自动化仪表
自動化儀錶
자동화의표
PROCESS AUTOMATION INSTRUMENTATION
2014年
3期
35-38
,共4页
故障诊断%高维矩阵%PCA%特征向量%极限学习机
故障診斷%高維矩陣%PCA%特徵嚮量%極限學習機
고장진단%고유구진%PCA%특정향량%겁한학습궤
Fault diagnosis%High dimensional matrix%Principal component analysis( PCA)%Eigenvector%Extreme learning machine
针对转子故障诊断日趋困难的问题,引入一种基于非线性流形学习和极限学习机( ELM)的转子故障诊断模型。基于转子振动信号时域与频域的14个特征参数构建高维矩阵,利用线性局部切空间排列( LLTSA)提取高维矩阵的特征向量,并投影到可视空间中。采用极限学习机作为分类器进行故障诊断。故障诊断实例验证了该模型的有效性和可行性,表明了该模型将成为故障诊断领域的发展方向。
針對轉子故障診斷日趨睏難的問題,引入一種基于非線性流形學習和極限學習機( ELM)的轉子故障診斷模型。基于轉子振動信號時域與頻域的14箇特徵參數構建高維矩陣,利用線性跼部切空間排列( LLTSA)提取高維矩陣的特徵嚮量,併投影到可視空間中。採用極限學習機作為分類器進行故障診斷。故障診斷實例驗證瞭該模型的有效性和可行性,錶明瞭該模型將成為故障診斷領域的髮展方嚮。
침대전자고장진단일추곤난적문제,인입일충기우비선성류형학습화겁한학습궤( ELM)적전자고장진단모형。기우전자진동신호시역여빈역적14개특정삼수구건고유구진,이용선성국부절공간배렬( LLTSA)제취고유구진적특정향량,병투영도가시공간중。채용겁한학습궤작위분류기진행고장진단。고장진단실례험증료해모형적유효성화가행성,표명료해모형장성위고장진단영역적발전방향。
The rotor fault diagnosis is an increasingly difficult problem. The rotor fault diagnosis model of nonlinear manifold learning and extreme learning machine( ELM) is cited. Based on 14 feature parameters of rotor vibration signals in time domain and frequency domain, the high dimensional matrix is built, by adopting linear local tangent space alignment ( LLTSA) algorithm the eigenvectors of high dimensional matrxs are extracted, and projected onto the visual space. The extreme learning machine algorithm is used as the classifier to do the fault diagnosis. The practical examples of fault diagnosis verify the effectiveness and feasibility of the model, and it is shown that this model will definitely become the developing direction of fault diagnosis field.