(53c) Simulation of the Semi-Dry Flue Gas Desulfurization in a CFB Using the Two-Fluid Model With a Heterogeneous Drag Model | AIChE

(53c) Simulation of the Semi-Dry Flue Gas Desulfurization in a CFB Using the Two-Fluid Model With a Heterogeneous Drag Model


Wang, X. - Presenter, Institute of Process Engineering, Chinese Academy of Sciences,China
Zhu, T., Institute of Processing Engineering,Chinese Academy of Sciences
Jing, P. F., Institute of Process Engineering, CAS

Semi-dry flue gas desulfurization (FGD) has been developed for low temperature in a circulating fluidized bed (CFB) reactor. The particle volume in the semi-dry FGD reactors (<5kg/m2•s) is much more dilute than the particle volume in the classical CFB reactors, such as the fluid catalytic cracking (FCC) reactors and the CFB combustion reactors. The sorbent for the semi-dry FGD is lime (CaO) which has unique fluidizing characteristics. However, previous studies of the numerical simulation of the CFB reactor mainly focused on the process of high particle volume, the reports on CFD modeling of the semi-dry FGD process are much rare and the controlling step is controversial currently.

The semi-dry FGD was carried out in a pilot scale CFB reactor with a venturi distributor for the gas acceleration. Experimental measurements of both fluidizing and FGD had been compared with the simulation results to verify the model accuracy. A comprehensive 3D model was developed based on the two-fluid model and involved reactants mass transfer, water evaporation and desulfurization.

Since the flow pattern in the CFB reactor is heterogeneous and has significant influence on the FGD process, the flow modeling of the cold state in CFB is necessary to get clearly comprehension. Compared to the homogeneous drag models, heterogeneous drag models could better describe the particle aggregation. In all heterogeneous drag models, the O-S model has an empirical factor representing the particle clustering effect with short computation time. Therefore, the O-S model was selected in this work. The O-S model was analyzed in terms of the pressure drop, the distribution of particle volume and the distribution of the velocity in the CFB by comparing with the experimental data and the computing results of the classical Gidaspow drag model. Results show that the O-S model is more accurate than the Gidaspow drag model, but there are still some differences from experimental data due to an inaccurate description of the particle volume in clusters.

The SO2 absorption process includes humidifying, evaporation, production of Ca(OH)2and neutralization reactions. The FGD simulation is based on the O-S drag model verified above. However, there are a lot of difficulties in the simulation process. First of all, the process of heat transfer relates to the humidifying and evaporation, which involves mass transfer and phase change of water, as well as the convective heat transfer near the wall. Previous studies considering heat transfer as an adiabatic process cannot simulate accurate results according to the different conditions in the real application. Besides, it is difficult and time-consuming to simulate such complicated process involving evaporation, ionization, dissolution, neutralization and deposition in the turbulent gas-solid flow. This method of microcosmic modeling is overloaded with details and might be inefficient for engineering prediction. Moreover, doubts on the controlling step still exist between the physical process such as mass and heat transfer and the chemical process such as dissolution and ionization.

The aim of this work is to develop a simplified model that can simulate the flow behavior coupling with the heat transfer and desulfurization reaction in the CFB reactor more efficiently. Simulation results of the desulfurization efficiency and reaction rate indicate that higher particle volume and more uniform particle distribution lead to higher desulfurization efficiencies. The controlling step is predicted to be either the dissolution of CaO or the mass transfer of SO2, corresponding to the different water spraying volume.

These considerations greatly improve the consistence of theoretical predictions with experiments and better explain the phenomena in the semi-dry FGD in the CFB. The model can also be used for design and optimization of the semi-dry FGD processes and guide industrial scale-up.