旋转机械运行状态数据库的模糊聚类研究

THE APPLICATION OF FUZZY SET CLUSTERING IN THE FORMATION OF DATABASE FOR RUNNING CONDITION OF ROTATING MACHINERY

旋转机械是化工、发电、高速运输工具动力系统的关键设备，为减少因定期维修或突发故障造成的经济损失，需对其进行实时状态监测，建立特定机组的历史故障或状态数据库。本文把旋转机械的振动频谱、润滑油温度等物理参量作为特征参量，利用模糊数学及多元分析原理对旋转机械状态监测系统数据库的建立进行了研究并将研究结果应用于实际机组，取得了很好的效果。

In this paper, the formation of database for running condition of rotating machinery is discussed on the base of ISODATA clustering method. The vibrational spectrum of rotating part and the temperature of lubrication oil are chosen to be characteristic parameters of running condition of machinery. The original samPle data are processed in order to apply ISODATA correCtly. Moreover, some improvement has been made towards ISODATA according to the characteristics of the sample data. The software is also designed for analysis of the sample data of practical machinery, and is proven to be effective in practice.