基于小波分析的液控缓闭止回蝶阀液压缸损坏故障的自动诊断方法  

Automatic Diagnosis Method for Hydraulic Cylinder Damage of Hydraulic Controlled Slow Closing Check Butterfly Valve Based on Wavelet Analysis

作　　者：罗妮娜 LUO Nina(Beijing Jinhe Water Construction Group Co.,Ltd.,Beijing 102200,China)

机构地区：[1]北京金河水务建设集团有限公司,北京102200

出　　处：《自动化应用》2023年第19期86-89,共4页Automation Application

摘　　要：为保证液控缓闭止回蝶阀的正常、稳定运行,需精准完成其液压缸损坏故障的自动诊断,因此,本文提出基于小波分析的液控缓闭止回蝶阀液压缸损坏故障的自动诊断方法。该方法采用小波包分析处理传感器采集的液压缸运行信号,并提取信号中的液压缸故障特征;通过主成分分析方法降低处理提取特征的维数后,将处理后的特征输入至深度置信网络模型中,通过模型学习和迭代训练,输出液压缸损坏故障的自动诊断结果。测试结果显示,该方法在神经元数量为12个的前提下,能有效完成信号中故障特征的提取,并可呈现时域特征结果;具有较好的液压缸损坏故障的自动诊断能力,能判断液压缸损坏故障程度。In order to ensure the normal and stable operation of the hydraulic control slow closing check butterfly valve,it is necessary to accurately complete the automatic diagnosis of its hydraulic cylinder damage fault.Therefore,this paper proposes an automatic diagnosis method for hydraulic cylinder damage fault of the hydraulic control slow closing check butterfly valve based on wavelet analysis.This method uses wavelet packet analysis to process the operating signals of hydraulic cylinders collected by sensors,and extracts the fault features of hydraulic cylinders from the signals.After reducing the dimensionality of the extracted features through principal component analysis,the processed features are input into a deep confidence network model.Through model learning and iterative training,the automatic diagnosis results of hydraulic cylinder damage faults are output.The test results show that this method can effectively extract fault features from the signal and present time-domain feature results with a minimum of 12 neurons.It has good automatic diagnosis ability for hydraulic cylinder damage faults,and judges the degree of hydraulic cylinder damage faults.

关 键 词：小波分析 缓闭止回蝶阀 液压缸 特征提取 特征降维 

分 类 号：TH137[机械工程—机械制造及自动化]