Abstract

In the present study Large Eddy Simulations (LES) are performed to investigate the ignition behavior of cavitating and non-cavitating n-dodecane flames. Inspired by the work of Tsang et al. (2019) [1] an LES specific dynamic dispersion model, that evaluates the dispersion velocity locally, is proposed. Although the model discards tuning of global model constants, excellent mixing predictions are obtained for all cases. The resulting model is extensively validated using inert Spray A conditions as defined by the Engine Combustion Network (ECN). Subsequently, it is applied to the larger orifices of Spray C and Spray D. A Flamelet Generated Manifolds (FGM) approach that takes the effect of scalar dissipation into account is adopted for combustion modeling. The coupling of the turbulence modeling approach and FGM shows excellent predictions of ignition characteristics on Spray C and Spray D, suggesting a minor effect of cavitation on ignition development. For the sake of understanding the injection-to-injection variations of LES, multiple realizations are performed. Based on the analysis of structure similarity index (SSI), it is found that a single realization is sufficient for global parameters such as ignition delay time (IDT) and lift-off length (LOL). However, different number of realizations are needed for different scalar fields. It is suggested that the temperature-based IDT is preferred for a single realization while a radially integrated intensity is needed for an OH-based IDT or LOL.

Highlights

  • The complex turbulent spray processes taking place in an internal combustion engine are driven by the mixing of fuel and air, leading to the chemical reactions that release a fuels energy

  • Based on the analysis of structure similarity index (SSI), it is found that a single realization is sufficient for global parameters such as ignition delay time (IDT) and lift-off length (LOL)

  • It is suggested that the temperature-based IDT is preferred for a single realization while a radially integrated intensity is needed for an OH-based IDT or LOL

Read more

Summary

Introduction

The complex turbulent spray processes taking place in an internal combustion engine are driven by the mixing of fuel and air, leading to the chemical reactions that release a fuels energy. The different configurations lead to significantly different flow development inside the nozzle, and have an impact on the morphological spray characteristics In this investigation, both the inert and reacting cases of Spray C and Spray D are simulated and compared by directly implementing spray parameters obtained in experiments, such as the discharge coefficient, spreading angle, etc. Dedicated investigations suggest that neglecting TCI leads to the failure of capturing the important physics during combustion processes, when higher temperatures or fuel-rich conditions are encountered [33,34,35,36] In this sense, here a presumed probability density function (PDF) using the top-hat filter is adopted to describe the local flame structures.

Basic equations and numerical methods
Gaseous phase modeling
Liquid phase modeling
The dispersion model
Case description
Computational setup
Governing equations for the non-reacting spray
Spray A validation
Performance on Spray C and Spray D
Simulation of reacting spray
Governing equations for the reacting spray
Turbulence-chemistry interaction
Simulation results
Ignition delay and flame lift-off
Findings
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call