液压系统油液分析信息融合故障诊断模型研究  

Research on Fault Diagnosis Model of Hydraulic System Oil Analysis Information Fusion

作　　者：王海涛 WANG Hai-tao(School of Advanced Manufacturing Technology,Guangdong Mechanical&Electrical Polytechnic,Guangzhou 510080,China)

机构地区：[1]广东机电职业技术学院先进制造技术学院,广东广州510080

出　　处：《液压气动与密封》2025年第3期83-89,共7页Hydraulics Pneumatics & Seals

基　　金：广东省普通高校特色创新项目(2020KTSCX225)。

摘　　要：采用油液分析技术对液压系统进行故障诊断和状态监测对保障其可靠运行,防止事故发生具有重要的经济和工程意义。受采样条件和测试费用的限制,油液样本相对较少;油液分析技术种类多,每种技术只能够在一定的知识范围内对液压系统的运行状态做出诊断,且准确率往往偏低。针对以上问题,采用信息融合思想,构建基于量子多种群遗传算法(QMPGA)参数寻优的最小二乘支持向量机(LSSVM)油液分析信息融合故障诊断模型。将该模型用于挖掘机液压系统故障诊断对比发现,该模型相较于BP神经网络、LSSVM和基本遗传算法(GA)参数优选的LSSVM模型,具有更高的故障诊断准确率,且对小样本情况下的故障诊断具有一定的参考价值。Fault diagnosis and condition monitoring of hydraulic system using oil analysis technology is of great economic and engineering significance to ensure its reliable operation and prevent accidents.Limited by sampling conditions and testing costs,the oil samples are relatively small,and there are many kinds of oil analysis technologies,each technology can only diagnose the operating state of the hydraulic system within a certain range of knowledge,and the accuracy rate is often low.To solve the above problems,a least square support vector machine(LSSVM)fault diagnosis model of oil analysis information fusion based on parameter optimization of quantum multi-population genetic algorithm(QMPGA)is constructed by using the idea of information fusion.Compared with BP neural network,LSSVM and GA-LSSVM,it is found that this model has a higher fault diagnosis accuracy,and this model has certain reference value for fault diagnosis in the case of small samples.

关 键 词：液压系统 油液分析 故障诊断 遗传算法 信息融合 

分 类 号：TH137[机械工程—机械制造及自动化] TH117