Longitudinal models are used to evaluate different vehicle-engine concepts with respect to driving behavior and emissions. The engine is generally map-based. An explicit calculation of both fluid dynamics inside the engine air path and cylinder combustion is not considered due to long computing times. Particularly for dynamic certification cycles (WLTC, US06 etc.), dynamic engine effects severely influence the quality of results. Hence, an evaluation of transient engine behavior with map-based engine models is restricted to a certain extent. The coupling of detailed 1D-engine models is an alternative, which rapidly increases the model computation time to approximately 300 times higher than that of real time.
In many technical areas, the Fourier transformation (FT) method is applied, which makes it possible to represent superimposed oscillations by their sinusoidal harmonic oscillations of different orders. In a first step, this paper presents a novel modeling concept for a 1D-engine-model with the Fourier transformation (FT) method, based on a mean value model providing the prerequisites for real-time capability, which has been developed at AUDI AG in cooperation with the Institute of Internal Combustion Engines and Automotive Engineering (IVK) at the University of Stuttgart. By means of the mathematical FT-concept, it is possible to represent pressure pulsations emerging inside the engine air path and so, to improve the standard of a mean-value model. Pressure pulsations have a strong influence on the engine’s internal behavior such as for the gas exchange of the cylinder, the responsive performance of the turbo charger, etc.
The quality of predictability of emission models strongly depends on the quality of its boundary conditions, delivered by the respective combustion calculated in a preliminary step. Models with simplified representation of the air path and combustion are not able to serve information of high enough accuracy for a subsequent emission prediction. In this case, emission models compensate for imprecise boundary values. The resulting disadvantage; the predictability of the model decreases such that a validation does not give any indication of whether an error derives from the emission, or the preliminary simulated combustion.
In a second step, this paper presents a NO-emission model which is based on a semi physical principle. The relevance of cycle specific in-cylinder values were investigated on their effects on NO-production. The functionality of each parameter was designed and calibrated on the basis of NO-measurements. An evaluation of the model for different validation engines showed a very high agreement for both stationary and transient engine operation.