基于C0复杂度和GG模糊聚类的轴承性能退化状态识别

Identification of Degradation State for Bearing Performances Based on C0 Complexity and GG Fuzzy Clustering

如何表征机械设备的性能退化程度,并对退化状态进行识别是机械设备故障预测中的关键问题。提出一种基于C 0复杂度和GG模糊聚类的退化状态识别方法。首先,以混沌logistics序列为例,对比验证C 0复杂度参数在复杂性表征以及运算速度方面的优势;然后,考虑退化状态在时间尺度的连续性,将时间参数映射到指数函数中,形成更符合性能退化过程的“弯曲时间参数”,并与C 0复杂度、有效值构建描述性能退化过程的三维特征向量;最后,采用GG模糊聚类方法对性能退化状态进行阶段划分,识别不同的退化状态,并选用分类系数、平均模糊熵以及序列离散度对聚类效果进行综合评价。实例分析表明:提出的三维特征向量既能够反映性能退化趋势,又能体现同一状态在时间尺度上的连续性;GG聚类算法与同类的GK,FCM算法相比,聚类效果更优。

It is a key issue in fault prognostic that how to characterize performance degradation degree and identify degradation state for mechanical equipments.An identification method for degradation state is proposed based on C0 complexity and GG fuzzy clustering.Firstly,the chaotic logistic sequence is taken as an example to comparatively verify advantages of C0 complexity parameter in complexity characterization and calculation speed.Then the continuity of degradation state on time scale is taken into account and time parameter is mapped to exponential function,forming“bending time parameter”which is more in line with performance degradation process.In order to describe performance degradation process,a three-dimensional feature vector is constructed including C0 complexity,RMS and curved time.Finally,the GG fuzzy clustering method is used to divide stages of performance degradation state and identify different degradation states.The classification coefficient,average fuzzy entropy and sequence dispersion are used to evaluate clustering effect comprehensively.The actual example analysis shows that the proposed three-dimensional feature vector reflects performance degradation trend and continuity for same state on time scale.Compared with similar GK and FCM algorithms,the GG clustering algorithm has better clustering effect.