Quality Of Steel Products Research Articles

HDFA-Net: A high-dimensional decoupled frequency attention network for steel surface defect detection

Accurately detecting surface defects is crucial for maintaining the quality of steel products. The existing methods often struggle with identifying small defects in complex scenes. To overcome this limitation, we propose a high-dimensional decoupled frequency attention network (HDFA-Net), which features a novel HDFA module. This module considers frequency domain feature information at both the channel and spatial levels and uniquely decouples feature representations into low-frequency components and high-frequency components, conveying global contextual information and highlighting local details, respectively. This innovative approach enables the network to apply distinct attention mechanisms to each frequency domain, significantly enhancing the ability to detect small defects. The HDFA-Net architecture comprises three main subnetworks: a backbone for feature extraction, a neck for multiscale feature interaction, and a head for precise defect localization. The experimental results demonstrate that HDFA-Net outperforms the state-of-the-art defect detection methods, highlighting the effectiveness of the HDFA module in improving the detection accuracy.

An efficient re-parameterization feature pyramid network on YOLOv8 to the detection of steel surface defect

In the field of steel production, the detection of steel surface defects is one of the most important guarantees for the quality of steel production. In the process of defect detection, there are problems regarding the noise of the acquisition background, the scale of defects, and the detection speed. At present, in the face of complex steel surface defects, realizing efficient real-time steel surface defect detection has become a difficult problem. In this paper, we propose a lightweight and efficient real-time defect detection method, LDE-YOLO, based on YOLOv8. First, we propose a lightweight multi-scale feature extraction module, LighterMSMC, which not only achieves a lightweight backbone network, but also effectively guarantees the long range dependence of the features, so as to realize multi-scale feature extraction more efficiently. Secondly, we propose lightweight re-parameterized feature pyramid, DE-FPN, in which the sparse patterns of the overall features and the detailed features of the local features are efficiently captured by the DE-Block, and then efficiently fused by the PAN feature fusion structure. Finally, we propose Efficient Head, which lightens the model by group convolution while its improves the diagonal correlation of the feature maps on some specific datasets, thus enhancing the detection performance. Our proposed LDE-YOLO obtains 80.8 mAP and 75.5 FPS on NEU-DET , 80.5 mAP and 75.5 FPS on GC10-DET. It obtains 2.5 mAP and 4.7 mAP enhancement compared to the baseline model, and the detection speed is also improved by 10.4 FPS, while in terms of the number of floating point operations and parameters of the model reduced by 60.2% and 49.1%, which is sufficient to illustrate its lightweight effectiveness and realize an efficient real-time steel surface defect detection model.

Inclusion Transfer Phenomenon under Unsteady Multiphase Flow during Rheinstahl–Heraeus Vacuum Treatment Process

Inclusion particles are extremely harmful to the quality of steel products, and Rheinstahl–Heraeus (RH) vacuum treatment is the key process to remove the inclusions from molten steel. By coupling analyzing the unsteady gas–liquid flow and decarbonization reaction, the inclusion mass/population conservation model is developed to explore the inclusion particles transfer phenomenon during RH vacuum treatment process. The coupling mathematical models are validated by the metallurgical experimental data. The results show that the predicted inclusion mass concentration in unsteady CO–Ar–molten steel multiphase system has significantly smaller relative error than that in steady Ar–molten steel flow. CO bubble plays a key role in the inclusion transfer phenomenon, which prompts the inclusion transport, the collision‐aggregation, and the removal process. At the 4th minute of vacuum treatment, the average inclusion volume concentration and the average inclusion number density in CO–Ar–molten steel are 72.75% and 64.62% of those in Ar–molten steel, respectively. The average inclusion characteristic radius in CO–Ar–molten steel is 7.33% larger than that in Ar–molten steel.

A Lightweight Strip Steel Surface Defect Detection Network Based on Improved YOLOv8

Strip steel surface defect detection has become a crucial step in ensuring the quality of strip steel production. To address the issues of low detection accuracy and long detection times in strip steel surface defect detection algorithms caused by varying defect sizes and blurred images during acquisition, this paper proposes a lightweight strip steel surface defect detection network, YOLO-SDS, based on an improved YOLOv8. Firstly, StarNet is utilized to replace the backbone network of YOLOv8, achieving lightweight optimization while maintaining accuracy. Secondly, a lightweight module DWR is introduced into the neck and combined with the C2f feature extraction module to enhance the model’s multi-scale feature extraction capability. Finally, an occlusion-aware attention mechanism SEAM is incorporated into the detection head, enabling the model to better capture and process features of occluded objects, thus improving performance in complex scenarios. Experimental results on the open-source NEU-DET dataset show that the improved model reduces parameters by 34.4% compared with the original YOLOv8 algorithm while increasing average detection accuracy by 1.5%. And it shows good generalization performance on the deepPCB dataset. Compared with other defect detection models, YOLO-SDS offers significant advantages in terms of parameter count and detection speed. Additionally, ablation experiments validate the effectiveness of each module.

TAFENet: A Two-Stage Attention-Based Feature-Enhancement Network for Strip Steel Surface Defect Detection

Strip steel serves as a crucial raw material in numerous industries, including aircraft and automobile manufacturing. Surface defects in strip steel can degrade the performance, quality, and appearance of industrial steel products. Detecting surface defects in steel strip products is challenging due to the low contrast between defects and background, small defect targets, as well as significant variations in defect sizes. To address these challenges, a two-stage attention-based feature-enhancement network (TAFENet) is proposed, wherein the first-stage feature-enhancement procedure utilizes an attentional convolutional fusion module with convolution to combine all four-level features and then strengthens the features of different levels via a residual spatial-channel attention connection module (RSC). The second-stage feature-enhancement procedure combines three-level features using an attentional self-attention fusion module and then strengthens the features using a RSC attention module. Experiments on the NEU-DET and GC10-DET datasets demonstrated that the proposed method significantly improved detection accuracy, thereby confirming the effectiveness and generalization capability of the proposed method.

Periclase-magnesium aluminate spinel ceramics filter with excellent purification efficiency on molten steel prepared from Mg(OH)2: Effect of fused MgO content

Reticulated ceramic filters are widely used in the field of steel purification to improve the quality of steel products. In this work, periclase-magnesium aluminate spinel ceramic filters with microporous MgO coating were fabricated by using Mg(OH)2 as raw material. The effect of fused MgO content on the microstructures, mechanical properties and filtration capacity of filters on Al2O3 inclusions in molten steel was comparatively studied. The results indicate that the particle packing and liquid phase sintering cooperated to promote the formation of neck connections between microporous MgO particles, which increased the cold compressive strength and thermal shock resistance of the filters. Furthermore, the results from the immersion test in molten steel demonstrate that an appropriate amount of high-temperature liquid phase and the relatively microporous structure of filter skeleton were favorable for filters to filtrate Al2O3 inclusion in molten steel. Overall, the filter with 30 wt% fused MgO had the best comprehensive performance with a cold compressive strength of 0.68 MPa, a residual retention of cold compressive strength after three thermal shock recycles of 79.07 % and a filtration efficiency of 79.06 % in reducing the total oxygen content of the steel.

Lightweight strip steel defect detection algorithm based on improved YOLOv7

The precise identification of surface imperfections in steel strips is crucial for ensuring steel product quality. To address the challenges posed by the substantial model size and computational complexity in current algorithms for detecting surface defects in steel strips, this paper introduces SS-YOLO (YOLOv7 for Steel Strip), an enhanced lightweight YOLOv7 model. This method replaces the CBS module in the backbone network with a lightweight MobileNetv3 network, reducing the model size and accelerating the inference time. The D-SimSPPF module, which integrates depth separable convolution and a parameter-free attention mechanism, was specifically designed to replace the original SPPCSPC module within the YOLOv7 network, expanding the receptive field and reducing the number of network parameters. The parameter-free attention mechanism SimAM is incorporated into both the neck network and the prediction output section, enhancing the ability of the model to extract essential features of strip surface defects and improving detection accuracy. The experimental results on the NEU-DET dataset show that SS-YOLO achieves a 97% mAP50 accuracy, which is a 4.5% improvement over that of YOLOv7. Additionally, there was a 79.3% reduction in FLOPs(G) and a 20.7% decrease in params. Thus, SS-YOLO demonstrates an effective balance between detection accuracy and speed while maintaining a lightweight profile.

Predicting Alloying Element Yield in Converter Steelmaking Using t-SNE-WOA-LSTM

The performance and quality of steel products are significantly impacted by the alloying element control. The efficiency of alloy utilization in the steelmaking process was directly related to element yield. This study analyses the factors that influence the yield of elements in the steelmaking process using correlation analysis. A yield prediction model was developed using a t-distributed stochastic neighbor embedding (t-SNE) algorithm, a whale optimization algorithm (WOA), and a long short-term memory (LSTM) neural network. The t-SNE algorithm was used to reduce the dimensionality of the original data, while the WOA optimization algorithm was employed to optimize the hyperparameters of the LSTM neural network. The t-SNE-WOA-LSTM model accurately predicted the yield of Mn and Si elements with hit rates of 71.67%, 96.67%, and 99.17% and 57.50%, 89.17%, and 97.50%, respectively, falling within the error range of ±1%, ±2%, and ±3% for Mn and ±1%, ±3%, and ±5% for Si. The results demonstrate that the t-SNE-WOA-LSTM model outperforms the backpropagation (BP), LSTM, and WOA-LSTM models in terms of prediction accuracy. The model was applied to actual production in a Chinese plant. The actual performance of the industrial application is within a ±3% error range, with an accuracy of 100%. Furthermore, the elemental yield predicted by the model and then added the ferroalloys resulted in a reduction in the elemental content of the product by 0.017%. The model enables accurate prediction of alloying element yields and was effectively applied in industrial production.

Three-Dimensional Imaging of Non-metallic Inclusions in Steel Using Ionoluminescence

Obtaining information on the morphology, size, distribution, and chemical composition of non-metallic inclusions in steel helps control the steel production process and the quality of the steel products. Two-dimensional analysis is commonly used to acquire this information; however, more accurate data can be obtained through three-dimensional analysis, leading to a better control of the quality of steel products and their production process. Currently, several techniques are proposed for the three-dimensional analysis of non-metallic inclusions; however, they are time consuming. Herein, the author presented a method to rapidly obtain three-dimensional images of non-metallic inclusions in steel using ionoluminescence (IL). A three-dimensional image of MgO·Al2O3 spinel inclusions was constructed based on two-dimensional IL images obtained every 10 min during argon–ion bombardment. The proposed IL imaging can cover an oval-shaped area of 1.17 mm × 0.26 mm on the semi-major and in the semi-minor axes, respectively, at a single measurement. Three-dimension images of MgO·Al2O3 spinel inclusions with sizes more than 20 μm can be obtained within 4 hours. Therefore, the IL imaging proposed here can provide a precise and rapid account of the effects of non-metallic inclusions on steel products and the steel production process.

ОПТИМІЗАЦІЯ КИСНЕВО-КОНВЕРТЕРНОГО ПРОЦЕСУ ЗАСОБАМИ АЛГОРИТМІВ МАШИННОГО НАВЧАННЯ

Traditional methods for optimising the BOF process in steelmaking are based on experience and rules of thumb, but do not always achieve maximum efficiency. As part of the Industry 4.0 concept, the use of machine learning algorithms is becoming a promising way to optimise production processes. They allow you to analyse large amounts of data collected by a variety of different sensors that operate during an industrial process and find complex relationships between parameters to achieve the best results. The created models can work automatically, analysing data in real time and responding to changes in the process. The use of machine learning algorithms can improve the accuracy of results and the quality of steel production by optimising the BOF process with a large number of parameters and their interrelationships.

Impact of functional coatings on microstructure and properties of periclase-magnesium aluminate spinel ceramic filter and its purification capacity on molten steel

Reticulated ceramic filters play an important role in the reduction of non-metallic inclusions in molten steel, which can improve the quality of steel products. In this work, periclase-magnesium aluminate spinel ceramic filters with Al2O3 and TiO2 functional coatings were fabricated by using microporous MgO powder, respectively. The effects of the functional coatings on their microstructure, properties and purification capacity on molten steel were studied. The filter with Al2O3 functional coating exhibited the best mechanical properties owing to the good rheological properties of the coating slurry, as well as the formation of in-situ spinel layer connection between the Al2O3 functional coating and the microporous MgO layer of the filter substrate. Compared with the filters without functional coating or with TiO2 functional coating, the filter with Al2O3 functional coating made more Al2O3 inclusions deposit on the filter surface by physical adsorption due to the same chemical composition between the functional coating and inclusions, which significantly improved the adsorption capacity of filter for Al2O3 inclusions in molten steel. Overall, the filter with Al2O3 functional coating had the highest cold compressive strength of 1.4 MPa, a residual retention of cold compressive strength after thermal shock of 65.71 % and an efficiency of 36.56 % in reducing the total oxygen content of the steel.

Stacking ensemble learning fused with whale optimisation algorithm based method for crown prediction of hot-rolled strip

Precision of strip shape is one of the important indexes to measure the quality of strip steel in hot rolling. Accurate prediction of strip shape can realise the timely adjustment of the production system, which is the important foundation for ensuring the quality and stable production of hot-rolled strip steel products. Given the poor accuracy of the traditional mechanism model, this paper proposes a stacking ensemble learning for crown prediction that consist of the random forest (RF), extreme gradient boosting (XGBoost), light gradient boosting machine (LightGBM), categorical boosting (CatBoost) and linear regression, whose hyperparameters are optimised by whale optimisation algorithm (WOA) Firstly, the key features are selected by analysing the feature importance and the Pearson correlation coefficient between rolling parameters. Then, the hyperparameters of the stacking ensemble learning are optimised by four meta-heuristic algorithms during modeling process, among which WOA has the best optimisation effect. Compared to RF, XGBoost, LightGBM, CatBoost and the stacking, proposed method has the highest prediction accuracy that the hit rate reaches 98.58% in the range of 5.0 μm of crown deviation. In addition, the effect of strip width, bending force, rolling force and roll shifting on crown are analysed based on Shapley additive explanations (SHAP).

특수강자재 구매요인이 만족도 및 행동의도에 미치는 영향

This study was conducted to investigate the mediating effect of satisfaction in the relationship between specialty steel purchase factors and behavioral intentions. As a result of statistically analyzing data collected through a survey, it was found that awareness and responsiveness among the factors for purchasing special steel materials had a significant effect on behavioral intentions by mediating immersion. In addition, among the factors for purchasing special steel materials, product quality, supply price, safety, and responsiveness were found to have a significant effect on behavior by mediating customer satisfaction. Through this result, the researcher suggests practical implications.

A lightweight image-level segmentation method for steel surface defects based on cross-layer feature fusion

Steel is widely used in the aerospace, machinery and automotive industries. Surface defects not only have a negative impact on the appearance of steel but also significantly reduce its wear resistance, high temperature resistance, corrosion resistance and fatigue strength. Therefore, the detection of steel surface defects is very important to improve the quality of steel production. The limited availability of surface defect samples in the industrial sector poses significant challenges for the accurate detection of defects in high-quality materials. In addition, the existing defect detection model is highly complex and not easy to deploy. To solve this problem, a lightweight defect detection network suitable for steel defects is proposed. The cross-layer feature fusion (CFF) in the design enables effective utilisation of multi-layer semantic features, facilitating the detection of small defects in steel. Secondly, a new loss function is designed to make up for the problems of small data volume and uneven data distribution. The experimental results demonstrate that the steel surface defect detection method proposed in this paper achieves the highest detection performance on widely used public datasets such as RSDDS, NEUS and NRSD-CR(test), while maintaining the lowest model complexity.

Development and Validation of Computational Fluid Dynamics Model of Ladle Furnace with Electromagnetic Stirring System.

Numerical methods are crucial to supporting the development of new technology in different industries, especially steelmaking, where many phenomena cannot be directly measured or observed under industrial conditions. As a result, further designing and optimizing steelmaking equipment and technology are not easy tasks. At the same time, numerical approaches enable modeling of various phenomena controlling material behavior and, thus, understanding the physics behind the processes occurring in different metallurgical devices. With this, it is possible to design and develop new technological solutions that improve the quality of steel products and minimize the negative impact on the environment. However, the usage of numerical approaches without proper validation can lead to misleading results and conclusions. Therefore, in this paper, the authors focus on the development of the CFD-based (computational fluid dynamics) approach to investigate the liquid steel flow inside one metallurgical device, namely a ladle furnace combined with an EMS (electromagnetic stirring) system. First, a numerical simulation of electromagnetic stirring in a scaled mercury model of a ladle furnace was carried out. The numerical results, such as stirring speed and turbulent kinetic energy, were compared with measurements in the mercury model. It was found that the results of the transient multiphase CFD model achieve good agreement with the measurements, but a free surface should be included in the CFD model to simulate the instability of the flow pattern in the mercury model. Based on the developed model, a full-scale industrial ladle furnace with electromagnetic stirring was also simulated and presented. This research confirms that such a coupled model can be used to design new types of EMS devices that improve molten steel flow in metallurgical equipment.

Digital-Twin-Based Monitoring System for Slab Production Process

The growing demand for high-quality steel across various industries has led to an increasing need for superior-grade steel. The quality of slab ingots is a pivotal factor influencing the final quality of steel production. However, the current level of intelligence in the steelmaking industry’s processes is relatively insufficient. Consequently, slab ingot quality inspection is characterized by high-temperature risks and imprecision. The positional accuracy of quality detection is inadequate, and the precise quantification of slab ingot production and quality remains challenging. This paper proposes a digital twin (DT)-based monitoring system for the slab ingot production process that integrates DT technology with slab ingot process detection. A neural network is introduced for defect identification to ensure precise defect localization and efficient recognition. Concurrently, environmental production factors are considered, leading to the introduction of a defect prediction module. The effectiveness of this system is validated through experimental verification.

Fitting optimization of steel product quality based on cubic splines

Fitting optimization of steel product quality based on cubic splines

To the question of improving the wear resistance of high manganese steel castings

Nowadays, one of the most significant tasks in production of equipment for mining industry is import substitution. In this regard, investigations aimed at improving the technologies of manufacturing high quality castings from wear resistance manganese austenitic steels seems to be relevant. The results of metallographic research of high manganese steel samples from foundries producing crushing plates and crushing cones, track links, hammers, shovel teeth, tooth caps etc. are given. Analysis of the microstructure of the presented castings showed that they are characterized by a coarse-grained structure with an austenite grain size of 1–2 points. In most cases, a thickened grain boundary, presence of manganese carbides along the grain boundaries and non-metallic inclusions are observed. The revealed mi-crostructure features of castings are one of the main indicators of low wear resistance of high manganese steel parts. The operational and technological factors affecting the wear resistance of castings are listed. Recommendations for improving the quality of steel products made from manganese steels are proposed, and the results of using original complex modifiers developed at NPP Technology Company in real manufacturing conditions are presented.

Improved back propagation neural network method for predicting sulfur content in hot metal

Blast furnace smelting is a traditional iron-making process. Its product, hot metal, is an important raw material for the production of steel. Steelmaking efficiency can be improved and steel product quality can be stabilized by using proper hot metal. Sulfur is an important indicator reflecting the quality of hot metal, it is necessary to establish an accurate prediction model to predict the sulfur content of hot metal, to effectively guide the production process. There is a non-linear relationship among the factors influencing the desulfurization effect during the blast furnace smelting process, and the back propagation neural network (BPNN) model has a strong ability to solve nonlinear problems. However, BPNN has the disadvantages of slow convergence speed and easy to fall into local minima. To improve the prediction accuracy, an improved algorithm combining Kmeans and BPNN is proposed in this paper. The study showed that compared with the BPNN model and case-based reasoning (CBR) model, the Kmeans-BPNN model has the lowest RMSE and MAPE values, which indicates a high degree of fit and a low degree of dispersion. The Kmeans-BPNN model has the largest HR value, which indicates the highest prediction accuracy. The proposed Kmeans-BPNN prediction model achieves a hit rate of 96%, which is 4.5% higher than before the improvement. It can effectively improve the prediction accuracy of hot metal sulfur content.

Cascaded adaptive global localisation network for steel defect detection

Defect detection is crucial in ensuring the quality of steel products. This paper proposes a novel deep neural network, cascaded adaptive global location network (CAGLNet), for detecting steel surface defects. The main objective of this study is to address the challenges associated with the irregular shape and dense spatial distribution of defects on steel. To achieve this goal, CAGLNet integrates a feature extraction network that combines residual and feature pyramid networks, a cascade adaptive tree-structure region proposal network (CAT-RPN) that eliminates the need for prior knowledge, and a global localisation regression for steel defect detection. This paper evaluates the effectiveness of CAGLNet on the NEU-DET dataset and demonstrates that the proposed model achieves an average accuracy of 85.40% with a fast frames per second of 10.06, outperforming those state-of-the-art methods. These results suggest that CAGLNet has the potential to significantly improve the effectiveness of defect detection in industrial production processes, leading to increased production yield and cost savings. Abbreviations: AT-RPN, adaptive tree-structure region proposal network; CAGLNet, cascaded adaptive global location network; CAT-RPN, cascade adaptive tree-structure region proposal network; CNN, convolutional neural network; DNN, deep neural network; EPNet, edge proposal network; FPN, feature pyramid network; FCOS, fully convolutional one-stage detector; FPS, frames per second; GMM, Gaussian mixture model; IoU, intersection-over-union; ROIAlign, region of interest align; RPN, region proposal network; ResNet, residual network; ResNet50_FPN, residual network and feature pyramid network; SABL, side aware boundary localisation; SSD, single-shot multiBox detector; TPE, Tree-structured Parzen estimator