MMK uses machine vision to assist in case iron production

Magnitogorsk Iron and Steel Works (MMK) has introduced a system that recognises cracks in the chutes in the casting beds of blast furnaces.

The system uses machine vision to provide continuous video monitoring of the condition of the shelters surrounding the entire length of the main and connective chutes.

Each casting bed is monitored by two video cameras. The video footage is loaded into a neural network, which evaluates the shelters around the chutes and displays their current condition in mnemonic diagrammes. When a change in the condition of a shelter is detected, the system automatically enters the date and time of the event, the percentage of the closure and an image of the chute into the database. If necessary, the system generates e-mails that are sent to the staff members responsible for making operational decisions.

The system monitors the output of cast iron and slag in the casting beds of blast furnaces Nos. 1, 2, 6, 9 and 10 in order to ensure process compliance and observance of industrial safety requirements, as well as to reduce fugitive emissions and improve air quality in the workplace.