Artificial intelligence: Very real possibilities
One of the ways Södra is keen to strengthen its competitive edge in the future will be through the use of Artificial intelligence (AI).
Andreas Darnell, System Engineer at our Värö mill, explains that AI has the potential for a range of applications from predictive maintenance to causal analysis and predictive control. It could bring a host of advantages for Södra as well as for our customers, from quality consistency and process reliability to environmental protection.
However, AI is no quick fix, as Darnell explains. “It can be tempting to buy off-the-shelf solutions but these only work for as long as everything stays the same in the mill. We are keen to take ownership of AI tools and develop them in house, working both with external experts but also using the high level of skills we already have within our organisation.”
To this end, Darnell and the team at Värö together with Johan Thor, Head of Analytics and Data Science at Södra, have spent the last months researching and trialling a range of AI tools with machine learning specialist Tenfifty. These initial trials were focused on developing a soft-sensor model for the pH of the wastewater treatment plant.
“We wanted to focus on developing process models from historical time series data capable of predicting the evolution of the state of the process, i.e. making them suitable for real-time monitoring and control. We have all this data sitting around, so why not make it work for us.”
Darnell now has a solution he is happy to launch which should result in more accurate control of the pH levels in the mill’s effluent.
“The possibilities for AI in the future are almost endless,” Darnell says. “If we take 10, 50 or even 100 small steps like this one across the process, they would add up to a substantial result in terms of productivity and efficiency for us and our customers.”
Show all content for topic.