Gao Rui, Li Jiaxuan, Wang Shuqing, Zhang Yan, Zhang Lei, Ye Zefu, Zhu Zhujun, Yin Wangbao, Jia Suotang
State Key Laboratory of Quantum Optics and Quantum Optics Devices, Institute of Laser Spectroscopy, Shanxi University, Taiyuan, China.
Collaborative Innovation Center of Extreme Optics, Shanxi University, Taiyuan, China.
Anal Methods. 2023 Mar 30;15(13):1674-1680. doi: 10.1039/d2ay02086f.
Calorific value is an important indicator to evaluate the comprehensive quality of coal, and its real-time and rapid analysis is of great significance for optimizing the coal blending process and improving boiler combustion efficiency. Traditional assays are time-consuming, and prompt gamma neutron activation analysis (PGNAA) and laser-induced breakdown spectroscopy (LIBS) have certain limitations. In this paper, a novel technique for ultra-repeatability measurement of coal calorific value by combining near-infrared spectroscopy (NIRS) and X-ray fluorescence (XRF) is proposed. In this NIRS-XRF technology, the former can stably measure organic components such as C-H and N-H that are positively correlated with the calorific value, while the latter can stably measure inorganic elements such as Na, Al, Si, Ca, Fe, and Mn that are negatively correlated with the calorific value. The combination of the two can greatly improve the measurement repeatability of coal calorific value. In the quantitative analysis algorithm, a holistic-segmented prediction model based on partial least squares (PLS) is proposed, that is, the holistic model is used to roughly predict the calorific value and determine the segment accordingly, and then the corresponding segmented model is used to accurately predict the calorific value. The experimental results show that the root mean square error of prediction (RMSEP), the average relative error (ARE), and the standard deviation (SD) of this method for predicting the calorific value of coal are 0.71 MJ kg, 1.18% and 0.07 MJ kg respectively. The measurement repeatability meets the requirements of the Chinese national standard. This calorific value measurement technology based on NIRS-XRF is safe, fast, and stable, providing a new way to optimize and control the utilization process of coal in coal washing plants, power plants, coking, and other industries.
发热量是评价煤炭综合质量的重要指标，对其进行实时快速分析对于优化配煤过程和提高锅炉燃烧效率具有重要意义。传统检测方法耗时较长，而瞬发伽马中子活化分析（PGNAA）和激光诱导击穿光谱法（LIBS）存在一定局限性。本文提出一种将近红外光谱（NIRS）与X射线荧光（XRF）相结合的煤炭发热量超重复性测量新技术。在这种NIRS-XRF技术中，前者能够稳定测量与发热量呈正相关的C-H和N-H等有机成分，而后者能够稳定测量与发热量呈负相关的Na、Al、Si、Ca、Fe和Mn等无机元素。两者结合可大幅提高煤炭发热量测量的重复性。在定量分析算法方面，提出一种基于偏最小二乘法（PLS）的整体-分段预测模型，即先用整体模型粗略预测发热量并确定分段，再用相应的分段模型精确预测发热量。实验结果表明，该方法预测煤炭发热量的预测均方根误差（RMSEP）、平均相对误差（ARE）和标准差（SD）分别为0.71 MJ/kg、1.18%和0.07 MJ/kg。测量重复性符合中国国家标准要求。这种基于NIRS-XRF的发热量测量技术安全、快速且稳定，为选煤厂、发电厂、焦化等行业优化和控制煤炭利用过程提供了新途径。
