Modeling of HCCI and PCCI Combustion Processes

Dan Flowers, Salvador Aceves

Lawrence Livermore National Laboratory

Robert Dibble
University of California, Berkeley

Randy Hessel
University of Wisconsin, Madison

Aristotelis Babajimopolous
University of Michigan

Work performed under the auspices of the U.S. Department of Energy by University of California, Lawrence Livermore National Laboratory under Contract W-7405-ENG-48.

Project Overview: HCCI Engines have potential for High Efficiency and Low Emissions Homogenous Charge Compression Ignition (HCCI) engines have many potential advantages:
Very Low NOx Emissions Very Low Particulate Emissions High Efficiency Lower Cost

We apply chemical kinetic modeling and experiments towards solving technical barriers of HCCI engines:
Combustion timing measurement and control Cylinder-by-cylinder combustion timing control Startup Fuel air ratio measurement and control Low Power Density Hydrocarbon and CO emissions

Approach: Fundamental and practical understanding of HCCI by judicious use of modeling and experiments Develop numerical tools that predict HCCI combustion with complex fuels in reasonable computational time (hours and days, not weeks and months)
O .O H2C H3C H3C C CH3 CH2 CH O Additive 1 H3C OH H2C H3C H2C O Additive 2 C CH3 HC O Additive 103 O O OH Additive 3 CH3 O. C H2C H3C H2C O HO Additive 5 H3C C H2C H3C H3C C CH3 C CH3 Additive 4 CH3 OH O Additive 204 O HO H2C H3C H3C H2C O C CH3 O. O CH2 H3C CH H3C O OH Additive 122 C CH2 O O. CH O O. HO O

High Resolution CFD

Fundamental Understanding + Design Guidance

H3C CH H3C

H2 C CH

H2 C

HO

Detailed Chemistry

Engine Experiments

Objective: Modeling and experiments applied to technical barriers towards practical HCCI engines Applying multizone detailed chemical kinetics models to understanding fundamentals of HCCI operation Extending and applying multizone modeling tools to handle non-homogeneous, but HCCI-like, combustion regimes: PCCI, SCCI Modeling and experiments to understand fundamental and practical issues in HCCI
HCCI with binary fuel mixtures using Carbon-14 tracing Studying chemi-ionization as a practical means for combustion timing measurement

We have long standing partnerships with industry and academia

Cummins
Currently involved in CRADA partnership Previous CRADA produced 2 joint papers

Caterpillar
Providing experimental support Close collaboration on natural gas HCCI

Sandia National Laboratories

Detailed analysis of experimental data, paper

Lund Institute of Technology

3 joint papers, collaboration on analysis

University of Wisconsin
joint work on KIVA analysis 6 joint papers

UC Berkeley
joint experimental and numerical work, >20 joint papers graduate students obtaining degrees on HCCI

U Michigan
PhD research conducted at LLNL

Our experiments and modeling advance both practical and fundamental understanding of HCCI and PCCI Characterized the role of turbulence on HCCI combustion through multizone modeling Completed development of Fully Integrated KIVA3Vmultizone (KIVA3V-MZ) model for simulation of HCCI and PCCI combustion Parallelized KIVA3V-multizone (KIVA3V-MZ-MPI) chemistry solver to significantly reduce computational time Developed rugged and low cost combustion timing sensor for HCCI combustion through ion sensing Investigated fuel and additive effects on HCCI combustion using C14 tracing of emission sources

What is the role of turbulence in HCCI engines?

Lund Institute studied HCCI operation in high and low turbulence engines These experiments showed that burn durations in higher turbulence configuration were significantly longer Debate raised about turbulence influence on the ignition process

Lower turbulence piston crown

Higher turbulence piston crown

Experiments show longer burn duration for higher turbulence geometry Our hypothesis: Chemistry timescales are far too rapid to be affected by local turbulence timescales, therefore turbulence plays only an indirect role through heat transfer

Our Multi-zone approach has proven accurate for HCCI combustion with wide range of fuels and engine geometries
KIVA3V is run until a specific Transition point Multi-zone HCT model used for simulating cycle after transition point

KIVA Temps mapped one-way onto multizone HCT

High resolution CFD with detailed chemistry at every grid point exceeds capability of todays computers
Possible Approaches:
Few grid points for CFD and Chemistry Many grid points for CFD, a few grid points for chemistry (Our approach) - works when chemistry and turbulence weakly interact ( e.g. HCCI) Engine Cycle Simulation Engine Cycle Simulation (single processor) (single processor) Axisymmetric CFD grid Axisymmetric CFD grid ~200 chemical species ~200 chemical species

1 Decade 1 Year 1 Month 1 Day 1 Hour

Kinetics CFD

We have analyzed HCCI operation both low and high turbulence geometries using the multizone approach

Flat crown (60K cells)

Square Bowl Crown (400K cells)

Kiva simulations show the flow pushed into the square bowl near TDC with high turbulence geometry

Very little change in swirl velocity occurs with flat-top piston geometry during compression

Flat piston

Square bowl piston

Substantial recirculation occurs in square-bowl piston, very little with flat-top

Flat piston

Square bowl piston

Higher turbulence yields more heat transfer and broader temperature distribution within the cylinder
0.045 0.04 mass fraction at 1K temperature bins 0.035 0.03 0.025 0.02 0.015 0.01 0.005 0 900

disc, peak heat release at 3 ATDC square, peak heat release at 3 ATDC

920

940

960 temperature, K

980

1000

Our multizone model shows that broader temperature distribution explains the lengthening of the burn duration

Heat release rate, flat piston

Heat release rate, square bowl

Conclusion: HCCI is dominated by chemical kinetics, turbulence plays an indirect role through heat transfer
12

10

solid line: experimental dotted line: numerical square disc

Burn duration, CAD

0 0 2 4 6 8 10 12

crank angle for peak heat release rate

Our Multi-zone approach has proven accurate for HCCI combustion with wide range of fuels and engine geometries KIVA is run until a specific transition point Multi-zone HCT model used for simulating cycle after transition point

KIVA Temps mapped one-way onto multizone HCT But great interest in partially stratified HCCI for better control and higher power output (PCCI, SCCI)

We extend this model to map between CFD and Multizone detailed kinetic solver throughout the engine cycle Fully Integrated KIVA and Multi-zone model (KIVA-MZ)
High resolution CFD solver handles mixing, advection and diffusion (>10K cells) Chemistry handled by multizone detailed kinetics solver (10-100 zones)

Solutions are mapped back and forth between solvers throughout the cycle

Significant reduction in computational time with CFD+multizone approach, but This approach still requires 50-100 chemistry zones Still computationally intensive, esp. with big mechanisms Multi-zone model is readily parallelized, parallel multizone chemistry solver now implemented within KIVA3V, yields near linear speedup with number of processors

Fully integrated KIVA3V-MZ-MPI handles chemistry part of simulation in parallel Fluid mechanics still handled with single processor Chemistry handled with additional processors, chemistry speedup scales almost linearly with number of processors (up to ~ number of zones)
Process 0 Processes 1 (N-1)

Fluid mechanics step

Zone1

Zone 2

Zone 3

Zone M

High resolution CFD with detailed chemistry at every grid point exceeds capability of todays computers
Possible Approaches:
Few grid points for CFD and Chemistry Many grid points for CFD, a few grid points for chemistry (Our approach) - works when chemistry and turbulence weakly interact ( e.g. HCCI) Engine Cycle Simulation Engine Cycle Simulation (single processor) (single processor) Axisymmetric CFD grid Axisymmetric CFD grid ~200 chemical species ~200 chemical species

1 Decade 1 Year 1 Month 1 Day Parallelization 1 Hour can shift kinetics cost dramatically lower

Kinetics CFD

Simulations were conducted to analyze Isooctane HCCI in the Sandia Engine (Experiments SAE 2003-01-0752)
5.88 mm (0.2314 in) 5.23 mm (0.206 in) 9.01 mm (0.355 in) 0.305 mm (0.0120 in)

102 mm (4.02 in)

Our fully integrated KIVA3V-MZ-MPI solver predicts pressure very well compared to Sandia experiments
90 80 Pressure (bar) 70 60 50 40 30 -10 -5 0 5 10 15 20 Crank Angle, Degrees
Phi=0.26 Experiment Phi=0.20 Experiment Phi=0.16 Experiment Phi=0.10 Experiment Phi=0.26 Simulation Phi=0.20 Simulation Phi=0.16 Simulation Phi=0.10 Simulation

KIVA3V-MZ-MPI emissions predictions are also in good agreement with the experiment
100 90 Fuel Carbon Into Emissions 80 70 60 50 40 30 20 10 0 0.1 0.12 0.14 0.16 0.18 phi 0.2 0.22 0.24 0.26
%Fuel C into CO [%] Experiment %Fuel C into CO2 [%] Experiment %Fuel C into HC [%] Experiment %Fuel into OHC [%] Experiment %Fuel C into HC [%] Simulation %Fuel into OHC [%] Simulation %Fuel C into CO [%] Simulation %Fuel C into CO2 [%] Simulation

Strength of the fully integrated KIVA3V-MZ-MPI solver is simulation of non-homogenous HCCI-like processes Using validated cases as a baseline, simulations conducted to investigate imposition of stratification on combustion
Comparison of equivalence ratio at 35 degrees BTDC (=0.26 Overall)

High Stratification Mild Stratification Uniform

HC and OHC lower with increased stratification, CO higher for some higher stratification conditions Consistent with experimental observations for early direct injection engines
8000 7000 6000 CO (ppm) 5000 4000 3000 2000 1000 0 0.1 0.12 0.14 0.16 0.18 0.2 Phi 0.22 0.24 0.26
600 400 0.1 0.12 0.14 0.16 0.18 0.2 Phi 0.22 0.24 0.26 1600 steep (HC)

steep shallow
HC or OHC (ppm)

1400 1200 1000 800

shallow (HC) uniform (HC) steep (OHC) shallow (OHC) uniform (OHC)

uniform

NOx emissions increase due to higher local equivalence ratio with increased stratification

14 12 10 NOx (ppm) 8 6 4 2 0 0.1 0.12 0.14 0.16 0.18 Phi 0.2 0.22 0.24 0.26 steep shallow uniform

KIVA3V-MZ-MPI simulation is currently being used to investigate PCCI with early direct injection Early Direct injection using KIVA spray model for diesel fuel 100 zone (with 100 processors) simulation with n-heptane diesel fuel surrogate

Summary: We continue to use our simulation capabilities to complement HCCI R&D in Industry, Nat Labs, Universities We now have the tools to computationally optimize the geometry, and fuel-air-EGR distributions in HCCI and HCCI-like engines KIVA3V-MZ-MPI is a computationally efficient and accurate tool coupling fluid mechanics and detailed chemical kinetics for prediction HCCI We have extended multizone tools to handle nonhomogeneous direct injected PCCI regimes Experiments, numerical modeling, and other unique capabilities applied to fundamental and practical issues of HCCI

Future Plans: We will continue to develop tools and analysis to guide HCCI engine design Continue applying KIVA3V-MZ-MPI to early direct liquid fuel injection PCCI Apply modeling to support optical engine PCCI experiments (Dec, Steeper) Investigate non-homogeneous engine operation towards higher load and lower HC and CO emissions Investigate effects of fuels and fuel mixtures on PCCI and HCCI operation Study practical strategies of combustion timing measurement and control for multi-cylinder engines

Recent and Upcoming Publications (1)

1) Analysis of the Effect of Geometry Generated Turbulence on HCCI Combustion by Multi-Zone Modeling SAE Paper 2005-01-2134. Collaboration: Lund Inst., U Wisconsin Highlights: Detailed study of Role of Turbulence in HCCI combustion 2) A Fully Integrated CFD and Multi-zone Model with Detailed Chemical Kinetics for the Simulation of PCCI Engines International Journal of Engine Research (in press). Collaboration: U Michigan, U Wisconsin Highlights: Development of approach for Fully Integrated Kiva-Multi-zone Detailed chemical kinetics code for simulation of HCCI and PCCI engines 3) Effect of the Di-Tertiary Butyl Peroxide (DTBP) additive on HCCI Combustion of Fuel Blends of Ethanol and Diethyl Ether SAE Paper 2005-01-2135. Collaboration: UC Berkeley, CAMS(LLNL) Highlights: Carbon 14 tracing investigating role of ignition enhancing additives on HCCI combustion with fuel mixtures.

Recent and Upcoming Publications (2)

4) Experimental and Numerical Investigation of Effect of Fuel on Ion Sensor Signal to Determine Combustion Timing in HCCI Engines, International Journal of Engine Research (in press). Collaboration: UC Berkeley Highlights: Developed ion formation model for HCCI engines for multiple fuels, verified model with experiments 5) EGR effect on Ion Signal in HCCI Engines, SAE 2005-01-2126. Collaboration: UC Berkeley Highlights: Experimental verification of ion formation chemical kinetics for HCCI with high EGR, for conditions ranging from very lean to slightly rich conditions 6) Combustion Timing in HCCI Engines Determined by Ion-Sensor: Experimental and Kinetic Modeling, 30th International Symposium on Combustion. Collaboration: UC Berkeley, U Heidelberg Highlights: First application and experimental verification of ion chemical kinetic mechanism applied to HCCI engines.

Recent and Upcoming Publications (3)

7) Investigation of HCCI Combustion of Diethyl Ether and Ethanol Mixtures Using Carbon 14 Tracing and Numerical Simulations, 30th International Symposium on Combustion. Collaboration: UC Berkeley, CAMS(LLNL) Highlights: Carbon 14 tracing used to investigate sources of emissions from dual component fuel blends in HCCI engines 8) Spatial Analysis of Emissions Sources for HCCI Combustion at Low Loads Using a Multi-Zone Model, SAE Paper 2004-01-1910. Collaboration: SNL, U Wisconsin, UC Berkeley Highlights: Detailed chemical kinetic analysis of geometry dependent emissions from an HCCI engine at low loads 9) Analysis of Homogeneous Charge Compression Ignition (HCCI) Engines for Cogeneration Applications, Proceedings of the ASME Advanced Energy Systems Division, 2004 Collaboration: Oregon State University Highlights: evaluation of the multiple advantages of HCCI for CHP

Recent and Upcoming Publications (4)

10) Thermal Management for 6-Cylinder HCCI Engine: Low Cost, High Efficiency, Ultra-Low NOx Power Generation, Proceedings of the ASME Internal Combustion Engine Division, May 2004 Collaboration: UC Berkeley Highlights: Characterization of HCCI combustion for stationary power 11) Improving Ethanol Life Cycle Energy Efficiency by Direct Combustion of Wet Ethanol in HCCI engines, Proceedings of the ASME Advanced Energy Systems Division Highlights: how HCCI combustion of wet ethanol can greatly improve energy balance for ethanol production 12) Source Apportionment of Particulate Matter from a Diesel PilotIgnited Natural Gas Fuelled Heavy Duty DI Engine, SAE Paper 2005-01-2034 Collaboration: University of British Columbia Highlights: C-14 analysis for identification of source of pollutants in a dual fuel engine

Other collaborative activities Ongoing participation at MOU meetings with vehicle and engine manufacturers Invited presentation at SAE HCCI symposium Invited presentation at ECI Lean Combustion Technology II workshop Several publications with National Laboratories and Universities in US and abroad:
SAE The Combustion Institute International Journal of Engine Research