Page 35 - Intelligent Digital Oil And Gas Fields

16 Intelligent Digital Oil and Gas Fields

1.5.3 Workflow Automation
Traditionally, geoscientists and various engineering disciplines (produc-
tion, reservoir, facilities, etc.) spent considerable time gathering data from
disparate sources for input into their mostly manual workflows. Engineers
generally use models developed in commercial software applications to
reproduce the oil production process. However, even these software models
required complex manual workflows that consumed engineers’ time, for
example: collecting data from different sources (spreadsheet, text, tables,
figures, historian, etc.); filtering data from noise; performing repetitive,
error-prone tasks to update models (e.g., manual data entry); reconciling
the data and calibrating the model; and running different scenarios of the
model.
Workflow automation uses high-level programming language routines
to connect these manual processes, so that models can be automatically
populated and updated. Automation is just part of the DOF requirement
for workflow construction. DOF solutions also require that engineering
workflows are intelligent enough to capture in real time alarms and alerts
to generate prompt actions, update engineering applications, and deliver
right-time monitoring, diagnostics, and process optimization that deliver
operations guidance at the field level.
Moreover, the workflows should have a predictive character and capa-
bility to foresee future operations issues. For these complex tasks, DOF
workflows must include sophisticated language program, like artificial
intelligence components such as pattern recognition, fuzzy logic, neural
networks, proxy models, and optimization supported with advanced multi-
variate statistical analysis that can generate reliable short- and long-term
forecasting. Chapter 4 discusses the concepts of these data analytics.
Chapter 5 discusses the main components of workflow automation,
which includes these key concepts:
• Workflow (WF) foundation and philosophy.
• WF types, such as single, integrated, automated, smart.
• Workflow focus including well-centric, task-centric, KPI-centric, and
facility-centric.
• Factors that control WFs, such as data- versus model-driven WFs.
• Physical models such as empirical, analytical, and numerical models to
serve data reconciliation.
• Virtual models such as virtual metering system when actual metering is
not available.