Development of a Recipe-based Strategy for Scrap Sorting
E. Le Blansch (TU Delft - Mechanical Engineering)
D.L. Schott – Mentor (TU Delft - Mechanical Engineering)
Y. Pang – Mentor (TU Delft - Mechanical Engineering)
S.E. Offerman – Graduation committee member (TU Delft - Mechanical Engineering)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
Steel production is essential for modern infrastructure and manufacturing, but remains highly resource- and emission-intensive. Increasing the use of recycled steel scrap is an important route towards more circular and lower-emission steel production. However, scrap streams contain individual pieces with varying chemical compositions, which limits their direct use in high-grade steel production. If residual or alloying elements exceed the limits of a target steel grade, scrap may be down-cycled or diluted with primary raw materials.
This thesis develops and evaluates a recipe-based strategy for scrap sorting that uses piece-level composition data to allocate scrap items to steel recipes. The proposed Sorting Decision Layer is intended to operate between sensor-based sorting equipment and the operator. It evaluates scanned scrap items according to mass, elemental composition, heap capacity, and recipe constraints derived from European Steel Standards. A heap is considered feasible when the mass-weighted average composition of the selected items satisfies the lower and upper elemental limits of a target steel recipe.
Because the allocation problem is highly combinatorial, several heuristic algorithms were developed and tested in a simulation framework using synthetically generated scrap data. Performance was evaluated using item conversion, mass conversion, heap utilisation, and economic value. The results show that a multi-pass greedy algorithm provides the best balance between solution quality and computational efficiency. Further experiments show that increasing heap capacity and the number of heaps improves mass conversion and value creation, but may reduce heap utilisation. A balanced configuration was found around seven heaps with a heap capacity of 12,000 kg, achieving approximately 78% mass conversion and 89% heap utilisation.
Overall, this thesis demonstrates that recipe-based scrap sorting is a promising strategy for increasing the value recovered from steel scrap. By combining sensor-based characterisation, steel recipe constraints, and scalable allocation algorithms, scrap sorting can shift from broad classification towards value-oriented recipe allocation.
help_outline