Process Unit Template for Smart Thinking

In preparing for this conversation, Peter and Monica wanted to communicate the importance of sharing a common nomenclature to describe production processes at the refinery. This was a critical issue: Currently, every unit had a different variable name, making it impossible to benchmark production performance and uncover operational constraints.

Peter noted, "The digital transformation team needs to develop a process unit template for the digital data infrastructure project." This modular, reusable template will define the data hierarchy and describe asset content for applications using the digital data infrastructure (see the box "About Process Unit Templates"). It will map the refinery data collected to applications that analyze operations and provide visibility into key areas, such as safety and security, energy utilization, process optimization, asset health management, quality improvement, and regulatory compliance.


A process unit template is an asset object model that contains the necessary contextual information about an asset as recorded by the digital data infrastructure. This information can be in the context of an important process event or the operational mode of an asset. The digital data infrastructure organizes the data and events for real-time and post hoc analysis. It enables raw data to be transformed into valuable information with the right level of detail. This is discussed in more detail in Chapter 4.

Time is a key variable for the digital data infrastructure project, used for key data aggregation, KPIs, KPI management, predictive analytics, and advanced data analysis. The digital data infrastructure enables the automated recording and analyzing of events, including event start and end times. It then transforms data within these events into actionable information. The key is to use the digital data infrastructure system to automate the capturing and analysis of events. Most relational databases do not handle time-series data well, so Proclndustries plans to use the EIDI with their production data model for machine learning (ML) and artificial intelligence (AI) (see the box "Challenges in Time-Series Data in Refinery Information Systems").


Despite the universal recognition that data is critical to smart operations, rarely does operational data inform decisions at all levels of the enterprise. Why?

  • • Operational sensors and systems produce massive volumes of data and often have process control automation systems that do not communicate with one another. Enterprise data records are often incomplete, fragmented, and frequently inaccessible to many users.
  • • As connected assets and increased connectivity lower the barriers to capturing even more data, most systems cannot scale to handle increased data volume.
  • • Traditional data archives lack contextual information that add value to data shared throughout the enterprise. Without context, valuable operational information often remains underutilized, sequestered, or unavailable to users unfamiliar with control-system naming conventions.
  • • Data collected and stored by isolated point solutions have disparate sources or formats. Reporting, calculations, and roll-ups require manual data entry, are prone to error, and are time and labor-intensive.
  • • Many interfaces that collect the data are at the asset location. Accessing information from remote sites or from centralized centers can be challenging, and delay or prevent timely use of valuable information.
  • • Connecting data from siloed point solutions, applications and data historians require skilled resources and customized solutions, which adds information technology (IT) complexity and cost.

With this technology, Proclndustries can isolate refinery events when there is a problem in one of the process units. For example, if a process unit experiences trouble for 15 minutes, workers can mark the start and end times of this event. They can also calculate the total amount of time the process unit is in trouble and estimate how that trouble affected the totals for the consumable variables that were configured in the process unit template.

Every shift can estimate the total amount of consumables (e.g., energy, water, and catalysts) used while deviating from target. They also can identify and totalize the times that the unit is down, idle, or under maintenance, and time on target. With the ability to track data for all of the variables and defined events of interest, they can calculate the monetary value of more quickly resolving problematic events.

< Prev   CONTENTS   Source   Next >