MHI-MS puts AI in control of corrugated-sheet warping
MHI-MS has developed and put into use an AI-based control device for corrugated-sheet production. Sensors collect moisture, temperature and bending measurements so a model can predict settings for the paper being processed. The system addresses adjustments previously dependent on experienced operators, as manufacturers handle more paper varieties and smaller production batches.
Artificial Intelligence··Night
Sensors guide production settings
MHI-MS, a machinery company in the Mitsubishi Heavy Industries group, has developed and put into use an artificial intelligence device for controlling warping in corrugated sheets. The electrical control equipment can be attached to its corrugators, the machines that make the sheets. Sensors measure moisture, temperature and bending, while a computer runs a prediction model connected to the machine-control equipment.[1], [2]
Different paper types need different settings
Corrugated-sheet production glues two outer liner sheets to a wavy middle sheet and dries the material at high temperature. Differences in moisture between the two faces can make the sheet bend. Standard settings do not suit every paper type, so experienced operators have adjusted the machines. MHI-MS says retirements and labour shortages complicate passing that knowledge to new staff, while smaller batches and more product varieties require more frequent adjustments.[1]
The model learns from each paper type
The device combines AI technology from Mitsubishi Heavy Industries’ research centre with MHI-MS warp controls. It automatically accumulates measurements for each paper type and learns suitable machine settings from them. Warped sheets can jam machines that assemble boxes, which the company identifies as a quality and production problem. MHI-MS plans to expand sales of the device in Japan and overseas.[1]