Identification of the Error Introduced in DI Diesel Engine Phenomenological Multi-Zone Models from Assumptions Related to the Initial Conditions at the Nozzle Exit 2010-01-0153

The past decade significant research effort has concentrated on the DI diesel engine due to stringent future emission legislation which requires drastic reduction of engine tail pipe pollutant emissions, mainly PM and NOx, without significant deterioration of specific fuel consumption. Towards this effort, the important role of modeling to investigate and understand the impact of various internal measures on combustion and emissions has been widely recognized. Phenomenological models can significantly contribute towards this direction because they have acceptable prediction capability and the advantage of low computational time. This enables the production of results, on a cycle basis, that indicate the effect of various parameters on both engine performance and emissions. Therefore their use can significantly reduce engine development time (i.e. reduction of experimental effort) and cost.

For performance prediction simple models i.e. single or two zone ones are adequate and well proven. However for emissions it is widely recognized that a more fundamental and detailed description for the air fuel mixing and combustion mechanism is necessary. Multi-zone phenomenological models offer a solution to this problem and provide a reasonable compromise between simple phenomenological and detailed CFD models.

The authors have developed in the past a multi-zone combustion model, which has been applied to simulate various, engine configurations. From the simulations, it has been revealed that the description of the air fuel mixing mechanism is crucial. The last for modern DI diesel engines is greatly affected from the fuel injection process and specifically the fuel injection rate which in multi-zone models is usually either assumed to be constant (i.e. mean value providing the actual injection duration) or variable (using simulated or measured data). The second approach is adopted in current advanced multi-zone models. However, these normally assume a constant injection pressure difference at the nozzle exit for the entire injection period, to avoid zone overlapping, which decouples the fuel mass injection rate from the instantaneous zone velocity and its momentum. This can introduce errors in the estimation of the air entrainment rate, which is based on momentum conservation, when the injection rate differs from a constant one. The highest effect is expected to exist at initiation and termination of injection where injection rate varies considerably. For this reason in the present work, an existing multi-zone model is used to investigate the effect of initial zone velocity on engine performance and emissions. There have been examined two main initial injection velocity schemes: constant injection velocity and variable injection velocity. In the second case (variable velocity) zone, overlapping is to be expected which normally leads to zone mixing. However, in the present investigation zone mixing is neglected because the objective is to examine the fuel air mixing and combustion mechanism on an individual zone basis. The derived results of performance and emissions from both initial zone velocity considerations are compared to measured values corresponding to the ESC cycle of a HD DI diesel engine. From this comparison, the error introduced from the assumption of a constant fuel injection velocity on engine performance and emissions is revealed.

Following this, an attempt is made to compensate for the error of air entrainment rate, introduced from the assumption of a constant fuel injection velocity, using a correction for individual zone air entrainment rate to account for its actual initial velocity (i.e. actual initial momentum). As shown from the results, it is possible to derive a physically based correction factor that allows the use of a constant injection velocity with minimum error on predicted engine performance, NOx and Soot emissions. This will assist the development of multi-zone models that describe properly the air fuel mixing mechanism and at the same time avoid zone overlapping which creates discrepancies in the resulting jet geometry and increases model complexity.