(737b) Population Balance of Hammer Milled Loblolly Pine in ‘Once-through’ and ‘Fractional Milling’ Configurations

In this work we develop a predictive population balance model for hammer milling of Loblolly pine , based on a probability and breakage function to predict the milled particle size distribution as a function of variation in mill tip speed, feed rate, initial particle size, and mill chamber retention screen size. These probability and breakage parameters are dependent on material properties like the fracture resistance and minimum specific energy. Batch scale experiments are used to determine material and mill parameters within the breakage law. A breakage matrix will be developed using fracture resistance and minimum specific energy, which are found from calibration tests with iterative regression to measured product size distributions as well as inputs for feed size, flow rate, screen classification efficiency, amount of recirculated material and the number of impacts that a particle experiences. High-speed camera imaging and velocimetry will be utilized to be illustrative of the later variable. In the hammer mill experiments, the initial particle size is obtained from analytical splits of the sample along with separations with an orbital screen. Screen opening is altered by interchangeable screen sizes while the mill speed is controlled with a VFD. and the mill feed rate is controlled via a loss-in-mass feeder. These variables are associated with the impact energy, number of impacts and the size of the particles under impact. This work reports on the imparted control over resulting particle size distribution as a function of controllable operational parameters (e.g. tip speed, feed rate, initial material size and retention screen size). In this work, results for both a ‘once-through’ configuration in addition to a recirculating fractional milling technique will be evaluated for degree of control over the comminution process, as well as potential energy savings and optimal configuration/recirculating rate. The model will provide an estimate for daughter particle size distribution which will then be validated by experimental results, and be able to calculate the number of impacts required to achieve desired classification. Determining the lowest number of impacts to achieve a specific particle size distribution will allow optimization on energy costs and milling configuration.