Numerical Simulation And Nonlinear Dynamic Analysis Of Gas-Solid Two-Phase Flow

Gas-solid fluidization technique finds wide applications in chemical, petrochemical and energy-related processes. In particular, more and more attention is focused on the high-velocity fluidized beds, e.g., risers and downers, due to their significant applications in industry. The representative processes include, for example, fluid catalytic cracking (FCC), coal combustion, pyrolysis of biomass, etc. However, the design and scale-up of such kind of reactors are still dependent on the empirical or semi-empirical knowledge because of the extreme complexity of gas-solid two-phase flows, e.g., non-uniform distribution of time-averaged flow parameters and complicated time-dependent behavior of system dynamics.The advanced computational fluid dynamic (CFD) simulation and nonlinear dynamic analysis are applied to study both the time-averaged and time-dependent behaviors of gas-solid two-phase flows in this work.The first part of this thesis is concerned with the experimental study and CFD simulation of downer reactors. Taken into account of the fast and dense gas-solid flow in downers, a turbulent gas- turbulent solid model was therefore proposed and deduced, i.e., a k-s turbulence model for gas phase, a kp turbulence model and a kinetic theory description of solid stresses characterized by granular temperature (0) for solid phase. A new concept to scale-up gas-solid flow in a duct was proposed, i.e., the assumption of the wall effect. It is assumed that the energy of gas-solid flow is dissipated at the wall due to the inelastic collisions between the particles and the wall; the wall only has a large influence on the hydrodynamics near it; and this effect weakens towards the center of the duct step by step. Due to the comprehensive consideration of gas-solid turbulent behavior and the collisions between particles, the k-s-G-kp model is suitable for describing the gas-solid pipe flow in wide operating ranges.A code was developed to solve the steady 3D flow fields of two-phase flowsnumerically. To validate the reliability of the model, the gas-solid flows in the developing and fully developed regions of a downer were simulated. The predicted results had satisfactory agreement with a large quantity of experimental data about axial and radial distributions of local solids fraction, local particle velocity and pressure in the literature. Furthermore, the scale-up effect in the fully developed region of a downer was predicted. Reasonable results were obtained. This provides a solid background in theory for the scale-up and industrialization of downers.Since a downer is specially applied in ultra-short contact processes, the inlet design of a downer is of the most significance. The hydrodynamics in the entrance of a downer with a solids jet process were studied using an optical fiber density probe and a Laser Doppler Velocimetry (LDV). The model was then used to simulate the gas-solid flow under this 2D flow field. The predicted results obtained quantitative agreement with the experimental data. This would be very helpful for the further design of complex inlet structures of downer reactors. Meanwhile, both the CFD simulation and the experimental study showed that the diffusion coefficient of particles is strongly correlated with the local solids fraction. The lower solids fraction corresponds to the larger radial diffusion coefficient. With the increase of solids fraction, the diffusion coefficient decreases and then levels off. This results in a new way to understand the solids mixing behavior in reactors.In the second part of this thesis, the local density fluctuations were investigated in the fluidized beds, involving the downer inlet region, the fully developed region in a riser and a downer, a turbulent bed, a fast bed and a pneumatic transport bed. The measured time series were analyzed using nonlinear dynamic methods. The results showed that the statistical and chaotic analyses of measured system dynamics reveal the characteristics of gas-solid flow in a different way from the time-averaged method.