基于小波和神经网络的石油井架损伤定位研究  

Study on Damage Location of Oil Derrick Based on Wavelet and Neural Network

作　　者：邹龙庆[1] 付海龙[1] 谢春强[1] 

机构地区：[1]大庆石油学院机械科学与工程学院,大庆163318

出　　处：《科学技术与工程》2009年第4期870-873,883,共5页Science Technology and Engineering

基　　金：黑龙江省自然科学基金(E200509)资助

摘　　要：针对石油井架单一损伤位置识别问题,结合小波和神经网络分析方法,建立两层BP神经网络,将小波包能量分析得到的归一化能量特征向量作为网络的输入向量,经过网络训练和仿真测试,证明其在识别单一损伤位置时,具有较好的实用性。同时对石油井架相似模型进行了锤击振动试验,说明基于归一化能量特征向量的损伤评价指标可以准确实现石油井架结构损伤位置的定位识别。According to the problem of single damage location of oil derrick, the two-layer BP neural network is build by the combination of wavelet and neural network method. The eigenvectors of normalized energy are as the neural network＇ s inputing vectors, which are obtained from the wavelet analysis. After network training and simulation test, it is concluded that the method above has better practicability to identify single damage location of oil derrick. The vibration experiment beating the indoor derrick model with a hammer is completed,which shows that the damage evaluation guide line based on eigenvectors of normalized energy can achieve to identify damage location of oil derrick well and truly.

关 键 词：石油井架 小波分析 神经网络 损伤 

分 类 号：TE923[石油与天然气工程—石油机械设备] TP391.76[自动化与计算机技术—计算机应用技术]