Composition Analysis Of Liquid Steel Slag Based On Laser-induced Breakdown Spectroscopy

As an important industrial raw material of our country,iron and steel plays an important role in many fields,and steel slag,as an important phase involved in various metallurgical reactions in the steelmaking process,has an important role in desulfurization,reduction and deoxidation.Steel slag plays a different role in steelmaking process due to its different composition,so the rapid analysis of steel slag composition is particularly important.At present,the commonly used detection methods include X-ray fluorescence spectroscopy,chemical analysis and other methods,but these methods can not achieve online analysis because of their limitations.Laser induced breakdown spectroscopy(LIBS)is a new detection technology,which has the advantages of simple and convenient,real-time online,small damage to the analysis sample,etc.The online analysis capability of this technology is not available in the above detection methods.In this thesis,the steel slag sample of a steel plant is taken as the research object,the laserinduced breakdown spectroscopy technology is used to detect liquid steel slag under high temperature conditions in real time,study the status of steel slag under different conditions and the factors affecting the quality of the spectral data of steel slag,find the appropriate conditions to collect the spectral data,and then conduct quantitative analysis of steel slag,laying the foundation for the subsequent improvement of the analysis accuracy of steel slag composition.The research contents of this thesis are as follows: The experimental conditions of liquid steel slag were analyzed,and the laser-induced breakdown spectroscopy was used to detect and analyze the steel slag under different conditions.After preprocessing the original spectral data,the state of steel slag under high temperature and the factors affecting the spectral data were analyzed according to the changes of the spectral data under different conditions.Finally,Fe,Si,Ca,Mg,Al and Mn in steel slag were quantitatively analyzed.The detailed research contents are as follows:The research background and significance of this thesis are introduced,the working principle and characteristics of laser-induced breakdown spectroscopy are expounded,and the research status of steel slag analysis by LIBS technology at home and abroad is analyzed.Design the experiment and analyze the experimental parameters.According to the experimental objectives and equipment functions,experiments of steel slag detection under different conditions were designed,and the indexes for evaluating the spectral data were introduced,which laid the foundation for the subsequent exploration of the factors affecting the spectral data of steel slag.Analyze the main factors affecting the spectral data in the experiment.After preprocessing the collected original spectral data,taking the stability and spectral strength of the spectral data as indexes,the influences of the composition of steel slag,temperature,trigger delay time,laser frequency,relative measurement distance and internal standard normalization on the spectral data were analyzed by box-graph method.Quantitative analysis of different elements in steel slag.The indexes of evaluating the performance of the model,root mean square error and goodness of fit,are introduced.It is proved that internal standard normalization can improve the stability of steel slag spectral data.The working principle of univariate analysis,partial least square method and neural network is introduced,and different elements in steel slag are quantitatively analyzed.Finally,the algorithm principle of recursive feature elimination is introduced.Through the feature selection of recursive feature elimination,the data dimension reduction and the removal of miscellaneous information are realized,and the partial least square method is used for quantitative analysis.