A Non-linearized PLS Model Based on Multivariate Dominant Factor for Laser-induced Breakdown Spectroscopy Measurements

摘要

A multivariate dominant factor based non-linearized PLS model is proposed.The intensities of different lines were taken to construct a multivariatedominant factor model, which describes the dominant concentration informationof the measured species. In constructing such a multivariate model, non-lineartransformation of multi characteristic line intensities according to thephysical mechanisms of lased induced plasma spectrum were made, combined withlinear-correlation-based PLS method, to model the nonlinear self-absorption andinter-element interference effects. This enables the linear PLS method todescribe non-linear relationship more accurately and provides thestatistics-based PLS method with physical backgrounds. Moreover, a secondaryPLS is applied utilizing the whole spectra information to further correct themodel results. Experiments were conducted using standard brass samples. Taylorexpansion was applied to make the nonlinear transformation to describe theself-absorption effect of Cu. Then, line intensities of another two elements,Pb and Zn, were taken into account for inter-element interference. The proposedmethod shows a significant improvement when compared with conventional PLSmodel. Results also show that, even compared with the already-improved baselinedominant-factor-based PLS model, the present PLS model based on themultivariate dominant factor yields the same calibration quality (R2=0.999)while decreasing the RMSEP from 2.33% to 1.97%. The overall RMSE was alsoimproved to 1.05% from 1.27%.