Reduction of Short-term Process Variations -- Root-Cause Analysis Based on Thermal Imaging Data

Important results from the project

This project aimed towards mitigation of short-term product and process variations in the Pulp and Paper industry. Current equipment can’t capture variation with periods less than 5 min. The goal was to investigate the feasibility of using thermal imaging to capture substantially faster variation and to identify their root-cause. The findings point clearly toward the feasibility of the intended system. The intermediate results caught commercial interest. The potential of the system goes far beyond and there are many relevant topics that needs to be investigated.

Expected long term effects

A demonstrator for identifying root-causes of product and process variations was developed, built, and successfully tested at the mill. It is now in use to improve the mills sustainability. A method to identify anomalies was developed and successfully tested on recorded data. The existing demonstrator for recording thermal data was further developed to fit the needs of root-cause identification and anomaly detection. The collaboration led towards the increase of digitalization, especially advanced data analysis at the mill.

Approach and implementation

Recorded data was used to develop the methods for root-cause identification and anomaly detection. Were necessary the recoding system has been further developed, and demonstrators developed. Students developed a graphical user interface to enable more people to participate in the evaluation of the root-cause identification. The methods were evaluated on recorded data, and the developed demonstrators at the mill trials with success.