Synthetic Gas Bench And Simulation Tool Characterisation of Three-Way Catalytic Converter

Abstract

A three-way catalytic converter (TWC) is a necessary tool in order to lower the emissions from stoichiometric combustion such the emissions from petrol-fueled vehicles. The TWC can efficiently reduce the three main pollutants: CO, hydrocarbons (HC) and nitrogen oxides (NOx). Once the TWC is used it may suffer from deactivation, which affects the performance of the catalyst. This master thesis was done in collaboration with Aurobay and the experiments were performed at the Competence Centre for Catalysis at Chalmers University of Technology. Three catalyst samples, one from a stabilised fresh commercial three-way catalyst and two from different parts of a rapidly aged TWC were investigated using flow-reactor experiments in a synthetic gas bench (SGB) to study the effect of the deactivation procedure. These results showed a clear axial distinction in the degree of deactivation in the catalyst. Following this was a simulation using the commercial simulation program Gamma Technology Suite (GT-suite), where the kinetic parameters of the model were calibrated and optimised against the experimental results of the stabilised fresh catalyst. For the optimisation, constant temperature steady state tests at three different temperatures were used, as well as oxygen storage tests to calibrate the reactions on the cerium oxide sites of the catalyst. The optimised model parameters were then validated against a SGB light-off test. The optimised simulation model was also tested against a SGB light-off test for the deactivated catalyst sample to investigate what measures had to be taken to correctly simulate the ageing procedure. Finally, the same model parameters were tested against values from a motor bench test performed at Aurobay, to study how the model would hold up against a more real life-like test. The parameters for the cerium oxide sites could not be optimised but instead left to their baseline parameters. In the end, the final optimised model showed good simulation results for the reduction of NOx but not as good results for the oxidation of CO and HC. When tested against the experimental values of the deactivated catalyst sample, the parameters from the optimised model performed better than the baseline parameters, despite the model being optimised on the stabilised catalyst. The same was seen for the motor bench experiments, however the simulated values still showed a substantial difference from the actual results. The optimisation processes show that the precious metal dispersion factor has a considerable effect on the final simulated results, and it might be worth to focus more on calibrating these parameters rather than optimising the kinetic parameters as was the main focus in this report. The thesis therefore concludes that further investigations are desired to quantify how crucial the need of unique experimental parameter setting actually is.