Avoiding false runaway prediction by properly describing axial conductive heat transfer in a low dt/dp industrial-scale packed bed reactor

Abstract

The impact of axial conductive heat transfer on the simulation and prediction of hot spots and, more particularly, runaway, within an industrial-scale wall-cooled packed-bed reactor with low tube-to-particle diameter ratio (dt/dp ∼ 3) has been assessed in this work. To this purpose, the effective axial conductivity (keff,z) was first determined from adiabatic experiments in a bench-scale packed bed in the absence of reaction. Next, non-adiabatic and non-isothermal bench-scale and industrial-scale packed bed experimental data were used in the assessment of the impact of axial heat conduction on the prediction of the temperature profile. The so-called Pseudo-local Internal Heat Transfer Approach (PL-IHTA) was used for the adequate description of the radial temperature gradients. Finally, an industrial-scale wall-cooled packed bed reactor for the highly exothermic catalytic oxidative dehydrogenation of ethane was modeled to assess the impact of the axial thermal conductivity on temperature profile simulations. When disregarding axial heat conduction or adopting literature-based values for keff,z, determined from either experiments at low dt/dp ratios (<8) and Rep ≤ 1000 or high dt/dp ratios (>8) and Rep ≤ 700 at a wide panel of operating configurations, runaway is simulated at conditions where it has not been observed experimentally or a negligible hot-spot prediction is obtained. When considering the keff,z determined for the specific reactor configuration in this work, hot spots are predicted at Rep of 700 and 1400, but no runaway. The discrepancies between experimental findings and temperature profiles simulated using literature-based values for keff,z indicate the need for a specific determination of keff,z in packed bed reactors with low dt/dp ratio to more accurately predict hot spots, resulting in more reliable reactor design and operation.