首页|基于便携式激光诱导击穿光谱的12Cr1MoV钢晶粒度等级评估

基于便携式激光诱导击穿光谱的12Cr1MoV钢晶粒度等级评估

扫码查看
原文链接
  • NETL
  • NSTL
  • 万方数据
激光诱导击穿光谱(LIBS)作为一种原子光谱技术,在金属失效检测领域表现出巨大的潜力.作为衡量金属老化重要指标之一的晶粒度等级,应用LIBS技术实现对其检测对失效特性预测具有重要意义.本文选用工业中广泛使用的12Cr1MoV钢为对象,利用便携式LIBS设备对金属的晶粒度等级进行评估.首先对光谱数据分别尝试应用主成分分析(PCA)、线性判别式分析(LDA)及递归特征消除(RFE)-LDA进行数据的降维处理,然后基于降维后的数据应用支持向量机(SVM)及多层感知机(MLP)分类算法建立金属晶粒度等级评估模型,并探究了数据经过RFE初步筛选后剩余特征数量对所建模型分类性能的影响.结果表明RFE与LDA组合方法,能较好地提升评估模型的分类准确度,同时发现在该组合的基础上进一步结合MLP分类算法所构建的模型分类准确率最高,达94.05％.该建模方案能有效地实现基于便携式LIBS设备的12Cr1MoV钢晶粒度等级评估.
Evaluation of grain size grade of 12CrlMoV steel based on portable laser-induced breakdown spectroscopy
Laser-induced breakdown spectroscopy(LIBS)is an atomic spectroscopy technique,and it has shown great potential in the field of metal failure detection.As one of the important indicators to measure metal aging,the detection of grain size grade by LIBS is of great significance for the prediction of failure characteristics.In this study,12CrlMoV steel,which is widely used in industry,was selected and the grain size grade of the metal was evaluated using a portable LIBS device.Firstly,principal component analysis(PCA),linear discriminant analysis(LDA),recursive feature elimination(RFE)-LDA were used to reduce the dimensionality of the spectral data.Then the metal grain size grade evaluation model was established based on the data after dimensionality reduction using the support vector machine(SVM)and multilayer perceptron(MLP)classification algorithms.The influence of the remaining features on the classification performance of the model after the preliminary RFE screening of the data was explored.The results showed that the combination of RFE and LDA could improve the classification accuracy of the evaluation model,and it was found that the model constructed by further combination with MLP classification algorithm had the highest classification accuracy,which reached 94.05％.The proposed modeling scheme could effectively realize the evaluation of grain size grade of 12CrlMoV steel based on portable LIBS equipment.

portable laser-induced breakdown spectroscopygrain size gradesupport vector machine(SVM)multilayer perceptron(MLP)12CrlMoV steelprincipal component analysis(PCA)linear dis-criminant analysis(LDA)recursive feature elimination(RFE)

陆盛资、董美蓉、唐飞强、王磊、尚子瀚、蔡俊斌

展开 >

广东省特种设备检测研究院,广东佛山 528251

华南理工大学电力学院,广东广州 510640

便携式激光诱导击穿光谱 晶粒度等级 支持向量机(SVM) 多层感知机(MLP) 12Cr1MoV钢 主成分分析(PCA) 线性判别式分析(LDA) 递归特征消除(RFE)

广东省市场监督管理局科技项目

2023CT02

2024

10.13228/j.boyuan.issn1000-7571.012388

冶金分析
中国钢研科技集团有限公司（钢铁研究总院） 中国金属学会

冶金分析

CSTPCD北大核心
影响因子：1.124
ISSN：1000-7571
年,卷(期)：2024.44(7)