摘要
为了研究离心泵不同程度磨损后叶片进口边的振动特性,在虚拟仪器平台搭建试验泵系统,选用IS-50-160-00单级单吸离心泵分析不同磨损程度下振动信号的时域和频域特征.以6个叶片离心泵为研究对象,分别破坏对称2,4,6个进口叶片进行试验,采集泵在不同磨损程度下径向、纵向、轴向和基座方向的振动信号,并进行时域和频谱分析.在时域上采用峭度分析方法,发现磨损后每个叶轮的峭度变化是一致的,通过检测基座方向和轴向的振动信号峭度,可诊断离心泵的磨损程度.利用峰值因子分析得出磨损后叶轮产生更高的尖峰,尤其是轴向峰值因子较大.在频域分析中采用最大熵谱估计方法,得出磨损后离心泵在高频信号中出现振动能量的变化,将高频段(2500~5000 Hz)的振动变化作为诊断磨损程度的特征频段,同时发现振动信号功率谱在基座方向和轴向的能量变化较大.
In order to study the characteristics of the vibration signal under different wear degrees of the inlet edge of centrifugal pump,based on the platform of virtual instrument,the IS-50-160-00 single-suction centrifugal pump was used as the research object,the vibration signal was conducted by means of time domain and frequency domain.The method of kurtosis analysis was taken in time domain analysis.The result shows the kurtosis of different wear impellers with the same variation degree.The kurtosis of vibration signal in the direction of base and axis was monitored to judge the wear degree of centrifugal pump.Based on the analysis of peak factors,there is a higher peak in the domain spectrum of wear impeller,especially in the direction of axis.The variation of vibration energy would appear at high frequency in the method of maximum entropy spectrum analysis.The variation of vibration energy at high frequency(2500-5000 Hz)can be used to determine the wear level of impeller.It can be found that there is the greatest variation of energy of PSD in the directions of base and axis.The cha-racteristics of the vibration signal of the centrifugal pump impeller under different wear degrees can provide reference for the monitoring and analysis of the centrifugal pump fault.
作者
尹江南
袁寿其
骆寅
孙慧
龚波
YIN Jiangnan;YUAN Shouqi;LUO Yin;SUN Hui;GONG Bo(National Research Center of Pumps,Jiangsu University,Zhenjiang,Jiangsu 212013,China)
出处
《排灌机械工程学报》
EI
CSCD
北大核心
2020年第2期139-144,共6页
Journal of Drainage and Irrigation Machinery Engineering
基金
国家自然科学基金资助项目(51409125)
江苏省自然科学基金面上项目(BK20171403)
江苏高校优势学科建设工程项目
关键词
离心泵
磨损
振动
时域分析
频域分析
centrifugal pump
wear
vibration
time domain analysis
frequency domain analysis