PNTZ Implements Artificial Intelligence for Pipe Geometry Control
Pervouralsk New Pipe Plant (PNTZ), a subsidiary of TMK (Pipe Metallurgical Company), has implemented a digital quality control project in its pipe rolling shop. The machine vision system includes a unique hardware and software complex and incorporates a range of proprietary solutions, enabling continuous end-to-end monitoring of pipe geometry parameters.
The digital solution was developed by specialists from the TMK Research Center in collaboration with PNTZ’s process engineering team. It is based on advanced scientific approaches and best practices in industrial artificial intelligence. The implementation has significantly reduced the human factor, improved measurement accuracy, and minimized deviations in the production of hot-rolled pipes.
High-precision machine vision cameras are installed at key sections of the TPA-160 pipe rolling unit—including the piercing, elongation, and reduction mills. A neural network model identifies the edge points of pipes and billets in real time, calculates their length and diameter, and synchronizes the data using time stamps. The information is transmitted to a centralized storage and processing system, ensuring continuous production monitoring and remote access from any device, including mobile platforms.
Approximately 2,000 annotated images of pipes and billets of various sizes were used to train the model. The hardware and software systems are adapted for metallurgical environments: the equipment is protected against high temperatures, thermal radiation, dust, moisture, and electromagnetic interference, ensuring stable and continuous operation.
“Incorporating machine vision at PNTZ is part of TMK’s digital strategy aimed at developing smart manufacturing. Today, all pipes and billets in the shop pass through this system. We have not just automated measurements—we have provided our personnel with a modern tool for monitoring critical parameters. This allows us to guarantee precise pipe geometry for our customers and significantly improve metal utilization by enabling real-time equipment adjustments,” said Vladimir Toporov, Managing Director of PNTZ.
PNTZ is implementing a comprehensive production digitalization program, including the use of big data analytics and machine learning technologies. As part of this initiative, solutions such as the “Digital Steelmaker Assistant” at the Iron Ozone 32 electric steelmaking shop, a fully integrated traceability system for ferroalloys, and an advanced production management system in pipe rolling and drawing shops have already been introduced, ensuring full tracking of all production stages across the entire process chain.
