Experimental and Kinetic Modeling Study of Aged Natural Gas Three-Way Catalyst Performance

<div class="section abstract"><div class="htmlview paragraph">The applicability of three-way catalyst (TWC) models for system-level aftertreatment simulations under transient operating conditions of natural gas engines depend on accurate integration of reaction kinetics as a function of the air-fuel equivalence ratio lambda(λ). A comprehensive global kinetic model has been developed for an aged commercial three-way catalyst (TWC), incorporating key reaction pathways including oxidation of CO, CH₄, C₂H₆, and H₂; reforming of CH₄ and C₂H₆; the water-gas shift reaction; and NO reduction via CO and H₂. The model also accounts for oxygen storage capacity (OSC) and its dynamic interaction with CO and H₂. To calibrate kinetic parameters, systematic bench-scale flow reactor experiments were conducted under lean, stoichiometric, and rich conditions. Performance metrics focused on CH₄ and C₂H₆ oxidation and reforming across varying O₂ and CO concentrations, and NO reduction with CO and H₂ under different oxygen levels. Experimental results revealed that CO suppresses the reforming of CH₄ and C₂H₆. NO conversion was observed between 150°C and 600°C, with H₂-driven reduction producing NH₃, N₂, and N₂O depending on lambda (λ). Under rich conditions, complete NO conversion occurred from 150°C, while lean conditions showed reduced NO conversion at elevated temperatures due to H₂ oxidation. NO reduction with CO initiated at 250°C, achieving full conversion under rich conditions. The model accurately captures the influence of λ on NO reduction with both H₂ and CO, predicts NH₃ formation under rich conditions, and simulates H₂ generation via the water-gas shift reaction above 400°C. It successfully reproduces λ sweep data (λ = 0.95–1.02) and demonstrates CO inhibition effects on H₂ oxidation and NO reduction. This global model is validated with dithering reactor data and qualitatively captures key trends in data which aids in catalyst sizing, calibration robustness and the system level modeling of end of useful life parts. Further validation of the current developed model with lean-rich cycle tests confirms the model’s ability to predict NO_x slip at the onset of rich cycles impacting the ability to accurately predict NO_x emissions during engine braking events in system level models.</div></div>