Characterization and Classification of Coals and Rocks Using Terahertz Time-Domain Spectroscopy

Abstract

Being the key unaddressed problem in unmanned mining condition, a new method for the coal-rock interface recognition was proposed in the study. Firstly, terahertz time-domain spectroscopy (THz-TDS) was employed to measure 10 kinds of coals/rocks which were common in China. Secondly, the physical properties of coals/rocks such as absorption coefficient spectra, refractive index, and dielectric properties in THz band were studied. The different responses in THz range caused by diverse components in coals/rocks were discussed, and the dielectric property of coals/rocks in THz band was well fitted by the Lorentz model. Finally, by the means of principal component analysis (PCA), support vector machine (SVM), and THz spectral data, the recognition rate of coals/rocks reaches to 100 % and the recognition rate of different bituminous coals reaches to 97.5 %. The experimental results show that the proposed method is fast, stable, and accurate for the detection of coal-rock interface and could be a promising tool for the classification of different bituminous coals.