摘要
为了监测和评估管线钢服役状态,以应用最为广泛的贝氏体+多边形铁素体双相结构X80管线钢作为研究对象,利用声发射技术对其塑性变形过程的损伤状态进行监测与评估。通过使用sym4系小波、Birge-Massart算法和小波包分解算法对声发射信号进行去噪和特征提取处理,并采用相关性分析结合LightGBM算法对提取后的特征进行选择。基于能量耗散模型,并考虑在塑性损伤过程中弹性模量衰减的因素,定义了损伤变量来定量表征塑性损伤程度。采用径向基函数神经网络实现了特征参量和损伤变量的映射关系。结果表明,该方法能够对服役中的管线钢进行监测和评估。
To monitor and evaluate the service status of pipeline steel,taking the most widely used dual-phase structure X80 pipeline steel of bainite and polygonal ferrite(B+PF)as the reaearch object,the acoustic emission(AE)technology was used to monitor and evaluate its damage state during plastic deformation process.The acoustic emission signal denoising and feature extraction were processed using sym4 wavelet,Birge-Massart algorithm and wavelet packet decomposition algorithm,and the extracted features were selected by correlation analysis and LightGBM algorithm.Based on the energy dissipation model and considering the attenuation of elastic modulus during plastic damage process,the damage variables were defined to quantitatively characterize the plastic damage degree.The radial basis function neural network was used to realize the mapping relationship between characteristic parameters and damage variables.The results show that the method can monitor and evaluate pipeline steel in service.
作者
宋宇佳
范利锋
杨学灵
苟建军
王森
秦淑淇
SONG Yu-jia;FAN Li-feng;YANG Xue-ling;GOU Jian-jun;WANG Sen;QIN Shu-qi(Transportation Institute,Inner Mongolia University,Hohhot 010021,China;School of Electronic Information Engineering,Inner Mongolia University,Hohhot 010021,China)
出处
《塑性工程学报》
CAS
CSCD
北大核心
2023年第9期204-210,共7页
Journal of Plasticity Engineering
基金
国家自然科学基金资助项目(51761030
52161022)
内蒙古自然科学基金资助项目(2019MS05081)
内蒙古大学自治区级大学生创新创业训练计划(202010126070)
内蒙古自治区高等学校“青年英才支持项目”(NJYT23045)。
关键词
X80管线钢
声发射检测
塑性损伤
小波分析
径向基神经网络
X80 pipeline steel
acoustic emission testing
plastic damage
wavelet analysis
radial basis neural network