基于SOM神经网络的液压设备区间观测与故障诊断研究  

Research on Hydraulic Equipment Condition Monitoring and Fault Diagnosis Based on SOM Network

作　　者：郝芳 陈丽娟 HAO Fang;CHEN Lijuan(College of General Education,Zhengzhou University of Economics and Business,Zhengzhou 451191,China)

机构地区：[1]郑州经贸学院通识教育学院,河南郑州451191

出　　处：《河南科技》2025年第6期44-48,共5页Henan Science and Technology

基　　金：郑州经贸学院青年科研基金项目“基于分布式区间观测技术在反馈控制和故障检测中的方法研究”(QK2327);河南省高等学校重点科研项目(24B413001)。

摘　　要：【目的】液压系统作为机械装备的核心组成部分,其故障诊断与健康管理对于预防出现严重后果至关重要。本研究旨在构建高精度的液压系统故障诊断模型,探索先进的液压系统健康管理方法,以确保机械设备的稳定运行和操作安全。【方法】本研究采用自组织映射(SOM)神经网络结合区间观测技术,对液压系统的关键性能指标(HFI)进行分析。SOM神经网络通过无监督学习机制,自适应调整网络参数与结构,揭示输入数据的内在规律。区间观测技术则通过对系统状态进行实时监控,构建状态的上下界估计,为故障诊断提供更为精确的依据。【结果】经过迭代训练的SOM神经网络结构对144种故障状态的液压冷却过滤系统进行了检测,故障诊断准确率达到了98.06%,准确率较高。【结论】本研究所提出的液压设备故障诊断模型不仅能够准确判断设备的工作状态,还具有较高的诊断性能,提高了整个机械装备的运行效率和安全性。[Purposes]As the core component of mechanical equipment,the fault diagnosis and health management of hydraulic system are very important to prevent serious consequences.The purpose of this study is to build a high-precision fault diagnosis model of hydraulic system,explore and realize advanced health management methods of hydraulic system,and ensure the stable operation and safe operation of mechanical equipment.[Methods]The key performance index(HFI)of hydraulic system was analyzed by using SOM neural network combined with interval observation technology.SOM neural network adaptively adjusts network parameters and structure through unsupervised learning mechanism,and reveals the inherent laws of input data.Interval observation technology can estimate the upper and lower bounds of the state by monitoring the system state in real time,which provides more accurate basis for fault diagnosis.[Findings]The iteratively trained SOM network structure was used to detect the hydraulic cooling filtration system in 144 fault states.The accuracy of fault diagnosis reached 98.06%,and the accuracy was high.[Conclusions]The fault diagnosis model of hydraulic equipment proposed in this study not only can accurately judge the working state of the equipment,but also has high diagnostic performance,thus improving the operation efficiency and safety of the whole mechanical equipment.

关 键 词：SOM神经网络 液压设备 故障诊断 区间观测 

分 类 号：TH133.33[机械工程—机械制造及自动化]