In a research project at Chalmers, post doc Farinaz Ebrahimian is creating a model that provides a detailed insight into probably the most important process in paper and pulp production, the one where the lignin in the wood fiber is separated from the cellulose. The very prerequisite for making paper and cardboard.
When her model is completed in early 2026, it will describe with high accuracy how the lignin content changes at different times in the process, which in turn creates the conditions for developing AI solutions that can optimize the process industry. The project is characterized by access to unique process data from one of Södra’s three pulp mills. Södra is an industrial partner in the Business Research School Resource Smart Processes, which also includes Farinaz’s project.
– The close collaboration with industry and their sharing of process data is crucial for this project. There is already some research in the field, but it is based on lab experiments and differs in some critical parts from the actual process. Now, thanks to process data, we can build a model that describes in detail the process that takes place at the factory and that provides unique knowledge about how the lignin content of wood chips changes at different times,” says Farinaz.
Examining the aspect of time as an important process parameter
The time aspect is the central parameter in this research project; it is one of countless parameters that are interrelated and affect the process in an extremely complex way. To begin with, Farinaz has identified the critical points in the process that need to be analyzed and then, with Södra’s help, she has collected samples. These are used to calculate the concentration of lignin and also various chemical substances that affect the process.
– In some parts of the process it is not possible to take samples, so I have supplemented the process data with experimental data by imitating the process in a lab environment. With all the data in place, it is possible to model how the concentration of lignin changes but also how different parameters affect and can be adjusted to achieve different results,” says Farinaz.
Bringing together basic research and industrial research needs
The project is a win-win for both industry and science. From the academy’s perspective, it provides important insights into where the focus of future research should lie in the process and opens the door to new research questions. For Södra and industry in general, the win is primarily about two things:
– Collaboration with academia is important both to train the experts of tomorrow and to produce results that can streamline and give us better control over the process. By participating in initiatives like this, we can bridge the gap that can arise between basic research and industrial research needs, where academia builds knowledge about our processes and how AI can be used in them. We see that interest in working with AI is growing internally, this is the beginning of a number of upcoming projects together with academia,” says Fredrik Wernersson Brodin, specialist in cellulose technology at Södra.
Several AI projects use Södra’s process data
Recently, two post-docs linked to AI research at Linköping University have also joined the graduate school to use Södra’s process data and experimental data from Farinaz’s project to drive AI development in the process industry.
– Hopefully we can show how close collaboration within academia and between academia and industry drives development forward – towards more efficient production and greater scientific understanding,” concludes Farinaz.
