Application of a new feature extraction and optimization method to surface defect recognition of cold rolled strips

Considering that the surface defects of cold rolled strips are hard to be recognized by human eyes under high-speed circumstances, an automatic recognition technique was discussed. Spectrum images of defects can be got by fast Fourier transform （FFF） and sum of valid pixels （SVP）, and its optimized center region, which concentrates nearly all energies, are extracted as an original feature set. Using genetic algorithm to optimize the feature set, an optimized feature set with 51 features can be achieved. Using the optimized feature set as an input vector of neural networks, the recognition effects of LVQ neural networks have been studied. Experiment results show that the new method can get a higher classification rate and can settle the automatic recognition problem of surface defects on cold rolled strips ideally.

Application of multi-scale feature extraction to surface defect classification of hot-rolled steels

Feature extraction is essential to the classification of surface defect images. The defects of hot-rolled steels distribute in different directions. Therefore, the methods of multi-scale geometric analysis （MGA） were employed to decompose the image into several directional subba^ds at several scales. Then, the statistical features of each subband were calculated to produce a high-dimensional feature vector, which was reduced to a lower-dimensional vector by graph embedding algorithms. Finally, support vector machine （SVM） was used for defect classification. The multi-scale feature extraction method was implemented via curvelet transform and kernel locality preserving projections （KLPP）. Experiment results show that the proposed method is effective for classifying the surface defects of hot-rolled steels and the total classification rate is up to 97.33%.

Stream Surface Strip Element Method and Simulation of Three-Dimensional Deformation of Continuous Hot Rolled Strip

A new method,the stream surface strip element method,for simulating the three-dimensional deformation of plate and strip rolling process was proposed.The rolling deformation zone was divided into a number of stream surface(curved surface)strip elements along metal flow traces,and the stream surface strip elements were mapped into the corresponding plane strip elements for analysis and computation.The longitudinal distributions of the lateral displacement and the altitudinal displacement of metal were respectively constructed to be a quartic curve and a quadratic curve,of which the lateral distributions were expressed as the third-power spline function,and the altitudinal distributions were fitted in the quadratic curve.From the flow theory of plastic mechanics,the mathematical models of the three-dimensional deformations and stresses of the deformation zone were constructed.Compared with the streamline strip element method proposed by the first author of this paper,the stream surface strip element method takes into account the uneven distributions of stresses and deformations along altitudinal direction,and realizes the precise three-dimensional analysis and computation.The simulation example of continuous hot rolled strip indicates that the method and the model accord with facts and provide a new reliable engineering-computation method for the three-dimensional mechanics simulation of plate and strip rolling process.

Application of automatic surface inspection system in cold-rolled strip production

The detection and classification of real-time surface defects play an important role in automotive sheet inspection and production. In this paper, an automatic surface inspection system （ASIS） based on signal processing in Baosteel NO. 4 cold-rolled plant is briefly presented. We demonstrate that the strip surface defect properties such as image, type, pitch, and position can be accurately calculated and classified by the automatic surface inspection system. In the manufacturing of the high-quality cold-rolled strips, it is necessary that the real-time surface defects can be detected and transferred by the automatic surface inspection system combined with annealing lines and recoiling lines.

“Quartus scale” of hot-rolled strip steels and its formation mechanism

The types and growth of various oxide scales formed during the different phases of the production of hotrolled strip steel products are reviewed. Similarities and differences between the ＂tertiary scale＂ on the surface of carbon steels at high temperatures and the oxide scale on pure iron are compared. The micro-structural features of the ＂final oxide scale＂ on the surface of strip steels at room temperature as well as the relationship between these features and the position of the steel coil （plate） and the subsequent processes of recoiling, temper rolling and trimming, etc. are summarized. The actual oxide scales retained on the commercial hot-rolled strip steels at room temperature have been proposed to define as ＂ quartus scale＂ for the first time. The micro-structural development and phase transformation of the initial ＂tertiary scale＂ during and after cooling and coiling are described. The reasons for the ＂tertiary scale＂ on carbon steels differing from the oxide scale formed on pure iron, and the major influencing factors in the formation of various types of ＂quartus scales＂ are analyzed from both thermodynamic and dynamic viewpoints. The development mechanism of ＂ quartus scales＂ is discussed and the potential effects of the ＂ quartus scale＂ state （thickness, constitution, structure and defects）, on the rusting and pickling properties of commercial hot-rolled strip steel, as well as on the mechanical properties of oxide scales are analyzed.

Complete Mill Simulation of the Rolling Process of 1660 mm Hot Strip Continuous Mills

The three-dimensional plastic deformations of strip are analyzed using the stream surface strip element method, the elastic deformations of rolls are analyzed using the influence coefficient method, the analyzing and computing model of shape and crown of 4-high mill was established by combining them, and the rolling process of 1660 mm hot strip continuous mills was simulated. The simulated results tally well with the experimental results. The modei and the method for simulation of shape analysis and control of hot strip mills were provided.

面向热轧带钢表面缺陷检测的YOLOv5算法优化分析

单阶段目标检测网络YOLOv5在处理热轧带钢表面缺陷的特征提取与感受特征融合时存在一定不足。文章提出一种适用于热轧带钢表面缺陷检测的优化YOLOv5算法,该算法通过IOUK-means++算法调整Anchor聚类锚框设定,并增加Dynamic Head目标检测头,引入通道注意力机制(C3_CA),同时结合Hard Swish激活函数与WIoU_Loss边界框回归函数,有效提高热轧带钢表面缺陷检测的综合精度。由NEU-DET数据集测试结果表明,相较于单阶段YOLOv5算法融合结果,优化后的YOLOv5网络模型的均值平均精度(mAP)可提高至75.7%,且网络约束率可有效提升6.1%。上述优化YOLOv5算法对热轧带钢表面缺陷位置勘定、分类指向与影响评估具有有益参考,同时也为金属表面的高精度筛检提供重要支持。

基于可变形卷积和注意力的带钢瑕疵识别方法

带钢生产中瑕疵检测的准确性对于其质量的保证有着至关重要的意义。本文针对带钢表面瑕疵检测中的瑕疵种类复杂、背景干扰严重、瑕疵样式间面积形态差异较大等问题,提出了一种基于ResNet50改进的结合多尺度变形卷积和注意力的带钢表面瑕疵分类识别方法。首先通过细化ResNet50中BottleNeck结构的卷积块为一组多尺度卷积以扩大感受野,然后引入可变形卷积代替组中的卷积核,使网络在训练中捕捉不同形态尺度的瑕疵特征。最后在网络中引入增强注意力模块,使网络可以关注到通道与空间之间的信息,将其联合起来从而关注到更重要的通道和空间位置。通过对比实验表明,本文提出的方法在带钢的表面瑕疵分类识别上精确度优于现有方法,可以应用于企业实际的带钢的工业生产中。

基于知识图谱推理的热轧带钢产品质量缺陷追溯

热轧带钢生产面临多工况、机理复杂、工艺参数繁多等问题,造成专家很难及时有效给出生产中导致质量缺陷的原因。由此提出一种基于知识图谱推理的质量缺陷追溯方法。首先通过可解释方法SHAP实现对随机森林模型预测结果的解释,并通过知识图谱将数据挖掘结果与工艺机理、专家经验等知识进行融合,进一步将图谱中表示工艺参数与质量参数依赖关系的子图映射到贝叶斯网络,推断不同工艺参数导致产品质量缺陷的后验概率。实际生产数据验证表明,针对不同生产工况,该方法能有效识别各个批次中导致质量缺陷的工艺参数,表现出良好识别率。

基于深度学习的钢板表面缺陷识别方法及系统开发

在热轧钢板的生产过程中,由于工况的多样性,钢板表面难免会形成不同类型的缺陷。针对传统方法对钢板表面缺陷的识别准确率和效率较低的问题,提出了一种结合残差网络和迁移学习的钢板表面缺陷识别方法。首先,构建基于迁移学习的神经网络框架,同时采用全局平均池化和Dropout等优化策略增加网络的泛化能力,提高分类精度。其次,预处理钢板表面缺陷数据集,并通过钢板表面缺陷识别实验对模型性能进行检验。实验结果表明,基于迁移学习的网络模型在测试集上的识别准确率可以达到98.33%。最后,设计并开发了钢板表面缺陷识别系统,采用Power Designer软件构建系统的物理数据模型,并完成系统架构的设计,将训练好的神经网络模型应用到系统中,有效提高了钢板缺陷分类识别的智能化和信息化水平。

不同除鳞工艺对热轧带钢热镀锌质量的影响

目的研究酸洗、湿法抛丸技术(Eco-Pickled Surface,EPS)、EPS+纤维刷工艺(EPS+plus)和干抛丸4种不同除鳞工艺对Q235B热轧带钢热镀锌质量的影响机理和规律,并将EPS和EPS+plus工艺处理后的热轧带钢在带钢热镀锌生产线上试制,实现大规模生产。方法采用粗糙度轮廓仪测量4种表面除鳞处理后的Q235B钢板表面粗糙度,利用扫描电镜分析表面形貌及氧化皮残留情况。通过实验室热镀锌实验,观察镀锌层表面形貌,利用场发射电子探针(EF-EPMA)分析锌层截面形貌及成分面分布;通过折弯实验评定锌层结合力;去除表面镀锌层,计算锌消耗量,观察脱锌后表面抑制层形貌,并用能谱对其进行成分分析。将Q235B热轧带钢在鞍钢金固EPS生产线上进行EPS和EPS+plus处理后,分别在热镀锌生产线进行镀锌和镀锌铝镁工业试制,并对生产试制卷进行力学、折弯及杯突检测,以评定质量。结果酸洗、EPS、EPS+plus和干抛丸4种工艺处理后钢板表面粗糙度分别为1.4442、3.3305、2.3964、4.3388μm。EPS处理试样表面未发现有氧化铁皮的残留,EPS+plus处理试样仅有极少量的氧化铁皮残留。EPS和EPS+plus工艺处理钢板热镀锌试样基体与锌层之间均能形成连续较致密的抑制层,锌层结合力评为1级,且锌消耗量较小,分别为148.3、123.5g/m^(2)。结论EPS和EPS+plus工艺处理后钢板镀锌质量优于酸洗和干抛丸工艺,EPS和EPS+plus工艺处理后在热镀锌生产线上试制的实验卷满足镀锌质量要求。

带钢热轧过程中残余应力的数值模拟

带钢内部纵向残余应力沿宽度分布不均会造成平坦度缺陷,严重影响带钢的板形质量在传统的残余应力解析计算方法中,大都忽略金属横向流动的作用为了解决这一问题,采用有限元方法,基于某1780mm热轧生产线,建立了三维弹塑性有限元计算模型,充分考虑金属横向流动对残余应力的影响,探究了残余应力在机架间的演变以及不同比例凸度变化下带钢的残余应力状态结果表明:残余应力在机架间发生反转金属横向流动会减小纵向应变差,从而达到减小纵向残余压应力的目的,当比例凸度增加至05%~115%时,残余压应力最小.

热轧酸洗板边部麻面缺陷的形成原因及控制策略研究

针对首钢某1580 mm热轧生产线生产酸洗板时出现边部麻面后,影响了后续酸洗质量的问题,对边部麻面缺陷的形成原因及机理进行了研究。根据缺陷的形态、位置,对比卷取机设备的控制时序,确定了导致边部麻面缺陷的设备,并对上下表面的边部麻面缺陷的不同点进行了研究。结果表明下夹送辊辊面状况对边部麻面缺陷几乎没有影响,上表面边部麻面缺陷产生的主要原因是上夹送辊边部锈蚀严重,出现麻坑而导致的。而下表面麻面缺陷产生的主要原因是压紧辊下的机架辊边部锈蚀严重,出现麻坑而导致。为此,提出了相关的改进措施,以有效控制边部麻面缺陷,改善酸洗板边部质量。

热轧Q235B钢翘皮缺陷特征及形成机制研究

翘皮缺陷是热轧Q235B带钢常见的典型表面缺陷之一,具有分布无规律性且深度较浅等特征,严重影响热轧带钢产品的表面质量。采用金相显微镜、扫描电镜和能谱仪相结合的方法对热轧Q235B带钢翘皮缺陷成因进行了研究分析,并提出了改进措施。结果表明,产生翘皮缺陷的主要原因包括钢坯表面产生裂纹,裂纹处氧化铁皮未除净,Si及Mn的二次氧化物与氧化铁皮在后续轧制过程中被带入钢板表面,部分结晶器卷渣进入铸坯表面。通过采取保护浇铸、优化浸入式水口插入深度、稳定拉坯速度及控制轧制过程中高压除鳞时的表面洁净等措施,可有效降低翘皮缺陷发生率。

冷轧厂薄带钢高效生产的问题及解决方案分析

热镀锌板具有抗腐蚀强、易于深加工、使用寿命长、成本低以及表面美观等特点,主要用于家用电器、汽车等产品的内部构件和面板应用。随着家电板、汽车板工业的发展,家电和汽车所用的镀锌板占全国钢产量的比例逐步提升。随着国内外家电板、汽车板轻质化和高质量的要求,客户对产品表面质量要求越来越严格。针对生产线在提速增产过程中出现的入口跑偏、工艺段带钢表面质量缺陷和薄带钢薄锌层产速受限问题,提出了相应的改进措施。优化后,不仅提高了产量,而且提高了产品质量。

一种基于CNN的带钢表面缺陷识别方法

带钢表面质量是衡量产品性能的重要指标,准确识别带钢表面缺陷是带钢生产过程的关键一环,剪除缺陷带钢对于提升带钢成材率具有重要意义。为提升带钢表面缺陷识别的准确率,构建了基于CNN的带钢表面缺陷识别模型,通过多个卷积层提取图像特征,从而自动识别缺陷类别,实现了端到端的带钢表面缺陷识别过程。实验结果表明,CNN模型对于带钢表面缺陷识别准确率达到了96.5%,识别一张图片时间仅为1.5 ms,基本满足了带钢缺陷识别要求。

基于决策级融合的曳引钢带表面缺陷检测方法

由于电梯曳引钢带的工况环境恶劣,钢带表面容易出现缺陷,而复杂昏暗的环境导致缺陷识别和定位困难。为此,提出了基于决策级融合的曳引钢带缺陷检测方法,利用迁移学习的方法微调YOLOv4及SSD的预训练模型,并将其应用于曳引钢带缺陷检测,得到不同的原始检测结果,利用所提方法在决策级融合YOLOv4及SSD模型的原始检测结果,以此提高钢带表面缺陷识别和定位的准确率。利用东北大学钢带表面缺陷公开数据集对所提方法的有效性和可靠性进行测试,缺陷识别和定位结果表明所提方法能够充分利用YOLOv4和SSD模型的原始检测结果,通过决策级融合初始检测结果取得较高的缺陷识别和定位准确性,准确率和召回率提高约15%,交并比达到0.6左右,这对于电梯曳引钢带的缺陷检测和维修保养有重要意义,并进一步保证电梯能够安全可靠运行。

热轧线材表面缺陷的图像纹理分析及其缺陷机制探究

热轧线材是金属材料加工中广泛应用的一种形式,其表面质量直接影响最终产品的性能和质量。针对常见的热轧线材表面缺陷,明确不同类型缺陷的特征及其对产品性能的影响。在热轧线材表面缺陷分析中采用的图像处理技术和算法,揭示工艺参数对表面缺陷的影响以及纹理特征与缺陷形成的关联性,为热轧线材表面质量控制提供理论支持。

铜镍锡合金表面水印缺陷的形成机制及控制措施研究

水印缺陷是铜镍锡合金在热处理及热轧生产时经常遇到的表面技术性难题,严重影响产品质量。本文利用光学显微镜、表面粗糙度仪、扫描电镜等设备对热处理热轧后的铜镍锡合金表面水印缺陷进行了分析。结果显示,水印缺陷处的粗糙度曲线显示异常,分布较多形状不规则的凸起啄点和微裂纹,热轧纹理紊乱。这是由于热处理的淬火油润滑性能不足,导致变形区局部润滑失效,轧辊与铜镍锡合金新生表面直接接触产生粘着,残留的浓缩乳化液吸附在缺陷中不易被吹扫清洗干净,经退火碳化后在视觉上表现为平行轧向的白条纹水印缺陷。通过研制缓蚀性和润滑效果更优异的新型乳化液,可解决热处理及热轧时铜镍锡合金表面水印缺陷问题。

热轧带钢麻面产生原因分析及控制措施

针对供冷轧用热轧带钢表面出现麻坑、麻面问题,文中结合德龙钢铁有限公司1250mm热轧生产线实际生产状况和机组情况,通过修订工艺制度、提高轧辊冷却水压力、启用机架带钢冷却水、调整水嘴型号和角度、改善轧辊材质等措施,有效的控制和减少了热轧低碳带钢的麻坑缺陷。