Sortierung von Aluminiumschrott: Methodische Ansätze zur Rohgutcharakterisierung, Abschätzung der Klassifizierungsgüte von LIBS-basierter Sortierung und Sortierungsstrategieoptimierung per künstlicher Intelligenz und Digitalisierung

As in other processing tasks, it is clear that both the systematic approach to raw material analysis and the presentation of the processing-related separation success are essential in the sorting of aluminium scrap. Only by acquiring this knowledge about the raw material and sorting equipment can a reliable statement be made about the process, which forms the basis for optimisation. To summarise the results of the raw material analysis, the material content distribution can be represented in a Henry-Reinhardt diagram. On the one hand, the sortability of the raw material can be derived from this, and on the other hand, the optimised limit content for the parameterisation of the sorting machine can also be determined. As a rule, a general chemical analysis of the sorted products obtained is not sufficient to show the sorting success of the processing technology when sorting scrap. It would be possible, with restrictions, to describe the sorting success by the material class yield into the corresponding products. For a more precise understanding, however, it is advisable to draw up a classification curve that describes how likely it is that a particle will be correctly assigned to the corresponding material class depending on its alloy element content at a defined threshold. Although the shape of the curve is similar to a division curve, its meaning is fundamentally different.

The latest developments in the field of artificial intelligence and neural networks make it possible to sort particles that basically have the same colours. In this specific case, CLARITY-AI (“Artificial Intelligence”) can be used to separate profile and casting qualities from wrought alloys based on VIS sensor technology. Here, the combination with the CLARITY-LIBS sorting machine is suitable for achieving maximum throughputs in the overall process. The data recorded simultaneously by the sorting machine can be stored and used in Binder + Co’s digital service called “b-connected”. This data is suitable for permanently monitoring the process and, if necessary, making adjustments to the sorting settings.