Artificial intelligence enhances mill processes
Rapidly developing technology continues to offer new opportunities to make mill processes run more smoothly and make fuller use of accumulated process data.
OLLI MANNINEN
For decades, highly automated pulp mills have been at the forefront of digitalisation. Digital applications continue to improve, boosting the internal efficiency and consistent quality of pulp production, as well as improving customer service. “Algorithmic machine vision, new measuring meth- ods and the efficient use of collected data are especially useful for improving the consistency of mill processes,” says Matti Toivonen , SVP, Technology at Metsä Fibre. This also allows Metsä Fibre to offer its customers var- ious ways to develop their business and new services. “The greatest challenge lies in keeping up with the rapid technological development, choosing the best solutions for our operations, and then combining these with our current ways of working.” IMPROVED PRODUCTION MANAGEMENT Toivonen says the most recent example is the digital twin introduced at the causticising plant of Äänekoski bioproduct mill. “It allows intelligent process predictions. It tells us where the processes are expected to settle and can adjust them to achieve the steadiest state possible. Intelligent prediction also lets us prepare for disruptions.” Predictability and consistency are vital elements in high-quality pulp production. “Thanks to greater computing power, we can now get the results of an intelligent prediction in a matter of seconds instead of the hours that it used to take.”
Machine vision, in turn, reveals areas that have not been visible before. “Machine vision applications are like tireless eyes that tell us what is happening to black liquor in the recovery boiler at a temperature of 1,000 degrees. Being able to monitor things 24/7 helps us to improve the managea- bility, reliability and consistency of production.” TOWARDS MORE HIGHLY AUTOMATED DATA ANALYSIS “One of our future goals is to use data to improve our internal efficiency, which will help us ensure that our customers get good and consistent production quality, as well as allowing us to predict the need for equipment maintenance and manage risks,” Toivonen explains. Another key field of development is the effort to de- crease process variations, supported by intelligent pre- diction using the digital twin, as well as algorithms and artificial intelligence. The third goal focuses on competence development. “Automation releases our human resources for more demanding brain work.” DATA MUST PROVIDE REAL ADDED VALUE As well as being used to improve the company’s own op- erations, data can also be made accessible to customers. Toivonen emphasises that digitalisation must provide real value through the whole value chain. “The data accumulated from pulp mills and forests can be used collaboratively with customers and partners to develop new products and services,” he says. Currently, an average of 40,000 data points are collect- ed from pulp mills and analysed largely by employees. Examples of data points include a machine’s tempera- ture, pressure or the position of a valve. “Our goal is to analyse at least 80 per cent of the data automatically in the future. Digitalisation will shape the content of our current tasks and will help us identify new abilities.” •
Matti Toivonen Senior Vice President, Technology at Metsä Fibre. He ensures that the best possible technology is used to develop quality and efficiency.
65
Powered by FlippingBook