This project was a collaboration between ArcelorMittal España, SSSA, ILVA, Helsinki University of Technology, Politecnico di Milano, RWTH IEKH Aachen University and University of Oviedo and is coordinated by ArcelorMittal España.
The aim of this project was to develop a system capable to determine the quality in the field of inclusions of steel before and during its production, in order to change the setups to improve it. Data from 4 Steel shops, 2 partners experts in metallurgical modelling and 3 partners with proven expertise in Data Based models worked together, including crossed evaluation in order to produce, validate and conclude the cleanliness model. Two ways of model development were carried out: classical thermodynamic calculation and data based analysis. Thermodynamical models provide good results for being the first approach to cleanliness models for selected cases. Data Mining models are capable to imitate classical models, improving their performance for more cases although it was demonstrated that it is not possible to manage all kind of steel with a single model.
Moreover, the need of improving reliability of inspection system output to be used as model input was identified.
In order to validate the models and get deeper knowledge on inclusions formation in selected steel grades, several sampling campaigns and inclusions analysis have been done. As expected, being the steel grade analyzed similar, similar results were obtained. Finally an user interface and requirements for integration of the developed models within the steel plant were also developed in the promising cases although it was not completely extended as the range of steels where models are applicable is limited.
Potential areas of exploitation for the results from this project have been highlighted.