發(fā)布時(shí)間：2023-05-24 22:32

　　針對(duì)傳統(tǒng)的滾動(dòng)軸承故障診斷方法難以提取軸承振動(dòng)數(shù)據(jù)有效特征的缺陷,提出一種基于平滑偽Wigner-Vill分布(smooth and pseudo Wigner-Ville distribution,SPWVD)和卷積神經(jīng)網(wǎng)絡(luò)(convolutional neural network,CNN)的網(wǎng)絡(luò)模型SPWVD-CNN。對(duì)振動(dòng)數(shù)據(jù)進(jìn)行平滑偽Wigner-Vill分布變換,將獲得的時(shí)頻圖進(jìn)行壓縮,作為CNN的輸入,利用遷移學(xué)習(xí)的思想進(jìn)行網(wǎng)絡(luò)訓(xùn)練,使得模型對(duì)于不同負(fù)載的數(shù)據(jù)具有良好的診斷性能,提高了網(wǎng)絡(luò)的泛化能力。實(shí)驗(yàn)結(jié)果表明:SPWVD-CNN對(duì)軸承故障數(shù)據(jù)的平均分類準(zhǔn)確率提升至99. 27%,總體性能優(yōu)于使用單一的CNN和其他傳統(tǒng)的故障診斷方法。

【文章頁數(shù)】：8 頁

【文章目錄】：
1 Theoretical basis of the method
    1.1 Smooth and pseudo Wigner-Vill distribution
    1.2 Convolution neural network
2 Modeling and analysis of fault diagnosis of rolling bearings based on SPWVD-CNN
    2.1 Description of the problem
    2.2 Fault diagnosis process
    2.3 Network training method
3 Case analysis
    3.1 Structure data sets
    3.2 Analysis of diagnostic performance of SPWVD-CNN under different working conditions
    3.3 Analysis of diagnostic performance of SPWVD-CNN and traditional methods
4 Conclusion



本文編號(hào)：3822409