模糊聚类分析在异常振动振源识别技术中的应用研究

Research on Application of Fuzzy Clustering Analysis in Distinguishing Technology of Abnormal Vibration Source

聚类分析是工程应用中模式识别的重要工具.现代化高楼等工程系统具有多振源且振动特征复杂的特点,模糊聚类分析可以使分类的结果更切合实际.本文以时间序列分析理论为基础,通过对振动信号建立时间序列模型,采用主分量分析法对模型参数进行特征量提取,将模糊聚类分析方法应用于高楼异常振动的振源的实时识别,对机电设备等装置造成的异常振动可以实现有效识别.

Clustering analysis is the important tool for pattern recognition in engineering application. Because the engineering system such as modern high building, has more vibration sources and complex vibration characters, the fuzzy clustering analysis can make the classification results more realistic. Based on time series analysis theory, this paper makes time series model of vibration signal and then picks characteristic quantity from the model parameters by principal component analysis. The fuzzy clustering analysis is used in the real-time distinguishing technology of abnormal vibration source in high building. It can distinguish the abnormal vibration made by equipment such as mechanical and electrical machines effectively.