首页 | 本学科首页   官方微博 | 高级检索  
相似文献
 共查询到19条相似文献,搜索用时 357 毫秒
1.
利用二维搜索计算法,通过对带钢表面缺陷位置的一列低精度检测数据分析,精确确定轧件表面缺陷的最小间距,从而快速找出连轧机缺陷轧辊的位置。  相似文献   

2.
通过对连铸坯和热轧板表面气孔缺陷影响因素的分析,确定了由上水口、氩封和浸入式水口3个位置吹入的氩气不能有效上浮而被凝固坯壳捕捉形成了连铸坯表面气孔缺陷。气孔基本分布在宽面靠近窄边位置,而连铸坯宽面中心和连铸坯宽面1/4位置气孔数量很少。采用不吹氩气的生产工艺可以消除连铸坯和热轧板表面的气孔缺陷。  相似文献   

3.
近20年来,为满足用户对表面质量的要求,基于机器视觉的钢铁表面缺陷识别系统得到了广泛的应用并取得了良好的效果。本文在机器视觉技术的基础上,利用层次聚类和K-means等聚类技术,设计了钢卷表面缺陷分布特征识别算法。本算法通过对钢卷表面进行网格化划分,根据缺陷位置,识别钢卷表面缺陷在头部、尾部、中心线等位置集中的特征。经实验证明,该算法可有效识别钢卷表面缺陷分布特征。  相似文献   

4.
陈沅  郭利宏 《包钢科技》2010,36(z1):21-24
包钢为了提高250 km/h钢轨的合格率,通过对影响钢轨合格率的主要因素进行分析,确定是轧疤、轧痕等表面缺陷,通过现场跟踪,判断出250 km/h钢轨表面轧疤、轧痕缺陷产生的位置,针对不同位置采取不同措施,同时,将这些轧疤、轧痕缺陷进行归纳总结,纳入生产管理,从而有效地控制了轧疤、轧痕缺陷,因而,随着表面缺陷的减少,250 km/h钢轨的合格率得到了大幅度提升,从而保证了250 km/h钢轨满足高速铁路用轨的要求。  相似文献   

5.
应用金相显微镜、扫描电镜及能谱仪对镀锌板表面亮点缺陷和对应基板位置的宏观、微观形貌以及缺陷区域化学成分进行分析,发现:亮点区域凹于周围表面,化学成分不存在异常;去除表面锌层后,亮点缺陷对应的基板处存在缺陷,缺陷区域存在凸起,大小与镀层表面亮点缺陷一致,化学成分无异常。因而认为镀锌层表面缺陷是由于基板本身存在缺陷而产生的。  相似文献   

6.
为了准确掌握冷轧钢卷经过各工序时的缺陷位置,查找缺陷产生的机组和原因,研发了钢卷缺陷位置判定技术。通过从后往前推算钢卷缺陷所在的表面位置、长度位置、宽度位置,定位缺陷发生的机组和具体位置。该技术的应用为分析缺陷产生原因、降低钢卷缺陷发生率、减少批次质量事故起到重要辅助作用。  相似文献   

7.
采用SEM、EDS等试验手段,研究了冷轧镀锌板典型的表面质量缺陷形貌及缺陷部位元素分布特征。结果表明:油脂类物质残留可导致冷轧镀锌板表面漏镀缺陷,氧化铁颗粒压入引起表面黑点缺陷,机械划伤可造成表面线状缺陷,保护渣卷入致使镀锌板表面出现孔洞缺陷。阐述了镀锌板各类表面缺陷的形成机理,并提出了相应缺陷的控制处理措施。  相似文献   

8.
目前钢轨表面缺陷的检查主要依靠人工目视觉检测的方式,但这种方式存在易疲劳、检测速度慢和检出率低等缺点。钢轨在线表面缺陷检测系统,通过高速、超高分辨率数字图像采集技术,获得高清晰钢轨表面图像;运用图像处理和检测技术,准确定位缺陷位置;利用现代人工智能技术,计算缺陷几何和灰度特征,对缺陷进行分类识别。  相似文献   

9.
介绍了IF带钢退火后表面出现的条状白斑缺陷、位置分布、缺陷部位,对缺陷的微观形貌和成分进行了分析。结合各工序跟踪及工艺条件状况,认为IF带钢表面白斑缺陷产生源头为热轧工序的氧化铁皮压入,确定了该缺陷的产生原因;提出增开机架间冷却水的改进措施,杜绝了IF钢白斑缺陷在生产过程中的出现,提高了产品质量。  相似文献   

10.
本文从缺陷的形状、颜色、触感等多个角度入手,按照热轧、酸连扎、退火的顺序详细介绍了热轧后、酸轧后、退火后柳叶状氧化铁皮的缺陷特征;为了确保缺陷形貌演变准确,利用表面检测设备对柳叶状铁皮进行各个工序的追踪,利用电镜设备对各个位置的铁皮进行光谱成分分析,精准推断出了柳叶状氧化铁皮缺陷的形貌演变。使用显微镜观察铁皮缺陷的微观形貌,从形成机理上确定铁皮缺陷的产生原因,并依据以往的生产经验从热轧和酸连轧工序入手确定柳叶状铁皮的控制方法。  相似文献   

11.
针对生产过程中出现的连铸钢坯表面裂纹缺陷,采用图像检测与小波分析方法模仿人工视觉对连铸钢坏表面图像信号进行模式识别检测。  相似文献   

12.
对超声波反射和透射的基本原理进行分析,结合焊缝缺陷特征和反射波的特征,归纳焊缝缺陷产生的原因和预防的有效方法。  相似文献   

13.
涡流探伤在钢管生产中的应用   总被引:1,自引:0,他引:1  
王珍 《包钢科技》2005,31(Z1):24-25
介绍了钢管涡流探伤的原理、特点;涡流探伤方法对钢管缺陷的检测及对检出缺陷的处理方法,影响涡流探伤的因素及涡流检测的局限性。  相似文献   

14.
This paper presents both numerical and experimental studies of the transient response of concrete I-girders with and without flaws subjected to transverse elastic point impact. These studies were aimed at evaluating the capability of using the impact-echo method for integrity testing of concrete I-girders. Numerical and experimental studies were performed first to gain an understanding of transient stress wave propagation in solid I-girder segments. Experimental studies were carried out on full-size reinforced-concrete I-girder segments containing flaws at known locations. Results obtained from these studies show that the transient response of a solid I-girder subjected to impact on the bottom surface is composed of a number of resonant frequencies caused by cross-sectional modes of vibration. Among these frequencies there is a predominant frequency, and its value depends on the cross-sectional geometry and dimensions of the bottom portion of the I-girder for a given P-wave speed in concrete. It is also shown that the presence of a flaw disrupts the modes of vibration. A shift of the predominant frequency to a lower value is a key indication of the presence of the flaw. In addition, multiple wave reflections between the impact surface and the surface of the flaw produce a large amplitude peak in the spectrum at the frequency corresponding to the depth of the flaw, and thus make it easy to locate the flaw.  相似文献   

15.
Mo-Cu合金具有高的导热系数和低的热膨胀系数,被广泛用作散热材料和电子封装材料.阐述了钼铜合金制备过程中容易产生的缺陷,对每一种缺陷产生的原因进行了详细的分析,并提出了相应的控制方法.试验结果表明,研究优化合理的工艺方法和工艺参数,可以避免合金中缺陷的产生,制备出性能优良的钼铜合金.  相似文献   

16.
 针对石油套管缺陷超声无损检测(NDT)中缺陷回波的特点,提出了一种基于小波包分解和支持向量机(SVM)的缺陷智能识别新方法。分析了Gabor、小波和小波包3种信号时频变换分解方法的特点,并进行了基于3种方法生成的特征数据可分性比较,确定了小波包分解方法效果最好。根据SVM解决分类问题的原理,采用SVM法对3种时频分解提取的缺陷信号特征数据进行识别。试验表明,基于小波包分解局部熵的特征提取结合SVM模式智能识别的组合方法,可应用于石油套管上的4种典型缺陷的识别。  相似文献   

17.
介绍了天津钢管公司能源部冷却水处理自动控制系统,阐述了PLC系统在水处理工艺应用过程中存在的缺陷和完善的方法.  相似文献   

18.
Prior studies on the time-of-flight diffraction (TOFD) method have focused primarily on ground smooth, clean surfaces of steel. In practice, however, the surface of an existing bridge element will be covered with rust or have several layers of paint. The main objective of this study was to evaluate the influence of the surface condition of steel elements (i.e., painted or rusted) on the ability of the method to accurately detect and size flaws. These objectives were met by performing a number of tests on plates with saw cuts or implanted fatigue cracks with different surface conditions. These included ground smooth and polished, rusted, and painted surfaces. The data show that rusted surfaces will reduce the amplitude of the ultrasonic signals, but they will not impair the ability of the TOFD method to detect and accurately size flaws. A painted surface will also cause a reduction in signal amplitude. More important, however, is the appearance of additional wave signals that could be interpreted as false indications. While these additional signals do not obscure the presence of actual flaws or affect the accuracy of the TOFD method to size the flaws, they make flaw detection more difficult. Based on the results of this study, recommendations for field inspection on rusted or painted surfaces using the TOFD method are provided.  相似文献   

19.
Place learning is impaired when a single plus maze is moved between adjacent locations 33-120 cm apart. This maze translation creates distinct start locations but maintains a single goal location with respect to distal cues. Hippocampal cell recording data suggest the majority of place fields are tied to apparatus boundaries, not to distal cues, when an apparatus is moved these distances to the left or right. Thus, rats may fail to appreciate the existence of multiple start locations with respect to distal cues when the maze is moved in this way and their start location on the surface is constant. Performance on the single plus maze problem was improved when texture cues were correlated with different start locations. Place learning was supported when multiple start locations were provided on a single large surface (double plus maze), even though rats did not explore the entire surface. Place learning was also supported when random extensions were added to a double plus maze such that start locations, relative to surface boundaries, were not informative as to goal location. This outcome suggests sensitivity to multiple start locations is required for distal cue use in translational place problems. (PsycINFO Database Record (c) 2010 APA, all rights reserved)  相似文献   

设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号