Your browser doesn't support javascript. This means that the content or functionality of our website will be limited or unavailable. If you need more information about Vinnova, please contact us.

Physics-informed AI for identification and control of process-product relations in papermaking

Reference number
Coordinator Kungliga Tekniska Högskolan - Kungliga Tekniska Högskolan Inst f fiber- & polymerteknologi
Funding from Vinnova SEK 1 500 000
Project duration December 2023 - November 2025
Status Completed
Venture Strategic innovation programme for process industrial IT and automation – PiiA

Important results from the project

The project developed a coherent method chain that combines IR thermography, signal/image processing and machine learning to identify CD and MD variations and recurring patterns in IR images of the paper web. A digital twin approach was developed where OpenFOAM simulations of the headbox/jet were be used to train a fast surrogate model that links process parameters to sheet structures. Software and analysis were developed based on Holmen´s full-scale data.

Expected long term effects

The expected long-term effect is a better understanding and controllability of process-product relationships in paper/board. The methods can provide faster diagnosis of periodic disturbances and profile problems, support optimised headbox settings, and, in the long term, enable more data-driven control. This can reduce waste, increase resource efficiency, stabilise quality, and provide a platform for follow-up projects and broader implementation in the Swedish process industry.

Approach and implementation

The project was carried out in three stages: (1) collection and processing of IR films from the full-scale machine and coupled process data, including establishing the ability to generate these continuously, (2) method development to separate non-periodic variations from CD and MD profiles and to identify recurring patterns, and (3) coupling to the flow in the headbox via OpenFOAM simulations and training of an ML surrogate model. The developed parts have been compiled in a Python pipeline.

External links

The project description has been provided by the project members themselves and the text has not been looked at by our editors.

Last updated 3 February 2026

Reference number 2023-04048