Recent years have witnessed a dramatic increase in global ethanol production, while cellulosic feedstock or the algae-based production approach make more sustainable ethanol production foreseeable in many countries. The ethanol produced will increasingly penetrate the markets not only as blending component, but also as main fuel component, boosting demand for flex-fuel vehicles.
One of the main challenges for flex-fuel vehicles is the cold start due to the poor vapor pressure of ethanol. This is detrimental to starting capability in DISI engines in particular, with increased cylinder wall wetting causing higher oil dilution.
The most efficient solution for DISI engines is a smart injection strategy, enabling fuel vaporization during injection in the compression stroke. But this requires optimum injection parameters such as injection timing, split ratio and rail pressure.
A quasi-dimensional engine start simulation model was developed to drastically reduce the experimental workload involved in cold start realization and optimization. It permitted efficient investigation of the impact of parameters such as fuel vaporization properties, injection parameters, cam profiles and boundary conditions such as temperature and pressure. The model was based on a one-zone approach and ignored the temperature differences in the combustion chamber during mixture formation, resulting in poor quantitative accuracy of the required fuel mass.
A new approach described in this paper has introduced a multi-zone model, considering different conditions for vaporization in the spray zone compared to pure air zone. Vaporization in the spray zone significantly reduces the temperature there, especially for fuels with high vaporization heat, e.g. alcohols, following the withdrawn heat of vaporization. The result is a notable delay in vaporization, an effect which must be considered by simulation tools.
The multi-zone approach was introduced to the existing one-zone simulation model. The model then formed the basis for estimating optimum injection parameters (timings, split ratios, pressures) for certain cold start temperatures.
Subsequent engine tests showed a good correlation between simulation and experimental results particularly for low temperatures. A multi-zone vaporization model that incorporates spray volume data is thus recommended for efficient cold start optimization.