FinlandManufacturing2023 – ongoingMetsä Board
Virtual sensor for paperboard quality
Metsä Board used a virtual sensor to predict paperboard quality before laboratory results arrived
Problem
Checking paperboard quality required collecting samples and waiting for laboratory analysis. That delayed production feedback.
Approach
Metsä Board developed a predictive virtual sensor with an IT partner. Laboratory results supplied the initial training data. The group ICT team helped establish a secure connection for transferring mill data to the partner.
Results
No published figures
Takeaway
The model depended on laboratory measurements and a new connection between mill systems and external developers. Metsä Board did not quantify the time saved or prediction error.
Categories
- Size
- Large enterprise
- Industry
- Manufacturing
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Partner co-development
- Evidence
- Company's own claim
Sources
Source summaries
Metsä Board's technical development manager described a working virtual sensor trained using laboratory results. He also described the data-transfer infrastructure built with the IT partner and group ICT team. The wider AI programme began in 2018; the sensor's launch date was not specified.
Full details are in the original sources.
