Keith Hanna from Fluent News recently interviewed Dr. Toshio Ishii, Manager of the Numerical Simulation Laboratory at JFE R&D Corporation in Japan about the trends and challenges facing the steel industry's use of CFD.

View the pdf of this article

Wastewater treatment plant of Nevers, France, certified ISO 14001, Courtesy of Générale des Eaux - Marcel Chevret

KH:The steel industry has been going through some very hard times during the last five years. How would you characterize the underlying trends in the industry?

Dr. Toshio Ishii, JFE R&D Corporation

TI: The global economic slowdown after the year 2000, plus the lowest price for steel per tonne in 20 years and a huge overcapacity in steel production have meant that the steel industry is in crisis worldwide, with many companies having to restructure and consolidate to survive. This has led to mergers involving many well-known steel manufacturers, like Hoogovens, British Steel, and Usinor (in Europe). Here in Japan, NKK and Kawasaki Steel merged to form JFE in April of this year. I believe that thirty American steelmakers have filed for bankruptcy recently, and the Organization of Economic and Commercial Development estimates that global steel production needs to be reduced by 100,000,000 tonnes per year for companies to become profitable again.

KH: Has this downturn affected research & development within your company?

TI: Yes. At JFE we merged our two separate NKK and Kawasaki Steel R&D Laboratories earlier this year, so that our research group has actually expanded. We currently have fifty engineers at JFE who are trained to use FLUENT. Numerical simulation is still seen as a cost-effective way of developing our steel technology, especially during the ideas stage where we are asked to carry out many evaluations of conceptual designs. As one of the top five steel companies in the world (based on tonnage), JFE focuses on producing high quality steels for the automotive and construction industries, primarily in Japan and the western Asia Pacific rim region. To maintain this high level of output, our group also carries out technology research projects for the whole company on many of our installed unit processes.

KH: CFD has been used in the steel industry for over 20 years now. How has its usage changed during that time?

JFE tundish model showing the free surface with gas bubbling effects

TI: Strangely enough, we are modeling the same CFD applications as we did 20 years ago – tundishes, casting molds, and furnaces! We were actually the second FLUENT customer in Japan in 1984, when we licensed FLUENT 2.7, the old structured, staggered grid code. Prior to that, we had been using in-house CFD codes, but we chose FLUENT mainly because of its ease-of-use and the fact that its modeling capabilities were all in a single module. In those days, our license gave us access to the source code, so we were able to customize the code to suit our needs. Today, even though we are doing the same simulations that we did back then, we can do them in greater depth and much faster than ever before. FLUENT 6.1 has better physics and chemistry models, plus we have developed our own specialized NO_x, MHD (magnetohydrodynamics), and enhanced solidification models that couple with the CFD solver.

KH: Can you describe an example of how you have customized FLUENT to better solve one of your steel applications?

TI: Yes. We worked with the worldwide Fluent organization, including developers at Fluent Europe in the UK, on a customized MHD model for FLUENT. The application of magnetic fields to continuous casting (concast) molds helps smooth out the liquid steel flows as they solidify. This revolutionary technology produces better quality sheet metal steels, which are used in the automotive industry. Unfortunately, it is very hard to get meaningful experimental measurements of liquid steel in a concast mold. Hence, a user-defined subroutine was developed to model the complex MHD flow patterns. The results have allowed us to grow our MHD casting technology to a very high level. We have also developed new modules for modeling solidification in steelmaking processes, and we have several ongoing collaborations with Fluent Asia Pacific.

JFE concast nozzle and mold flow with MHD effects included in the FLUENT model; the free surface and electromagnetic braking effects in the steel flow are illustrated

KH:What are the CFD issues you currently see in the steel industry?

TI: Today, a typical CFD simulation is approximately 1.5 million cells in size, and we often need to run many transient cases for parametric analyses concurrently on the same hardware. This can cause a bottleneck, and we hope that advances in hardware technology, as well as faster, more accurate CFD solvers will become standard. We see Fluent as the premier CFD software house in our area, and we believe that Fluent will continue to deliver improvements in the solver, as well as more complex physics and chemistry, in the years to come.

KH: What do you foresee as the factors affecting CFD usage in the steel industry in the long term?

TI: I see several driving forces in the steel industry once it stabilizes again, some of which are external and some of which are internal. A significant external factor will be environmental and legislative trends. The Kyoto Protocol on CO₂ targets mandates that the steel industry reduce its levels of CO₂ production, or find new technologies to reduce the levels and recycle the gas better in its processes. An internal factor will be materials development, in which new application areas for steel are identified. In the automotive industry, for instance, we see many new materials, such as plastics, glass, ceramics, and composites emerging and replacing steel, but also combining with steel. These new hybrid, steel-based materials will pose interesting manufacturing and CFD modeling challenges in the not too distant future. I also foresee the need for more coupling of computer aided engineering (CAE) methodologies with CFD in steel industry applications. For instance, refractory blocks used in the linings of steel vessels experience enormous thermal stresses during normal operation, and we use codes like ADINA to model these effects. However, we need to couple them with FLUENT better. I also believe that the industry needs to use Design of Experiment DOE) software in our CAE experiments. Certainly, CFD is here to stay in the steel industry and will continue to play an indispensable role in JFE's R&D operations.

FluentNEWS