计算机应用
計算機應用
계산궤응용
COMPUTER APPLICATION
2014年
z2期
184-185,229
,共3页
夏梦莹%刘啸奔%陈严飞%马虎强%郑伟
夏夢瑩%劉嘯奔%陳嚴飛%馬虎彊%鄭偉
하몽형%류소분%진엄비%마호강%정위
主成分分析%自组织映射神经网络%柴油机%故障诊断
主成分分析%自組織映射神經網絡%柴油機%故障診斷
주성분분석%자조직영사신경망락%시유궤%고장진단
Principle Component Analysis ( PCA)%SOM Neural network%diesel engine%fault diagnosis
提出了一种新型的柴油机故障诊断方法,该方法使用主成分分析( PCA)法对故障样本降维,有效提取故障样本主要特征,在此基础上,将其作为输入使用自组织映射( SOM)神经网络进行训练得到故障识别网络。400组模拟故障数据的测试表明,两者结合的方法能有效提高网络的训练速度,获得满意的故障识别率。
提齣瞭一種新型的柴油機故障診斷方法,該方法使用主成分分析( PCA)法對故障樣本降維,有效提取故障樣本主要特徵,在此基礎上,將其作為輸入使用自組織映射( SOM)神經網絡進行訓練得到故障識彆網絡。400組模擬故障數據的測試錶明,兩者結閤的方法能有效提高網絡的訓練速度,穫得滿意的故障識彆率。
제출료일충신형적시유궤고장진단방법,해방법사용주성분분석( PCA)법대고장양본강유,유효제취고장양본주요특정,재차기출상,장기작위수입사용자조직영사( SOM)신경망락진행훈련득도고장식별망락。400조모의고장수거적측시표명,량자결합적방법능유효제고망락적훈련속도,획득만의적고장식별솔。
A new method for diesel engine fault diagnosis was proposed here. Principal Component Analysis ( PCA) was used to decrease the dimension of the fault data in the first step. Based on the result of the first step, a Self-Organizing Map ( SOM) neural network was trained to identify the fault type. At last, a series of simulation fault data was tested by the network. From the test result, some important conclusions were gotten: the combination of PCA and SOM neural network can improve the training rate of the network, and this new method gets a satisfied fault recognition rate.