OMK implemented an AI-based service for scrap metal quality control
United Metallurgical Company has introduced a digital service based on machine vision and neural networks at the Vyksa OMK plant, which automatically determines the type, quality, and cleanliness of incoming steel scrap. The service has been launched in the casting and rolling complex, where scrap metal used for steel production is supplied
For this, industrial cameras have been installed in the workshops, continuously capturing photos and videos of incoming raw materials and sending data to a system that analyzes each unloaded layer of raw material in all railcars or trucks
Several computer vision models have been implemented in the process: the first model analyzes the entire video stream and identifies layers of unloaded scrap. The second model evaluates the scrap layers in real time for contamination and sends automatic warnings that allow unloading to be stopped automatically. The third model assesses how well the raw material complies with the GOST standard stated in the documents
The service is also currently being trained to recognize and block the unloading of certain categories of prohibited items that may end up in scrap, such as potentially explosive objects like gas cylinders, barrels, and others. In the near future, the technology will be integrated into the production process
Based on the analysis of scrap deliveries, the service generates a report for each unloaded transport vehicle. All reports with layer-by-layer photos of each railcar’s unloading are stored and can be provided to suppliers for certification justification
The solution will help optimize and speed up the unloading process of railcars with raw materials, save time for employees involved in acceptance, effectively resolve disputes in communication with suppliers, and shorten the product acceptance time
"Optimization is at the core of all processes related to automation in production. Machine vision has removed quality control inspectors from the operating zone of lifting mechanisms and accelerated quality control without involving highly valuable specialists, who are in short supply at any plant. We delegate routine processes to machines and allow workers to be where their presence is truly necessary. However, one should not think that machines will replace people 100%. A person also monitors how the system operates. This can be called a mutually beneficial synergy that saves costs and improves the quality of incoming raw materials," concluded Ilya Dzyub, Chief Architect for Digital Technology Development at OMK IT
Comments
No comment yet.