Time-lapse methods are becoming increasingly useful in understanding reservoir dynamics. Any parameter that changes slowly over time can be studied by comparing sequential measurements. These measurements may be made in wells or remotely on the surface and can provide useful information to update reservoir models used to manage a field. Particularly successful offshore, time-lapse surface seismic, popularly called 4-D, provides an example of such remote measurements that have matured into a commonly used technique.

Onshore time-lapse seismic is more difficult. The critical parameter affecting 4-D quality is measurement repeatability. Complete replication of source and receiver positions is often challenging due to construction of surface facilities such as roads, buildings, wells, and pipelines during the interval between surveys. Near-surface changes such as those due to seasonal fluctuations in the water table level add to the complexity of acquiring repeatable surface seismic data onshore.

The challenge

After 26 years of cold production, South Oman’s onshore Haradh formation was chosen for a steam-injection program to reduce viscosity and drive additional hydrocarbons to production wells. The field consists of a four-way dip closure. Lying about 1,000 m (3,281 ft) beneath the surface, the reservoir is about 200 m (656 ft) in thickness. The Haradh formation holds the bulk of the oil reserves in the field. Three cross-well seismic profiles and a 3-D vertical seismic profile (VSP) survey were acquired in a time-lapse mode. The baseline surveys were conducted in 2009, with the first monitor surveys following a year later. A subsequent second monitor 3-D VSP survey was conducted in 2012. At the same time, time-lapse surface seismic was piloted to evaluate its repeatability and ability to image the steam flood. The recognized risks are strong surface reflections from shallower horizons, weak target reflectivity, poor repeatability of source and receiver positions, near-surface changes, and considerable surface “noise” from ongoing drilling and production activities. These risks make standard 4-D surface seismic acquisition unlikely to usefully image the steam flood.

Time-lapse 3-D VSP

Truck-mounted vibrators crossed the field, stopping as close to each baseline source position as was physically possible given the surface obstructions. VSP acquisition was performed by Reservoir Imaging Inc. Due to changes in surface facilities between the baseline and repeat surveys, 70% of the source locations were replicable. Acceptance criteria stated that deviations from baseline source points of more than 2 m (6 ft) were grounds for rejection of the measurement. Nevertheless, a good representation of the steam-front movement was obtained with this reduced number of source points.

Ultimately, 2,443 colocated shots out of a possible 3,500 shot locations were used for a time-lapse 3-D VSP analysis. The shots were recorded downhole by a three-component 109-level tool, with 73 receiver levels eventually used for time-lapse processing and imaging using Shell internal and Schlumberger Omega software packages.

Survey-to-survey repeatability was measured using the root mean square (RMS) repeatability ratio (RRR) attribute, defined as the RMS of the difference normalized by the RMS of the baseline and monitor data in a given window, in this case 60 m (197 ft). RRR was checked at a strong reflector above the reservoir where no changes were expected. The difference was found to be only 0.1 in the area of interest, giving confidence to the significance of 4-D changes observed within the reservoir (RRR of 0.3).

Time-lapse cross-well tomography

Fortunately, field geometry lent itself to cross-well evaluation. Six producing wells were arrayed in a hexagonal pattern around a single steam-injector well. Temperature logging was run on a monthly basis in three observation wells located within the hexagonal pattern. Cross-well seismic, which uses downhole sources and receivers, is not constrained by changes in surface facilities and, therefore, maximizes the potential repeatability of the acquisition.

To perform the cross-well surveys, piezoelectric sources and hydrophones were run by Schlumberger in the producing wells over the interval from 800 m to 1,200 m (2,625 ft to 3,937 ft). This resulted in three intersecting profiles across the steam-injection pattern, giving excellent insights as to the propagation of the steam front in the reservoir. Time-lapse differences in acoustic wave travel times as small as 0.5 millisecond were detected. These were attributed mainly to changes as steam or hot water replaced the heavy oil.

Data integration helps confirm results

Because the various measurement physics are somewhat dissimilar, anomalies that are evident in one method may not be seen in the other and vice versa. This suggests an integrated approach. RRR volumes were used in the analysis of 4-D differences since VSP amplitudes are sensitive to offset-dependent illumination, amplitude balancing, and amplitude-vs.-offset effects. An RRR attribute map of the top Haradh formation was integrated with cross-well tomography and time-lapse seismic images to show the full picture of time-lapse differences. Hot wells are indicated by red circles, while cold wells are in white. Arrows indicate the extent of anomalies derived from cross-well seismic. The configuration of the reservoir top relief suggests that the lighter steam would be expected to migrate in a northerly direction.

The RRR attribute map does not show changes in the observation well O8, but cross-well data indicate differences there. It is possible that this is the result of impedance changes in a thin layer of the reservoir. It is postulated that changes at the thin edges of the zone are more likely to be seen on the cross-well time-lapse images because of the cross-well image’s greater vertical resolution. However, the areal extent of anomalies derived from cross-well and VSP is consistent with those observed on temperature surveys.

The path forward

The results of integrated interpretation of various seismic and nonseismic surveillance results led the asset team to the decision of perforating steam injectors about 100 m (328 ft) deeper. The second monitor 3-D VSP was acquired in 2012 to evaluate feasibility of this technology to track steam injection at greater depths for busy and changing field environments. A new anomaly was revealed at the depth of perforation interval. Its outline suggests more uniform heat propagation from the newly drilled injector than in 2010, but still north is the preferential direction.

The use of 3-D VSPs for field-scale monitoring requires a more economic solution, which is given by emerging distributed acoustic sensing (DAS) technology. DAS uses in-well fiber-optic cables for recording seismic waves. DAS has a number of advantages over geophones, such as the possibility of permanent installations and super-long receiver arrays (up to a few kilometers of fiber). Simultaneously with the second monitor 3-D VSP, an extensive experimental program was conducted to evaluate the feasibility of recording usable 3-D VSP with DAS for fibers strapped to tubing. This deployment type is particularly useful in a high-temperature brownfield environment where the fibers are retrofitted to existing wells and could be easily replaced in case of damage. It turned out that even though raw DAS VSP is noisier than geophone VSP, the migrated images are of comparable quality. Given these results, a field-scale DAS 3-D VSP monitoring program is being planned for the coming years.

Time-lapse measurements can allow early prediction of the potential success of EOR projects. While improvement in field recovery factors typically justifies the investment, it is prudent to implement a production management scheme to maximize profitability. The use of cross-well tomography integrated with 4-D VSP imaging and temperature monitoring data has proven to be effective, with the 4-D VSP repeatability levels comparable to those of marine seismic acquisition. By avoiding the potential noise interference from onshore surface activities, downhole measurements have exhibited resolution and time-lapse fidelity superior to that of traditional land seismic imaging in an active and changing field. With DAS acquisition solutions becoming available, field-scale time-lapse 3-D VSP monitoring programs are being implemented.