Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
Daniel Flowers, Salvador Aceves, Joel Martinez, Francisco EspinosaLawrence Livermore National Laboratory
and Robert Dibble
University of California, Berkeley
DEER MeetingNewport, RI
August 26, 2003
Detailed Modeling of HCCI and PCCI combustion and Multi-cylinder HCCI Engine Control
We have developed and continue to improve the most advanced analysis tools for HCCI combustionThrough our collaborations we have applied these tools to a
wide variety of industrially significant questions:– What effect does turbulence have on HCCI combustion?
• Completed study of Lund TD100 engine for low turbulence geometry, now conducting simulations on high turbulence geometry
– What are the low load limits to HCCI operation?• Study of HCCI low-load operation limits in Sandia engine
– Does mixing affect CO emissions in HCCI combustion?• Extended multi-zone model includes mixing effects –
framework for general tool to analyze of HCCI, PCCI, SCCI• Analyzed post-combustion mixing and CO formation in
HCCI engine
Our detailed chemical kinetics codes and optimization software guide our experimental control efforts
• What control methods are viable for multi-cylinder HCCI engines?– Designed and tested a generic control system for internal
EGR control of cylinder by cylinder variation in 4-cylinder engine
– Implementing closed loop feedback control for cylinder-by-cylinder phase balancing
Outline
Through our collaborations we have applied these tools to a wide variety of industrially significant questions:– What effect does turbulence have on HCCI combustion?
• Completed study of Lund TD100 engine for low turbulence geometry, now conducting simulations on high turbulence geometry
– What are the low load limits to HCCI operation?• Study of HCCI low-load operation limits in Sandia engine
– Does mixing affect CO emissions in HCCI combustion?• Extended multi-zone model includes mixing effects – framework
for general tool to analyze of HCCI, PCCI, SCCI
• Analyzed post-combustion mixing and CO formation in HCCI engine
What is the role of turbulence in HCCI engines?
• Lund Institute studied HCCI operation in high and low turbulenceengines
• These experiments showed that burn durations in higher turbulence configuration were significantly longer
• Debate raised about turbulence influence on the ignition process• We conjecture that chemistry timescales are far too rapid to be
affected by local turbulence timescales
Lower turbulence piston crown
Higher turbulence piston crown
We have completed simulations with the flat-top (lower turbulence) geometry that agree very well with experiments
Flat piston (Disc) p=2 bar, Φ=0.40
0
10
20
30
40
50
60
70
80
90
100
-10 -5 0 5 10 15 20 25 30CAD
Pres
sure
, bar
T=469T=463T=457T=453T=451T=450P_dcT451P_Dc443P_DcT436P_Dc429P_DcT433P DcT431
Solid lines: numericalDotted lines: experimental
Simulations have been completed for the low turbulence case, high turbulence case is in progress
Flat crown (60K cells) Square Bowl Crown (400K cells)
Outline
Through our collaborations we have applied these tools to a wide variety of industrially significant questions:– What effect does turbulence have on HCCI combustion?
• Completed study of Lund TD100 engine for low turbulence geometry, now conducting simulations on high turbulence geometry
– What are the low load limits to HCCI operation?• Study of HCCI low-load operation limits in Sandia engine
– Does mixing affect CO emissions in HCCI combustion?• Extended multi-zone model includes mixing effects – framework
for general tool to analyze of HCCI, PCCI, SCCI
• Analyzed post-combustion mixing and CO formation in HCCI engine
Experiments in the Sandia HCCI engine show very inefficient combustion as equivalence ratio is reduced
Experimental isooctane HCCI data from the Sandia Engine
0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 0.24 0.26 0.28Equivalence ratio
0
10
20
30
40
50
60
70
80
90
100
Fuel
Car
bon
into
em
issi
ons,
(%)
CO2COHCOHC
Simulations agree well with experiment
0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 0.24 0.26 0.28Equivalence ratio
0
10
20
30
40
50
60
70
80
90
100
Fuel
Car
bon
into
em
issi
ons,
(%) Solid lines: experimental
Dashed lines: numerical
CO2COHCOHC
Our multi-zone model gives insight into the reduced combustion efficiency as equivalence ratio decreases
Complete Combustion
High CO
Fuel, IHC, CO
Unburned fuel
φ=0.26
φ=0.16
φ=0.10
φ=0.04
Our multi-zone model gives insight into the reduced combustion efficiency as equivalence ratio decreases
Complete Combustion
High CO
Fuel, IHC, CO
Unburned fuel
φ=0.26
φ=0.16
φ=0.10
φ=0.04
Our multi-zone model gives insight into the reduced combustion efficiency as equivalence ratio decreases
Complete Combustion
High CO
Fuel, IHC, CO
Unburned fuel
φ=0.26
φ=0.16
φ=0.10
φ=0.04
Our multi-zone model gives insight into the reduced combustion efficiency as equivalence ratio decreases
Complete Combustion
High CO
Fuel, IHC, CO
Unburned fuel
φ=0.26
φ=0.16
φ=0.10
φ=0.04
Outline
Through our collaborations we have applied these tools to a wide variety of industrially significant questions:– What effect does turbulence have on HCCI combustion?
• Completed study of Lund TD100 engine for low turbulence geometry, now conducting simulations on high turbulence geometry
– What are the low load limits to HCCI operation?• Study of HCCI low-load operation limits in Sandia engine
– Does mixing affect CO emissions in HCCI combustion?• Extended multi-zone model includes mixing effects – framework
for general tool to analyze of HCCI, PCCI, SCCI
• Analyzed post-combustion mixing and CO formation in HCCI engine
The original design of the multi-zone model is appropriate for modeling truly homogeneous HCCI combustion
Original Multi-zone Model: One-way mapping from Kivato Kinetic Solver
Enhancement to multi-zone model gives us the capability to handle stratified combustion (PCCI, SCCI)
Enhanced Model: Kinetic solver results are mapped back onto Kiva Grid
Kiva can now be used to account for mixing and heat transfer between zones for the whole cycle
The enhancements in the model allow us to move beyond “truly homogeneous” HCCI combustion
• Model is generally applicable to cases where heat release and mixing are weakly coupled
– PCCI = Premixed Charge Compression Ignition• High Internal EGR
– SCCI = Stratified Charge compression Ignition• Early Direct Injection
Accounting for mixing in multi-zone model yields much improved CO and HC emissions (SAE 2003-01-1821)
Experimental work
• What control methods are viable for multi-cylinder HCCI engines?– Designed and tested a generic control system for internal EGR
control of cylinder by cylinder variation in 4-cylinder engine– Implementing closed loop feedback control for cylinder-by-
cylinder phase balancing
Controlling cylinder-by-cylinder combustion timing is very important for multi-cylinder HCCI engines
20
25
30
35
40
45
50
55
-10 0 10 20 30Crank Angle (degrees)
Pres
sure
(bar
)Cylinder 1Cylinder 2Cylinder 3Cylinder 4
A generic, low cost surrogate to variable valve timing is being used for controlling cylinder to cylinder variations
Individual cylinder EGR Control by Exhaust Throttling
Surprising Exhaust Throttle Effects on start of combustion (SOC)
1. SOC delayed by backpressure of cylinder !
2. Constant fuel flow intake (non injected), and naturally aspirated air intake may explain SOC retardation.
3.0
4.0
5.0
6.0
7.0
-10.0 0.0 10.0 20.0
CAD
pres
sure
(10
bar
)
1 Throttled
Unthrottled
Cylinder 1 Response to Throttling
Surprising Exhaust Throttle Effects on SOC
Amazingly, throttling of other cylinders advances SOC of cylinder 1 !?!
3.0
4.0
5.0
6.0
7.0
-10.0 0.0 10.0 20.0
CAD
pres
sure
(10
bar
)
2 Throttled
1 Throttled
Unthrottled
Cylinder 1 Response to Throttling
Any Exhaust Throttle affects SOC in Cylinder 1
Backpressure of cylinders 3 and 4 also increases the SOC in cylinder 1
3.0
4.0
5.0
6.0
7.0
-10.0 0.0 10.0 20.0
CAD
pres
sure
(10
bar)
Cylinder 1 Response to Throttling
2 Throttled
1 Throttled
3 4
Unthrottled
Future analysis and experimental work will focus on cylinder interactions and how to control them