Datanami
EnterpriseAI
HPCwire Japan
QCwire
HPC & AI Wall StreetAug. 26, 2020 — ArcelorMittal Global Research and Development Center in East Chicago is once again the recipient of two prestigious High-Performance Computing for Manufacturing Program (HPC4Mfg) grants through a Department of Energy (DOE) initiative. The DOE announced $3.3 million in federal funding for 11 manufacturing and materials projects aimed at solving technical challenges in manufacturing and materials development in the United States.
“These awards from the DOE engage national labs and industry to benefit industry and ultimately conserve energy,” said Mark Atkinson, senior director, ArcelorMittal Global Research and Development, East Chicago. “High-performance computing (HPC) capitalizes on the computational horsepower that exists in the national labs that we do not have at ArcelorMittal. Together, we can solve complicated problems and accelerate product development to meet our stringent quality and sustainability goals.”
Grant one: “Ab-initio Guided Design and Materials Informatics for Accelerated Product Development of Next Generation Advanced High Strength Steels.”
ArcelorMittal is partnering with Lawrence Livermore National Laboratory (LLNL), and Argonne National Laboratory (ANL) to establish a method for rapid product development of next generation high strength steels. ArcelorMittal’s typical product development lifecycle equates to approximately 4-6 years, which consists of developing new chemistries and testing the chemistries many times until proven successful. The labs, along with ArcelorMittal, will use HPC at the atomistic scale and artificial intelligence (AI) with the aim of reducing the development time.
“This can be a tedious cycle and we are always looking for new ways to cut down on the product development time to bring new automotive steel grades faster to our customers,” added Brian Lin, research engineer, product development, ArcelorMittal Global Research and Development, East Chicago.
ArcelorMittal has generated copious amounts of data and chemistries over the years that product development engineers use to develop future grades of steel.
“We cannot always utilize that data to the best of our abilities. That is where introducing HPC simulations guided by input from earlier experimental data and AI would be beneficial for us,“ noted Brian. “We will be using AI to identify correlations between chemistry, processing and properties from the HPC simulations to ultimately speed up the product development lifecycle.”
Energy savings would be realized by eliminating portions of the plant’s trial phases. Testing coils is extremely energy and cost intensive if the trial coils do not meet the property requirements.
Grant two: “Reduced Order Modeling and Performance Prediction for Steel Refining Ladle Processing via High Performance Computing.”
ArcelorMittal is collaborating with Oak Ridge National Laboratory (ORNL) and Purdue University Northwest to reduce the yield loss caused by impurities forming in the refining ladle process. In the steelmaking process, impurities must be eliminated to achieve optimal product performance. The only way to remove impurities is when the steel is molten. This study, taking place at the ladle at Indiana Harbor’s #3 steel producing operation, will examine techniques to produce cleaner steels for our customers. The goal is to conserve energy for ArcelorMittal USA by increasing the probability of melting steel correctly the first time.
“We have a long history working with Purdue University Northwest and they have vast expertise in the ladle process as well as simulating that process,” said Pallava Kaushik, manager, steelmaking and casting, ArcelorMittal Global Research and Development, East Chicago. “This is perfect collaboration with ArcelorMittal representing the operations, Purdue Northwest representing academia and ORNL providing computer devices. We are expecting usable results which can be effectively implemented during the ladle refining process in multiple plants in our company.”
HPC will allow faster calculations to evaluate multiple scenarios in a shorter time with the end goal of achieving process efficiency, which will reduce energy consumption at ArcelorMittal.
About ArcelorMittal
ArcelorMittal is the world’s leading steel and mining company, with a presence in 60 countries and primary steelmaking facilities in 18 countries. In 2019, ArcelorMittal had revenues of $70.6 billion and crude steel production of 89.8 million metric tons, while iron ore production reached 57.1 million metric tons. Our goal is to help build a better world with smarter steels. Steels made using innovative processes which use less energy, emit significantly less carbon and reduce costs. Steels that are cleaner, stronger and reusable. Steels for electric vehicles and renewable energy infrastructure that will support societies as they transform through this century. With steel at our core, our inventive people and an entrepreneurial culture at heart, we will support the world in making that change. This is what we believe it takes to be the steel company of the future. For more information about ArcelorMittal please visit corporate.arcelormittal.com or usa.arcelormittal.com.
Source: ArcelorMittal
