基于模糊聚类的水电机组轴心轨迹多重分形特征识别

Multi-fractal feature recognition for shaft centerline orbit of hydropower units based on fuzzy clustering

本文应用多重分形特征谱来描述水电机组轴心轨迹图形特征,结合常见水电机组轴心轨迹形状探讨了轴心轨迹多重分形谱特征提取实现方法;利用模糊C-均值聚类方法对多种故障的轴心轨迹多重分形特征值进行了聚类识别,其结果表明该方法可在多轴心轨迹样本情况下达到很好的分类识别,符合水电机组智能诊断的需要。

In this paper, multi-fractal spectrum is used to represent the graphics features of axis orbits of hydropower unit and to achieve feature extraction from common orbit shapes of unit axis. Fuzzy c-means clustering method is used to identify the muhi-fractal characteristics of clustering variety of fault value of orbits. Results show that this method produces better classification identification of the multi-axis samples and meets the needs of intelligent diagnosis of hydropower generating units.