Flow uniformity characterization in catalytic converters under turbulent inlet conditions
Projektarbete, avancerad nivå
In the sectors utilizing the combustion of fossil fuels, catalytic converters are commonly employed to reduce dangerous gases generated by combustion engines. For greater efficiency, it is critical to understand the exhaust flow homogeneity inside the converter. This work attempts to evaluate the added value of using Detached Eddy Simulation (DES) simulation over Reynolds-Averaged Navier-Stokes (RANS) simulation for analysis of turbulent flow inside catalytic converters. The meshing and simulation of the domain was carried out using ANSYS FLUENT and ANSYS Workbench. Initially a RANS simulation with k-ω SST turbulence model was carried out for a specific case to understand the behavior of the time averaged flow field with modeled turbulence and its effect on uniformity index. The results from the RANS simulation were, together with definitions of turbulent length scales, used to develop a mesh for DES simulation. In this process it was found that in order to resolve a sufficient amount of turbulence upstream of the monolith inlet, the DES simulation required a considerably finer grid than the RANS simulation. The DES simulation with k-ω SST model was used to simulate multiple retention times, attempting to achieve a quasi steady flow with resolved velocity fluctuations. However due to constraints on computational power, simulating until statistical convergence was not possible. The time averaged quantities were extracted from the DES simulation in order to make a fair comparison with RANS results. The uniformity indices of RANS and DES were then compared throughout the monolith of the catalytic converter. A difference of ∼1-1.5% in uniformity index was found from this comparison. Further more it was found that the resolved turbulence in the DES generates fluctuations of ∼3-6% in the uniformity index based on instantaneous quantities. The work concludes that for the exact case simulated, the possibility of increased accuracy when predicting uniformity index with DES instead of RANS does not outweigh the increased computational cost. However for different flow conditions, with larger mean velocity magnitude and less uniform flow, the measured difference between RANS and DES could make a larger difference for prediction of catalytic conversion efficiency. Hence in these cases it is possible that DES simulations are worth their computational cost. Whether this is the case or not would have to be examined through further research before coming to any definitive conclusions.
Catalytic converters , Flow uniformity , Detached-eddy simulation , Reynolds averaged navier stokes simulations