摘要
针对故障诊断中单一来源信号特征信息表征不充分以及深度神经网络调参复杂、构建难度大等问题,提出了一种基于声振特征融合和改进级联森林的离心泵故障诊断方法。首先,对多个传感器采集的声振信号进行小波包去噪,提取降噪信号的时域特征、频域特征和小波包能量特征。利用核主成分分析(kernel principal component analysis,KPCA)对声振信号特征进行特征融合与数据降维,得到特征矩阵。在深度级联森林的基础上引入极端随机森林构建级联层,并添加XGBoost预测器提升模型性能,得到改进级联森林模型。利用改进的级联森林模型进行故障分类,试验结果表明,该方法能够有效识别离心泵的故障类型,并且声振信号特征融合相比于单源信号特征能够有效提升诊断精度。
Aiming at the problems of insufficient characterization of signal feature information from a single source in fault diagnosis and the complexity of tuning parameter of deep neural network and the difficulty of constructing it,a centrifugal pump fault diagnosis method based on acoustic vibration feature fusion and the improvement of cascade forest is proposed.First,wavelet packet denoising is performed on the acoustic vibration signals collected by multiple sensors to extract the time domain features,frequency domain features and wavelet packet energy features of the degraded signals.Using kernel principal component analysis(KPCA),the acoustic vibration signal features are fused and the data are downsized to obtain the feature matrix.On the basis of the deep cascade forest,the extreme random forest is introduced to construct the cascade layer,and the XGBoost predictor is added to improve the model performance,so as to obtain the improved cascade forest model.The improved cascade forest model is used to classify the faults,and the experimental results show that the method can effectively identify the fault types of centrifugal pumps,and the fusion of acoustic vibration signal features can effectively improve the diagnostic accuracy compared with single-source signal features.
作者
厉强国
陈品
陈剑
LI Qiangguo;CHEN Pin;CHEN Jian(Institute of Noise and Vibration Engineering,Hefei University of Technology,Hefei 230009,China;Anhui Automotive NVH Technology Research Center,Hefei University of Technology,Hefei 230009,China)
出处
《组合机床与自动化加工技术》
北大核心
2025年第2期217-221,共5页
Modular Machine Tool & Automatic Manufacturing Technique
关键词
离心泵
故障诊断
特征提取
声振融合
改进级联森林
centrifugal pump
fault diagnosis
fault feature extraction
sound and vibration fusion
improved cascade forest
