机构地区: 广东省科技干部学院
出 处: 《中国造纸学报》 2003年第1期110-113,共4页
摘 要: 利用近红外光谱法测量蒸煮过程黑液中木素含量的变化 ,从而在线测定出纸浆卡伯值。通过一元线性回归、多元线性回归和主成分回归分析显示 :3种测量模型均具有较好的学习和预测精度 ,其中主成分回归模型具有较高的预测鲁棒性。 On-line determination of pulp kappa number with near-infrared spectroscopy (NIR) during cooking is introduced. Three sloichiometric methods, i.e. unary liner regression (ULR), multivariable linear regression (MLR) and principal component regression (PCR) are applied for data processing. The results show that the predicting models determined by the three methods all have good learning ability and predicting accuracy of kappa number, and the best robust model is determined by principal component regression.