基于激光诱导击穿光谱与随机森林识别GCr15钢的硬度

Hardness Characterization of GCrl5 Steel Based on Laser-Induced Breakdown Spectroscopy and Random Forest

通过热处理工艺制备了6种不同回火温度的GCr15钢样品,通过对样品的光谱特性进行分析,发现谱线强度的比值与样品的硬度之间存在线性相关性,并且这种相关性的大小与所选择的分析谱线有关。因此,提出将激光诱导击穿光谱与随机森林算法相结合的方法,对样品的硬度进行研究。利用两种特征选择方法建立了不同的随机森林模型,结果表明,当基于主成分分析选择特征时,所建立的随机森林模型不能对样品的硬度进行正确的识别,而当基于变量重要性提取特征时,所建立的随机森林模型能有效识别样品硬度,并且调节随机森林模型的参数可以使模型的性能得到进一步提高。研究结果表明,激光诱导击穿光谱与随机森林算法相结合是一种新颖的硬度测量技术,可以应用于工程中钢的性能评估。

Objective Bearing is an important part of modern machinery and equipment,in which the working conditions are much complicated.Therefore,there exist extremely high requirements for the mechanical performance of bearings.As one of the important properties of bearings,hardness needs to be tested to meet these requirements during the production process.The traditional hardness test is a mostly destructive one,which requires the structural destruction of the original component,and the use of mechanical means is done to cut and sample the parts to be tested for the laboratory analysis.This test method is not only slow and heavy,but also damages the samples.The laser-induced breakdown spectroscopy(LIBS)technology,as an atomic emission spectroscopy technology,has the advantages of fast analysis speed,almost no sample processing,and less damage to samples,which is widely used in environmental monitoring,material characteristic research,coal quality analysis,and other fields.The interaction between laser and substance and the plasma characteristics are affected by the physical and chemical material characteristics of the samples to be tested.This effect causes the difference in the characteristics of emission spectra.This difference in the material spectra can affect the characteristics of the material.However,although the spectral line intensity ratio can be used to characterize the hardness of the samples,there are still some shortcomings such as difficulty in spectral line selection and low accuracy.Therefore,in recent years,with the development of machine learning,combining algorithms with LIBS can make up for the shortcomings of the traditional LIBS analysis methods.As an integrated learning method,the random forest(RF)algorithm can efficiently extract and generalize spectral features.This research attempts to use LIBS combined with the RF algorithm(LIBS-RF)to estimate the hardness of steel samples.Methods In order to measure the hardness of GCrl5 steels,a method combining LIBF and RF is proposed to study the hardness of the samples.First,six GCrl5 steel samples with different tempering temperatures are prepared by the heat-treatment process,and an LIBS experimental device is built.A 10 × 10 lattice is scanned with 300 laser pulses per point on each sample surface to obtain the original spectral data.Then the spectral characteristics of the samples are analyzed,and a correlation between the spectral line intensity ratio and the hardness is established.Finally,a RF model is established based on the principal component analysis(PCA)method and the variable importance method.Results and Discussions The RF model is established based on the PCA method,and the result is shown in Fig.8.It can be found that the RF model is established after PCA dimensionality reduction.The highest prediction accuracy for sample S2 is 69%(20/29)followed by 52%(12/23)for sample S3,and the rest prediction accuracy rates of the samples are all low.The prediction accuracy rates of samples S5 and S6 are almost zero.RF is used to reduce data dimensionality based on the importance of features and extract the effective spectral features.The prediction results of the model are shown in Table 5.Compared with the model based on the PCA method,the accuracy rate is significantly improved,and its average accuracy reaches 0.96.In addition,through the study on the effects of the number of decision trees(Fig.9)and the number of random features(Fig.10)on the performance of the model,it can be found that the number of decision trees and the number of random features are not that the larger the better.With the increase of the number of decision trees and the number of random features,the prediction accuracy of the model first increases to a certain value and then remains relatively stable.Conclusions In this paper,the laser-induced breakdown spectroscopy and random forest are combined to study the GCr15 steel samples with different hardness.Through the linear relationship between the spectral line intensity ratio(I_(Fe Ⅱ316.786 /I_(Fe Ⅰ 1375.745) and I_(Cr Ⅱ 482.413)/I_(Cr Ⅰ1302.067))and the sample hardness,the experimental results show that the spectral line intensity ratio of the matrix element to the alloy element and the hardness of the sample present a linear correlation,in which the linear correlation between I_(Fe Ⅱ 316.786/I_(Fe Ⅰ375.745) and hardness is higher,indicating that the method of the spectral line intensity ratio has a certain dependence on the selection of spectral lines.The LIBS-RF method is proposed to estimate the hardness of the sample.First,PCA is used to reduce the dimensionality of the original data,and subsequently a random forest model is established.It is found that the hardness of the sample can not be effectively predicted.Then,the random forest is used to select feature spectra based on the importance of the variables to establish a random forest model.The results show that the prediction accuracy of the LIBS-RF model based on the full-spectrum data is lower than that based on the partial characteristic spectrum data.This is because the full-spectrum data also contains a lot of noise and other redundant information,which also participates in the training of the model and results in a decrease in model accuracy.In addition,it is found that as the number of decision trees and the number of random features increase,the prediction accuracy of the model increases to a certain value and remains relatively stable.Based on this,the parameters of the model can be adjusted reasonably while satisfying the accuracy requirements,the complexity of the algorithm is reduced,and the efficiency of the algorithm is improved.Above all,as a novel hardness measurement technology,LIBS-RF has the advantages of simpler and faster than the traditional LIBS hardness measurement,and the results here provide a theoretical basis for engineering practice applications.