Discrete element method model of the first break wheat milling process

Abstract

It is a well-known phenomenon that the break-release, particle size and size distribution of wheat milling are functions of machine operational parameters and grain properties. Due to the non-uniformity in characteristics and properties of wheat kernel, the kernel physical and mechanical properties may affect the size reduction process. The discrete element method (DEM) is a numerical modeling technique that can be used to study and understand the effect of physical and mechanical properties of a material on processing. The overall objective of this study is to develop a DEM model of the 1st break wheat milling process. In this study, different physical and mechanical properties of wheat mill streams were determined for using as the input parameters in DEM model development. The particle size and size distribution (PSD), true, bulk and tapped density, young’s modulus, coefficient of static and rolling friction, and coefficient of restitution were measured for wheat kernel, 1st break and flour from hard red winter (HRW), hard red spring (HRS), and soft red winter (SRW) wheat. Overall moisture content was found to have a greater significant effect on the physical properties i.e. density and PSD of the mill streams than material properties i.e. Young’s modulus, coefficients of friction and coefficient of restitution. The DEM model of 1st break wheat milling was developed using both single and multi-sphere approaches. The single sphere approach simulated the size reduction of a spherical cluster of bonded particles with mono-sized particles. The model was simulated for hard red winter (HRW) wheat milling at 16% moisture levels and validated using lab scale milling trials giving a PSD of 437.73 m with a percent deviation of prediction of 235.37. The deviation of prediction was reduced to 192.29 with a mean PSD of 371.52 m by conducting sensitivity analysis by modifying the shear modulus and coefficient of restitution values. In the multi-sphere approach, a bonded cluster resembling a wheat kernel in shape and size was used to simulate the milling process. The model predicted a 1st break PSD of 412.65 µm which had a deviation of 185.89 from lab scale and 156.78 from plant scale milling. The model could however satisfactorily predict the variation in PSD from 1st break milling with moisture content with reasonable accuracy. Future capabilities using the model include performing additional sensitivity analysis to understand the effect of other mechanical properties of wheat on the 1st break PSD. It can also be used to improve the 1st break release during wheat milling.