基于电磁检测的火车车轮残余应力定量预测  

Quantitative Prediction of Residual Stress for Railroad Wheels Based on Electromagnetic Detection

作　　者：夏鹏 魏志辉 XIA Peng;WEI Zhi-hui(Central Iron&Steel Research Institute,Beijing 100081,China;NCS Testing Technology Co.,Ltd.,Beijing 100081,China)

机构地区：[1]钢铁研究总院,北京100081 [2]钢研纳克检测技术股份有限公司,北京100081

出　　处：《精密成形工程》2022年第9期136-141,共6页Journal of Netshape Forming Engineering

基　　金：国家重点研发计划(2018YFF01012300)。

摘　　要：目的为了突破火车车轮残余应力有损测试的局限性、实现车轮残余应力的准确定量预测,研究电磁参量特征值的遴选过程并建立相关模型。方法对比分析单一电磁参量(磁巴克豪森噪声或增量磁导率)和电磁参量融合(磁巴克豪森噪声和增量磁导率)的检测方法,通过逐步回归对电磁参量的特征值进行遴选,利用多元线性回归方法构建残余应力的定量预测模型。结果基于单一电磁参量建立的应力预测模型,其残余应力预测值与实际值的偏差超过±15 MPa的允差范围;基于电磁参量融合建立的应力预测模型,其残余应力预测值与实际值的偏差均在±15 MPa的允差范围之内。结论采用电磁融合方法建立的多元线性模型,可以有效提高模型精度、实现火车车轮残余应力的定量预测。The work aims to study the selection of characteristic values of electromagnetic parameters and the establishment of models to break through the limitation of destructive testing of residual stress of railroad wheels and realize the accurate and quantitative prediction of the residual stress of the wheels.Based on the testing methods of single electromagnetic parameter(Magnetic Barkhausen noise or incremental permeability)and electromagnetic parameter fusion(Magnetic Barkhausen noise and incremental permeability),the eigenvalues of electromagnetic parameters were selected by stepwise regression,and the quantitative prediction model of residual stress was constructed by multiple linear regression.It was found that the stress prediction model established based on single electromagnetic parameter had an experimental verification results of residual stress all exceed the tolerance range of±15 MPa;The experimental results of the model established based on electromagnetic parameter fusion were within the tolerance range of±15 MPa.In conclusion,the multivariate linear model established by electromagnetic fusion method can effectively improve the model accuracy and realize the quantitative prediction of residual stress of railroad wheels.

关 键 词：火车车轮 残余应力 电磁检测 多元线性回归 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP212[自动化与计算机技术—控制科学与工程]