Aiming to reach circularity and resource efficiency, the metal industry pushes towards recycling secondary products instead of producing from primary material. This increases the use of scrap, which can bring about several benefits but could also come at the expense of the materials' quality and the potential loss of valuable resources. The above is mainly a result of the likely accumulation of unwanted elements throughout the recycling process, which may dissolve and become difficult to extract from the metal's melt, turning into what is known as tramp elements. This study focuses on the opportunity to limit the accumulation of unwanted elements before they end up in the molten solution. Taking an exploratory approach with the use of observation and expert interviews, this study examined where and how unwanted elements enter the recycling system. Eight element types were identified and categorised after the intentionality of their addition and their desirability in the end product. Thereafter, this article proposes a taxonomy based on the way in which these are present inside the furnace before melting, suggesting that the manner elucidates the entry points where impurities are introduced in the recycling stream. By introducing a taxonomy, this study aims to pave the way for developing strategies and research on how to minimise or prevent the presence of these elements in recycled metals, thereby increasing the quality of the recycling process.

While steel is the most recycled material, it is often downcycled due to the accumulation of unwanted elements. These not only diminish its quality but also represent a resource loss. Previous research has looked into the causes behind the accumulation of tramp elements in the metal and the resource loss stemming from current recycling processes. Nevertheless, these rarely consider the complex interactions between industrial practices, resource flows and system constraints, nor examine time delays in the system. This exploratory study is the first attempt to use Causal Loop Diagrams to bring together and map these interdependencies, incorporating time delays to better understand the accumulation of unwanted elements and its impact on the future of the steel recycling industry. The findings suggest that persisting with current practices risks creating accumulations of overly-contaminated, unusable scrap, thereby jeopardising the future of metal recycling.

Uncertainties with respect to the chemical composition of scrap limit its suitability as an input to recycling. This study offers an alternative approach in dealing with this concern and explores the hypothetical case where this uncertainty is nonexistent. The effect of fully knowing the scrap composition is simulated using an optimization software adopted to scrap-based, stainless-steel production. Through the systematic implementation of this information-driven model in the studied cases, the results suggest that with access to perfect information, recycling incentives can be realized. Essentially, the steel scraps’ consumption increased since it was possible to select and combine scrap quantities with varying composition profiles to achieve the targeted product compositions. This also meant that elements already in the scrap were allocated in a manner that was less dependent on pure alloy additions. Being able to demonstrate the value of information on scrap composition could rationalize upgrades on current scrap management systems.

The current work presents an optimization approach for improving stainless steel's performance and environmental impact. A Taguchi-based grey relational analysis is used to determine the optimal alloy design of six stainless steels (304, 316L, 904L, 2304, 2205, 2507). The alloy design of a multivariable system (C, Mn, Cr, Ni, N, Mo, and Cu) is initially formulated based on the Taguchi orthogonal array. It combines the multiperformances (pitting corrosion resistance, proof strength, tensile strength, and greenhouse gas emissions) to determine a numerical score and suggests the optimal combination of alloy contents. The results are further analyzed using a range analysis to detect the significance of each alloy parameter and its level. The results show that nitrogen plays the most important role in determining the stainless steel's combined performance. Besides, the optimum alloy design consists of a high content of nitrogen, chromium, molybdenum, and copper combined with a low content of nickel. The study has a focus on proposing the systematic approach, Taguchi-based grey relational analysis, to optimize the stainless steel's different mechanical properties with respect to environmental impact.

The present work focuses on predicting the steel melt temperature following the vacuum treatment step in a vacuum tank degasser (VTD). The primary objective is to establish a comprehensive methodology for developing and validating machine learning (ML) models within this context. Another objective is to evaluate the model by analyzing the alignment of the SHAP values with metallurgical domain expectations, thereby validating the model’s predictions from a metallurgical perspective. The proposed methodology employs a Random Forest model, incorporating a grid search with domain-informed variables grouped into batches, and a robust model-selection criterion that ensures optimal predictive performance, while keeping the model as simple and stable as possible. Furthermore, the Shapley Additive Explanations (SHAP) algorithm is employed to interpret the model’s predictions. The selected model achieved a mean adjusted (Formula presented.) of 0.631 and a hit ratio of 75.3% for a prediction error within ±5 °C. Despite the moderate predictive performance, SHAP highlighted several aspects consistent with metallurgical domain expertise, emphasizing the importance of domain knowledge in interpreting ML models. Improving data quality and refining the model framework could enhance predictive performance.

The present work aims to answer three essential research questions (RQs) that have previously not been explicitly dealt with in the field of applied machine learning (ML) in steel process engineering. RQ1: How many training data points are needed to create a model with near-upper-bound predictive performance on test data? RQ2: What is the near-upper-bound predictive performance on test data? RQ3: For how long can a model be used before its predictive performance starts to decrease? A methodology to answer these RQs is proposed. The methodology uses a developed sampling algorithm that samples numerous unique training and test datasets. Each sample was used to create one ML model. The predictive performance of the resulting ML models was analyzed using common statistical tools. The proposed methodology was applied to four disparate datasets from the steel industry in order to externally validate the experimental results. It was shown that the proposed methodology can be used to answer each of the three RQs. Furthermore, a few findings that contradict established ML knowledge were also found during the application of the proposed methodology.

The current nature of steel design and production is a response to meet increasingly demanding applications but without much consideration of end-of-life scenarios. The scrap handling infrastructure, particularly the characterization and sorting, is unable to match the complexity of scrapped products. This is manifested in problems of intermixing and contamination in the scrap flows, especially for obsolete scrap. Also, the segmentation of scrap classes in standards with respect to chemical compositions is based on tolerance ranges. Thus, variation in scrap composition exists even within the same scrap type. This study applies the concept of expected value of perfect information (EPVI) to the context of steel recycling. More specifically, it sets out to examine the difference between having partial and full information on scrap composition by using a raw material optimization software. Three different scenarios with different constraints were used to appraise this difference in terms of production and excess costs. With access to perfect information, production costs decreased by 8–10%, and excess costs became negligible. Overall, comparing the respective results gave meaningful insights on the value of reestablishing the compositional information of scrap at the end of its use phase. Furthermore, the results provided relevant findings and contribute to the ongoing discussions on the seemingly disparate prioritization of economic and environmental incentives with respect to the recycling of steel.

Bioenergy with carbon capture and storage (CCS) in iron and steel production offers significant potential for CO2 emission reduction and may even result in carbon-negative steel. With a strong ambition to reach net-zero emissions, some countries, such as Sweden, have recently proposed measures to incentivise bioenergy with CCS (BECCS), which opens a window of opportunities to enable the production of carbon-negative steel. One of the main potential applications of this route is to decarbonise the iron reduction processes that account for 85 % of the total CO2 emission in the iron and steel plants. In this study, gasification is proposed to convert biomass into biosyngas to reduce iron ore directly. Different cases of integrating the biomass gasifier, Direct Reduced Iron (DRI) shaft furnace, and CCS are evaluated through process simulation work. Based on the result of the work, the proposed biosyngas DRI route has comparable energy demand compared to other DRI routes, such as the well-established coal gasification and natural gas DRI route. The proposed process can also capture 0.65-1.13 t of CO2 per t DRI depending on the integration scenarios, which indicates a promising route to achieving carbon-negative steel production.

This work presents a process model developed based on mass and energy conservation to assess high carbon ferrochrome production from cradle to gate through four supply routes: (1) a conventional submerged arc furnace (SAF), (2) a closed submerged arc furnace with preheating (CSAF+PH), (3) a closed submerged arc furnace with 60% prereduction (CSAF+PR60%) and (4) a direct-current arc furnace (DCAF). The energy requirements are between 40 and 59 GJ/t FeCr (74–111 GJ/t Cr), and the greenhouse gas (GHG) emissions range between 1.8 and 5.5 tCO2-eq/t FeCr (3.3–10.3 tCO2-eq/t Cr). The upgrading of coal-powered SAF process to a closed furnace CSAF+PH and CSAF+PR60% contributes to an emission reduction of 23% and 18%, respectively. Moreover, the use of hydro-powered electricity leads to a further emission reduction of 68% and 47%, respectively. For CSAF+PR process, the GHG emissions can be reduced by 14% when increasing the pre-reduction ratio from 30% to 80% and decreased by 10% when charging hotter feed from 100 °C to 1000 °C. The proposed process model is feasible in generating site-specific inventory data and allowing for parameter studies as well as supporting companies to improve the transparency of the environmental performance in the FeCr value chain.

A two-dimensional CFD model was developed to simulate the fluid flow and temperature distribution inan AOD nozzle using a mixture of oxygen and nitrogen as the fluid phase, aiming at predicting how the outlet gas properties are influenced by the inlet pressure, inner nozzle length/diameter and process gas composition. The proposed mathematical model assumes a steady, non-isothermal flow condition, using the realizable k-ε turbulence model to describe the gas phase. A mesh sensitivity analysis was performed where predictions were compared to experimental data. The results show that the gas properties are mainly dependent on the inlet pressure, nozzle length/diameter and heating condition but less dependenton the composition of the gas mixture. This fundamental model can be applied to provide a process specified boundary condition for gas blowing when simulating a multiphase flow in a full-scale AOD converter.