摘要
为了高效、精准地检测煤炭灰分质量分数,开展基于光谱分析技术的研究,利用微型光纤光谱仪采集煤样光谱数据,结合多元散射校正(MSC)、主成分分析(PCA)和BP神经网络方法,建立了灰分检测模型。结果表明,PCA+BP神经网络模型显著提高了灰分检测精度,验证集均方根误差(RMSEP)降至1.24%,相关系数达0.998。研究解决了光谱噪声校正、复杂煤样适应性和灰分检测模型鲁棒性问题,实现了煤炭灰分的高效、实时检测,为煤质分析提供了技术支持。
In order to efficiently and accurately detect the quality fraction of coal ash content,research based on spectral analysis technology was carried out.A micro fiber optic spectrometer was used to collect coal sample spectral data,and a ash content detection model was established by combining multiple scattering correction(MSC),principal component analysis(PCA),and BP neural network methods.The results showed that the PCA+BP neural network model significantly improved the accuracy of ash detection,with the root mean square error(RMSEP)of the validation set reduced to 1.24%and the correlation coefficient reaching 0.998.The research has solved the problems of spectral noise correction,adaptability to complex coal samples,and robustness of ash content detection models,achieving efficient and real-time detection of coal ash content and providing technical support for coal quality analysis.
作者
张彬彬
Zhang Binbin(New Energy Geological Team of Hebei Coalfield Geological Bureau,Xingtai Hebei 054000,China)
出处
《山西化工》
2025年第9期91-92,109,共3页
Shanxi Chemical Industry
关键词
煤炭灰分检测
光谱分析技术
数据预处理
偏最小二乘法
BP神经网络
coal ash content detection
spectral analysis technology
data preprocessing
partial least squares method
BP neural network
