No abstract

A constant volume chamber system was used to characterize the ignition and combustion of a diesel fuel. Fuel is injected in the chamber through a standard diesel common rail injection system. Injection parameters and initial chamber conditions were varied: fuel injection pressure (600 to 1000 bar), injection duration (500 to 3000 μs), initial temperature (808 to 923 K), initial pressures (6 to 21 bar), and oxygen mole fractions (15 to 30% diluted in nitrogen). Multiple measurements confirmed the repeatability of the measurements. Most of the experiments showed a two-stage combustion process, especially at low chamber pressure and highly premixed conditions. This is characteristic of large paraffinic hydrocarbons such as those composing around two-thirds of commercial diesel fuels. Combustion pressure data from diesel fuel have shown a moderate effect of injection pressure and injection duration on the ignition delays, especially when compared to the effect of other operation parameters. A strong effect of initial pressure and temperature was also observed. Increasing the initial pressure leads to nonlinear decreases in the ignition delay, this effect being especially strong at low pressures. Further, an increase in the oxygen mole fraction advances the onset of ignition, this effect being more intense at the lowest oxygen mole fraction in the oxidant. Two separate regions with different temperature-dependence functions were detected, indicating two different fuel oxidation chemistries. Arrhenius-type correlations were obtained for both regions, including the most influential parameters, to predict the firstand second-stage ignition delays.

New diesel engine combustion concepts, such as the homogeneous charge compression ignition (HCCI), have encouraged the development of ignition delay correlations allowing the reliable prediction of the chemical auto-ignition time of the fuel, which depends on its oxidation kinetics. These correlations permit one to design and optimize the most adequate control strategies leading to both low emission levels of pollutants and proper engine performance. Although some diesel ignition correlations can be found in the literature, most of them consider the physical delay time and have been obtained only for high temperature values, which are usual at the end of the compression stroke in traditional diesel engines. However, there is still a significant lack of information regarding the complex three-stage oxidation (low-, intermediate-, and high-temperature range) of a diesel fuel under HCCI conditions. Thus the proposal of delay time correlations for the three temperature ranges is the objective of the present work. The correlations have been assumed as Arrhenius-type equations and consider the effect of the main parameters affecting the auto-ignition time, i.e. pressure, temperature, and equivalence fuel/air ratio. The adjustment coefficients have been calculated by using multiple linear regression and least-squares techniques, and a very good fit between modelling-predicted and correlation-predicted delay values has been obtained ( R2 higher than 0.95 in all cases).

The integral method proposed by Livengood and Wu has been traditionally used to predict the occurrence of knock on spark-ignition engines. Due to its simplicity and low computational demand, this is a method of great interest for the prediction of another autoignition phenomenon, such as the onset of combustion in compression ignition or homogeneous charged compression ignition engines. However, the simplicity of the method is a consequence of the restrictive assumptions considered during its development, which may limit the applicability of the equation. In this study, the validity of the correlation proposed by Livengood and Wu has been evaluated at different initial operative conditions under pure homogeneous charged compression ignition combustion mode for fuels with practical interest (hydrogen, methane, ethanol and n-heptane). The integral method has shown very good prediction capability for the fuels, which do not present two-stage heat release (hydrogen, methane and ethanol) except in those cases when the onset of combustion is very delayed. When cool flames appear (as in the case of n-heptane), the integral method overpredicts the autoignition times since it does not consider the first stage of heat release. In these cases, the prediction of the integral method may be improved if the whole combustion process is considered as two individual processes. This approach shows fairly good prediction capacity although it is unpractical since the simulation of the second-stage combustion requires the previous calculation of the composition of the mixture and the temperature increase at the end of the first stage. Finally, two alternatives to the original integral method are tested which keep its simplicity and universality while taking into account both first and second heat release, one of them showing better results than the original Livengood and Wu equation.

No abstract