Decarbonization of steelmaking has stagnated while it has a considerable share of global greenhouse gas emissions and a growing demand. Digitalization is seen as a viable option to reduce emissions and costs of the sector in the near term and life cycle assessment (LCA) as a comprehensive framework to evaluate changes in production practices. In this study, we analyze the potential impact of using optimization algorithms to improve the operation of a steelmaking plant in Spain. Specifically, we study the potential effects of optimizing the sequence in which steel is produced to minimize losses during casting. The global warming (GW) impacts and economic costs are quantified using a dynamic LCA model, considering uncertainty and temporal variability using an open-source LCA framework. The results indicate, on average, modest savings in costs and are inconclusive regarding GW emissions. Most of the cost savings come from a reduction in the use of additives and electricity, which are wasted when the steel is scrapped during casting. The methodological framework has proven useful in quantifying and interpreting the potential effects of digitalization. The implemented solution, tested in an industrial setting, allows an automated evaluation of production at the plant using the LCA model, facilitating the use of sustainability criteria in decision-making.
The study tests how an optimisation algorithm could improve production planning at a Spanish steelmaking plant that uses an electric arc furnace. Researchers compared the plant’s usual, human-led sequencing of steel grades with an algorithm designed to reduce losses during continuous casting. They connected production data to a dynamic LCA model, allowing impacts and costs to be assessed across different planning periods while accounting for uncertainty and variation over time.
In the modelled periods, the algorithm reduced the share of steel scrapped during casting, but the resulting change in global warming impacts was small and inconclusive. Cost savings were more consistent but modest, averaging about 0.31%; much of the saving came from using less electricity and fewer additives that would otherwise be lost with scrapped steel. The findings suggest that digitalisation can support incremental efficiency gains, while its climate benefit depends on the plant’s operations and the assumptions used in the assessment.
The authors also demonstrate a practical way to link operational data, optimisation tools and an LCA model for decision support. They note that further case studies are needed before drawing conclusions about the wider steel industry.
The study also distinguishes temporal variability from model uncertainty, helping readers judge how robust the observed differences are.
