As a critical parameter in electromagnetic detection for coal mine safety, the dielectric permittivity of coal is influenced by its structural characteristics. In addition, electromagnetic waves with terahertz(THz) frequencies have extensive application potential in coal mining. Therefore, in this study, the terahertz dielectric permittivity of coals with varying metamorphism degrees was measured using the terahertz time-domain spectroscopy, and its evolution mechanism was investigated through grey relational analysis based on proximate components. Experimental results indicate that the Landau-Lifshitz-Looyenga model provides the most accurate description of the dielectric behavior of coal tablets. Both the real and imaginary parts of the terahertz dielectric permittivity exhibit a decreasing trend with increasing metamorphism degrees. Due to the highest grey correlation, moisture and ash are the main factors affecting the dielectric properties of coal. Specifically, the decrease in moisture content with increasing coal rank primarily accounts for the decline in permittivity. Ash content is the key factor causing the increase in the real part of non-stick coal and imaginary part of long-flame coals due to its high dielectric permittivity and loss. These findings can enrich the theoretical understanding of coal's dielectric behavior and offer a valuable reference for the application of electromagnetic detection technologies in coal mining.

In view of the current situation that there are many but no uniform coal molecular modeling methods, this paper explores the structuve of lignite macromolecules and optimizes them from a microscopic perspective by selecting physical methods that have less impact on molecular structure.A relatively simple coal macromolecule modeling method was put forward.Based on the experimental methods of elemental analysis, ¹³C-NMR and XPS, molecular dynamics simulation software method was used to analyze and study lignite and build molecular structure models.The results show that the aromatic structure of lignite macromolecules is mostly pentaphenyl.The fat carbon structure is mostly methylene and methine, and the alkane chain is mostly cycloalkanes.Oxygen atoms are mostly ether bond oxygen, followed by carboxyl oxygen and carbonyl oxygen.Nitrogen atom exist in the form of Pyridinic nitrogen N-6.After the optimization, the structure of the coal molecular model is more compact and the energy is obviously reduced.As the main component of the non-bonding potential energy and the main factor to maintain the stability of the coal molecular structure, the van der Waals potential energy has the most obvious change.In order to construct coal molecular cell, periodic boundary conditions were added.According to the energy variation, the cell density was 1.2 g/cm³ and the total energy was 1 140.624 kJ/mol, which was basically consistent with the actual situation and verified the effectiveness of the modeling method.This study provides a method for the direct understanding of the macromolecular structure of coal, and is of great significance for the mechanism research and prevention of coal and gas outburst, coal spontaneous combustion and other disasters.