Page 138 - Intelligent Digital Oil And Gas Fields

102 Intelligent Digital Oil and Gas Fields

• Implementation of large-scale, Big Data-driven advanced analytics, inte-
grated into role-centric, relevant time workflows.
• Delivery of holistic ability for capture, classification, integration, and
interpretation of all the relevant and disparate data sources (geological,
engineering, production, equipment, performance, etc.), regardless of
the origin or structure.
• The ability to understand advanced analytical trends and correlation
models to quickly and efficiently unlock the “hidden” knowledge from
all data sets—from small data to large scale and complex data as well as
from historic repositories and databases or from fast streaming data.
DOF systems have been used in the E&P industry for several decades and have
beencommonlyknownfordeliveringonthepromiseofgettingtherightdata,
to the right users, and at the right time, for effective asset decision-making,
maximized recovery, and improved operational efficiency. However, the
expansionofBigData,evolutionoftheInternetofthings(IoT)andintegration
of intelligent, virtual sensors requires rapid transformation to an evolving con-
cept of data-driven DOF systems. These trends introduce challenges in the
areas of DOF system architecture, data architecture, and data analytics and
invite the following questions, which this chapter aims to answer:
• What data architecture is needed in the data-driven DOF to accommo-
date the ever-increasing demand to leverage the real-time sensor data,
that is, the IoT, across the asset?
• If real-time analytics are a must, what are the challenges related to the
quality of sensor data and its integration with the historical data for
closed-loop analytics and how do we overcome them?
• How will disparate streaming data and even unstructured data, regardless
of structure and origin, be integrated and analyzed and how will it be
used in real-time automation systems monitor and act?
• What place does data analytics have in the DOF and how can data-driven
models be seamlessly positioned and integrated with the physics models?
Data integration is the first step to use data generated by various sources.
It is the common notion that the engineers spend up to 70% of their time
searching for data, performing data QA/QC, and reformatting data for ana-
lytics and modeling routines. Moreover, engineers spend anywhere between
10 and 20 working days manually collecting performance data for annual
reservoir performance reviews. The complexity of data types by activity,
incumbent in data-driven DOF operations is captured in Table 4.1.
In a DOF environment, data integration must occur quickly to generate
value, but this requirement creates challenges as data volumes grow and vary