摘要
本研究旨在采用近红外光谱法快速预测新闻纸厂脱墨浆中胶黏物含量。从新闻纸厂脱墨浆生产线中采集到103个脱墨浆样品,以筛选法测得脱墨浆中胶黏物含量,采用近红外光谱仪漫反射方式在12500～4000 cm-1波数范围内采集相应样品的光谱,利用化学计量学软件建立样品胶黏物含量和光谱数据之间的相关性模型。结果表明:对原始光谱进行矢量归一法预处理后,选择7502.2～5446.3 cm-1和4601.6～4246.8 cm-1波数区间,用偏最小二乘法(PLS)和完全交互验证方式建立的校正模型和外部验证预测模型的相关系数R2分别为0.892和0.914,校正标准差SEC为0.027,预测标准差SEP为0.027,这表明脱墨浆中胶黏物含量和其近红外光谱之间存在较好的相关性,该基于近红外光谱的校正模型可用于快速预测新闻纸厂脱墨浆中的胶黏物含量。
A rapid measurement method to predict the stickies content of deinked pulp was developed with near-infrared spectroscopy in this study. One hundred and three deinked pulp samples with different stickies content were collected from a newsprint mill. We measured their stickies content by screening method. The spectra of these pulp samples were collected by NIR spectroscopy in range of 12500 -4000cm-1, and the relationship between the stickies content and the spectra of pulp sample was established by chemometrics software. After Vector normalization pretreatment of the original spectra, the stickies content of the deinked pulp samples was quantified using partial least squares （ PLS1 ） statistical analysis and full cross validation in the range of 7502.2 - 5446.3cm-1. and 4601.6 -4246.8cm-1. The correlation coefficient R2 of calibration model and prediction model were 0. 892 and 0. 914, respectively. The standard error of calibration （SEC） and standard error of prediction （SEP） were 0. 027 and 0. 027, respectively. It was found that there was a good correlation between the stickies content and the NIR spectra. Thus, it is feasible that the stickies content of deinked pulp could be predicted rapidly by NIR spectroscopy-based model.
出处
《造纸科学与技术》
北大核心
2010年第4期50-53,64,共5页
Paper Science & Technology
基金
国家自然科学基金(30771689)资助项目
关键词
胶黏物
脱墨浆
近红外光谱法
新闻纸
stickies
deinked pulp
Near Infrared Spectroscopy
newsprint
