摘要
在船舶运输、石油化工等需要广泛使用各类型管道的行业中,管道的结构健康监测(structural health monitoring, SHM)对于工业系统的安全稳定运行意义重大。在基于超声导波的管道裂纹等级识别方面,建立了一个与实际管道基本一致的有限元模型,通过添加噪声的方式合成了更接近实际检测的导波数据。基于包含不同管道裂纹等级的有限元仿真数据库,提出了一种基于多尺度一维卷积神经网络(multi-scale one dimensional convolution neural network, MS-1DCNN)的管道裂纹等级识别模型,该模型以端到端的方法,将原始波形信号直接作为输入,无需专门设计信号降噪及特征提取算法。试验结果表明,该模型相较于传统机器学习方法在噪声环境下对管道裂纹等级的识别具有较高精度,并通过实物管道试验,验证了该模型在管道结构健康监测中的有效性。
In shipping, petrochemical and other industries needing to widely use various types of pipelines, structural health monitoring(SHM) of pipelines is of great significance for safe and stable operation of industrial systems. Here, for pipeline crack grade identification based on ultrasonic guided wave, a finite element model being basically consistent to actual pipeline was established, and guided wave data closer to actual detection were synthesized through adding noise. Based on the finite element simulation database containing different pipeline crack grades, a pipeline crack grade identification model based on multi-scale one dimensional convolution neural network(MS-1DCNN) was proposed. The model could use the end-to-end method to directly take the original waveform signal as input without specially designing signal denoising and feature extraction algorithms. Test results showed that the proposed model has higher accuracy compared with the traditional machine learning method to identify pipeline crack grade under noise environment;the effectiveness of the proposed model in pipeline SHM is verified through actual pipeline tests.
作者
唐若笠
张尚煜
伍文君
TANG Ruoli;ZHANG Shangyu;WU Wenjun(School of Naval Architecture,Ocean and Energy Power Engineering,Wuhan University of Technology,Wuhan 430063,China)
出处
《振动与冲击》
EI
CSCD
北大核心
2023年第5期183-189,共7页
Journal of Vibration and Shock
基金
国家自然科学基金青年基金项目(51709216)。
关键词
超声导波
有限元仿真
卷积神经网络
裂纹等级
ultrasonic guided wave
finite element simulation
convolutional neural network(CNN)
crack grade