基于WSGMD和RCMFDEn的柱塞泵运行状态识别  

Operation Status Identification of Plunger Pump Based on WSGMD and RCMFDEn

作　　者：马桂林 张昌杰 梁洪基 MA Gui-lin;ZHANG Chang-jie;LIANG Hong-ji(Zhengzhou Branch,China Nuclear Power Engineering Co.,Ltd.,Zhengzhou 450000,China)

机构地区：[1]中国核电工程有限公司郑州分公司,河南郑州450000

出　　处：《液压气动与密封》2025年第2期86-95,共10页Hydraulics Pneumatics & Seals

摘　　要：针对柱塞泵故障信号易受到噪声干扰与识别难度高的问题,提出一种基于加权辛几何模态分解(WSGMD)和精细复合多尺度模糊散布熵(RCMFDEn)的柱塞泵运行状态识别模型。该模型首先将采集到的柱塞泵振动信号通过辛几何模态分解方法分解为多个分量信号,然后依据线性峭度准则筛选出有效的分量信号,并进一步提出归一化自相关能量比指标对筛分信号进行重构;其次,利用精细复合多尺度粗粒化算法改进模糊散布熵用于挖掘故障信号的多尺度特征,继而计算每个重构信号的精细复合多尺度模糊散布熵值并将其作为特征向量;最后,将提取的特征向量输入到经改进灰狼优化算法优化后的核极限学习机中进行状态识别。柱塞泵故障实验台的实测数据分析结果表明,该模型不仅能够准确地识别出不同故障的状态特征,且相较于其他方法具有更高的识别准确率,对于实际工程应用具有一定参考及借鉴价值。Aiming at the problem that the fault signal of plunger pump is easy to be disturbed by noise and difficult to identify,In this paper,a model based on WSGMD and RCMFDEn for identifying the running state of a plunger pump is proposed.Firstly,the collected plunger pump vibration signal is decomposed into several signal components by symplectic geometry mode decomposition method,and then the appropriate signal components are screened according to the linear kurtosis criterion,and the normalized autocorrelation energy ratio index is further proposed for signal reconstruction.Secondly,the refined composite coarse graining algorithm is used to improve the fuzzy dispersion entropy to mine the multi-scale features of the fault signal,and then the refined composite multiscale fuzzy dispersion entropy value of each reconstructed signal is calculated and used as the feature vector.Finally,the extracted feature vectors are input into the kernel extreme learning machine optimized by the improved gray wolf optimization algorithm for state recognition.Plunger pump fault test bench experimental data analysis results show that this method not only can accurately identify the different state characteristics of the fault,and has higher recognition accuracy compared with other methods,with a certain reference and reference value for practical engineering applications.

关 键 词：运行状态识别 柱塞泵 加权辛几何模态分解 精细复合多尺度模糊散布熵 核极限学习机 

分 类 号：TH137[机械工程—机械制造及自动化]