The cold start catalyst warm-up operation is seen as one of the most important modes in Direct Injection Spark Ignition (DISI) Engine operation. When the catalyst is cold the engine out emissions become the tailpipe out emissions, so it is vital for the catalyst to obtain its working temperature as quickly as possible. A very high exhaust temperature can be achieved with a very retarded ignition - the engine can be made to operate at no load with a close to wide open throttle. With a retarded ignition, a split injection strategy has been shown to improve combustion stability which is critical for the trade-off between tailpipe emissions and vehicle idle stability. The spray guided DISI engine has a multi- hole injector centrally located in the chamber with the spark plug. For catalyst heating operation, the first injection occurs during induction, which forms a relatively well mixed but lean mixture in the cylinder before ignition, and the second injection occurs close to a retarded ignition, which produces a stratified fuel rich mixture in the central region of the combustion chamber near the spark plug. Combustion initialization is found to be sensitive to spark plug protrusion and orientation, injector orientation and 2nd injection timing relative to ignition. High tension current and voltage measurements have been taken in order to characterize the effect of the 2nd injection timing on both the breakdown and the glow phase of the arc discharge. Both phases are shown to be influenced by the timing of the 2nd injection. The richer mixture causes the breakdown voltage to increases while the airflow entrained in the 2nd injection has been shown to stretch the spark and in the worst case extinguish it prematurely. In-cylinder spray imaging by Mie scattering has been taken with frame rates up to 6000 fps, with high speed video photography of chemiluminescence and soot thermal radiation. Tests have studied the effect of the spark plug orientation and injector orientation, with timing sweeps for the phasing of the second injection. The images show interaction of a fuel jet with the earth electrode, stretching of the arc, variable location for the start of combustion and significant cycle-by-cycle variations with the same operating point leading to normal combustion, slow combustion and misfiring cycles. Spectroscopic measurements have confirmed the presence of OH *, CH * and C2*; emissions lines, and their relative magnitude compared to soot radiation. Filtering for CH * has been used with a photo-multiplier tube. These signals show the arc discharge, the delay between the arc and the kernel growth and (depending on the timing of the 2nd injection) small kernels which do not subsequently fully develop and can cause misfiring cycles. Unburned hydrocarbon emissions have been measured with a fast-response FID, so that emissions can be related to: misfiring cycles, slow burning cycles (0 < GMEP <0.5), and normal cycles. These measurements show that only the misfiring cycles lead to significant unburnt hydrocarbon emissions. The misfire mechanism depends on the timing of the 2nd injection. When the 2nd injection ends at the spark, no kernel is seen for a misfiring cycle. However, a kernel is shown to grow in the lean background mixture indicating that the misfire mechanism, when the 2nd injection ends close to the spark, is that the local air/fuel ratio is too rich for the onset of combustion. However, when the 2nd injection is significantly retarded from the spark a different misfire mechanism is present. A small kernel is shown to exist between the spark and the arrival of the fuel from the 2nd injection. For the misfiring cycle, this kernel is extinguished early, possibly due to an interaction between the kernel and the 2nd injection.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:558407 |
| Date | January 2010 |
| Creators | Twiney, Benjamin W. G. |
| Contributors | Stone, R. |
| Publisher | University of Oxford |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
Page generated in 0.0036 seconds