首页|基于传感信号采集的电控发动机振动故障监测方法

基于传感信号采集的电控发动机振动故障监测方法

扫码查看
原文链接
  • 国家科技期刊平台
  • NETL
  • NSTL
  • 万方数据
通过调理振动信号可以更高效地监测振动故障.为此,提出基于传感信号采集的电控发动机振动故障监测方法.首先,搭建电控发电机传感信号采集与处理架构,通过放大传感信号增益、滤波和转换信号模数的方式处理待监测信号,为提高监测准确性奠定可靠的数据基础.通过小波包分解与重构,获取信号的时域参数和小波能谱熵,并构建三维特征量.然后,利用"一对一"分解策略优化孪生支持向量机,构造多元分类器,使其更适用于振动故障监测这一多类别分类问题,再输入待监测信号的特征量,通过确定故障类别实现持续性监测.仿真结果表明:该方法训练耗时的最大值仅为 897 ms,对于转子摩擦振动、不平衡振动等 5 种类型故障的监测准确率始终在 97%以上,在缩减训练样本后准确率仍保持在 90%以上.
Vibration Fault Monitoring Method of Electronically Controlled Engine Based on Sensor Signal Acquisition
Vibration faults can be monitored more efficiently by modulating vibration signals.Therefore,a vibration fault monitoring method for electronically controlled engine based on sensor signal acquisition is proposed.First of all,the sensing signal acquisition and processing architecture of the electric control generator is built,and the signal to be monitored is processed by amplifying the sensing signal gain,filtering and converting the signal modulus,so as to lay a reliable data foundation for improving the monitoring accuracy.Through wavelet packet decomposition and reconstruction,the time domain parameters and wavelet energy spectrum entropy of the signal are obtained,and the three-dimensional feature quantity is constructed.Then,the"one-to-one"decomposition strategy is used to opti-mize the twin support vector machine,and a multivariate classifier is constructed to make it more suitable for the multi category classifi-cation problem of vibration fault monitoring.Then,the characteristic quantity of the signal to be monitored is input,and the continuous monitoring is achieved by determining the fault category.The simulation results show that the maximum training time of the proposed method is only 897 ms,and the values of the monitoring accuracy of five types of faults such as rotor friction vibration and unbalanced vibration are always above 97%,and the values of the accuracy remain above 90%after reducing the training samples.

signal and information processingvibration fault monitoringcollect sensing signalelectronically controlled enginesignal conditioningsignal conversionwavelet energy spectrum entropytwin support vector machine

马晓、郑晅、柴艳娜

展开 >

长安大学信息与网络管理处(教育技术与网络中心),陕西 西安 710064

长安大学能源与电气工程学院,陕西 西安 710064

信号与信息处理 振动故障监测 传感信号采集 电控发动机 信号调理 信号转换 小波能谱熵 孪生支持向量机

江苏省发改委新兴产业发展专项

KY2020037

2024

10.3969/j.issn.1004-1699.2024.04.016

传感技术学报
东南大学 中国微米纳米技术学会

传感技术学报

CSTPCD北大核心
影响因子：1.276
ISSN：1004-1699
年,卷(期)：2024.37(4)
  • 12