Reciprocating Internal Combustion Engines

Reciprocating Internal Combustion Engines Prof. Rolf D. Reitz, Engine Research Center, University of Wisconsin-Madison 212 Princeton-CEFRC Summer Program on Combustion Course Length: 9 hrs (Wed., Thur., Fri., June 27-29) Hour 8 Copyright 212 by Rolf D. Reitz. This material is not to be sold, reproduced or distributed without prior written permission of the owner, Rolf D. Reitz. 1 CEFRC8 June 29, 212

Short course outine: Engine fundamentals and performance metrics, computer modeling supported by in-depth understanding of fundamental engine processes and detailed experiments in engine design optimization. Day 1 (Engine fundamentals) Hour 1: IC Engine Review,, 1 and 3-D modeling Hour 2: Turbochargers, Engine Performance Metrics Hour 3: Chemical Kinetics, HCCI & SI Combustion Day 2 (Spray combustion modeling) Hour 4: Atomization, Drop Breakup/Coalescence Hour 5: Drop Drag/Wall Impinge/Vaporization Hour 6: Heat transfer, NOx and Soot Emissions Day 3 (Applications) Hour 7: Diesel combustion and SI knock modeling Hour 9: Automotive applications and the Future 2 CEFRC8 June 29, 212

Overview of Optimization Techniques Enumerative or exhaustive Calculus or gradient-based local methods which search in the neighborhood of current design point Random global methods such as genetic algorithms (GA) which typically converge on a global optimum Univariate (one-factor-at-a-time) Design of Experiments (DOE) Two-level factorial designs (main and interaction effects) Response surface methods (RSM) Statistical model building 3 CEFRC8 June 29, 212

Genetic Algorithms Individuals are generated through random selection and a population is produced A model is used to evaluate the fitness of each individual The fittest individuals are allowed to reproduce A new generation is formed - mutations are allowed through random changes The fitness criteria thins out the population and the most fit solution is achieved over successive generations Senecal & Reitz, SAE 2-1-189 4 CEFRC8 June 29, 212

Implementation of Algorithm Binary representation of parameters X - genes X 1 X 2 X 3 Goldberg, 1989 Carrol, 1996 Senecal, 2 11111 111 111 - chromosome gene string Precision X i,max X 2 1 i,min Evaluate merit f (X) for each generation member - identify fittest Binary tournament selection Bit-swapping Cross-over Parent 1 Parent 2 11111 111111 111 111111111 11111 111111111 111 111111 Descendants Bit-flipping Random mutation 5 CEFRC8 June 29, 212

Optimization Methodology Multi-Objective Genetic Algorithm Nonparametric Regression Technique 28 All Citizens Produced Pareto Citizens.3.3.28 24 2 196 NOx [g/kgf].25.2.26.24.22 GISFC [g/kw-hr] 192.1.2.3.4.5 NOx [g/kgf].6.7.2.4.6 1..8 Soot [g/kgf].15.5 1.5 swirl ratio 2.5 3.5 4.5.6.7.8 bowl diameter.9.2.18.16 Simultaneous optimization of many objectives [1] No merit function required to drive search Pareto front offers more information than a single optimum Coello Coello & Pulido, 21 Regression technique suitable for handling irregular and undesigned data sets (e.g., GA data) [2] Utilizes otherwise discarded optimization data Captures magnitude of effects AND the shape of their response Liu, IJER 26 6 CEFRC8 June 29, 212

Example Optimization - piston bowl design Parameters and Objectives Optimize: NOx Soot ISFC 7 Geometry Parameters: Pip height Bowl diameter of bowl bottom 4 curvature control points Injector Spray Angle Swirl Ratio Genzale, SAE 27-1-119. 7 CEFRC8 June 29, 212

Pareto Front Designs Bowl geometry or injection targeting trends? 3 28 All Citizens Pareto Citizens NOx Soot GISFC 68% 77% 15% swirl =.7 26 24 GISFC [g/kw-hr] NOx 57% Soot 6% GISFC % NOx Soot GISFC swirl = 1.4 45% 3% 2% swirl = 3.1 Genzale, SAE 27-1-119. 22 increasing swirl 2 18.1.2 NOx Soot GISFC.3 NOx [g/kgf] 5% 42% 6%.4 swirl = 3.1.5. 2. 1.6 1.2.8.4 Soot [/kgf] 8 CEFRC8 June 29, 212

Regression Used to Identify Dominant Design Parameters Regression fits performed for each design on the Pareto front 3 dominant design parameters identified: 1. Spray angle 2. Swirl ratio 3. Bowl diameter s oot [g/kgf] 2 1.5 1.5 soot [g/kgf].8.6.4 soot [g/kgf] 1.4 1.2 1.8 5 55 6 65 7 75 8 85.5 spray angle 1.5 2.5 3.5 swirl ratio 4.5.2.6.65.7.75.8.85 bowl dia meter (% bore).9 4.5 3.5 2.5 1.5 swirl ratio.5.6 5 55 6 65 7 75 8 85 spray angle.2.4.6 pip height (% bowl depth) Genzale, SAE 27-1-119. 9 CEFRC8 June 29, 212

Regression Used to Understand Parameter Effects.8 2 swirl = 3.1 NOx Soot GISFC 45% 3% 2% soot [g/k gf] 1.5 1.5 5 55 6 65 7 75 8 85.5 spray angle 1.5 2.5 3.5 swirl ratio 4.5 soot [g/kgf].6.4.2.6.65.7.75.8.85 bowl diameter (% bore).9 4.5 3.5 2.5 1.5 swirl ratio.5 Response Surface Observations: An optimal spray angle is predicted. Increased swirl ratio is predicted to enhance soot reduction near the optimal spray angle. Increases soot emissions at narrow spray angles. Increased swirl ratio is predicted to decrease soot at all bowl diameters. Genzale, SAE 27-1-119. 1 CEFRC8 June 29, 212

Optimization of Low Temperature Combustion - LTC Increased interest in new combustion regimes HCCI, PCCI, MK - offers simultaneous reduction of NOx and soot Challenges High CO, HC High loads Transients Klingbeil, SAE 23-1-341 NOx EGR Soot 11 CEFRC8 June 29, 212

Combustion Optimization - Fuel and EGR Selection HCCI simulations used to choose optimal EGR rate and PRF (isooctane/n-heptane) blend 6 bar 9 11 barimep IMEP 1 At 6, 9, and 11 bar IMEP 13 rev/min 8 18 g/kw-hr 1 bar/deg. 2 5.6 1 19 g/kw-hr bar/deg. bar/deg. 21 6 PRF PRF [-] As load is increased the minimum ISFC cannot be achieved with either neat diesel fuel of neat gasoline Predicted contours are in good agreement with HCCI experiments 16 bar/deg. 25 4 Net ISFC [g/kw-hr] MISFIRE MISFIRE 23 24 18 19 23 21 19 22 18 g/kw-hr 2 17 19 g/kw-hr 1 2 3 4 5 6 EGRRate Rate [%] EGR [%] Kokjohn, SAE 29-1-2647 12 CEFRC8 June 29, 212

Charge preparation optimization Premixed and Direct Injected fuel blending Desirable to use traditional diesel type injector Large nozzle hole (25 μm) Wide angle (145 included angle) KIVA + Multi-Objective Genetic Algorithm (MOGA) Fuel reactivity and EGR from HCCI investigation (9 bar IMEP) Global PRF = 65 EGR rate = 5% Five optimization parameters Minimize two objectives Wall film amount PRF Inhomogeneity Simulations run to 1 BTDC 21 generations with a population size of 24 Kokjohn, SAE 29-1-2647 Kokjohn & Reitz ICLASS 29 PRF Inhomogeneity [-].3.28.26.24.22.2.18 Inj. 1 Pressure Inj. 2 Pressure SOI 1 SOI 2 Fuel split NSD PRF ncells i 1 1 to 15 bar 1 to 15 bar IVC to (SOI2-2) ºATDC -5 to -3 ºATDC ~ 1 % Diesel Fuel m PRF PRF i i GLOBAL PRF ncells GLOBAL i 1 2 Results Film (%) All Solutions.45 PRF Inhomogeneity Pareto Solutions.19 1 2 3 4 5 6 7 8 Wall Film [% of Total Fuel] m i Parameters Inj. Pres. 1 (bar) 115 Inj. Pres. 2 (bar) 555 SOI1 ( ATDC) -67 SOI2 ( ATDC) -33 Fraction in first pulse.64 13 CEFRC8 June 29, 212

Optimized Reactivity Controlled Compression Ignition (RCCI) Port injected gasoline Optimized fuel blending in-cylinder Direct injected diesel Gasoline Injection Signal Squish Conditioning Ignition Source -8 to -5-45 to -3 Crank Angle (deg. ATDC) Diesel Kokjohn, SAE 29-1-2647 Gasoline Diesel 14 CEFRC8 June 29, 212

Heavy- and light-duty ERC experimental engines Engine Heavy Duty Light Duty HD LD Engine CAT SCOTE GM 1.9 L Displ. (L/cyl) 2.44.477 Bore (cm) 13.72 8.2 Stroke (cm) 16.51 9.4 Squish (cm).157.133 CR 16.1:1 15.2:1 Swirl ratio.7 2.2 IVC ( ATDC) -85 and -143-132 EVO( ATDC) 13 112 Injector type Common rail Nozzle holes 6 8 Hole size (µm) 25 128 Engine size scaling Staples, SAE 29-1-1124 15 CEFRC8 June 29, 212

Experimental validation - HD Caterpillar SCOTE 9 Speed (rpm) Effect of gasoline percentage 13 EGR (%) 43.5 Intake Temp. ( C) 32 Intake pressure (bar) Gasoline (% mass) Diesel inject press. (bar) 1.74 76 82-58 SOI2 ( ATDC) -37 Fract. diesel in 1 st pulse.62 IVC (ºBTDC)/Comp ratio 143/16 89 8 SOI1 ( ATDC) 14 82% 12 Pressure [MPa] Equivalence ratio (-) Experiment Simulation 14 1 8 12 89% 76% 1 Neat Diesel Fuel 89% Gasoline 6 8 6 Neat Gasoline 4 4 2 2-3 -2-1 1 2 3 Crank [ ATDC] Hanson, SAE 21-1-864 Computer modeling predictions confirmed Combustion timing and Pressure Rise Rate control with diesel/gasoline ratio Dual-fuel can be used to extend load limits of either pure diesel or gasoline 16 CEFRC8 June 29, 212 Apparent Heat Release Rate [J/ ] IMEP (bar)

RCCI high efficiency, low emissions, fuel flexibility Indicated efficiency of 58±1% achieved with E85/diesel Emissions met in-cylinder, without need for after-treatment Considerable fuel flexibility, including single fuel operation Diesel can be replaced with <.5% total cetane improver (2-EHN/DTBP) in gasoline - less additive than SCR DEF NOx [g/kw-hr] Soot [g/kw-hr] Gross Ind. Efficiency Heavy-duty RCCI (gas/gas+3.5% 2-EHN, 13 RPM) Heavy-duty RCCI (E-85/Diesel, 13 RPM) Heavy-duty RCCI (gas/diesel 13 RPM).3.2.1..3.2.1. 57 54 51 48 45 Splitter, SAE 21-1-2167; Hanson, SAE 211-1-361 HD Target (~21 Levels) HD Target (~21 Levels) 4 6 8 1 12 14 16 Gross IMEP [bar] 17 CEFRC8 June 29, 212

Dual fuel RCCI combustion controlled HCCI Heat release occurs in 3 stages (SAE 21-1-345, 212-1-375) Cool flame reactions result from diesel (n-heptane) injection First energy release occurs where both fuels are mixed Final energy release occurs where lower reactivity fuel is located Changing fuel ratios changes relative magnitudes of stages Fueling ratio provides next cycle CA5 transient control 2 95 Cool Flame PRF Burn Primarly n-heptane n-heptane + entrained iso-octane Iso-octane Burn Delivery Ratio [% iso-octane] AHRR [J/o] 15 RCCI Primarly iso-octane 1 5-2 -1 Crank [ ATDC] 1 2 o Kokjohn, IJER 211 18 9 85 CA5=2 ATDC 8 75 7 65 RCCI SOI = -5 ATDC 6 55 8 9 1 11 12 13 14 15 16 17 18 Intake Temperature [oc] CEFRC8 June 29, 212

Understanding RCCI Combustion Process Location B with dummy plug installed Optical Cylinder Head common rail injector fiber to FTIR Location A with optics installed Port Fuel Injector common rail fuel spray Splitter, SAE 21-1-345 19 CEFRC8 June 29, 212

Understanding RCCI Combustion Process 88 Pressure Pressure Press ure [MPa] Pressure [MPa] [MPa] Location B 4 4 Experiment Simulation 32 32 66 24 24 44 16 16 22 8 8-2 -2-1 -15-1 -5-5 55 1 1 15 15 Heat ]] Heat Release Release Release Rate Rate [J/ [J// [J [J/ 1 1 2 2 Crank [ ATDC] ATDC] -11-7 -3 16 3 deg deg degatdc ATDC ATDC Products Reactants Reactants Location A Experimental in-cylinder FTIR measurements of combustion process at two locations Spectra shows different fuel species at locations A and B, a result of the reactivity gradient Fuel decomposition and combustion products form at a slower rate at location B, extending combustion duration B A 23 27 31 31 35 35 39 39 Wavelength Wavelength (nm) (nm) Splitter, SAE 21-1-345 2 CEFRC8 June 29, 212

RCCI optical experiments Engine Bore x stroke Displacement Cummins N-14 13.97 x 15.24 cm 2.34 L Geometric compression ratio 1.75 RCCI experiments in Sandia heavy-duty optical engine LED illumination through side windows to visualize sprays Images recorded through both pistoncrown and upper window Crank-angle-resolved high-temperature chemiluminescence with high-speed CMOS camera Short-wave pass filter to reject longwavelength (green through IR) soot luminosity GDI Iso-octane 1 bar 7x15 micron Common-rail n-heptane 6 bar 8x14 micron Inc. Ang. 152 Kokjohn, SAE 212-1-375 21 CEFRC8 June 29, 212

RCCI combustion luminosity imaging Load: 4.2 bar IMEP Speed: 12 rpm Intake Temperature: 9 C Intake Pressure: 1.1 bar abs. Bowl window Kokjohn, ILASS-211 22 GDI SOI: -24 ATDC CR SOI: -57 /-37 ATDC Equivalence ratio:.42 Iso-octane mass %: 64 Squish (upper)june window CEFRC8 29, 212

Comparison of dual fuel RCCI and single fuel LTC Kokjohn, A&S 211 Engine specifications Three engines operating with different forms of LTC combustion Diesel LTC1 Ethanol PPCI2 Dual-Fuel RCCI 3 Cummins N14 Scania D12 CAT 341 Displacement (cm3) 234 1966 244 Stroke (mm) 152.4 154 165.1 Bore (mm) 139.7 127.5 137.2 Con. Rod (mm) 34.8 255 261 CR (-) 11.2 14.3:1 16.1.5 2.9.7 8 8 6 196 18 25 Case Engine Swirl Ratio (-) Number of nozzles Nozzle hole size (μm) 1. Singh, CNF 29 2. Manente, SAE 21-1-871 3. D. A. Splitter, THIESEL 21 23 CEFRC8 June 29, 212

Comparison with single fuel LTC Diesel LTC Single early injection at 22 BTDC 16 bar injection pressure Diluted intake (~6% EGR) Ethanol PPCI Single early injection at 6 BTDC 18 bar injection pressure No EGR Dual-fuel RCCI Port-fuel-injection of low reactivity fuel (gasoline or E85) Direct-injection of diesel fuel Split early injections (SOI1 = 58 BTDC and SOI2 = 37 BTDC) 8 bar injection pressure Liquid Fuel Vapor Fuel Liquid Fuel Vapor Fuel Liquid Fuel Vapor Fuel Kokjohn, A&S 211 24 CEFRC8 June 29, 212

Dual-fuel RCCI Kokjohn, A&S 211 allows higher load operation E85-diesel RCCI combustion has larger spread between most reactive (lowest RON) and least reactive (highest RON) 12 1 Pressure [MPa] Comparison of gasoline-diesel and E85diesel dual-fuel RCCI combustion For fixed combustion phasing, E85-diesel DF RCCI exhibits significantly reduced RoHR (and therefore peak PRR) compared to gasoline-diesel RCCI E85 and Diesel Fuel Gasoline and Diesel Fuel 8 6 E85 & Diesel - Experiment E85 & Diesel - Simulation Gasoline & Diesel - Experiment Gasoline & Diesel - Simulation 4 2 7.3.2.15 AHRR [J/deg] Mass Fraction [-].25 6 E85 & Diesel Gasoline & Diesel RON Distribution at -2 ATDC Gasoline & Diesel Fuel.1 4 3 2 1 E85 & Diesel Fuel.5. 5 85 9 95 1-2 15-15 -1-5 5 1 Crank [ ATDC] RON [-] 25 CEFRC8 June 29, 212 15 2

Comparison between diesel LTC, ethanol PPCI, and RCCI Evolution of key intermediates:.1 Reaction progress CH2O 1E-3 fuel CH 2O OH second stage 1E-4 OH combustion 1E-5 E85-diesel RCCI combustion shows a staged consumption of more reactive diesel fuel and less reactive E85 Ethanol and gasoline are not consumed until diesel fuel transitions to second stage ignition Mole Fraction [-] first stage combustion Diesel LTC Diesel.1 Ethanol PPCI C2H5OH 1E-3 OH 1E-4 CH2O 1E-5.1 C2H5OH 1E-3 ic8h18 1E-4 Diesel Dual-Fuel RCCI E85 & Diesel Fuel CH2O 1E-5-3 Kokjohn, A&S 211 26-2 OH -1 1 Time [ms ATDC] 2 CEFRC8 June 29, 212 3

Comparison between diesel LTC, ethanol PPCI, and RCCI 27 Kokjohn, A&S 211 Diesel LTC Ethanol PPCI Dual-Fuel RCCI (E85 & Diesel) 35 AHRR [% Fuel Energy/ms] Diesel LTC Earliest combustion phasing and most rapid energy release rate High reactivity of diesel fuel requires significant charge dilution to maintain appropriate combustion phasing (12.7% Inlet O2) Ethanol PPCI Low fuel reactivity and charge cooling results in delayed combustion Sequential combustion from leanhigh temperature regions to richcool regions results in extended combustion duration Dual fuel RCCI Combustion begins only slightly later than diesel LTC Combustion duration is broad due to spatial gradient in fuel reactivity Allows highest load operation due to gradual transition from first- to second-stage ignition 3 25 Diesel LTC 2 15 1 Dual-Fuel RCCI Ethanol PPCI 5-3 -2-1 1 2 Time [ms ATDC] RCCI Engine Experiments Hanson SAE 21-1-864 Kokjohn IJER 211 Kokjohn SAE 211-1-357 CEFRC8 June 29, 212 3

Single fuel RCCI RCCI is inherently fuel flexible and is promising to control PCI combustion. Can similar results be achieved with a single fuel and an additive? Splitter et al. (SAE 21-1-2167) demonstrated single fuel RCCI in a heavy-duty engine using gasoline + Ditertiary-Butyl Peroxide (DTBP) SAE 21-1-2167 2-Ethylhexyl Nitrate (EHN) is another common cetane improver 35 Estimated CN Contains fuel-bound NO and LTC results have shown increased engine-out NOx (Ickes et al. Energy and Fuels 29) 4 EPA 42-B-4-5 Extrapolated 3 25 Concentrations from SAE 21-1-2167 2 Kaddatz SAE 212-1-111 28 DTBP EHN 2 4 6 8 Additive Concentration [Vol %] CEFRC8 June 29, 212 1

Comparison of E1-EHN and Diesel Fuel Engine experiments performed on ERC GM 1.9L engine Diesel fuel and splash blended E1-3% EHN mixtures compared under conventional diesel conditions E1+EHN SOIc = -11.25 Pinj = 5 bar (5.5 bar IMEP, 19 rev/min) Diesel fuel injection parameters adjusted to reproduce combustion characteristics of E1+EHN blend Diesel Fuel SOIc = -7.9-11.25-9.25 Pinj = 9 5 bar Ignition Differences Diesel fuel SOI must be retarded to match ign. (Consistent with lower CN) Mixing Differences Diesel fuel injection pressure must be increased by 4 bar to reproduce premixed burn Kaddatz SAE 212-1-111 29 Diesel Fuel SOIc = -11.5 Pinj = 5 bar SOIc = -9.25 Pinj = 5 bar SOIc = -7.9 Pinj = 9 bar CEFRC8 June 29, 212

Comparison of E1-EHN and Diesel Fuel Diesel fuel and E1-EHN compared under conventional diesel conditions (5.5 bar IMEP, 19 rev/min) CDC operation with matched AHRR Diesel fuel injection parameters adjusted to reproduce combustion characteristics of E1+EHN blend For CDC operation, E1+EHN and diesel fuel show similar NOx and soot EPA 21 Kaddatz SAE 212-1-111 3 CEFRC8 June 29, 212

Diesel/Gasoline and E1+EHN RCCI PFI E1 and direct-injected E1+3% EHN compared to gasoline diesel RCCI operation Combustion characteristics of gasolinediesel RCCI reproduced with E1 E1+3%EHN Adjustment to PFI percentage required to account for differences in ignitability 12 Operating Conditions DI Fuel PFI Fuel Net IMEP (bar) Engine Speed (RPM) 24 Premixed Fuel (% mass) 2 Common Rail SOIc( ATDC) E1+EHN Diesel E1 Gasoline 5.5 19 84 69-42 SOIC (degatdc) Gasoline-diesel (84%) Pressure [bar] E1+EHN/E1 (69%) 8 16 6 12 4 8 2 4-3 -2-1 1 2 Heat Release Rate [J/deg] 1-32 to -52 Injection Pressure (bar) 8 5 Intake Temperature (C) 65 Boost Pressure (bar) Swirl Ratio EGR (%) 1.3 1.5 3 Kaddatz SAE 212-1-111 CA [degatdc] 31 CEFRC8 June 29, 212

Performance of E1 and E1+EHN RCCI Parametric studies performed to optimize efficiency of single-fuel RCCI at 5.5 and 9 bar IMEP E1/E1+EHN E1/Diesel Using a split-injection strategy, performance characteristics of single-fuel + additive RCCI are similar to those of dual-fuel RCCI Peak efficiency data for E1/E1+EHN shows higher NOx emissions, but levels meet EPA mandates Soot is very low for all cases Kaddatz SAE 212-1-111 32 CEFRC8 June 29, 212

Additive Consumption Estimate 12 Light-duty drive cycle average is 55% PFI fuel (i.e., 45% additized fuel) 3% additive level EHN volume is ~1.4% of the total fuel volume Similar to DEF levels Assuming 5 mpg and 1, mile oil change intervals, additive tank must be ~2.7 gallons 1 IMEP g [bar] SAE 21-1-151 8 6 4 2 4 2 5 Size shows relative weighting 3 1 1 15 2 Speed [rev/min] 25 3 Assumes 5 mpg and 1, mile oil change interval 33 CEFRC8 June 29, 212

Summary and Conclusions CFD modeling can be integrated with efficient optimization techniques for improved engine design New combustion strategies can be discovered using CFD-optimization Reactivity Controlled Compression Ignition strategy explained and validated with engine experiments Dual fuel and single-fuel (with additive) RCCI demonstrated to provide combustion control using optimized blending of port- and direct-injected fuels RCCI offers high thermal efficiency and meets EPA NOx and soot emissions mandates in-cylinder, without the need for after-treatment 34 CEFRC8 June 29, 212