摘要
针对在役斜拉桥索只能进行无损检测的特点,设计了基于漏磁检测的缆索缺陷检测系统.实测结果表明,缺陷产生的漏磁信号往往附有大量的噪声信号,导致信号特征不显著.为此,应用信号平滑、小波包差分超限滤波和BP神经网络模式识别等信号分析方法,对采集信息进行滤波处理,以获取比较显著的特征量化信号.仿真结果表明,经过滤波处理后的漏磁信号特征明显,为桥梁斜拉索的断丝缺陷检测和分析提供了可靠依据.
An inspection system is offered for stay cables by magnetic leakage field, due for which the nondestructive test only. The measured results have showed that the MLF signals for cables fault is very unclear with too many noisy signals, so the analytical method of smooth process, wavelet difference transfinite values filter and BP neural network pattern recognition is applied for measured data to obtain the distinct character signals. The simulation results showed that the default signals have a distinct character by that filter process, so that the credible proof will be given about detection and analysis to the broken wires for the stay cable.
出处
《武汉理工大学学报（交通科学与工程版）》
2013年第6期1255-1258,共4页
Journal of Wuhan University of Technology（Transportation Science & Engineering）
基金
广东省交通运输厅科技项目资助(批准号:2011-02-031)
关键词
斜拉桥索
信号平滑滤波
信噪分离
漏磁检测
stay cable
smooth filter signals
separation from signal and noise
magnetic leakage field inspection
