Previously I have compared the model having a true CSTR emulsion phase with the Dasila et al. results using a plug flow model of the dense bed. However, the two models used different kinetics for the regenerator and different decay functions in the riser. In order to obtain a more equivalent comparison, I have built a model similar to the Dasila model, except I used the tanks method to integrate the plug flow models in the regenerator and riser. This comparison should better highlight the differences due to the flow assumptions.
The comparison approach
The plug flow model has been shown to result in much lower coke level for the regenerated catalyst. If I were to use the same coke dependent decay function for both models, the conversions and yields in the risers would be greatly different. Since my objective is to determine the affects of the flow assumptions for the regenerator, I adjusted the parameters in the decay function for the plug flow model until I obtained nearly equivalent conversions and yields for the two riser models.
For the regenerator, both models used the same kinetic model for the oxidation reactions.
The table of results shows that both riser models produced the same amount of coke and thus the regenerators oxidized the same amount of coke.
The major differences are highlighted. (1) The coke levels are an order of magnitude apart. (2) The after burn temperature difference is much higher with the CSTR emulsion model.
The differences in the flue gas compositions are due to the different coke levels. The plug flow model produced a higher CO/CO2 ratio than the CSTR model. This resulted in a slightly lower bed temperature for the plug flow model, even though the same amount of oxygen was consumed.
The increased after burn for the CSTR model is due to extra oxygen available in the freeboard. This results in a higher fraction of the coke being oxidized in the freeboard region where there is less heat capacity to absorb the heat of reaction.
The two highlighted differences are the best indicators for selecting which model is better for a given reactor. A high after burn temperature difference indicates gas bypassing, probably via a bubble or jet phase. A high coke level most likely indicates significant gas backmixing in the dense bed.
Thus far, I have examined (1) a model with backmixing and gas bypassing, and (2) a model with no backmixing and no gas bypassing. Other models can be devised: (3) backmixing and no bypassing, and (4) plug flow with bypassing. Model (3) should produce high coke level and low after burn. Model (4) should produce low coke level and high after burn.