Page 289 - Intelligent Digital Oil And Gas Fields

Integrated Asset Management and Optimization Workflows 237

linked to a common database, span long production and acquisition
times, and have not been adequately and consistently quality checked.
4. For large-scale IAM projects, the reservoir simulation run times can still
represent a major bottleneck. The fine-scale integrated reservoir simula-
tion models can easily exceed 100 million grid cells. Even when upscaled,
the model size usually remains well above a few tens of millions grid cells.
As recently indicated in Maucec et al. (2017), an ensemble-based uncer-
tainty quantification (UQ) and AHM workflow alone, embedded in the
large-scale IAM project (grid-size: approximately 34 million cells;
24 uncertainty scenarios; production period: approximately 55years)
can easily require the simulation model to run sequentially over
20–30h, usingthousands of CPU cores. Ideally, the UQ-IAMworkflows
in large-scale IAM projects will be executed on dedicated HPC frame-
works, which in practice can be frequently difficult to achieve.
5. Last but not the least, demonstrating the value of IAM application in
large-scale mature fields can sometimes be quite challenging. IAM in
mature fields can be considered to have less analytical value because of
lowering operating surface pressures, already existing facilities, known
well performance, and well-described subsurface geology. However,
when applied to mature fields, the IAM workflows can aid in more accu-
rate estimation of the remaining reserves, reduce risks through better
understanding of the interaction between subsurface and surface, indi-
cate real opportunities for optimization scenarios, and facilitate better
decisions for field development planning or well operations.
The push toward the DOF of the future will require a new generation of
IAM workflows that seamlessly integrate the plethora of software tools
for modeling, simulation, prediction, and optimization of an asset’s perfor-
mance. Moreover, the next-generation IAM must be able to operate under
the extreme conditions of integrating unprecedented complexity of data in
terms of the volume and transfer speed. And finally, teams will have to orga-
nize in new ways around the processes.
We have investigated several IAM concepts that may be considered the
future of IAM workflows for DOF applications. Zhang et al. (2006) describe
the model-based framework for oil production forecasting and optimiza-
tion. Some of the key objectives of the proposed design are: generic, reus-
able, and flexible IAM framework architecture; adoption of a variety of assets
and workflows; centralization and transparency of a common database; uni-
fication of information that handles the disparity in data formats; tool inte-
gration through loose coupling; and standards-based implementation.
Zhang et al. (2006) propose the use of domain-specific modeling language