Estimating the grain size of microgranular material using laser-induced breakdown spectroscopy combined with machine learning algorithms

Recent work has validated a new method for estimating the grain size of microgranular materials in the range of tens to hundreds of micrometers using laser-induced breakdown spectroscopy (LIBS). In this situation, a piecewise univariate model must be constructed to estimate grain size due to the complex dependence of the plasma formation environment on grain size. In the present work, we tentatively construct a unified calibration model suitable for LIBS-based estimation of those grain sizes. Specifically, two unified multivariate calibration models are constructed based on back-propagation neural network (BPNN) algorithms using feature selection strategies with and without considering prior information. By detailed analysis of the performances of the two multivariate models, it was found that a unified calibration model can be successfully constructed based on BPNN algorithms for estimating the grain size in the range of tens to hundreds of micrometers. It was also found that the model constructed with a prior-guided feature selection strategy had better prediction performance. This study has practical significance in developing the technology for material analysis using LIBS, especially when the LIBS signal exhibits a complex dependence on the material parameter to be estimated.