Study on LIBS Standard Method via Key Parameter Monitoring and Backpropagation Neural Network

This paper proposes a method based on key parameter monitoring and backpropagation neural network to standardize LIBS spectra, named KPBP. By the laser output energy plasma flame morphology using algorithm fit spectral intensity, KPBP standardizes segments containing characteristic lines. study first conducted experiments spectra of pure aluminium, monocrystalline silicon, zinc optimize model then performed standardization lines GSS-8 standard soil sample. The intensity relative deviations (RSDs) Al 257.51 nm, Si 298.76 Fe 406.33 nm dropped from 12.57%, 16.60%, 14.10% 3.40%, 3.20%, 4.07%, respectively. Compared with internal normal variate method, obtained smallest RSD. also used GSS-23 sample Beijing farmland conduct optimization experiments. RSD was still significantly reduced, proving that has stable effects wide applicability improve repeatability analysis.