MMK to use machine vision technology to remove defects from rolled steel

Magnitogorsk Iron and Steel Works (MMK) has launched a system that can detect and classify defects on the surface of cold-rolled steel during the production process. The system displays an accuracy rate of at least 95% when it comes to detecting and identifying defects.

To carry out the project, MMK purchased equipment that uses its own unique proprietary software. The entire system was installed and configured by specialists from MMK-Informservice, United Service Company and R&D MSTU, as well as staff from MMK’s rolling mill No. 11.

Camera images are fed into the system, which uses machine vision technology and a server equipped with high-performance graphics processors to find and identify over 22 different defects on the surface of rolled products, including fractures, rust spots, strip breakage and coil breaks. The system automatically sorts them based on various parameters – class, size, quantity, etc. – and maps the location of defects on the surface of the metal. Data on any deviations discovered are entered into the manufacturing execution system for subsequent use in product certification, and an archive of defects found over the past three years is being created.

The innovative solution has been introduced for use in the inspection unit at rolling mill No. 11 and the continuous annealing unit. This significantly reduces employee workload, increases the mill’s productivity and improves quality control. Start-up operations are currently under way in the continuous annealing unit. Similar systems will be put in place for use on the Company’s continuous pickling line and continuous hot-dip galvanising line in November 2023.

MMK plans to scale up the system in the near future for use in its thick plate mill.