声阵列监测下轴向柱塞泵噪声空域滤波与故障分类  

Acoustic Signal Array Based Spatial Filtering and Fault Location for Axial Piston Pump

作　　者：王长林 孙俊杰 钟永腾 WANG Chang-lin;SUN Jun-jie;ZHONG Yong-teng(School of Information Technology,Jiangsu Open University,Nanjing 210036;College of Mechanical and Electrical Engineering,Wenzhou University,Wenzhou 325035)

机构地区：[1]江苏开放大学信息工程学院,江苏南京210036 [2]温州大学机电工程学院,浙江温州325035

出　　处：《制造业自动化》2025年第4期54-60,共7页Manufacturing Automation

基　　金：国家自然科学基金青年科学基金(51505339);浙江省自然科学基金青年基金(LQ16E050005)。

摘　　要：轴向柱塞泵长期工作在高速、高压的恶劣环境下,其内部的关键零件常发生故障,导致整个液压系统失效。采用自制非接触式麦克风阵列(声阵列)研究了轴向柱塞泵故障噪声特征分析、定位与故障分类。首先,基于二维传感器阵列建立了噪声定位与空域滤波模型,实现噪声源定位下的特定方位滤波,将空域滤波后的时域信号及其频域信号作为典型故障样本;然后,用支持向量机(SVM)替代原来的分类器,并利用空域滤波后的典型故障样本训练一维CNN-SVM故障分类模型;通过柱塞、配流盘、斜盘和回程盘等四个关键部件预制故障的实验研究,实现了轴向柱塞泵运行状态下的故障源定位,并结合CNN-SVM模型对故障进行了分类,其中时域样本的测试集准确率为89.5%,频域样本的测试集准确率为93.5%。Axial piston pump works for a long time in the harsh environment of high speed and high pressure,and its key parts inside often fail,resulting in the failure of the whole hydraulic system.In this paper,fault noise source location method and fault characteristics analysis of axial piston pump are proposed using a self-made noncontact microphone array(acoustic array).Firstly,a signal source location and spatial filtering model based on array are established.Secondly,four kinds of faults such as plunger fault,plate fault,swash plate fault and return plate fault are designed and studied experimentally.Finally,the time domain index is extracted by the filtered signal using the location results.The experimental results show that the proposed method can locate the fault source,and the noise signal index after spatial filtering can reflect the fault characteristics of different parts and guide the fault diagnosis of axial piston pump effectively.

关 键 词：轴向柱塞泵 故障噪声 声阵列定位 空域滤波 时域指标 

分 类 号：TH133[机械工程—机械制造及自动化]