[go: up one dir, main page]

0% found this document useful (0 votes)
131 views26 pages

Internal Combustion Engine Simulations

Download as pdf or txt
Download as pdf or txt
Download as pdf or txt
You are on page 1/ 26

ANSYS Strategy for

Internal Combustion Engine Simulations

Ellen Meeks, ANSYS, Inc.


IC Engine Products at ANSYS
• Workbench – ICE System (WB-ICE)
ANSYS Workbench based ICE analysis
• FORTÉ
Full 3D CFD analysis for IC engine designers
• Model Fuels Library
Best-validated fuel models in the market (~60)
• CHEMKIN PRO
Ignition, flame-speed tables, zonal models
• CHEMKIN-PRO REACTION WORKBENCH
Fuel analysis and mechanism reduction
The ANSYS strategy for
IC engine simulations is: FORTÉ CFD
• We are investing in the ANSYS FORTÉ solution:
– Integrating RD Products into ANSYS Workbench and WB-ICE
– Integrating with FLUENT for conjugate heat transfer (CHT)
– Developing new capabilities
– Performing extensive verification and validation tests
– Continuing development of fuel models & fuel modeling tools

• FLUENT (with WB-ICE) remains a supported option for cold-flow


and scale-resolved turbulence simulation
A brief history of FORTÉ CFD
• Partnered with Profs. Rolf Reitz and Chris Rutland
– State of the art spray models
– Engine design expertise
• Reaction Design
– Combustion kinetics and kinetics solvers
• Model Fuels Consortium experience
– 7+ year, industry-led effort
– Comprehensive library of fuel models for
engine combustion simulation
• Now part of ANSYS, Inc.!
– Full suite of analysis tools and integration
Key advantages of FORTÉ CFD
• Set up tailored to IC Engines
• Automatic mesh generation
– On-the-fly, dynamic mesh motion
• Well validated fuel-combustion
kinetics with fast solver
– Capture fuel effects
– Predict ignition, emissions
– Predict flame-front location and
knocking in SI engines
• State-of-the-art spray and spray-
wall models
– Capture physics accurately and robustly
FORTÉ provides auto-generated,
Cartesian immersed-boundary mesh
• Dynamic mesh from CAD:
– CAD STL data read directly into FORTÉ
– WB-ICE can be used to provide geometry
• E.g., Fluent *.msh file read in => surface mesh
– FORTÉ generates the mesh on-the-fly
• With automated refinement around valve gaps
and small features
• Preview the mesh in the Simulation Interface

• Flexible mesh options:


– Fully automatic Cartesian meshing
– Body-fitted sector meshing for
quick diesel design studies
• Sector Mesh Generator included
• ICEM-CFD can also be used
FORTÉ integration with WB-ICE allows
quick geometry preparation from CAD
FORTÉ in ANSYS Workbench

Mesh
Preview in
FORTÉ
Interface

CFD-Post 3D with
FORTÉ results
Advanced solution techniques allow
practical turn-around with chemistry
700
CPU Time in Chemistry (Hours) Traditional CFD
600 FORTÉ CFD Accurate fuel models
usually have at least
500 this level of detail
400

300

200

100

< 10 hours 0
0 200 400 600 800
Size of Chemistry Model (Number of Species)

A good chemistry mechanism is key to predicting engine combustion behavior


Example: Dual-fuel heavy-duty engine
• RCCI Simulation with 470-species for Diesel & Gasoline
• CPU time (16 CPUs): < 4 hours

CO

CO and UHC agree with experiment


From: K. Puduppakkam, et al., SAE 2011 World Congress
Spark ignition and flame propagation
use state-of-the-art modeling
The spark ignites a kernel of gas The flame (typically <~0.01 mm thick)
that is very small at first propagates through the unburned gas

Flame
represented
by the G=0
Iso-surface

Particles are used to track the initial Afterwards, the level-set (G-eq.) method
kernel growth and flame location takes over in tracking the flame front
The flame propagation responds to local turbulence
conditions and chemistry is solved throughout
SAE 2003-01-0722, 2003-01-1843, 2006-01-0243, 2007-01-0165
FORTÉ spray model captures all the
physics with reduced mesh dependence
• Even with a coarse mesh,
the spray characteristics are
well represented
– Allows quick exploratory
studies as well as detailed
– Removes need to
iteratively test for mesh
“convergence”

From N. Abani, A. Munnannur, and R. D. Reitz, ILASS 2006


(Data from Naber and Siebers, SAE960034)
FORTÉ includes advanced spray-wall interactions

• Important to many engine simulations:


– Cold start of PFI engines
– Fuel/air mixing in DISI and GDI engines
– Soot from small-bore diesel and
GDI engines
• FORTÉ includes physics for:
– Droplet stick, rebound
– Droplet splash
– Film spread, separation
(stripping)
Work-in-progress for FORTÉ and WB-ICE

• Optional gradient-driven mesh refinement


• Multi-cylinder simulations
• Conjugate Heat Transfer coupling with Fluent
• Full automation & scripting through WB
– Parameterization of geometry inputs
– Optimization of geometry as well as other
parameters
Summary of key capabilities in FORTÉ
• Ease of Use
– Automatic mesh generation with moving boundaries
– Workflow specifically for engine combustion simulations
• Accuracy
– Detailed kinetics tied to Model Fuels Library to capture fuel effects
– Track flame front location using level-set method, DPIK for SI engines
– State-of-the-art spray and spray-wall models
– Capture correct physics on practical mesh
– Direct prediction of ignition, emissions, particle size
• Speed
– Fast and accurate chemistry solver
– Parallel (MPI) simulation using automatic domain decomposition
• Integration with other simulation
– Integration with WB-ICE, ANSYS Workbench for fluid-structural interactions
Applying FORTÉ to Design Simulations:
Importance of the Fuel Model
Changing engine-design requirements
shift the focus to the fuel model
Fuel model Injector nozzle-flow

Air flow with EGR


Spray/droplet break-up mixing
Wall-spray interaction

Mesh representing geometry


Fuel droplet vaporization of system

THE MODEL
Gas flow equations
• Computational mesh
Fuel-air combustion
chemistry model • Governing equations
• Physics models Heat transfer and wall
functions
• Chemistry models
Time-accurate numerical • Numerical methods Turbulence model
solution methods
Moving boundary
treatment
Model Fuels Library (MFL)
• The MFL enables detailed fuel modeling
– Represent complex “real” fuels with
appropriate model fuels
– Extensively validated chemistry
• Results from the MF Consortium:
– 7 years and $7 million invested
– Involved 20 companies
• The MFL is now continuously updated and
maintained through a subscription program
– Internal expertise adds & improves models
– Advised by world-wide experts
A model fuel should match real fuel properties
Set target characteristics
Desired Class composition
Heating value Model Fuel Library
Fuel
Octane / Cetane # aromatics
Properties H/C ratio, O content olefins
Boiling points c-paraffins

Sooting threshold index i-paraffins


n-paraffins

CHEMKIN Reaction Workbench


Select surrogate components
Optimize blend to match properties

n-heptane
Surrogate fuel
Iso-octane composition
19% 1-pentene Fuel
45% mc-hexane
15% m-xylene Model for
3% ethanol
1%
Simulation
15%
A good chemistry mechanism is key to
describing the fuel combustion behavior
The engine is a chemical plant:
Fuel
Chemical reactions
between molecules

Horsepower
Heat release
Air
(O2 + N2)
CO2 , H2O
CO, NOX ,PM
Unburned HC
Many other species
The Model Fuel Library is a large set of
validated fuel-component mechanisms
~ 60 Validated Fuel Models
Biodiesel Components • Common core
chemistry
Diesel and Jet-Fuel Surrogates • Consistent rules
for reaction
Gasoline Surrogates
rates
Gaseous Fuels
• Validated fuel-
blending
+ behavior
Detailed Particle-based and
Pseudo-gas Soot Models
With a good fuel model, we match emissions in
high-performance Diesel engines
• Detailed (7-component fuel model), 500 gas-
phase species, multi-step soot model
• Combustion and emissions trends captured
with no change to model or model approach

Joint publication with BMW:


Puduppakkam, et al.
SAE Powertrain Fuels and
Lubes, 2014
Detailed kinetics also enable
prediction of soot in GDI engines
7-component gasoline fuel model
SAE 2014-01-1135, Naik, et al.. 230 species mechanism
Full cycle simulations using RNG k-eps

Data from IFPEN,


Soot in Pires da Cruz et al.

laser plane
FORTÉ can predict knocking tendency
• Detailed chemistry predicts end-gas auto-ignition
• Pressure sensor monitors, just like the test stand
16 0.8

Band-pass filtered pressure (MPa)


Position 1
14 0.6
12 0.4

Pressure (MPa)
10
0.2
8
0.0
6
PPmax -0.2
4
2 -0.4
Pressure sensors in 0 -0.6
FORTÉ simulation match -20 -10 0 10 20 30 40
Crank Angle (oATDC)
experimental setup

• Allows exploration of trade-offs FORTÉ calculates Knock Intensity

– Fuel effects, spark timings, higher compression ratios, boost, etc.

SAE 2007-0Cal1-0165
Knocking can also be visualized in FORTÉ
• White/Black grid line intersections shows flame front location
• Contour plot shows temperature distribution in the engine
No knocking observed Significant knocking observed

Flame
Front

Auto-ignited end
gas ahead of the
flame front
● Intake Pressure = 0.9bar ● Intake Pressure = 1.2bar
● Spark Timing = -32CA ATDC ● Spark Timing = -32CA ATDC
24
Better fuel model = better predictions
• Multi-component fuel model captures knock in the
engine where a single-component fuel does not
FORTÉ Prediction of Knock Tendency - Surrogate
Comparison
20.0 1,400

Same engine 18.0 Pressure PRF ON=87


1,200
Pressure iso-octane ON=100
conditions 16.0
HRR PRF ON=87
14.0 HRR iso-octane ON=100 1,000
Pressure (MPa)

The only 12.0

HRR (J/deg)
800

difference is the 10.0


600
fuel model 8.0

6.0 400
4.0
200
2.0

0.0 0
660 680 700 720 740
Crank Angle (deg)

25
Overall Summary
• ANSYS is committed to advancing IC Engine Simulation
• FORTÉ is the core of the ANSYS IC engine strategy
– New capabilities and WB integration well underway
– Integration with Fluent for CHT
– Continuous improvement on performance and
scalability
– Validation and verification is an active and continuous
process
• The Model Fuels Library is key to accurate combustion,
and for soot and knocking predictions
• We are working closely with customers to fine-tune our
long-term plan and address critical short-term needs

You might also like