摘要
氢(H)作为煤中的主要非金属元素,与煤的热值息息相关,是评价煤质特性的重要指标。将可以实现煤质特性多指标快速分析的激光诱导击穿光谱(laser-induced breakdown spectroscopy,LIBS)技术应用到煤中H元素的定量分析中。基于H元素在煤中赋存形态及结构特点,提出以氢原子、碳原子、碳分子光谱信息作为主导因素的燃煤H元素多元线性回归(multiple linear regression,MLR)模型。同时考虑LIBS基体效应对测量准确度的影响,进一步采用偏最小二乘回归(partial least squares regression,PLSR)、支持向量机回归(support vector machine regression,SVR)对主导因素建立的MLR模型的预测偏差进行修正。结果表明,利用PLSR、SVR模型对主导因素MLR模型产生的预测偏差进行修正,可以提高煤中H含量预测准确度。其中,基于主导因素MLR结合SVR偏差修正模型有助于提高H元素定量分析的精确度和准确度。
Hydrogen(H) is the main element in coal, which is closely related to the calorific value of coal, and it is an important index for evaluating coal quality. In this work, the laser-induced breakdown spectroscopy(LIBS) technique, which can realize rapid analysis of coal qualities, was applied to the quantitative analysis of H element in coal. Based on the existing form and structural characteristics of H in coal, it was proposed to establish the multiple linear regression(MLR) model by combining the atomic hydrogen, carbon and carbon molecular spectrum as the dominant factors. According to the influence of LIBS matrix effect on the measurement accuracy, it was further proposed to use partial least squares regression(PLSR) and support vector machine regression(SVR) to correct the deviations generated by the MLR model. The results show that the PLSR and SVR models that used to correct the prediction deviation of the dominant factor MLR model, can improve the prediction accuracy of H content in coal. Among them, the model based on the dominant factor MLR combined with the SVR deviation correction can improve the prediction precision and accuracy of H content in coal.
作者
董美蓉
韦丽萍
张勇升
陆继东
DONG Meirong;WEI Liping;ZHANG Yongsheng;LU Jidong(School of Electric Power,South China University of Technology,Guangzhou 510640,Guangdong Province,China)
出处
《中国电机工程学报》
EI
CSCD
北大核心
2020年第20期6617-6624,共8页
Proceedings of the CSEE
基金
国家自然科学基金项目(51976064)。
关键词
激光诱导击穿光谱
煤
氢
主导因素偏差修正
laser-induced breakdown spectroscopy
coal
hydrogen
dominant factors combined with deviation correction
