Page 393 - Intelligent Digital Oil And Gas Fields

The Future Digital Oil Field 331

9.4 AUTOMATION AND REMOTE CONTROL

Dr. Demyanov stated that analyzing, inferring, and making decisions
based on the new monitoring (see Nanosensors above) would be the next
task that would require a step-change level of data analytics implemented
into reservoir surveillance workflows. Can we speculate on the transition
toward artificial intuition embedded in DOF technology so the O&G digital
transformation could result in a self-thinking DOF ecosystem? Demyanov
notes that artificial intelligence and cognitive systems have been widely used
in mining large data domains and supporting expert judgment by elicitation
of vital information patterns from data. Such data-driven methods have been
historically implemented in seismic processing before gaining a wider rec-
ognition in other O&G application domains. Nowadays, the cognitive
approach is intensively implemented in many diagnostic systems that require
interaction with human activity/technology, such as medicine, electronics,
etc. There is great potential in implementing data-driven cognitive learning-
based technology in DOF systems. The present challenges lie with a tech-
nological level of storing and mining the right amount of information at the
acceptable cost level. Furthermore, the cognitive systems to be developed to
mine reservoir operational information should embed experiential learning
based on decades of manual operational success and failures. Such technol-
ogy already exists (e.g., in mechanical manufacturing) where mechanisms
are able to capture and learn from the operators’ experiences. However,
the more complex the mechanisms are, the more elaborate and sophisticated
the cognitive learning needs to be, especially given the vast uncertainty in
hydrocarbon reservoirs and the variability of many different options and pos-
sibilities of events.
Demyanov concludes that there is a great potential in linking the step
change in data acquisition in the DOF era with the novel data-driven
learning-based workflows—as the second is the solution to extract the added
value from the first. It is essential to view the high-level problems of where
the value is in the O&G operations and what are the bottlenecks where the
value is depreciated for certain reasons. Thus, the efforts need to be aimed at
those bottlenecks to explore the opportunities of intelligent data-driven
workflows that would be aimed at gaining the qualitative step change in
the way the DOF systems are operated. Present-day practices and technol-
ogy allow monitoring social behavior and reveal certain dependencies and
even predict behavior events based on the monitored response of a swarm