An image auxiliary method for laser-induced breakdown spectroscopy analysis of coal particle flow

Abstract

The inevitable unstable fluctuations of particle flow affecting the stability and repeatability of laser–material interaction which contribute to invalid spectral data are an important interference factor for unsatisfactory quantitative analysis via LIBS. In this study, a spectral data screening method based on plasma image information is proposed to identify the effective LIBS spectrum from coal particle flow, aiming to improve the accuracy of online analysis of particle flow. The image auxiliary method was established by investigating the correlation between the plasma image, the laser–particle interaction and spectra, and compared with the existing numerical screening methods: the signal-to-noise ratio (SNR) method and standard deviation (SD) method to further verify its effectiveness. The correlation analysis result showed a good linear relationship between the spectral line intensity and plasma image area with a determination coefficient (R²) of 0.91 and the correlation between the image and excitation effect. Based on this discovery, spectral data of particle flow were filtered according to image information and optimized the threshold setting. After screening invalid spectral data under the optimal threshold value, spectral quality is significantly improved that the average RSD of typical spectral line intensity of coal (C I 247.86 nm, H I 656.24 nm, O I 777.34 nm, N I 742.36 nm, Si I 288.15 nm, and Al I 308.20 nm) was significantly decreased. Moreover, the coefficient of determination (R²) of the quantitative models for the volatile content, calorific value and ash content built after the image auxiliary method screening process reached 0.999, 0.981 and 0.982, which exhibited superior analytical performance compared with the SD and SNR methods. The image auxiliary method showed the application feasibility and prospect for improving the quantitative analysis of coal particle flow.