JOURNAL OF IRON AND STEEL RESEARCH INTERNATIONAL

钢铁研究学报(英文版)

Home | About Journal | Editorial Board | Instruction for Authors | Subscriptions | Advertisement | Contacts Us | Chinese

　

Office Online

　

　

　

　

　

　

Editor-in-chief

　

Journal Online

　

　

Forthcoming Articles

　

　

　

　

Download Articles

　

　

Quick Search

Advanced Search

	2023 Vol. 30 No. 05
	Published: 2023-06-24



	Original Paper

Original paper

	851	Shu-yi Zhou, Xiao-yan Liu
		A simple image-based method for online moisture content estimation of iron ore green pellets
		Moisture content (MC) is an important quality metric of iron ore green pellets in pelletizing process in ironmaking industry. Current image-based methods for MC estimation may result in big errors for gray-scale pellet images captured under various lighting conditions. We proposed a simple image-based method to improve the MC estimation accuracy by illumination correction and linear regression modeling. Firstly, the illumination of the pellet image was transformed into the reference illumination by use of a color checker chart and piecewise linear interpolation, so that the influence of different illuminations could be greatly reduced. By experimental analysis, it was found that MC is approximately adversely proportional to the average intensity of the transformed images. A simple model for MC prediction was then established by linear fitting. Experiments demonstrated that the proposed method has good robustness to different lighting conditions and achieves the best performance in metrics of mean square error, mean absolute error and maximum absolute error in comparison with five state-of-art MC estimating methods. Application on a working disk pelletizer shows that the proposed method can predict well the change of moisture content with time, and its computing efficiency can satisfy the requirement for online MC monitoring during the pelletizing process.
		2023 Vol. 30 (05): 851-863 [Abstract] ( 73 ) [HTML 1KB] [PDF 0KB] ( 141 )

	864	Shu-han Liu, Wen-qiang Sun, Wei-dong Li, Bing-zhen Jin
		Prediction of blast furnace gas generation based on data quality improvement strategy
		The real-time energy flow data obtained in industrial production processes are usually of low quality. It is difficult to accurately predict the short-term energy flow profile by using these field data, which diminishes the effect of industrial big data and artificial intelligence in industrial energy system. The real-time data of blast furnace gas (BFG) generation collected in iron and steel sites are also of low quality. In order to tackle this problem, a three-stage data quality improvement strategy was proposed to predict the BFG generation. In the first stage, correlation principle was used to test the sample set. In the second stage, the original sample set was rectified and updated. In the third stage, Kalman filter was employed to eliminate the noise of the updated sample set. The method was verified by autoregressive integrated moving average model, back propagation neural network model and long short-term memory model. The results show that the prediction model based on the proposed three-stage data quality improvement method performs well. Long short-term memory model has the best prediction performance, with a mean absolute error of 17.85 m3/min, a mean absolute percentage error of 0.21%, and an R squared of 95.17%.
		2023 Vol. 30 (05): 864-874 [Abstract] ( 69 ) [HTML 1KB] [PDF 0KB] ( 118 )

	875	Jia-wei Guo, Dong-ping Zhan, Guo-cai Xu, Nai-hui Yang, Bo Wang, Ming-xin Wang, Geng-wei You
		An online BOF terminal temperature control model based on big data learning
		The development of basic oxygen furnace (BOF) intelligent steelmaking model based on artificial intelligence and big data is the focus of international research and development. In the view of the current situation that the BOF cannot continuously detect the composition and molten steel temperature, combined with the monitoring results of the high-definition and high-brightness camera at the converter mouth, an online BOF terminal temperature control model is established based on big data learning case-based reasoning model and expert system model. The on-site online operation shows that the model can effectively improve the ‘‘flying lance’’ phenomenon and the splashing condition, the stability and safety of smelting process are better than that of artificial smelting, the ‘‘flying lance’’ rate decreases from 39.2% to 0, the early splashing rate decreases from 21.4% to 13.3% and the late splashing rate decreases from 81.25% to 56.7%. When the temperature fluctuation is controlled at ± 15 oC, the hit rate of the terminal temperature under the automatic control of the model is 90.91%.
		2023 Vol. 30 (05): 875-886 [Abstract] ( 93 ) [HTML 1KB] [PDF 0KB] ( 145 )

	897	Yang Han, Ze-qian Cui, Li-jing Wang, Jie Li, Ai-min Yang, Yu-zhu Zhang
		Cascade model for continuous prediction of silicon content of molten iron with coupled state variable nodes
		With the goal of achieving advanced and multi-step prediction of silicon content of molten iron in the blast furnace ironmaking process, a path adaptive optimization seeking strategy coupled with simulated annealing algorithm and genetic algorithm was proposed from the perspective of innovative intelligent algorithm application. It was further coupled with wavelet neural network algorithm to deeply explore the nonlinear and strong coupling relationship between the information of big data samples and construct a cascade model for continuous prediction of silicon content of molten iron with the intelligent research results of state variables such as permeability index as the node and silicon content forecast as the output. In the model construction process, the 3r criterion was used for non-anomaly estimation of abnormal data to build a time-aligned sample set for multi-step forecasting of iron content, the normalization method was used to eliminate the influence of dimensionality of sample information, and the spearman correlation analysis algorithm was used to eliminate the time delay between state variables, control variables, and silicon content of molten iron in the blast furnace smelting process. The results show that permeability and theoretical combustion temperature as the key state variable nodes have real-time correlation with the silicon content of molten iron, and there are accurate forecasting results on the optimal path with the endpoint of molten iron silicon content prediction. The path finding based on the improved genetic algorithm of simulated annealing has good effect on the downscaling and depth characterization of sample data and improves the data ecology for the application of wavelet neural network algorithm. The accuracy of the real-time continuous forecasting model for the silicon content of molten iron reaches 95.24%; the hit rate of continuous forecasting one step ahead reaches 91.16%, and the hit rate of continuous forecasting five steps ahead is 87.41%. This model, which can realize the nodal dynamics of state variables, has better promotion value.
		2023 Vol. 30 (05): 897-914 [Abstract] ( 62 ) [HTML 1KB] [PDF 0KB] ( 125 )

	915	Ran Liu, Zhi-feng Zhang, Xin Li, Xiao-jie Liu, Hong-yang Li, Xiang-ping Bu, Jun Zhao, Qing Lyu
		Hot metal quality monitoring system based on big data and machine learning
		The system of hot metal quality monitoring was established based on big data and machine learning using the real-time production data of a steel enterprise in China. A working method that combines big data technology with process theory was proposed for the characteristics of blast furnace production data. After the data have been comprehensively processed, the independent variables that affect the target parameters are selected by using the method of multivariate feature selection. The use of this method not only ensures the interpretability of the input variables, but also improves the accuracy of the machine learning process and is more easily accepted by enterprises. For timely guidance on production, specific evaluation rules are established for the key quality that affects the quality of hot metal on the basis of completed predictions work and uses computer technology to build a quality monitoring system for hot metal. The online results show that the hot metal quality monitoring system established by relying on big data and machine learning operates stably on site, and has good guiding significance for production.
		2023 Vol. 30 (05): 915-925 [Abstract] ( 63 ) [HTML 1KB] [PDF 0KB] ( 134 )

	926	Guang-da Bao, Ting Wu, Duo-gang Wang, Xiao-bin Zhou, Hai-chuan Wang
		Multi-model coupling-based dynamic control system of ladle slag in argon blowing refining process
		Since the current slagging of argon blowing refining process is relatively fixed, which cannot adapt to the fluctuation of converter smelting process, it poses the problems of poor metallurgical property of refining slag and a large amount of molten heel. An optimization system coupled with multiple models was proposed to dynamic control the ladle slagging in the argon blowing refining process. It can compile the optimal dynamic slagging scheme in real time under the guarantee of deoxidation performance and reasonable fluidity. The argon blowing refining slag composition range of CaO/Al2O3 = 1.3–1.7, CaO/SiO2 = 6–12, w(MgO) = 2%–6% was determined based on FeO activity and liquidus temperature by equilibrium thermodynamic calculation. In addition, it demonstrated better performance in the viscosity prediction task of the presented Visual Geometry Group 16-like one-dimensional convolutional neural network deep learning algorithm versus the Random Forest ensemble learning algorithm, as the adjusted coefficients of determination were 0.9712 and 0.9637, respectively. After the system was applied in operation, the argon blowing refining process was stable, and the steel yield was enhanced, which promoted the intelligent steelmaking level while achieving the cost reduction and efficiency improvement.
		2023 Vol. 30 (05): 926-936 [Abstract] ( 57 ) [HTML 1KB] [PDF 0KB] ( 129 )

	937	Gong-hao Lian, Qi-hao Sun, Xiao-ming Liu, Wei-miao Kong, Ming Lv, Jian-jun Qi, Yong Liu, Ben-ming Yuan, Qiang Wang
		Automatic recognition and intelligent analysis of central shrinkage defects of continuous casting billets based on deep learning
		The internal quality inspection of the continuous casting billets is very important, and mis-inspection will seriously affect the subsequent production process. The UNet-VGG16 transfer learning model was used for semantic segmentation of the central shrinkage defects of the continuous casting billets. The automatic recognition accuracy of the central shrinkage defects of the continuous casting billets reaches more than 0.9. We use the minimum circumscribed rectangle to quantify the geometric dimensions such as length, width and area of the central shrinkage defects and use the threshold method to rate the central shrinkage defects of the continuous casting billets. The results show that all the testing images are rated correctly, and this method achieves the automatic recognition and intelligent analysis of the central shrinkage defects of the continuous casting billets.
		2023 Vol. 30 (05): 937-948 [Abstract] ( 60 ) [HTML 1KB] [PDF 0KB] ( 146 )

	949	Long Zhang, Sai-fei Yan, Jun Hong, Qian Xie, Fei Zhou, Song-lin Ran
		An improved defect recognition framework for casting based on DETR algorithm
		The current casting surface defect detection algorithms suffer from poor small target defect recognition and imbalance between detection performance and detection time. An improved algorithmic framework for casting defect detection was proposed based on the DEtection TRansformer (DETR) algorithm. The algorithm takes ResNet with an efficient channel attention (ECA)-Net module as the backbone network. In addition, based on the original algorithm architecture, dynamic anchor boxes, improved multi-scale deformable attention module, and SIoU loss function are introduced to improve the sensitivity of transformer structure to input location information and scale size, and the small target defect detection performance is effectively improved. The recognition performance of the algorithm in a self-built casting defect dataset was studied. The improved DETR algorithm has 97.561% accuracy in recognizing two defects, namely sandinclusion and notch, with the detection rate being improved by 65.854% and 17.073% compared with the original DETR and you only look once (Yolo)-V5, respectively. This algorithm verifies the applicability of the transformer architecture target detection algorithm for casting defect detection tasks and provides new ideas for detecting other similar application scenarios.
		2023 Vol. 30 (05): 949-959 [Abstract] ( 81 ) [HTML 1KB] [PDF 0KB] ( 140 )

	960	Le-bao Song, Dong Xu, Peng-fei Liu, Jin-hang Zhou, Hui-qing Yan, Jing-dong Li, Hai-nan He, Hai-jun Yu, Xiao-chen Wang, Quan Yang
		A novel mechanism fusion data control method for slab camber in hot rolling
		For asymmetric plate shape, control over the hot rolling process mainly depends on the subjective judgement and personal experience of the operator as there are great deviations and much instability in hot rolling. Unfortunately, the intrinsic mechanisms and sensitivity affecting characteristic parameters and variables in the asymmetric rolling process remain understudied. Therefore, a novel mechanism fusion data control method for slab camber in hot rolling using dimensional analysis and data-driven technique was proposed. The approach of dimensional analysis was used to establish a mathematical model and analyse the main parameters affecting the slab camber of the rough rolling. Subsequently, the established mathematical model combined with the data-driven techniques was employed to accurately predict the slab bending value. Furthermore, the superiority and effectiveness of the proposed model were demonstrated by a comparison with three regression models. Finally, the proposed control strategy was successfully applied in a 1580 mm hot rolling industrial process. The automatic control results show that the hit rate of slab cambers in different sizes from 10 to 30 mm is improved, and the quality stability of intermediate slab is significantly improved.
		2023 Vol. 30 (05): 960-970 [Abstract] ( 55 ) [HTML 1KB] [PDF 0KB] ( 131 )

	971	Qian-qian Dong, Qing-ting Qian, Min Li, Gang Xu
		Monitoring and diagnosis of complex production process based on free energy of Gaussian–Bernoulli restricted Boltzmann machine
		Online monitoring and diagnosis of production processes face great challenges due to the nonlinearity and multivariate of complex industrial processes. Traditional process monitoring methods employ kernel function or multilayer neural networks to solve the nonlinear mapping problem of data. However, the above methods increase the model complexity and are not interpretable, leading to difficulties in subsequent fault recognition/diagnosis/location. A process monitoring and diagnosis method based on the free energy of Gaussian–Bernoulli restricted Boltzmann machine (GBRBM-FE) was proposed. Firstly, a GBRBM network was established to make the probability distribution of the reconstructed data as close as possible to the probability distribution of the raw data. On this basis, the weights and biases in GBRBM network were used to construct F statistics, which represents the free energy of the sample. The smaller the energy of the sample is, the more normal the sample is. Therefore, F statistics can be used to monitor the production process. To diagnose fault variables, the F statistic for each sample was decomposed to obtain the Fv statistic for each variable. By analyzing the deviation degree between the corresponding variables of abnormal samples and normal samples, the cause of process abnormalities can be accurately located. The application of converter steelmaking process demonstrates that the proposed method outperforms the traditional methods, in terms of fault monitoring and diagnosis performance.
		2023 Vol. 30 (05): 971-984 [Abstract] ( 39 ) [HTML 1KB] [PDF 0KB] ( 133 )

	985	Jia-qiang Chen, Shu-zong Chen, Chang-chun Hua, Cheng Jia, Cheng Qian
		Extended-state-observer-based robust torsional vibration suppression for rolling mill main drive system with input saturation
		A robust torsional vibration suppression strategy is proposed for the main drive system of the rolling mill subject to uncertainties, disturbances and input saturation. With given model information incorporated into observer design, an extended state observer that relies only on roller speed measurements is developed to estimate the system states and lumped uncertainties of the rolling mill main drive system. To handle the motor torque saturation, an auxiliary signal system with the same order as the plant is constructed. The error between the control input and plant input is taken as the input of the constructed auxiliary system, and a number of signals are generated to compensate for the effect of the motor torque saturation. Furthermore, a robust output feedback controller is introduced to obtain better transient and steady-state performance of the rolling mill main drive system and the stability of the closed-loop system is strictly proved via Lyapunov theory. Finally, comparative simulations are performed to verify the effectiveness and superiority of the proposed control strategy.
		2023 Vol. 30 (05): 985-993 [Abstract] ( 59 ) [HTML 1KB] [PDF 0KB] ( 148 )

	994	Yang-huan Xu, Dong-cheng Wang, Bo-wei Duan, Hong-min Liu
		Data-driven flatness intelligent representation method of cold rolled strip
		A high-accuracy flatness prediction model is the basis for realizing flatness control. Real flatness is typically reflected as the strain distribution, which is a vector. However, it is difficult to obtain ideal results if the real flatness is directly used as the output value of the flatness intelligent prediction model. Thus, it is necessary to seek an abstract representation method of real flatness. For this reason, two new intelligent flatness representation models were proposed based on the autoencoder of unsupervised learning theory: the flatness autoencoder representation (FAR) model and the flatness stacked sparse autoencoder representation (FSSAR) model. Compared with the traditional Legendre fourth-order polynomial representation model, the representation accuracies of the FAR and FSSAR models are significantly improved, better representing the flatness defects, like the double tight edge. The optimal number of bottleneck layer neurons in the FAR and FSSAR models is 5, which means that five basic patterns can accurately represent real flatness. Compared with the FAR model, the FSSAR model has higher representation accuracy, although the flatness basic pattern is more abstract, and the physical meaning is not clear enough. Furthermore, the accuracy of the FAR model is slightly lower than that of the FSSAR model. However, it can automatically learn the flatness basic pattern with a very clear physical meaning for both the theoretical and real flatness, which is an optimal intelligent representation method for flatness.
		2023 Vol. 30 (05): 994-1012 [Abstract] ( 35 ) [HTML 1KB] [PDF 0KB] ( 133 )

	1013	Yun-jian Hu, Jie Sun, Huai-tao Shi, Qing-long Wang, Jian-zhao Cao
		Roll gap prediction in acceleration and deceleration process of cold rolling based on a data-driven method
		Severe fluctuation of the effective roll gap in the acceleration and deceleration section of the cold rolling process is a significant factor causing thickness deviation. However, the conventional roll gap compensation method and control strategy do not meet the stringent strip quality requirements. The roll gap model in the acceleration and deceleration process is studied to increase the thickness control precision. In order to improve model accuracy, a roll gap prediction method based on data-driven is proposed. Given the complexities of the cold rolling process, the extreme gradient boosting (XGBoost) method is used to predict the roll gap model as the rolling speed changes. Meanwhile, support vector regression and neural network-based methods are taken to evaluate and compare the prediction performances. Based on the field data, the simulation experiments are carried out. It demonstrated that the prediction performance of the proposed method outperformed the other two methods. The values of root mean square error, determination coefficient value, mean absolute percentage error and mean absolute error obtained from the XGBoost model were equal to 0.000346, 0.952, 7.02, and 0.00028, respectively. In addition, the proposed method analyzed the contribution rates of the rolling affecting parameters on the roll gap. The data showed that in the controllable rolling parameters, the rolling speed is the most impacting factor that disturbs the roll gap model in the acceleration and deceleration process, which can provide a useful direction for actual roll gap adjustment.
		2023 Vol. 30 (05): 1013-1021 [Abstract] ( 48 ) [HTML 1KB] [PDF 0KB] ( 126 )

	1022	Li Wang, Song-lin He, Zhi-ting Zhao, Xian-du Zhang
		Prediction of hot-rolled strip crown based on Boruta and extremely randomized trees algorithms
		The quality of hot-rolled steel strip is directly affected by the strip crown. Traditional machine learning models have shown limitations in accurately predicting the strip crown, particularly when dealing with imbalanced data. This limitation results in poor production quality and efficiency, leading to increased production costs. Thus, a novel strip crown prediction model that uses the Boruta and extremely randomized trees (Boruta–ERT) algorithms to address this issue was proposed. To improve the accuracy of our model, we utilized the synthetic minority over-sampling technique to balance the imbalance data sets. The Boruta–ERT prediction model was then used to select features and predict the strip crown. With the 2160 mm hot rolling production lines of a steel plant serving as the research object, the experimental results showed that 97.01% of prediction data have an absolute error of less than 8 lm. This level of accuracy met the control requirements for strip crown and demonstrated significant benefits for the improvement in production quality of steel strip.
		2023 Vol. 30 (05): 1022-1031 [Abstract] ( 49 ) [HTML 1KB] [PDF 0KB] ( 134 )

	1032	Xiao-ya Huang, Biao Zhang, Qiang Tian, Hong-hui Wu, Bin Gan, Zhong-nan Bi, Wei-hua Xue, Asad Ullah, Hao Wang
		Machine learning study on time–temperature–transformation diagram of carbon and low-alloy steel
		Time–temperature–transformation (TTT) diagram plays a critical role in designing appropriate heat treatment process of steels by describing the relationship among holding time, temperature, and quantities of phase transformation. Making predictions for TTT diagrams of new steel rapidly and accurately is therefore of much practical importance, especially for costly and time-consuming experimental determination. Here, TTT diagrams for carbon and low-alloy steels were predicted using machine learning methods. Five commonly used machine learning (ML) algorithms, backpropagation artificial neural network (BP network), LibSVM, k-nearest neighbor, Bagging, and Random tree, were adopted to select appropriate models for the prediction. The results illustrate that Bagging is the optimal model for the prediction of pearlite transformation and bainite transformation, and BP network is the optimal model for martensite transformation. Finally, the ML framework composed of Bagging and BP network models was applied to predict the entire TTT diagram. Additionally, the ML models show superior performance on the prediction of testing samples than the commercial software JMatPro.
		2023 Vol. 30 (05): 1032-1041 [Abstract] ( 57 ) [HTML 1KB] [PDF 0KB] ( 125 )

	1042	Xing-qi Jia, Feng-hua Lu, Kai Yang, Shi-long Liu, Chun Yu, Wei Li, Xue-jun Jin
		An optimization of harmonic structure nickel-saving cryogenic steel via combinatorial high-throughput experiment
		A combinatorial high-throughput experiment (HTE) was used to optimize composition and process of nickel-saving cryogenic steel. A gradient temperature heat treatment method with a high linear distribution of heat treatment temperature using customized graphite sleeve direct current heating was used in the combinatorial HTE, which enhanced the richness of the sample library for the single preparation of the 102 level component process variables. Cryogenic steel with excellent mechanical properties was optimized using this combinatorial HTE, and the Ni content was reduced from the traditional 9 to 5.6 wt.% by using Mn instead of Ni. The heterogeneous structure architecture strategy and strengthening and toughening mechanism of the harmonic structure induced by intrinsic heat treatment of additive manufacturing were revealed. Taking the composition process optimization of Ni-saving cryogenic steel as an example, the boosting ability of combinatorial HTE in the research and development of new metal materials was proposed.
		2023 Vol. 30 (05): 1042-1049 [Abstract] ( 62 ) [HTML 1KB] [PDF 0KB] ( 148 )

	1050	Xuan-dong Wang, Nan Li, Hang Su, Hui-min Meng
		Prior austenite grain boundary recognition in martensite microstructure based on deep learning
		Grain size determination is essential in producing and testing iron and steel materials. Grain size determination of martensitic steels usually requires etching with picric acid to reveal the prior austenite grain boundaries. However, picric acid is toxic and explosive and belongs to hazardous chemicals, which makes it difficult for laboratories and testing institutions to obtain. A new experimental method was developed to use Nital etchant instead of picric acid. The deep learning method was used to recognize the prior austenite grain boundaries in the etched martensite microstructure, and the grain size could be determined according to the recognition result. Firstly, the polished martensite specimen was etched twice with Nital etchant and picric acid, respectively, and the same position was observed using an optical microscope. The images of the martensitic structure and its prior austenite grain boundary label were obtained, and a data set was constructed. Secondly, based on this data set, a convolutional neural network model with a semantic segmentation function was trained, and the accuracy rate of the test set was 87.53%. Finally, according to the recognition results of the model, the grain size rating can be automatically determined or provide a reference for experimenters, and the difference between the automatic determination results and the measured results is about 0.5 level.
		2023 Vol. 30 (05): 1050-1056 [Abstract] ( 49 ) [HTML 1KB] [PDF 0KB] ( 129 )

	887	Yu-xiao Liu, Yan-wu Dong, Zhou-hua Jiang, Yu-shuo Li, Wei Zha, Yao-xin Du, Shu-yang Du
		XGBoost-based model for predicting hydrogen content in electroslag remelting
		An Xtreme Gradient Boosting (XGBoost)-based endpoint hydrogen content prediction model was proposed for the electroslag remelting process, the data collected in the field were pre-processed, and the characteristic variables of the physical parameters related to the variation of hydrogen content in the electroslag remelting process were selected by machine learning analysis and metallurgical mechanism. The kernel ridge regression model, ridge regression model, XGBoost model, support vector regression model and gradient boosting regression model were developed and validated using the electroslag remelting data collected from the steel mills, and the model structure and parameters were adjusted several times. The prediction accuracy of hydrogen content was compared horizontally. The XGBoost model was validated for the test set with the following hit rates: 70.59%, 82.35% and 100% for the endpoint hits at the allowable hydrogen content error of ± 0.05 × 10^–6, ± 0.10 × 10^–6 and ± 0.50 × 10^–6, respectively.
		2023 Vol. 30 (05): 887-896 [Abstract] ( 43 ) [HTML 1KB] [PDF 0KB] ( 117 )

钢铁研究学报(英文版)

　

News

　

　

·	《钢铁研究学报》第二届青年编委招募

·	Call for Papers for the Special Issue on “Corrosion and wear of materials in extreme environments”

·	Call for Papers for the Special Issue on “Resources Recycling of Solid Wastes from Ironmaking and St

·	Call for Papers for the Special Issue on "Numerical Simulation of Multiphase and Multiscale Metallur

·	Call for Papers for the Special Issue on “Application of AI in steelmaking and ferrous materials”

　　　　　　　　　　More

　

Author Center

　

　

Instruction for Authors

　

　

Copyright Agreement

　

Links

　

　　　　　　　　　　More

Copyright © JOURNAL OF IRON AND STEEL RESEARCH INTERNATIONAL　
Supported by: Beijing Magtech