外文资料翻译.docx

某纺织用收线轮加工用精调对刀仪设计(CAD图纸和说明书资料)

收藏

压缩包内文档预览:(预览前20页/共38页)
预览图 预览图 预览图 预览图 预览图 预览图 预览图 预览图 预览图 预览图 预览图 预览图 预览图 预览图 预览图 预览图 预览图 预览图 预览图 预览图
编号:100229364    类型:共享资源    大小:5.83MB    格式:ZIP    上传时间:2020-10-23 上传人:机****料 IP属地:河南
45
积分
关 键 词:
纺织 收线 工用 设计 CAD 图纸 说明书 资料
资源描述:

喜欢这个资料需要的话就充值下载吧。。。资源目录里展示的全都有预览可以查看的噢,,下载就有,,请放心下载,原稿可自行编辑修改=【QQ:11970985 可咨询交流】====================喜欢就充值下载吧。。。资源目录里展示的全都有,,下载后全都有,,请放心下载,原稿可自行编辑修改=【QQ:197216396 可咨询交流】====================

内容简介:
毕业设计(论文)外文资料翻译设计(论文)题目: 某纺织用收线轮加工用精调对刀仪 学生姓名: 学 院: 机械与电气工程学院 班 级: 学 号: 指导教师: 外文出处: 年2月28日1外文资料翻译译文(约3000汉字):基于机器视觉方式,开发的计算机数控车削机床内插补状态监测系统Wei-Heng Sun 1 and Syh-Shiuh Yeh 2关键词:数控车床,监测,插补,工具,状态,计算机数字控制摘要 :本研究使用机器视觉方法开发了一种机上车刀插入状态监控系统,用于监控计算机数控(CNC)切削过程中的刀具状态。该系统可以识别四个外部车刀插入条件,即断裂,积屑瘤(BUE),崩刃和后刀面磨损。这项研究还使用周围的光源和补光设计了用于刀片尖端的视觉检查系统,可以将其安装在车床机床上,以克服环境对捕获的刀片图像的环境影响,以进行后续图像处理。在图像捕获期间,光源的强度会发生变化,以确保测试插入物具有适当的表面和尖端特征。这项研究实现了外部轮廓构造,刀片状态区域捕获,刀片磨损区域判断以及用于监视和分类刀片条件的计算。然后根据垂直侧面,水平刀片和垂直刀片线修剪插入图像。捕获磨损区域的图像,以使用灰度值直方图监视后刀面或碎屑磨损。使用磨损区域图像作为判断正常磨损或过磨损状况的评估指标来计算磨损量。测试了机上刀片状态监测,以确认所建议的系统可以判断刀片断裂,BUE,崩裂和磨损。结果表明,崩刃的标准偏差和磨损量分别占平均值的0.67和0.62,从而确认了系统运行的稳定性。关键词:机器视觉机上监控;工具插入条件;计算机数控;车床1.简介机械零件的质量取决于加工工具的精度和切削工具的磨损状况。例如,Fernndez-Valdivielso等。1分析了刀片的几何特征对工件表面完整性的影响,并开发了一种间接方法来确定在加工难切削合金时能获得最佳性能的刀片的几何特征。Pereira等。2考虑了刀片和工件之间的界面上的磨损条件,并提出了一种冷却剂结构,该结构结合了低温冷却和最少的润滑量,以提高刀具寿命和工件表面完整性。因此,为了提高产品质量,机械零件制造商必须了解切削工具在实际加工过程中的服务行为(根据在机切削工具状态监视系统确定),以便能够分析刀具寿命并确定是否需要更换切削刀具3,4。车削过程中的刀片磨损形成机理包括磨损,扩散,氧化,疲劳和粘附磨损。如图1所示,侧面,断裂和堆积边缘(BUE)在常规切割过程中最常见,并且大多数磨损,碎裂,断裂和堆积边缘(BUE)在通常集中在刀尖和刀腹的一般切削过程中最常发生5-8。因此,本研究对这四个条件进行了分类,并通过目测检查了插入条件。由于刀片与工件接触的部分之间的腐蚀,切削刀片逐渐发生侧面磨损。过度的切削力通常会导致切削刀片的脆性断裂。但是,由于在加工过程中工件与刀片之间的接触区域处温度很高,所以BUE(这种现象是指加工后的材料在离开刀片边缘时堆积并可能从刀片上带走一部分材料的现象。 ,从而引起刀片边缘),它可能会从刀片边缘脱离,并可能造成一部分破裂和碎裂。图1.四种插入状态表格。(a)侧面磨损;(b)碎裂;(c)骨折;(d) BUE车削过程中有两种插入状态检查:一种是间接检查,外部传感器反馈分析机数据9,10,另一种是直接检查,测量切削刀具的状态11, 12。间接检查分析数据以估计刀具状态;根据参考来分析某些机器状态,这意味着系统地评估了切削状态,从而取代了经验丰富的操作员的判断,从而减少了人为错误并提高了生产自动化的能力13。例如,切削刀具的磨损情况根据切削噪声或振动的差异进行分析14,15,通过测量切削温度和切削力的变化来监控切削工具16,17,并使用机器功率或切削力来分析切削状态电流变化信号18。所有这些方法都使用感测信号进行检查分析。最近,通过电荷耦合器件(CCD)相机进行间接检查已变得很流行。它通过捕获图像中的工件表面纹理来分析切削刀具,以确定切削刀具是否磨损,并根据工件表面纹理和表面粗糙度的变化进行判断19-23。一些研究集中于融合多个传感器和图像的视觉信息以进行进一步的工具状态监控24,25,或使用不同的算法模型进行分析以实现更准确的监控和评估26-29。根据上述参考文献,可以通过分析机器信息的变化来获得切削工具的状态。然而,这种间接检查有时会在外部传感环境的影响下降低系统的精度30。因此,需要一种直接检查方法来分析切削刀具状态的变化。直接检查通过直接观察切削刀具的实际情况来分析加工中的问题。一些文档使用声音,光线或探针来建立切割工具模型,以观察切割工具的状态25,31,32。但是,这种测量设备相对复杂并且不适合现场检查。另一种方法是使用电荷耦合器件(CCD)相机捕获工具图像并分析切削工具的状态。有关切削工具状态的分析有两种,一种是通过外部轮廓和轮廓检查分析磨损状况33,通常用于监视外部轮廓磨损状态以判断切削工具是否仍可使用。与用于分析和检查加工工件的表面纹理以确定刀具状态的间接检查方法相比,直接检查方法是通过刀具的表面纹理或表面粗糙度分析来判断切削刀具的状态。加工后的边缘34,因为它提供了详细的加工信息,因此可用于更详细地检查切削刀具和机床状态。CCD相机的一般视觉检查分析了切削刀具的不同位置,例如,一些研究根据火山口磨损进行了分析35,36,而另一些研究则根据侧面磨损情况进行了分析37。可以通过视觉检查收集有关切削刀具的大多数状态信息。换句话说,可以从图像获得外部轮廓的变化。Giusti等。38提出了一种切削刀具磨损的视觉检查方法,Rangwala和Dornfeld 39提出了使用神经网络分析磨损状态的方法,许多学者相继提出了其他相关的检查方法26。关于磨损特征的优化方法,Kurada和Bradley 40提出了使用梯度算子来计算纹理特征的方法,其中使用八边形矩阵计算磨损区域边界特征搜索,并通过亮度差和距基体中心的径向距离,以确定最佳磨损特征的位置。在预处理过程中对原始图像进行了平滑处理,以减少不规则性的干扰。对于特征计算,通过图像阈值转换像素值以获得实际磨损强度并确定磨损量的变化。Yuan等。41提出了一种获取平均图像的新滤波方法,并提出了一种基于小波变换的边缘检测新方法。选择小波函数后,将生成一个新的小波函数,用于描述图像的灰度变化。换句话说,可以避免噪声干扰,以获得更好的边缘特征以及磨损区域,宽度,长度,可以测量磨损区域的中心位置。Wang等。42提出了一种图像处理程序,它不同于基于恒定阈值的传统方法。在这种方法中,考虑了从粗到精的策略。首先,获得针对搜索候选者的磨损底边缘点的阈值图像。然后,基于矩不变性的与阈值无关的边缘检测方法用于确定磨损边缘。为了缩短计算时间,首先定义一个关键区域,然后在后续过程中仅将该区域视为关注区域;因此,规避了基于阈值的磨损特征检测方法。Li等。43在刀具磨损监测中使用了仿生的脉冲耦合中性网络(PCNN),并根据灰度强度高于的条件,使用空间邻域像素和类似的灰色像素簇来分割刀具磨损的二值图像。工具的主体和工具磨损领域的背景。Shahabi和Ratnam 44使用原始图像的外部轮廓测试工具图像的对齐方式,然后使用中值滤波,形态学运算和阈值算法来减少由于切割工具未对齐,微细齿的存在而导致的系统误差。灰尘颗粒,振动以及环境光的强度变化。目的是确定刀架的位置和定位误差,以确保无需精确对准刀具就可以检查切削刀具的磨损。Pfeifer和Wiegers 45使用光源变化来确定不同光源下的磨损边缘特征。虽然光的变化会影响切削工具磨损边缘的阴影变化,但实际的边缘位置不会随光源而变化。因此,可以使用高通滤波器和阈值图像获得不同光源下的刀具磨损图像信息,并且可以通过重叠确定强边缘的位置来获得递归边缘位置,以滤除由于污染物引起的误识别并减少污染物和阴影变化对检查系统的影响。Barreiro等。46使用不同的矩作为描述符来说明工具磨损图像,然后使用有限混合MCLUST模型将工具磨损条件分类为低磨损,中磨损和高磨损类别。此外,通过基于线性和二次判别分析的方法对监测结果进行了验证,该分析基于最先进的图像处理结果,Alegre等人。47开发了一种程序,通过使用k最近邻和多层神经网络来确定工具更换的时间。DAddona和Teti 48使用图像标准化过程在切割测试中获得具有标准尺寸和像素密度的图像。然后,对反向传播神经网络进行优化,并使用标准化的切削刀具图像来估计刀具的磨损状况。与现有的研究结果不同,本研究分析了刀片的状态,并使用融合轮廓和纹理检查方法来构建更准确的评估和判断系统,该系统适用于机上自动检查并消除了检查过程中的环境问题。构造了可用于CNC车床的视觉检查系统,该系统由CCD相机和用于捕获插入图像的镜头,用于保护照相设备的保护盒以及用于避免废料飞溅的外围电路和组件组成。切割过程中的切削液。视觉检查系统具有清洁空气管,该清洁空气管向插入件喷射空气以清洁被检查插入件的表面,从而减少了后续图像处理的问题并提高了插入件状态判断的准确性。本研究中设计的视觉检查系统具有周围光源和用于刀片尖端的补光灯,以确保可以在变化的照明条件下分析刀片条件。当调节光源以确定刀片和刀片尖端的位置时,如果现场工具对准不准确,则可适用于刀片状态监测,从而提高了机器中图像识别的可行性。调整光源强度,并在变化的强度下捕获插入图像以进行检查分析。可以减少任何外部环境变化对插入状态监视结果的影响,并且本研究设计的系统可以在不同环境中获得准确的结果。通常由于插入条件的变化而导致的图像曝光不足或日晒也得到了改善。这项研究分析了捕获的具有不同特征的刀片图像,并且还可以检查外部车刀的常见刀片状况,包括断裂,BUE,崩刃和磨损。可以根据纹理特征分布对结果进行量化。本研究对处于不同状态的刀片进行了机上刀片状态监测实验,结果表明,本研究设计的刀片状态监测系统适用于计算机数控(CNC)车床机床,用于正确,稳定地识别刀片断裂, BUE,碎裂和磨损情况。因此,这项研究的贡献包括开发了一种机载刀片状态监视系统,该系统可用于一次识别四个刀片状态-断裂,BUE,崩刃和后刀面磨损。开发具有不同光源的可安装视觉系统,以便在机器上捕获可在不同照明条件下进行精确分析的高质量插入图像。开发轮廓和纹理融合检查方法以减少环境问题并在检查过程中准确识别插入条件。本文的结构如下。第2节介绍了本研究中使用的实验系统和相关设备,以及机器视觉检查系统的硬件体系结构设计。第3节介绍了本研究中设计的插入物图像捕获过程,以及插入物尖端的周围光源和补光灯的使用。第4节介绍了本研究中设计的刀片状态监测分类过程,包括刀片外部轮廓构造,刀片状态区域捕获以及磨损区域判断和计算。第5节介绍了插入条件监视的实验过程和结果。通过数控车床对机载刀片状态进行监控的实验验证了该系统的可行性和稳定性。第6节总结了本文。2.实验系统与设备介绍使用外部车刀对本研究中使用的CNC车床进行了测试,如图2所示。将测试外部车刀安装在车床刀架上,并通过计算机数字控制器将车床刀架移动到位于车床刀轴上方的外观检查系统,以监控插入状态。在此,通过移动转塔来调整每个插入位置,以使关注区域聚焦,以减少捕获图像的模糊。此外,在实验期间,通常使用安全门帽来保护操作员处于关闭状态,从而使转弯区域的环境可以减少来自外部环境的影响。本研究使用GigE DFK 23GP031彩色工业相机,其图像分辨率为2592 X 1944(15 fps)。图3a示出了照相机的硬件组合。镜头是Myutron HS3514J CCTV镜头,结合了双镜头以拍摄特征图像,并且90度反射镜可以调节摄像机的角度。由于机器内部结构的速度限制,并考虑到实际加工环境造成的潜在污染,本研究设计了一种可视检查系统,该系统可以安装在车床中,如图3b所示。如图3a所示,相机硬件的保护盒可防止机器中的切屑飞溅,从而减少镜头上切削液的污染。为了捕获清晰的刀片图像,清洁空气管向刀片喷射空气以进行清洁。保护盒配备有可调节亮度的周围LED光源,并用环氧树脂覆盖以提供保护。保护盒将补光灯扩展到要检查的刀片尖端(尖端光源)。在保护盒基座上设置了两个磁性基座,用于将保护盒固定在机床上,以监控在机插入状态。图2.用于实验的CNC车床。(a)转弯区;(b)炮塔结构图3。 图3.可安装在本研究设计的车床内部的外观检查系统。(a)工业相机和镜头相关组件;(b)外观检查系统保护箱。3.插入图像捕获过程在图像捕获期间,补光灯用于检查的插入物,并且改变光源强度以确保不同的插入物具有适当的特征强度。这项研究使用了位于不同位置的两个光源,如图3b所示。就周围的光源而言,强光照射测试插入物以获得其表面形状和面积特征。在嵌件尖端的补光方面,尖端状态功能得到增强,以方便以后对捕获的图像进行处理和分析。如图4a所示,在本研究设计的刀片图像捕获过程中,在高强度的周围光源下拍摄刀片以捕获刀具侧面曝光图像。然后,使用高强度的周围光源和插入头的补光来捕获插入图像,如图4b所示。在此,曝光图像可以顺序用于确认刀片位置,增强几何特征并增强磨损特征。拍摄曝光图像后拍摄特征图像。首先,关闭刀片尖端的补光灯,并调节周围光源的强度,以获得适当的侧面特征图像,如图5a所示,然后再对刀片尖端的补光灯强度进行调整。如图5b所示,调整来获得合适的针尖特征图像。在此,根据拍摄图像的平均阈值自动地进行光源强度的调整。参照曝光图像,如图4所示,特征图像用于分析不同的插入条件,并可用于插入条件的分类过程,包括插入轮廓构造,状态区域捕获以及磨损判断和计算。图4.捕获的侧面并插入曝光图像。(a)侧面曝光图像;(b)插入曝光图像图5.捕获的侧面和尖端特征图像。(a)侧面特征图像;(b)提示功能图片。4.插入状态监视分类过程4.1.插入外轮廓构造首先,使用图4a中的侧面曝光图像确定侧面轮廓特征。本研究使用灰度图像阈值确定侧面轮廓特征,如图6aa所示。同样,图4b中的插入轮廓特征和灰度图像阈值用于确定插入轮廓特征,如图6b所示。在这里,阈值是250。使用直线霍夫变换确定图6中阈值图像中的线,如图7所示。侧面轮廓曝光阈值图像确定垂直侧面线和水平刀片线,而插入轮廓曝光阈值图像确定垂直刀片线。可以修整阈值图像,并沿着图7中的水平刀片线(图7a)和垂直刀片线(图7b)旋转,以构建完整的刀片外轮廓阈值图像,如图8a所示。根据图7a中的垂直侧面线,完整的刀片外轮廓阈值图像分为两个块,如图8b所示:前端前端下侧(块B)和刀片后端下侧(块A),用于后续的插入条件特征识别。图9通过参考完整的刀片外部轮廓阈值图像显示了修剪后的刀片图像的结果。图6.捕获的曝光插入灰度图像的阈值操作结果。(a)侧面曝光阈值图像;(b)插入曝光阈值图像。图7.阈值图像行。(a)侧面轮廓特征的垂直侧面线和水平叶片线;(b)插入曝光图像中的垂直刀片线。图8.完成的刀片外部轮廓阈值图像和块划分。(a)完整的刀片外轮廓阈值图像;(b)插入外部轮廓块分隔。 图9.修整后的刀片外部轮廓图4.2。插入状态区域捕获 图7a中的水平刀片线可用于判断刀片是否有断裂或BUE。图10a中的刀片阈值图像是在图9的灰度图像阈值化处理之后获得的。这里,使用11进行的腐蚀和膨胀操作x 11菱形结构元素用于清除几何特征。沿着水平刀片线对插入阈值图像进行分割,以获得图10b中的插入断裂区域和图10c中的插入BUE区域,其中计算断裂区域和BUE区域的像素区域以判断插入断裂或BUE状态。 图10.刀片断裂和BUE状态的判断。(a)插入阈值图像;(b)插入断裂带;(c)插入BUE区域。 如果通过本研究中设计的刀片状态监测系统确定的刀片状态未归类为断裂或BUE,则开始侧翼磨损判断过程。首先,对图9中的修剪后插入的外部轮廓图像执行灰度转换。本研究使用平均图像RGB值进行灰度处理。在将刀片外部轮廓图像转换为灰度图像之后,使用Sobel运算符进行刀片边缘检测以获得良好的刀片边缘特征。如图8b所示,然后将刀片的外轮廓块进行分割,并去除刀片后端的下部区域(块A),以分割侧面磨损特征的位置,如图11a所示。为了便于修整侧面磨损部件以进行后续判断和计算,如图9所示,执行了噪声消除,对比度拉伸过程,腐蚀和膨胀操作的计算,并获得了侧面磨损区图像,如图所示在图11b中。此处,使用3 X 3盒型低通滤波器和21 X 21盒型中值滤波器来抑制噪声。最后,根据图11b对图9中的刀片外轮廓图像执行修整操作,并在图11cis中获得侧面磨损区域图像。 图11.修整后的刀片磨损区域的实际图像。(a)侧面磨损特征区;(b)侧面磨损区的范围;(c)侧面磨损区图像。4.3.磨损区域判断与计算 可以根据修剪后的侧面磨损区域图像对侧面磨损或碎屑磨损状态进行分类,如图11c所示。图12显示了侧面磨损和崩刃磨损之间存在显着差异。侧面磨损是在加工过程中由于切削刀片和工件之间的摩擦而产生的撕裂,因此,侧面磨损的表面特征大部分是连续且均匀的。但是,由于切屑磨损是由于异常加工过程导致的尖端断裂,所以切屑表面相对粗糙。这项研究分析了磨损区域实际图像的表面特征的连续性,以将刀片磨损区域识别为侧面磨损或碎裂磨损,如图11c所示。将图11c转换为灰度图像后,可以获得所有像素的灰度值直方图,如图13a所示。像素数量明显大于侧面磨损图像的像素灰度值直方图分布,如图13b所示。因此,将大于预设阈值的像素数除以整个磨损区域的像素数的计算值百分比,以将磨损区域识别为侧面磨损或碎裂磨损,即百分比(碎裂率)将大于预设阈值的像素数与整个磨损区域的像素数之比作为判断的基础。此外,本研究使用磨损区域图像上下边界的像素长度来计算磨损量。使用磨损区域图像转换像素单位,如图14所示,其中磨损区域图像像素的转换长度为0.007 mm,磨损区域图像的上下边界的像素长度为184个像素,因此,转换后的磨损量为1.288毫米。5.实验监视插入条件为了验证本研究中提出的机载刀片状态监视系统的可行性,图15中显示了安装在车床机床上以进行刀片状态监视实验的外观检查系统。该研究使用了20种处于不同状态的二手刀片进行实验。结果显示在表1和图16中。此处,使用的刀片是在切削速度(130-150 m / min),切削进给速度(0.2-0.3 mm / rev)和切削深度的车削后收集的切割(2- -3 mm)。工件材料为中碳钢,插入材料为碳化钨。笔记本电脑使用Intel Core i7-4720HQ,2.6 GHz CPU和64位Microsoft Windows 10操作系统来实施整个系统,因此每个监控任务1S所需的时间约为13 S 1n,其中2.75 S on平均,是确定插入条件所必需的。为了进一步减少执行每个监视任务所需的时间,可以使用具有更快CPU的计算机来实施该系统。表1示出了处于不同状态的刀片的判断结果。碎裂率设置为50以进行监控,磨损量设置为0.3 mm以识别过度磨损。根据结果,本研究开发的系统可以正确识别测试刀片的各种刀片条件。表1列出了三种类型的BUE插入件,其中两种具有轻微的BUE。因此,可以说这项研究根据预设的BUE阈值准确地识别了BUE状态。 刀片状态监视系统可以识别不同的刀片状态,其运行稳定性是评估的关键。由于外部环境和光源强度的变化,插入条件测试和计算的结果将有所不同。这项研究反复测试了同一刀片,以验证刀片状态监测系统的稳定性,实验结果列于表2,其中计算了刀片磨损区域的磨损量以进行比较分析。重复该实验10次,记录每个实验的碎屑率和磨损量,并使用计算出的平均值和标准偏差检查系统稳定性。实验结果表明,切屑速率分析具有较大的标准偏差,这表明结果存在较大差异。即使算法和光源系统的操作步骤相同,也可以根据侧面磨损区图像的像素灰度值直方图分布来计算碎片率,即使图像捕获的每个时刻都受到光源变化和灰度值直方图的影响磨损图像的分布发生变化。尽管如此,切屑率的标准偏差仅为平均值的0.67,并且可以计算出此插入状态监视系统的切屑率计算的稳定性。就磨损量结果而言,磨损量的标准偏差仅为平均值的0.62,换句话说,该标准偏差小于两个像素。因此,可以计算出该插入状态监视系统在计算磨损量时的稳定性。因此,上述实验结果可用于验证本文设计的插入状态监测系统和计算方法的可行性和稳定性。 这项研究开发了一种机上刀片状态监测系统,以识别四个外部车刀插入状态。断裂,BUE,碎裂和侧面磨损。实验结果表明,所开发的监控系统能够成功识别出四种刀片状态,如图16所示,当难以通过标准磨损测量来精确测量磨损量时,该系统可用于识别刀片状态。但是,由于安装在车床机床内部的已开发的外观检查系统的视角与标准磨损测量设备的视角不同,因此计算出的磨损量可用于指示磨损状况,无法与使用标准磨损测量设备获得的测量结果进行比较。6。结论 机床切削过程中使用的切削刀具的状态将明显影响机床零件的制造质量。因此,本研究开发了一种用于数控车床的车刀刀片的机载刀片状态监测系统,并使用机器视觉方法检查车刀刀片的常见侧面磨损,崩裂,断裂和BUE状态。这项研究与现有的研究方法和结果不同,因为它将机器视觉方法与轮廓和纹理检查相融合以分析刀片状态,从而消除了刀片检查过程中的环境问题,从而建立了更准确的机上车刀插入条件监视系统。 为了将CCD摄像机和镜头固定在CNC车床中以执行机上插入条件的目视检查过程,设计了一种目视检查系统,该系统带有保护盒,清洁空气管和两个光源。保护盒可以避免在切割过程中碎片飞溅和镜片上的切削液污染,同时清洁空气管将空气吹向刀片,以清除表面污染物。使用具有可变光强度的插入物的尖端的周围光源和填充光来分析照明条件的变化对插入物状态的目视检查的影响。在插入件图像捕获过程中,将更改周围光源和不良光线的强度,以确保测试插入件具有适当的特征强度。周围的光源使用强光照射插件表面以获得表面形状和区域特征,而补光则增强了笔尖状态特征,以利于后续捕获的图像处理和分析。本研究中设计的刀片状态监测分类过程包括刀片外部轮廓构造,刀片状态区域捕获以及磨损区域判断和计算。刀片的外部轮廓构造使用曝光图像确定外部轮廓特征,然后根据该外部轮廓特征建立垂直侧面线,水平刀片线和垂直刀片线。可以对插入图像进行修剪,以用于后续的插入条件特征识别。就刀片状态区域捕获而言,根据外部轮廓特征线识别刀片断裂区域和BUE区域,并对刀片外部轮廓图像进行修整,以获取刀片磨损区域的实际图像。为了进行刀片磨损区域的判断和计算,基于修整后的侧面磨损区域图像中所有像素的灰度值直方图来识别侧面磨损或碎屑磨损,并使用切削刃的上下边界的像素长度来计算磨损量磨损区域图像,用作识别刀片的正常磨损或过度磨损状态的参考索引。最后,将处于不同状态的刀片用于机上刀片状态监测实验,以确认本研究中设计的系统可以识别刀片的断裂,BUE,碎屑和磨损状态。另外,由于外部环境和光源的变化有时会影响图像处理结果,因此在本研究中测试了机载刀片状态监视系统的操作稳定性。反复进行实验,记录实验结果中崩刃率和磨损量的平均值和标准偏差,作为评估系统运行稳定性的基础。实验结果表明,光源变化确实会影响切屑率和磨损量的计算。切屑率的标准偏差仅为平均值的0.67,而磨损量的标准偏差为平均值的0.62(标准偏差小于2个像素),从而验证了系统运行的稳定性。作者贡献:W.-HS,S.-SY,调查;监督,S.-SY资金来源:这项研究由台湾科学技术部根据MOST 104-2221-E-027-132和M(ST 103-2218-E-009-027-MY2)进行部分资助。 致谢:作者要感谢SRAM台湾公司的代表与研究团队进行的有益讨论。作者特别感谢SRAM台湾公司的Meng-Hui Lin的有益讨论。 利益冲突:作者声明没有利益冲突。2外文资料原文(与课题相关,至少1万印刷符号以上):Using the Machine Vision Method to Develop an On-machine Insert Condition Monitoring System for Computer Numerical Control Turning Machine ToolsWei-Heng Sun 1 and Syh-Shiuh Yeh 2,*1.Institution of Mechatronic Engineering, National Taipei University of Technology, Taipei 10608, Taiwan; sun245689Department of Mechanical Engineering, National Taipei University of Technology, Taipei 10608, Taiwan2.* Correspondence: ssyeh.tw; Tel.: +886-2-2771-2171Abstract: This study uses the machine vision method to develop an on-machine turning tool insert condition monitoring system for tool condition monitoring in the cutting processes of computer numerical control (CNC) machines. The system can identify four external turning tool insert conditions, namely fracture, built-up edge (BUE), chipping, and flank wear. This study also designs a visual inspection system for the tip of an insert using the surrounding light source and fill-light, which can be mounted on the turning machine tool, to overcome the environmental effect on the captured insert image for subsequent image processing. During image capture, the intensity of the light source changes to ensure that the test insert has appropriate surface and tip features. This study implements outer profile construction, insert status region capture, insert wear region judgment, and calculation to monitor and classify insert conditions. The insert image is then trimmed according to the vertical flank, horizontal blade, and vertical blade lines. The image of the insert-wear region is captured to monitor flank or chipping wear using grayscale value histogram. The amount of wear is calculated using the wear region image as the evaluation index to judge normal wear or over-wear conditions. On-machine insert condition monitoring is tested to confirm that the proposed system can judge insert fracture, BUE, chipping, and wear. The results demonstrate that the standard deviation of the chipping and amount of wear accounts for 0.67% and 0.62%, of the average value, respectively, thus confirming the stability of system operation.Keywords: machine vision; on-machine monitoring; tool insert condition; computer numerical control; turning machine tools1. IntroductionThe quality of mechanical parts is dependent on the accuracy of the machining tools and the abrasion conditions of cutting tools. For instance, Fernndez-Valdivielso et al. 1 analyzed the effects of geometrical features of inserts on workpiece surface integrity and developed an indirect method for determining the geometrical features of inserts that achieve the best performance in machining difficult-to-cut alloys. Pereira et al. 2 considered the abrasion conditions on the interface between an insert and a workpiece, and proposed a coolant structure that combines cryogenic cooling and the minimum quantity of lubrication to improve tool life and work piece surface integrity. Thus, to improve the quality of products, mechanical part manufacturers must be aware of the service behaviors of cutting tools in the actual machining process, as determined from the on-machine cutting tool condition monitoring system, to be able to analyze tool life and decide whether the cutting tool needs to be changed 3,4. The insert wear formation mechanism in the turning process comprises abrasion, diffusion, oxidation, fatigue, and adhesion wear. As shown in Figure 1, flank,fracture, and built-up edge (BUE) occur most frequently in general cutting processes and are mostlywear, chipping, fracture, and built-up edge (BUE) occur most frequently in general cutting processes concentrated at the tool tip and tool flank 58. Therefore, these four conditions are classified in this study, and the insert condition is reviewed by visual inspection. Flank wear gradually occurred cutting insert owing to the erosion between the portions of the insert in contact with the workpiece. Excessive cutting force can usually lead to brittle fracture of a cutting insert. However, due to the high temperature at the contact area between the workpiece and the insert during the machining processes, the BUE (the phenomenon that the machined material builds up on break away from the insert edge and could carry a portion of material from the insert, thereby causing the insert edge) occurs and it could break away from the insert edge and could carry a portion of fracturing and chipping.Figure 1. Four insert status forms. (a) Flank wear; (b) Chipping; (c) Fracture; (d) BUEThere are two types of insert condition inspections in turning processes: one is indirect inspection, .where the external sensors feedback the analytical machine data 9,10, and the other is direct inspection, where the cutting tool status is measured 11,12. Indirect inspection analyzes data to estimate the cutting tool status; some machine states are analyzed according to a reference,which means that the cutting status is systematically evaluated, thus replacing the judgment of experienced operators to reduce human errors and enhancing the ability of production automation 13.For example, the cutting tool wear condition is analyzed based on the difference in the cutting noise or vibration 14,15, the cutting tool is monitored by measuring the changes in cutting temperature and cutting forces 16,17, and the cutting status is analyzed using the machine power or current variation signal 18. All these methods use sensing signals for inspection analysis. Lately, indirect inspection by a charge-coupled device (CCD) camera has become popular. It analyzes the cutting tools by capturing the workpiece surface texture in images to determine whether the cutting tool is worn,judged according to the changes in the workpiece surface texture and surface roughness 19-23. Some studies have focused on the fusion of multiple sensors and visual information of images for further tool status monitoring 24,25 or used different algorithm models for analysis to implement more accurate monitoring and evaluation 26- -29. According to the aforementioned references, the status of cutting tools can be obtained by analyzing machine information variations; however, such indirect inspection sometimes reduces the accuracy of the system under the effect of the external sensing environment 30.Therefore, a direct inspection method is required to analyze the changes in the status of cutting tools.Direct inspection analyzes the problems in machining by directly observing the practical situation of the cutting tool. Some documents use sound, light, or a probe to build a cutting tool model to observe the status of the cutting tool 25,31,32. However, such measurement equipment is relatively complicated and unsuitable for on site inspection. Another method uses a charge coupled device (CCD) camera to capture tool images and analyze the status of cutting tools. There are two types of analysis regarding the status of cutting tools, one is to analyze the wear condition by outer contour and profile inspection 33, which is generally used to monitor the outer profile wear status to judge whether the cutting tool is still workable. In comparison with the indirect inspection methods that are used to analyze and inspect the surface texture of a machined workpiece to determine the tool status, the direct inspection method is to judge the status of the cutting tools by surface texture or surface roughness analysis of the tool edge after machining 34, which is applied for a more detailed inspection of the cutting tool and the machine states, as it provides detailed machining information. The general visual inspection of a CCD camera analyzes the different locations of a cutting tool, for example, some studies have implemented analysis according to crater wear 35,36, whereas others have implemented it according to the flank wear condition 37. A majority of the status information regarding a cutting tool can be gathered by visual inspection; in other words, changes in the outer profile can be obtained from the images. Giusti et al. 38 proposed a visual inspection method for cutting tool wear, Rangwala and Dornfeld 39 proposed using a neural network to analyze wear status, and many scholars successively proposed other related inspection methods 26. Regarding the methods for optimization of wear features, Kurada and Bradley 40 proposed using gradient operators to calculate texture features, where the wear region boundary feature search was calculated using an octagonal-shaped matrix, and the slope was established by the brightness difference and radial distance from the matrix center to determine the location of optimized wear features. The original image was smoothed during preprocessing to reduce the interference of irregularities. For feature calculation, the pixel value was converted by image thresholding to obtain the actual wear intensity and determine the change in wear amount. Yuan et al. 41 proposed a new filtering method to obtain average images and proposed a new edge detection method based on wavelet transform. When the wavelet function is selected, a new wavelet function is generated that describes the gray change of the image. In other words, noise interference can be avoided to obtain better edge features and the abrasion region, width, length, and center location of abrasion region can be measured. Wang et al. 42 proposed an image processing procedure, which is different from the traditional method based on constant thresholding. In this method, a rough-to-fine strategy is considered. First, the thresholding images are obtained for the search candidates wear bottom edge points. Then, the threshold-independent edge detection method, based on moment invariance, is used to determine the wear-edge. To shorten the computing time, a critical area is defined first, and only this area is taken as the region of interest in subsequent processes; thus, evading the threshold-dependent wear features detection method. Li et al. 43 used the pulse-couple neutral networks (PCNN) of bionics in cutting tool wear monitoring and used the spatial neighbor and similar gray clusters of pixels to segment the binary image of tool wear according to the condition that the gray intensity is higher than the body of the tool and background in the field of tool wear. Shahabi and Ratnam 44 used the external profile of the original image to test the alignment of the tool image and then used median filtering, morphological operations, and thresholding algorithms to reduce the system errors resulting from cutting tool misalignment, the presence of micro-dust particles, vibrations, and the intensity variations of ambient light. The aim was to determine the tool holder position and positioning error to ensure that cutting tool wear could be inspected without precision tool alignment. Pfeifer and Wiegers 45 used light source changes to determine wear-edge features under different light sources. While light changes can influence the shadow changes of the cutting tool wear-edge, the actual edge location does not vary with the light source. Thus, cutting tool wear image information under different light sources can be obtained using high-pass filter and thresholding images and the recurrent edge location can be obtained by overlapping to determine the location of a strong edge to filter out the misrecognition due to contaminants and reduce the effects of contaminants and shadow changes on the inspection system. Barreiro et al. 46 used different moments as descriptors to illustrate the tool wear images and then used a finite mixture MCLUST model to classify tool wear conditions into low-, medium-, and high-wear classes. Furthermore, the monitoring results were validated through the use of linear and quadratic discriminant analysesBased on the image processing results of the cutting edge, Alegre et al. 47 developed a procedure to determine the time for tool replacement through the use of k-nearest neighbors and a multilayer neural network. DAddona and Teti 48 used an image standardization process to obtain images with standard size and pixel density during cutting tests; then, the back-propagation neural network was optimized and used to estimate tool wear conditions with standardized cutting tool images.Differing from existing research findings, this study analyzes insert statuses and uses fusion contour and texture inspection methods to build a more accurate evaluation and judgment system, which is applicable to on-machine automatic inspections and eliminates the environmental problems during inspection. A visual inspection system that can be used in CNC turning machine tools is constructed, which consists of a CCD camera and a lens for capturing insert images, a protection box to protect the photographic equipment, and a peripheral circuit and components, to avoid scrap splashes and cutting fluids during the cutting processes. The visual inspection system has a cleaning air tube, which jets air toward the insert to clean the surface of the inspected insert, thus, reducing the problems of subsequent image processing and increasing the accuracy of the insert condition judgment. The visual inspection system designed in this study has a surrounding light source and a fill-light for the tip of the insert to ensure that the insert condition can be analyzed under changing lighting conditions. When the light source is adjusted to determine the location of the blade and the tip of the insert, it is applicable to insert condition monitoring if on site tool alignment is not accurate, thus enhancing the feasibility of image recognition in the machine. The light source intensity is adjusted and the insert image is captured under varying intensities for inspection analysis. The effect of any external environment changes on the insert condition monitoring result can be reduced and the system designed in this study can obtain accurate results in different environments. Image underexposure or solarization that generally result from changes in the insert condition are also improved. This study analyzes captured insert images with different features and the common insert conditions of an external turning tool, including fracture, BUE, chipping, and wear, can also be inspected. The analysis of the results can be quantized according to the texture feature distribution. This study conducts on-machine insert condition monitoring experiments with inserts in different states and the results show that the insert condition monitoring system designed in this study is applicable to computer numerical control (CNC) turning machine tools for correct and stable identification of insert fracture, BUE, chipping, and wear conditions. Contributions of this study therefore include development of an on-machine insert condition monitoring system that can be used to one-time identify the four insert conditionsfracture, BUE, chipping, and flank wear. development of a mountable visual system with different light sources to on-machine capture good-quality insert images that can be exactly analyzed under different lighting conditions. development of a contour and texture fusion inspection method to reduce environmental problemsand to accurately identify insert conditions during inspection.The structure of this paper is as follows. Section 2 describes the experimental system and related equipment used in this study, along with the hardware architecture design of the machine vision inspection system. Section 3 describes the insert image capture process designed in this study and the usage of the surrounding light source and fill-light for the insert tip. Section 4 describes the insert condition monitoring classification process designed in this study, including the insert outer profile construction, insert status region capture, and wear region judgment and calculation. Section 5 describes the experimental process and results of insert condition monitoring. The experiment on the on-machine insert condition monitoring by a CNC turning machine tool validates the feasibility and stability of this system. Section 6 summarizes this paper.2. Introduction to the Experimental System and EquipmentThe CNC turning machine tool used in this study, shown in Figure 2, is tested using an external turning tool. The test external turning tool is mounted on the turning machine tool turret and the turning machine tool turret is moved by a computer numerical controller to the visual inspection system placed above the turning machine tool spindle for insert condition monitoring. Here, each insert position is adjusted by moving the turret such that the region of interest is focused in order to reduce the blurring of captured images. Moreover, during the period of experiments, the security door hat is usually used to protect operators was closed so that the turning zone environment can reduce the influence from external environments. A GigE DFK 23GP031 color industrial camera, with an image resolution of 2592 X 1944 (15 fps), is used in this study. Figure 3a shows the camera hardware combination; the lens is a Myutron HS3514J CCTV lens, combined with a double lens to capture the feature image and the 90-degree reflecting mirror can adjust the angle of the camera. Due to the pace constraints of the internal structure of the machine and considering the potential contamination resulting from the actual machining environment, this study designs a visual inspection system that can be mounted in turning machine tools, as shown in Figure 3b. The protection box for the camera hardware, as shown in Figure 3a, prevents the cutting scrap in the machine from splashing, thus,reducing the contamination of cutting fluid on the lens. To capture sharp insert images, the cleaning air tube jets air toward the insert for cleaning. The protection box is equipped with a surrounding LED light source with adjustable brightness and is covered with epoxy resin for protection. The protection box extends the fill-light for the tip of the insert to be inspected (tip light source). Two magnetic bases are set up at the protection box base to fix the protection box in the machine tool for on-machine insert condition monitoring.Figure 2. CNC turning machine tool for experiment. (a) Turning zone; (b) Turret structureFigure 3. Cont. Figure 3. Visual inspection system mountable inside turning machine tools designed in this study.(a) Industrial camera and lens related components; (b) Visual inspection system protection box.3. Insert Image Capture ProcessDuring image capture, the fill-light is used for the inspected insert and the light source intensity is changed to ensure that different inserts have appropriate feature strength. This study uses two light sources in different positions as shown in Figure 3b. In terms of the surrounding light source, a strong light irradiates the test insert to obtain its surface shape and area features. In terms of the fill-light for the tip of an insert, the tip status feature is enhanced to facilitate later processing and analysis of he captured image. In the insert image capture process, as designed in this study, the insert is shot under a high-strength surrounding light source to capture the tool flank exposure image, as shown in Figure 4a. Then the insert image is captured using a high-strength surrounding light source and fill-light for the insert tip, as shown in Figure 4b. Here, the exposure images can be sequentially used to confirm the insert position, enhance geometry features, and strengthen wear features. The featured images are captured after the exposure image capture. First, the fill-light for the tip of an insert is closed and the surrounding light source intensity is adjusted to obtain appropriate flank feature images,as shown in Figure 5a, and then the intensity of the fill-light for the tip of an insert is adjusted to obtain appropriate tip feature images, as shown in Figure 5b. Here, the adjustment of light source intensity is automatically performed depending on the average thresholding value of the captured images. Referring to the exposure images, as shown in Figure 4, the feature images are used to analyze different insert conditions and can be utilized in the classification process of the insert conditions,including insert profile construction, status region capture, and wear judgment and calculation.Figure 4. Captured flank and insert exposure images. (a) Flank exposure image; (b) Insert exposure image Figure 5. Captured flank and tip feature images. (a) Flank feature image; (b) Tip feature image.4. Insert Condition Monitoring Classification Process4.1. Insert Outer Profile Construction First, the flank profile feature is determined using the flank exposure image in Figure 4a.This study uses grayscale image thresholding to determine the flank profile feature, as shown in Figure 6aa. Similarly, the insert profile feature in Figure 4b and grayscale image thresholding are used to determine the insert profile feature, as shown in Figure 6b. Here, the thresholding value is 250.The lines in the thresholding images in Figure 6 are determined using straight-line Hough transform,as shown in Figure 7. The flank profile exposure thresholding images determine the vertical flank line and horizontal blade line, while the insert profile exposure thresholding images determine thevertical blade line. The thresholding image can be trimmed and rotated along the horizontal bladeline (Figure 7a) and vertical blade line (Figure 7b) in Figure 7 to construct a complete insert outerprofile thresholding image, as shown in Figure 8a. According to the vertical flank line in Figure 7a, thecomplete insert outer profile thresholding image is divided into two blocks, as shown in Figure 8b: tipfront-end underside (block B) and insert backend underside (block A) for subsequent insert conditionfeature recognition. Figure 9 shows the results of the trimmed insert images by referring to thecompleted insert outer profile thresholding image.Figure 6. Thresholding operation result of captured exposure insert grayscale image. (a) Flank exposurethresholding image; (b) Insert exposure thresholding image.Figure 7. Lines of thresholding images. (a) Vertical flank line and horizontal blade line of flank profile feature; (b) Vertical blade line in insert exposure image.Figure 8. Completed insert outer profile thresholding images and block division. (a) Completed insert outer profile thresholding image; (b) Insert outer profile block division. Figure 9. Trimmed insert outer profile image.4.2. Insert Status Region Capture The horizontal blade line in Figure 7a can be used to judge whether the insert has a fracture or BUE.The insert thresholding image in Figure 10a is obtained after the grayscale image thresholding process of Figure 9. Here, the erosion and dilation operations with the 11 x 11 diamond-shaped structuring element are used to clear the geometry features. The insert thresholding image is segmented along the horizontal blade line to obtain the insert fracture zone in Figure 10b and the insert BUE zone in .Figure 10c, where the pixel areas of the fracture zone and BUE zone are calculated to judge the insert fracture or BUE status. Figure 10. Judgment of insert fracture and BUE statuses. (a) Insert thresholding image; (b) Insert fracture zone; (c) Insert BUE zone. If the insert condition, as identified by the insert condition monitoring system designed in this study, is not classified as fracture or BUE, the flank wear judgment process begins. First, the grayscale transformation is implemented for the trimmed insert outer profile image in Figure 9. This study uses the average image RGB values for grayscale processing. After the insert outer profile image is converted into a grayscale image, the Sobel operator is used for insert edge detection to obtain a good insert edge feature. The insert outer profile blocks are then segmented, as shown in Figure 8b, and the lower region at the back end of the insert (block A) is removed to segment the location of the flank wear feature, as shown in Figure 11a. To facilitate the trimming of the flank wear part for subsequent judgment and calculation, the computation for noise removal, contrast stretching process, erosion, and dilation operations are implemented, as shown in Figure 9, and the flank wear zone image is obtained,as shown in Figure 11b. Here, the 3 X 3 box-pattern low-pass filter and the 21 X 21 box-pattern median filter are used for noise suppression. Finally, the trimmed operation is implemented for the insert outer profile image in Figure 9 according to Figure 11b and the flank wear zone image in Figure 11cis obtained. Figure 11. Actual image of trimmed insert wear region. (a) Flank wear feature zone; (b) Range of flank wear zone; (c) Flank wear zone image.4.3. Wear Region Judgment and Calculation The flank wear or chipping wear status can be classified according to the trimmed flank wear zone image, as shown in Figure 11c. Figure 12 shows that there is a significant difference between the flank wear and chipping wear. The flank wear is the tear resulting from the rub between the cutting blade and workpiece in the machining process, thus, the flank wear surface features are mostly continuous and even. However, as chipping wear is the tip breakage resulting from abnormal machining processes,the chipping surface is relatively rough. This study analyzes the continuity of surface features for the actual image of a wear region to identify the insert wear region as flank or chipping wear, as shown in Figure 11c. The grayscale value histogram of all pixels can be obtained after Figure 11c is converted into a grayscale image, as shown in Figure 13a. The number of pixels is obviously larger than the pixel grayscale value histogram distribution of the flank wear image, as shown in Figure 13b. Therefore, the number of pixels larger than the preset threshold value is divided by the calculated value percentage of the number of pixels of the overall wear region to identify the wear region as flank or chipping wear.In other words, the percentage (chipping rate) of the number of pixels larger than the preset threshold value to the number of pixels of the overall wear region is taken as the basis of judgment. Moreover,this study uses the length of the pixels of the upper and lower boundaries of the wear region image to calculate the wear amount. The pixel unit is converted using the wear region image, as shown in Figure 14, where the conversion length of the wear region image pixels is 0.007 mm and the length in he pixels of the upper and lower boundaries of wear region image is 184 pixels, thus, the converted wear amount is 1.288 mm.5. Experiment Monitoring Insert ConditionTo validate the feasibility of the on-machine insert condition monitoring system proposed in this study, the visual inspection system mounted on the turning machine tool for insert condition monitoring experiments is shown in Figure 15. This study uses twenty used inserts in various states for experimentation and the results are presented in Table 1 and Figure 16. Here, the used inserts were collected after turning with cutting speed (130-150 m/ min), cutting feed rate (0.2-0.3 mm/ rev),and depth-of-cut (2- -3 mm). The workpiece material is medium carbon steel and the insert material s tungsten carbide. The laptop computer with an Intel Core i7-4720HQ, 2.6-GHz CPU, and 64-bitMicrosoft Windows 10 operating system was utilized to implement the whole system so that the time required tor each mon1torlng task 1S approximately 13 S 1n Which 2.75 S, on average, are required for the identification of insert conditions. To further reduce the time required for each monitoring task,a computer with a faster CPU could be used to implement the system. Table 1 shows the judgment results of the inserts in different states. The chipping rate is set at 50% for monitoring and the wear amount is set as 0.3 mm for identifying over-wear. Based on the results, the system developed in this study can correctly identify the various insert conditions of the test inserts. Table 1 presents three types of BUE inserts, where two of them have slight BUE. Thus, it can be said that this study identifies the BUE status accurately according to the preset threshold of BUE. The insert condition monitoring system can identify different insert conditions and its operational stability is a key point of evaluation. Due to the changing external environment and light source intensity, there will be different results for insert condition tests and calculations. This study repeatedly tests the same insert to validate the stability of the insert condition monitoring system and the experimental results are shown in Table 2, where the wear amount of the insert wear region is calculated for comparison analysis. The experiment is repeated 10 times, the chipping rate and wear amount of each experiment are recorded, and system stability is checked using the calculated mean value and standard deviation. The experimental results show that the chipping rate analysis has large standard deviation, which signifies that there is a large variation in the results. The chipping rate is calculated according to the pixel grayscale value histogram distribution of the flank wear zone image,even though the algorithm and light source system operating procedure are identical, each moment of image capture is affected by the light source change and the grayscale value histogram distribution of the wear images changes. Despite all this, the standard deviation of the chipping rate is only 0.67% of the average value and the stability of the chipping rate calculation of this insert condition monitoring system can be calculated. In terms of wear amount results, the standard deviation of wear amountis only 0.62% of the average value, in other words, the standard deviation is lower than two pixels.Hence, the stability of this insert condition monitoring system in calculating wear amount can becalculated. Therefore, the aforementioned experimental results can be used to validate the feasibilityand stability of the insert condition monitoring system and calculation method designed in this study. This study developed an on-machine insert condition monitoring system to identify four external turning tool insert conditions; fracture, BUE, chipping, and flank wear. The experimental results demonstrate that the developed monitoring system can successfully identify the four insert conditions.Moreover, as shown in Figure 16, the developed system can be used for identifying the insert conditions when it is difficult to measure the wear amount precisely using standard wear measurement methods.However, because the view angle of the developed visual inspection system that is mounted inside the turning machine tools is different from the view angle of standard wear measurement devices, the calculated wear amount, which is used to indicate the degree of wear conditions, cannot be compared with the measurement results obtained using standard wear measurement devices.6. Conclusions The status of cutting tools used in the cutting processes of machine tools will obviously influence the manufacturing quality of machine parts. Therefore, this study develops an on-machine insert condition monitoring system for the turning tool insert of CNC turning machine tools and uses the machine vision method to inspect the common flank wear, chipping, fracture, and BUE statuses of turning tool inserts. This study differs from the existing research methods and outcomes as it fuses the machine vision method with contour and texture inspections to analyze the insert status.This eliminates the environmental problems in the insert inspection process to build a more accurate on-machine turning tool insert condition monitoring system. To fix the CCD camera and lens in the CNC turning machine tool to carry out the on-machine insert condition visual inspection process, a visual inspection system with a protection box, cleaning air tube, and two light sources is designed. The protection box can avoid the scrap splash and contamination of cutting fluid on the lens during the cutting processes, while the cleaning air tube jets air toward the insert to clean off surface contaminants. A surrounding light source and a fill light for the tip of an insert with variable light intensities are employed to analyze the effect of change in lighting conditions on the visual inspection of the insert status. In the insert image capture process,the intensity of the surrounding light source and ill-light is changed to ensure that the test insert has appropriate feature strength. The surrounding light source uses strong light to irradiate the insert surface to obtain the surface shape and area features, while the fill-light enhances the tip status feature to facilitate subsequent captured image processing and analysis. An insert condition monitoring classification process designed in this study includes insert outer profile construction, insert status region capture, and wear region judgment and calculation. The insert outer profile construction uses the exposure image to determine the outer profile feature, and then the vertical flank line, horizontal blade line, and vertical blade line are established according to this outer profile feature. The insert image can be trimmed for subsequent insert condition feature recognition. In terms of insert status region capture, the insert fracture zone and BUE zone are identified according to the outer profile feature lines and the insert outer profile image is trimmed to obtain the actual image of the insert wear region. For insert wear region judgment and calculation, the flank wear or chipping wear is identified based on the grayscale value histogram of all pixels of the trimmed flank wear zone image.The wear amount is calculated using the pixel length of the upper and lower boundaries of the wear region image, which are used as the reference index to identify the normal wear or over-wear status of the insert. Finally, inserts in different states are used for on-machine insert condition monitoring experimentation to confirm that the system designed in this study can identify insert fracture, BUE,chipping, and wear statuses. In addition, as the changes in external environment and light source sometimes influence the image processing result, the operational stability of the on-machine insert condition monitoring system is tested in this study. The experiment is conducted repeatedly and the average value and standard deviation of the chipping rate and wear amount in the experimental results are recorded as the basis for evaluating the operational stability of the system. The experimental results show that the light source variation does influence the calculation of chipping rate and wear amount. The standard deviation of the chipping rate is only 0.67% of the average value, while the standard deviation of wear amount is 0.62% of the average value (standard deviation lower than 2pixels), thus validating the stability of system operation.Author Contributions: Investigation, W.-H.S., S. -S.Y.; Supervision, S.-S.Y.Funding: This research was funded in part by the Ministry of Science and Technology, Taiwan, R.O.C., under Contract MOST 104-2221-E-027-132 and M( ST 103-2218-E-009-027-MY2. Acknowledgments: The authors would like to thank representatives from the SRAM Taiwan Company for their useful discussions with the research team. The authors especially thank to Meng- Hui Lin (SRAM Taiwan Company) for his beneficial discussions. Conflicts of Interest: The authors declare no conflict of interest.Referencesl. Fernndez- Valdivielso, A.; Lpez De Lacalle, L.N,; Urbikain, G,; Rodriguez, A. Detecting the key geometrical features and grades of carbide inserts for the turning of nickel-based alloys concerning surface integrity.Proc. Inst. Mech. Eng. Part C J. Mech. Eng. Sci. 2016, 230, 3725- -3742. CrossRef2. Pereira, O.; Rodrguez, A.; Fernndez Abia, A.I.; Barreiro, J.; Lpez de Lacalle, L.N. Cryogenic and minimum quantity lubrication for an eco-efficiency turning of AISI 304. J. Clean. Prod. 2016, 139, 440- 449. CrossRef3. Yu, J. Machine tool condition monitoring based on an adaptive Gaussian mixture model. J. Manuf. Sci. Eng. Trans. ASME 2012, 134, 031004. CrossRef4. Jones, B.E. Sensors in industrial metrology. J. Phys. E Sci. Instrum. 1987, 20, 11-1116. CrossRef5. Avinash, C.; Raguraman, S.; Ramaswamy, S.; Muthukrishnan, N. An Investigation on Effect of Workpiece Reinforcement Percentage on Tool Wear in Cutting Al-SiC Metal Matrix Composites. In Proceedings of theASME International Mechanical Engineering Congress and Exposition, Seattle, WA, USA, 11-15 November2008; pp. 561- -566.6. Ee, K.C.; Balaji, A.K.; Jawahir, I.S. Progressive tool-wear mechanisms and their effects on chip-curl/ chip-form in machining with grooved tools: An extended application of the equivalent toolface (et) model. Wear 2003,255, 1404- -1413. CrossRef7. Nordgren, A.; Melander, A. Tool wear and inclusion behaviour during turning of a calcium-treated quenched and tempered steel using coated cemented carbide tools. Wear 1990, 139,
温馨提示:
1: 本站所有资源如无特殊说明,都需要本地电脑安装OFFICE2007和PDF阅读器。图纸软件为CAD,CAXA,PROE,UG,SolidWorks等.压缩文件请下载最新的WinRAR软件解压。
2: 本站的文档不包含任何第三方提供的附件图纸等,如果需要附件,请联系上传者。文件的所有权益归上传用户所有。
3.本站RAR压缩包中若带图纸,网页内容里面会有图纸预览,若没有图纸预览就没有图纸。
4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
5. 人人文库网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对用户上传分享的文档内容本身不做任何修改或编辑,并不能对任何下载内容负责。
6. 下载文件中如有侵权或不适当内容,请与我们联系,我们立即纠正。
7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。
提示  人人文库网所有资源均是用户自行上传分享,仅供网友学习交流,未经上传用户书面授权,请勿作他用。
关于本文
本文标题:某纺织用收线轮加工用精调对刀仪设计(CAD图纸和说明书资料)
链接地址:https://www.renrendoc.com/paper/100229364.html

官方联系方式

2:不支持迅雷下载,请使用浏览器下载   
3:不支持QQ浏览器下载,请用其他浏览器   
4:下载后的文档和图纸-无水印   
5:文档经过压缩,下载后原文更清晰   
关于我们 - 网站声明 - 网站地图 - 资源地图 - 友情链接 - 网站客服 - 联系我们

网站客服QQ:2881952447     

copyright@ 2020-2025  renrendoc.com 人人文库版权所有   联系电话:400-852-1180

备案号:蜀ICP备2022000484号-2       经营许可证: 川B2-20220663       公网安备川公网安备: 51019002004831号

本站为文档C2C交易模式,即用户上传的文档直接被用户下载,本站只是中间服务平台,本站所有文档下载所得的收益归上传人(含作者)所有。人人文库网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私,请立即通知人人文库网,我们立即给予删除!