基于自相关分析的离心泵振动信号的分段方法

Segmentation Method Based on Autocorrelation Analysis for Vibration Signals of Centrifugal Pump

离心泵是石油工程领域重要的设备之一。分析离心泵运行过程的振动信号、提取反映设备运行状态的有用信息是目前的研究热点。以自吸空调泵为试验研究对象,采集不同负载下该泵运行的振动信息,运用小波去噪与自相关性分析进行振动信号的分段处理,并结合分段平均周期图与功率谱估计方法,提取相应特征。研究表明,采用自相关方法可以对离心泵的振动信号进行合理分段,获得完整的周期信号,从而有效防止了信号信息的混叠与缺失,为信号的特征提取、识别提供了便利。

Centrifugal pump is one of the important equipment in petroleum engineering. Analyzing the vibration signals during pump operation and extracting useful information reflecting operation status of equipment are current research hotspot. With the self - priming air conditioning pump as the experimental research object,the vibration information of this pump under different load is collected,and the vibration signals are segmented by adopting wavelet de - noising and autocorrelation analysis. In addition, the corresponding features are extracted with the method combining segmented average periodogram and estimation of power spectrum. The studies show that the autocorrelation method can reasonably segment the vibration signals of the centrifugal pump to obtain complete periodic signal, effectively prevent the aliasing and lack of information signal, as well as provide convenience for feature extraction and recognition of signals.