纸浆卡伯值的在线测量数学模型

Mathematical Model for On-Line Measurement of Pulp Kappa Number

从Kubelka-Munk基本方程和纸浆中残留木素的光谱特性出发推导出纸浆漫反射理论模型,进而根据纸浆残余木素与纸浆卡伯值的本质关系推导出在线实时测量纸浆卡伯值的理论模型.根据纸浆中残留木素的光谱吸收特性,选择F390,F560和F840为自变量,建立了"三波长"纸浆卡伯值光谱测量简化模型.该模型结构简单,物理意义清晰,能够揭示模型自变量与卡伯值之间的线性关系.由于从机理模型变为简化模型经过了多次线性近似和简化,造成了纸浆卡伯值与反射光谱间的非线性信息缺失.为此,结合纸浆卡伯值光谱测量简化模型和自适应神经模糊推理系统(ANFIS),建立了纸浆卡伯值在线光谱测量的融合模型,经过样本训练后,融合模型预测结果稳定,精度高.

Based on the Kubelka-Munk equation and the spectrum features of the residual lignin in pulp, the theoretical model describing the pulp diffuse reflection is established. According to the internal relationship between the residual lignin in pulp and the pulp Kappa number, the model for the on-line measurement of pulp Kappa number is presented. Then, according to the absorbing spectrum of the residual lignin in pulp, F390, F560 and F840 are used as independent variables to build a simplified measuring model named ＂tri-wavelength＂. The simplified model is simple in structure and clear in physical description, and it can reveal the linear relationship between the Kappa number and the independent variables. However, the proposed model omits the nonlinear relationship between the Kappa number and the reflected spectrum due to the simplification and linearization in the transformation from theoretical model to simplified one. In order to solve this problem, the simplified measuring model of Kappa number is combined with the adaptive neural-fuzzy inference system （ANFIS） to build a fusion model for the on-line measurement of Kappa number. After the sample training, the proposed fusion model is of good stability and high precision.