Ground Biomass Sieve Analysis Simulation by Image Processing and Experimental Verification of Particle Size Distribution

摘要

Sieve analysis is the standard method of determining particle size distribution of particulate material such as ground biomass particles; however to produce a balanced distribution of fractions it involves a lot of guesswork in sieve set selection, inthis paper we propose an alternative image processing method of sieving simulation with experimental verification. An Image J user-coded plugin to measure particles dimensions and analyze their size distribution from flatbed scanner captured images wasdeveloped. Hammer-milled dry pine first thinning, and elephant grass biomass were the test materials. The plugin used pixel-marching method for determining the particle dimensions and AS ABE Standard S319.3 for analyzing particle size distribution. The developed sieving simulation plugin had several features, among which are capability to i) choose the number of sieves, ii) suggest set of sieves for equal distribution of material, Hi) choose standard sieves set from list, iv) analyze size distribution,and v) output results in graphical and tabular formats. Plugin-suggested sieves, based on particle lengths, eliminated the guesswork involved in sieve selection. Observed deviation between the experimental and simulated sieving were attributed to the difference in densities of individual fractions and the "fall-through" effect of longer particles through smaller sized sieve openings. Because of this situation, it is proposed that the image processing sieving simulation plugin producing the "true separation" of particles can totally replace mechanical sieving and analysis, especially when dealing with particles fibrous in nature (UW 1). The plugin application was accurate (96.6%), fast (more than 450 particles/s), cost effective, reduced the drudgeryof sieve analysis, and can be readily applied to other particulate systems..