船电技术
船電技術
선전기술
2012年
6期
9-11
,共3页
船用柴油机%故障%信息%融合
船用柴油機%故障%信息%融閤
선용시유궤%고장%신식%융합
Marine Engine%Fault diagnosis%Information%fusion
为了解决船用柴油机故障诊断中基于单传感器信息的方法诊断精度低的缺点，应用神经网络原理，提出了一种基于气缸压力，缸盖振动信号和燃油压力等多传感器信息融合的喷油器故障诊断新方法。通过提取船用柴油机工作过程故障三种信号的八个特征值，按正常和五种故障状态的构造学习样本文集和检验样本文集，对输入进行归一化处理，该方法能有效地提高其故障诊断精度。
為瞭解決船用柴油機故障診斷中基于單傳感器信息的方法診斷精度低的缺點，應用神經網絡原理，提齣瞭一種基于氣缸壓力，缸蓋振動信號和燃油壓力等多傳感器信息融閤的噴油器故障診斷新方法。通過提取船用柴油機工作過程故障三種信號的八箇特徵值，按正常和五種故障狀態的構造學習樣本文集和檢驗樣本文集，對輸入進行歸一化處理，該方法能有效地提高其故障診斷精度。
위료해결선용시유궤고장진단중기우단전감기신식적방법진단정도저적결점，응용신경망락원리，제출료일충기우기항압력，항개진동신호화연유압력등다전감기신식융합적분유기고장진단신방법。통과제취선용시유궤공작과정고장삼충신호적팔개특정치，안정상화오충고장상태적구조학습양본문집화검험양본문집，대수입진행귀일화처리，해방법능유효지제고기고장진단정도。
In order to solve the non-linear and uncertain faults of engine, a new in-cylinder engine faults diagnosis method based on multi-sensor information fusion is presented by using the theory of neural network. Through extracting eight features of three faults signals in the operating process of the marine diesel engine, three kinds of signals are sampled and analyzed using the pressure in cylinder, liberation of cover and the fuel pressure packet methods. Through this method, the fault diagnosis accuracy is improved effectively.
