航空发动机结构与系统的复杂性导致轴承的故障诊断方法通常面临特征提取与模式识别的困难。针对以上不足,考虑实际工程诊断的实时性与准确性,提出了一种新的基于转子位移概率密度信息(probability density information of rotor displac...航空发动机结构与系统的复杂性导致轴承的故障诊断方法通常面临特征提取与模式识别的困难。针对以上不足,考虑实际工程诊断的实时性与准确性,提出了一种新的基于转子位移概率密度信息(probability density information of rotor displacement,PIRD)的航空发动机轴承智能故障诊断方法。其主要对一维卷积神经网络(1-dimensional convolutional neural network,1DCNN)模型进行改进,在传统的卷积层前面增加了PIRD的提取层,可以提取转子振动位移信号的概率密度信息,有效地降低了数据的冗余度,同时保留了故障监测的重要指标。提出的PIRD-CNN诊断模型保留了1DCNN端到端的故障诊断优势,将该模型在航空发动机试验台产生的轴承故障数据进行测试,其对轴承故障诊断精度可达96.58%,与基准研究相对比表明,PIRD-CNN能够快速且更加精准地诊断航空发动机轴承的故障。展开更多
Effects of substrate materials on the properties of nanofibrous polyaniline（PANI） film prepared by the pulse galvanostatic method（PGM） were investigated.The chronopotentiograms of aniline polymerization in 0.3 mol...Effects of substrate materials on the properties of nanofibrous polyaniline（PANI） film prepared by the pulse galvanostatic method（PGM） were investigated.The chronopotentiograms of aniline polymerization in 0.3 mol/L aniline+1 mol/L HNO3 aqueous solution showed that the anodic potential on Pt and Ru electrodes rised to 880 and 850 mV quickly when the mean current density was 1.0 mA/cm2.The potential turned to（750 mV） after about 40 s and kept constant until the experiment was over.When the aniline polymerization occurred on stainless steel（SS） and Al electrodes,the anodic potential climbed to 1 100 and 1 700 mV respectively,and after about 55 and 250 s it turned to 750 mV.The scanning electron microscopic images demonstrated that the PANI films on Pt,Ru,SS and Al prepared by PGM all exhibited a similar fibrous morphology with a diameter of 80-100 nm.Therefore,it can be considered that the PGM polymerization of aniline on different substrate materials was markedly distinct when the surface of electrode was not covered by PANI completely.The cyclic voltammograms and electrochemical impedance spectroscopy results showed that PANI films on various substrates presented different electrochemical reactivities in an aqueous aniline free solution of（1 mol/L） HNO3.展开更多
文摘航空发动机结构与系统的复杂性导致轴承的故障诊断方法通常面临特征提取与模式识别的困难。针对以上不足,考虑实际工程诊断的实时性与准确性,提出了一种新的基于转子位移概率密度信息(probability density information of rotor displacement,PIRD)的航空发动机轴承智能故障诊断方法。其主要对一维卷积神经网络(1-dimensional convolutional neural network,1DCNN)模型进行改进,在传统的卷积层前面增加了PIRD的提取层,可以提取转子振动位移信号的概率密度信息,有效地降低了数据的冗余度,同时保留了故障监测的重要指标。提出的PIRD-CNN诊断模型保留了1DCNN端到端的故障诊断优势,将该模型在航空发动机试验台产生的轴承故障数据进行测试,其对轴承故障诊断精度可达96.58%,与基准研究相对比表明,PIRD-CNN能够快速且更加精准地诊断航空发动机轴承的故障。
文摘Effects of substrate materials on the properties of nanofibrous polyaniline（PANI） film prepared by the pulse galvanostatic method（PGM） were investigated.The chronopotentiograms of aniline polymerization in 0.3 mol/L aniline+1 mol/L HNO3 aqueous solution showed that the anodic potential on Pt and Ru electrodes rised to 880 and 850 mV quickly when the mean current density was 1.0 mA/cm2.The potential turned to（750 mV） after about 40 s and kept constant until the experiment was over.When the aniline polymerization occurred on stainless steel（SS） and Al electrodes,the anodic potential climbed to 1 100 and 1 700 mV respectively,and after about 55 and 250 s it turned to 750 mV.The scanning electron microscopic images demonstrated that the PANI films on Pt,Ru,SS and Al prepared by PGM all exhibited a similar fibrous morphology with a diameter of 80-100 nm.Therefore,it can be considered that the PGM polymerization of aniline on different substrate materials was markedly distinct when the surface of electrode was not covered by PANI completely.The cyclic voltammograms and electrochemical impedance spectroscopy results showed that PANI films on various substrates presented different electrochemical reactivities in an aqueous aniline free solution of（1 mol/L） HNO3.
