PSO-ICA算法及其在机械故障诊断中应用.pdfVIP

振动与冲击 第29卷第S期 JOURNALOFVIBRATl0NANDSHOCK、，01．29No．$2010 PSO．ICA算法及其在机械故障诊断中的应用 李豫川，柳小勤，伍星，迟毅林 (昆明理工大学机电工程学院，昆明650093) 摘要：提出一种基于独立分量分析(IcA)和微粒群优化(PSO)相结合的算法，该算法以分离信号的负熵为目标函数，采用 微粒群优化算法进行全局、并行寻优。该算法针对微粒群算法在达到一定精度后全体微粒爬坡能力低下的不足，提出了迭代 次数与分离精度相结合的终止标准，提高了算法的计算速度与灵活性。仿真及实测信号分离结果表明该算法对混合信号具备 较好的盲分离能力。 关键词：微粒群优化；独立分量分析；故障诊断 中图分类号：TH212；TH213．3文献标识码：A ANDITS MACHINEFAULT PSO．ICAALGoRITHMAPPLICATl0N0F DIAGNOSIS L1 Yilin Yuchuan·LIUXiaoqin。WUXing，CHI ofMechanicalElectrical of (Faculty Engineering，KunmingUnivemityScienceTechnology，Kunming650093) Abstract：An on swarm ofindependentcomponent algorithm analysis(ICA)basedparticle optimization(PSO)isproposed，which takesthe of asthecontrastfunctionand in and swarm separatedsignal negen订opy optimizesglobalparallelwaybyparticle optimization． usedto the in drawbackofPSOthatwhentheswarmto factoris ofswarm,andviewofthe acertain Convergence updatevelocity gets termination with and haveaweak of criteriacombinediterationnumber is precision,it capacityclimbing，a separationprecisionproposed, which the and of simulationand showthatthePSO-ICA has to improves flexibilityalgorithm．The experiment algorithmhighability speed the source fromtheirmixtures． separateoriginalsignals Key swarm component words：particleoptimization；independentanalysis；faultdiagnosis 振动信号作为设备状态