(215g) Scale-up of Corn Stover Comminution Population Balance in Knife Mills

Biomass derived energy has a significant role in solving the energy crisis as the need for renewable and clean energy rises, Biorefinery operations and new process technologies must be optimized and demonstrated to enable this developing industry. Behind drying of moisture-laden biomass resources, comminution is often the highest energy consuming unit operation as well defining the format of the feedstock. This, in turn, is largely responsible for the feeding and handling performance as well as convertibility in terms of biological access or thermal transport, etc. In this work we develop a predictive population balance model for knife milling of corn stover stalks, based on a breakage probability and breakage function to predict the milled particle size distribution as a function of variation in mill tip speed, moisture content, initial particle size, and retention screen size. This PBM was first developed and calibrated using bench-scale experimental data and then scaled up to characterize breakage behavior in a larger pilot-scale mill that represents realistic comminution processes prevalent in industry. Model fit parameters were then obtained through minimization of squared error for each experiment. Calibration experiments were performed with corn stover stalks (15-30 mm diameter) in a bench-scale knife mill. The bench-scale PBM was used to predict product particle size distributions at each condition experimentally tested at the pilot scale. Using the model parameter fitting results, statistical regression was performed to evaluate dependance of material breakage parameters (material resistance to fracture, minimum impact energy for fracture) on initial moisture content and rotor speed. These developed prediction expressions were then used to obtain global fit parameters for all the different processing conditions, and to get a predictive basis for operation of the pilot-scale knife mill. Validation experiments were then performed with the pilot-scale mill and a good predictive match was obtained using the developed PBM. Particle size and distribution shape metrics matched within 80% accuracy between experimental measurements and model predictions. Particle size metrics that were studied included 10, 50 and 90% passing diameters and distribution shape metrics studied were relative span, kurtosis and skewness.