Moscow, 1 July 2025 – RUSAL, one of the world's largest aluminum producers, is implementing artificial intelligence to monitor the aluminum melt preparation. The preparation includes up to 20 operations, and a specially trained neural network monitors compliance with the optimal sequence of operations. The technology using AI reduces the risk of obtaining a low-quality alloy.

The RUSAL Engineering and Technology Centre (RUSAL ETC) developed a system for monitoring melt preparation using machine vision together with employees of the Foundry Directorate of the Taishet Aluminum Smelter (TaAZ). Tests at TaAZ demonstrated the accuracy and efficiency of the system.

"A specially trained neural network monitors the melt preparation, the personnel actions and the equipment operation through cameras in the casthouse, and displays prompts on a specially installed monitor based on the operational standard sheet. The system helps the operator to achieve the necessary alloy chemistry and reduce the risk of producing low-quality metal. The plan is to implement this technology at all aluminum smelters of the Company," says Victor Mann, Technical Director at RUSAL.

"Depending on the alloy complexity and the type of aluminum ingot cast, preparation includes from 7 to 20 operations, each of which shall be performed at a strictly defined moment in compliance with certain process parameters. The system, which operates using machine vision, reduces the impact of human error as it helps the operator not to miss key moments during alloy preparation, and to better identify the areas of dross on the hot melt surface to improve the efficiency of its cleaning," remarks Mikhail Grinishin, Director of Production Automation at RUSAL ETC.