基于神经网络的管道损伤识别分析  

Analyze on damage identification of pipeline based on neural network

作　　者：马燚[1] 

机构地区：[1]沈阳市建设工程质量检测中心,辽宁沈阳110004

出　　处：《山西建筑》2017年第35期22-24,共3页Shanxi Architecture

摘　　要：利用有限元软件对管道模拟,获得模态参数作为神经网络的输入样本,采用BP,RBF网络的方法对损伤前后的某管道工程进行分析,对比并探讨两种方法在管道损伤识别中的训练结果,实现了对管道的损伤定位和损伤程度的识别。结果表明对于小样本径向基神经网络(RBF)较BP网络有较大的优势。The modal parameters as damage identification are received by the FEM. The pipeline engineering damaged and undamaged is analyzedthrough the BP and RBF network method, and the results of the pipeline damage identification of these two methods are discussed. Theidentification of the pipeline damage location and damage degree are realized, and the results indicate that the RBF neural network is more advantagethan the BP network to algorithm for small sample.

关 键 词：径向基(RBF)神经网络 BP神经网络 管道 损伤识别 

分 类 号：U177[交通运输工程]