Comparison of plug flow & CSTR bed models
I have completed the parametric studies with two models. The models differ only in the flow assumptions for the emulsion phase in the regenerator...one assumes plug flow and one assumes a CSTR for the gas. Both models assume a bubble phase exists and that there is no mass exchange between the bubble phase and the emulsion phase. Thus, no reaction occurs in the bubble phase and therefore the bubble phase air merely bypasses the bed before it enters the freeboard region.
The effects of the operating variables...air rate and temperature, oil rate and temperature, and catalyst circulation rate were all examined. Since the model differences center on the gas flow in the regenerator, I will report the effect of air rate in this post.
The coke levels are slightly different, as reported previously. The trends appear different also, but the differences are very minor.
The most important trends are the product yields in the riser reactor.
Again, the trends are very similar. Both models produce nearly equal slopes for the yield curves.
The trends for the other operating variables were also very similar for the two models. Based on the similar trends, it appears that either model can be used for optimization. My choice would be the CSTR model.
The highest gasoline yield (blue curves) is at the lowest air rate. This air rate is at the rate where the onset of a bubble phase was assumed. Higher air rates increase the amount of air bypass. Both models indicate that increasing the air rate beyond the bubble phase onset is undesirable from the standpoint of gasoline production.
The mathematical models used in these comparisons are much simplified from the tanks models previously used. Now, I use differential equations for all of the plug flow regions. The CSTR model contains one nonlinear mass balance.
The tanks models made it easy to include mass exchange between the bubble and emulsion phases. I will try to add that feature back to the ODE models.
Next: an important model assumption