Bagasse is a very useful feedstock to produce biofuels and chemicals. In order to importantly enhance its digestibility, pretreatment is required. To succesfully identify ranges of pretreatment conditions that importantly enhance pretreatment selectivity, a trustable kinetic model for this operation is required. In this work, models for lignin and carbohydrates degradation are used to calculate pretreatment conditions that result in desired biomass composition. This work is innovative because it applies an iterative method that uses a D-optimal experimental design to calculate model parameters for both lignin and carbohydrates degradation. It is known that the least squares estimator for parameters minimizes the variance of mean-ubiased estimators. In the estimation theory for multiparameter models, minimizing the variance of the information matrix (i.e., the inverse matrix of the variance-matrix) corresponds to maximizing the information available from the experiments. D-optimal design achieves this by maximizing the determinant of the information matrix. This criterion results in unprecedent reduction in the volume of the confidence ellipsoid for the parameters. Furthermore, this experimental design proved to be a powerful tool to substantially reduce the number of experimental runs.
Eventhough the method can be extrapolated to any type of pretreatment, this work covers oxidative lime pretreatment using hydrogen peroxyde as oxydative agent with temperature raging from 40°C to 80°C, peroxide concentration varying from 1% v/v to 4%v/v, an excess of lime loading of 0.5 g (calcium hydroxide)/g (dry biomass) and a reaction time from 30 to 120 minutes.
Keywords: d-optimal, experimental design, kinetic model, optimal conditions, delignification.
See more of this Group/Topical: Sustainable Engineering Forum