基于矫正磁化和机器学习的管道应力检测技术研究  

Research on pipeline stress detection technology based on corrective magnetization and machine learning

作　　者：党赫 张来斌[1,2] 樊建春 高富民[1,2] 李睿 陈闻澜 DANG He;ZHANG Laibin;FAN Jianchun;GAO Fumin;LI Rui;CHEN Wenlan(College of Safety and Ocean Engineering,China University of Petroleum(Beijing),Beijing 102249,China;Key Laboratory of Oil and Gas Safety and Emergency Technology,Ministry of Emergency Management,Beijing 102249,China;PipeChina Science and Technology Research Institute,Tianjin 300457,China)

机构地区：[1]中国石油大学(北京)安全与海洋工程学院,北京102249 [2]应急管理部油气生产安全与应急技术重点实验室,北京102249 [3]国家管网集团科学技术研究总院分公司,天津300457

出　　处：《中国安全生产科学技术》2025年第4期27-33,共7页Journal of Safety Science and Technology

基　　金：国家重点研发计划项目(2022YFB3207600)。

摘　　要：为解决管道因初始磁化状态差异而导致的磁法定量应力检测难题,使用自研应力检测装置,对有无施加矫正磁化场条件下的受拉X80钢试样表面正交向多维磁信号进行采集并分析,基于提取的磁特征参数,采用冠豪猪算法优化的支持向量回归(CPO-SVR)建立应力预测模型。研究结果表明:带有直流偏置的正弦交变矫正磁化场能够有效削弱试样初始磁化状态差异,重复试验的最大偏差仅为未施加矫正磁化场条件下的24.0%;磁探头的差分测量模式不仅能够提高信号测量精度,还能够消除因试样边缘效应带来的信号干扰;所建立的应力预测模型表现出良好的精度,最小绝对误差为3.907 3 MPa。研究结果可为油气管道磁法应力检测技术发展提供工程参考。In order to solve the problem of quantitative magnetic stress detection in pipelines arising from difference in the initial magnetization state,the orthogonal multi-dimensional magnetic signals on the surface of tensile X80 steel specimens with and without corrective magnetization field were collected by a self-developed stress detection device and analyzed.A stress prediction model was established using the support vector regression optimized with the crown porcupine optimization algorithm(CPO-SVR)based on the extracted magnetic characteristic parameters.The results show that the sinusoidal alternating corrective magnetization field with DC bias can effectively weaken the difference in the initial magnetization state of specimen,and the maximum deviation of repeated test is only 24.0%of that without the corrective magnetization field.The differential measurement mode of magnetic probes can not only improve the signal measurement accuracy,but also eliminate the signal interference due to the edge effect of specimen.The stress prediction model presents good prediction accuracy,with a minimum absolute error of 3.9073 MPa.The research results can provide engineering reference for developing the magnetic stress detection technology of oil and gas pipelines.

关 键 词：管道安全 应力检测 力磁耦合 激励磁场 机器学习 

分 类 号：X937[环境科学与工程—安全科学]