Rusal told about the use of AI at production facilities

"Rusal" has implemented a system of computer vision models based on AI-technologies at its production facilities, the company's press service said.

The Bogoslovsky aluminum smelter has put into operation a system of online monitoring of the granulometric composition of sinter - raw material for alumina production - on the conveyor belt using AI solutions developed by the company's specialists. The technology is based on a set of computer vision models, including detectors, an algorithm for determining the size of granules with the conversion of pixels into millimeters, as well as combining granules into fractions with certain sizes according to the detected layer.

The development presented by Rusal's experts helps to improve production processes, provides online monitoring and control of the particle size distribution as it moves along the conveyor belt to the next process, automatically identifies deviations from the required characteristics and suggests corrective actions to the operator. Prior to the introduction of the AI solution, sinter composition control was performed manually by daily sampling.

"The Alumina Division is actively exploring and implementing digital technologies where they really make a difference and help to improve productivity. The AI-based sinter inspection system is able to detect defects at early stages of production, allowing us to respond quickly to problems and raise product quality standards. This is just one of the projects that we are actively implementing at both mining and production enterprises," said Yakov Itskov, Director of Rusal's alumina business.

The digital solution is planned to be replicated at all production sites of Rusal's Alumina Division.

This solution is not the only AI-based solution in Rusal's portfolio of in-house developments. In 2023, the company received the national award "Leaders in Artificial Intelligence" for its technology for monitoring electrolysis shops using machine vision. The domestic innovative system will appear at the Krasnoyarsk, Bratsk, Novokuznetsk, Irkutsk and Volgograd aluminum smelters by 2027.

Another solution is a specially trained neural network model that calculates the chemical composition of the electrolyte in the electrolyzer and adjusts the technological process. The project makes it possible to increase productivity in aluminum smelting.

The company has also developed AI technology for container control. It allows the time required to send a container from the terminal to load it with finished products to be reduced by several times.

Rusal also uses its own AI technology to analyze the microstructure of aluminum ingots. It is based on the application of machine vision and neural network models and reduces the time required to analyze each sample by several times, thus saving employees from a large amount of routine work.

"Rusal is implementing dozens of projects to digitalize production, including the use of neural networks, machine vision, and big data analysis. Artificial intelligence relieves our employees of a huge amount of routine operations and saves their time, increases the accuracy and efficiency of technological processes," said Victor Mann, Rusal's technical director.

As Mikhail Mishustin, head of the Russian government, stated during the plenary session of the Digital Almaty 2024 forum, by 2030 the effect from the introduction of AI into the Russian economy, even at an average rate, may exceed 10 trillion rubles. So far, digitalization affects different industries in different ways: for example, in 2024, according to Strategy Partners, only 24% of domestic companies in the oil and gas and metallurgical sectors will implement AI solutions.