Discrete element method as an approach to model the wheat milling process

Authors: PATWA, ABHAY - Mennel Milling Company; AMBROSE, R.P. KINGSLEY - Purdue University; Casada, Mark

Submitted to: Powder Technology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/20/2016
Publication Date: 9/1/2016
Publication URL: http://handle.nal.usda.gov/10113/5565954
Citation: Patwa, A., Ambrose, R., Casada, M.E. 2016. Discrete element method as an approach to model the wheat milling process. Powder Technology. 302:350-356. doi:10.1016/j.powtec.2016.08.052.

Interpretive Summary: Models of the wheat milling process save time and reduce the effort required to control each variable independently; however, previous statistical models were limited and an accurate model of individual particle behavior during the milling process is needed. We developed preliminary discrete element method (DEM) models for the wheat roller milling first break stream, conducted lab-scale validation milling trials, and studied the effect of moisture content (from 12% to 16% wet basis) on the process. The first model, with a spherical kernel and uniform bond strength between all particles in the kernel, deviated greatly from the experimental results because all bonds broke when the force on the kernel from the mill rolls reached the threshold to break one bond. The second model, with a true kernel shape and variable bond strength between particles, showed only a small improvement compared to the first model because the bonds still tended to all break near one threshold force instead of yielding the variable-sized clusters seen in the laboratory milling results. The model was able to correctly predict the change observed in first break stream particle size with moisture content. Future research could improve the model predictions by accounting for a distributed bond strength that better represents the wheat kernel in terms of its structural components: endosperm, bran layers, and germ.

Technical Abstract: It is a well-known phenomenon that break-release, particle size, and size distribution of wheat milling are functions of machine operational parameters and grain properties. Due to the non-uniformity of characteristics and properties of wheat kernels, the kernel physical and mechanical properties affect the size reduction process. This research tested the functionality of the discrete element method (DEM) to simulate the 1st break wheat milling process. DEM models of 1st break wheat milling were developed using both spherical-shaped and kernel-shaped particle models. The models simulated hard red winter (HRW) wheat milling at 16% moisture content and were validated using lab scale milling trials. The first approach simulated the size reduction of a spherical cluster of mono-sized particles with uniform bond strength throughout the kernel. This spherical-shaped kernel model resulted in an average particle size of 438 µm with a deviation of prediction of –177%. The prediction error was reduced to 144% with a mean PSD of 372 µm by modifying the shear modulus and coefficient of restitution values. With the kernel-shaped model, a bonded cluster resembling a wheat kernel in shape and size was used with a random distribution of particle bond strengths in the kernel. This model predicted a 1st break particle size of 413 µm, which had a deviation of 139% from the lab scale milling results. However, the model satisfactorily predicted the variation in particle size distribution from 1st break milling.