Modeling of maize breakage in hammer mills or different scales through a population balance approach

IVANA M. COTABARREN; ALEJANDRO G. CHIARAVALLE ; JACQUELINA C. LOBOS DE PONGA; JULIANA PIÑA

Londres

Otro; APM Forum 2021; 2021

The grinding of maize grains is an important process in the industry of poultry feed, beingthe most commonly used cereal in diet worldwide. Hammer mills are usually used for thisprocess because of their simplicity, easy operation and low maintenance cost. However,the grinding process in hammer mills itself is complex and not completely understood.Hammer mills are impact-type crushers that comprise a rotating shaft fitted with fixed orpivoted hammers and mounted in a cylindrical chamber. The particles are fed into thechamber by gravity and exit the cylinder when they are small enough to pass through ascreen located at the bottom. Comminution in hammer mills is influenced by designvariables as the rotor shaft configuration (i.e., vertical or horizontal shaft), the hammerdesign and placement (i.e., distance between hammers and screen), the screen openingsize and type (i.e., round, square, mesh type), and by operating variables as material feedrate and rotor speed. The resulting particle size distribution depends not only on materialproperties but also on design and operation variables of the mill.In this work, a simulator for a grinding process is presented. The simulator is based on apopulation balance model (PBM) that allows quantifying changes in particle sizedistributions (PSDs) and the mass geometric mean diameter (Dgw) as a function of milloperation variables (screen opening, mill rotor speed and feed rate).This model was developed completely in the gPROMS® Model Builder, usingexperimental data collected from a pilot-scale mill to fit the parameters of the breakageand classification functions, considering the effects of rotor speed and screen openingsize. The gPROMS® in-built parameter estimation tool was used for parameter fittingand statistical test analysis. The model was validated with both pilot- and industrial-scales steady-state data, which also included variations in feed rates. The developedmodel was used to predict the dynamic behavior of the product mass geometric meandiameter and mill hold-up as a function of disturbances in the feed rate and the rotorspeed. These results are highly valuable since they indicate the feasibility of using thefitted model to predict with confidence new operating points, study industrial millingperformance and optimize hammer mill operation reducing the need to carry outexpensive and time-consuming experimental tests.