The project involves further development of an interface prototype that integrates both historical and real-time process data. The goal is to support operators in making well-informed decisions during the pulping process by visualising complex relationships that are otherwise difficult to interpret.
– The interface enables AI and machine learning to be used as practical tools in everyday operations, for example to understand how different process settings affect pulp quality, says Elmira Zohrevandi.
A key aspect of the research is transparency and user involvement. Operators will be able not only to use the decision support, but also to understand, adapt and expand the model using their own data. In the long term, this may lead to improved process efficiency, reduced energy consumption and better transfer of knowledge within the industry.
In its motivation, the Gunnar Sundblad Research Foundation highlights the project’s ambition to use interactive visualisation and machine learning to make hidden process relationships visible and actionable in industrial practice.
