Experimental Validation of Extended Stratification Model Part B: Coal-Ash Segregation Studies in a Batch Jig Operation

Abstract

King’s equilibrium stratification model has been extended to incorporate particle size effects along with particle density effects, for bed stratification of particles, in unit operations like jigs, sluices and Richert cone separators. The focus here is to validate this model with five sets of coal-ash separation data. The data is generated using a laboratory Harz jig. The feed composition of the coal is expressed in terms of particle size and density, as a bivariate distribution, which becomes an input to the model. The feed particles segregate during jig operation in accordance with the fluidization pulse characteristics. The model has two basic parameters to describe segregation of the entire bed, namely, γ and δ. These parameters can be optimized for each test by solving a set of differential equations equal to the number of size-density species present in the feed and subjected to the constraints of the model. The partition coefficients arising from the model are matched to the experimentally measured values at various cut heights of the bed. The ability of the two parameters γ and δ to define the measured partition surfaces across the bed height is reasonably well, which indicate the efficacy of the extended stratification model. The model defines the segregation pattern in terms of both particle size and density across the bed height. The optimized parameters γ and δ for the five tests show that specific stratification coefficient, \(\beta_{i}\), can either increase, decrease or remain constant with increasing particle size which depends on the pulse water flow rate. Within the limited test data collected, the parameters γ and δ show an increasing trend with increasing batch jig pulse water flow rate. The average bed density profile across the bed height either gradually decreases or shows a stepwise decrease, which is a common feature of a stratified jig bed. Partition surface fits at various cut heights show that the parameters \(c\) and \(\rho_{p}\) are almost constant in each set of experiment.