Nozzle clogging is widely present in low-carbon steel, stainless steel, rare earth steel and other steel. Numerical simulation of nozzle clogging during continuous casting can reveal the formation process of deposits and quantitatively analyze the influence of various factors on clogging. The results provide theoretical guidance for controlling and mitigating nozzle clogging. However, current simulations are performed off-line. Furthermore, numerical simulation can predict the location where clog deposits break off, the size of the broken fragments, as well as their distribution and inheritance within the cast strand. To date, researchers have developed corresponding mathematical models based on the four main formation mechanisms of nozzle clogging, and the findings have been validated through both laboratory and industrial experiments. Nevertheless, nozzle clogging incidents still occur occasionally in practice. There is currently a lack of real-time prediction methods for the evolution of clogging. Future studies on numerical simulation of nozzle clogging may focus on real-time forecasting of clogging development and prediction of the distribution of broken clog materials in the cast product.

Crack breakout often occurs during the continuous casting process of peritectic steel, which seriously restricts the high-efficiency and high-quality production of peritectic steel continuous casting. Crack breakout mainly originates from solidification defects of the initial solidified shell in the mold. The phase transformation shrinkage of peritectic reaction forms air gaps, leading to uneven heat transfer and uneven thickness of the solidified shell, and crack breakout is prone to occur at the weak parts of the shell. This paper reviews the formation mechanism and conditions of depressions in peritectic steel continuous casting slabs, summarizes the typical characteristics and main influencing factors of slab corner crack breakout, wide-face off-corner breakout and wide-face center crack breakout, and also summarizes the influence laws of steel composition and process factors on crack formation. Aiming at the frequent crack breakout of peritectic steel, this paper puts forward requirements for improving crack breakout from aspects of process factors, mold flux properties and mold design, and points out that establishing a crack breakout risk prediction model for the continuous casting process based on the key influencing factors of crack breakout, and conducting early warning and timely intervention on accident precursor characteristics, is the development direction of future breakout prediction models.

Aiming at the prevalent issue of carbon segregation during the continuous casting of 82B steel, a comprehensive model integrating fluid flow, heat transfer, solidification, and electromagnetic stirring was developed to analyze carbon segregation. The model was validated through measurements of slab surface temperature, magnetic field strength of the Final Electromagnetic Stirring (F-EMS), and actual segregation patterns in the cast slab. The effects of F-EMS were investigated under a fixed frequency of 5 Hz and various current intensities—400 A, 420 A, 450 A, and 470 A. Results indicate that, under the given conditions, the solidification fraction at a cross-section 8.06 m below the meniscus reached 78%. As the stirring current increased, the degree of central carbon segregation decreased gradually, with the optimum value of 1.00±0.01 achieved at 450 A. Industrial trials confirmed that adjusting the F-EMS current effectively mitigates central carbon accumulation in 82B slabs. However, excessive current (470 A) promotes excessive dispersion of carbon, resulting in negative segregation.

Fe-Mn-Al-C medium manganese steel has great application prospects in automotive lightweighting due to its low density, high strength and high ductility. However, the peritectic phase transition that occurs during the solidification of medium manganese steel can easily lead to the formation of surface longitudinal cracks in continuously cast slabs. Therefore, in this paper, Fe-5.59Mn-0.064Al-0.155C-0.022Ti (mass fraction/%) medium manganese steel was taken as the research object, and the effects of Ti and cooling rate on the evolution of solidification microstructure and the behavior of peritectic phase transition of medium manganese steel were studied with combination of thermodynamic calculation via Thermo-Calc and in-situ observation via high-temperature confocal scanning laser microscopy. The influence mechanism of Ti on the peritectic solidification process was clarified, and the growth kinetics of the solidification microstructure was determined. The results show that the solidification process of medium manganese steel bearing Ti is L→(L+δ)→(L+δ+γ)→(L+γ)→γ, and it belongs to hypereutectic steel. The δ grains are cellular, and their growth rate increases and their size decreases with the increase of cooling rate. During the peritectic reaction, the γ phase nucleates at the δ/L phase interface, contacts each other to form austenite grain boundaries, and finally completely separates the L phase and δ phase. During the peritectic transformation, the migration rate of the δ/γ phase interface is much greater than that of the γ/L phase interface. With the increase of cooling rate, the peritectic reaction temperature decreases. The addition of Ti effectively promotes the nucleation of δ grains and inhibits their growth, making the δ/L phase interface smooth and preventing it from developing into a dendritic morphology. Smaller δ grains increase the velocity of γ phase enveloping δ phase during the peritectic reaction process and the velocity of γ phase engulfing δ phase during the peritectic transformation process, effectively avoiding the occurrence of massive-like transformation.

To optimize the flow field inside the mold during continuous casting of round billets and enhance both surface and internal quality of the cast product, this paper investigates the multi-hole swirling flow nozzle technology. Experiments were conducted using a water model with a geometric scale ratio of 1∶1.625 to examine the flow behavior and meniscus fluctuations in a ø650 mm round billet mold under various conditions, including different numbers of side holes, different casting speeds, and with/without mold electromagnetic stirring (M-EMS). Comparisons were made with a conventional straight nozzle. The results indicate that the multi-hole swirling nozzle significantly reduces the impact depth of molten steel, facilitates low-superheat casting, and contributes to the refinement of the solidification structure. Increasing the number of side holes weakens the downward backflow generated by the nozzle discharge and suppresses the characteristic “inverted triangle” flow pattern at the center of the mold, thereby delaying the transport of solute elements to the central region and reducing centerline segregation. Furthermore, the multi-hole swirling nozzle helps maintain a symmetrical flow field distribution along the longitudinal section, preventing flow bias that may occur when M-EMS is used alone. When combined with M-EMS, the coupled swirling flow effectively stabilizes the meniscus, reducing the fluctuation amplitude from ±3.0 mm with M-EMS alone to within ±1.5 mm. This study provides experimental evidence supporting the application of multi-hole swirling flow nozzle technology in improving the quality of continuous cast round billets.

Surface slag entrapment is a typical defect in continuous casting slabs. Online prediction of slag positions followed by grinding removal can reduce the occurrence of downstream rolled plate defects caused by the hereditary nature of slag. Existing machine learning-based slag prediction models lack a systematic approach to parameter optimization, resulting in high training accuracy but unreliable performance in practical test applications. To address this, prediction models were developed using Support Vector Machine (SVM), Random Forest, and Adaptive Boosting (AdaBoost). The slab sample dataset was preprocessed using the Z′-Score method to handle outliers, and SVM was identified as the most accurate model. A relational model was further established between the test sample accuracy metric—F₂ score (f)—and the training sample metrics: false positive rate (t₁), false negative rate (t₂), and F₂ score (t₃). The model is defined as:f(t)=0.62-1.19t₂t₃-0.89t₃²-0.4t₂²-0.35t₁t₃-0.28t₁t₂, where t₁, t₂, and t₃ are standardized feature values. The Particle Swarm Optimization (PSO) algorithm was applied to optimize model parameters, using f(t) to evaluate the best particle positions during optimization. The optimal parameter set achieved a false positive rate of 22.9%, a false negative rate of 18.6%, and an F₂ score of 0.727 on test samples. This study highlights the relationship between training and test accuracy of the slag prediction model. By leveraging this relationship, model parameters can be systematically optimized based on training performance—even when test accuracy is unknown—to improve practical prediction performance, thereby enabling more effective identification of slag-containing slabs and enhancing product quality stability.

As the primary feedstock for X65 pipeline steel, the continuous casting slab is prone to macrosegregation, which is one of the key defects limiting the performance of steel materials. To control solute segregation during the continuous casting process, a multiphase solidification model was established by coupling melt flow, macroscopic heat transfer, microstructural solidification evolution, and solute transport, with full consideration of the soft reduction process at the solidification end. The model was applied to systematically investigate the influence mechanisms of superheat and casting speed on slab macrosegregation. The computational accuracy was validated through carbon-sulfur analysis of industrial slabs. The results show that as the superheat increases from 14 ℃ to 30 ℃, columnar dendrite growth is enhanced and the equiaxed grain zone narrows. For every 8 ℃ increase in superheat, the equiaxed zone at the slab center decreases by 0.6-0.8 mm, and the segregation index increases by approximately 1.39%. The effect of casting speed on segregation control depends on the matching between the solidification end and the soft reduction zone: at 0.8 m/min, the solidification end (18.46 m) occurs before the reduction zone, resulting in ineffective reduction and a segregation index of 1.151; at 1.1 m/min, the solidification end (23.55 m) falls entirely within the reduction zone, where the reduction effect is most effective, yielding the lowest segregation index of 1.071; at 1.2 m/min, the solidification end (27.95 m) extends beyond the reduction zone, where insufficient coverage causes the segregation index to rise slightly to 1.082. This study clarifies the quantitative relationship among superheat, casting speed, and central segregation under fixed soft reduction conditions, providing a theoretical basis and process optimization guidance for the precise control of macrosegregation in slab continuous casting.

To meet the stringent non-magnetic requirements for 316L austenitic stainless steel in applications such as nuclear power and mobile devices, it is essential to precisely control the residual ferrite content in the cast slab. This study investigates a high-nickel, high-nitrogen 316L austenitic stainless steel slab produced industrially. Using metallographic examination and electron backscatter diffraction (EBSD), the morphology, distribution, and content of ferrite across the thickness direction at the slab centerline were characterized, along with the identification of secondary precipitated phases. Thermodynamic simulations were employed to analyze the equilibrium solidification process, and the predictive accuracy of various empirical models for solidification mode and ferrite content was evaluated. The results indicate that the ferrite morphology varies from granular near the surface to short rod-like and finally to a semi-network morphology toward the center, predominantly along austenite grain boundaries. The ferrite content (volume fraction)ranges from 0.14% to 1.47%, with the highest value observed at the center, meeting the non-magnetic criteria. The main secondary precipitates were identified as chi-phase and sigma-phase. Near the surface, ferrite precipitates as discrete particles at austenite grain boundaries, with partial transformation into chi-phase. The high cooling rate at the surface inhibits complete transformation, resulting in a coupled ferrite/chi-phase microstructure. In contrast, the semi-network ferrite in the central region has largely transformed into sigma-phase. Although thermodynamic calculations predict single-phase austenitic solidification (A-mode), experimental observations confirm austenite-ferrite (AF-mode) solidification. Among five chromium/nickel equivalent formulas and the WRC-1992 diagram evaluated, only the Hull formula accurately predicts the ferrite content (1%-2%), consistent with measured results. Other models either overpredict (3% or 7%-10%) or incorrectly predict fully austenitic solidification.

Nb-containing non-quenched and tempered cold-heading steel MFT9 exhibits high strength, excellent impact resistance, and low production cost, making it a promising candidate for automotive fasteners. However, during the continuous casting of MFT9 blooms, both surface transverse cracks and central porosity defects occur. To address these issues, this study systematically investigates the precipitation behavior of Nb-rich particles in MFT9 steel, which are prone to induce transverse cracking in the bloom. A mathematical model was also applied to optimize the continuous casting process parameters. The results show that the large precipitates formed during casting are primarily NbC. Notably, when the temperature drops below 1 000 ℃, the number of particles larger than 120 nm increases significantly, aggravating the formation of surface transverse cracks. Hence, it is crucial to maintain the bloom surface temperature above 1 000 ℃. Numerical simulation results indicate that, under the existing equipment conditions, the casting speed should be controlled within 1.3-1.5 m/min, the specific water flow rate should be maintained between 0.8-1.0 L/kg, and casting with low superheat is recommended.

To optimize the solidification uniformity of 304 austenitic stainless steel continuous casting slabs, this study employs numerical simulation to investigate the influence of nozzle arrangement patterns in the secondary cooling zone on the solidification process. By establishing a simulation model aligned with actual production parameters, the solidification behavior of slabs under four distinct nozzle arrangement schemes was systematically compared. The results demonstrate that different nozzle arrangements significantly alter the cooling intensity at various surface positions of the slab, thereby affecting internal temperature gradient distribution and the progression of the solidification front. As the nozzle arrangement schemes were progressively optimized, the solidification uniformity of the slab markedly improved. Through comparative analysis of stepwise optimization schemes, this study quantitatively elucidates the regulatory mechanisms of secondary cooling nozzle arrangements on solidification uniformity. It provides a clear optimization pathway and theoretical foundation for enhancing the internal quality of 304 stainless steel slabs by improving solidification uniformity.

High-strength steel is widely used in automotive frames and shipbuilding plates, among other applications. The presence of irregular and large-sized inclusions is one of the primary factors detrimental to its performance. Due to their advantageous properties, rare earth elements are commonly employed to modify inclusions in steel. To investigate the underlying modification mechanisms, this study examines the influence of rare earth Y on the hot ductility and austenite grain size of steel through experimental observations. Additionally, the evolution of non-metallic inclusions under varying Y contents was analyzed using thermodynamic calculations. The results indicate that the addition of rare earth Y significantly enhances the hot ductility of steel. The reduction of area (Z) increases by 21.9%, representing a 94.0% improvement. The fracture form of the steel changes from brittle fracture to plastic fracture, and the austenite grain size decreases by 47%. Rare earth Y promotes the transformation of MnS, Al₂O₃, and their complex inclusions into rare earth-containing inclusions such as Y₂O₃, Y₂O₂S, YS, and Y₂S₃. This modification reduces the aspect ratio of the inclusions by 19%. Al₂O₃ inclusions are almost entirely modified, and the formation of MnS inclusions is reduced by 64%. The evolution mechanism of inclusions proceeds as follows: in the absence of rare earth, the steel contains MnS, Al₂O₃, and (MnS+Al₂O₃) inclusions. With the addition of 0.015 mass% rare earth Y, rare earth inclusions form preferentially, serving as nucleation sites for MnS and Al₂O₃, which are subsequently fully modified into rare earth inclusions.

To investigate the fluctuation behavior of the steel-slag interface in a wide-thick slab continuous casting mold, a three-dimensional mold model was established. The computational differences between the Realizable k-ε model and the Large Eddy Simulation (LES) model in simulating the steel-slag interface behavior were compared and validated against water model experiments. The results indicate that the LES model captures transient flow structures and interface fluctuations more accurately than the Realizable k-ε model, showing better agreement with experimental observations. As the interfacial tension increases from 1.0 N/m to 1.4 N/m, the flow velocity of molten steel at the interface first increases and then decreases, with the most pronounced variation occurring at the quarter-width location near the narrow face. The interface fluctuation is minimized at an interfacial tension of 1.4 N/m.

To address the issue that breakout prediction models often struggle to achieve high accuracy under conditions of small sample size, nonlinearity, and high-dimensional training data, this paper proposes a breakout prediction algorithm for continuous casting based on PCA-PSO-SVM. The method first employs Principal Component Analysis (PCA) to reduce the dimensionality of collected continuous casting process parameters, extracting comprehensive indicators to alleviate data complexity. It further integrates Particle Swarm Optimization (PSO) to optimize the parameters of the Support Vector Machine (SVM), thereby mitigating the issues of strong parameter dependency and limited generalization capability typical of traditional SVM in industrial data classification. Experimental tests using real-world data from a steel plant demonstrate that the proposed algorithm reduces computational load while enhancing prediction accuracy, achieving a breakout prediction rate of up to 95.56%.

The novel fluorine-free CaO-Al₂O₃-TiO₂-based mold flux is designed to suppress interfacial reactions between high-titanium steel and the flux, ensuring stable flux performance during continuous casting. The physicochemical properties of the mold flux directly influence the surface quality of the cast strand and the smooth operation of the casting process, with lubrication behavior being a critical aspect of its metallurgical function. Key indicators such as melting temperature, viscosity, and breaking temperature are essential in characterizing lubrication performance; however, their experimental determination is often time-consuming and costly. To address this, the melting temperature, viscosity, and breaking temperature of the mold flux were systematically measured and analyzed. Based on a comprehensive dataset, predictive models for these parameters were developed. The findings reveal that Na₂O and B₂O₃ lower the melting temperature, while the w(CaO)/w(Al₂O₃) ratio and SiO₂ raise it. In contrast, TiO₂, BaO, Li₂O, and MgO exhibit dual effects. The w(CaO)/w(Al₂O₃) ratio, BaO, Na₂O, B₂O₃, Li₂O, and MgO reduce viscosity, whereas SiO₂ increases it. TiO₂ again shows a dual influence on viscosity. BaO, Na₂O, B₂O₃, and SiO₂ decrease the breaking temperature, while the w(CaO)/w(Al₂O₃) ratio, TiO₂, and MgO increase it; Li₂O presents a dual effect. The predictive models for melting temperature, viscosity, and breaking temperature exhibit mean absolute percentage errors of 1.60%, 8.43%, and 0.63%, respectively, demonstrating their strong predictive capability for the lubrication behavior of CaO-Al₂O₃-TiO₂-based mold fluxes. These models offer an intuitive reference and technical support for the development of mold fluxes used in high-titanium steel continuous casting.

High carbon high silicon wear-resistant ball steel B3 with ø125 mm cracked during the ball drop test and in service. The cause was analyzed by using scanning electron microscopy and metallographic examination, and main reason is center residual shrinkage cavity and network carbide in large bar due to incomplete welding of severe center shrinkage cavity and center segregation in bloom during the hot rolling process. Combined with numerical simulation and industrial experiments, the continuous casting process was optimized to improve center shrinkage cavity, center porosity and center segregation in 390 mm×510 mm cross section bloom for improving the central density and reducing network carbide level with ø125 mm bar. The results show process optimization strategy with lower superheat degree, secondary cooling with smaller water ratio, M-EMS and F-EMS with higher stirring intensity, soft reduction with large reduction amount can significantly improve the center quality of high carbon high silicon wear-resistant ball steel B3 bloom. Casting speed is 0.43 m/min. Stirring current and frequency of M-EMS and F-EMS was 450 A/1.5 Hz and 600 A/8.0 Hz respectively. Soft reduction is implemented in No.3- No.6 withdrawal and straightening unit and effective reduction amount is 16 mm. After the above optimized process is applied in the industrial optimization experiment, center carbon segregation index in bloom is below 1.07 and center shrinkage cavity is below 0.5 level, thus center compactness of the hot-rolled ø125 mm large bar is significantly improved and flaw detection eligibility rate reaches above 98%. Eventually, crack defect of high carbon high silicon wear-resistant steel balls have been resolved.

In high-titanium steels, [Ti] is highly prone to react with SiO₂ and Al₂O₃ in the tundish flux, resulting in deteriorated metallurgical properties, imprecise composition control, and reduced steel cleanliness. To address these issues, a systematic study was carried out on the physical properties of the flux, its reactivity with molten steel and refractory materials, as well as its stability after absorbing inclusions. Based on the findings, a low-reactivity, high-stability tundish flux system based on (BaO+CaO)-Al₂O₃-TiO₂ was developed and successfully applied in the continuous casting of high-titanium steel containing 0.3 mass%-0.4 mass% of Ti. During the industrial trials, the total oxygen content (T[O]) in the tundish was controlled within 12×10^-6-16×10^-6, the nitrogen content ([N]) varied by no more than 2×10^-6, and the titanium loss rate was only 0.52%, ensuring smooth operation of the continuous casting process.

Tundish flux plays crucial roles in the continuous casting process, including thermal insulation, isolation of air, and purification of molten steel. It′s purification function becomes particularly important when casting high-cleanliness steels. In this study, a calcium aluminate-based tundish flux with high basicity was investigated. Thermodynamic calculations were performed using Factsage to provide theoretical guidance for the composition design and optimization of such fluxes, aiming to enhance their steel-purifying performance. The results demonstrate that, to effectively suppress the corrosion of tundish dry refractories, the MgO content in the flux should be no less than 6 wt.%. The C/A ratio (mass ratio of CaO to Al₂O₃) should be increased as much as possible—without compromising the melting behavior—to improve the flux's capacity for dissolving and absorbing inclusions. Increasing the temperature helps reduce the viscosity of the liquid slag, thereby facilitating the dissolution and adsorption of inclusions.

A three-dimensional coupled model integrating fluid flow, heat transfer, and solidification was developed to systematically investigate the evolution of center solidification in high-strength 40Mn steel slabs. Industrial trials were conducted to further explore the influence mechanism of roll gap contraction on center segregation. The results indicate that the double-roll flow pattern inside the mold leads to non-uniform shell growth across the wide face of the slab, with a thicker central region and thinner sides. This non-uniformity in initial solidification is further exacerbated during the secondary cooling process, ultimately resulting in a “W”-shaped distribution of the solidification endpoint. The application of roll gap contraction technology reduced the center segregation level from B1.0 to C1.0, demonstrating a noticeable improvement. However, the “W”-shaped solidification profile attenuates the effectiveness of roll gap contraction in different zones, leading to variations in the degree of center segregation along the centerline. To enhance the consistency of the center solidification behavior, it is recommended to optimize the flow field structure in the mold and rationally adjust the nozzle layout and spray angles in the secondary cooling zone. These measures would promote more stable and uniform solidification, thereby improving the internal quality of the slab.

To address transverse cracks observed at the upper surface edge (100-200 mm from the edge) of rolled extra-thick plates made of pressure vessel steel 13MnNiMoR at a domestic steel plant, the formation mechanism and countermeasures for such edge cracks in continuous casting slabs were investigated through microscopic examination and hot ductility simulation. The results reveal that the matrix microstructure near the edge cracks consists of coarse grains, with a pro-eutectoid ferrite film—ranging in thickness from several tens to hundreds of micrometers—present along the prior austenite grain boundaries. Cracks initiate and propagate within this ferrite film, with their roots embedded in an un-cracked ferrite zone. Significant decarburization is observed around the cracks, while no precipitated phases are detected. Under the conventional secondary cooling conditions of the current slab casting process, the upper surface edge region (100-200 mm from the edge) experiences severe local overcooling in the straightening section, where the surface temperature drops to about 850 ℃—below the lower limit of the measured ductile-to-brittle transition temperature (DBTT) range of 880-890 ℃. The presence of pro-eutectoid ferrite films along austenite grain boundaries significantly reduces the material's plasticity. In the straightening segment, this low-temperature brittle zone is subjected to both high stress concentration (exceeding the material's strength limit) and additional tensile strain due to equipment misalignment (e.g., roll gap deviation or poor arc alignment). These combined stresses act on the brittle ferrite film, leading to intergranular cracking. It is concluded that the transverse cracks in 13MnNiMoR extra-thick slabs primarily result from the combined effects of casting machine precision, grain boundary strength, and cooling conditions, which extend during rolling into large-scale defects. Based on the DBTT range (880-890 ℃) identified via hot ductility tests and insights from a fully coupled thermal-mechanical model—which highlighted issues such as low temperature, high stress concentration, and additional strain caused by equipment inaccuracy in the straightening section—adjustments were made to the casting machine alignment and cooling process. A new controlled cooling strategy was implemented in production, along with tighter control of nitrogen and aluminum content during steelmaking. These measures significantly improved the crack resistance of the slab edges, reducing the incidence of edge cracks in 13MnNiMoR slabs from 15.39% to 1.67%.

A three-dimensional Euler-Lagrangian numerical model was established based on a six-strand tundish to investigate the influence of the ladle change process on the escape behavior of inclusions. The results indicate that during ladle change casting, the escape rate of inclusions is the lowest in the emptying stage and reaches its maximum in the filling stage, due to changes in the flow behavior of the molten steel. Under the same initial tundish weight, a higher filling speed leads to a higher maximum escape rate of inclusions. When the filling speed increased from 4.70 t/min to 11.75 t/min, the maximum escape rate of 100 μm inclusions increased by 2.45 times. At the same filling speed, a lower initial tundish weight results in a higher maximum escape rate of inclusions. When the initial tundish weight was reduced from 40 t to 30 t, the maximum escape rate of 10 μm inclusions showed no significant change, while that of 50 μm inclusions increased from 0.115 to 0.132 (a 14.8% increase), and that of 100 μm inclusions rose from 0.029 to 0.047 (a 62% increase). In production practice, the filling speed during the ladle change process was optimized, and an automatic control system for tundish weight was developed. As a result, the proportion of off-spec wire rods due to inclusion size was reduced from 7.7% to 2.5%.

To investigate the flow behavior in the mold during continuous casting of 230 mm thick slabs under current operating conditions and to identify the causes of minor longitudinal surface cracks, a full-scale physical model of the mold was developed. This study systematically examined the effects of key process parameters—including submerged entry nozzle (SEN) geometry, casting speed, and immersion depth—on the fluid flow and meniscus fluctuation within the mold. The results demonstrate that SEN dimensions significantly influence the flow pattern: increasing both the port size and the inner bore diameter reduces the kinetic energy of the exiting jet, decreases the velocity of the upper recirculation zone, and effectively suppresses level fluctuations. Excessive meniscus fluctuation was identified as the primary factor leading to slab cracking. Based on comprehensive flow field analysis, the optimal SEN configuration for current operational conditions was determined to feature ports sized 66 mm×90 mm and an inner bore diameter of 80 mm.

Taking a domestically produced HSLA steel as the research object, this study applied a dual-process approach combining narrow-window refining process control with constant casting speed and accurate rare-earth wire feeding into the mold. The recovery rate, content, and distribution of rare earth in the slab were systematically investigated. Through comprehensive sampling and comparative testing, the effects of rare earth addition on the internal cleanliness, segregation behavior, and solidification structure of the slab were evaluated. The results show that the refining-continuous casting dual process enables a rare earth recovery rate exceeding 80%, with the residual rare earth content in the steel maintained above 200×10^-6, uniformly distributed across different slab locations. This method significantly reduced both the average total oxygen content and its fluctuation in the slab, indicating improved steel cleanliness. After adding a high concentration of rare earth, conventional inclusions such as Al₂O₃, MgO and MnS were modified into rare earth-containing inclusions (e.g., RE-O, RE-S, RE-Al-O, RE-S-O). The dominant inclusion size range shifted from above 10 μm to below 5 μm, and the morphology changed from elongated or chain-like to spherical or spindle-like. Compared with the rare earth-free slab, the number density of inclusions increased in the rare earth-treated slab, while the area density decreased, indicating that rare earth modifies, refines, and disperses inclusions, effectively reducing large heterogeneous inclusions in the steel matrix and enhancing microstructural continuity. The uniform distribution of rare earth also notably suppressed the development of columnar crystals, increased the equiaxed crystal ratio, and refined the dendrite arm spacing. This refined and homogeneous solidification structure alleviated solute element enrichment toward the slab center, reducing segregation and cracking tendency. These improvements help minimize banded structure formation during rolling and enhance the strength, toughness, and fatigue resistance of the steel through grain refinement.

This study systematically investigates the formation mechanisms of non-metallic inclusions in 18CrNiMo7-6 steel used for wind power gearbox components. A multi-technique approach combining in-situ monitoring, macrostructure observation, water-immersion ultrasonic testing, industrial computed tomography (CT), and scanning electron microscopy (SEM) with energy dispersive spectroscopy (EDS) was employed to comprehensively characterize the morphology, chemical composition, and spatial distribution of inclusions. The results demonstrate that by optimizing continuous casting parameters—including precise control of pouring temperature, improved cooling water distribution, and application of electromagnetic stirring—the macrosegregation of carbon was significantly mitigated, with its variation range reduced to 0.033%. This optimization notably enhanced the homogeneity of material properties. Exogenous inclusions were primarily attributed to refractory erosion or slag entrainment, among which MgAlO-type inclusions played a critical role in the formation of large composite inclusions and macroscopic defects. In contrast, endogenous inclusions mainly originated from deoxidation reactions, solidification segregation, and secondary oxidation processes.