Data Analytics Innovations Improve Refinery Operations

March 16, 2016 – As published in the AFPM 2016 Show Daily (Day 3)

Refineries often find themselves data rich and information poor, with millions of data points stored in historians and no easy way to extract value from this information (FIG. 1: Traditional data analytics tools such as spreadsheets are general purpose and not specifically designed for extracting value from historian data.). This situation is changing; Seeq and other companies have recently introduced data analytics software specifically designed to enable insights and extract value from the time-series data stored in historians such as OSIsoft PI, Honeywell PHD and InfoPlus 21. There are several examples of how data analytics solutions are being applied to address specific refinery issues.

Better energy management delivers bottom line benefits. In a refinery, energy accounts for a significant share of operating costs, so even small gains in efficiency go straight to the bottom line. However, most of the low-hanging fruit has been picked, leaving the more difficult and challenging projects that can achieve significant energy savings.

 AFPM 2016 Show Daily

Refinery energy inefficiency sources include waste heat, steam systems and furnaces. Other sources of inefficiency are found in atmospheric distillation, vacuum distillation, delayed coker, hydrotreater, hydrocracker, catalytic reformer and alkylation units. Each operating unit produces huge volumes of data that is usually stored in information systems owned and managed by the operating disciplines. In many cases, this information goes unused.

A leading refiner is addressing this issue by deploying Seeq software to eliminate the excess effort required to visualize time-series data and align it with other contextual information sources. The refinery engineers can now search for historical patterns and uncover inefficiency signatures, process problems and potential equipment failure. More than 100 opportunities to reduce and suppress energy loss have been identified and addressed. What used to take days or weeks of analysis can now be accomplished in minutes or hours.

Increasing profits by using lower cost feedstock. A major refiner in the Midwest was looking for ways to increase profitability, and found they needed to make quicker decisions to exploit changing market conditions. Despite having what they thought were best-of-breed tools, decisions were taking too long and market opportunities were being missed.

Midwest refineries have typically been designed to accept lighter crude feedstock. Heavier feedstock occasionally trades at a discount, creating an opportunity for refiners to improve margins by sourcing this lower-priced supply. However, the information systems within many refineries make it difficult to make timely decisions about expected yields when accepting off-spec feedstock.

To make an informed decision, engineers need to quickly answer the following questions: 

  • How far off-spec are my column products?
  • How much does this deviation cost?
  • How do the costs compare to the profit opportunity?
  • What changes need to be made to process the feedstock?
  • Is there a set of operation specifications to process the crude? 

To answer these questions, data was imputed into modeling tools, with results transferred into complex home-grown spreadsheets for analysis. With this traditional approach, missing data was a common occurrence, requiring the engineer to use a best-guess approach to fill in the gaps, or spend additional time to gather historical data from another point in time to rerun the model.

The refinery’s engineers can now use Seeq to perform advanced analytics and extract actionable insights from their data in a timely manner (FIG. 2: Data analytics software specifically designed to interact with the time-series information found in historians allows process engineers and other specialists to quickly analyze and compare temperature and other process parameters.). Using context-based search capabilities similar to modern web search engines like Google, internal operating properties such as flow data, column temperature and pressure, heat exchanger inlet and outlet temperatures, pressure drops and boiling point curves are quickly gathered over a specified time period.


AFPM 2016 Show Daily Figure 2

Data is automatically checked for inconsistent and missing information and quickly reconciled if necessary. Unit of measure consistency is also ensured, eliminating the need for spreadsheet transformations. Search results are fed directly into the refinery model to analyze historical operating states for crude feedstock changes under current operating conditions, and cost-benefit scenarios are stored for each time period. Results can be recalled and visually compared against each other, providing an overview of the process through time.

Predictive maintenance cuts costs and increases uptime. The reliability team at another refinery was seeking to avoid unplanned equipment outages. Maintenance of equipment was required periodically, but performing these tasks on a calendar basis was inefficient. Some items were serviced too frequently, adding to downtime and costs, while others weren’t serviced often enough, negatively impacting quality. Refinery maintenance management system information was frequently out of date.

Applying Seeq to data already resident in a historian, plant maintenance personnel now use pattern-based anomaly detection techniques to identify the actual state of equipment or a process. This identifies precursor events to unscheduled equipment outages and anticipates emerging process problems.

With real-time trending in place, equipment performance is monitored continuously, preemptive measures are taken when maintenance is required, and maintenance intervals are stretched out wherever possible.

The data analysis required to predict problems before they occur is quite complex, as it requires looking at real-time and historical data in context with work orders, notifications, incidents, operator logs, alarms and events, corrosion, inspection and equipment data to build a picture of conditions that cause failure.

Process and reliability engineers most familiar with the equipment and the operating units perform this complex analysis using Seeq, creating data signatures by using information from sensors monitoring equipment known to be in good running order, tracking deviations from these signatures, and automatically alerting plant personnel before problems occur.

The refinery’s engineers now have affordable data analysis tools to prevent unplanned capacity losses.

Breaking away from traditional data analytics tools. Refineries often possess all the information they need to improve operations within their data historians, but creating insight from this information can be difficult, expensive and time-consuming using traditional data analytics tools.

Seeq now offers data analytics software specifically designed to interact with the time-series information found in historians. Because this software is designed for just one task, as opposed to a general purpose tool like a spreadsheet, it provides a faster and less expensive solution.



Bring discovery and decision support to your organization.

Read More

Request a Demo

Get insight from your data.

Read More


Webinars, case studies, and short videos to get you up and running with Seeq.

Read More


Seeq is adding value in many industries.

Read More


We welcome partnership

Read More


We are entrepreneurs with a passion for developing innovative technology.

Read More