Process Improvement through Visualization

For process engineers, the EIDI provides real-time and historical data that is vital to troubleshooting and improving the refinery's unit processes. Plant engineers can perform many simple debottlenecking efforts through efficient use of trend displays, specialized charts, and event analysis. The EIDI real-time data management system provides standard, configurable visualization tools for many of these needs. Following are some examples:

  • • Performance trends show targets and current measurements used for root cause analysis (Figure 5.8).
  • • Holistic dashboards and refinery-wide displays use drill-down capability to unit-specific data.
  • • Event-framed displays compare event activity against other events or against an ideal event.


Real-time data trend display for root cause analysis.

  • • Metric predictions are used in variance analysis, comparing current versus expected performance.
  • • Statistical quality control (SQC) and SPC charts can be used. SPC helps ensure that the process is operating normally (Gaussian distribution). If the process moves more than 3 sigma in variance, automatic alerts will be generated because the process is drifting outside its normal variance. There are a number of statistical methods to model the processes and to define faults if the process has deviated from normal, expected ranges.
  • • Linear (x-y) and scatter plots show the results of two time-based data sets, so that a pattern can be deduced.
  • • Actual performance is tracked in comparison to expected or predicted results (e.g., process model predictive data).
  • • Data is classified and aggregated using Pareto charts, displaying cumulative totals for priority analysis.

The refinery engineers and other knowledge workers were then able to extract data, such as the example trends shown in Figure 5.8, and import that and any other EIDI data into Microsoft Excel (Figure 5.9). The EIDI menu in Excel provided a variety of options to populate their spreadsheets with current values, historical archived values, derived values, or statistical data on the archived values. Figure 5.9 demonstrates an example of preformatted or ad hoc Excel reports.

Some visualization tools, such as multivariate charts and fishbone diagrams are deployed outside the real-time data management system, but can leverage EIDI data, such as

  • • Cause-and-effect trends (fishbone diagrams);
  • • Multivariable charts (key metrics vs. manipulated variables); and
  • • Pie charts, bubble charts, and other business-oriented charts.


Creating Excel reports using EIDI data.

Using analytics to create these visualizations, refinery operating people can embed acceptable operating ranges and monitor real-time behavior, using real-time notification alerts when the process strays outside preconfigured upper and lower control limits. This is generally an early warning indication that the process unit is in trouble. Concurrently, there are mechanical limits used for equipment monitoring. The goal is to integrate production and maintenance strategies for optimal performance, that is, goals of no excess process deviations and no unscheduled downtime events. These displays and dashboards are available on desktop computers, laptops, or on mobile devices (tablets, smartphones) connected to the industrial network via tools like a virtual private network (VPN) or other secure access mechanism.

< Prev   CONTENTS   Source   Next >