Artificial Intelligence AI Detects the Smallest Changes in Steel

With around 5,000 types of steel, the manufacturing process depends on nuances. In order to create new properties, the steels are analyzed using various imaging methods. A research team at Saarland University has now trained an AI with microscopy analysis data from 10,000 steel samples to detect the smallest changes in the steel.

In the production of steel and other metals, each production step has an effect on the internal structure, known as the "microstructure" by materials scientists. This is changed by the chemical composition, the rolling process or heat treatments. "The microstructure of steel is extremely complex and varies greatly depending on the desired property. However, even the smallest differences must be recognized and correctly classified under the microscope or in computer tomography. Our AI-supported process now does this automatically," explains Frank Mücklich, Professor of Functional Materials at Saarland University (Germany).

10,000 Material Samples of Different Steels Analyzed And Recorded

Years of research were required to train the artificial intelligence so that it could not only recognize different patterns in the material structure, but also analyse them objectively. "Several dissertations were written at my department, all of which were interdisciplinary in nature. We brought scientists from the Max Planck Institute for Informatics and the German Research Center for Artificial Intelligence on board, who transferred their methods of machine learning and AI to materials science," explains Frank Mücklich, who also heads the Steinbeis Research Center for Materials Science (MECS). Thanks to the long-standing cooperation between this transfer institute and the Saarland-based (Gerany) steel company Dillinger, the scientists were able to analyze around 10,000 material samples of different steels on a micro, nano and atomic scale and record them in a comprehensive database.

Extend Material Data to All Other Metals And Ceramics

The Steinbeis Transfer Institute MECS has now entered into a strategic partnership with the Swiss company Imagic Bildverarbeitung AG to enable industrial companies to carry out their analyses independently on the basis of this database in future. This company develops software for microscopy, image analysis and image data management. "We offer this company the so-called ground truth, i.e. verified and reliable data that is suitable for training artificial intelligence and achieving correct results with it. Up to now, this material data has related to steel grades and various metals, but we also want to extend this to all other metals and ceramics," explains Frank Mücklich.

The materials scientist wants to keep the expert knowledge on imaging processes for materials on the Saarbrücken campus in order to offer his graduates highly qualified jobs. "Several of my former doctoral students are already working at the Steinbeis Research Center MECS, which we spun off from the university 15 years ago, and are contributing their expertise from their research activities," says Professor Mücklich.

Smart-Factory

Access solutions for the Smart Factory - Industry 4.0