In automation control for metallurgical unmanned overhead cranes,a critical challenge lies in balancing operational speed and anti-swing measures to enhance crane efficiency while ensuring safe operation. A metallurgical unmanned crane anti-swing full-range speed control method was pro- posed based on controller parameter switching derived from human expertise under multi-run states in this article. The method first plans the full-range speed control for the overhead crane linked by a large and small vehicle under anti-swing conditions. Then anti-swing control is optimized by switching PID parameters based on human crane operation experience under different running states,forming an integrated feedforward-feedback hybrid control for the unmanned overhead crane. Simulation experi- ments demonstrate that,compared to the commonly used manual operation of large and small vehicle linkage control,this method reduces overall operation time by 24. 9% . Field experiments show thatthe maximum swing angle is controlled within the range of 0. 45°-0. 85 °,and positioning accuracy is maintained within 10 mm. The relevant indicators are in a leading position in similar domestic scenar- ios. This method has been successfully applied and validated in the steel industry in China.

Sintering is a pre-process of blast furnace ironmaking,and the quality of sintered ore directly affects the quality and quantity of hot metal in the ironmaking process. Intelligent prediction and control of key parameters plays an important role in improving the quality of sintering ore in the sintering process. Firstly,the flowchart is introduced and its process characteristics is analyzed. Then, the predictive modeling research status for quality indicators and state parameters in the sintering process are reviewed. On this basis,the control methods about burn through point and ignition temperature in detail is illustrated. Finally,conclusions and prospects for the predictive and control modeling of key parameters in the sintering process are made.

An online thickness prediction algorithm for strip steel was proposed to address issues of strong coupling and low accuracy in the thickness mathematical model. Firstly,the rolling data is used to establish an online thickness prediction model based on light gradient boosting machine (LightGBM) model. Then improved bat algorithm (IBA) is applied to improve the LightGBM model parameters,and a self-learning system is deployed to optimize the results. Finally,the predicted results are compared with the actual thickness to verify the accuracy of the prediction model. The experimental results show that the online thickness prediction algorithm can quickly and accurately predict the strip thickness. When the IBA-LightGBM model was used to predict the 2 mm,3 mm,4 mm,and 5. 65 mm  strips,root mean square error ERMS (RMSE) can be controlled within 11. 0 μm,11. 5 μm,11. 6 μm and 16. 4 μm respectively. The results can improve the accuracy of the thickness mathematical model for hot rolling strip and enhance the level of the thickness control system.

With the deepening of the transformation of the steel industry towards automation and intelligence,intelligent equipment systems based on machine vision have been widely applied due to their outstanding advantages of low cost,high accuracy,stability and reliability. Based on extensive practical experience in industrial sites,this article studies and designs an intelligent equipment software platform for the steel industry. It adopts a layered structure,builds underlying libraries,development templates,communication protocols,instruction sets,etc. as the basic support,and provides essential software functional modules in the intelligent equipment system to assist in the rapid construction of new applications. The platform module involves a comprehensive process of data acquisition,storage, image algorithm processing,device control,data display,and fault diagnosis,fully considering the structural requirements and functional support for intelligent equipment in industrial scenarios. Through this software platform,the development time of specific intelligent equipment applications can be greatly reduced,and the requirements for application implementation can be met while ensuring system stability.

The blast furnace ironmaking process is of great significance in the iron and steel industry. However,due to the complexity of the blast furnace ironmaking process and the existence of high temperature,high pressure,and complex physical and chemical reactions,it is still a major challenge to establish an effective process monitoring model. Aiming at the requirement of multi-step real-time prediction of hot metal key quality indicators in the process of blast furnace ironmaking,this paper refers to the theory of continuous learning (CL) that simulates the human brain's hippocampus and neocortex,and constructs a CL-based time series data pre diction modeling framework based on multihead self-attention mechanism,one-dimensional residual neural network (ResNet),long short term memory network (LSTM) and multi-layer perceptron (MLP) in order to realize the multi-step realtime prediction task of hot metal key quality indicators of the blast furnace ironmaking process. The experimental results show that the proposed CL-based method is superior to the traditional deep learning models,achieves higher prediction accuracy,and with the increase of prediction time step,the proposed CL-based model shows strong robustness.

The continuous deepen of the global Industry 4. 0 revolution has driven the development of computationally intensive industrial applications such as smart steel. Steel production,characterized by diverse processes, complex procedures, and concentrated high-temperature and high-pressure equipment,necessitates real-time monitoring and analysis of production status to optimize production processes. Due to the opaque nature of each process,with real-time status being difficult to accurately obtain,digital twin (DT) technology offers a solution by facilitating real-time simulation and transparency in steel production,thus improving efficiency and quality. A multi-layer collaborative scheduling framework for industrial networks based on DT is presented,aiming to enhance resource management and network security. Challenges and explores solutions for novel resource scheduling and network security under DT-enabled industrial Internet of things (IIoT) is analyzed. Moreover,the exploration  of cross-layer resource scheduling strategies attracts extensive focus,facilitated by twin data at both the physical layer and data link layer. Finally,it points out the emerging trends in DT-enabled smart steel resource management,offering valuable insights for the digital transformation of the steel industry.

At present , a neWWaVe of technological reVolution is emerging globally , represented by in- dustrial technologies such as artificial intelligence , big data and industry 4 . 0 .  The traditional chemical plant control model can no longer meet the  needs  of digital  manufacturing.   Typical  problems  such  as complex production processes and difficulty in collaboratiVe operations in the steel industry haVe gradu- ally emerged.  Enterprises also haVe uneVen information leVels  and serious information island phenome- na.  In this regard , the factory and production line based on emerging technologies such as digital tWin technology , internet of things , machine learning and Video intelligent recognition , accesses Video ima- ges , sensor monitoring data , system  control  data , external  draWing  information  and  management  sys- tems , and the status of production line  equipment and materials  in real time Were  established , Visual linkage of  data  information  Within  the  production  line  Was  displayed ,  and  information  islands  Were eliminated.  At the same time , artificial intelligence algorithms Were used to merge and splice the com- plex and scattered Video images , and splice the production line monitoring images.  For large-scale key areas , real-time monitoring Was achieVed in the form of“ one picture ”.

Sintering is one of the important processes in blast furnace ironmaking. Using intelligent technology to accurately predict the state,quality and other parameters during the sintering process to control the sintering process is crucial for reducing production costs,optimizing the sintering process, and improving production safety. Firstly,the role of big data platforms in sintering parameter prediction technology was analyzed. Secondly,the application of intelligent technology in sintering production process was compared and analyzed from the aspects of sintering endpoint position prediction, sintering material layer permeability prediction,sintering endpoint temperature prediction,FeO content prediction,drum strength prediction,ignition temperature prediction,bellows valve opening prediction,and economic and technical index prediction. The development process of FeO content prediction from manual observation method to mathematical modeling method and then to artificial intelligence method,revealing the development status and evolution law of intelligent sintering prediction technologybasedonbigdata.Inaddition,theshortcomingsanddevelopment trendsof intelligent sinteringpredictiontechnology insinteringproductionpredictionwerediscussed.

Improving the operation efficiency of cranes in the steelmaking area can reduce energy consumption for transportation while effectively linking the preceding and succeeding processes, which is of certain value for green production,cost reduction,and efficiency increase. In this regard, this article proposed a crane scheduling optimization method driven by simulation modeling and machine learning. Firstly,multi-agent is used to establish a production simulation model for the steelmaking area,which is driven by historical production plans and crane scheduling workflows. Subsequently,the simulation model is run multiple times to obtain a large number of high-quality crane operation samples through built-in sample evaluation formulas. Finally,a random forest model is employed to learn from the samples and obtain a machine learning model for matching cranes with transportation tasks. Experimental analysis shows that applying the machine learning model to crane scheduling decisions can increase the proportion of effective transportation time,thereby reducing energy consumption losses caused by mismatched transportation tasks,path avoidance,etc. This advantage is particularly significant under heavy production loads. Furthermore,the crane scheduling machine learning model is decoupled from the steelmaking plan,exhibiting high flexibility in practical app lications.

In order to achieve more accurate calculation of alloy addition amount,a comprehensive al- gorithm based on genetic algorithm ( GA) optimization for BP neural network was adopted. During the training process of the BP neural network,the refining starting composition of the molten steel was used as input parameters for the BP neural network model. Then,the fitness function of the GA was used to optimize and adjust the weight and threshold of BP neural network,predicting the refining endpoint steel composition of the LF furnace. By comparing the prediction results of BP neural net- work algorithm and GA-BP neural network algorithm, it was found that the GA-BP algorithm has smaller mean absolute error (MAE) and mean square error (MSE) ,and the prediction results aremore accurate and basically consistent with the actual steel composition,indicating that this model can be used in production. Based on the GA-BP neural network model,the alloy addition amount was determined according to the composition of the molten steel at the beginning of LF refining and con- trol composition requirements. After deploying a predictive model in the 140 ton ladle LF refining sys- tem of a steel plant and tracking 188 furnace data,the difference between the actual amount of alloy added and the predicted amount was within ±30 kg,then the prediction accuracy of high manganese alloy was 91. 3% ,high chromium alloy was 90. 4% ,ferrosilicon alloy was 90. 2% ,and the predic- tion accuracy of carburant was 91% in furnaces,from which can guide the determination of alloy ad- dition amount in the actual refining process. 

The prediction of the endpoint carbon content and molten steel temperature in converter steelmaking is of great significance for the precise control of molten steel composition and temperature and the improvement of product quality. Due to the high dimension, high noise and strong nonlinearity of the converter steelmaking process data, it is difficult to obtain high-quality data. Direct modeling not only has a low hit rate but also is prone to overfitting. To address this problem, this paper proposes a random forest prediction model based on adaptive SMOTE data enhancement technology was proposed. Firstly, the feature selection of the endpoint carbon content and molten steel temperature iswas performed by recursive elimination method. Secondly, the adaptive SMOTE algorithm iswas used to enhance the original data. Finally, random forest iswas used to predict the endpoint carbon content and molten steel temperature respectively. The actual industrial data shows that the prediction hit rate of the end-point carbon content within the target error value ±0.02% is 88.9%, and the prediction hit rate of the end-point molten steel temperature within the target error value ±20°C is 92.3%, which is significantly improved, and provides a reference for end point prediction and control of converter steelmaking.

Aiming at the outstanding problems of insufficient digitization of iron making process,low degree of intelligence,lack of a unified intelligent platform,and far from adapting to the development needs of intensification,digitization and intelligence,Angang Steel Co.,Ltd. established a big data centre for intensive control of blast furnace with blast furnace group as the core and covering other processes. The big data centre breaks the information silos of each regional information system,releases the effectiveness of data. The blast furnace group and subsidiary process form a centralized control and management centre for data sharing and efficient collaboration,realizing the blast furnace process upgrading from an intelligent unit to an intelligent platform. At the same time,the intelligent application model of blast furnace is built,which realizes the visualized intelligent monitoring of the safe production and operation of blast furnace,and guides the production operation of blast furnace,as well as improves the digitization and intelligence level of the production,technology and management of the blast furnace of Angang Steel Co., Ltd.

In order to further improve the intelligent manufacturing level of production planning and scheduling at Laiwu Iron and Steel Group Yinshan Steelmaking Plant,a systematic study was conducted on the complex scheduling plan of 4 furnaces,4 machines,and non matching furnace machines. By analyzing the operation cycle and production operation mode of the steelmaking refining continuous casting process,four bottom level matching principles were explored and summarized,including the dynamic total balance principle,furnace machine matching principle,non interrupted pouring principle,and space proximity principle. Based on the four principles of matching logic,a furnace machine matching scheduling model was established,which achieved automatic scheduling of steelmaking plans for multiple furnaces to multiple machines in the non matching state of the furnace machine. This ended the production scheduling mode that was completely centered on dispatchers and driven by manual experience,greatly improving scheduling efficiency and planning accuracy. In addition,the dynamic scheduling model of the furnace machine based on the matching logic of the four principles has strong flexibility and universality,and can be applied to more steel enterprises.

Iron ore sintering is a key preliminary process in blast furnace ironmaking. Online realtime intelligent sensing of key production indicators in the sintering process is one of the key technologies to achieve green,low-consumption,and efficient development of the sintering process. However, on one hand,traditional methods for measuring sintering production indicators (such as FeO content) are time-consuming and difficult to meet the needs of real-time control. On the other hand,the sintering process data has nonlinearity,multi-source heterogeneity,and time lag,which poses a great challenge to improving modeling accuracy. To this end,this paper proposes a multi-source heterogeneous  data fusion and real-time sensing technology for the sintering industry. This study uses a multi-source heterogeneous information fusion method to extract shallow and deep features respectively through expert knowledge and the FcaNet model based on discrete cosine transform (DCT) for the cross-sectional image data of the sintering machine collected by the infrared thermal imaging camera,achieving feature level and data level fusion. In the prediction task,the two-dimensional convolution block is applied to the time series data,and FcaBlock is used as the convolution block for feature extraction, which effectively extracts the frequency component information of the time series data. On the iron ore sintering data set of a real steelmaking plant,the prediction accuracy and stability of this model are superior to existing models,and it significantly improves the online real-time sensing of key quality indicators of the sintering process.

Crane scheduling in steelmaking and continuous casting section is a typical multi-machine and multi-task constraint problem,which is of great significance to the connection and forward flow of each process and the control of the production rhythm of the whole steelmaking plant. In order to im- prove the efficiency of crane operation and ensure the stability of production,a crane scheduling mod- el based on spatio-temporal heuristic algorithm was established by analyzing the process flow and the constraints of crane job scheduling in steelmaking and continuous casting section. Partition rules,task allocation rules,collision avoidance rules and state update rules are designed to characterize the oper- ation process of the crane. The model is solved by a heuristic method,which can avoid the high com- putational complexity of traditional theoretical methods in large-scale problem solving. The results show that the model can provide a reasonable scheduling scheme and effectively avoid the spatio-tem- poral conflict during the operation of the crane. Compared with the actual production scheduling method,the performance of the crane under the heuristic algorithm model has been improved,and the operation efficiency of the crane has been improved.

To address the issues of low detection accuracy and slow speed in the inspection of slab on steel continuous casting production lines,an algorithm named DMS-YOLOv8 (YOLOv8 with depthwise separable convolution,multi-pooling,and SE-EMA) that combines machine vision,image processing,and deep learning was proposed. This algorithm replaces standard convolution with depthwise separable convolution based on reverse residual and multi-scale pooling,reducing the burden of redundant networks,decreasing memory usage,and improving computational speed. The mixed attention mechanism SE-EMA allows the model to selectively focus on and weight information from different parts when processing input data,enhancing the model's expressive power. Finally,by conducting comparative analysis and ablation experiments on the self-made slab dataset and the PASCAL VOC2012 dataset,the study validates that the proposed method can effectively enhance slab detection accuracy in real-time operating conditions while ensuring a certain level of efficiency,laying the foundation for efficient production in subsequent hot rolling processes.

According to the analysis of the functional requirements of blast furnace bottom filtration process (normally called OCP) and the structure of bridge grab-crane,the intelligent control system of grab crane was proposed. The system adopts the technology of automatic positioning of crane and grab bucket,frequency conversion speed control,anti-sway grab,wireless network communication, video surveillance,knowledge learning algorithm,etc. ,to achieve the intelligent control of the grab crane. The system divides the filter tank into grid according to the process requirements, and adopts an efficient slag grabbing strategy to automatically control the grab bucket to complete the grabbing and slag discharge operations. The practical application results of the intelligent control system of grab crane in the bottom filtration process of many blast furnaces show that the operation rate of the system is more than 99%,which is 32. 5% higher than the efficiency of manual slag grabbing,reducing staff 66. 7%,and the maintenance cycle is doubled.

In order to solve the widely existing problem of excessive head thickness deviation of the first strip after finishing roll change in hot strip mill,a hybrid feature selection method based on variance selection,mutual information,L1 /2 regularization combined with experts'experiences was proposed. The method is used to select features from historical production data of the first strip after roll change in a 1 700 hot strip mill in China,and the feature selection results are used as the training set of GA-BP neural network-based head thickness deviation prediction model for the first strip coil after finishing roll change. A series of experiments are carried out on the model,and mean absolute percentage error (MAPE),mean square error (MSE),coefficient of determination (R2) and other indicators are used as model evaluation criteria. The results show that the hybrid feature selection method  proposed in this paper has significantly improved the prediction accuracy of the model after training compared with the traditional mathematical feature selection method. Through testing on data samples of different steel grades and different thickness intervals of the main products of a steel mill,it is verified that the model has a high prediction accuracy with a certain degree of generalizability. The method has good application prospects in production practice.

Under the background of green and intelligent construction in the steel industry , traditional single-process quality control systems become insufficient to meet the demands of smart steel manufac - turing.  They need to eVolVe into full-process integrated process control systems.  These systems need to transition from  mere  model  calculations  to  collaboratiVe  Whole-process  process  model  manufacturing , production data  mining ,  and  intelligent  system  decision-making ,  self-adaptation ,  and  self-learning.  with this objectiVe in mind , the system architecture , research and deVelopment content , and anticipa- ted Vision for the deVelopment and application of intelligent process control systems and process models across  four  dimensions : intelligent  egaipment , horizontal  process  floW, Vertical  information  transfer , and time Were explored.

The hanging of blast furnace is one of the abnormal conditions which can easily occur during blast furnace smelting. Because the hanging diagnosis in actual production mainly depends on the experience of the operator,the personal subjectivity is large and the transmission is poor,and it is easy to misjudge and lead to a large fluctuation of the furnace condition. With the development of computer technology,data-driven process monitoring theory has gradually matured,and fault diagnosis based on machine learning has been widely used in complex industrial production and has achieved good diagnostic results. Therefore,this study analyzed the hanging phenomenon and its formation causes,and conducted data cleaning on the initial feature set based on expert experience. Recursive feature elimination (RFE) algorithm was used to screen valid variables and determine their importance. The gradient boosting decision tree (GBDT) was established according to the selected key indexes for real-time hanging diagnosis. The results show that the accuracy rate of GBDT classification by integrated learning voting mechanism is more than 90% ,showing strong robustness,that is,the diagnosis model has excellent performance in both accuracy and computational efficiency. Based on the diagnosis model,the intelligent diagnosis system can help the BF operators to monitor the hangings in real time,so as to ensure the stable running of the BF.

Due to the increasing requirements of current steel product quality,a stacking algorithm based on multiple machine learning models was proposed to improve on-site production while reducing slag defects caused by high-frequency data fluctuations within the continuous casting. Firstly, high-frequency feature data which extracted from the continuous casting mold is segmented into subsequences by sliding window,important features in each window are extracted from the fluctuation patterns of the high-frequency data. Secondly,the locations of the slag defects which identified by the hot-rolling surface inspector are corresponded to the window locations. Finally,a stacked classification algorithm that combines multiple machine learning models is utilized for defect prediction. Comparison experimental results show that the stacking classification model performed better than each single machine learning model and voting classification model,and the overall prediction performance and  generalization ability is better than others. Currently,this slag defect analysis model based on the stacked ensemble algorithm has been deployed on a hot-rolling production line. The comparison of online prediction results with actual data demonstrates that the occurrence and location of slag defects can be predicted in advance by the model to improve the efficiency of slag defect analysis and new approaches and insights are given for investigating the causes of slag defects.

The camber defect in the hot rolling rough rolling area affects the shape quality and rolling stability of the strip steel. The research on online detection and control technology of camber has always been a hot topic of industry. This study develops an online detection and control technology for camber based on image recognition. By installing high-definition cameras in front and behind the stand to obtain strip steel images,image processing technology is used to measure and recognize the contour shape of the strip steel. Combined with actual measurement data,equipment data,model setting data,etc. during the rolling process,a camber adjustment model is constructed,online automatic feedforward control function for camber is implemented. This technology has been put into online application on a hot rolling production line in China,with an accuracy rate of over 95% for adjusting the camber direction. It can effectively reduce production costs and labor intensity,improve the automation level of the production line,and has high application value and development prospects.

Addressing the challenges of the heavy workload involved in scrap steel classification and the lack of uniform grading standards,this paper employs machine learning to identify and compare scrap steel images for determining the scrap steel grade. A scrap steel classification dataset is constructed,and the Siamese neural network is utilized to train the dataset,selecting the optimal weights to enable the model to accurately distinguish different types of scrap steel. The similarity between the scrap steel benchmark and the image to be tested is compared using the Siamese network calculation method. Based on the similarity results,the scrap steel grade is determined. When the similarity approaches 1,the scrap steel shapes are considered similar. When the similarity approaches 0,the shapes are deemed different,allowing for the determination of the distribution of scrap steel types in the image. Experimental results demonstrate that the method of scrap steel classification using similarity comparison and the Siamese neural network exhibits excellent accuracy and reliability. Compared with traditional manual classification methods,this approach not only significantly improves classification efficiency but also achieves standardization and consistency in scrap steel grading.

Traditional mechanical properties prediction and optimization methods are mostly based on experience and mechanisms,and do not fully consider the value contained in the data. One of the current research hot spots is how to explore the linear and nonlinear transfer relationship between steel performance and related process parameters,construct high-precision performance prediction models, and achieve process optimization. Based on the high-dimensional process quality dataset of the throughout manufacturing process of hot rolled strip,a performance optimization method for hot rolled strip steel was proposed that integrated machine learning performance prediction model and Shapley additive explanation (SHAP) interpretation framework. This method first uses maximal information coefficient (MIC) metrics to select effective variables that have a significant impact on mechanical performance indicators from high-dimensional process parameters dataset. Then,by comparing the prediction accuracy of performance prediction models based on multiple output support vector regression (MSVR),support vector regression (SVR),and random forest,the optimal performance prediction model is selected. Finally,based on the SHAP interpretation framework and optimal prediction model,process parameter evaluation is conducted to measure the quantitative impact of each process parameter on the final performance,and the operational variables are adjusted according to the results of SHAP analysis to verify the effectiveness of performance optimization. The experimental results indicate that the performance optimization method proposed in this paper can significantly improve performance indicators according to demand,and has guiding significance for mechanical performance control in steel production processes.

In view of the shortcomings of the existing carbon content detection methods of continuous casting billets,the predecessors tried to establish an exponential function model for predicting the carbon content of high-carbon steel in continuous casting billets based on the corresponding relationship between the grayscale of the macrostructure image and the carbon content,but the degree of carbon segregation is more difficult to characterize for low-carbon steel. The partial segregation degree of low carbon steel is obvious,and the maximum segregation index is greater than 3. 0,so it is necessary to carry out efficient characterization of carbon content. A typical low-carbon steel continuous casting billet sample was selected and an exponential function model for carbon content prediction based on  macrostructure image grayscale was established. The R-square coefficient of the function fitting result was 0. 62,and the average relative deviation (ARD) was 29. 7% . Then,a carbon content prediction neural network model based on the color parameters of macrostructure images was established,and the ARD of the training results was 19. 5% . Finally,a comprehensive model for carbon content prediction was established based on the characteristics of the prediction results of the neural network model and the exponential function model. The ARD between the prediction results of the test set of comprehensive model and the results of electron probe detection was 14. 27% . The relative deviation between the average value of the prediction results and the average value of the electron probe detection results is 3. 43% . Compared with the commonly used carbon content detection methods,the error has basically reached the same order of magnitude,and some prediction results have lower errors than the commonly used detection methods. Since the macrostructure image and its color information acquisition process are simple to operate,the cost is low,the pixel scale can be at the micron level,and the acquisition range can be targeted at the entire continuous casting billet section or large area. This model can provide guidance and lay the foundation for the automatic fine detection and evaluation of carbon segregation in similar steel billets,which is meaningful for the corresponding automatic and digital intelligent analysis.

The crane is an important transportation tool and core execution equipment for auxiliary operations connecting the upstream and downstream processes of the steelmaking-continuous casting section. Its scheduling level directly affects the logistics transportation efficiency and production rhythm of the steelmaking-continuous casting section. This article introduced the research status and progress of crane scheduling in the steelmakingcontinuous casting section in China. Firstly,the characteristics and existing problems of crane scheduling in the steelmaking-continuous casting section were introduced and analyzed. Secondly,the research methods and advantages and disadvantages of crane scheduling in the steelmaking-continuous casting section were summarized. Finally,the currently developed crane scheduling management system for the steelmaking-continuous casting section was introduced and summarized,providing research ideas and theoretical guidance for further achieving digital management and intelligent operation of the crane scheduling task process in the steelmaking- continuous casting section,and improving the integration level of production and scheduling in steelmaking-continuous casting section.

The use of metallurgical robots to replace manual labor to engage in some dangerous and repetitive labor has gradually become one of the important elements of the construction of unmanned production lines in major factories. In some application scenarios,the robot忆s end-effector needs to be in close contact with the workpiece. When an abnormal situation occurs,the robot may trigger a collision alarm,and require the operator to manually recover the robot,greatly affecting the production efficiency and posing a large safety risk. For those applications that require contact work,based on the Rapid language of ABB series of robots,a collision adaptive control program and anti-collision monitoring along with distance compensation program were developed. Then without manual intervention, the robot can automatically carry out the next action,to ensure the normal operation of the production process. Both solutions have been put into use in the steel plant,the effect is good,and improve the level of intelligence in the steel plant.

To address the issue of unknown molten iron yield prior to tapping in blast furnaces,which leads to inefficiencies in transportation and scheduling of molten iron ladles,a prediction model about molten iron yield constructed and trained using the GA-XGBoost algorithm was proposed. After testing and comparing multiple models,the proposed method demonstrates a certain advantage in predicting molten iron yield from a multi-feature dataset,achieving an accuracy of 89. 64% within a ±10 t error range. Firstly,the missing and anomalous values in the experimental dataset is corrected. After normalization,the structured data is used for model training. The grey relational analysis method is then used to identify the key influencing factors of molten iron yield,and redundant parameters are removed based on process principles. In the end,15 feature variables are selected to form the input vector for model construction. Additionally,to quantify the impact of different operational parameters on molten iron yield,the SHAP calculation framework is employed,providing data support for regulating parameter in blast furnaces. This study achieves the prediction task of molten iron yield of blast furnace based on furnace characteristics,contributing to more efficient blast furnace regulation to improve production. Furthermore,by leveraging the prediction results,workers can pre-plan the transportation routes for the ladles,reducing heat dissipation and thus improving the cost-efficiency of blast furnace smelting.

Under the situation of no great innovation in metallurgical process,the optimization of material flow and energy flow directly is facing bottlenecks. With the rapid development of intelligent big data,through the research and development of intelligent information system,the information flow and control flow of iron and steel production are optimized to realize process reengineering,particularly, through changes in control flow,integration and optimization can be influenced across positions, units,operating areas,and even departments within the factory. Under the joint action of intelligent information system and process reengineering,a three-level technical framework of " intelligent decision-intelligent centralized control-intelligent operation at site" is established,which is practice of the new generation of intelligent manufacturing technology system HCPS in steel industry,and ensuring the intelligence and precision of the entire steel production has improved efficiency,reduced manual labor intensity,and enhanced intrinsic safety. The integrated solution of intelligent manufacturing of iron and steel has been applied in iron and steel enterprises,the number of operating areas has been reduced by 60%,operating positions has been reduced by 40% ,the coke ratio of blast furnace has been reduced by 25 kg/ t,the coal ratio has been increased by 24 kg/ t,and the output has been increased by 2.2%. The optimization of information flow and control flow,makes the optimization of material flow and energy flow,which constructs a new mode of intelligent manufacturing of steel industry.

With the rapid development of manufacturing industry,more and more data management problems are gradually exposed,such as difficulties in data aggregation,lack of unified representation of data,and difficulty in data integration,which reduce data utilization and make it difficult for enterprises to analyze data effectively. Aiming at the above problems,a data space platform for the whole life cycle of steel rolling process is designed. A six-dimensional data model is proposed to describe the metadata of steel rolling process,the data entity association network model is completed by constructing the metadata model of steel rolling process,and the tasks of transformation,storage,management,query and analysis of multi-source heterogeneous data are realized. The class of rolling process metadata is determined by data analysis of the obtained hot strip rolling process data,and a six-dimensional data model is proposed on the basis of data unified modeling technology,and a rolling process metadata structure model is constructed according to the relationship among class and data. 摇 According to the data relationship and related attributes,an ontology model is constructed to complete the data entity association network model;tracking the information of data flow,data storage,data processing and display based on metadata through data consanguinity analysis,and realizing data traceability and data query. Finally,a knowledge graph of data relations and attributes for rolling process is constructed,which is convenient for users to query the association relationships of various entities in the data space more intuitively and conveniently,manage various metadata in the database,and complete the construction of the rolling process data space.

KR desulfurization is a typical approach for hot metal pretreatment. With the increasing demand of low sulfur steel products,achieving stable control of end sulfur content in molten iron and subsequently reducing the overall cost of desulfurization processes are of paramount importance. A critical aspect of this process is the ability to predict whether the desulfurization endpoint complies with the required standards. Therefore,a modeling method combining a sample class balance processing method based on cost-sensitive strategy and Bayesian-optimized extreme gradient boosting(XG-Boost) algorithm with a binary classification analysis method for solving end sulfur prediction problem in KR desulfurization process is proposed. Firstly,the end sulfur content of the desulfurization data is processed into conforming or non-conforming two categories based on the binary classification analysis method. The category sample weights are adjusted based on the cost-sensitive strategy to alleviate the imbalance issue for constructing the feature dataset. Then,using actual production data from a steel company,the model is trained via cross-validation with cost-sensitive strategy and Bayesian-optimized XGBoost with the optimal parameters are selected based on Macro-F1 metric to form the final desulfurization conformity prediction model for the KR process,achieving the data prediction for desulfurization conformity and non-conformity targets. Experimental results comparing with support vector machine(SVM) and back propagation neural network(BPNN) prediction models show that the proposed method can effectively deal with the imbalance issue in desulfurization data,showing good practical effects in desulfurization conformity prediction.

The cooling rate of the head and tail in the inner and outer rings is higher than that in the middle of the coil,and the performance of the head and tail cannot meet the technical requirements and the cutting rate of the product is higher. In order to improve the yield of high strength steel products,U-shaped cooling control strategy is used. At the early stage of U-cooling research and development,the mathematical model did not meet the requirements of U-cooling setting for multi-varieties and multi-processes in high strength steel production,and the coiling temperature control was not accurate. According to the existing problems in production,the coiling temperature control (CTC) model is upgraded,the setting function of single steel is added,and the temperature compensation function of transition section is developed,which solves the problems such as unqualified performance of high strength steel head and tail,and realizes high efficiency and low cost production of hot rolled high strength duplex steel.

As a key index to evaluate the shape quality of medium and heavy plate,the plate crown plays an important role in determining the market competitiveness of steel companies. With the devel- opment of computer technology,intelligent control has become the most focused subject in steel in- dustry. However,artificial intelligence models of predicting plate crown still lacks interpretation and are difficult to effectively guide the practice. Based on problems above,a medium and heavy plate crown prediction model based on data mining and multi-model fusion was proposed. Firstly,rolling data was projected onto spaces which have less dimensions,which solved the multi-source heteroge- neity problem of data. Secondly, samples were expanded by using the improved synthetic minority oversampling technique ( SMOTE) method based on Bootstrap,which solved the problem of low mod- el learning performance and accuracy caused by imbalance data. Then,an integrated model was pro- posed,which used predicted value of Ridge regression model as the main value and prediction of BPneural network model as the deviation. The experimental results show that mean square error (MSE) of this model is 0. 004 mm 2 ,and absolute error within 100 of model is more than 95% . On one hand, the range of prediction error is further reduced and the prediction accuracy of the medium and heavy plate crown is improved through multi-model fusion. On the other hand,the changes of feature varia- bles affect the plate crown is quantitatively obtained by reversely deducing plate crown prediction for- mula,which not only made data reflect and effectively guide the actual production,but also improved interpretability of plate crown prediction model as well. This paper provided a strong reference for control of plate crown and useful new ideas for future development and research of related fields.

To improve the accuracy of parameter prediction for the leveling process of high-strength plate,enhance production efficiency and product quality,the particle swarm optimization ( PSO) al- gorithm was introduced to optimize the weights and biases of the BP neural network,and the PSO-BP model was constructed. Through the training and testing of 550 sets of actual production data,the re- sults show that the PSO-BP model significantly outperforms the traditional method in terms of predic- tion accuracy. In the optimization process,combining the Levenberg-Marquardt (LM) algorithm with the PSO algorithm,the optimal network weights and thresholds are systematically searched for by iter- atively updating the positions and velocities of the particles. This method effectively overcomes the problems of BP neural network that is easy to fall into local optimal solutions and slow convergence speed. The experimental results show that the root mean square error (RMSE) and error rate of the PSO-BP model on the test set are improved by 0. 075 and 5. 5% ,respectively,which indicates thatthe model possesses excellent adaptability and reliability,and also shows that the results of this re- search are of great practical significance for parameter optimization in industrial production.

Blast furnace burden distribution is one of the crucial aspects of blast furnace ironmaking.  Due to the airtight nature of the top loading equipment , it is not possible to intuitiVely and accurately obserVe the actual distribution of furnace burden in the furnace throat , such as the distribution of bur- den layers and ore-to-coke ratio.  During production , blast furnace operators typically rely on the distri- bution of cO2  , coal gas temperature , or gas floWrate in the coal gas at the furnace throat for upper ad- justments.  With the adVancement of detection technology , operators can utilize infrared cameras , ther- mal imaging , laser technology , and radar technology to perform more precise upper adjustments.  HoW- eVer , these technologies are only capable of capturing information pertaining to the  surface of the fur- nace material , and they fail to proVide a detailed understanding of the distribution information regard- ing the material layer and ore-to-coke ratio.  Based on the mathematical model of blast furnace burden distribution , the research comprehensiVely applies cutting-edge technologies such as numerical simula- tion , supercomputing , and artificial intelligence.  Through big data cloud platforms and intelligent neu- ral netWorks , an intelligent monitoring system for the distribution of the entire blast furnace burden lay- er Was constructed.  This system can achieVe online tracking and simulation of the material distribution process , assisting blast furnace operators in understanding the  distribution of furnace materials.   After the application of the system , Various economic and technical indicators of the blast furnace haVe been significantly improVed.   compared  With  the  past ,  the  monthly  output  has  increased by  an  aVerage  of 4. 7% , the  fuel  ratio  has  decreased  by 2. 37  kg/tFe ,  and  the  Vanadium  recoVery  rate  is  all  aboVe 8o% .

With the rapid development of the engineering machinery manufacturing industry,the heat treatment process of steel plates has higher requirements for plate shape. The quenching process is a key factor in determining the shape quality of heat-treated plates. Since steel plates experience both phase change stress and thermal stress during the quenching process,defects are prone to appear on quenched plates,affecting the quality of the final product shape. Therefore,there is an urgent need for a method that can identify the shape of steel plates to provide meaningful guidance for controlling the quenching shape of steel plates. To address the shape recognition problem,a quenching process shape recognition neural network based on an attention mechanism and lightweight design was proposed. Additionally,the plate shape image is processed using K-means to enhance defect characterization. Tests were conducted on the slab shape dataset,and the experimental results demonstrate that this lightweight network module has higher recognition accuracy mean and a single inference speed im- provement of 20-50 ms compared to other networks. 

Continuous annealing is a pivotal process in the production of high value-added cold-rolled strip steel. Efficiently allocating strip steel across different continuous annealing lines and determining the optimal processing sequence are essential not only for enhancing production efficiency and prod- uct quality but also for ensuring the timely delivery of final products. To address these challenges,a mixed integer programming model was established. This model accounts for the distribution and se- quencing of strip steel on parallel continuous annealing production lines,incorporating process regula- tions and production organization requirements to achieve on-time order fulfillment. Furthermore,an improved tabu search algorithm was developed to expediently solve industrial-scale problems. This al- gorithm introduces multiple neighborhoods and dynamically alternates among them to balance explo- ration and exploitation in search processes. In the small-scale instance tests,the proposed algorithm’ssolutions had an average deviation of 5. 46% from the optimal solutions. In the large-scale instance tests,the proposed algorithm’s solutions showed an average improvement of 24. 95% over the heuris- tic solutions. Based on four weeks of actual data,the test results indicate that intelligent scheduling, compared to manual scheduling,improved on-time contract delivery rates by 11. 2% and reduced the number of transitional material usages by 9. 8 times. 

Based on the 1 580 mm hot rolling production line of a certain factory,in response to the problem that traditional thickness models cannot accurately reflect actual thickness,an improved stochastic configuration network ( SCN) based strip thickness prediction model was proposed. Firstly, from the perspective of rolling mechanism, the reasons for thickness fluctuations in hotrolled strip products were analyzed. Secondly, based on the original SCN, a stochastic configuration network based on hunting prey optimization algorithm ( HPO-SCN) and a stochastic configuration network based on hunting prey optimization algorithm and orthogonal triangular decomposition ( HPO-QR-SCN) were proposed. Then,using onsite measurement devices,parameters of three different thick ness specifications of strip steel products were collected to form a database of strip thickness. Parame ters related to export thickness were selected as input values for the model,and the Pauta rule was used to preprocess the original rolling data. SCN,HPO-SCN, and HPO-QR-SCN prediction models were established,and their prediction results were compared. The experimental results show that the proposed HPO-QRS-CN thickness prediction model has the shortest prediction time and the highest accuracy,with a model determination coefficient of 0. 963 8. At the same time,based on the model with the best predictive performance,the influence of rolling force and roll gap on the exit thickness of the strip steel was tested,and the results were in line with actual physical laws. The asymptotic be havior of the model was tested,with a root mean square error ERMS (RMSE) of 0. 059 8,indicating good approximation performance. 

As a key equipment in the hot-rolled Wide and thick plate production line , it is a complex system With multidimensional factors coupled With each other.  Especially during operation , due to the strong coupling and nonlinear characteristics  of Various  influencing  factors , the  rolling  mill  operation process is in a“ black box ”state , Which makes it difficult to accurately grasp the rolling mill operation status in real time and formulate preVentiVe measures.  This has become  a bottleneck problem that re- stricts the high-precision and stable rolling of the hot-rolled Wide  and thick plate  production line.   To solVe this industry problem , data-driVen monitoring and analysis technology for the status of hot-rolled Wide  and thick plate rolling mills Was introduced.  Based on deep perception of Various data in the roll- ing process , big data algorithms are applied for data correlation analysis , and a rolling process dynam- ic model and rolling mill status online monitoring system platform With multi-source data-driVen fusion mechanism are constructed.  After being applied in releVant production lines , the accuracy of the sickle bending offset model has exceeded 95% , and the Vibration speed of the precision rolling mill has been reduced by 64 . 8%  ~72 . 8% .   It can effectiVely  identify 90%  of the rolling mill state data , effectiVely improVing the real-time accurate perception of the rolling mill state and the ability to analyze and pre- dict abnormal states.

In the automatic gauge control ( AGC) system for hot rolling,a combination of pressure AGC and monitoring AGC is often used for control. In the case of a large preset deviation,monitoring AGC is prone to causing excessive adjustment amplitude of the final stand roll gap due to its fast ad- justment and hysteresis characteristics,resulting in wave phenomenon. Based on actual rolling process conditions,a monitoring AGC control method based on multiple factors such as bending roll control and loop control was proposed. Based on the relative reduction rate of each stand,upper and lower limits of bending roll force adjustment,adaptive range of loop angle,variable monitoring proportion adjustment range,and monitoring AGC hysteresis deviation,a monitoring AGC control method for pre- cision rolling units was constructed. Practical applications have shown that the monitoring AGC con- trol strategy has achieved the expected results,increasing the speed and stability of strip head rolling, and effectively improving the overall thickness hit rate of the strip.