压力容器气体泄漏中的FSWT和SVM联合检测方法  

FSWT and SVM Combined Detection Method for Gas Leakage in Pressure Vessels

作　　者：李鹏[1,2,3] 陆一 杨佳康 徐永凯 LI Peng;LU Yi;YANG Jiakang;XU Yongkai(Jiangsu Key Laboratory of Meteorological Observation and Information Processing,Nanjing University of Information Science&Technology,Nanjing Jiangsu 210044,China;Jiangsu Collaborative Innovation Center on Atmospheric Environment and Equipment Technology,Nanjing University of Information Science&Technology,Nanjing Jiangsu 210044,China;School of Automation,Wuxi University,Wuxi Jiangsu 214105,China)

机构地区：[1]南京信息工程大学,江苏省气象探测与信息处理重点实验室,江苏南京210044 [2]南京信息工程大学,江苏省大气环境与装备技术协同创新中心,江苏南京210044 [3]无锡学院自动化学院,江苏无锡214105

出　　处：《电子器件》2021年第4期1011-1018,共8页Chinese Journal of Electron Devices

基　　金：江苏省重点研发计划社会发展项目(BE2015692)。

摘　　要：及时发现压力容器气体泄漏并判别其程度对避免发生安全事故具有重要意义,从气体泄漏源的声信号特征分析角度,从超声范围内提出一种基于频率切片小波变换和支持向量机的气体泄漏故障检测方法。利用频率切片小波变换对声波传感器采集的声信号进行预处理,分析泄漏声信号局部特征并选取特征明显的观测范围进行重构,从多域分量中提取能够区分是否存在泄漏以及泄漏程度的代表性特征,利用Relief-F算法计算并选取最优特征作为支持向量机的输入并对压力容器的气体泄漏进行识别和分类。经实验分析,采用所筛选出的最具鉴别性的6种特征对是否存在泄漏的判断准确率达到99.75%,对泄漏程度分类的准确率达到94.68%。结果表明该方法的气体泄漏检测准确率高,有助于后续实时检测系统的开发和泄漏孔定位研究。The detection of gas leakage in pressure vessels and the identification of its degree to avoid the occurrence of safety accidents is of great significance.A gas leak detection method based on frequency slice wavelet transform and support vector machine in ultrasonic range from the perspective of acoustic signal characteristic analysis of gas leakage source is proposed.The frequency slice wavelet transform is used to preprocess the acoustic signals collected by acoustic sensors.The local characteristics of leakage signal are analyzed,and the observation range with obvious characteristics is selected for inverse transformation.From the multi-domain components of the acoustic signal,the representative characteristics can be extracted to distinguish whether there is leakage and the extent of leakage.The Relief-F algorithm is used to calculate and select the best feature as the input of support vector machine to identify and classify the pressure vessel gas leakage.The experimental analysis indicates that the accuracy rate of leakage detection is 99.75% and leak classification is 94.68% by using the six most discriminative features.The result shows that this method has high accuracy in gas leak detection,which is helpful to the development of leakage source location and real-time detection system.

关 键 词：气体泄漏检测 超声 频率切片小波变换 阈值滤波 支持向量机 

分 类 号：TM933[电气工程—电力电子与电力传动]