针对天然纤维素清洁制浆新型连续生产工艺, 提出了采用近红外漫反射光谱测定天然纤维素（棉、 木浆粕）清洁浆料中α-纤维素含量。 收集了142个天然纤维素清洁浆料样品, 采用GB/T 9107—1999方法（化学分析方法）测定其α-纤维素含量。 通过粉碎预处理提高样品的均匀性, 继而压入旋转杯采集光谱。 采用簇类独立软模式（SIMCA）方法建立了有效的棉浆粕和木浆粕的分类模型, 模型识别率达到100%。 基于偏最小二乘（PLS）法分别建立的全部样品以及分类棉、 木浆粕的α-纤维素含量定量校正模型相关系数分别为0.954, 0.911和0.839, SEP分别为2.4%, 1.2%和1.6%, 模型预测精密度与GB方法的允差接近, 表明该方法是可行的, 且操作简单, 分析速度快, 对提高天然纤维素清洁浆料α-纤维素含量分析效率和指导其连续生产具有积极意义。

A new near infrared diffuse reflectance spectroscopy method is proposed to rapidly detect α-cellulose content of natural cellulose (plant fiber: cotton, wood) pulp in a new clean pulping process. One hundred forty two samples were collected and their α-cellulose content data were determined by standard method GB/T 9107-1999. The samples were homogenized by grinding pretreatment to improve spectroscopy measurement accuracy. Effective classification models were built by SIMCA, with the total correct identification. Using partial least squares (PLS) quantitative calibration, α-cellulose of the whole and separate cotton and wood pulp was established, with the correlation coefficients of 0.954, 0.911, 0.839, SEP, 0.024, 0.012 and 0.016, respectively .The repeatability results obtained by the new method are in agreement with the results from GB/T 9107-1999. The new method is feasible for determining α-cellulose content of natural cellulose (plant fiber: cotton, wood) in clean pulping process.