Recently, remarkable advances have been made in statistical analyses based on deep-learning techniques. Applied studies of deep learning have been reported in various industrial fields, including the iron and steel-making industries. The production of iron and steel requires a variety of processes, such as the processing of ingredients, iron-making, casting, and rolling. Consequently, the data acquired from them are diverse, and various tasks exist that can be assisted by deep-learning algorithms. Hence, providing a summary of the application is helpful for researchers specializing in information science to grasp the current trend of applied studies on deep learning techniques and for researchers specializing in each field of the iron and steel-making industry to understand what types of deep learning techniques are being utilized in other specialized fields. Therefore, in this study, we summarize the current studies on the application of deep learning in the iron- and steel-making fields by organizing them into several categories of processes and analytical methodologies. Furthermore, based on the results, we discuss future perspectives on the development of deep-learning techniques in this field.

The effect of MgO/CaO ratio on the viscosity and free running temperature of chromium-containing high-titanium blast furnace slag (CaO–SiO₂–Al₂O₃–MgO–TiO₂–Cr₂O₃) was investigated by a rotating crucible viscometer. When the ternary basicity with a fixed (CaO+MgO)/SiO₂ ratio of 1.41 and the temperature was fixed, the MgO/CaO ratio had an obvious influence on the viscosity of slags. Increasing MgO/CaO ratio from 0.34 to 0.44 caused a slight decrease in the viscosity of the slag, and had an opposite effect when MgO/CaO ratio was more than 0.44. The XRD measurements showed that the technology of ‘‘replacing CaO with MgO’’ has an effect on the precipitation temperature of perovskite phase and spinel phase. According to the Raman spectroscopy results, with the increase of MgO/CaO ratio from 0.34 to 0.44, the DOP decreased, and then increased as the MgO/CaO ratio increased from 0.44 to 0.56.

Thermal mixing during the gas stirring operation and arc heating in a steel ladle is analyzed through the modern tools of a physical model using PIV (Particle Image Velocimetry) and thermal PLIF (Planar Laser Induced Fluorescence), whose velocity and temperature fields were used to fine-tune and validate a multiphase Eulerian two-phase mathematical model. Agreement on both fluid dynamics and thermal evolution is reasonably good between experiments and the predictions obtained by the mathematical model of the physical model. The analysis coming from the numerical model validated by the physical model measurements included the thermal mixing and energy efficiency of single nozzle injection in centric and eccentric (4/5R) gas injection. It turned out that energy efficiency in the centric gas injection is 20% more efficient than in eccentric injection. Then, under the same heat flux provided, the maximum temperature of the water in the centric gas injection would be higher than the maximum temperature reached in the eccentric mode with the same gas flow rate. Good heat transfer happens when the heat source impinges in a fluid region with high circulation and turbulent dispersion.

This study examines the effect of container size on coal briquette’s internal structure using the Discrete Element Method. It found that when the frictional resistance between particle and wall was large and the inner diameter small, the difference in particle filling ratio between the upper and lower parts of the briquette was significant. Conversely, with a larger inner diameter, this difference nearly disappeared. The distribution of contact force indicated that the frictional force’s inhibiting effect on force transmission lessened with a larger container’s inner diameter. The study also revealed that the height of the container affects the briquette’s internal structure, and these results can be summarized by the container’s height to diameter ratio. Essentially, a larger ratio led to a linear increase in the difference in filling ratio between the upper and lower parts of the briquette.

Metallurgical industries often discharge slag containing valuable elements that are poorly utilized when producing copper alloys and silicon-manganese alloys. To improve the utilization rate, in this study, a method to mix copper slag with water-quenched silicon-manganese slag and CaO for roasting and modification was proposed. In this work, FactSage 8.0, DSC-TG, and XRD were used to examine the phase change during the modification process and investigate the impacts of the CaO content, roasting temperature, and holding time on the modification effect. The results showed that the addition of water-quenched silicon-manganese slag and CaO could effectively promote the transformation of fayalite to (Mn, Mg, Fe)Fe₂O₄, with the highest conversion rate occurring at a 10% CaO content. An increase in the temperature and prolongation of the time facilitated fayalite transformation, but excessive temperature or time could result in iron loss. The optimal recovery rate and iron grade were achieved with roasting at 1400°C for 60 min. This method can provide a concentrate suitable for producing copper-containing antibacterial stainless steel and wear-resistant cast iron, and the tailings can be used to produce ceramic materials.

Air-quenched electric arc furnace slag (AEAFS) is a black sphere or spheroid particle prepared by an air quenching theology using electric arc furnace steelmaking slag as raw materials, possessing the characteristics of small particle size, moderate density and high hardness Combined with the tight supply and demand of the existing abrasive market and the continuous increase in price, AEAFS is tried to be used as a free abrasive for sandblasting processing according to its physical characteristics. In order to make sure that the AEAFS meets the requirement of free abrasive blasting, it is necessary to conduct a comprehensive and in-depth analysis of its physical and chemical properties. The research shows that the AEAFS is a spherical particle with weak magnetism and particle size being mainly 2.8 mm (accounting for more than 90%). Its Vickers hardness is in the range of 600–1000 HV; its compressive strength is between 20 and 465 N and increases first and then decreases with particle size. The water content is more than 0.019%, except that the particle size is less than 0.5 mm. All the others meet the requirements of ISO-11126-6: 2018 standard. The content of f-CaO is between 1.122% and 1.612% increasing with the particle size, AEAFS has good chemical stability and weak acid resistance. In summary, AEAFS meets the performance requirements of the medium used in the sandblasting process and is a potential alternative product for sandblasting abrasives.

In the continuous casting process, the temperature of liquid steel in tundish determines the casting speed and secondary cooling conditions, and then influences the billet quality. It’s very important to measure the temperature of liquid steel in tundish quickly and accurately. However, the initial response lag of blackbody cavity sensor is inevitable since the time is required for the sensor inner wall and the liquid steel reaching thermal equilibrium by heat transfer. In this paper, in order to eliminate the initial response lag of sensor, a heat transfer model of sensor is established. The heat transfer characteristics and cavity integral emissivity of sensor with different depths immersed into liquid steel are analyzed. The analytical solution of sensor temperature is derived by separation of variables method and superposition principle, and is verified by the actual temperature measurement data. Then an innovative method of liquid steel temperature rapid identification is deduced and validated by the actual measurement data. The results show that the initial response lag of sensor is greatly shortened and the temperature measurement efficiency is improved. This study provides a theoretical method for improving the initial response speed of sensor.

Friction stir welding (FSW) is expected to be applied as a welding technique of materials with relatively high melting temperature such as steel materials. Silicon nitride is one of the inexpensive and attractive tool materials for FSW of the thick steel plate. Therefore, in this study, the capability of the silicon nitride tool without groove scroll to weld a low carbon steel plate with a thickness of 15 mm was investigated. The suitability of a tool shape was confirmed by FSW of a thick A5052 plate using a SKD61 tool with same shape as the silicon nitride tool. The defect-free welded specimen of the thick steel plate was obtained using the silicon nitride tool under the optimum welding condition. The silicon nitride tool could be used for FSW of the 15 mm thick steel plate until the welding length of 200 mm without breaking the tool. The groove defect area in the stir zone of the thick steel plate was decreased with decreasing of the tool rotation speed and tool tilt angle. Especially, the tool tilt angle was effective to increase the heat input and the material flow velocity. It is considered that the defect-free weld specimen of the thick steel plate was obtained to sufficient material supply to the RS of the stir zone by decreasing tool tilt angle to 1°.

A brace-type seismic damper made of an Fe-15Mn-11Cr-7.5Ni-4Si alloy solidified in the ferrite-austenite (FA) mode and SN490B steel, which can be constructed via welding, was proposed. To realize the proposed seismic damper, gas metal arc welding was applied to produce similar FMS/FMS fillet welds and dissimilar FMS/SN490B fillet weld joints. Based on the Schaeffler diagram, similar and dissimilar welding consumables were designed such that the fillet weld metal solidified in the FA mode without solidification cracking. Sound similar fillet welded joints were obtained using two types of welding consumables with different Cr/Ni equivalent ratios although both the similar fillet weld metals had a coarse columnar austenite grain structure. These displayed higher tensile strengths (716–736 MPa) and marginally lower elongations (67–70%) than the FMS alloy. Moreover, a similar fillet weld metal with a chemical composition almost identical to that of the FMS alloy exhibited a remarkable low-cycle fatigue life (5740 cycles). This was shorter than that of the FMS alloy (9351 cycles) owing to the easier formation of α’-martensite. A dissimilar fillet welded joint with a chemical composition within the austenite region was produced without solidification cracking. The dissimilar fillet weld metal showed high tensile strength (867 MPa) and total elongation (61%). These were comparable to those of similar fillet weld metals.

In the gas wiping process used in hot-dip galvanizing, the coating thickness has two thinning limits. The first is the limit due to splashing of the liquid film of molten zinc, and the second is the thinning limit of the wiping capacity of the equipment.

In this study, we investigated the possibility that wiping efficiency is reduced by the effect of zinc solidification due to gas jet cooling by conducting a gas wiping experiment under various temperature conditions.

A galvanized steel strip with a width of 100 mm was immersed in a molten zinc bath in the air atmosphere. The steel strip was heated by induction heating or a gas burner, and the wiping gas was also heated.

The results clarified the fact that high temperature conditions improved gas wiping efficiency. It is suggested that high wiping efficiency is prevented by an increase in viscosity due to an increasing solid volume fraction in the liquid zinc film surface caused by microscopic solidification. In addition, it was also found that the development of the initial alloy layer reduced the amount of liquid phase, which inhibits wiping.

The microstructure of the B1-type TiC formed during solidification and its mechanical properties were investigated using arc-melted Fe–Ti–C ternary alloys. The TiC formed at relatively high temperatures in the liquid as the primary phase exhibited a dendritic shape. With decreasing temperature and/or decreasing Ti and C content in the liquid, the morphology of the TiC changed to a cubic shape with a {001}_TiC habit plane, a plate shape with a {011}_TiC habit plane, and a needle shape with a preferential growth direction of <001>_TiC. The morphology of the TiC was characterized by the anisotropy of its surface energy and its growth rate. The cubic shape with a {001}_TiC habit plane was formed as a result of the {001}_TiC surface exhibiting the lowest surface energy among the TiC surfaces. However, the plate shape with a {011}_TiC habit plane and the needle shape with a <001>_TiC preferential growth direction likely formed because the slowest and fastest growth rates corresponded to the <011>_TiC and <001>_TiC directions, respectively. At room temperature, the alloy with dendritic TiC was fractured in the elastic deformation region because TiC exhibited no plastic deformation. However, the results obtained at 800°C suggested that the TiC exhibited plastic deformability and that the alloy with the dendritic TiC was also plastically deformed.

Atomic-scale lattice defects beneath various hydrogen embrittlement fracture process regions, i.e., different fracture modes, were compared for martensitic steel. The crack initiation region containing quasi-cleavage (QC) and intergranular (IG) fracture, the crack propagation region mostly consisting of IG fracture, and the final fracture region completely composed of microvoid coalescence (MVC) were extracted from the same fracture surface and analyzed by low-temperature thermal desorption spectroscopy (L-TDS). The relative shapes of the L-TDS curves were different at the sampling positions, demonstrating at the atomic-scale that the types and numbers of lattice defects formed vary depending on the fracture process and fracture mode.