Page 398 - Intelligent Digital Oil And Gas Fields

336 Intelligent Digital Oil and Gas Fields

clusters to speed the CPU time. However, the E&P industry faces many
challenges to integrate Big Data with big models, including:
• Enough storage capacity to submit more than 100 realizations/scenarios
to the cloud.
• Parallel multiple simulation jobs without increasing cost.
• CPU scalability and acceleration to reduce CPU.
• Budget constraints. Technology for hardware and software is available,
with literally thousands of economic options for cloud and cluster
environments.
• Decisions about which real-time data should be integrated into big res-
ervoir models. Monthly production data are enough for 3D reservoir
modeling. However, the water and gas breakthrough could occur in
weeks. The simulator is capable of predicting when fluid breakthrough
will happen and what action should be taken to prevent it.
Do we really need to integrate Big Data with big models? In many exper-
iments, we observed potential discoveries and insights that were not
observed with upscaled processes. Stochastic analysis is the key to run Big
Data-big model to explore the impact on production forecast and oil recov-
ery, especially when uncertainty plays a fundamental role.
Fig. 9.5 is a schematic for the integration of a big reservoir model and
Big Data, applying production data to update the model, running many
scenarios for production forecasts, generating intuitive diagnostic and anal-
ysis of production downtime, extracting data for data analytics, and show-
ing where to drill, complete, and optimize well production performance.
It will be one common platform to capture real-time data into the model to
generate scenarios rapidly and rank economic decisions; the future plat-
form will provide intuitive workflows without coding or mapping individ-
ual properties to connect different software applications. It is envisioned
that the platform will generate cognitive diagnostics to rank solutions
according to events and well issues. Models will integrate physics-based
and data-driven responses.

9.6.2 Optimizing Optimization and the “Closed Loop”
Chapter 6 presents the state of the industry in optimization and introduces
the closed-loop concept. The future DOF will harness and benefit funda-
mentally from advancements in the process optimization technology.
According to Pallav Sarma, an integrated monitoring and control approach
known as model-based closed-loop optimal control has to be implemented