| Kappa number is used to quantify the residual lignin content in a pulp sample. This average value does not quantify the kappa number variability on the single fiber level. Kappa variability, however, has consequences in unit operations downstream from the digester, such as bleaching and papermaking.; The University of Washington pulping model predicts the single fiber kappa distribution of a kraft pulp. In its previous one dimensional form the model correctly predicts the skewed nature of the kappa distribution but misses the broad nature of kappa distribution near the mean. It appears that there is another source of pulp uniformity that occurs on fiber scale. Understanding this underlying heterogeneity is essential to accurately model the kraft pulping process.; The objectives of this research are the following: (1) To investigate sources of pulp non-uniformity on the fiber scale; (2) To revise University of Washington mathematical pulping model to accurately predict single fiber kappa distributions; (3) To improve statistical methods for quantification of single fiber kappa distributions.; Experimental studies show that fiber scale kappa heterogeneity in softwood kraft pulp comes from the random fiber natural chemistry and structure, and the difference in pulping rate between earlywood and latewood. The fibers' random nature results in a Gaussian single fiber kappa distribution with a COV∼0.16 for single wood species pulp. A Gaussian distribution with a COV of 0.20 can be used to model the single fiber kappa distribution of a softwood pulp resulting from fiber scale heterogeneity, accounting for both the random fiber nature and differences in earlywood and latewood.; University of Washington kraft pulping model, which accounts for the effects of chip and digester scale sources of heterogeneity, was revised to a three dimensional form incorporating chemical diffusion and heat transfer in three dimensions and fiber scale pulp uniformity and into the simulation. This new three dimensional model predicts single fiber kappa distributions significantly better than the previous 1 dimensional model.; Several statistical methods were investigated to better quantify the fiber kappa distributions. COV (Coefficient Of Variation) and COMAD (Coefficient Of Median Absolute Deviation) were found to be good statistics defining the overall pulp uniformity and the peak breadth of the kappa distribution curve. The Quantile-Quantial Plot is an effectively intuitive method to compare two pulps' uniformities. |
