2023年第5期仪表技术与传感器InstrumentTechniqueandSensor2023No．5基金项目:山东省自然科学基金(ZＲ2020QE158)收稿日期:2022－08－31基于双树复小波包变换的钢丝绳断丝损伤信号特征信息提取研究朱良1，孙艺哲1，谢波1，刘艳蕊1，井陆阳2(1．火箭军士官学校，山东潍坊262500;2．青岛理工大学机械与汽车工程学院，山东青岛266520)摘要:针对钢丝绳断丝损伤信号特征信息难以有效提取的问题，提出一种基于双树复小波包变换与奇异值分解相结合的时频域特征信息提取方法。首先将钢丝绳断丝损伤信号采用双树复小波包变换为等长的频带子信号，构造时频域空间状态矩阵，然后采用奇异值方法提取各频带子信号的奇异值，组成表征各类损伤状态的特征向量，得到钢丝绳断丝损伤信号的特征信息矩阵。采用距离可分离性判据与传统时域特征信息提取方法相比较，结果表明双数复小波包变换与奇异值分解相结合的特征信息提取方法具有更强的分类能力。关键词:钢丝绳;特征信息;双树复小波包变换;奇异值分解;距离可分离性判据;分类能力中图分类号:TH865文献标识码:A文章编号:1002－1841(2023)05－0090－07ＲesearchonFeatureInformationExtractionofBrokenWireDamageSignalofSteelWireＲopesBasedonDualTreeComplexWaveletPacketTransformZHULiang1，SUNYi-zhe1，XIEBo1，LIUYan-rui1，JINGLu-yang2(1．PLAＲocketForceNCOCollege，Weifang262500China;2．SchoolofMechanical＆AutomotiveEngineering，QingdaoUniversityofTechnology，Qingdao266520China)Abstract:Aimingattheproblemthatitisdifficulttoextractthecharacteristicinformationofthebrokenwiredamagesignalofsteelwireropeseffectively，atime-frequencydomaincharacteristicinformationextractionmethodbasedonthecombinationofdualtreecomplexwaveletpackettransformandsingularvaluedecompositionwasproposed．Firstly，thebrokenwiredamagesignalofsteelwireropeswastransformedintoequallengthfrequencybandsubsignalsbydualtreecomplexwaveletpacket，andthetime-frequencydomainspatialstatematrixwasconstructed．Thenthesingularvaluemethodwasusedtoextractthesingularvalueofeachfrequencybandsignal，formedthecharacteristicvectorrepresentingvariousdamagestates，andobtainedthecharacteristicinformationmatrixofwirebreakdamagesignalofsteelwireropes．Thedistanceseparabilitycriterionw...
