tom: contours of gas volume fraction) (legend not shown due to confidentiality constraints). Reproduced in colour plate section between pages 210 and 211.

the main requirement in the reactor engineering of FCC regenerators are therefore, to limit the extent of after-burning and to ensure adequate regeneration of spent catalyst. Use of computational flow modeling within the framework of a multilayer modeling strategy to address these issues is discussed below.

9.5.1. Modeling Approach

Gas-solid flow and reactions occurring in an industrial FCC regenerator present severe challenges to flow modelers and reactor engineers. Several attempts have been made to model gas-solid flows based on an analogy with the kinetic theory of gases. A detailed review of these attempts is given in Chapter 12. It is sufficient to state here that none of the available models for simulating dense bubbling beds, were able to predict the continuous bubbling with corresponding satisfactory estimations of bubble volume fractions. Ranade (1996) reported that in many cases Eulerian-Eulerian computational models predicted eventual carry-over of all solids from the regenerator

Flue gas circuit

Flue gas circuit

at high superficial gas velocities. Though there has been some progress in simulating bubbling fluidized beds (see, for example, van Wachem et al., 1999), it may be necessary to use a multilayer modeling strategy to simulate industrial bubbling fluidized bed reactors. Ranade (1998) used a three-layer approach, which is shown schematically in Fig. 9.24.

FIGURE 9.24 Multilayer modeling of FCC regenerator.

A regenerator performance model based on a mixing cell framework is used as a first layer. The model is used to gain an understanding of the overall behavior of the FCC regenerator. The sensitivity of the regenerator performance to a variety of operating conditions was studied. The model was used to fit some of the kinetic constants by comparing model predictions with plant data. A second modeling layer comprised a bubble-bubble interaction model to track bubble trajectories. This Lagrangian model was used to understand the coalescence of bubbles and to estimate the bubble size distribution within the dense bed. A third, CFD-based modeling layer was developed to simulate complex macroscopic flow and reactions in a FCC regenerator. The flow of information between these layers is not unidirectional. There has to be significant interaction and exchange of information during the development and application of these three modeling layers to obtain as much information about the FCC regenerator as possible. The application of these three modeling layers to develop a comprehensive understanding of a FCC regenerator is described below.

9.5.2. Regenerator Performance Model

In view of the possibility of limiting mass transfer from the voids phase, a heterogeneous model (comprising two phases: emulsion and voids (or bubbles)) was used to simulate the dense bed. The homogeneous model was used for the dilute bed region. The generalized mixing cell framework used in the model is shown in Fig. 9.25. It can be seen that the framework allows the flexibility of independent selection of appropriate mixing in the dense phase, voids phase and dilute bed. Mass balances were written for bubble (void), emulsion and dilute bed regions of the regenerator (Utikar and Ranade, 1997). Model equations were incorporated into a user-friendly code called MoBB (Model for Bubbling Beds), to simulate regenerator performance.

Preliminary numerical experiments were carried out to select an appropriate value of time step. To illustrate the typical results obtained using MoBB, a sample of results is shown in Fig. 9.26. It was observed that the regenerator attained steady state in approximately 30 min. The coke on regenerated catalyst reduced from 1% initially to about 0.35% at steady state. The emulsion temperature increased from 723 to 932 K. The oxygen mole fraction at the outlet was about 0.003 and that of carbon dioxide was 0.146. In addition to the prediction of these outlet parameters, the simulation model also provided information about the variation of concentrations and temperatures within the regenerator. The influence of air flow rate and coke on spent catalyst (CSC) on oxygen breakthrough from the dense bed is shown in Fig. 9.26. It is interesting to note that exit oxygen concentration (from the dense bed) does not

ND: number of mixing cells in dilute bed

NE: number of mixing cells in emulsion phase

0 0

Post a comment