Project Details
Projekt
Multidimensional probabilistic characterization of slag materials for the optimization of cooling, comminution and separation processes, using statistical image analysis supported by machine learning
Applicant
Professor Dr. Volker Schmidt
Subject Area
Mechanical Process Engineering
Term
since 2021
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 470322626
An interdisciplinary approach of different fields (including metallurgy, tomographic imaging, mineral processing, as well as data-driven analytics and modeling) is necessary to improve the recyclability of valuable metals in slag materials. A particularly important point is the adjustment of the slag structure during the formation of slags to improve the performance of potential downstream processes, which requires a deep understanding of the interplay of process parameters and structural/compositional descriptors of slags. Achieving this goal necessitates a profound comprehension of quantitative relationships between process parameters and the slag's structural/compositional descriptors, requiring an in-depth characterization of slag materials through microscopic imaging. Such analysis is pivotal for accurately quantifying the impact of process parameters on the slag's microstructure and composition, alongside its functional properties like separation behavior. The project aims to enhance recyclability of slags by means of the following three primary objectives: (a) Characterization of slag materials: Employing statistical image analysis and machine learning for a multidimensional characterization of slags, using descriptor vectors of morphology, texture, and chemical composition. This includes advanced image processing techniques (convolutional and generative adversarial networks) followed by multivariate stochastic modeling. (b) Characterizing the impact of cooling, comminution and separation of slag material on the output of these processes, by deriving quantitative process-structure-property relationships. This involves the mapping of process parameters onto structural/compositional parameters through non-linear regression models, facilitating the prediction of slag properties after processing. (c) Application of process-structure-property relationships for process optimization to achieve desired slag properties with reduced experimental efforts. Namely, process parameters will be identified, which lead to predefined, desired structural/compositional parameters of slag materials. In this way both structural recommendations and optimized process parameters will be made available to partner groups within SPP 2315, which can contribute to the reduction of experimental efforts for the optimization of cooling, comminution and separation processes.
DFG Programme
Priority Programmes
