HPC Cloud-based simulation of steel casting
Ergolines’ goal is the development of equipment supporting the production of flawless steel alloys with metallurgical properties able to satisfy an ever increasingly quality-oriented market.This case study addresses the problem of slag carry-over from the ladle to the tundish which is a serious problem in steel casting and which can lead impurities in steel or poor ladle yield. The simulation of slag carry-over requires the use of HPC.
Start date: 01/07/2013
Duration in months: 18
Problem Description
In the field of continuous casting there is an increasing industrial demand for the development of new technologies for preventing slag transfer from the ladle to the tundish. Such an event may cause a breakout, that is the breaking of the solid skin of the solidifying cast products, which results in hazardous dispersion of liquid steel within the industrial plant. Ladle-slag monitoring is currently performed by operators on an empirical basis. Given the relevance of both safety and the economic implications of a breakout, there is a significant demand for an effective, automated system for ladle-slag monitoring.
Goals
Virtual prototyping
Challenges
Develop an effective detection system, it is necessary to correlate the vibrational signal with the fluid dynamics of the system. Such a correlation requires a complex, detailed simulation, which can only be carried out on an HPC system.
Innovation results
Dedicated HPC-based simulations followed by case experimental validation have provided Ergolines with key insights into the physics of the system and into different ladle-emptying mechanisms. As a result, it has been possible to establish a correlation between the shroud vibrational signal and the fluid dynamics of the system.
Business impact
The proposed slag monitoring technology can improve steel plant control and yield by detecting slag passing through the shroud, 60% of that lost steel can be saved. For a per medium sized steel plant saved amount up to 670,000€ saved per year.