可见/近红外光谱分析秸秆-煤混燃物的秸秆含量(英文)

Qualitative and quantitative analysis of straw content in straw-coal blends using Vis/NIR spectroscopy

快速检测秸秆-煤混燃物对生物质混燃发电中补贴政策的制定具有重要意义。该研究采用可见/近红外光谱法定性判别秸秆、煤和秸秆-煤混燃物,定量分析秸秆-煤混燃物中秸秆含量。收集并制备秸秆样品 80 个(粒径小于 80 mm)、煤样品 9个(粒径小于 10 mm),制备秸秆质量分数为 70%～99%的秸秆-煤混燃物样品 120 个(混燃物 1)、秸秆分数含量为 1%～30%的秸秆-煤混燃物样品 120 个(混燃物 2)。使用 FOSS NIRS DS 2500 型光谱仪获取样品光谱。分别使用偏最小二乘判别法(PLS-DA)建立定性分析模型,使用改进的偏最小二乘法(MPLS)建立定量分析模型。结果显示,在秸秆和混燃物 1之间进行判别,使用 1100～2 500 nm 谱区,正确判别率为 90.00%;在煤和混燃物 2 之间进行判别,使用 400～2 500 nm 谱区,正确判别率为 71.88%;定量分析混燃物 1 和混燃物 2 中秸秆含量,相对分析误差分别为 2.32(400～2 500 nm 谱区)和1.48(400～1 100 nm 谱区)。研究结果表明,1 100～2 500 nm 谱区较适合秸秆和混燃物 1 之间的判别,该谱区同样适合定量分析混燃物 1 中秸秆含量。400～1 100 nm 谱区较适合煤和混燃物 2 之间的判别,该谱区同样适合定量分析混燃物 2 中秸秆含量。可见/近红外光谱结合化学计量学是快速定性和定量分析大粒度秸秆-煤混燃物的可行方法。

Analysis of biomass-coal blends is crucial for developing appropriate subsidy policies for biomass co-firing power generation. Visible and near infrared reflectance spectroscopy （Vis/NIRS） was used for qualitative analysis of straw, coal and straw-coal blends, and quantitative analysis of straw content of straw-coal blends. A total of 80 straw samples （particle size less than 80 mm）, 9 coal samples （particle size less than 10 mm）, 120 straw-coal blends samples with straw content varying from 70% to 99% （blends1） and 120 straw-coal blends samples with straw content varying from 1% to 30% （blends2） were prepared. Vis/NIRS spectra of samples were recorded using a FOSS NIRS DS 2 500 spectrometer. Partial least squares-discriminant analysis （PLS-DA） was used for qualitative analysis while modified partial least squares （MPLS） was used for quantitative analysis. The correct classification percentages （CCP） of straw versus blends1 and coal versus blends 2 were 90.00% （using region 1 100-2 500 nm） and 71.88% （using region 400-2 500 nm）, respectively. The ratio of standard error of performance to standard deviation （RPD） for quantitative analysis of straw content of blends1 and blends2 were 2.32 （400-2 500 nm） and 1.48 （400-1 100 nm）, respectively. It was concluded that the spectral region from 1100 to 2500 nm is more suitable for classification of straw versus blends1, and the region is also applied in quantitative analysis of straw content of blends1. Likewise, the suitable spectral region for classification of coal versus blends2 is from 400 to 1 100 nm, which is also appropriate for quantitative analysis of straw content of blends2. Vis/NIRS along with chemometrics may be a feasibility way for qualitative and quantitative analysis of straw-coal blends with large particle size.