神经网络信息融合用于电梯故障诊断的研究  

Study on fault diagnosis of elevator based on neural network information fusion technology

作　　者：李端[1] 艾永乐[1] 

机构地区：[1]河南理工大学电气工程与自动化学院,河南焦作454000

出　　处：《计算机工程与应用》2010年第14期231-234,共4页Computer Engineering and Applications

基　　金：国家高技术研究发展计划(863)No.2008AA01Z103;河南省国际合作基金资助(No.0646660006);河南理工大学实验室开放基金(No.SKJB07033)~~

摘　　要：针对传统电梯故障诊断系统实时性有限、故障定位准确率低等问题,将多信息融合技术引入到电梯故障诊断中来,建立了基于模糊神经网络和D-S证据理论相结合的故障诊断模型。为了提高神经网络的训练速度和推广能力,采用了正则化算法对BP网络算法进行修改,并且利用D-S证据理论对神经网络的诊断结果进行决策融合,仿真结果表明了此方法有效地提高了故障诊断的准确率。Researchers have studied on elevator fault diagnosis,but timeliness is limited,and can＇t discover reason of the faults accurately.To solve these problems,this paper introduces the multi-information fusion technology to the fault diagnosis of elevator control system.The fault diagnosis model：A combination of fuzzy neural network and D-S evidential theory is set up.The BP neural network＇s algorithm has been revised in order to enhances its study speed and the ability for precise diagnosis from the D-S evidential theory is fully made use of.The simulation of the fault phenomena has the good effect of this fault diagnosis system.

关 键 词：故障诊断 信息融合 模糊神经网络 D-S证据理论 

分 类 号：TU857[建筑科学]