Skip to main content

Predicting thermal excursions during in situ oxidative regeneration of packed bed catalytic fast pyrolysis catalyst

Publication Type
Journal Name
Reaction Chemistry & Engineering
Publication Date
Page Numbers
888 to 904

DOE’s Bioenergy Technology Office (BETO) has funded development of an ex-situ process for Catalytic Fast Pyrolysis (CFP) that uses a packed bed of catalyst to upgrade pyrolysis vapors produced in an upstream reactor. The current catalyst champion is a Pt/TiO2 formulation in the form of 0.5 mm diameter spheres. This catalyst has high activity and also allows for long time on stream before requiring oxidative regeneration to remove coke. The current design concept for the upgrading system makes use of a swing reactor system which allows coke deactivated beds to be regenerated in-situ while other beds remain on-line for vapor upgrading. Since the regeneration is done in-situ, it has to be performed very carefully to avoid irreversible deactivation and/or physical degradation of catalyst pellets due to localized thermals: excessive temperatures, temperature gradients, or heating/cooling rates. Collectively these are referred to as “thermal excursions”.

Two finite element computational models have been built in COMSOL Multiphysics® 5.5 to assist in scaling up the regeneration process from a lab scale packed bed with 100 g of catalyst to a pilot scale packed bed with 2 kg of catalyst and internal cooling tubes. One of the most important parameters in the models is the effective thermal conductivity of the catalyst pellets. Based on transient measurements of average outlet temperature and effluent CO2 concentration during regeneration of the lab scale catalyst bed, and using an assumed coke profile and activation energy, this paper demonstrates that an entire range of models with widely differening combinations of effective thermal conductivity and wall heat transfer coefficient can fit the lab scale data equally well. However, for the modeled regeneration of the upscaled 2 kg bed, the study reveals that each combination gives a somewhat different prediction for location and magnitude of thermal excursions. The differences between simulations using four best-fit parameter pairs depend on the amount of heat removed by the process and cooling gases: the more heat removal is shifted to the cooling gas, the greater the differences. Over the entire pilot scale simulation space, the span in thermal excursions is huge. For example,: the maximum temperature gradients in the bed range from a low of 30 ⁰C/cm to almost 3,000 ⁰C/cm!

Since minimizing thermal excursions is crucial to the success of the packed-bed ex-situ CFP process, this study points to the need for additional thermocouples in the lab scale bed, additional variations in process conditions, and careful bed dissections to determine the true coke profile. The model would also benefit greatly from direct measurements of the effective thermal conductivity of the catalyst pellets .