摘要
用于机械摩擦副磨损状态监测的多传感器融合系统，是本世纪８０年代末问世的一项新技术．在讨论了融合系统的构成和不同传感器信息的预处理之后，提出了一种基于自组织神经网络的多传感器融合系统．将这种融合系统用于实际机械摩擦副磨损状态监测的结果证明，其正确识别率明显地比单一传感器的高，容错能力强，而且由于自组织神经网络有在线学习的功能。
The multisensor fusion system used for monitoring the wear state of the mechanical friction pairs is a new technique which come out in 1980s. After discussing the construction of fusion system and the pretreatment of different sensor information, it is suggested that the multisensor fusion system based on the self form nervous net. The results of monitoring the wear state of actual mechanical friction pairs with this fusion system showed that the correct recognition rate of the multisensor is higher than that of single sensor, and its allowance ability is also stronger. Owing to the self form nervous net have the on line function, the system can be used for monitoring on line.
出处
《摩擦学学报》
EI
CAS
CSCD
北大核心
1996年第4期367-370,共4页
Tribology
关键词
传感器
融合系统
神经网络
磨损
摩擦副
sensor fusion system nervous net wear monitoring failure diagnosis
