Study On Algorithms For Quantitative Analysis Of Laser Induced Breakdown Spectrum

Laser-Induced Breakdown Spectroscopy(LIBS),a typical emission spectrum tech-nology,has attracted the attention of the industry because its advantages of no sample pretreated,remote measurement,rapid detection,multi element simultaneous detection and so on.Specifically,LIBS can obtain the qualitative or quantitative information of the sample by analyzing the spectral lines of the plasma spectrum.LIBS technol-ogy can be divided into calibration LIBS and calibration-free LIBS(Calibration-Free Laser-Induced Breakdown Spectroscopy,CF-LIBS)based on its different implemen-tations.At present,the calibration LIBS has been applied to many fields,but it has matrix effect and sometimes the calibration sample is difficult to prepare.Thus,in recent years,CF-LIBS has attracted more and more attention from researchers.CF-LIBS formulates a mathematical model to represent the physical properties of the laser plasma,which takes matrix effect as a physical property into account.Therefore,there is no matrix effect,which can be more widely applied than the calibration LIBS.But at the moment,the measurement accuracy of CF-LIBS has not reached a satisfactory degree.The researchers have been optimizing CF-LIBS by stucdying the physical prop-erties of the plasma,upgrading the experimental equipment,changing the experimental environment,and improving the numerical quantitative analysis,etc.This dissertation mainly focuses on improving the quantitative analysis algorithm.Calibration LIBS and CF-LIBS are both considered.Starting from the calibration LIBS,this dissertation briefly introduces the exper-imental system of LIBS technology,the acquisition of experimental spectrogram and the commonly used spectral preprocessing algorithms.In order to accelerate the exper-imental analysis and the selection of the quantitative analysis algorithms for different samples,this dissertation implements the common calibration LIBS algorithm library,and combines with the existing spectral preprocessing software,which lays the founda-tion for the follow-up experiment.After completing the algorithm library and the spectrum preprocessing script,the test results of the current calibration LIBS quantitative analysis algorithm are compared on the BS carbon steel samples,and some experimental conclusions for this series of samples are obtained,and some suggestions for the selection of the calibration LIBS quantitative analysis are given.Improving the CF-LIBS algorithm is imminent.Based on the traditional CF-LIBS algorithm,this dissertation proposes a PSO-based CF-LIBS algorithm.The above prob-lems are modeled from the optimization point of view,and the reasons for the selection of intelligent optimization algorithms are explained,and a simplified CF-LIBS algo-rithm is proposed after the constraint conditions are relaxed.