非接触旋转密封故障声发射信号的诊断与识别研究

打开文本图片集

关键词：故障诊断；声发射；非接触旋转密封；状态识别；深度学习 中图分类号： TH165+.3 文献标志码：A DOI：10.16385/j.cnki.issn.1004-4523.202307046

Research on diagnosis and identification of acoustic emission signals fornon-contactrotary seal failures

CHEN Jinxin 1，2 ，LUJunjie¹，DINGXuexing²，XU Jie²，DING Junhua 2 ， GAO De¹ （1.Ningbo KeyLaboratory of Advanced Seal Technology，Universityof NingboTech，Ningbo 3150o，China; 2.School of Petrochemical Engineering，Lanzhou University of Technology，Lanzhou 73Oo5O，China）

Abstract：Toaddresstheisseofunclearandchalengingidentificationofnoncontactrotatingsealfult signals，this studyestab lishedanexperimentalplatformandacousticemissiontestingsystem.It involved monitoringacousticemisionsignalsduringVari ousoperationalconditions，including normaloperationandsix typicalfaultscenariosof noncontactrotatingseals.Atotalof 14000 feature samples wereefectivelycollcted.Byapplying the Bayesianoptimizationalgorithmandincorporatingcontinuous wavelettransform，anadaptiveconvolutional neuralnetwork clasification modelwasconstructed.Subsequently，thediagnostic performanceofthefault recognitionmodel was analyzedusingconfusion matrices andt-distributed stochasticneighborembedding. Theresearchresultsdemonstratethat thismodelsuccesullyclasifiesndidentifiesevendiferentoperationalconditionsofoncontactrotatingseals，icludingnoraloperation，dryfrction，mixedlubrication，springfilure，endfacepiting，localspngfil ure，and end-face scratching，with an average recognition accuracy of 99.7023% . This achievement underscores the capability of effectivelyisolatingandidentifngsealfaultsourcesfromacousticemissonsignalsofnncontactrotatingsealsinnonstatioary， complex，andoverlappingenvironments，therebyestablishingasolidtheoreticalfoundationforpracticalengineeringpliations.

Keywords： fault diagnosis;acoustic emission;non-contact rotary seal;condition recognition;deep learning

密封装置直接决定了重大装备的热控效果和运行效率，其中机械密封在先进工业设备密封装置中的使用率达到约 90%[1] 。（剩余12688字）