Professional Documents
Culture Documents
00053004practical Issues of 4D Seismic Reservoir Monitoring What An Engineer Needs To Know
00053004practical Issues of 4D Seismic Reservoir Monitoring What An Engineer Needs To Know
surveys by the anticipated quality improvement compared to the risk assessment is a staged decision-point process with increasing
existing 3D images of structure, fault definition, and stratigraphy in degrees of technical sophistication.
the reservoir zone. This reflects the general seismic technology The first step is a quick spreadsheet evaluation of the first-order
advancements made in the past decade or more. Although image seismic and reservoir variables critical to the success of any 4D
quality improvement is not strictly a business driver for 4D seismic project.15 An engineer can expect the success of 4D seismic if a
on its own, it can certainly help move many 4D project candidates specific reservoir passes three critical tests. Is the reservoir rock
into more attractive economics when the data acquisition and highly compressible and porous? Is there a large compressibility
processing satisfy multiple exploration and development objectives. contrast and sufficient saturation changes over time between the
fluids to be monitored? Is it possible to obtain high-quality 3D
Will 4D Seismic Work in My Reservoir? seismic data in the area? The spreadsheet expands upon these three
Before undertaking a 4D seismic project, a feasibility and risk questions by focusing on the most important reservoir and seismic
assessment study should be performed to estimate the likelihood parameters. The important reservoir parameters are dry rock com-
that 4D seismic will be able to image the desired production effects pressibility, fluid compressibility contrast, fluid saturation change
and to constrain a 4D project decision analysis. The feasibility and (sweep), porosity, and predicted seismic impedance change. The
surveys, perhaps spaced at multiyear intervals, one needs to con- ining how small the 4D seismic difference anomalies are outside the
sider carefully whether the remaining lifetime of the field offers a reservoir zone compared to anomalies within the reservoir zone,
valid time window and economic basis to follow up with a per- and by seeing how well anomalies within the reservoir zone
manent seismic monitoring installation. correlate with the geologic model and engineering production data.
Practical issues related to 4D seismic data processing include The third major 4D seismic risk, after data acquisition and
enhancing repeatability and maintaining processing consistency processing, is seismic interpretation. Given that we have deter-
among multiple seismic datasets, preserving true relative amplitude mined a 4D seismic anomaly to be real and not merely an artifact
information, reducing turnaround time for processing new vintages of data acquisition and processing, these anomalies need to be
of 4D seismic data, and post-stack cross-equalization of processed interpreted in terms of meaningful reservoir changes, such as oil
3D seismic image cubes before interpretation. Image consistency is saturation, pressure, temperature, rock properties, or all the above.
enhanced by a post-migration image-processing technique called To do this, the 4D seismic data must be reconciled with the
cross-equalization, which is critical when seismic surveys are not reservoir geology model and the engineering production data
acquired or processed identically. We find that no matter how through seismic simulation. This requires a revamping of the
carefully we acquire and process 4D seismic data, cross-equaliza- traditional reservoir management workflow.
tion is almost always needed to identify and enhance 4D seismic We show, in Fig. 13, a conventional reservoir management
anomalies. Fig. 11 shows that the difference image above the process workflow, including reservoir geology and production data.
reservoir zone in two West Africa time-lapse surveys is very noisy The main work in the flow focuses on building a reservoir model,
before cross-equalization. Ideally, there should be no difference obtaining a production-history-matched flow simulation, making
except in the reservoir zone in the lower left of each panel. This reservoir management decisions based on the results, acquiring new
figure shows that any differences in the reservoir zone are likely to data, updating the model, and iterating the loop. Often, seismic is
be meaningless artifacts of nonrepeatable noise. After cross-equal- used sparingly in this workflow only at the outset to define the
ization, the difference section on the right in Fig. 12 shows much initial reservoir container shape, and perhaps as a soft constraint on
more coherent differences associated with the reservoir zone and the population of the model with heterogeneous reservoir proper-
low-energy incoherent noise above the reservoir. These reservoir ties. Fig. 14 shows our proposed new model for reservoir man-
anomalies now correlate meaningfully with known pressure and agement that includes the 4D seismic data as a reservoir model
fluid movement caused by water injection on the left downdip constraint through seismic simulation. Now, seismic is an integral
flank. Engineers should be wary of false anomalies in 4D seismic part of the reservoir management process. A successful reservoir
images arising from acquisition nonrepeatability and time-lapse model not only has to match the geology and engineering produc-
processing inconsistencies. Also, engineers should be wary of the tion data at well locations, it also has to match the 4D seismic
fact that all seismic data cross-equalization algorithms may not information in the interwell volume regarding fluid contact move-
perform equally well, because cross-equalization is a relatively new ment, bypassed oil, progress of injected fluids, pressure compart-
processing technique that is still rapidly evolving in the seismic mentalization, and the hydraulic properties of faults and fracture
industry. The engineer can assess these processing risks by exam- zones. We call this new workflow process seismic history match-
data, and seismic data, especially when a good production history We have performed a cost analysis that is summarized in Table
match alone may be ambiguous. The lesson is that 4D seismic 2. Our approach is as follows. We considered more than 15
anomalies must be reconciled with some version of the seismic Chevron-operated 4D projects around the globe, representing many
history-matching workflow in Fig. 14. If they are not, 4D seismic geographies, reservoir geology types, and production methods. We
anomalies may be misinterpreted as production effects when, in categorized these as Indonesia, West Africa (Angola and Nigeria),
fact, they could be false anomalies related to seismic noise, acqui- West Texas (with similar values for the San Joaquin basin in
sition nonrepeatability and processing inconsistencies, unexpected California), North Sea, Gulf of Mexico (shelf), and Deepwater.
rock, fluid, and pressure influences on the seismic response, and Next, we considered the typical costs of well work in each of those
could represent gross inconsistencies with the reservoir model and six production environments, including recompletion, sidetrack or
its constraining data. redrill, and a new well. We considered mainly vertical or deviated
wells, because horizontal wells can have highly variable costs and
What Does 4D Seismic Cost? are not drilled in all environments. Then, we considered what the
We now discuss cost issues related to 4D seismic monitoring. fundamental unit of seismic acquisition is in each environment and
Ideally, we should be discussing value. However, it will take the its cost. For example, on land, 1 sq mile (640 acres) of seismic data
industry some time and many case studies to gather reliable is a typical minimum amount of seismic data to acquire. However,
statistics on the value of 4D seismic for many reservoir types and offshore, 9 sq miles [one outer continental shelf (OCS) block] is the
production methods. Because cost is half of the value equation and typical minimum amount of seismic data to acquire. Then, we
the first item to hit the budget, we focus on cost issues for now, with calculated what the typical amount of well coverage is in each
the knowledge that each 4D project must pass some first-order environment for the seismic area covered. Finally, we decided
business driver and anticipated payout test before it is implemented. which type of well work was most likely to be performed in each
TABLE 2—RELATIVE COST (IN THOUSANDS OF U.S. DOLLARS) OF 4D SEISMIC TO WELL WORK (SEE APPENDIX FOR
DETAILS)