基于小波分析的传感器故障诊断方法仿真研究  

Simulation Research on Fault Diagnosis Method of Sensor Based on Wavelet Analysis

作　　者：张伟鹏 李光升 李国强 Zhang Weipeng;Li Guosheng;Li Guoqiang(Department of Control Engineering,Academy of Armored Forces Engineering,Beijing 100072,China)

机构地区：[1]陆军装甲兵学院控制工程系,北京100072

出　　处：《计算机测量与控制》2018年第4期39-43,48,共6页Computer Measurement &Control

摘　　要：装甲车辆数据采集控制系统在使用过程中,其传感器部分容易发生故障且难于直接检测和准确定位;目前对传感器故障诊断的技术中,解析冗余的方法需要精确的数学模型,诸如专家系统等智能化诊断方法都需依赖数学模型,计算量大,实现起来较为困难;为此,提出一种基于小波变换的方法对传感器进行故障诊断;在研究小波分析基本理论的基础上,选择连续小波变换的方法实现故障诊断;利用Matlab/Simulink搭建模型进行仿真试验,区分传感器输入信号不恒为零和恒为零两种情况讨论,对传感器系统输入输出信号做连续小波变换,求出输入输出信号奇异值,去除由于输入突变导致的极值点进而准确定位故障点;仿真结果表明,小波分析的方法能够准确测量传感器的故障,而且该方法具有简单、快速、依赖系统模型程度低、诊断效果好的优点。Armored vehicle data acquisition control system in use,the sensor part of the prone to failure and difficult to directly detect and accurately location.At present,in the technology of sensor fault diagnosis,the method of analyzing redundancy needs an accurate mathematical model.Intelligent diagnosis methods such as expert systems depend on the mathematical model,which requires a large amount of calculation and is difficult to realized.Therefore,a method of fault diagnosis based on wavelet transform is proposed.On the basis of studying the basic theory of wavelet analysis,the method of continuous wavelet transform is selected to realize the fault diagnosis.Using Matlab/Simulink to build a model to simulate the experiment,the sensor input signal is not always zero and constant zero two cases discussed,the input and output signals of the sensor system continuous wavelet transform,the input and output signal singular value,remove the input mutation Resulting in the extreme point and then accurately locate the point of failure;The simulation results show that the wavelet analysis method can accurately measure the sensor fault,and the method has the advantages of simple,fast,low dependence on the system model and good diagnosis effect.

关 键 词：小波变换 传感器故障诊断 SIMULINK仿真 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]