The ever-changing demands from consumers for more fuel efficient vehicles necessitates automotive manufacturers developing ever more radical proposals for engine design.
The latest trend from manufacturers is engine ”downsizing”; where a smaller, more efficient, engine is pressure-charged to recoup the lost full-load performance. An effect of
downsizing is that engines now run at significant intake pressures. Additionally, implementation of simulation techniques as an integral part of the research and design process is becoming commonplace in the automotive industry. Whether or not current models can be considered reliable for combustion prediction at significantly elevated pressures, such as those experienced by downsized engines, is a main focus of the current work, predominantly assessed through how the crank-resolved in-cylinder pressure traces from prediction, compare to experiment. Experimental data was provided by four different engines: two Jaguar Land Rover multi-cylinder engines, one naturally aspirated, the other heavily downsized, and two University of Leeds bespoke research engines operating under naturally aspirated, or high pressure conditions representative of downsized engines.
It was seen that the combustion models were not able to accurately predict combustion at different pressures without adjustment of the turbulence quantities, namely
the length scale used to define the ”after-burning” process. Additionally, it is known that variability of combustion limits the performance of engines significantly; a better understanding of variability which may lead to mitigation methods would result in significant efficiency gains. The magnitude of variability in the four different engines is investigated in this work as well as the current capability in predicting the variability. Once a mean cycle
was successfully matched, the variability of the engine cycles is accurately predicted for all engines and conditions by a random-number model, invoking variability on two parameters, u0 and �, with Gaussian distribution. Moreover a novel method for assessing variability is proposed and is employed in a study to assess the influence of combustion variability at different stages of combustion. The variability of the combustion event was
seen to be a strong function of the very early stages of combustion.
The propensity for autoignition is also known to increase with operating pressure and temperature. The predictive capability, of two autoignition models, was assessed
against experimental data. It was observed that a chemical kinetic based autoignition model was more able to predict autoignition over all engines, vis-´a-vis an empirically
based model. In addition, it was also seen that a variability of autoignition within the engine existed, which was independent of burning rate. Finally, ”knocking” cycles were identified within the heavily supercharged, multi-cylinder research engine which defined the calibration limit of the engine. It was seen, however, that the frequency of pressure oscillations was beyond that traditionally seen for knock cycles.
Altmetric
Altmetric