ILASS 2008

Sep. 8-10, 2008, Como Lake, Italy

Paper ID ILASS08-6-10
Paper ID ILASS08-A033

NOZZLE MOMENTUM EFFICIENCY DEFINITION, MEASUREMENT AND IMPORTANCE

FOR DIESEL COMBUSTION
Godfrey Greeves*, Gavin Dober*, Simon Tullis*, Nebojsa Milovanovic* and Stefan Zuelch*
*Delphi Diesel Systems, Gillingham, Kent, UK
[email protected]

ABSTRACT
An understanding of fuel injection related effects on conventional diffusion-controlled diesel combustion and soot emissions is reviewed. From
this it is concluded that at normal hot engine operating conditions the liquid phase length of the diesel fuel spray /jet is relatively short. The
review identifies the importance of the axial momentum input by the nozzle to the fuel spray/jet. This is a key factor determining the rate of
air/fuel mixing for control of soot formation in the fuel vapour jet. A new parameter named nozzle momentum efficiency (ηn) is defined. This is
essentially a measure of the efficiency of a nozzle in converting the fuel injection pressure into axial momentum in the fuel spray/jet.
Measurements of ηn from a newly developed test rig are presented. The ηn values for different spray-hole shapes show an increase for more
convergent spray-hole shapes and a correlation with measured HSDI engine exhaust smoke levels. The values of ηn are found to decrease at part
needle lift conditions explaining the soot emission benefits seen with the fast opening and closing characteristics of an advanced piezo direct-
acting injector. The new test rig is also used to measure an initial jet angle of the emerging fuel spray/jet and the measurements show that this
initial jet angle decreases with higher momentum efficiency such as produced with convergent spray-hole shapes. Also the initial jet angle
increases at part needle lift conditions. Both ηn and the initial jet angle are factors that affect the spray/jet penetration and rate of air/fuel mixing in
the combustion system. The advancement in understanding of these nozzle spray jet characteristics will contribute to future diesel fuel injection
nozzle design and manufacture.

INTRODUCTION the definition of a new nozzle parameter named the “nozzle

momentum efficiency”.
The continuing advancement of diesel engine combustion A test rig has been developed to measure the nozzle
and fuel injection systems is required to meet future more momentum efficiency to the desired accuracy and initial test
stringent vehicle emissions regulations for particulate and results obtained with this test rig are presented.
NOx as well as demands for lower fuel consumption and CO2.
There have been substantial advances in fuel injection IMPORTANT ASPECTS OF THE DIESEL FUEL
technology providing the capability for much higher fuel SPRAY/JET FOR CONTROL OF SOOT EMISSION
injection pressures that together with high levels of EGR have
allowed combined reductions in engine-out emissions of soot Soot formation and emission is very sensitive to the fuel
and NOx [1,2,3]. These advances include new injector injection and combustion system design and operating
concepts such as a two-actuator EUI system for heavy-duty parameters. It is important to advance the understanding of the
engines [1] and common rail injection systems [3] with fast key controlling processes. The diesel fuel spray atomisation,
acting solenoid injectors and the latest directed-acting piezo fuel evaporation, air/fuel mixing, ignition, combustion, heat
injector technology for light-duty engines. release and emission formation are a complex combination of
The diesel fuel injection nozzle has also been the subject of multiphase physical and chemical processes involving
a continuous design and development improvement process hundreds of chemical species.
that includes development of more flow-efficient nozzles. This One approach is to construct detailed 3-D CFD computer
paper is concerned with the diesel fuel injection nozzle and models and sub models that aim to simulate this complex
certain characteristics of diesel fuel sprays that are relevant to sequence of processes. Currently this approach cannot predict
the control of engine exhaust emissions. engine results to the required level of accuracy for all design
First a review is given of important aspects of diesel fuel and operating variables of interest. In particular the very dense
sprays in terms of what the engine combustion system and in liquid phase part of the diesel fuel spray/jet is not modelled
particular the soot emission responds to. This paper considers well by the usual CFD atomisation and evaporation models
conventional diffusion type diesel combustion with fuel and requires new approaches including the way the mass
injection timings near to top dead centre (TDC) crank angle exchange is done between the Lagrangian liquid phase and the
such as used with high injection pressures and high levels of Eulerian gas phase co-ordinate systems [4].
EGR for low soot and NOx emission. Also the context is light- An alternative/parallel approach used here is to start from
duty passenger car HSDI engines and heavy-duty truck DI measured results obtained from different types of experiment
engines that use multi-hole nozzles with a conventional and then by a technique somewhat like reverse engineering
petroleum-based automotive diesel fuel. deduce and model these important processes that control soot
formation and emission. In the end the measured engine
From the review it is concluded that the axial momentum in emission result is what matters to achieve the required future
the fuel spray/jet is an import characteristic and this leads to emission regulation. So the approach used here is to identify
1
1
 those key in-cylinder processes that explain the measured optical engine gives a good simulation of these liquid length
changes of the engine–out emissions with various changes of results and trends in the optical engine as in Figure 1.
fuel injection and engine design and operating variables. Then Also in an independent study liquid lengths have been
construct models of the important processes and devise test measured in a heated high pressure spray chamber for a wide
techniques to measure certain aspects of these processes. This range of conditions [7]. In this case a scaling law model [8]
provides a well-validated combined experimental/theoretical has been developed for the hot quiescent charge conditions in
understanding of the means to achieve the next reduction in the spray chamber and the results obtained are in agreement
emission level relative to a given baseline engine technology with the experimental results and trends in the spray chamber
and system. for quiescent charge conditions.
The following is based on an understanding of in-house Both of these liquid length measurements in an optical
experimental data from optical engine studies and from the engine and in a hot high-pressure spray chamber and the
responses of single-cylinder and multi-cylinder engines to a corresponding modelling results show that the maximum
large number of fuel injection and engine design and operating liquid length is a function of the spray hole diameter, gas
variables as well as computer modelling of diesel fuel sprays density, gas temperature and the thermal and vapour pressure
and combustion. properties of the fuel [5,8]. The observed maximum liquid
length is correctly modelled by an assumption of local thermal
Air/Fuel Mixing – Liquid phase equilibrium (liquid, vapour and entrained charge) and the local
concentration of fuel (liquid + vapour). There is no evidence
Figure 1 shows an evaporating fuel spray/jet image taken in of a dependence on droplet size since neither of these
an optical engine similar to that of [5] using a nitrogen charge simulations requires an input of droplet size to predict the
to suppress combustion to give a clear view of the details of experimental liquid lengths.
air/fuel mixing. This illustrates the nature of the air/fuel The initial angle of the liquid fuel jet as it emerges from the
mixing process for an evaporating diesel fuel spray/jet in hot spray hole is however important and this is an input to the
end of compression conditions for a swirling charge as in the SPRY2 model of the fuel spray/jet [6]. This initial jet angle
piston bowl of a HSDI engine. The coloured lines and marks can change with nozzle design parameters and affects the
in Figure 1 are an over-plot of calculated results from an in initial rate of air entrainment into the liquid phase of the fuel
house phenomenological computer model of the fuel spray/jet jet.
called SPRY2 [6] as shown in Figure 2 and will be referred to
later. JET EQUATIONS:
The dark region of the image of the fuel jet in Figure 1 is (1) Turbulent mixing CHAMBER
CHAMBER
CHAMBER
CROSS FLOWFLOW
(2) Axial and transverse CROSS FLOW
the only place where liquid phase fuel exists. This region is momentum (AIRSWIRL)
(AIR
(AIR SWIRL)
SWIRL)
V, ρ
ρ
V, ρρaaa Contra-rotating
Contra-rotating
too optically dense for optical techniques to penetrate and is (3) Liquid fuel Contra-rotating
Contra-rotating
vortices
vortices
vortices
evaporation vortices
likely to be a mixture of fuel ligaments, droplets, fuel vapour Y
JET
JET
JET
JET
RESULTS:
and entrained gas. The end of this dark image represents the (1) Jet trajectory
SECTION
SECTION
SECTION
SECTION
V
VVyyy
end of the liquid phase of the evaporating diesel fuel spray/jet. (2) Tip penetration v time
V y

(3) Liquid core length Vx

Measurements of this maximum liquid phase length of the fuel (4) Air/fuel mixing
b

V Um
Um
Um
DENSE
DENSE
DENSE
NOZZLE
NOZZLE y
LIQUID
LIQUID AXXXISIS
IS
IS VVxxx
SPRAY HOLE
SPRAY
SPRAY HOLE
HOLE A
ETTT A V
V
CORE
CORE
CORE JJJEE x
40 d,
d, Uooo,,, ρρρoooo
d, U
U
Vxxx
V
V

Fuel
Fuel Liquidcore
core Fuel
Fuel
Fuel
End of 30 Fuel Liquid
Liquid core
flow Vapour
Vapour
R10
liquid vapour jet
flow length
length
length
Vapour
Vapour
Jet
Jet
Jet
phase
20 Figure 2 Phenomenological model of fuel spray/jet (SPRY2)

10
An important conclusion from the optical engine
experiments and the validated phenomenological model [5,6]
0 is that the maximum liquid lengths are much shorter than the
-40 -30 -20 -10 0 10 20 30 40
distance to the piston bowl wall for the spray hole diameters
-10 and conditions that exist in automotive diesel DI engines of
interest (LD, MD, HD). This conclusion can also be made
-20 from the high-pressure chamber measurements and scaling
law model [7,8] when a spray hole size is used as for today’s
-30 modern diesel DI engines. Also in both these optical engine
and hot high-pressure spray chamber studies it was found that
-40
no fuel droplets escaped beyond the dense liquid phase region
referred to above.
Figure 1 Shadowgraph/Schlieren image of This does not mean that the atomisation characteristic of the
evaporating diesel fuel spray/jet taken in optical engine diesel fuel spray is unimportant. Instead it can be said that at
with superimposed SPRY2 spray model calculations, normal hot diesel engine conditions with near TDC injection
1000 bar rail pressure, 78 bar cylinder pressure and standard automotive diesel fuel the atomisation appears to
spray/jet show that it reaches a maximum quasi-steady length be more than adequate, liquid lengths are short and other
after the maximum lift of the nozzle needle is reached [5]. The factors become controlling.
SPRY2 model of the fuel spray/jet (Figure 2) applied to the Also from observations with diesel ignition and combustion
hot end of compression swirling charge conditions in the in an optical engine with a single well-defined fuel spray/jet
2
2
[5] it was concluded that it is too cold for soot to form in the well-validated representation of the airborne part of the diesel
liquid phase of the jet. In particular soot did not form in the fuel spray/jet as observed in the optical engine.
fuel jet until well after the end of the maximum liquid length. Figure 3 shows the SPRY2 model calculations for a HSDI

4200 rpm, 100% load 1000 rpm, 0% load

Fuel element LI PHI Fuel element LI PHI

- End of liquid 1 0 0.64 - End of liquid 1
30 - PHI = 1.2 2 0 0.88 - PHI = 1.2 2 0 0.81
3 0 0.99 3 0 1.21
25 4 0 1.11 4 0 1.42
5 0 1.26 5 0 1.48
20 6 0 1.43 6 0 1.27
7 0 1.56 7 0 0.81
15 8 0 1.38 8
10
PLAN VIEW ( mm)

5
0
-5
Initial jet angle
-10
0
Piston Piston

SIDE VIEW ( mm)

positions as positions as -5
fuel jet fuel jet
elements elements -10
impinge impinge
FSQ = 0.407 FSQ = 0.0 -15

CALCULATED FUEL SPRAY TRAJECTORY - Multiple 4 CALCULATED FUEL SPRAY TRAJECTORY - Multiple -20
degree elements 0.5 degree elements

40
-25
0 10 20 30 0 10 20 30 40
DISTANCE FROM CYL/INJ AXIS (mm)
Figure 3 Spray/jet model calculations for an HSDI engine, full load rated speed condition on LHS, engine idle
condition on RHS, blue dots are the calculated maximum liquid phase lengths for each of the injected fuel elements
engine. The two diagrams on the left-hand side (LHS) are for a
Air/fuel mixing – Vapour phase fuel jet and SPRY2 model full-load rated speed engine condition (4200 rpm). The upper
diagram is a plan view of one sector of the piston showing the
In Figure 1 the mottled image is the fuel vapour jet. This trajectory of one of the fuel sprays. The lower diagram is a
mottling is caused by the variations in fuel vapour corresponding side elevation view. The blue solid dots
concentration differences between adjacent turbulent eddies represent the maximum liquid phase length for each of the
which cause refractive index changes and light steering in the injected fuel elements and these liquid lengths are much
Schlieren/shadowgraph image. shorter than the distance to the piston bowl wall. The two
The SPRY2 model in Figure 2 has been developed in– diagrams on the right-hand side (RHS) show a similar
house over a period of years. It calculates the liquid and calculation for a hot engine idling condition that is an opposite
vapour phases of the fuel jet, the jet trajectory and air/fuel extreme of the engine speed/load range. At this condition the
mixing in a hot swirling air charge. In Figure 1 the lines/marks maximum liquid lengths are longer because of the lower end
calculated by the SPRY2 model are the upstream edge (yellow of compression temperature and lower charge density (low
line), the jet centreline (red), downstream edge (blue) and end boost pressure), but the liquid lengths are still much shorter
of the liquid phase (blue circle). These lines/marks show a than the distance to the piston bowl wall.
good general agreement with the observed image taken from
the optical engine. This comparison image is for one particular Air/fuel mixing – Rate of vapour phase air/fuel mixing and
set of conditions. A similar level of agreement between the soot formation
calculation model and the optical images is achieved over a
wide range of conditions including nozzles with different An earlier phenomenological fuel jet-based computer
spray hole diameters and different injection pressures, model of diesel combustion and a semi-empirical sub model of
injection timings, charge densities, charge temperatures, levels soot formation showed that it correctly calculated the main
of air swirl and fuel types with widely different fuel vapour trends of the measured exhaust soot emission from engines
pressure characteristics. This is achieved with no change of the over a range of engine and fuel injection variables [9]. This
model constants and means that the SPRY2 model gives a calculation model provided a useful concept of soot formation
3
3
 in diesel combustion and in explaining the observed soot assumptions are needed to derive the axial momentum input
emissions from engines. In particular high molecular weight by the nozzle. There is a need to obtain a more direct measure
gaseous soot precursors form rapidly in the gas phase fuel-rich of the axial momentum input by the nozzle to the combustion
fuel vapour regions of the fuel jet when these are at a system to advance understanding and take account of different
sufficiently high temperature [10]. As an injected fuel parcel detailed nozzle design effects.
travels along the fuel jet centreline more of the hot compressed
air charge is entrained into the jet. This added entrained charge NOZZLE MOMENTUM EFFICIENCY DEFINITION
causes an increase of the mixture temperatures promoting soot AND MEASEMENT TECHNIQUE
formation but eventually further along the jet there is
sufficient added oxygen to prevent soot formation through A comparison of interest here is the effect of different
oxidation of the gaseous soot precursors. This means that there nozzle internal flow details and in particular the effect of
is a certain range of fuel-rich air/fuel ratios (e.g. φ 1.4 – 1.2) different spray hole shapes for the same injection pressure and
along the fuel jet where soot forms most rapidly. the same fuel mass flow. It is found that the engine-out soot
Consequently the amount of soot formed from an injected emission can be significantly reduced with a nozzle having
fuel parcel is very dependent on the residence time of when it convergent spray holes compared with a nozzle having parallel
is in a fuel-rich high-temperature condition and hence on its spray holes. From the review and the SPRY2 model above it is
rate of air/fuel mixing as it passes through the combustion concluded that these smoke differences should be related to
system. This explains the very strong effect that an increase of the axial momentum input from the spray hole.
fuel injection pressure has on reducing soot emission since it
increases the axial fuel jet velocity and reduces the residence Nozzle momentum efficiency definition
time of the gas phase fuel-rich mixtures [1].
The SPRY2 model has been used to calculate the rate of An efficient nozzle that minimises soot formation should be
air/fuel mixing corresponding to the measured engine exhaust one which maximises the axial momentum from the spray hole
soot emissions as shown in Figure 4 [1]. The x-axis values are for a given fuel injection pressure and fuel mass flow. In order
derived from the rate of change of local air/fuel ratio as each to capture this property of a nozzle a new parameter named
injected fuel parcel passes through a fuel-rich air/fuel ratio of nozzle momentum efficiency is defined here below.
9 (RAF9) in the fuel jet. This calculated rate of air/fuel mixing The nozzle discharge coefficient (Cd) for an individual
shows a good correlation with the measured exhaust soot for a spray hole can be defined as follows.
range of different levels of injection pressure and nozzles with
different spray hole diameters. Cd = (actual spray-hole mass flow) / (theoretical spray-hole
Consequently an understanding of the basis of the SPRY2 mass flow)
model and its input/result helps to define the important
features for fast fuel/air mixing and low soot emission. In & / m& th
Cd = m
particular it is the fuel jet entrainment process that drives the = m& / ( ρf . Ah . Uth ) (1)
rate of air/fuel mixing. The axial fuel jet momentum that is
input by the nozzle spray hole is important because this drives Uth = √ [ 2 (P1-P2) / ρf ] (2)
the turbulent fuel jet entrainment process.
For a momentum efficiency value to be useful in comparing
For the SPRY2 model calculations, the fuel mass flow and
engine results there is a need to hold the fuel mass flow
the axial momentum input from the spray hole are derived constant at a given fuel injection pressure. Otherwise for a
1.6 PEAK INJECTION PRESSURE 1600
given total fuel quantity delivered to the combustion system
the injection period will change which means the smoke
1.4 1400 comparison is no longer valid because of the other in-cylinder
effects that changes of injection period will have on soot
1.2 1200
formation. Thus both the theoretical momentum flux and the
Peak Injection Pressure (bar)

1.0 1000
actual momentum flux need to be based on the actual or
measured fuel mass flow.
Smoke (FSN)

8x0.18 Timing swing NOP = 325 bar

8x0.18 NOP swing TDC
0.8
8x0.20 Timing swing NOP = 325 bar
800 The new definition for the nozzle spray-hole momentum
8x0.20 NOP swing TDC
8x0.20 Increasing NOP, constant EOI
Correlation of measured smoke results
efficiency (ηn) is given here as follows.
0.6 600
with calculated rate of air/fuel mixing
SMOKE
0.4 400 ηn = (actual spray-hole momentum flux) / (theoretical spray-
hole momentum flux based on the actual fuel mass flow)
0.2 200

0.0 0
ηn = M& / M& tham (3)
11 12 13 14 15 16 17 18 19
Calculated Rate of Air/Fuel Mixing (RAF9)

& . UeffM / ( m& . Uth) = UeffM / Uth

ηn = m (4)
Figure 4 Correlation of HD engine (with EUI) exhaust smoke
with rate of air/fuel mixing calculated by SPRY2 model [1] ηn = M& / ( m
& . √ [ 2 (P1-P2) / ρf ] ) (5)

from various inputs. These include the nozzle flow measured The definition for ηn given by equation (5) compares the
in a test rig and/or the nozzle discharge coefficient (Cd) and a actual momentum flux to a theoretical momentum flux based
Bernoulli velocity calculated from the injection pressure on the actual mass flow m & for a spray hole.
difference across the nozzle. With the use of these data certain The effects that lead to a nozzle spray-hole discharge
coefficient Cd, as defined by equations (1) and (2), of less than
4
4
1.0 include a vena-contracta (this effect restricts the use of the from this a relatively small change in the nozzle momentum
full geometric area of the spray hole) and flow losses that efficiency can have a large effect on the soot emission.
reduce the effective exit velocity such as from secondary Consequently for momentum efficiency measurements to
flows. Further effects are a non-square velocity profile at the be useful there is a need to measure ηn to a target accuracy of
minimum spray-hole area and cavitation in the spray hole if it 1%.
exists at the minimum area since this reduces the effective fuel
density. Also a choked condition means that the mass flow is Measurement of nozzle momentum efficiency
no longer a direct function of (P1 - P2).
Some of these effects will also play a part in reducing the Axial spray momentum can be measured by directing the
momentum efficiency, as defined in equations (3-5), below fuel spray on to a fixed plate that is arranged to measure the
100%. This will mean that nozzle design changes that increase impact force on the plate necessary to destroy all of the axial
Cd may also tend to improve ηn. Nevertheless by considering momentum in the fuel spray/jet [11]. In this type of technique
theoretical nozzles it can be argued that the Cd value for a the sprays from a dynamically operating injector are aimed
nozzle cannot be assumed to directly represent the momentum perpendicularly towards a plate with a force transducer. The
efficiency of a nozzle. For example, if a square axial velocity dynamically varying electrical output signal from the force
profile equal to the Bernoulli velocity existed in a central transducer with each injection event can be used to give a
region of the spray hole with zero velocity outside this region spray force versus time diagram for each of the fuel sprays.
then the momentum efficiency would be 100%, while the This type of technique is useful to indicate differences in the
discharge coefficient would be less than unity. In general if dynamic characteristics of different injector designs. For
both Cd and ηn are expressed as ratios then the values of these example results in [12] show the benefit that the advanced
parameters should be related as follows. piezo directing-acting common rail injector gives over the
solenoid injector in terms of a much more square spray
Cd < ηn < = 1.0 (6) momentum diagram.
Dynamic spray momentum rigs measure the varying spray
In essence it can be said that Cd is particularly important for force during a single injection event with the needle opening
the fuel metering function of the nozzle while ηn is important and closing. Because of the relatively small spray forces
for the nozzle function that gives efficient conversion of fuel involved that must be measured in the presence of rig
vibrations and unevenness of the turbulent structure in the
spray itself, it is difficult to get a high level of absolute
Nozzle P1 Measurement accuracy. Also to derive momentum efficiency it is necessary
chamber to measure spray force and mass flow rate simultaneously at
Spray hole different stages of the nozzle needle lift. This is very difficult
axis to do with a dynamic injection event with the uncertainty of
the needle lift value at any instant in time. Consequently the
m& , M& P2 absolute value of ηn for each spray hole cannot be measured to
the required accuracy of the order of 1% as specified above.
Fp
A first step towards this objective is to measure at steady-
P2
state conditions for preset nozzle needle lifts since it is the
Aperture efficiency of the nozzle flow paths that are of interest here
rather than the dynamic operation of the injector as a whole.
Figure 5 shows a schematic of the new in house developed
Force plate steady-flow momentum rig that is the subject of a patent
application [13]. This spray momentum rig is designed to
measure spray force, fuel mass flow, momentum efficiency
and initial jet angles.
In this test rig a fuel spray/jet emerges from a nozzle spray
hole with a steady upstream fuel pressure P1, a steady
downstream ambient pressure P2 and for a fixed position of
the lift of the nozzle needle within the nozzle body. The force
measurement plate is arranged to be perpendicular to the
Figure 5 Schematic of new steady-state flow rig to
nozzle spray hole and fuel jet axis. The fuel spray to be
measure spray force, mass flow and momentum
measured enters through the centre of the aperture in a
efficiency
measurement chamber cover and impacts normally on the
spray target. This target is fitted with a high accuracy steady-
injection pressure into axial momentum in the fuel jet. state force sensor to measure the force Fp.
A nozzle design with non-tapered cylindrical spray holes By considering a momentum control box it can be shown
with a sharp spray-hole entry and exit would have a nozzle that the actual axial momentum flux M& , leaving the nozzle
flow discharge Cd value in the region of ~ 0.7 and a nozzle spray hole, is equal to the difference in measured force (Fp) on
spray hole momentum efficiency ηn of ~ 0.8. An advanced the target plate between with and without the fuel spray/jet.
nozzle design with convergent spray holes would have a Cd This gives the following:
value in the region of ~ 0.8 and an ηn value of ~ 0.9 giving
about 50% reduction of the engine exhaust soot emission. So
M& = m
& . UeffM = Fp (7)

5
5
 spray hole
P1 – Upstream fuel pressure
- Mom entum Efficiency P2 – Downstream pressure for spray hole
1.00
ρf – The fuel density at entry to the spray hole
- Cd

It should be noted that this definition of nozzle spray hole

Momentum Efficiency and Cd - (ratios)

0.90
momentum efficiency is particularly convenient because,
unlike the nozzle discharge coefficient Cd as defined in
0.80 equations (1-2), it does not require the minimum geometric
spray hole area Ah or diameter to be known or measured. The
0.70 latter can be difficult to do with certain shapes of spray hole, if
the spray hole surface is lacquered from running on an engine
0.60
or at part needle lift conditions.
This new steady-state flow rig (Figure 5) measures the
spray force and in addition has a measurement chamber with
0.50
an aperture. For a multi-hole nozzle this separates and collects
the spray that is being measured to give simultaneous spray
0.40 force and spray mass flow measurements to derive the
momentum efficiency as equation (8). Some special features
0.30 of the test rig give inherent damping characteristics that
0 0.1 0.2 0.3 0.4 remove rapid time-varying unevenness in the spray itself to
give a consistent and repeatable measurement of the total
Needle Lift (m m )
Figure 6 Momentum efficiency and Cd versus needle lift spray force. The measurement chamber can be inclined in both
for hole 1 of nozzle 8921-3 horizontal and vertical planes as indicated in Figure 5 relative
to the spray hole axis. For momentum efficiency
Fp is the required opposing axial force on the force measurements the measurement chamber axis is aligned with
measurement plate to bring the axial momentum of the fuel jet the spray hole axis.
flowing along the fuel spray/jet path to zero and UeffM is an
actual effective velocity relevant to the momentum flux. The Measurement of initial jet angle
momentum efficiency as defined by equation (5) above
combined with equation (7) gives: This test rig can also be used to measure an initial angle of
the fuel spray/jet as it emerges from the spray hole. This is
ηn = & ) / √(2(P1-P2)/ρf)
(Fp/ m (8) done by measurement of the decline in spray force as the
measurement chamber is deviated away from the axis of the
The equation (8) can be used to calculate the nozzle spray-hole fuel/spray jet. At first the force will drop very slowly with the
momentum efficiency from measurements of the following. cosine of the angle between the measurement chamber and the
spray/jet, but then much more quickly as the aperture of the
Fp – Force measured on a plate to destroy all of the axial measurement chamber cuts off the entry of the spray/jet cross
spray/jet momentum section. The angle over which the force declines from 90% to
m& – Mass flow rate of fuel measured to emerge from the 10% of its value is a measure of the spray/jet angle.

Nozzle Momentum Efficiency = (measured spray momentum) /

(theoretical spray momentum for measured mass flow)

92
Nozzle Momentum Efficiency

Sac A - different
90 spray hole
shapes
88 Initial jet angle
effect?
Sac B - different
(%)

86 sac shape

84
Sac A - 0.6,
82 0.8,1.0 hole
length
80
0 0.5 1 1.5 2
Smoke (FSN)

Figure 7 Correlation of measured engine smoke with measurements of nozzle momentum efficiency
for different nozzle designs – single-cylinder HSDI engine, 2800 rpm, 10.8 bar BMEP

6
6
RESULTS FOR NOZZLE MOMENTUM EFFICIENCY be indistinct. As an alternative the new test rig as in Figure 5
has been used to measure the initial jet angle results given
Figure 6 shows some sample momentum efficiency and Cd below. This was done by variation of the angle of the
values measured with this new test rig. These measurements measurement chamber such that the aperture gives a
were taken for an upstream injection pressure of P1 = 100 bar progressive cut off of the jet section and spray force as
and atmospheric back pressure for a range of needle lift described above. It should be made clear that this initial jet
conditions. The momentum efficiency values decrease with a angle is measured close to and within 5 mm of the spray hole
reduction of the nozzle needle lift and the momentum exit (as illustrated in Figure 3). The intention is to capture the
efficiency values are greater than the Cd values as stated in output from the spray hole rather than the wider jet angles
equation (6). At higher engine load conditions the nozzle seen further downstream (Figure 3) as these latter are more to
needle lifts to a full lift condition and the greater part of the do with the gas phase jet angles and are a function of the
fuel is injected at full needle lift. Of particular interest is the entrained gas charge density etc.
extent to which momentum efficiency values at full lift Figure 8 shows some results obtained for the initial jet
explain changes of exhaust smoke levels for different nozzle angles of two different nozzles. These results show that the
design details. initial jet angle increases as the needle lift decreases. This
Figure 7 shows the exhaust smoke values measured on a means that at low needle lift the wider initial jet angle will
single-cylinder HSDI engine plotted versus the measured cause the liquid jet velocity to initially decay more rapidly
momentum efficiency values at a full needle lift condition. along the axis of the fuel spray/jet. This combined with the
The engine test results were taken for the same injection lower momentum efficiency at low needle lift (Figure 6)
pressure and for nozzles with the same flow, so that the means that the part of the fuel spray injected at low needle lift
differences in smoke should be the effect of the different spray will be much less penetrating than at full needle lift. A low
hole shapes and sac shapes. These results show there is a penetrating part of the fuel spray is deflected rapidly by the air
significant reduction of the engine exhaust smoke for nozzle swirl in the combustion chamber. This leads to interference
designs with higher momentum efficiency (as expected from with the adjacent fuel sprays leading to fuel-rich areas and an
Figure 4). additional source of soot formation.
The nozzle designs with different spray-hole shapes include These effects explain one of the benefits of the direct acting
the effect of spray holes with parallel and convergent tapers piezo injector [12]. This injector has very fast needle opening
and entrance rounding. Also shown is the effect of a different and closing characteristics. This minimises the quantity of fuel
sac design and different spray hole lengths. A separate trend that is injected at low needle lifts (with a wider initial jet
line is indicated for the results with different spray hole angle) and with low momentum efficiency, to give significant
lengths. A possible explanation for this is that the different reductions in emissions.
spray hole lengths can change the initial fuel spray/jet angle as
it emerges from the spray hole. This angle is an additional 6
Spray/Jet Angle (degrees)

spray hole input characteristic to that of momentum efficiency

since it has a separate effect on the fuel jet and air/fuel mixing 5

4
7785 Nozzle 5464 Nozzle
3
10
2
9
1
Spray/Jet Angle (degrees)

8
7 0
6 82 84 86 88 90 92
5
Momentum Efficiency (%)
4
Figure 9 Measurements of initial spray/jet angle versus
3
measured momentum efficiency at full needle lift for
2 five different nozzle designs
1
Measurements of the initial jet angle at full lift are shown in
0
Figure 9 for five different designs of nozzle. In particular
0.0 0.1 0.2 0.3 0.4 spray-hole shapes that give higher momentum efficiency, as
Needle Lift (mm ) described in Figure 7, also tend to have smaller initial jet
angles. This means that the lower engine exhaust smoke
Figure 8 Initial spray/jet angle versus needle lift values observed with higher momentum efficiency spray-hole
process. For example, in the SPRY2 model mentioned earlier, shapes (Figure 7) may be influenced also by the narrower jet
the initial jet angle is an independent input to that of the input angles.
spray momentum by the spray hole.
CONCLUSIONS
RESULTS FOR INITIAL JET ANGLE
From a review of optical engine experiments and a
phenomenological computer model of the diesel fuel spray jet,
It is difficult to get an accurate measure of the initial jet in a swirling air charge, it is concluded that the liquid phase
angle by optical techniques since the edge of the spray/jet can
7
7
part of the fuel spray jet is very dense and relatively short at (A/F ratios)/ms
normal hot engine operating conditions. Soot forms further UeffM Effective velocity for spray momentum m/s
downstream in hot fuel-rich regions of the fuel vapour jet and Uth Theoretical or Bernoulli velocity m/s
is reduced by faster rates of air/fuel mixing. The axial fuel jet ηn Momentum efficiency ( M& / M& tham ) ratio or %
momentum input by the nozzle spray hole drives the fuel jet ρf Liquid fuel density kg/m3
entrainment process and the speed of air/fuel mixing.
A new parameter named nozzle momentum efficiency is REFERENCES
defined. This is essentially a measure of the efficiency of a
nozzle in converting the fuel injection pressure into axial [1] G. Greeves, S. Tullis and B. Barker, Advanced Two-
momentum in the fuel spray/jet. In order for this parameter to Actuator EUI and Emission Reduction for Heavy-Duty
be useful to evaluate the effect of different nozzle design Diesel Engines, SAE 2003-01-0698.
details a new technique was needed to measure momentum [2] S. Tullis and G. Greeves, HSDI Emission Reduction
efficiency to an accuracy of ~1%. Results with a new in-house with Common Rail FIE, Seminar on Fuel Injection
technique show the following. Systems, I.Mech.E., S492/S18/99.
[3] D. Schoeppe, P. Spadafora, N. Guerrassi, G. Greeves, D.
(1) The momentum efficiency increases for nozzles with Geurts, Diesel Common Rail Technology for Future
more convergent spray-hole shapes and entrance High Power and Low Emission Standards, Dresden
rounding. 2005.
(2) Observed engine exhaust smoke levels decrease with [4] N. Sterno, G. Greeves, S. Tullis, X. Jiang and H. Zhao,
nozzles having higher momentum efficiency. Improvements of the KIVA Dense Spray Modelling for
(3) The momentum efficiency decreases and the initial HSDI Diesel Engines, Fuels and Emissions Conference,
spray jet angle increases as the nozzle needle lift is January 2007, Cape Town, SAE 07SFL-118.
reduced. This means that fuel injected at low needle [5] K.R. Browne, I.M. Partridge and G. Greeves, Fuel
lift will have lower penetration and tend to be Property Effects on Air/Fuel/ Mixing in an Experimental
deflected much more rapidly by the air swirl. The Diesel Engine, SAE 860223.
latest direct-acting piezo injector has much faster [6] I.M. Partridge and G. Greeves, Interpreting Diesel
needle opening and closing that minimises the Combustion with a Fuel Spray Computer Model,
quantity of fuel injected at low needle lifts giving Conference on Combustion in Engines and Hybrid
emissions benefits. Vehicles, I.Mech.E., C529/025/98.
(4) The nozzle design affects the initial jet angle at full [7] B.S. Higgins, J.M. Meuller and D.L. Siebers,
lift with higher momentum efficiency nozzles giving Measurements of Fuel Effects on Liquid-Phase
a narrower initial jet angle. Penetration of DI Sprays, SAE 1999-01-0519.
[8] D.L. Siebers, Scaling Liquid-Phase Fuel Penetration in
The advancement in understanding of these nozzle spray jet Diesel Sprays Based on Mixing Limited Vaporisation,
characteristics will contribute to future diesel fuel injection SAE 1999-01-0528.
nozzle design and manufacture. [9] I.M. Khan, G. Greeves and C.H.T. Wang, Factors
Effecting Smoke and Gaseous Emissions from Direct
ACKNOWLEDGEMENTS Injection Engines and a Method of Calculation, SAE
730169.
The authors would like to thank Dr Detlev Schoeppe, [10] S. Hong, M.S. Wooldridge, H.G. Im, D.N. Assanis and
Director of Engineering, Delphi Diesel Systems for permission H. Pitsch, Development and Application of a
to publish this paper. Thanks are also due to the authors’ Comprehensive Soot Model for 3D CFD Reacting Flow
colleagues and in particular K.R. Browne for his support in the Studies in a Diesel Engine, Combustion and Flame 143
experimental work. (2005) 11-26.
[11] J.M. Desantes, R. Payri, F.J. Salvador and J. Gimeno,
NOMENCLATURE Measurements of Spray Momentum for the Study of
Cavitation in Diesel Injection Nozzles, SAE 2003-01-
Symbol Quantity SI Unit 0703.
[12] D. Schoeppe, M. Hardy, S. Zuelch, D. Geurts, Exceeding
Ah Minimum geometric area of the spray hole m2 Customer Expectations with the Innovative Direct
Cd Nozzle spray-hole discharge coefficient ratio Acting Diesel Fuel Injection System: Unique
Fp Force measured on a plate to destroy all Advantages of a Breakthrough Technology, 29.
of the axial spray/jet momentum N International Wiener Motorensymposium 2008.
FSQ Fraction of injected fuel from part of jet section [13] Patent Application GB 0713678.1
going directly over top of piston bowl ratio
m& Actual mass flow rate exiting the spray hole kg/s
m& th Theoretical mass flow kg/s
M& tham Theoretical momentum flux exiting the spray
hole based on the actual mass flow kg m/s2
M& Actual momentum flux exiting the spray hole
kg m/s2
P1 Upstream or source fuel pressure N/ m2
P2 Downstream or back pressure N/ m2
RAF9 Rate of air/fuel mixing through an air/fuel ratio of 9

8
8