Abstract
Steady-state, transient and dithering characteristics of emission conversion efficiencies of three-way catalysts on natural gas IC engine were investigated experimentally on a single-cylinder CFR engine test bench. Steady-state runs were conducted as references for specific engine emission levels and corresponding catalyst capacities. The steady-state data showed that conversion of HC will be the major problem since conversion of HC was effective only for a very narrow range of exhaust mixture. Unsteady exploration runs with both lean-to-rich and rich-to-lean transitions were conducted. These results were interpreted with a time scale analysis, according to which a qualitative oxygen storage model was proposed featuring the difference between oxygen absorption and desorption rates on the palladium catalysts. Additionally this oxygen storage model also revealed information about parameter selection for dithering to achieve optimal emission operation where all three, HC (mostly CH4), CO and NOx are simultaneously reduced. Dithering with different midpoint Lambda values, frequencies and amplitudes were further investigated. The results showed that the optimal dithering Lambda midpoint resided in slightly rich engine charge. An optimal intermediate frequency was found corresponding to the characteristic time of the emission-favorable Lambda zone as suggested by the oxygen storage model. An intermediate optimal amplitude was also found due to that extra HC and CO emission coming from the richer side lowered down the overall conversion efficiencies when dithering with higher amplitudes. The difference between emission data after the first catalyst and after both indicated that more catalyst capacities are critical for HC conversion.
Original language | English (US) |
---|---|
Pages (from-to) | 1246-1252 |
Number of pages | 7 |
Journal | SAE International Journal of Engines |
Volume | 8 |
Issue number | 3 |
DOIs | |
State | Published - Apr 14 2015 |
ASJC Scopus subject areas
- Automotive Engineering
- Fuel Technology