Research On The Motion Characteristics Of Particle Agglomeration In Fast Fluidized Bed Based On Image Method

In order to further understand the formation,development and disintegration mechanism of clusters and the momentum transfer process between gas and solid phases in a fast fluidized bed(FFB),improve the numerical simulation method of gas-solid fluidization,realize accurately simulate and predict the movement of clusters in a fast fluidized bed(FFB),experimental research,theoretical analysis and numerical simulation are combined to study the particle clustering characteristics in a fast fluidized bed in this paper systematically.The dynamic process of clustering of glass beads,river sand and FCC particles in a fast fluidized bed is recorded by the high speed camera.Based on the self-compiled programs,particle clusters were identified,located and tracked automatically,then the structure and motion characteristics,such as the cluster concentration,shape,size,lateral and longitudinal moving were studied.Based on the experimental data and theoretical analysis,the cluster concentration model and cluster size model were established,the drag force model based on the multi-scale minimum energy theory(EMMS)was modified.Combing the modified drag model with MP-PIC algorithm,the three-dimensional mathematical and physical model of the fast fluidized bed was established,the gas-solid flow and particle clustering behavior in the fast fluidized bed was simulated,and the experimental data were compared with the model results for verification.The main conclusions are as follows:Based on the built-up cold fast fluidized bed test device,under a wide operating range(U_g=4.5-7.5 m/s,G_s=20-250 kg/m~2·s),the glass beads,river sand and FCC particles characteristics of clustering in the gas-solid flow field were studied.High-speed camera and image acquisition technology were used to record the particle clustering behavior and dynamic movement process in the riser and realize automatic recognition of clusters based on the k-means++clustering algorithm.Parameters,such as the shape,size,concentration,centroid position and so on,of all clusters were obtained in the observation window at the time of shooting through image processing.The cross-correlation algorithm was used to process the time series images and the movement information such as the trajectory and speed of the cluster material core was obtained.The results show that:The movement trajectory of the clusters in the riser conforms to the"ring-nucleus flow"model.The vertical velocity of the clusters is approximately symmetrically distributed in the horizontal direction with a"higher center and lower side walls".The clusters have lateral displacement in the riser and they are most significant in the transition zone.Moreover,the size of the clusters in the horizontal direction is approximately symmetrically distributed with a"larger side walls and smaller centers"and the shape of those in the center area is closest to a circle.Under the same working conditions,the larger the particle size and sphericity,the easier it is to form large-sized clusters.Compared with type B particles,the clusters formed by type A particles have higher concentration,larger size and more irregular shapes.Under a wide time-average particle concentration(?_s=0.02-0.46),the rationality of the QC-EMMS cluster concentration model was verified through experimental data and the overall non-uniformity index?was determined to be 0.39,thus we established the cluster concentration formula.The two assumptions of the cluster size formula in the bubble model"the volume fraction of the solid phase and the dilute phase are equal to the volume ratio of the cluster to the bubble and the absence of particles in the bubble"were revised and the cluster size formula was established.The established cluster concentration formula and size formula were substituted into the EMMS drag force model to modify it.Combined the drag force model before and after the correction with the MP-PIC algorithm,the three-dimensional fast fluidized bed gas-solid flow mathematical model was established,the fast fluidized bed riser and the whole field were stimulated respectively,and the accuracy of the corrected drag force model was vertified compared with the test data.The results show that:Taking the modified drag force model,the time-average particle concentration and gas-solid two-phase velocity obtained are closer to the experimental data.In addition,it can well predict the mesoscopic gas-solid flow structure such as the concentration,shape and size of clusters.