Pdf Modeling of High-pressure Turbulent Spray Combustion under Diesel-engine-like Conditions
Open Access
- Author:
- Bhattacharjee, Subhasish
- Graduate Program:
- Mechanical Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- September 18, 2012
- Committee Members:
- Daniel Connell Haworth, Dissertation Advisor/Co-Advisor
Stephen R Turns, Committee Member
Savas Yavuzkurt, Committee Member
Andre Louis Boehman, Committee Member - Keywords:
- Turbulence
Combustion
TCI
WSR
PDF - Abstract:
- This thesis is focused on understanding the extent to which turbulent fluctuations in composition and temperature influence global ignition characteristics (e.g., ignition delays and liftoff lengths) and flame structure for high-pressure, transient, autoigniting spray flames under diesel-engine-like conditions. Turbulent spray flames for two single-component fuels (n-heptane and n-dodecane) are simulated. The modeling framework is a hybrid Lagrangian-particle/Eulerian-mesh probability density function (PDF) method. This framework allows for arbitrarily large chemical mechanisms, and features Lagrangian-based spray breakup and dispersed-phase models, soot models, and an optically thin radiation model. The influence of turbulent fluctuations is explored by comparing results from the PDF method (which explicitly accounts for turbulent fluctuations) with those from a model that neglects the influence of turbulent fluctuations on local mean chemical reaction rates (a well-stirred reactor – WSR – model) for the same chemical mechanism. Computed results are compared with experimental measurements that are available through the Engine Combustion Network. Here a 40-species mechanism has been adopted for n-heptane, and a 103-species chemical mechanism for n-dodecane. Overall, it is found that for conditions that correspond to robust diesel combustion (e.g., high initial temperatures, high initial pressures and/or high oxygen concentrations) the computed liftoff lengths and ignition delays for the WSR and PDF models are close to each other, and both are in good agreement with experiments. For less robust conditions (e.g., low initial temperatures and/or low oxygen concentrations), the computed liftoff lengths and ignition delays from the two models can be significantly different, and the results from the PDF model are generally in better agreement with measurements. The differences between the two models are especially apparent for n-dodecane at low initial temperatures. For n-dodecane at an initial temperature of 800 K, the WSR model fails to ignite, while the PDF model shows a distinct two-stage autoignition process and the computed ignition delay and liftoff length are within 30\% of the experimental values. For n-dodecane at an initial temperature of 900 K, the WSR model predicts an ignition delay that is three times higher than the measured value, while the PDF model prediction is within 5\% of the measurement. For both fuels and for all initial conditions, the WSR and PDF models produce significantly different turbulent flame structures, and the differences are greater for lower initial temperatures and/or oxygen concentrations. The WSR model produces a laminar-like flame structure, whereas the PDF model produces a broader turbulent flame brush that is qualitatively more consistent with what is expected for a turbulent flame, and with what is observed in the experiments. While it has been shown in the literature that some global characteristics (e.g., ignition delays and liftoff lengths) of high-pressure turbulent spray flames can be captured using a WSR model and a chemical mechanism that has been tuned for this purpose, the present results suggest that this approach effectively amounts to changing a model for one physical process (chemical rate coefficients) to account for deficiencies in modeling a different physical process (turbulent fluctuations in composition in temperature). It is expected that by properly accounting for turbulent fluctuations, it will be possible to develop a model that can be applied over a broad range of engine-relevant conditions without changing the model coefficients.