径向柱塞式液压马达故障实验及其智能诊断方法  

Hydraulic Motor Fault Experiment and Its Intelligent Diagnosis Method

作　　者：李涛 徐玮 谯自健 李国平[1] LI Tao;XU Wei;QIAO Zi-jian;LI Guo-ping(School of Mechanical Engineering and Mechanics,Ningbo University,Ningbo,Zhejiang 315211;Ningbo Helm Tower Noda Hydraulic Co.,Ltd.,Ningbo,Zhejiang 315200;Laboratory of Yang jiang Offshore Wind Power,Yangjiang,Guangdong 529599)

机构地区：[1]宁波大学机械工程与力学学院,浙江宁波315211 [2]宁波恒通诺达液压股份有限公司,浙江宁波315200 [3]阳江海上风电实验室,广东阳江529599

出　　处：《液压与气动》2023年第11期46-54,共9页Chinese Hydraulics & Pneumatics

基　　金：国家自然科学基金青年项目(52205569);浙江省自然科学基金(LQ22E050003);阳江海上风电实验室联合基金(YJOFWD-OF-2022A08);宁波市自然科学基金(2022J132)。

摘　　要：针对径向柱塞式液压马达故障数据源较少、故障特征与振动信号之间对应关系不明等问题,运用某企业柱塞式液压马达扭矩测试平台收集其故障振动信号,对比径向柱塞式液压马达不同故障类型之间的振动数据特点,分析故障特征与振动信号之间的对应关系,丰富径向柱塞式液压马达故障诊断数据集,并提出基于深度卷积神经网络的径向柱塞式液压马达智能故障诊断方法。径向柱塞式液压马达故障实验结果表明:相比于正常振动信号而言,滚子和定子裂纹及磨损故障振动信号烈度较大,时域统计指标变化较明显,且频谱中有清晰的故障特征及其倍频成分;提出的智能故障诊断方法可以高效识别径向柱塞式液压马达9种健康状态,诊断精度高达99.85%,为径向柱塞式液压马达出厂前测试的智能诊断奠定了数据和理论基础。Aiming at the problems of few fault data sources of radial plunger hydraulic motors and unclear correspondence between fault characteristics and vibration signals,a company's plunger hydraulic motor torque test platform was used to collect its fault vibration signals and compared with radial plunger hydraulic motors.The characteristics of vibration data between different fault types of hydraulic motors,analyze the corresponding relationship between fault features and vibration signals,enrich the fault diagnosis data set of radial piston hydraulic motors,and propose a radial plunger based on deep convolutional neural network Type hydraulic motor intelligent fault diagnosis method.The results of the radial piston hydraulic motor fault experiment show that compared with the normal vibration signal,the vibration signal intensity of the roller and stator crack and wear fault is greater,the statistical index changes in the time domain are more obvious,and there are clear faults in the frequency spectrum features and their frequency multiplication components;the proposed intelligent fault diagnosis method can efficiently identify nine health states of the radial piston hydraulic motor,and the diagnostic accuracy is as high as 99.85%,which lays the foundation for the intelligent diagnosis of the radial piston hydraulic motor before delivery data and theoretical basis.

关 键 词：径向柱塞式液压马达 智能故障诊断 数据集 振动信号 深度卷积神经网络 

分 类 号：TH137[机械工程—机械制造及自动化] TH17