The proposed paper deal with a novel approach to evaluate the work hardening in the roll bite using the Orowan hot rolling equilibrium equation, under the assumption of a known rolling pressure distribution along the contact arc and the plane strain plasticity condition. This formulation leads to a first-order differential equation, d sigma/d epsilon = Omega sigma, wherein work hardening is directly proportional to equivalent stress (related to dislocation density and underlying metallurgical complexity) and a function Omega depending only on strain through the rolling pass geometry. The implementation of this approach in the incremental plasticity framework with iterative calculation scheme, makes it possible to reconstruct the stress-strain path of the material along the deformation history and to identify metallurgical transitions during hot rolling from deviations in the stress-strain response or changes in the work-hardening rate. The off-line predictive module is based on a machine-learning model trained on a large database comprising the results of the real-time module together with additional information on chemical composition, microstructural features, and tensile properties of the final products. A key advantage of the proposed incremental plasticity approach is its ability to drastically reduce the number of required constitutive parameters, thereby improving robustness, transferability, and suitability for real-time industrial implementation.
A hot rolled steel strip that seemingly comes out as flat after the final rolling pass might potentially end up with flatness issues after it has been coiled. It is not easily understood which mechanisms in the coiling process are causing flatness issues. It is known from a material perspective that a combination of high stresses and temperatures can cause stress relaxations and creep deformations when the time in this state is long enough. A hot coiled steel strip at 600 degrees C with a mass of 27 tonnes will take long time to cool down and it is uncertain whether stress recovery and creep behaviour have an impact on the final flatness. To investigate this, a three-dimensional thermo-mechanical finite element model with a creep material model is used to simulate the influence of creep deformations on final shape. It is, on one hand found that a relatively complex stress profiles are developed through the strip thickness when coiling, with compressive and tensile stresses beyond the yield stress, and that the tensile stresses recover unsymmetrically on one side of the strip midplane. On the other hand, is it also found that the creep deformations are only a fraction of the plastic deformations caused by the mechanical work during the coiling process. Hence, it is concluded that creep mechanisms play no, or possibly a marginal role in the shape of the final shape. Whereas stress relaxations result in a stress neutral profile in the coiled state, which may cause shape variation after uncoiling and post processing.
In steel product manufacturing, the demand for thermo-mechanical controlled processing (TMCP) is rising to achieve various desired properties. However, uneven temperature distribution during cooling can cause shape defects. Therefore, accurately predicting thermal deformation is vital for optimizing cooling conditions. This task is challenging because thermal deformation results from complex interactions among strain, heat transfer, and phase transformations. In particular, transformation plasticity significantly influences final product quality, affecting shape and residual stresses. This study developed a thermal deformation prediction model that incorporates transformation plasticity through multi-phase transformations. The model's accuracy and validity were assessed by comparing experimental results with simulations of camber in plate samples subjected to one-sided spray cooling. The findings revealed discrepancies between experiments and simulations that excluded transformation plasticity, whereas including it led to good agreement. This prediction method enhances the understanding of TMCP processes.
QSP-DUE direct casting and rolling technology allows three production modes: coil-to-coil, semi-endless, and full endless. These modes enable the production of various steel grades and strip formats, meeting end-user needs. The maturity of the technology is evidenced by concrete applications. The process offers a wide product mix, catering to various market needs, with perspective towards automotive exposed sector with dedicated features and production strategies. At the same time the commitment to sustainability is evident through the electrification efforts, setting new standards for high-quality, sustainable steel production, driving innovation, and meeting diverse market demands.
To optimize the utilization of roller straightening machine, achieve an even and controlled distribution of the total load across the drive rollers, avoid over-dimensioning of the drives, and protect the machinery from overload, the required drive torque is distributed specifically across the drives using a control system. A new, optimized Load Sharing Control (LSC) has been developed by interdisciplinary cooperation within the SMS group. This new LSC features the elimination of master-slave operation. As a result, it enables control when not all rolls are in operation during pulling in and out. It is independent of the process parameters (roll diameter, roll adjustment) and redistributes the total drive torque to the single drives according to a predefined pattern. Initial functional tests were conducted on forces and drive torques calculated with the aid of finite element (FE) simulations of the roller straightening processes and with available measurements. The successful implementation of the newly developed LSC in four industrial Compact Roller Straighteners (CRS (R)) in the fields of railway rail, medium and heavy section production conclusively demonstrates its effectiveness. The newly optimized LSC offers benefits such as reduced total drive capacity, decreased roll wear, reduced experimental effort during commissioning and improvement of product quality. Furthermore, it can be implemented on cantilever, horizontal, vertical and combined straightening machines as well.
A simple yet effective method has been developed to simulate macro-segregation during continuous casting. The model focuses on a thin slice moving at the casting speed, incorporating temperature and solid/liquid fraction profiles. The model incorporates micro-segregation and two key flow mechanisms: density-driven flow, governed by mass balance, and cavity-driven flow, which occurs as liquid and equiaxed crystals fill the solidification cavity forming at the center during solidification. Density-driven flow leads to positive macro-segregation near the center in both slabs and blooms, but the effect is significantly stronger in blooms. This is attributed to their higher Flow Contribution Ratio (FCR), which quantifies the relative contribution of Y-axis flow to overall flow. The higher FCR is likely influenced by greater solidification shrinkage. The cavity-driven flow model successfully explains the observed positive segregation peak at the center and the negative segregation peak at the total shrinkage location. The model's predictions closely match experimental data and provide valuable insights for optimizing casting parameters and soft reduction patterns to minimize center-segregation. With its fast computation speed, the model is well-suited for both offline analysis and real-time implementation, making it a valuable tool for research and industrial applications.
The processes of electric steelmaking are complex and difficult to control to achieve sustainable production. To strive towards competitiveness and green transformation, steelmakers apply the Electric Arc Furnace (EAF) to circulate scrap into new products. This saves resources compared to iron-ore-based production, enabling both circular economy and green energy sources. However, efficient EAF operation faces difficulties in state monitoring and control decisions. Fortunately, the control can be facilitated with optimization based on Artificial Intelligence (AI) and Digital Twins (DT). Still, DT accuracy can suffer from input data fluctuation or coverage limitations in development and validation datasets. The fluctuation stems from the environment-related variation, especially the scattering and not-exactly-known chemical composition of secondary raw materials. For decision support, this work suggests a DT framework with Federated Learning (FL) for multi-plant schemes, focusing on electric steelmaking. The framework can deliver both historical data and message-oriented online data to the DTs. It builds upon a container orchestration system (Kubernetes) for software lifecycle management and resource scaling. Importantly, the framework implements FL to exploit network-wide knowledge. That is, the DTs share knowledge with a centralized server that aggregates a global model distributed to the participants, broadening data diversity. Still, all data remain local, which preserves privacy. The framework applies FL for two types of process DTs, EAF and the subsequent Ladle Furnace (LF). FL can optimize EAF parameters although EAF is not AI but a dynamic model. Conversely, the LF model is composed of a set of neural networks. The results from a prototype system with actual data prove the concept. Firstly, the DTs accurately estimate process variables online, such as the chemical composition and temperature. Secondly, FL experiments indicate potential for model parameter optimization and enhanced performance. Besides, the framework concept is applicable for even more DTs and across industries.
This study investigates the effect of starter-block geometry, expressed by the diameter-to-height (D/H) ratio, on crystallographic orientation and creep rupture behavior of CMSX-4 single-crystal superalloys. Directional solidification was performed using the vertical Bridgman process at a constant withdrawal rate of9 mm/min with starter blocks of D/H = 0.3, 0.5, and 0.7. The resulting misorientations from the preferred [001] direction were 6.3 degrees, 21.7 degrees, and 3.3 degrees, respectively, indicating that crystallographic orientation does not vary monotonically with the D/H ratio. Complementary thermal simulations performed using ProCAST showed distinct transitions in solidification-front stability across the selected geometries, providing a physical explanation for the non-monotonic trend in misorientation. Creep rupture life at 850 degrees C and 560 MPa was measured to be 97.8, 24.6, and 122.6 h for the three samples, demonstrating a strong inverse dependence of creep resistance on crystallographic deviation. Fractographic analyses revealed more severe interdendritic cavitation and porosity coalescence in the highly misoriented specimens, confirming cavitation-dominated failure. In the best-oriented sample (3.3 degrees), EDS measurements near creep cavities showed Re depletion in the gamma/gamma ' matrix and the formation of Re-rich secondary precipitates, indicating diffusion-driven segregation and stress-assisted precipitation during creep. Overall, the results show that crystallographic alignment, solidification-front stability, and porosity control are jointly critical for optimizing the high-temperature creep performance of CMSX-4 single crystals.
Continuous casting machine (CCO) is fundamental to produce steel at reasonable costs. The tundish represents the main source of steel for this kind of facility, acting as a buffer between the ladle and the nozzles feeding casting profiles. This component significantly influences the quality and yield of continuous casting operations. Our study focuses on leveraging Computational Fluid Dynamics (CFD) to track particle behaviour within the tundish. By simulating the flow patterns, we aim to optimize tundish design in terms of steel cleanliness. We defined two main scenarios that have different impact on molten steel flow and inclusion transport, with the intention to minimize defects and enhance overall efficiency. Our approach involves tracking individual particles-both in terms of trajectory and residence time-allowing us to understand their impact on product quality. Intense postprocessing has been done with python to perform analysis on a structured framework, making comparison between the two cases much easier. Finally, the head reason is to reveal by simulations how inclusions move and could accumulate within the steel flow. This kind of knowledge prompt strategies to reduce inclusions in the final product.
Solidification plays a fundamental role in metallurgical processes, such as continuous casting of steel billets, as it greatly influences the final microstructure, in terms of chemical segregation, grains morphology and size. Segregation of solute elements occurs because of their partitioning between the liquid and the growing solid. The liquid interdendritic regions are enriched of solute elements rejected by the growing dendrites. Such segregation cannot be fully mitigated through subsequent hot rolling processes. Common evidence of this undesirable phenomenon is the banded microstructure observed in hot-rolled products, characterized by alternating longitudinal bands of ferrite and pearlite. Normalization can reduce the banded microstructure, but micro-segregation remains difficult to completely homogenize. Therefore, hot-rolled wire rods produced from steel billets solidified under different conditions exhibit varying mechanical properties. In this context, the study presents a comparative analysis of two low-carbon steel billets with similar chemical composition but produced under different casting conditions with two different casting machines. The analysis investigated the differences in macro and micro-structure of the two billets by measuring the secondary dendrite arm spacing (SDAS) and the segregation index. The mechanical properties of wire rods derived from the billets were assessed by tensile tests and micro-hardness Vickers measurements. The microstructures (of billets and wire rods) were analyzed by optical and scanning electron microscopy (SEM) equipped with an energy dispersive spectrometer (EDS).
Decarbonizing steelmaking is one of the greatest challenges facing the steel industry today. Electrifying steel production is a pivotal step in reducing CO2 emissions. Over the coming years, the installed base of high-power EAF is expected to grow significantly, which will affect power quality but also EAF performances. To address these challenges, GE Vernova has developed an innovative solution. The Direct Feed system connects directly to the grid, enabling precise and highly stable electrode current regulation. Design of the Direct Feed system will be presented with performance results derived from simulations and on-site measurements. Key outcomes, including improvements in EAF flicker reduction and operational performance, are highlighted.
HIYIELD project applies advanced technologies to enhance circular economy practices by increasing scrap usage in steel production and reducing reliance on pig iron from coal-fired blast furnaces. The project's objectives are structured across three industrial demo cases, each addressing a critical aspect of scrap utilization. In demo case 1, industrial-scale trials were conducted to optimize scrap sorting through mechanical, physical, and sensor-based separation techniques. A hammer mill-based process achieved a ferrous yield of 99.5% purity, with a magnetic separation efficiency of 91%. In addition, a laser scanner system was implemented for real-time scrap volume estimation, improving charge optimization for steelmaking. A Deep Learning (DL) based classification model was developed to enhance automated scrap recognition, integrating Electric Arc Furnace (EAF) process data and real-time imaging for improved material characterization. In demo case 2, industrial trials were conducted to optimize the identification, classification, and processing of pre-consumer scrap using X-Ray Fluorescence (XRF) based separation and DL-based models. The implementation of the Digital Scrap Information Card (DiSC) enabled efficient data exchange between suppliers and consumers, ensuring accurate scrap tracking. Furthermore, a DL-based scrap identification system utilizing Self-Supervised Learning (SSL) models for automated scrap classification was developed, improving scrap assessment. In demo case 3, High-Speed Sampling (HSS) and an analysis system were developed for direct on-site characterization of liquid steel. The chemical compositions obtained from the combined HSS, and conventional lollipop sampling system were analysed, showing strong agreement between the two sampling methods. This high level of consistency confirms the accuracy and reliability of HSS sampling for immediate steel analysis. The project's findings support increased scrap usage in steelmaking, enhanced process efficiency, and reduced environmental impact, aligning with the EU's long-term decarbonisation and circular economy goals.
Hardness testing is a key procedure in materials science for evaluating mechanical properties and process quality. Traditional Vickers hardness measurement relies on manual identification of indentation diagonals, a process that is slow, subjective, and prone to variability. This work introduces a deep learning-based pipeline for fully automated Vickers hardness measurement, combining instance segmentation via Mask R-CNN with sub-pixel geometric fitting for diagonal extraction. A dataset of 403 micrographs of samples under loads from 10 gf to 2000 gf was assembled and annotated for training and validation. Hyperparameter optimisation was performed using a Taguchi design of experiments, and the final model achieved near-perfect segmentation accuracy (overall AP approximate to 90.5%) on the test set. Measurement accuracy was assessed against manual ground truth, yielding mean relative errors of 1.6-1.9% for the two diagonals, with most cases within 2-3%. These results demonstrate that the proposed system provides robust detection, high metrological precision, and reproducible performance across diverse imaging conditions, paving the way for reliable, high-throughput hardness testing in industrial and research settings.
In the primary steel production, hot metal produced in the blast furnace (BF) is fed via a ladle to the Basic Oxygen Furnace (BOF), where it is converted to liquid steel. During the metallurgical operations, slag can form and solidify on the refractory walls during tapping. These oxide deposits must be periodically removed, to ensure the regularity of BOF operations over time. Cleaning operations slow the process, as they require dedicated plant operations. For this reason, a collaboration between RINA-CSM and ADI was set up to find a solution to shorten the BOF "cleaning times", by managing the injected o8xygen flow. The investigated solution (slag-cutting baffle) is based on two countermeasures: an operational one, jet "upstream" flow management, and one "downstream", thanks to an appropriately designed baffle, adaptable to the heads of the oxygen lances, to properly guide the jet. To comply with the required effects without impacting on internal lining safety, the oxygen jet blown must meet certain requirements. First, it must be oriented and concentrated within a "blade" shape, for a "compact" stream. Furthermore, the jet in the BOF must be fast enough to provide for slag melting, but without local velocity "hot spots", harmful to internal lining integrity. Therefore, different configurations were designed for the deflecting system (walls, slits), based on the geometric characteristics of the oxygen lances tips. The study presented hereinafter shows the approach to the problem, the configurations designed, the evaluation criteria of the expected performance, and the results of the computational fluid dynamics (CFD) simulations carried out to verify jet performance. This study made it possible to identify critical issues in the initial configurations, and then to fix them, with solutions considered reliable and industrially applied.
The achievement of C-lean and sustainable steelmaking processes is one of the challenges of the European Green Deal to target the climate neutrality by 2050. In this context, electric steelmaking is investigating alternatives to improve the sustainability of its production routes. The use of alternative carbon-bearing materials in electric arc furnaces and the replacement of natural gas with green hydrogen in related burners are two promising solutions. However, investigations are fundamental to assessing the viability of different technological solutions and their possible combination by avoiding unexpected process and product issues. Therefore, next to industrial trials, simulations are important to broaden the investigation, as they enable exploration of process configurations and the use of materials that are also quite far from conventional practices and that cannot be directly investigated through experiments for economic and practical constraints, such as material unavailability and high costs. The contribution focuses on the results of pilot trials and simulations done with an updated flowsheet model of the electric steelmaking route.
Advancements in modelling are transforming the metallurgy sector, providing more precise tools for quality control and process optimization. In this context, within the framework of an EU-financed dissemination project (METACAST, Research Fund for Coal and Steel), a comprehensive review of solidification modelling in continuous casting and of the research landscape in Europe and worldwide has been performed. Modelling steel continuous casting and solidification is essential for the accurate optimization of process parameters, such as casting speed and secondary cooling, in order to minimize defects like cracks and segregation. The use of advanced numerical simulations allows the prediction of dendritic structures and improves the quality of the final product. The research examines the fundamentals of thermodynamics, solidification kinetics, and fluid dynamics, and highlights the interplay among heat flow, mass transfer, and thermal stresses, showing their relevance in predicting microstructure formation and defect control. Techniques such as numerical simulation and thermal analysis are used to predict the formation of porosity and shrinkage, thereby improving the quality of the final product. The integration of advanced technologies, such as artificial intelligenceand machine learning, is opening new frontiers in solidification modelling, allowing for greater precision and adaptability of models to various production processes. The scope of the work was to map solidification models and research groups present in Europe by gathering statistical data on their countries of origin and on the areas of interest of the developed models, and to identify common lines between them and compare EU expertise within the global context.
This paper presents a novel, physically informed machine learning (ML) framework for the accurate modeling and visualization of steel microstructure evolution during annealing. Utilizing a chained support vector regression (SVR) architecture with optimized hyperparameters, the model sequentially predicts key microstructural states, ensuring metallurgical consistency. The process begins by forecasting recrystallization fraction (RF) kinetics, which is critically constrained by the classical Johnson-Mehl-Avrami-Kolmogorov (JMAK) model. The resulting JMAK-corrected RF then serves as a fundamental input to subsequent SVR models, which forecast the average grain size (AGS) and, finally, essential image-based microstructural features (mean and standard deviation of pixel intensity). This chained approach inherently prioritizes physically sound outputs, avoiding the consistency issues of isolated predictions. A unique visualization methodology is introduced, which selects and maps the closest experimental inverse pole figure (IPF) maps to the predicted states. This robust, multi-stage framework establishes a powerful, data-driven tool for simulating complex material evolution, thus minimizing the need for extensive experimental operations in materials design and process optimization.
The aim of the present work is to evaluate the localized corrosion behaviour of an AISI 316L stainless steel alloy produced using different additive manufacturing technologies, comparing it with products manufactured through conventional way. Potentiodynamic polarization and critical pitting temperature (CPT) tests were carried out. The results of the potentiodynamic polarization tests showed differences in breakdown potentials between materials produced with additive technologies and those produced by traditional methods. The CPT tests yielded results consistent with those observed in the potentiodynamic tests. The results of CPT tests are in agreement with the potentipdynamic tests, confirming that the pitting corrosion behaviour is strictly dependent upon the additive manufacturing technologies. The article stresses once more the importance of defining suitable protocols for the corrosion qualification of alloys obtained by additive manufacturing that take into account the specificity of the production process and the treatments performed on the alloys themselves.
The most commonly used alloys in orthodontic wires are stainless steel and titanium due to their corrosion-resistant properties. Whilethesealloysgenerallyexhibitgoodcorrosion behavior in mildlyaggressiveenvironments, unexpected breakages and allergic reactions can sometimes occur. This is mainly due to the degradation of the passive film of orthodontic wires, which is caused by the complex and variable conditions within the oral cavity. The composition of human saliva and the duration of exposure can differ significantly depending on the patient. This study evaluated the corrosion resistance properties of two commercially available orthodontic wires: 304 steel and a nickel-titanium alloy. Their electrochemical behavior was investigated using cyclic potentiodynamic polarization curves recorded in artificial saliva at 37 degrees C and a pH of 5.5. The results showed that 304 stainless steel orthodontic wires exhibit variable corrosion resistance properties, ranging from excellent to poor. Furthermore, when the passive film on 304 stainless steel breaks down, the repassivation potential is lower than the corrosion potential. This indicates that damage can easily propagate, leading to significant failure and the release of metal ions. This poses risk of allergies. On the other hand, the nickel-titanium orthodontic wires never showed localized corrosion behavior, demonstrating superior corrosion resistance properties compared to 304.
Binder Jetting (BJ) is an additive manufacturing technology that can produce alloy components at a higher speed and resolution than other systems. However, the localized corrosion resistance properties of BJ stainless steel are lower than those obtained with conventional manufacturing processes. This study aims to evaluate the effects of passivation treatments on the corrosion behavior of as-sintered 17-4 PH samples fabricated through BJ. The samples were treated with four acidic solutions: 15%, 20%, 40% Eta NO3, and 40% HNO3 + 1% HF. The localized corrosion resistance properties were evaluated through Cyclic Potentiodynamic Polarization (CPP) tests in a neutral pH sodium chloride electrolyte. The treatments significantly enhanced the localized corrosion resistance properties of as-sintered BJ samples, determining the typical passive anodic behavior of the CPP curve, which was not shown in the untreated samples. Moreover, higher concentrations of HNO3 (40%) improved the pitting corrosion resistance, while adding HF was ineffective. The study paves the way for broader industrial applications of additive-manufactured stainless steel, emphasizing the critical role of improving localized corrosion resistance properties through passive treatments.
