Selection of Spectral Variables in Multivariate Calibration of C, Mn, Si, Cr, Ni, and Cu Concentrations in Low-Alloy Steels by Laser-Induced Breakdown Spectroscopy

Multivariate models for calibration of C, Mn, Si, Cr, Ni, and Cu concentrations from low-resolution emission spectra (190–440 nm, resolution 0.4 nm, spectral step 0.1 nm) by the least-squares method were developed in sets containing from 31 to 39 reference samples of low-alloy steels. Three methods of selection of spectral variables are considered, namely, method of ranking spectral variables by their correlation coefficient with the sought parameter, a successive projection algorithm, and an original modification of the method of searching combination moving window. The partial least-squares model with selection of spectral variables by searching combination moving window for C is quantitative (root-mean-square deviation 0.004%, residual deviation in the test sample 23.4 in the concentration range from 0.13 to 0.43%). The concentration calibrations are also quantitative for Mn (0.04% and 5.2 in the range of 0.47–1.15%), Si (0.003% and 20.7 in the range of 0.15–0.33%), Cr (0.04% and 3.1 in the range of 0.09–0.43%), and Ni (0.01% and 4.8 in the range of 0.05–0.25%). The calibration for Cu in the concentration range of 0.06–0.26% is qualitative (0.04% and 1.4).