期刊
CLUSTER COMPUTING-THE JOURNAL OF NETWORKS SOFTWARE TOOLS AND APPLICATIONS
卷 20, 期 2, 页码 1691-1702出版社
SPRINGER
DOI: 10.1007/s10586-017-0748-0
关键词
Marine diesel engine; Fault diagnosis; Support vector machine; Association rule mining
资金
- National Natural Science Foundation of China [61673129]
- Natural Science Foundation of Heilongjiang Province of China [F201414]
- Fundamental Research Funds for the Central Universities [HEUCF160418]
This paper proposes a novel approach for marine diesel engine fault diagnosis based on data mining. First, a marine diesel engine is divided into four subsystems according to the basic structure and fault features of the engine, which overcomes the difficulty of fault diagnosis caused by the structure complexity. Second, fault diagnosis of each subsystem is performed accordingly by employing the support vector machine algorithm, and the classification models are trained using the historical fault features. Third, association analysis of the fault features is achieved efficiently by applying the association rule mining algorithm and the implied relationship of faults among subsystems are mined using the transaction database of faults. A simulation system is implemented in Matlab to achieve real-time marine diesel engine state monitoring and fault diagnosis.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据