InSpecScrap
RUNNING TIME:
04/2023
—
03/2025
Total project duration:
2 Years
The project
Steel scrap is needed as an important secondary raw material in steel production. The use of scrap as a raw material means resource conservation and a reduction in CO2 emissions of up to 75%. When using scrap, precise knowledge of the quality is essential. In the project, an innovative material characterization using digital AI-based methods for the spectroscopic detection of impurities and scrap quality, as well as an optimization of the scrap composition with regard to specifications for the required steel grade is researched.
Our activities in the project
Through InSpecScrap, we have the opportunity to significantly expand our existing expertise in the areas of data management, hyperspectral image analysis, machine learning/data analytics and to deepen this know-how in a basic research project. The results can be used to develop new methods for the creation of more accurate training data for future AI-based classification methods, which adds another crucial aspect to the digitalization chain of the steel industry.
Research groups
Downloads
Auftraggeber
Universität Graz
Project partners
K1-MET Metallurgisches Kompetenzzentrum
KNOW-Center
Graz University of Technology - Institute of Computer Graphics and Vision
Project details
The aim is to investigate the innovative digital sensor-based characterization of steel scrap, which uses multispectral analysis to capture images of the examined scrap samples in order to determine the spectrum of impurities contained in the scrap, scrap quality and valuable metals.
Machine learning is the key to optimized image evaluation and the detection of contaminants through the intelligent coupling of AI and sensor technology.
The methods will be validated and evaluated in use cases of scrap characterization with a focus on the analysis of industrially used scrap samples and the application of the sensor concept in the industrial environment of two steel producers.
One of the objectives of InSpecScrap is also to generate proposals for the optimized scrap batch mix for specific steel grades and to investigate the effects of the optimized scrap mix on the metallurgical reactions taking place in the process.
As a result, a roadmap is created that includes all technically relevant aspects along the scrap value chain (use of digital methods, separation of impurities, automation of the scrap industry, sensor combinations for metallic secondary raw materials).
