Advances in streamer technology are reducing uncertainties in repeat seismic surveys.

The growing use of time-lapse (4-D) seismic technology to monitor changes due to production within the reservoir has necessitated a change in thinking as to the fidelity required from the seismic measurement.
In 4-D seismic production monitoring, changes in fluid saturation and pressure are inferred from changes in the measured seismic signal. These inferred changes then may be mapped over the field to assist in identifying swept and unswept reservoir zones. Time-lapse monitoring requires that changes in the seismic signal associated with production be significantly greater than the uncertainty associated with repeating the seismic signal measurement itself. Here, "uncertainty" means the variation in seismic signal unrelated to changes in the reservoir. Several factors may contribute to uncertainty in the repeat seismic measurement (Figure 1). Uncertainties introduced by the acquisition system may be grouped into three major components: source, receivers and positioning. Uncertainties of more than -20 dB to -30 dB in the signal strength typically may be considered significant for time-lapse seismic surveys.
To reduce the uncertainties in the time-lapse seismic signal, the processing chain usually includes steps to cross-equalize successive time-lapse seismic measurements in an attempt to minimize differences in the signal outside of the zone affected by production. Clearly, it would be preferable to remove the root cause of these differences, where possible, by modification to the acquisition system.

Source-related uncertainties
Variations in air-gun array source signatures come from two groups of causes: first, variations in array output due to changes in gun timing, pressure, source geometry and gun dropouts; second, variations in sea wave height resulting in changes to the source "ghost" (i.e., the reflection of the emitted source signature from the sea surface). Combining the direct source output and the source ghost forms the complete source signature.
Processing methods are available to correct for many of these variations in source signature if high-fidelity, near-field measurements of the individual gun outputs are available.

Receiver-related uncertainties
In conventional streamers, groups of about 24 sensors within a 52-ft (16-m) length are summed within the system electronics to form an analog group. Typical variations in the grouped sensor output can become significant in the context of time-lapse monitoring. Next-generation acquisition systems overcome these variations by recording the outputs from individual high-fidelity sensors along the streamer (with capacity for several thousand individual sensor recordings per streamer) together with their individual calibration factors.
During towed streamer acquisition, movement of the streamers through the water creates considerable amounts of noise. This streamer noise is characteristically high-amplitude, relatively low-frequency and with low propagation. The streamer noise varies considerably as a function of time, and hence, unless removed, may be a major contributor to variations in the time-lapse signal.
With conventional analog groups, attenuation of the streamer noise is often incomplete, resulting in breakthrough of sporadic, high-amplitude noise bursts. With finely sampled single-sensor recordings, it is possible to employ digital processing techniques, in the form of noise-adaptive filters, to more effectively remove this noise.

Positioning-related uncertainties
Positional uncertainty contributes to significant uncertainty in the final time-lapse signal, from incorrect positioning of the final image and more subtly by deteriorating the final image. Typical uncertainties in receiver positioning, relying on conventional front and tail-buoy global positioning system networks linked by compasses sited along the streamer, may be as large as 50 ft (15 m) for typical streamer lengths. In contrast, a full acoustic network, as is available in next-generation acquisition systems, would yield a receiver position uncertainty of 15 ft (3 m), which is more in line with the expected image resolution at reservoir level.
Spatial variations in the geology overlying the reservoir also can impact the recorded seismic signal. To prevent this imprint of the overburden from varying between time-lapse datasets, source and receivers ideally should be in the same position from one survey to the next. For the source, towed near the seismic vessel, control of positions from one survey to the next is not too difficult. But currents and variable sea conditions can cause the streamer to be laterally displaced by hundreds of meters. In conventional systems, the best that can be done to repeat streamer positions is to shoot in similar tidal and sea conditions and rely on vessel steering for coverage. Next-generation acquisition systems position the streamers with steering devices along the streamer, allowing it to be steered sideways to better match positions obtained in a previous survey.
Repeatability case study
Demonstration of the repeatability of the seismic experiment for time-lapse monitoring purposes may be achieved through acquisition of repeat lines during a period in which there have been no changes in the reservoir. One such example, from a recent time-lapse 3-D survey in the North Sea, is shown in Figure 2.
These lines were acquired with the WesternGeco Q acquisition system, featuring the developments for improved source, receiver and positioning repeatability. The 3-D survey was acquired using streamer steering to maintain near-zero feathering of the streamers, facilitating positional repeatability in subsequent surveys. Survey-wide feathering was maintained between -1° and 1°, compared to -8° to 8° degrees achieved with heritage data over the field. The difference between the repeat lines is shown on the right. Little, if any, information relating to the reservoir zone geology is contained within the difference plot, the majority of the energy being related to scattering from the overburden. The difference between time-lapse datasets commonly is measured by the normalized root mean square difference (NRMS), a measure of the difference in energy contained within two datasets. The NRMS difference in this example is about 20%, a significant improvement over values for other North Sea datasets of about 40%.
The enhancements to seismic time-lapse signal fidelity, achieved through addressing the fundamental limitations of conventional system design, have a significant impact on the repeatability of the seismic measurement. The reduction in time-lapse uncertainty achieved by next-generation systems will significantly impact seismic time-lapse measurements for reservoir monitoring.

Comments