Development and industrial application of LIBS-XRF coal quality analyzer by combining PCA and PLS regression methods

Abstract

Rapid and stable analysis of coal quality for fine management of coal is essential for the clean and efficient utilization of coal in thermal power plants. In this work, a software-controlled coal analyzer with laser-induced breakdown spectroscopy (LIBS) coupled with X-ray fluorescence spectroscopy (XRF) was developed, which includes an LIBS analysis module, XRF analysis module, sample feeding module, control module and operating software. The instrument not only plays to the strengths of LIBS in multi-element analysis, but also inherits the advantages of XRF in high stability analysis, so that it can be used in power plants for rapid and continuous quality analysis of coal tablets. Based on chemometric regression methods using principal component analysis (PCA) and partial least squares regression (PLS), as well as the spectra of hundreds of coal samples, quantitative prediction models were established, and the industrial test and performance evaluation of the instrument were completed in the Shanxi Yangguang Power Plant in China. The experimental results showed that the R² of the prediction models for calorific value, ash, volatiles and sulfur were 0.973, 0.986, 0.977 and 0.979, respectively, the RMSEs were 0.26 MJ kg⁻¹, 0.68%, 0.33% and 0.13%, respectively, the RMSEPs were 0.62 MJ kg⁻¹, 1.46%, 0.23% and 0.19%, respectively, and the average SDs were 0.11 MJ kg⁻¹, 0.49%, 0.15% and 0.09%, respectively. The models showed good accuracy and stability, and the repeatability of the measurements of coal quality all met the requirements of national standards, and thus, could meet the needs of power plants. This work provides a new idea for the increasingly mature application of LIBS in coal analysis in various industrial sites.