基于信息融合的校车自动变速器故障诊断新平台与方法  

On New Platform and Method of School Bus Automatic Transmission Information Fusion Fault Diagnosis

作　　者：李爽[1] 李培[1] 

机构地区：[1]郑州旅游职业学院信息工程系,郑州450009

出　　处：《西南师范大学学报（自然科学版）》2015年第9期49-54,共6页Journal of Southwest China Normal University(Natural Science Edition)

摘　　要：为保证校车安全,在强电磁干扰环境下实现高速大容量的数据传输、信息融合与故障诊断,建立了基于光纤CAN(Controller Area Network)网络的数据通信和信息融合的故障检测综合平台,通过该平台实现了节点中各传感器数据的实时采集和融合诊断处理.将BP神经网络(BPNN)和D-S(Dempster-Shafer)证据进行有机结合,然后将BPNN的初级诊断结果处理后,作为D-S证据的基本可信度分配,诊断结论通过D-S二次融合后输出.采用该方法在校车自动变速器的故障诊断中进行实验,结果表明,该平台具有很高的可靠性,且该融合诊断方法能够对故障数据的冗余和互补信息进行有效的处理,引入BPNN和D-S证据之后,综合诊断结果的准确性和可靠性比单一神经网络有了很大的提高.In order to ensure the safety of school buses and achieve a high-speed and large-capacity data transfer,information fusion and fault diagnosis under the environment of strong electromagnetic interference,an integrated fault detect platform for data communication and information fusion has been established based on CAN network,which can help realize a real-time data collection of all sensors as well as fusion diagnosis within nodes.BP Neural Network（BPNN）and D-S evidence have been integrated organically.The primary BPNN diagnosis results after processed are distributed as the basic reliability of D-S Evidence,and then after the secondary fusion with D-S,the diagnosis will be output.The method has been taken as an experiment in automatic transmission fault diagnosis of school buses.It shows that this platform is of high reliability and this fusion diagnosis method can effectively process the redundancy of fault data as well as complementary information.After the introduction of BPNN and D-S Evidence,integrated diagnosis results of this method are more accurate and reliable than that of the single neural network.

关 键 词：CAN通信 信息融合 故障诊断 BPNN D-S证据 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]