【作者】 詹松；

【导师】 胡献国；

【作者基本信息】 合肥工业大学 ， 机械电子工程， 2004， 硕士

【摘要】 机器中磨损磨粒的铁谱分析技术因其高效、经济的优点，在机械设备的状态检测、故障诊断和预防性维修等方面都得到了广泛的应用。但是，传统铁谱技术操作的复杂性和判断主观性又提出必须尽快解决铁谱诊断智能化的问题，所以磨损磨粒的自动识别技术是近几十年来国内外学者十分关注并致力加以研究的前沿课题。本论文基于摩擦学原理和计算机技术，通过对铁谱片进行数字化图像预处理，采用对谱片图像的平滑、滤波以及阈值二值分割等方法，计算出谱片中特征磨粒的一些特征参量，如面积、周长、粒度以及纵横比等。依据这些特征参量建立必要的数据库，并以标准磨粒特征参量的数据库作为参考系，采用灰色理论中的灰色关联度算法，编写C语言程序，对这些特征磨粒的类别进行分析、识别和判断，从而可以实现它们所属类别的划分。通过对实际油样中制取的图谱进行分析与处理，结果表明，本系统可以实现对三种常见的磨粒进行识别，即对正常磨粒、球状磨粒以及切削磨粒进行分类，验证了该软件的自动识别效果与人工识别相当，具有一定程度的智能效果，为铁谱仪数字化系统打下了一定的基础。更多还原

【Abstract】 Ferrography analysis technology for wear debris has been applied broadly in many aspects, such as inspection of machine operation state, failure diagnosis and preventive maintenance, because of its advantages of good effect and economical. However, the issue how to realize the automatization in ferrograph diagnosis has been brought forth owning to manipulative complexity and subjectivity in the judgement for traditional ferrography technology. Thus, the automatic recognition technology for wear debris is a focus research topic and many researchers, whether in overseas or domestic, have paid more attention on studying it last decades.The characteristic parameters of effective debris in a Ferrograph have been calculated in the present thesis, like area, perimeter, aspect ratio and granularity, in which some methods have been adopted such as smoothing, filtering and thresholding and so on, according to tribological theories and computer technologies and digital image preprocessing. Then, an essential database has been founded with these characteristic parameters. Using the related gray relational grade algorithms, a set of software system with C language has been programmed to analyze, recognize and judge the classification of unknown characteristic wear debris which is referring to the characteristic parameters database of standard wear debris. As a matter of fact, three types of wear debris can be classified by the present software programmed in this research, that is, normal, spherical and cutting debris can be identified. By analyzing and processing the actual ferrograph of wear debris from a used oil sample, the experimental results show that the effects of automatic recognition are equal to those of manual recognition, and the automatization of ferrographical diagnosis has been realized simply and partly, which will be helpful to improve the intelligence of the digital Ferrography system.更多还原