Impressive results when steel meets Artificial Intelligence
Swedish history is closely connected to iron. For over a thousand years, its people have mined and refined steel, in a process that has gone from extracting the iron by hand, to what is today a highly automated process. But now, it’s time for the steel industry to take yet another step into the future: HiMat Engineering is bringing artificial intelligence into the mix. This not only saves time and money, but is also better for the environment.
The extraction of iron ore is closely connected to Swedish history. From the manual mining that in the Middle Ages helped Sweden to enter the European community, to the
18th century, when wrought iron made up over two thirds of Swedish export, to the technological development of the 20th century, mining and the export of iron ore and steels have played vital roles in the Swedish economy. Today, the Swedish steel industry is highly specialized and high-tech, and is world-leading in a number of areas. But at KTH, a number of researchers realized it was time for the Swedish steel industry to take yet another step into the future.
The result of 40 years of research
Together, the team behind HiMat Engineering has over 40 years of experience of research on metallic materials and material modelling. They realized that there was a need to up the efficiency of how industrial enterprises develop new materials, a process that today is both costly and time-consuming. Since 2008, the founding team has focused specifically on this area, including research within the Vinn-Excellence centre Hero-m at KTH.
Conducting a single test in a steel millcan today cost around 1MSEK, and often one test isn’t enough to reach a satisfactory result. Therefore, HiMat Engineering developed HEDDA, a digital platform for data-driven materials development. Using HEDDATM, companies can conduct virtual experiments, instead of wasting time and money on tests in mills. Additionally, this process is more environmentally friendly, since fewer tests mean less emission of greenhouse gasses. This way, HiMat Engineering can streamline the process for materials development and production, and also optimize the materials’ performance to make their products more competitive. The data-driven methodology is based on machine learning, a sub-group of artificial intelligence, used to find correlations between important parameters in large amounts of data.
Materials development meets AI
HiMat Engineering is today one of very few companies focusing on machine learning to make materials development and manufacturing more efficient. Right now, the focus is on launching more tools on their digital platform, and to showcase its functionality in a number of pilot studies.
However, the team has already produced proof that the platform works. Right now, they are in the process of patenting their first material developed using HEDDATM: a stainless steel that has twice the lifespan of similar steels, and that additionally has been developed in record time. Together with KTH and three European partners, they have applied for EU funding to bring the material to the market.
- We get excellent support from KTH Innovation and our coach Daniel Carlsson, who has helped us develop our business, but also to secure funding, says Peter Hedström, Head of Development at HiMat Engineering.
As one of nine KTH startups, HiMat Engineering has just finished the KTH Innovation pre-incubator program, after making it through all three phases. The pre-incubator is a year-long program for early, tech-based KTH projects that want to accelerate their development.
- The next step includes entering the Swedish market on a larger scale. We will also focus on developing our business model, and continue building our team, to be ready to accelerate in 2019.