To main content

UnWarp: Understanding warping corrugated board with machine learning

UnWarp: Understanding warping corrugated board with machine learning

Warped cardboard is a major quality issue in the production of corrugated cardboard. It reduces the productivity in the converting units, causes reduced quality on the final product, and increases waste. In UnWarp we are using machine learning to understand when the cardboard becomes warped in and provide guidelines for minimising the warp.

Photo: Peterson

At the VPK Peterson plant in Sarpsborg approximately 2 100 000 m2 (or ~300 soccer fields) corrugated board is scrapped every year. On top of the direct waste, the scrap is also a source of increased transportation of raw materials, increased energy use, and unnecessary CO2 emission. VPK Peterson works to reduce the scrap, and in particular at the new plant under construction in Halden. This has led to a pre-project from the Regionale Forskningsfond Oslofjordfondet with SINTEF Digital as research partner. In this pre-project SINTEF will apply machine learning to understand when the cardboard becomes warped. The long-term goal is to provide guidelines for minimising the warp at the new plant.

Published 19 February 2019
Research Scientist

Project duration

2019 - 2019