A Neural Fuzzy Model Applied to Hydrogen Peroxide Bleaching of Non-wood Soda Pulps

A neural fuzzy model was used to examine the influence of pulp bleaching variables of empty fruit bunches from oil palm (EFB) and Hesperaloe funifera, such as soda concentration (0.5-3%), hydrogen peroxide concentration (1-10%) and processing time (1-3 h), on Kappa number, brightness and viscosity. The experimental results are reproduced with errors below 10% and 15% for EFB and H. funifera, respectively. Bleaching pulp simulation permits to obtain optimal values of the operating variables, so that the properties of bleached pulps will only slightly differ from their best values, while the lower values of the operating variables will save chemical reagents, energy and plant size. Thus, if applying 0.5% soda and 3% peroxide for 3 h, it is possible to get a pulp with a brightness of 74.9% and a viscosity of 716 mL/g, for EFB pulp, and of 63.3% and 584 mL/g, respectively, for H. funifera pulp.