Rusal Implements AI Technology to Accelerate Aluminum Ingot Analysis

Rusal, one of the world's largest aluminum producers, has introduced artificial intelligence technology to analyze the microstructure of aluminum ingots. Developed by the Rusal Engineering and Technology Center (Rusal ITC), the proprietary system reduces the analysis time from several hours to just 15 minutes, significantly minimizing routine work for employees.

The new technology employs machine vision and neural network models to automatically analyze cylindrical ingots, which are critical for applications in mechanical engineering, construction, and other industries. The quality of these ingots directly affects the durability of pressing equipment and the final product.

According to Viktor Mann, Rusal’s Technical Director, the neural network evaluates eight microstructure parameters, including grain size and the number and size of inclusions. While traditional microscopic analysis by laboratory technicians can take between 1.5 to 4 hours, the AI system produces comprehensive reports in just 15 minutes. Mann highlighted this advancement as an example of industrial AI enhancing production technology.

Currently, the technology is operational in the Rusal ITC laboratory, with plans to expand its use to aluminum plants for finished product analysis. Mikhail Grinishin, Director of Production Automation at Rusal ITC, noted that separate neural network models were trained for each parameter using a dataset of digital microscope images. These images were annotated by laboratory specialists to ensure precise training. The AI achieves accuracy comparable to that of human technicians, but with greater speed and repeatability, reducing the influence of human error.