Accurate elemental analysis of alloy samples with high repetition rate laser-ablation spark-induced breakdown spectroscopy coupled with particle swarm optimization-extreme learning machine

Abstract

High repetition rate laser-ablation spark-induced breakdown spectroscopy (HRR LA-SIBS) coupled with particle swarm optimization-extreme learning machine (PSO-ELM) was developed to realize quantitative elemental analysis of aluminum alloy accurately. A compact Nd: YAG laser operated at 1 kHz was used as the ablation laser and a spark discharge was utilized to enhance the plasma emission. The characteristic lines of the elements were selected as input variables and the parameters of ELM were optimized by the PSO algorithm. The PSO-ELM calibration models were established. The correlation coefficient of prediction set reached to 0.9996, 0.9972 and 0.9999, the root mean square error of prediction set decreased to 0.0138%, 0.0045% and 0.0095% for Mg, Cr, and Cu analysis, respectively. Better predictive performance has been demonstrated in comparison with the univariate and SVM methods. HRR LA-SIBS coupled with PSO-ELM can give convenient, rapid and accurate elemental analysis of alloy samples. Moreover, it will extend the potential applications of HRR LA-SIBS in intelligent manufacturing and on-line monitoring in the future.