Improved measurement in quantitative analysis of coal properties using laser induced breakdown spectroscopy

Abstract

By selecting the appropriate spectral pre-processing combined with regression algorithms, the quantitative analysis schemes for each indicator were determined to improve the measurement of coal properties using LIBS.

A hybrid method combining discharge-assisted laser induced breakdown spectroscopy with wavelet transform for trace elemental analysis in liquid targets

“…Multivariate regression algorithms, such as partial least squares regression (PLSR), support vector regression (SVR), and random forest regression (RFR), have been introduced to LIBS and chemical analysis. [52][53][54][55][56] Compared to conventional calibration methods, the multivariate regression algorithm has a good tolerance to outliers and noise and can overcome interference factors such as the matrix effect and overlapping spectral peak effect. 53 Here, PLSR and SVR multivariate regression algorithms are applied to the quantitative analysis of Ba elements in the spiked lubricating oils.…”

“…[52][53][54][55][56] Compared to conventional calibration methods, the multivariate regression algorithm has a good tolerance to outliers and noise and can overcome interference factors such as the matrix effect and overlapping spectral peak effect. 53 Here, PLSR and SVR multivariate regression algorithms are applied to the quantitative analysis of Ba elements in the spiked lubricating oils. The establishments of the algorithm models are as following: (1) Obtain the raw spectral data of Ba elements in the range of 441-467 nm, and the total number of 2048 wavelengths are considered as the variables.…”

Atomic spectrometry update – a review of advances in environmental analysis

“…Analysis by LIBS has great potential for the determination of coal properties such as calorific value, ash, volatile content and C and H contents. Zhang et al 263 developed a set of calibration schemes with the aim of improving the figures of merit of such measurements to meet industrial needs. The selection of an appropriate spectral pre-processing method combined with multivariate calibration models improved the accuracy and precision of each index of coal properties.…”

Quantitative Analysis of Coal Quality by a Portable Laser Induced Breakdown Spectroscopy and Three Chemometrics Methods

“…In this study, 59 commonly used coal samples in China power plants were tested by a lab-made field-portable LIBS instrument and the quantitative analysis of ash content, volatile matter and calorific value was carried out by three chemometrics models [20,21]: partial least square regression (PLSR) [22,23], supported vector regression (SVR) [24,25], and random forest (RF) [14]. Necessary data pretreatment and data partitioning methods were also established.…”