Time to "eat crow"
I previously reported that the plug flow and CSTR bed models produced greatly different coke levels. I finally realized that the results were very sensitive to even small differences in the flue gas results which I was trying to keep constant. This lead to an "apples to oranges comparison."
I decided that a better comparison was to use identical kinetics in both models and also identical decay functions. Under this comparison, the two models produce more similar results, as shown below.
The Plug vs. CSTR comparison
Now, the coke level comparison is not consistent. Without gas bypass (i.e. no bubble phase), the CSTR produced the higher coke level. With bypass, the plug model produced the higher value. The bed temperatures, however are consistently lower for the CSTR, but the difference decreases as gas bypass is added. These inconsistencies demonstrate the non-linear complexity of the system.
The two models produced identical after burn values. The results demonstrate that after burn temperature differences greater than a few degrees are an indication of gas bypassing.
I haven't made the series of runs that vary the process conditions (e.g. air and oil feed rates). It will be interesting to see how the two models compare regarding trends because the trend prediction is the main objective for the model.