I found an error in the carbon balance...a single incorrect subscript. This changed the results and made my last post invalid, so I have deleted it.
The results for the base case now show an order of magnitude higher level of coke on the catalyst and higher temperatures throughout.
The higher coke levels were what I had expected when I started the model. The CSTR emulsion phase in this model should have a lower overall reaction rate than the plug flow model by Dasila, et al. because of the lower oxygen level. To achieve the coke removal needed to balance the coke produced in the riser reactor, the coke level in the regenerator has to increase to compensate. In addition to the differences in the flow models used by me and Dasila, I have used different kinetics for the regenerator. This change also could contribute to the different coke level.
The high temperatures compared to Dasila's results are due to the more complete oxidation provided by the kinetics I have used. Note that the CO2/CO ratio is nearly the inverse of the Dasila results. Since Dasila's temperatures match the plant data values, the kinetics she used are more appropriate for that reactor.
My model's temperature difference between the dense bed and the flue gas appears to be typical for regenerators that operate in the complete combustion mode.
Effect of air rate
Increasing the air flow to the regenerator increases the temperatures in the system and increases the coke oxidation. For this model, I have assumed a modest level of mass exchange between the bubble phase and the emulsion phase. Thus, increasing the air rate does allow more oxygen to enter the emulsion phase even though the air flow to the emulsion phase is fixed according to the model assumptions.
The oil conversion and gasoline yield also increase with increasing air rate.
The flue gas composition varies as shown below. The decreasing CO2/CO ratio is apparently due to the differences in activation energies for the reactions. The reaction with the highest activation energy is the oxidation of coke to CO.
A model has been completed that can incorporate the results of a tracer study on the regenerator. The regenerator model is integrated with a simple model of the riser. Improvements to the riser model are certainly possible. For example, the model assumes that the catalyst and gas have the same velocity. Instead, a slip velocity could be used. Also, additional detail could be provided for oil vaporization and heat exchange between the catalyst and the fluids in the riser. I believe that these changes would have to be based on theoretical models rather than experimental studies. Given the uncertainty regarding the kinetics of the oil mixtures, I'm not certain that additional detail in the model is justified.
The 5 lump model has been used for the riser. The model can be easily modified for other lumping schemes.
The regenerator flow model assumed is simple. If the tracer study indicates one of the more complex models is needed (see earlier posts on tracer studies), the model structure using a set of "tanks" will make modification easy (:-)
Dasila, Prabha K, Indranil Choudhury, Deoki Saraf, Sawaran Chopra, and Ajay Dalai. “Parametric Sensitivity Studies in a Commercial FCC Unit.” Adv. in Chem. Eng. Sci. 2012, no. January (2012): 136–49.