Feasibility study on quantitative analysis of coal content in co-firing biomass-coal blends by near infrared spectroscopy

Rapid quantitative analysis of coal content in co-firing biomass-coal blends is greatly important for drawing up reasonable precept about biomass co-firing power generation subsidy. The use of near infrared reflectance spectroscopy (NIRS) predicting coal content in straw-coal blends was investigated in this study. A total of 81 straw-coal blends samples with the coal content from 1% to 9% were prepared and separated into a calibration set and a validation set according to 7/9 and 2/9 ratio. Spectra were scanned by a FT-NIR spectrometer. The regression coefficient method (RC) and the genetic algorithm (GA) were used for spectral region selection. Quantitative analysis models for coal content were established by partial least squares (PLS). The results showed that variables for models decreased from 3, 001 to 141, relative coefficient (R), standard error of prediction (SEP) and the ratio of performance to standard deviation (RPD) in validation were 0.95, 0.80% and 3.31 by using RC-PLS method, respectively. When using GA-PLS, the variables for models decreased from 3, 001 to 11, R, SEP and RPD in validation were 0.90, 1.17% and 2.59, respectively. It is concluded that NIRS with RC-PLS or GA-PLS is feasible for fast quantitative analysis of the coal content in co-firing biomass-coal blends.