Clearwater Paper: not your run of the mill PI System deployment
Clearwater Paper is no stranger to the PI System. They've used the solution for a number of years, and the PI System is constantly collecting real-time data from each of their paper mills. As the largest provider of private label tissue paper to retail grocery chains in the United States, and a manufacturer of bleached paperboard, their paper mills press and cut paper on a massive scale, and at all different grades. PI Tags located on each mill feed real-time data for every single grade into the PI System so Clearwater can understand asset performance.
For each tag, a target and limit is typically set in order to identify if a particular function is operating in line with past performance. If there is a deviation, a notification is sent so the issue can be fixed before catastrophic failure occurs. However, according to Troy Foede, the West Region Black Belt at Clearwater Paper, it was difficult to put targets and limits on the thousands of tags given the sheer volume, so the majority of tags were simply collecting information.
At the 2017 OSIsoft Users Conference in San Francisco, Foede and John Antanies, Founder at Envoy Development, detailed how the combination of the OSIsoft PI System and Envoy Process Monitor (EPM) gave them increased visibility into tag data. Rather than manually set targets and limits on each of those tags, Clearwater opted to aggregate this data using the EPM to see which tags were operating outside of statistical averages.
Gaining Visibility into PI System Tag Data Using Statistical Averages
Having visibility into these averages is incredibly important because even the slightest adjustments can drastically affect mill performance. These averages can identify deviations such as if a setting is slightly off or if a part is about to fail. If left untreated, these deviations can cause a break, or pause in production, can make paper quality suffer, cause other manufacturing issues, or even the catastrophic failure of a paper mill.
While many of us took statistics during college, any mention of CUSUMs or variances were often met with eye rolls and questions as to how this would be relevant in the “real world.” Now, fast forward a few years (or decades), and statistics are becoming increasingly important as companies look to better understand operational equipment performance. For Clearwater, applying statistics to PI Tag data is exactly what allowed them to understand the thousands of tags that were feeding data into their PI System and quickly identify the reasons behind performance deviations in their paper mills.
In statistics, CUSUM charts show inflection points in data, and those inflection points indicate a change. When Clearwater looked at mill CUSUM performance charts, they notice inflection points that signified breaks in production on October 5th, and another on December 6. Using the Envoy Excel Add-In, they looked at all 3000 tags that were present on the paper mill. By sorting the tag data by percent deviations, they were able to identify which ones were operating outside of the normal range. Of the 3000, 1025 had deviations during that period. When they filtered the remaining tags by CUSUM inflection points, they were quickly able to identify the two tags that were responsible for the break.
Those two tags indicated that a mill operator has changed settings on the equipment. Clearwater was able to work with that operator to ensure that settings were correct in the future, which can help prevent costly production breaks.
A Comprehensive Solution
Clearwater uses a combination of PI DataLink, PI Asset Framework, PI Vision, and PI ProcessBook SQC to gain visibility into their operational data. With the help of the Envoy EPM, they have even greater visibility into their PI System tags. With this comprehensive solution, they can identify potential issues before they happen, which saves them time and money, and keeps paper production mills up and running.