基于LIBS光谱实现燃煤企业碳排放量高精度快速预测

High precision and fast prediction of carbon emissions from coal-fired enterprises based on LIBS spectroscopy

准确预测煤炭的碳排放数据是实现中国“双碳”目标的重要基础。在煤炭入炉前完成碳排放量数据和煤质指标的精确、快速预测,将转变目前“先排放,后检测”的碳排放数据计算方式,对落实“双碳”战略意义重大。以煤炭碳含量和煤质指标的精准预评估为研究目标,通过自行搭建激光诱导击穿光谱(Laser Induced Breakdown Spectroscopy,LIBS)系统和LIBS光谱统计分析,利用主成分分析-偏最小二乘(PCA-PLS)结合小样本算法对训练集进行回归训练,建立了一种煤炭的碳排放预测模型,并通过在新疆A电厂和山东B电厂中的实际应用,证明了LIBS系统在煤质指标分析中的可靠性。结果表明:所提出的预测方法与传统方法相比绝对误差最大为0.0256 t,表现出较好的碳排放预测能力。所提方法可以用于燃煤企业碳排放量的精准预测,为实现煤炭碳排放和煤质指标的预评估提供了一种较为高效的技术解决路径。

The accurate prediction of coal carbon emissions is an important basis for achieving China’s“dual carbon”goal.Completing the accurate and fast prediction of carbon emissions and coal quality indicators before coal entering the furnace will transform the current method of calculating carbon emissions data from“emission first,then detection,”and is essential for implementing the“dual carbon”strategy.Our study aims to accurately evaluate the coal carbon content and coal quality indicators.We build a laser-induced breakdown spectroscopy(LIBS)system and use principal component analysis-partial least squares(PCA-PLS)combined with small sample algorithms to regressively train the training set.A prediction model of coal carbon emission is established to demonstrate the reliability of the LIBS system in the analysis of coal quality indicators through its practical application in the A power plant in Xinjiang and the B power plant in Shandong.The results show that the proposed prediction method has a maximum absolute error of 0.0256 t compared to the traditional method,demonstrating good prediction capabilities for carbon emission.It can be adopted for the accurate prediction of carbon emissions in coalfired enterprises and provides a relatively efficient technical solution for the preassessment of coal carbon emissions and coal quality indicators.