Case of Nippon SteelAccelerating DX at steelmaking sites
NIPPON STEEL and NEC accelerate DX at steel manufacturing sites by
utilizing AI to build equipment status monitoring infrastructure at steelworks
Introduction
Nippon Steel Corporation has adopted the AI analysis software "NEC Advanced Analytics - Invariant Analysis" (Note 4), which utilizes NEC Corporation's AI technology "System Invariant Analysis Technology", to build an infrastructure for monitoring equipment conditions at its steel mills. In January 2021, a long-term operational test of online monitoring of equipment status began in the Kimitsu Works. any time.
In the Kimitsu Works, Nippon Steel Corporation has verified and confirmed that this software can detect signs of trouble in advance by using offline data for a highly difficult-to-resolve trouble that required 10 days to find and confirm the root cause.
How the technology works
This software can automatically detect signs of "unusual" abnormalities based on measurement data obtained every 100ms in real time, thus preventing operation stoppages due to trouble and product quality deterioration due to equipment defects. Furthermore, by learning from "usual conditions" during operation, the system does not require past trouble data and can detect unknown problems that have never occurred. In addition, the system is responsive enough to detect abnormalities in the steel manufacturing process, and the user can customize the pre-processing of the data used to make the determination.
Real Time Monitoring and Good Interface with Client System
- High-speed analysis abnormality judgment (real-time monitoring at 100 ms)
- High detection accuracy
- Communication interface is flexible and can be customized by the user
- Versatility with an eye on future equipment development, ease of analysis and evaluation
NEC's AI prevent problems
As part of these efforts, NEC's AI will be used to detect signs of abnormalities in the operation and quality status of in-process manufacturing equipment at the hot rolling mill in the Kimitsu plant in real time. By learning and modeling the behavior of facilities and equipment based on more than 2,000 measurement data (current, temperature, pressure, control signals, etc.), including data collected from 500 physical sensors installed throughout the manufacturing process, AI will prevent problems before they occur and improve the efficiency of facility inspection and operation monitoring.
- * A platform that can develop and apply AI by analyzing various types of data, equipped with high computing power capable of advanced image analysis and deep learning, which is maintained by Nippon Steel Corporation so that any staff member who promotes the development and application of advanced IT can quickly analyze the data at any time.
- * This case introduced here is done between Nippon Steel and NEC. Our company has not been involved with this case.