Study On The Method Of Quantitative Analysis Of Metal Elements In Liquid Steel Based On LIBS

Laser Induced Breakdown Spectroscopy is characterized by plasma spectroscopy technique for quantitative analysis of a classic. The technique has the advantages of sample non-sample preparation,in situ detection, multi-element simultaneous detection, real-time remote non-contact measurement,wide detection range of elements and short detection time. Its prospects in the steel smelting is very broad. At this stage, the research in the fields of iron and steel smelting technology is mainly in the direction of the analysis algorithm. With the results of previous studies and the use of new data processing algorithms, liquid steel metallic elements are analyzed quantitatively.Firstly, the principle of LIBS technology is described. Describing the plasma from generation to the annihilation characteristics of plasma characteristic spectral lines in the whole process and analysising of the factors affecting the plasma spectral line are introduced. A set of experimental platforms for industrial production are set up on this basis. The platform has been put into the steel liquid to avoid the interference of the air environment. The focal length adjustment device is installed on the platform, which improves the effect of laser focus distance on the experimental data.Electromagnetic shielding equipment installed on the platform, reduce the complicated industrial environment on the influence of experimental equipment.Background noise of the experimental data is removed after the characteristic spectrum line selected and the smoothing method is used for filtering. Use wavelet transform method for automatic search and qualitative analyze of measurement elements. By means of partial least squares based on LIBS on the analysis of liquid steel analysis principle and lists the advantages and disadvantages. The analysis method is more advantages than the traditional method by analysis. For the lack of PLS, the method of stepwise regression and weighted average method is improved. Through the analysis of the Mn data, the accuracy of the partial least squares algorithm and the weighted average partial least squares algorithm based on stepwise regression are analyzed. The linear correlation coefficients of the two curves are 0.991,0.946 and 0.994,0.978 and the relative error is 6.231%,7.611% and5.510%,6.512% respectively. The results show that the improved WPLS algorithm is more accurate.