Article

Data-driven chemistry control: See the future 5 paper reels ahead

A modern paper machine generates huge amounts of data from different points in the production process. But process adjustments for improved performance, efficiency, and end-product quality are still often based on human judgement and the mill operators’ personal experiences. Why?

Data-driven chemistry control: See the future 5 paper reels ahead

“Today, data from the papermaking process is still not efficiently turned into actionable information. It is typically scattered in multiple data collection systems, which aren’t always communicating with each other to create a real-time view of the process. This leaves the full potential of the data untapped,” Juha Rintala, Manager of Digital Applications at Kemira, explains.

He gives an example. Already for a couple of decades, Kemira has developed on-line monitoring tools for chemistry applications. With real-time chemistry usage and performance metrics, they enable smart chemistry management and optimized chemistry control and diagnostics. “If we combine this information with performance and condition data from other points in the papermaking process, we can identify chemistry-related correlations more broadly. And with the use of predictive analytics and machine learning technologies, we can discover phenomena that impact the overall performance and can even predict and prevent costly problems.”

New times ahead for quality control

Currently, the quality control at a paper or a board mill relies mostly on reactive measures. Issues with end-product quality are detected from samples taken from the paper or carton reels, which leads to a significant delay – and hours’ or even days’ worth of poor-quality production – before root causes for the issues can be identified and corrective actions can be taken. But what if you could see hours ahead into the future, predict the process conditions that have an effect on the quality of the manufactured board, and take preventative actions before problems occur?

“With predictive analytics, this can be done already today. For example, a board mill that is utilizing our latest digital solution can reliably predict the deposit risk level in their production five board reels into the future and predict and prevent resulting quality issues and defects on the manufactured board,” Juha says. “The data-driven chemistry control can identify and highlight process disturbances that are around the corner, such as the formation of deposits, and give the mill operators time to take corrective action before the board quality is affected.”

Kemira’s measurements from the chemical state of the process, together with information from the paper machine’s other process units reveal correlations and enable new data-driven ways to control the process chemistry and improve end-product quality.

With the help of the predictive tool, the board mill can avoid thousands of tons of out-of-spec production. Beyond the improved board quality, the continuously balanced wet end chemistry stabilizes the process, which leads to improvements also in production availability and performance.

With the help of the predictive tool, the board mill can avoid thousands of tons of out-of-spec production.

A 0.5% increase in the overall equipment effectiveness (OEE) might not seem like a lot, but for example for a folding-box board machine producing 400,000 tons a year, this small nudge in OEE could already mean a 2.5 million revenue increase.

“And this is just a starting point; machine performance can be boosted more when hidden opportunities for improving process efficiency, product quality, and raw material usage are discovered,” Juha states.

Building on the chemistry expertise

Predictive analysis is also used to improve paper and board machines’ wet-end stability. Currently, a fine paper mill is working together with Kemira on a project to prevent breaks and to increase runnability and efficiency on their machine. Production process abnormalities and causes of breaks are identified from the customer’s historical data with machine learning technologies.

“A holistic view of the performance data and process conditions paints a clear picture of the overall process health. With the predictive model, this enables us to provide the mill production managers and operators with valuable information to make the right decisions in their daily work,” Juha says and continues: “Deep understanding of the chemical state of the process helps mitigate many risks: runnability issues, breaks, defects in the end product… Data-driven chemistry control makes visible the previously unseen cause-and-effect relationships in the process.”

The ability to react more quickly matches the complex and ever-changing nature of chemical processes.

Kemira’s new digital solutions build on the extensive chemistry and application expertise, process know-how, and the portfolio of patented solutions for measuring and controlling process chemicals. The KemConnect platform has already more than 500 connected customers and 48,000 data measurement points.

“In the end, it is a new way to manage the process. It is not unusual that certain wet-end chemicals are run at standard dosages, and when issues occur, the dosage is changed. The data-driven approach, together with eventually automated chemistry management, allow for constant micro-adjustments based on the real-time performance information and reliable predictions on the process conditions. The ability to react more quickly matches the complex and ever-changing nature of chemical processes,” Juha summarizes.

Share

You might also be interested in

We use cookies to improve your experience on our website. Read more about our cookie policy and privacy statement.