Fundamental Combustion Characteristics of Gasoline Compression Ignition (GCI) Fuels S. Mani Sarathy, Clean Combustion Research Center, KAUST
Acknowledgments Curran et al. Farooq, Javed, Abbad, Chen, Selim, Ahmed, Naser, Singh, Bhavani Shankar, Mohamed, Atef, Manaa, Roberts, Chung Pitz, Mehl, Westbrook Oehlschlaeger et al. Dagaut Kukkadapu, Sung Sponsors Hansen
What is KAUST? Aleppo Founded in 2009 on the shores of the Red Sea Graduate study only research-based University International privately operated instituion in Saudi Arabia (~80 nationalities) 3
Engines and Fuels Easy emission control Lower efficiency High-octane gasolines (AKI* 90) SI CI Expensive emission control Higher efficiency Diesel fuel PPC/ GCI High efficiency and better emissions Low-octane gasolines (AKI 70) AKI = RON+MON 2 Slide credit: Tamour Javed/Bengt Johansson HCCI Bad control Can use almost any fuel 4
1,2,4-trimethylbenzene decalin 1-methylnaphthalene tetralin n-dodecylcyclohexane n-hexadecane n-dodecane 2,9-dimethyldecane 2-methylpentadecane 3-methyldodecane n-alkanes branched alkanes cycloalkanes aromatics 1. Molecular Level Fuel Characterization Fuels Light Gases Diesels Solid fuels Naphthas Lubricants Synthetic fuels Gasolines Heavy fuel oils Oxygenates 2. Surrogate Fuel Formulation Reproduces target properties of real fuel H/C ratio, functional groups, molecular weight, ignition 4. Experimental Testing 3. Chemical Kinetic Modeling 5. Predict Combustion Coupled kinetic/fluid models 6. Fuel/Engine Design 5
WTT & TTW emissions reductions Fuels for advanced gasoline engines Optimum fuel for GCI engines are in 60 85 octane range Low carbon emissions [SAE 2013-01-2701] Low fuel consumption (BSFC) [SAE 2012-01-0677] Lower regulated emissions [SAE 2014-01-2678] Additional benefits low aromatics: better H/C ratio, low engine-out soot Slide credit: Tamour Javed 6
Fuels for Advanced Combustion Engines FACE Gasolines Only sold in 55 gal barrels RON 70 to 97 Sensitivity 0 to 11 Aromatics 0 to 35% Collaborative research program led by KAUST with LLNL, UConn, RPI, UC Berkeley, CNRS... - Acquisition of 6 FACE fuels (A, C, F, G, I, J) - Compositional Analysis - Testing in ST, RCM, and JSR at different facilities - Formulation of suitable surrogates, modeling and validation - Kinetic analysis 7
Fundamental Data for GCI Fuels GCI fuels are tested at wide range of combustion conditions Ideal reactors Laminar Flames Engines Ignition Devices 8
Surrogate formulation methodology Optimization of palette species blend by matching target properties (Ahmed et al.) Ahmed et al., Fuel 143 (2015) 290-300. Target Properties H/C ratio Density RON & MON DCN Carbon type mole fraction (DHA, PIONA, NMR) Distillation curve iso-alkanes iso-pentane (2-methylbutane) 2-methylhexane Chemical Kinetic Models n-alkanes Aromatics Try it: cloudflame.kaust.edu.sa n-butane n-heptane Alkenes 1-hexene 1-hexene Cycloalkanes iso-octane (2,2,4-trimethylpentane) Toluene 124-trimethylbenzene 1,2,4-trimethylbenzene cyclopentane 9
Public Position The purpose of models is not to fit the data but to sharpen the questions. -Samuel Karlin 10
Private Position The purpose of models is to fit the data. 11
Summary of GCI Fuel Ignition Studies Ignition of GCI Fuels and Surrogates Shock tube and rapid compression machine ignition delay and species measurements Low-octane Fuels (Light naphtha AKI 64 and FACE I AKI 70) Mid-octane Fuels (FACE A and C AKI 84) High-octane Fuels (TPRF surrogates and wide range of high octane gasolines AKI 91) Understanding surrogate complexity requirements using targeted experiments and chemical kinetics analysis PRF, TPRF and multi-component surrogates 12
Light Naphtha Fuel Characterization Detailed hydrocarbon analysis (DHA) and octane testing (RON & MON) were done at Saudi Aramco R&DC Low octane (RON = 64.5, MON = 63.5), highly paraffinic (> 90% paraffinic content) fuel Light naphtha RON 64.5 MON 63.5 Sensitivity 1 H/C ratio 2.34 Avg. mol. wt. 78.4 n-alkanes 55.4 iso-alkanes 35.9 Cycloalkanes 6.7 Aromatics 1.32 Slide credit: Tamour Javed (Javed et al, PROCI 2016) 13
Multi-component Surrogate Formulation LN-KAUST surrogate composition Species mol% 2-methylbutane 0.25 2-methylhexane 0.1 n-pentane 0.43 n-heptane 0.12 Cyclopentane 0.1 Slide credit: Tamour Javed (Javed et al, PROCI 2016) 14
Comparison of Experimental Data with Surrogate Simulations f = 0.5 f = 1 f = 2 LN-KAUST and PRF 64.5 simulations are in good agreement with each other and with data at high temperature and NTC region At low temperatures, PRF 64.5 simulations are more reactive by a factor of two specially at f = 1 and 2 Slide credit: Tamour Javed (Javed et al, PROCI 2016) 15
Low Temperature Rich Conditions: Experiments and Simulations f = 2 Further targeted experiments reveal same trends at low temperatures LN-KAUST simulations and experiments are in good agreement with light naphtha data PRF 64.5 simulations and data are around a factor of two faster Multi-component surrogate (LN-KAUST) works better over a broad range of test conditions Slide credit: Tamour Javed (Javed et al, PROCI 2016) 16
FACE I Measurements Fuel / air, f = 1, P = 20 bar FACE I FG-I PRF 70 surrogate surrogate RON 70.3 70.7 70 MON 69.6 68.4 70 Sensitivity 0.7 2.3 0 Avg. mol. wt. 95.5 98.9 109.7 n-alkanes 14 12 33 iso-alkanes 70 72 67 Cycloalkanes 4 6 0 Aromatics 5 4 0 Olefins 7 6 0 FACE I exhibits full NTC behavior in 750 850 K range PRF 70 captures the reactivity of FACE I Slide credit: Tamour Javed (unpublished) PRF 70 marginally faster ( 25 %) at low temperatures 17
Low Temperature Octane Dependence Sensitivity (S) = RON MON Fuels with S < 7 exhibit weak octane dependence on ignition delay times Fuels with S > 7 exhibit octane dependence on ignition delay times 18
Slide credit: Nimal Naser (unpublished) Low Octane GCI Study Fuel Light naphtha PRF 65 FACE I gasoline PRF70 RON 64.18 65 70.3 70 S (=RON-MON) 0.61 0 0.7 0 Density (kg/m 3 ) 642 689 688 690 H/C 2.34 2.26 2.25 2.26 Description Specification Injector type Common rail piezo-injector Injector model Bosch (0445116030) Fuel inj. pressure 300 bar Injector holes 7 Nozzle hole diameter 0.18 mm Spray included angle 142 Fuel injector Piston Combustion chamber
Mass of fuel for constant CA50 Mass of PRF 65 to achieve constant CA50 of 4 CA atdc Slide credit: Nimal Naser (unpublished) CA50 of different fuels using same mass as PRF 65 at corresponding SOI
Equivalence ratio distribution for PRF 65 and LN PRF 65 LN Equivalence ratio distribution on the piston bowl surface at 1 CA atdc (above) side view of piston bowl (middle), T-f map colored with OH mass fraction with SOI at 19 CA btdc Slide credit: Nimal Naser (unpublished)
Slide credit: Bengt Johansson CFD on the fuel stratification with injection time
Predicting ignition quality from NMR spectra DCN of 71 pure compounds and 54 blends was collected/ measured using IQT. Dataset was used to study the relationship between CN/DCN and 8 structural parameters 1) Paraffinic CH 3 groups 2) Paraffinic CH 2 groups 3) Paraffinic CH groups 4) Olefinic CH-CH 2 groups 5) Naphthenic CH-CH 2 groups 6) Aromatic C-CH groups 7) Molecular weight 8) A new parameter called as Branching Index (BI) Slide credit: Abdul Jameel (Energy Fuels 2016)
NMR Based Model DCN= 21.71 + 0.2730 paraffinic CH 3 wt % +0.5645 paraffinic CH 2 wt % +0.2393 paraffinic CH wt % 1 H NMR spectra 0.0031 olefinic CH CH 2 wt % +0.3238 naphthenic CH CH 2 wt % +0.2481 aromatic C CH wt % +0.2484 Molecular weight 20.27 BI Try it: cloudflame.kaust.edu.sa Slide credit: Abdul Jameel (Energy Fuels 2016)
Predictive capability The model was validated with 22 real fuel mixtures (gasoline / diesel) and 59 blends of known composition. Slide credit: Abdul Jameel (Energy Fuels 2016)
Summary Both physical and chemical kinetic properties of GCI fuels control combustion performance Surrogates used for CFD simulations need to capture both physical and chemical kinetic features (depending on engine operating mode). Fuel design based on first principles of combustion chemistry is possible. 26
MAKE COMBUSTION GREAT AGAIN 27
ขอบค ณ 谢谢 Thank you Danke நன ற 감사합니다 شكرا Obrigado Merci Grazie ありがとう شکريا takk Dank u धन यव द ευχαριστώ Efharistó Спасибо Gracias Děkuju hvala Dziękuję köszönöm Tack Cảm ơn bạn terima kasih teşekkür ederim Kiitos متشکرم mani.sarathy@kaust.edu.sa http://cpc.kaust.edu.sa
Temperature (K) Ignition Delay Time (s) 3500 3000 2500 2000 1500 1000 1.E-01 1.E-02 1.E-03 Fuel design from chemical kinetics const. vol. simulations 20 atm, stoichiometric fuel/air mixtures 825 K MON-like 700 K RON-like RON=94, S=5.6 RON=97, S=11 1 1.1 1.2 1.3 1.4 1.5 1000/T (1/K) 20 atm, 700 K, phi=1 FGG-Temp FGF-Temp FGG-OH FGF-OH FGG-HO2/1000 1.E-01 1.E-03 1.E-05 1.E-07 1.E-09 1.E-11 1.E-13 500 1.E-15 0 0.005 0.01 0.015 0.02 0.025 Time (s) Mole Fraction Temperature (K) 3500 3000 2500 2000 Higher sensitivity fuel displays less NTC behavior; less reactive at RON-like and more reactive at MON-like. At RON-like conditions, fuel components that control OH radical pool are rate controlling At MON-like conditions, fuel components that drive OH and HO 2 radical coupling are important 20 atm, 825 K, phi=1 1.E-01 1.E-03 1.E-05 1.E-07 1.E-09 1500 FGG-Temp FGF-Temp 1.E-11 FGG-OH 1000 FGF-OH FGG-HO2/1000 1.E-13 FGF-HO2/1000 500 1.E-15 0 0.005 0.01 0.015 0.02 0.025 Time (s) Mole Fraction Modeling rationalizes nonlinear blending effects (source/sink interactions) Aromatic/alcohol and aromatic/naphthenic couplings Sarathy et al, Combust Flame 2016 29
RON, MON, and S correlations IDT = a * P ^ -N 30 25 9(8.2) TRF89.3(11.1) Ignition delay time (ms) 20 15 10 5 Engineering correlations can be made using simulated ignition delay times (79 fuels in training set) Reaction path analysis shows the effects of fuel composition (PIONA) on radical source/sink Pressure dependence of a ignition delay is correlated to sensitivity such that quantitative predictions can be made RON (S) MC90.9(-0.2) MC90.5(2.5) TRF89.1(3.5) MC90.9(8.2) TRF89.3(11.1) 0 10 20 30 40 50 Pressure (bar) Pressure Exponent (N) 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 850 K, 50 bar in Air Phi 1.0 0 1 2 3 4 5 6 7 8 9 10 11 12 Fuel Sensitivity (S) Singh, Badra, Mehl, Sarathy, Energy Fuels 2016