基于信息熵的液压状态监测系统故障诊断方法  

Fault diagnosis method of hydraulic condition monitoring system based on information entropy

作　　者：王佳晨 张宇来 岑岗[1] WANG Jia-chen;ZHANG Yu-lai;CEN Gang(School of Information and Electronic Engineering,Zhejiang University of Science and Technology,Hangzhou 310023,China)

机构地区：[1]浙江科技学院信息与电子工程学院,浙江杭州310023

出　　处：《计算机工程与设计》2021年第8期2257-2264,共8页Computer Engineering and Design

基　　金：国家自然基金青年科学基金项目(61803337)。

摘　　要：为提高液压状态监测系统故障诊断的准确度,提出一种基于IGCS-K2算法的液压状态监测系统传感器故障诊断方法。通过结合信息几何理论与K2评分搜索策略优化贝叶斯网络结构的生成方法,利用运转正常的传感器数据形成贝叶斯网络模型并对传感器最新数据进行预测,通过预测值与观测值的对比判断传感器是否存在故障。实验结果表明,优化后的贝叶斯网络结构生成方法具备相应的理论支持与实践证明,结构准确率与方差优于各类传统方法,该方法可以应用于液压状态监测系统传感器故障诊断中,结果优异。To improve the accuracy of fault diagnosis in hydraulic condition monitoring system,a sensor fault diagnosis strategy based on IGCS-K2 algorithm was proposed.The building method of Bayesian network structure was optimized by combining information-geometric theory and K2 score search strategy,and functioning sensor data were used to build a Bayesian network model and the latest sensor data were predicted.Whether the sensor was faulty was judged by comparing the predicted value with the observed value.Experimental results show that the Bayesian network structure building method has corresponding theoretical and practical proof,and its structural accuracy and variance are better than that of traditional methods.In the meantime,this method can be applied to the sensor fault diagnosis of hydraulic state monitoring system,and achieve excellent results.

关 键 词：贝叶斯网络 信息几何因果推断 信息熵 传感器故障诊断 液压状态监测系统 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]