15 March 2025, Volume 35 Issue 3

    

• Select all
  |

Expert Forum
• Select
  Research on low-carbon development strategies in steel industry:Review and prospect

  YIN Ruiyu, SHANGGUAN Fangqin, CUI Zhifeng

  China Metallurgy. 2025, 35(3): 1-15. https://doi.org/10.13228/j.boyuan.issn1006-9356.20250058

  Abstract ( ) XML Related Articles File   Knowledge map   Save

  The changes of China's annual crude steel output, average daily crude steel production on a monthly basis over the past two years, end-of-month rebar prices, direct steel exports, indirect exports of steel products, and CO₂ emissions were reviewed and analyzed. It is concluded that China's crude steel production is generally in a situation of oversupply, and the steel industry has entered a downward phase of reduction fluctuations. Under the premise of no significant increase in crude steel output, China's steel industry can be considered to have entered a CO₂ emission stabilization period. The study explores targets and measures for total crude steel output control from perspectives including future production forecasts, supply-side structural reforms, and adjustments to import/export policies. By integrating projections of future crude steel output and scrap resources, it is proposed that under the goals of "Carbon peak and Carbon neutrality", China's steel industry will gradually develop three typical manufacturing processes:BF-BOF long process, full scrap EAF process, and hydrogen reduction-EAF process. The evolutionary alternation of these three processes is discussed, emphasizing that China's steel industry should leverage the "Carbon peak and Carbon neutrality" context to guide scrap resources toward EAF process, thereby gradually optimizing the sector's ferrous resource structure, product structure, and process structure. Pathways for enhancing manufacturing continuity are explored through interface technology optimization, dynamic precision design, and full-process intelligentization. Finally, through the construction and analysis of a "dual carbon" analytical model for the steel industry, the study identifies controlling and reducing crude steel output as the most effective decarbonization measure, with process structure optimization in steel plants being equally critical.
Review
• Select
  Progress in GH4738 alloy development

  CUI Tianliang, HOU Weixue, YANG Chengbin, WANG Guanglei, TANG Chao, QU Jinglong, DU Jinhui

  China Metallurgy. 2025, 35(3): 16-33,112. https://doi.org/10.13228/j.boyuan.issn1006-9356.20240649

  Abstract ( ) XML Related Articles File   Knowledge map   Save

  GH4738 alloy is a high-performance nickel-based superalloy. It plays a key role in aviation, aerospace and petrochemical industries due to its excellent high temperature strength and excellent creep and fatigue properties. The basic research progress of GH4738 alloy in the past half century is reviewed (from 1973 to present), including its composition, smelting process, homogenization treatment, hot deformation behavior, heat treatment and development trend. The innovative achievements of a series of engineering preparation technologies, such as the synergistic increase of Al and Ti content, the substitution of W for Mo to improve the comprehensive performance, the development of a triple smelting process and the control technology of inclusions and harmful elements, the establishment of a thermal deformation constitutive equation and recrystallization grain structure evolution model, and the application of finite element numerical simulation in the microstructure prediction of large-size complex forgings, are described, and the future development is prospected based on the current research situation. The aim is to meet the urgent needs of high performance, high reliability and long life of the new generation of aero-engine materials and to promote the further promotion of GH4738 alloy.
• Select
  Current situation and research progress of recycling process for scrap copper in China

  WANG Zhanhao, LIU Shuo, YANG Siyuan, BAO Shenxu

  China Metallurgy. 2025, 35(3): 34-43,75. https://doi.org/10.13228/j.boyuan.issn1006-9356.20240686

  Abstract ( ) XML Related Articles File   Knowledge map   Save

  With the increasing demand for domestic copper resources, there is a large gap between the production and consumption of refined copper, and the recycling of scrap copper has become a necessary way to solve the shortage of copper resources. Scrap copper can be divided into three categories according to copper grade, namely high, medium and low grade, in which the high-grade scrap copper can be recycled by direct utilization or one-stage process, while the medium and low grade scrap coppers are mainly recycled by two-stage process, flotation and hydrometallurgical process. The recycling process of different types of scrap copper was discussed in detail, and the principles, advantages and disadvantages of various processes were summarized by analyzing the implementation cases of pyrometallurgy, hydrometallurgy and flotation methods for scrap copper. Based on the current situation of copper scrap recycling in China, improving the classification standards and recycling system of copper scrap can effectively enhance the comprehensive utilization of scrap copper resources.
Special Contribution of Metallurgy Talent
• Select
  Research progress of direct reduction of vanadium titanium magnetite and resource utilization of titanium-bearing slag

  LI Lanjie, ZHANG Caidong, LIU Yang, ZHAO Suxing, ZHAO Jianqi, XU Kun

  China Metallurgy. 2025, 35(3): 44-54,64. https://doi.org/10.13228/j.boyuan.issn1006-9356.20240790

  Abstract ( ) XML Related Articles File   Knowledge map   Save

  Vanadium titanium magnetite is a featured strategic metal resource in China. The existing mainstream of vanadium extraction at home and abroad is the blast furnace ironmaking-converter process, which has high carbon emissions and low recovery rate of vanadium and titanium. A large amount of titanium element enters blast furnace slag and accumulates over 100 million tons, which not only wastes resources, but also significantly increases the pressure of environmental protection. The existing non-blast furnace smelting processes to make up for the shortcomings of the blast furnace process are introduced, including direct reduction process and smelting reduction process, and the process flow and application status are summarized. The principles, application status, advantages and disadvantages of titanium extraction methods of titanium-containing slag at home and abroad, such as modified enrichment, carbonization-chlorination and hydrometallurgical extraction, are summarized. In addition, the future development for comprehensive utilization of vanadium titanium magnetite is prospected. It is considered that the full-process technology of hydrogen-based shaft furnace reduction-electric furnace melting separation-comprehensive utilization of titanium slag for vanadium titanium magnetite holds tremendous development prospects. This technology can strongly support China's vanadium titanium steel industry in achieving the "dual carbon" goals.
Monographic Study
• Select
  Spatiotemporal distribution of multicomponent burden segregation in charging process of a blast furnace

  TENG Zhaojie, CHEN Jian, YANG Yizhang, LI Chenrui, WU Jianlong, WU Shengli, XU Jian

  China Metallurgy. 2025, 35(3): 55-64. https://doi.org/10.13228/j.boyuan.issn1006-9356.20240697

  Abstract ( ) XML Related Articles File   Knowledge map   Save

  The eccentric arrangement of hoppers in the top of a bell-less blast furnace with parallel hoppers leads to the segregation of burden particles during the charging process. Although the quantitative characterization and formation mechanisms of this segregation have been systematically studied, the changes in segregation characteristics over time and space during the continuous movement of burden particles are rarely reported. Therefore, firstly, a 3D mathematical model of the charging system for a large bell-less blast furnace with parallel hoppers was established. Then, the discrete element method (DEM) was used to simulate and quantitatively analyze the mass distribution of multicomponent burden particles sequentially at the conveyor belt, hopper, and furnace throat. Subsequently, the spatiotemporal distribution of the segregation characteristics during the charging process was summarized. The results show that the segregation of burden particles within the hopper is mainly influenced by the layering on the conveyor belt and the charging sequence. Burden charged earlier tends to accumulate at the bottom of the hopper, while burden charged later predominantly occupies the top. Radially within the hopper, burden in the lower layers of the conveyor belt tends to concentrate in the central region of the hopper, while burden in the upper layers tends to accumulate near the hopper walls. At the furnace throat, the segregation of burden particles is mainly influenced by the charging operation rather than directly by the layering and timing on the conveyor belt. However, due to the segregation of burden particles within the hopper and the differences in discharge sequences from various regions, significant radial differences in the mixing uniformity distribution of burden at the furnace throat are observed. Therefore, the sequence and structure of the burden on the conveyor belt can affect the mass distribution of the charge at the throat by determining the segregation inside the hopper, thereby improving the charge structure and improving the overall performance of the blast furnace, which provides a new optimization direction for blast furnace operations.
• Select
  Thermodynamic simulation of gas generation reactions containing F, K, and Na in blast furnaces

  CAO Linxiang, LUO Guoping, WANG Ziwei, SUN Quan, BAI Xiaoguang, LIU Shuguang, LIU Jingquan

  China Metallurgy. 2025, 35(3): 65-75. https://doi.org/10.13228/j.boyuan.issn1006-9356.20240647

  Abstract ( ) XML Related Articles File   Knowledge map   Save

  In order to reveal the formation rules of fluorine, potassium and sodium gases in the process of blast furnace smelting, the No.7 blast furnace at Baotou Steel Group was taken as an example, and the blast furnace was segmented into four distinct thermal regions along its vertical axis. According to the main chemical reactions in different regions, the mixed charge composition and gas composition at any temperature along the height of blast furnace were obtained by using the difference calculation method through the amount of raw material and fuel, the composition of raw material and fuel and the composition of hearth gas. The gas pressure loss in different temperature regions and the gas pressure at any temperature of blast furnace were determined by combining the Ergun equation, the mixed ore droplet experiment with the hot blast pressure of blast furnace and the top gas pressure of blast furnace. Based on the composition of mixed burden, gas composition, gas pressure and reaction temperature at any temperature of blast furnace, FactSage7.1 thermodynamic software was used to simulate the formation reactions of fluorine (F), potassium (K) and sodium (Na) gases in different temperature regions of blast furnace. The simulation results show that in the range of 200-800℃, the production of F, K and Na gases is less, mainly HF and a small amount of KF. In the range of 800-1 025℃, the amount of gas containing F, K and Na increases. In the soft melt zone of 1 025-1 365℃, the maximum production of KCN, NaCN and KF increases significantly, reaching 1.74, 1.06 and 1.08 kg/t, respectively. In the drip zone of 1 365-1 500℃, the amount of alkali metal gases NaCN, KCN, K and Na is the largest. This study can lay a foundation for further understanding the influence of cyclic enrichment of fluorine, potassium and sodium gases on the properties of ore, coke and lining.
• Select
  Energy-mass transport model of smelting process in iron bath smelting reduction vessel

  JIA Lijun, QIN Deyue, JIANG Shan, LI Baokuan

  China Metallurgy. 2025, 35(3): 76-85,122. https://doi.org/10.13228/j.boyuan.issn1006-9356.20240612

  Abstract ( ) XML Related Articles File   Knowledge map   Save

  HIsmelt is a representative process in smelting reduction ironmaking. The elimination of pelletizing, sintering and coking processes to reduce pollutant emissions is a prominent feature of the process. Based on the material balance and energy balance, the energy-mass transport model in iron bath smelting reduction vessel was established. In the model, the effects of water gas reaction on mass transport and energy transport were emphatically considered. The effects of reaction with or without water gas on the composition of molten iron, iron slag and flue gas were compared. The effects of the water-gas reaction, both with and without water gas, on the compositions of molten iron, slag, and flue gas were analyzed. Compared to actual production data, the model's prediction accuracy is improved. In addition, the energy-mass transfer model demonstrates greater rationality compared to the model that excludes the coal-water gas reaction. The model can calculate the production of molten iron, the production of iron slag, the volume of flue gas, the composition of molten iron, the composition of iron slag and the composition of flue gas. The results of the model calculations are in agreement with the actual production data. On this basis, it is concluded that with the increase of molten iron temperature, the production of hot metal, the content of carbon and sulfur in molten iron, and the content of manganese in iron slag gradually decrease, while the content of ferrous oxide in iron slag gradually increases. However, the production of iron slag demonstrates a fluctuating pattern, reaching a minimum at 1 748 K, and yielding 0.47 t of iron slag per ton of molten iron. The development of an energy-mass transport model serves as a robust foundation for accurate production forecasting.
• Select
  Theoretical analysis and practice on improving ladle free-opening rate

  ZHOU Yu, JU Yinjun, ZHANG Tianshu, WANG Rongkun, LI Wanming

  China Metallurgy. 2025, 35(3): 86-93. https://doi.org/10.13228/j.boyuan.issn1006-9356.20240564

  Abstract ( ) XML Related Articles File   Knowledge map   Save

  When the ladle cannot open automatically,an oxygen burning operates on the sintered layer generates a large amount of oxidation products that enter the tundish,causing secondary oxidation of the molten steel. Therefore,improving the ladle free-opening rate is of great importance for enhancing the purity of molten steel. The influence of the area and thickness of the filler sand sintered layer, the type of filler sand, the temperature of the molten steel, and the holding time in the ladle on the ladle free-opening rate is investigated. The mechanisms by which the area and thickness of the sintered layer affect the ladle free-opening rate were theoretically analyzed. Research shows that for every 1% increase in the radius of the sintered layer,the stress it experiences increases by 1.2%,making the sintered layer more prone to cracking and resulting in a higher free-opening rate. When the temperature of the molten steel is less than 1 630℃ and the holding time is less than 180 minutes,the free-opening rate is higher when using chrome-based filler sand. Conversely,when the temperature of the molten steel is above 1 650℃ and the holding time exceeds 180 minutes,the free-opening rate is higher when using zirconium-based filler sand. The main factors influencing the free-opening rate, listed in order of importance, are the area of the sintered layer, followed by the temperature of the molten steel and the holding time, and finally the type of filler sand. After implementing improvements such as increasing the area of the sintered layer and using different types of filler sand for different steel grades,the ladle free-opening rate may increase to 99.7%. The research results are of great significance for improving the theoretical basis of free-opening rate and increasing the production efficiency and product quality of steel enterprises. It also provides reference value for the ladle free-opening rate of high alloy, high purity steel grades with extended smelting processes, such as pipiline steel and bearing steel.
• Select
  Effect of CaO-Al₂O₃-MgO-SiO₂-CaF₂ refining slag on desulfurization behavior of high-manganese steel

  YAN Shihao, WANG Zhitong, WANG Weijian, ZHANG Xuewei

  China Metallurgy. 2025, 35(3): 94-102. https://doi.org/10.13228/j.boyuan.issn1006-9356.20240660

  Abstract ( ) XML Related Articles File   Knowledge map   Save

  To study the desulfurization behavior of CaO-Al₂O₃-MgO-SiO₂-CaF₂ slag in ladle refining process for high-manganese steel,the effect of CaO/Al₂O₃ mass ratio in CaO-Al₂O₃-5%MgO-5%SiO₂-10%CaF₂(mass fraction) refining slag on the desulfurization efficiency and the rate controlling step of the desulfurization of ZGMn13 high manganese steel at 1 873 K was investigated. The highest desulfurization efficiency of 97.9% achieved when the CaO/Al₂O₃ mass ratio in the slag was 1.8,while desulfurization effect was the worst when the CaO/Al₂O₃ mass ratio was 0.2. The mass transfer coefficient in the slag k_s was calculated as between 1.38×10^-7 m/s and 1.99×10^-6 m/s,and the mass transfer coefficient in the steel k_m was 7.85×10^-5 m/s. In addition,the overall mass transfer coefficient k_o of sulfur in high-manganese steel increases with the increase of CaO/Al₂O₃ mass ratio in the slag and gradually levels off. The change in ko is not significant when the CaO/Al₂O₃ mass ratio in the slag is between 1.2 and 1.8(the change is 8.7×10-6 m/s). This shows that the rate controlling step of desulfurization process was the transfer of sulfur in the slag when the CaO/Al₂O₃ mass ratio was lower than 1.2,while it changed to the mass transfer of sulfur in steel when the desulfurization reaction of the CaO/Al₂O₃ mass ratio was between 1.2 and 1.8. Compared to low alloy steels,the diffusion of sulfur in the steel was reduced due to the higher content of manganese in the steel.
• Select
  Prediction of hot-rolled strip plate crown based on PCA-DWT-SVR

  XIA Zhihui, LI Weigang, MAO Yunfei

  China Metallurgy. 2025, 35(3): 103-112. https://doi.org/10.13228/j.boyuan.issn1006-9356.20240641

  Abstract ( ) XML Related Articles File   Knowledge map   Save

  Convexity of hot-continuous-rolled strip steel plates is an important indicator for assessing the quality of plate shape,which is characterized by multivariate,nonlinear and hereditary properties. In order to improve the prediction accuracy and modeling speed of the convexity of strip steel plate,there is an urgent need to improve the method. Aplate convexity prediction model by Support Vector Regression (SVR) combining Discrete Wavelet Transform (DWT) and Principal Component Analysis (PCA) is developed. Firstly,the outliers are removed by Pauta's criterion,and then the data are smoothed by discrete wavelet transform,and the dimensionality is reduced by Principal Component Analysis (PCA),and the important variables are selected by using the load matrix,and the main variables are screened out by combining the variables that have significant influence on the convexity of the plate in the rolling process;and The processed data is then fed into the Support Vector Regression (SVR) model to predict the convexity of the plate through the optimization of the lattice parameter. the data is entered into the support vector machine regression model to generate predictions after optimizing the grid parameters to predict the plate convexity. Finally,four different regression models are constructed,including K-nearest neighbor (KNN) regression model,gradient boosted decision tree (GBDT) regression model,extreme gradient boosted (XGBoost) regression model,and random forest (RF) regression model,and they are compared. The experimental results show that the PCA-DWT-SVR model reduces the data to 15 dimensions,which greatly reduces the prediction time,and at the same time,the evaluation indexes of this model,The metrics include the root mean square error (E_RMS),the mean absolute error (E_MA),and the correlation coefficient (R) are 3.26,2.11,and 0.921,respectively,which are optimal among the models,and the absolute prediction errors for more than 98% of the sample points are within 10 μm and the absolute prediction error is within 10 μm for more than 98% of the sample points. The model not only achieves high accuracy prediction of plate convexity,but also significantly shortens the modeling and prediction time,which provides an important reference for the quality control of plate convexity and related research.
• Select
  Effect of aging process on microstructure and mechanical property of Cu age-hardening steel

  ZHANG Zhengyan, WANG Sencai, CHAI Feng, LUO Xiaobing, CHAI Xiyang

  China Metallurgy. 2025, 35(3): 113-122. https://doi.org/10.13228/j.boyuan.issn1006-9356.20240617

  Abstract ( ) XML Related Articles File   Knowledge map   Save

  To explore the aging characteristics of Cu age hardened steel,master the aging process that matches the best strength and toughness,and guide production practice,the microstructure,impact fracture morphology and nanoscale Cu-rich precipitates of tested steels with different thermal history were characterized using scanning electron microscopy (SEM),high-resolution transmission electron microscopy (HRTEM) and atomic chromatography (APT). And the precipitation kinetics of Cu-rich precipitates was calculated using Thermo-Calc calculation software. The high strength toughness matching of the tested steel under different thermal history states was also investigated and analyzed. The results show that the microstructure of the hot-rolled tested steel is granular bainite,containing a large number of large-sized M/A islands,resulting in poor strength and low-temperature toughness;after quenching,the mixed microstructure of bainite and martensite can be obtained,exhibiting good matching of strength and toughness,but poor plasticity;after quenching and aging heat treatment,the strength of the tested steel first increases and then decreases with the increase of aging temperature. After quenching and aging heat treatment,the strength of the tested steel first increases and then decreases with the increase of aging temperature. When the aging temperature is below 450℃,it is in an under aged state with lower strength.When the aging temperature reaches 450℃,the strength reaches its peak,but at this time,the low-temperature toughness is the lowest.When the temperature exceeds 450℃,the strength decreases and the toughness increases. When the temperature reaches 650℃,both the strength and toughness decrease.With the increase of aging time at 620℃,the size of substructure and lath as well as the size of Cu-rich precipitates in steel show increasing trend,resulting in a gradual decrease in the strength and and a gradual increase in toughness of the tested steel,and it is find that excellent strength toughness matching can be achieved under the condition of aging at 620℃ for 120 minutes. Compared to the aging time,the aging temperature has a more significant impact on the properties of steel.
• Select
  Hot deformation behavior of GH99 alloy based on Arrhennius, Johnson-Cook,and Fields-Backofen constitutive models

  WANG Wenjingzi, MA Xin, PENG Yuhan, JIANG Xueyu, ZHOU Ge, ZHANG Haoyu, ZHANG Siqian, CHEN Lijia

  China Metallurgy. 2025, 35(3): 123-134. https://doi.org/10.13228/j.boyuan.issn1006-9356.20240530

  Abstract ( ) XML Related Articles File   Knowledge map   Save

  To investigate the hot deformation behavior of the GH99 alloy under various conditions,the hot compression simulation tests were conducted. These tests covered a deformation temperature range from 1 020℃ to 1 170℃ and a strain rate range from 0.001 to 1 s^-1,based on the obtained stress and strain data,Arrhenius,modified Johnson-Cook,and Fields-Backofen constitutive models were established to comprehensively predict EBSD characterized the rheological stress values of the alloy and the deformation microstructure. The results showed that the Arrhenius constitutive model predicted stress values slightly deviated from experimental values under medium and low-temperature conditions compared to high-temperature conditions. The modified Johnson-Cook constitutive model can effectively predict the rheological stress values of alloys under medium-temperature conditions. Under low-temperature conditions,the Fields-Backofen constitutive model has high prediction accuracy. Dynamic recrystallization is the primary softening mechanism of GH99 alloy. As the deformation temperature increases and the strain rate decreases,the degree of dynamic recrystallization and grain size gradually increase.
Energy Saving and Environmental Protection
• Select
  Life cycle carbon footprint assessment of automotive IF steel products

  LI Weidong, FANG Xiaoqing, MA Guangyu, CAI Qiuye, ZHANG Yuhang, SUN Wenqiang

  China Metallurgy. 2025, 35(3): 135-143,154. https://doi.org/10.13228/j.boyuan.issn1006-9356.20240632

  Abstract ( ) XML Related Articles File   Knowledge map   Save

  To meet the increasingly strict low-carbon requirements for automotive structural materials and evaluate the degree of greenization of interstitial-free (IF) steel, a lifecycle assessment (LCA) model for IF steel from "cradle to gate" was established based on the unit process and process flow of IF steel production. The carbon footprint (CF) results of IF steels were obtained, and the main influencing factors of the CF were analyzed. The sensitivity of CF to the influencing factors was further analyzed, and the corresponding carbon reduction proposals were formulated based on the sensitivity results. The results show that the CF of IF steel is 2.78 kg/kg, and the upstream process composed of raw material processing and transportation accounts for only 22.98%, while the production process accounts for 77.02%, among which the blast furnace (BF) plant and sinter plant contribute the largest CF. The CF of the BF plant mainly comes from the use of coke, the combustion of blast furnace gas (BFG) and the combustion of pulverized coal injection, and the CF contribution of the sinter plant mainly comes from the use of coke and electricity. The sensitivity analysis results show that the CF is highly sensitive to the proportion of sinter, the proportion of self-produced coke and the recovery rate of BFG. The development of self-soluble pellet technology, intelligent coke production and optimization of intelligent BFG scheduling are helpful for iron and steel industry to achieve green and low-carbon IF steel production.
• Select
  Numerical simulation on pneumatic atomized water for granulation and heat recovery of molten slag

  ZHAO Kai, BAI Xiaojun, WANG Dianxin, ZHEN Changliang, BAI Gengchen, CHEN Wei

  China Metallurgy. 2025, 35(3): 144-154. https://doi.org/10.13228/j.boyuan.issn1006-9356.20240457

  Abstract ( ) XML Related Articles File   Knowledge map   Save

  The traditional granulation and heat recovery of slag has defects such as poor granulation effect,low heat recovery efficiency,and high resource consumption. It is urgent to developing new quenching media. Air,water,and pneumatic atomized water were chosen as new quenching media. Fluent software is used to simulate the interaction process of quenching media and slag,compare the effects of granulation and heat recovery,analyze the mechanism and influencing factors,and then propose the optimization direction of process parameters. The results show that the flow field outside the dual-medium atomization nozzle has a large atomization radius,a large jet velocity and a long jet distance. The optimized structural parameters are as follows:the shrinkage half cone angle is 40°,the throat diameter is 5 mm,the expansion section length is 13 mm,and the expansion half cone angle is 5°. The Pneumatic atomized water significantly improves the effect of granulation. The proportion of particle size below 1 mm in the slag increased by 8%,and the proportion of particle size above 4 mm decreased by 3%. Meanwhile,the atomized water increases the temperature reduction of the slag by 200 K,which provides effective support for improving the heat exchange efficiency. The simulation results of the cooling and solidification process show that the gas velocity and water velocity of pneumatic atomized water are two key factors affecting the granulation and heat transfer effect. The process with water velocity of 15 m/s and gas velocity of 65 m/s has a better granulation effect and a higher degree of solidification. Compared with the traditional treatment method,the granulation effect and heat recovery efficiency are enhanced,which provides theoretical and technical support for metallurgical enterprises to carry out slag heat recovery and resource utilization.
Intelligentization Technology
• Select
  Intelligent acceptance model for steel scrap based on dense small object detection

  WANG Chao, XIAO Pengcheng, ZHU Liguang, ZHU Rong

  China Metallurgy. 2025, 35(3): 155-167. https://doi.org/10.13228/j.boyuan.issn1006-9356.20240449

  Abstract ( ) XML Related Articles File   Knowledge map   Save

  Scrap steel is a recyclable green resource critical to the sustainable development of the steel industry. The grade and composition of scrap steel significantly affect production costs and steel quality. Therefore,classification and grading of scrap steel before smelting have become key areas of research in the steel industry. To address the limitations of traditional manual classification methods in tE_RMS of fairness and safety, the CSIL-Net model for scrap steel quality inspection was developed based on object detection technology. First,a multi-camera scrap steel image acquisition system was established to collect and annotate images of various types of scrap steel. Through image augmentation,a comprehensive scrap steel dataset was created. Then,building on the YOLOv5 model,a small object detection layer P2 was added to improve detection performance for small scrap steel targets in complex scenes. The P5 detection layer was removed to simplify the model,leading to the design of a novel three-layer detection network (IL:Improved Layer). The Bi-directional Feature Pyramid Network (BiFPN) was introduced to enhance the model's ability to fuse cross-scale features and transmit information. Soft Non-Maximum Suppression (Soft-NMS) was applied to mitigate missed detections in cases of dense scrap steel overlap. Finally,the enhanced scrap steel dataset was used to train and validate the model,evaluating key metrics such as network depth,complexity,and inference time. Experimental results showed that the average precision of the CSIL-Net model across all categories increased from 88.8% to 96.3%,with an inference speed of just 21.8 ms per image. Compared to manual methods,the model demonstrates significant advantages in both accuracy and speed,making it well-suited for real-world applications to address challenges of real-time detection and fairness in scrap steel grading.
Metallurgy Talent
• Select
  LI Lanjie

  China Metallurgy. 2025, 35(3): 168-168.

  Abstract ( ) Download PDF ( ) HTML ( ) XML Related Articles File   Knowledge map   Save
• Select
  QI Jianghua

  China Metallurgy. 2025, 35(3): 169-169.

  Abstract ( ) Download PDF ( ) HTML ( ) XML Related Articles File   Knowledge map   Save