Russian Steelmakers Use AI to Optimize Metal Pouring

A major Russian steel plant has deployed an AI-powered video analytics system that cuts metal waste, reduces worker risk, and delivers significant savings.

At the West Siberian Metallurgical Plant in Novokuznetsk, engineers have implemented an AI-driven video analytics system to monitor molten metal levels in ladles. The new technology is designed to reduce material loss, lower refractory consumption, and improve worker safety, according to Metalinfo.ru.

Previously, steelworkers relied on visual estimation to gauge the height of molten metal in the ladles. That often led to errors—either leaving metal in the converter or causing overflow, both of which raised costs and equipment wear. Now, a neural network analyzes thermal and optical camera data in real time, identifying the temperature of slag, refractories, and liquid metal, and then calculates the optimal fill level. Data processing takes less than one second.

The plant expects the system to save around 28 million rubles per year. It also improves working conditions by reducing exposure to heat radiation, gas, and dust. Workers can now monitor the steel pouring process remotely using screens.