Lean Lifted Flame Combustion (LLFC)
Non-sooting combustion under diesel-like conditions is known to be achievable in several ways; e.g., lean fuel-air charge combustion, use of oxygenated fuel, high injection pressure, small injector orifice, cooling ambient gas temperature, EGR and so on. The concept behind those techniques share the same goal; an enhancement of fuel-air mixing and accordingly lean mixture charge formation prior to partially premixed combustion phase where soot precursors start to form. To this end, increased mixing residence time and mitigated charge stratification level are prerequisite; i.e., lean lifted flame combustion (LLFC) is one such potential strategy as mixing-controlled combustion technique that may avoid soot forming process.
Seeking a means of LLFC strategy, a new spray injection configuration, termed ducted fuel injection (DFI), was invented by Charles Mueller [1, 2], who leads the research team at Sandia National Laboratory. The idea is that fuel spray is injected through a small duct installed at a nozzle offset distance (G) downstream of injector nozzle as illustrated on the left figure. By doing so, the injected fuel travels through the small channel maintaining its axial momentum without being dragged down by ambient air entrainment; improving turbulent mixing through the duct channel. This method ultimately is designed to enhance premixing combustion and mitigate soot formation.
The presence of duct channel could limit over-mixing at the spray periphery and lead to cooler mixture due to cold duct wall, resulting in more premixed charge in favor of low soot conditions. Furthermore, the higher scalar dissipation rate potentially generated at the duct exit could make it difficult for the premixed combustion to initiate at the duct outlet, thereby enabling secondary turbulent mixing with ambient air downstream of the duct exit.
Mueller et al. [2] conducted initial proof-of-concept study on DFI and demonstrated the potential benefits by using the constant-volume combustion vessel (CVCV) at Sandia Combustion Research Facility (CRF). The tested spray flame was visualized by their optical diagnostics of natural luminosity and OH* chemiluminescence. The features of more distanced lifted flame and non-sooting combustion are evidenced by hot soot incandescence as shown below.
The first-of-its-kind computational study on DFI concept was attempted during my dissertation [3] study at Georgia Tech. The study used CONVERGE CFD software for modeling the mixing and chemical reaction process. The purpose of the study is to demonstrate the LLFC concept and its effect on soot formation. To this end, a non-reacting condition (zero oxygen) and a reacting simulation (21% oxygen) under ECN Spray A injection condition (Tamb = 900 K) [4] were performed using the Reynolds-Averaged Navier-Stokes (RANS) simulation technique with the turbulent stress term approximated through a relevant model. The combustion process was modeled by well-stirred reactor (WSR) concept and the soot formation process was approximated by using Hiroyasu-NSC soot oxidation model [5].
The figure on the left illustrates the difference in spray-driven flow field between free-spray (upper half figure) and DFI spray (lower half figure) configurations under non-reacting ambient gas condition. It is noted that the streamlines appear to keep entering into the spray jet via ambient air entrainment in the free-spray jet; thus the fuel-air mixing is continuously sustained throughout the jet, whereas the presence of duct prevents the air entrainment from feeding the spray with ambient gas momentum along the duct length; then the secondary fuel-air mixing is allowed downstream of duct exit. Since in the DFI spray the liquid jet may have maintained its inertia to some extent without being disturbed by the ambient air flow, the jet exiting out the duct should convect further downward at higher velocity than the free-spray configuration.
The left figure displays the lifted spray flames contoured by temperature and soot concentration captured at 1.3 ms after start of injection. The ducted spray flame is found to be further lifted downward compared to the free-jet spray, yielding weak concentration soot quantity across the entire reacting spray jet. This effect is attributed to the leaner and less stratified mixture formation prior to premixed combustion phase, since the duct may have promoted the secondary mixing after the spray jet penetrates through the duct channel. The cross-sectional average soot mass is also quantified along the spray axis and presented below:
One of the major outcomes of my dissertation research was highlighted by Mueller's research team in their pursuit of proof-of-concept study [6]. Using GTFOAM, the OpenFOAM-based library I developed, they carried out numerical simulations to investigate mixing enhancement through the DFI strategy. Their simulation results showed excellent agreement with experimental measurements (see left), establishing a strong basis for understanding DFI-driven mixing improvements and its potential role in reducing soot formation.