Willkommen Welcome Bienvenue Split Injection for CNG Engines Patrik Soltic, Hannes Biffiger Empa, Automotive Powertrain Technologies Laboratory
Motivation CNG engines are gaining on importance in the stationary and in the mobility sector (availability/costs of natural gas, CO 2 advantage versus crude oil product, low pollutant emissions with comparably simple exhaust gas treatment, full compatibility with alternative methane such as biogas or synthetic NG) Methane is a very good fuel for internal combustion engines because of its knock resistance Methane is hard to ignite ( high ignition energy/voltage needed for high BMEP operation, especially for lean or EGR concepts) and the early flame development phase is slow It is known that the addition of small amounts of hydrogen massively enhances inflammability, combustion stability and increases engine efficiency Up to now: port fuel injected concepts are dominating nearly perfect premixed air/fuel (& EGR) Main question: is there an advantage for stratified concepts (with or without hydrogen addition)?
State of Research Pure methane DI λ=1 (no EGR) operation on a Volkswagen EA111 engine cyl2+3 cyl1+4 DI in direction of the spark plug shows a massive acceleration of combustion (1) This is mainly a turbulence (not a stratification) effect (2) Pure DI is problematic because large gas volumes have to be injected which needs very special injectors (1) Soltic P, Egli R, Mauke D, Wright Y M, Bach C, Strömung, Gemischbildung und Verbrennung bei Methandirekteinblasung im homogenen λ = 1 Betrieb : Simulationen und Versuchsergebnisse, in 6- Tagung Gasfahrzeuge, 26.-27.10.2011, Stuttgart, 2011. (2) Schmitt M, Hu R, Wright Y M, Soltic P, Boulouchos K, Multiple Cycle LES Simulations of a Direct Injection Natural Gas Engine, Flow, Turbul. Combust., 2015.
Question: is there an advantage if only a small amount is directly injected (into a premixed environment)? 8bar Basis PFI M100 PFI M85H15 PFI M75H25 PFI methane methane/hydrogen 8bar H2 DI DFI H100 PFI M100 Massflow meter 30bar PFI methane Hydrogen Direct fuel injection 8bar PFI CH4 DI DFI M100 PFI M100 30bar Druckregler PFI methane
Experimental Approach Swissauto 0.25 l single cylinder engine on test bench (bore 75mm, stroke 56.5mm, compression ratio 12.5, operated at 3500 rpm and WOT)
Experimental Approach air PFI DI
Experimental Approach Swissauto 0.25 l single cylinder engine on test bench (bore 75mm, stroke 56.5mm, compression ratio 12.5, operated at 3500 rpm and WOT) Single-hole DI injector with bent injection direction 8 Side-mounted DI injector (between intake and exhaust valves) with four injection angles experimentally covered
Fuel Properties and Test Cases PFI DI total total CH 4 [vol%] H 2 [vol%] CH 4 [vol%] H 2 [vol%] CH 4 [mass%] H 2 [mass%] CH 4 [energy%] H 2 [energy%] 100 0 0 0 100 0 100 0 92.5 0 0 7.5 99 1 98 2 85 15 0 0 98 2 95 5 80 0 0 20 97 3 93 7 75 25 0 0 96 4 91 9 91 0 9 0 100 0 100 0 same energy split PFI/DI The addition of H 2 to methane leads to only very minor reductions of the expected power output
Theory: Injection of Gases What do we expect from DI of H 2 and CH 4 in terms of penetration If the pressure-ratio is critical and above, an underexpanded jet is produced (1) For our case Direct injection species / amount Injection duration H 2 / 7.5 vol% (1 mass% or 2.4 energy%) 0.36 ms (72%) H 2 / 20 vol% (3 mass% or 7 energy%) 0.70 ms (140%) CH 4 / 9 vol% (9 mass% or 9 energy%) 0.50 ms (100%) (2) (3) Jet penetration (for same backpressure) (4) SS HH2 = pp 1/4 rrrrrrrr,hh2mm HH2 tt 1/2 HH2 SS CCCC4 pp rrrrrrrr,cccc4 MM CCCC4 tt CCCC4 =0.6 The H 2 jets penetrates considerably less then the CH 4 jet (1) Müller, F., Schmitt, M., Wright, Y., and Boulouchos, K., "Determination of Supersonic Inlet Boundaries for Gaseous Engines Based on Detailed RANS and LES Simulations," SAE 2013-24-0004 (2) Birch, A. D., Brown, D. R., Dodson, M. G., and Swaffield, F. The structure and concentration decay of high pressure jets of natural gas. Comb.Sc. and techn., 36(5-6), 249-261 (1984) (3) Bonelli, F., Viggiano, A., Magi, V., A Numerical Analysis of Hydrogen Underexpanded Jets Under Real Gas Assumption, Journal of Fluids Engineering, 135, 1-11 (2013) (4) Biffiger, H., Soltic, P., Effects of split port/direct injection of methane and hydrogen in a spark ignition engine, International Journal of Hydrogen Energy 40, 1994-2003 (2015)
Results: Combustion Stability For injection direction towards spark plug (CWR0) DI of H 2 led to better combustion stability when injected early, DI of CH 4 led to better combustion stability when injected late DI of H 2 does not show an advantage compared to premixed CH 4 /H 2 fuel Late DI of CH 4 allows much leaner global combustion than for pure PFI
Results: Early Flame Phase For the initial combustion phase (ignition to 5% fuel mass burned) Especially at lean conditions, the early phase of combustion is strongly accelerated by late DI toward the spark plug (CWR0) Very late injection (EOI 50 CA btdcf) is optimal for CH 4, but not for H 2 -> the directly injected H 2 seems not to reach the spark plug
Results: Early Flame Phase Ignition setting (for the example 7.5 vol% H 2 and CWR0) MBT ignition settings vary considerably for different injection settings
Results: Later Flame Phases The later combustion phases depend also on the injection settings, but not that pronounced as the early flame phases
Results: Efficiency Ignition always MBT, Variation of Lambda H2 addition to the fuel shows the best efficiencies, DI does not show an advantage Split injection of CH4 leads to the highest output, the efficiency behavior does not show an advantage But: all this was done on an engine, not optimized for DI, engine adaptions, accompanied by CFD simulations, would be necessary to unveil the real potential
Results: NOx Emissions Ignition always MBT, Variation of Lambda DI in direction of the spark plug leads to fuel-richer zones which increses NOx at globally lean conditions (But: all this was done on an engine, not optimized for DI)
Results: HC (Methane) emissions Ignition always MBT, Variation of Lambda DI in direction of the spark plug shows disadvantages regarding HC emissions (most probably due to flame quenching effects) But: all this was done on an engine, not optimized for DI
Conclusions Split PFI/DI injection gives new degrees of freedom for mixture formation of gaseous fuels Small amounts of DI are possible with comparably small injectors To exploit the full potential, the engine would have to be laid out accordingly CFD investigations would help to understand and optimize the processes involved (e.g. swirl/tumble flow) Acknowledgments Competence Center Energy and Mobility for financial support Empa Nanoscale Materials Science Laboratory and Oerlikon Balzers for coating of the DI injectors