近红外光谱结合区间偏最小二乘法快速分析污泥中蛋白质含量  

Rapid analysis of protein content in sludge using NIR spectroscopy combined with iPLS method

作　　者：李卫华[1] 田士玉 唐玉朝[1] 黄显怀[1] 刘绍根[1] 杨引浩 LI Weihua TIAN Shiyu TANG Yuchao HUANG Xianhuai LIU Shaogen YANG Yinhao(Key Laboratory of Anhui Province of Water Pollution Control and Wastewater Reuse, School of Environment and Energy Engineering, Anhui Jianzhu U- niversity, Hefei 230601, China)

机构地区：[1]安徽建筑大学环境与能源工程学院,水污染控制与废水资源化安徽省重点实验室,合肥230601

出　　处：《环境工程学报》2017年第3期1819-1824,共6页Chinese Journal of Environmental Engineering

基　　金：国家自然科学基金资助项目(51378017,51278002);国家水体污染控制与治理科技重大专项(2012ZX07103-001)

摘　　要：采用近红外光谱表征了不同来源污泥样品,并应用主成分投影实现了污泥样品的聚类分析。在此基础上,结合凯氏定氮法的蛋白质含量分析结果,采用区间偏最小二乘法(i PLS)选择光谱建模区间为4 382~3 996 cm^(-1),建立了近红外光谱反演污泥中蛋白质含量的数学模型,模型相关系数为0.973,交叉验证均方差为1.321,预测值与真实值的误差范围为^(-1)0.0%~8.52%,说明近红外光谱可以实现污泥样品的快速定性和定量分析。Different sludge samples were characterized using near infrared（NIR） spectroscopy,and cluster analysis was carried out using the principal components projection method based on the difference in the NIR spectra.Furthermore,combined with the protein contents in the sludge samples determined by the Kjeldahl method,the modeling spectral region of 4 382 to 3 996 cm^-1 was chosen to establish a correlation model between the NIR spectra and the protein content by the interval partial least squares（iPLS） method.The correlation coefficient was 0.973.The root mean stand error of cross-validation（RMSECV） was 1.321.And the relative errors between the predicted and the real values ranged between- 10.0%and 8.52%,which implied that NIR spectroscopy could be used for rapid identification of the origin of the sludge sample and quantification of the protein content in the sludge samples.

关 键 词：凯氏定氮法 近红外光谱 区间偏最小二乘法 污泥 蛋白质含量 

分 类 号：X705[环境科学与工程—环境工程]