波纹管内部脱浆状态的超声检测方法  

Ultrasonic detection method of desizing state inside corrugated pipe

作　　者：朱耀东 韩庆邦 林泽宙 尹琳丽 ZHU Yaodong;HAN Qingbang;LIN Zezhou;YIN Linli(College of IOT Engineering of Hohai University,Changzhou 213022,Jiangsu,China)

机构地区：[1]河海大学物联网工程学院,江苏常州213022

出　　处：《声学技术》2021年第2期210-215,共6页Technical Acoustics

基　　金：声场声信息国家重点实验室开放课题研究基金(SKLA201913);国家自然科学基金(11574072)。

摘　　要：为了检测波纹管内部不同脱浆状态的缺陷问题,提出了一种采用变分模态分解(Variational Mode Decomposition,VMD)与蚁群神经网络相结合的超声检测方法。将检测到的回波信号进行变分模态分解,将分解后信号的多尺度样本熵作为特征参数,输入到蚁群神经网络中进行检测。实验结果表明利用VMD分解方法与蚁群神经网络相结合可以对波纹管内部横向缺陷进行有效判断。In order to detect the defects of desizing states inside corrugated pipes,an ultrasonic detection method based on variational mode decomposition(VMD)and ant colony neural network is proposed.The detected echo signal is decomposed by VMD method,and the multi-scale entropy of the decomposed signal is taken as the characteristic parameter and input into the ant colony neural network for detection.The experimental results show that the combination of VMD method and ant colony neural network can effectively judge the transverse defects in corrugated pipes.

关 键 词：变分模态分解(VMD) 多尺度样本熵 蚁群神经网络 

分 类 号：TB551[理学—物理]