贝叶斯优化TQWT 参数在轴承故障诊断中的应用.pdfVIP

贝叶斯优化TQWT参数在轴承故障诊断中的应用

摘要：针对可调品质因子小波变换(TunableQ-factorwavelettrans-form，TQWT)采用网格有哪些信誉好的足球投注网站和优

化算法调参存在评估计算代价高的问题，提出基于贝叶斯优化TQWT参数的故障诊断算法。通过

贝叶斯优化算法在TQWT参数空间内求取熵-峭综合目标函数最优解，据此设置TQWT参数分解

轴承故障信号，选择熵-峭指标最小值对应子带信号，经TQWT逆变换后进行包络解调分析，最终由

重构信号包络谱判别轴承故障类型。仿真实验和实测轴承信号分析表明，该算法可以准确提取轴

承故障特征频率信息，实现早期故障诊断。

关键词：贝叶斯优化；TQWT；熵-峭指标；故障诊断

ApplyingBayesianOptimizationofParametersofTunable

Quality-FactorWaveletTransformtoBearingFault

Abstract：Itiscostlytousethegridsearchandoptimizationalgorithmtotunetheparametersoftunablequality-factorwavelet

transform(TQWT).AmethodforbearingfaultdiagnosisbasedontheBayesianoptimizationofTQWTparameterswasproposed.

Theoptimalsolutionoftheentropy-kurtosissyntheticobjectivefunctionwassolvedbyusingtheBayesianoptimizationalgorithm

inthespaceofTQWTparameters,accordingtowhichtheTQWTparametersweresettodecomposetheoriginalbearingfault

signals.Thesub-bandsignalwiththeminimumvalueoftheentropy-kurtosisindexwasselectedtoreconstructitsfeaturesignals

withtheinverseTQWTtransform,andthesignalwasthenprocessedwithanenvelopedemodulationalgorithm.Thetypeofbearing

faultwasjudgedwiththereconstructedfeaturesignalenvelopespectrum.Thesimulationresultsontheactuallymeasuredbearing

vibrationsignalsandtheiranalysisshowthattheproposedmethodcanaccuratelyextractthecharacteristicfrequencyinformation

onfaultanddiagnosebearingfaultsatanearlystage.

Keywords：bayesianoptimization；TQWT；entropy-kurtosisindex；faultdiagnosis

轴承是旋转机械的核心部件，研究早期故障信护设备安全稳定运行十分重要。

号特征，识别故障类型，采取合理的应对措施，对维可调品