Research and Analysis of Coal Volatile Matter and Calorific Value Based on Laser-Induced Breakdown Spectroscopy

Abstract

Laser-induced breakdown spectroscopy (LIBS) technology has demonstrated significant application value in coal industrial process monitoring due to its advantages of rapid, in-situ, and multi-element synchronous detection. This study systematically investigates a rapid detection method for coal volatile matter and calorific value based on LIBS technology, addressing the issues of low efficiency and time-consuming processes associated with traditional coal quality analysis methods. Conventional methods for selecting characteristic spectral lines from coal samples are time-consuming and exhibit poor accuracy. Therefore, this study employs variable importance in projection (VIP) to screen key variables in coal samples. Firstly, a full-spectrum analysis model for volatile matter and calorific value was established based on partial least squares regression (PLSR). The coefficients of determination (R²) for the training and test sets of volatile matter and calorific value were 0.9610 and 0.9485, and 0.8972 and 0.9312, respectively. The root-mean-square error of cross-validation (RMSECV) and the root mean square error of prediction (RMSEP) were 0.3678% and 0.5204%, and 0.2999 MJ/kg and 0.2515 MJ/kg, respectively. Subsequently, VIP variable screening was applied for PLSR modeling, resulting in improved R² values of 0.9773 and 0.9569, and 0.9263 and 0.9348, respectively. The RMSECV and RMSEP decreased to 0.2804% and 0.4759%, and 0.2538 MJ/kg and 0.2450 MJ/kg, respectively. This research results indicate that the established analytical framework can be extended to other solid fuel quality detection fields, playing a significant role in achieving intelligent monitoring of industrial processes.