and surface data is resolved through an optimized joint migration scheme. Joint migration images preserve all relevant sedimentological features in the area of our test.

Nodes can successfully be merged to surrounding streamer data to constitute a valid geophysical data set for time lapse studies over key areas of our fields.
It is important to notice that a series of important issues are still being evaluated, namely:
• The repeatability of a node acquisition
• The optimization of pre-processing (noise reduction, data interpolation, multiple attenuation) and imaging workflows (multi azimuth tomography) to take full advantage of the wide azimuth nature of the node data.
• The added value of four-component data for fluid-related reservoir studies.
We are today focusing on these subjects in the continued evolution of this promising technology towards full-scale operations.

References

Boelle, J.L., Granger, P.Y., Ceragioli, E., Crouzy, E., Lefeuvre, F. [2005] Autonomous 4-C Nodes used in infill areas to complement streamer data. A deepwater case study – 75th Meeting, Society of Exploration Geophysics, Expanded Abstracts.

Acknowledgments

The authors would like to thank Total EP Angola for supporting the operations offshore during the 2004 pilot acquisition and the European Association of Geoscientists and Engineers (EAGE) for authorizing the publication of this material. This is a revised version of EAGE paper B042, “Filling the gap: Integrating nodes and streamer data for geophysical monitoring purposes,” EAGE 68th Conference & Exhibition, Expanded Abstracts.

Comments