Statistic model for predicting cluster movement in circulating fluidized bed (CFB) risers

Abstract

Abstract This study aims to develop a statistic model which can predict the cluster movement in a circulating fluidized bed (CFB) riser accurately and fast. The statistic model takes the CFD–DEM results as samples to make the statistical analysis of the particle movement in a two dimensional (2D) CFB riser. The model combines a stochastic cluster developing model (SCDM) with a Markov chain model (MCM) of particles. The Markov process of particles, the image recognition of clusters and the stochastic cluster moving and shaping sub-models are introduced in detail. Four representative cases with different fluidized air velocities are simulated using the CFD–DEM and the SCDM–MCM separately. Results show that the computing speed of the SCDM–MCM is approximately 60 times faster than that of the CFD–DEM, and the SCDM–MCM successfully simulates the development and the noticeable rectangular-shaped structures of clusters, which is consistent with the description of the cluster shaping sub-model. Besides, the SCDM–MCM introduces the instantaneous disturbance of clusters to the movement of particles, which is reflected in the good comparison between the SCDM–MCM and the CFD–DEM simulated particle RTDs and particle mixing curves.