轴向柱塞泵气穴现象故障诊断研究  

Fault Diagnosis of Cavitation Phenomenon in Axial Piston Pump

作　　者：杨景昆 张纯[1] 强熠宇 戴吉勇 关栋 秦永法[1] YANG Jingkun;ZHANG Chun;QIANG Yiyu;DAI Jiyong;GUAN Dong;QIN Yongfa(School of Mechanical Engineering,Yangzhou University,Yangzhou Jiangsu 225127,China)

机构地区：[1]扬州大学机械工程学院,江苏扬州225127

出　　处：《机床与液压》2025年第6期125-130,共6页Machine Tool & Hydraulics

基　　金：国家自然科学基金青年科学基金项目(52005433);江苏省(扬州大学)研究生科研与实践创新计划项目(SJCX23_1904)。

摘　　要：在高速高压工况下,柱塞泵的气穴现象不可避免。针对传统气穴现象故障检测方法中存在效果差和迟滞性的缺点,提出一种结合实验研究和信号处理的新型气穴检测系统,并采用KOA-CNN-BiLSTM-Attention深度学习算法对采集的信号进行训练,实现对轴向柱塞泵气穴强度的检测。通过CNN提取数据中的高维空间特征,利用开普勒优化算法(KOA)与CNN相互配合来增强关键特征的表现能力,再通过BiLSTM提高信号的时序性,最后利用注意力机制进行优化。结果表明:该方法对轴向柱塞泵气穴故障状态识别精度达到98%,其效果明显优于其他方法,可以更加准确地识别轴向柱塞泵气穴故障状态。Under high speed and high pressure conditions,the cavitation phenomenon of plunger pump is inevitable.In view of the shortcomings of poor effect and hysteresis in traditional fault detection methods of cavitation phenomenon,a new cavitation detection system combining experimental research and signal processing was proposed.The KOA-CNN-BiLSTM-Attention deep learning algorithm was used to train the collected signals and detect the cavitation strength of the plunger pump.High-dimensional spatial features were extracted from the data using a convolutional neural network(CNN).Subsequently,the Kepler optimization algorithm(KOA)was used to cooperate with the CNN to enhance the representation capability of key features.A bidirectional long short-term memory(BiLSTM)network was then applied to improve time sequence of the signals,finally attention mechanism was used for optimization.The results show that this method achieves a cavitation fault state recognition accuracy of 98%for axial plunger pumps,which is significantly better than other approaches and can identify the fault state of axial plunger pump more accurately.

关 键 词：柱塞泵 信号处理 开普勒优化算法 深度学习 

分 类 号：TH322[机械工程—机械制造及自动化]