RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2
Project

Intechcom digitalizes NLMK equipment

Customers: Novolipetsk Steel Mill, NLMK

Novolipetsk; Metallurgical industry

Contractors: Intechcom
Product: Video analytics (projects)

Project date: 2023/05  - 2023/11

2023: Digitalization of rolling mill

Intechcom completed a project for PJSC NLMK based on machine vision technology. Intechcom announced this on December 19, 2023.

source = Intechcom

The main equipment in NLMK's hot rolling shop is the rolling mill 2000, namely a 1.5 km conveyor. Difficulties arose at the site where three coilers turn a hot strip of metal into 800-900 rolls per day: in case of violation of positioning, the rolls come out substandard. Manual adjustment took time from the operator, in addition, it could not always be carried out on time.

The location of the rolls and the progress of the conveyor, as well as the movement of the turntable and the quality of the winding, are monitored by 11 webcams and a 2D scanner. The scanner transmits pictures of the roll profile and, using specially developed software, determines the coiling quality, the number of falling turns and their height. Based on this data, the operator makes a decision on the correction of the roll, determines the malfunctions of the coils in a timely manner and adjusts them. Webcams collect information about the position of the roll from the moment it appears and send signals to the conveyor to speed up or slow it down. Information on the roll position and its data are displayed on the mnemonic diagram in the application interface, and also sent to the enterprise APCS.

File:Aquote1.png
The most difficult and interesting part of the project is the control of turning the table with the roll, - said Savva Vorobyov, technical director of Virsine. Since internal position sensors cannot be installed in the table, the angle of rotation is determined by the cameras and a corresponding signal is sent to change the position of the table. Control takes place in real time and requires 100% machine vision quality, which is difficult to provide in production conditions. Therefore, to train the systems, a virtual copy of the production site was modeled and non-standard conditions were programmed, including smoke or overlay of the view. I will add that if the cameras are damaged, the system is automatically transferred to manual control.
File:Aquote2.png

A number of hours were allocated for the direct implementation of the system at NLMK, since the production process at the enterprise is almost continuous.

File:Aquote1.png
Before that, of course, it was necessary to get all the required safety clearances, etc. This is another reason why not every company is able to switch from pure analysis of the collected data to use, - said Sergey Astakhov, head of the Intechcom branch in Lipetsk, - You need to feel what is happening at the enterprise, be able to dismantle, install and commission in a short time, carefully introduce new systems so that they work without comments as part of the whole.
File:Aquote2.png