摘要
模糊聚类、粗糙集理论、灰色系统理论等相关技术曾被广泛应用于设备故障诊断中,但是模糊聚类只能对已知样本做出决策,不具有柔性,不能通过已知信息和聚类结果对问题所涉及领域内的新样本的类别做出决策;粗糙集理论不能处理连续变量;而灰色系统理论无法去除故障诊断中冗余的特征参数,不能区分各特征参数的重要性,因而制约了它们在故障诊断中的应用。在本文中,这几种理论被有机地结合起来,应用于设备故障诊断中。在故障诊断过程中,首先利用模糊c均值聚类对样本的参数进行离散化处理,求得各类别的聚类中心,接着基于粗糙集原理对设备特征参数进行约简,去除冗余参数,定量确定各特征参数的重要程度,然后根据约简的特征参数和各参数的重要程度,利用灰色关联分析的方法确定各种标准故障状态与目前设备状态的关联度,从而找到设备的故障所在之处。在本文最后部分通过实例证明,将模糊c均值聚类、粗糙集理论和灰色系统理论结合起来,应用于设备的故障诊断中是一种行之有效的方法,为智能故障诊断提供了理论基础。
Fuzzy cluster, grey system theory and rough set theory have been extensively used in fault diagnosis. Unfortunately, fuzzy cluster can only classify the known samples, the flexibility is very limited, it can not make decision on new samples based on the clustering result according to the known information even if the new samples belong to the correlative domain, grey system theory cannot reduce the superfluous attributes, nor can it determine the relative importance of the attributes while rough set theory cannot solve the problem with the continuous variables. These drawbacks restrain their application in fault diagnosis. These theories are combined together seamlessly and applied to fault diagnosis in the paper. During the process of fault diagnosis, the values of the attributes are discretized through fuzzy c-means cluster and cluster center is obtained, the rough set theory is used to reduce superfluous attributes and quantitatively determine the relative importance of the attributes, and then grey correlation analysis is used to calculate the grey correlation degrees of all the standard fault states with respect to the current state according to reduced attributes and their relative importance, so that the fault can be found. At the end of the paper, an example is used to demonstrate the feasibility of the method put forward in the paper. The method lays the foundation for the intelligent fault diagnosis.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2006年第10期1270-1275,共6页
Chinese Journal of Scientific Instrument
