振动与冲击
振動與遲擊
진동여충격
JOURNAL OF VIBRATION AND SHOCK
2014年
3期
89-93
,共5页
故障诊断%特征提取%小波包分解%局部保留投影%高斯混合模型
故障診斷%特徵提取%小波包分解%跼部保留投影%高斯混閤模型
고장진단%특정제취%소파포분해%국부보류투영%고사혼합모형
fault diagnosis%feature extraction%wavelet packet decomposition (WPD )%locality preserving projection (LPP)%Gaussian mixture model (GMM)
小波包分解（WPD）能够将非平稳信号在低频和高频上同时分解以有效反映信号潜在的特征信息，而局部保留投影法（LPP）在降维的同时保留了信号的局部特征信息。结合上述特点，给出了选取信号小波包分解后形成全部节点的谱能量，作为表征信号的特征，采用LPP提取降维特征进行模式识别的方法进行设备故障分类研究。在多组不同轴承故障及同故障不同损伤程度的多类别数据集上进行了实验，实验结果验证了这种方法的有效性。
小波包分解（WPD）能夠將非平穩信號在低頻和高頻上同時分解以有效反映信號潛在的特徵信息，而跼部保留投影法（LPP）在降維的同時保留瞭信號的跼部特徵信息。結閤上述特點，給齣瞭選取信號小波包分解後形成全部節點的譜能量，作為錶徵信號的特徵，採用LPP提取降維特徵進行模式識彆的方法進行設備故障分類研究。在多組不同軸承故障及同故障不同損傷程度的多類彆數據集上進行瞭實驗，實驗結果驗證瞭這種方法的有效性。
소파포분해（WPD）능구장비평은신호재저빈화고빈상동시분해이유효반영신호잠재적특정신식，이국부보류투영법（LPP）재강유적동시보류료신호적국부특정신식。결합상술특점，급출료선취신호소파포분해후형성전부절점적보능량，작위표정신호적특정，채용LPP제취강유특정진행모식식별적방법진행설비고장분류연구。재다조불동축승고장급동고장불동손상정도적다유별수거집상진행료실험，실험결과험증료저충방법적유효성。
Wavelet packet decomposition (WPD)can be used to decompose a non-stationary signal within low and high frequency fields to reflect effectively the potential feature information of the signal.Locality preserving projection (LPP)can be used to retain local features of an analyzed signal during its dimension reduction.Here,combining these features,the spectral energy of all nodes after WPD was given as a characterization of the analyzed signal and LPP was used to extract characters after dimension reduction for pattern recognition of machine faults.The effectiveness of the proposed method was verified by using several multi-class data sets of bearing faults with different fault types and damage levels.
