Twr Symposium (lnternaticmal)on Combustion/The (]ombustioll lustittltc, 1984/pp.

201-208

M E A S U R E M E N T S OF SWIRLING F L O W IN A D I R E C T INJECTION
DIESEL ENGINE UNDER MOTORED AND FIRED CONDITIONS

KEIYA NISHIDA, AKIRA MURAKAMI ant) HIRO HIROYASU

Department of Mechanical Engineering, University of Hiroshima
Saijocho, Higashi Hiroshima, 724, Japan

Swirling flow in a cylinder of a direct injection diesel engine was investigated under mo-
tored and fired conditions. Air flow velocity in a motored engine was measured by a laser
doppler velocimeter (LDV). Profiles of tangential and axial velocities of the swirling flow,
and a local swirl ratio, which was defined as the ratio of the local tangential velocity to an
engine revolving speed, were obtained. The combustion process in a fired engine was ob-
served by high speed photography. Local swirl ratios of a flame during the combustion pro-
cess were measured by tracing the flame motion on the combustion photographs. Tile history
of the swirl ratio of a flame during the combustion of a small amount of fuel could be con-
sidered to be a standard for histories of swirl ratios of a flame inside a cavity radius.
The local swirl ratios at various positions in a cylinder show different histories under both
motored and fired conditions. The value is larger inside a cavity radius than outside it. There-
fore, it can be considered that the swirling flow in a cylinder consists of two regions, that
is, the regions inside and outside the cavity radius. The comparison of measured results
under motored and fired conditions shows that the swirl motion inside the cavity radius is
accelerated by the effect of combustion, especially by the effect of pre-mixed combustion in
an early stage of diesel combustion. These experimental results can be utilized to verify the
results of mathematical modeling for the swirling flow in a cylinder.

Introduction ratios was made, and the effect of combustion on

swirling flow was considered.
In recent years, small and high-speed direct in-
jection (D.I.) diesel engines have been used in light
duty vehicles because of their superior fuel econ- Experimental
omy. However, there still remain many problems
related to the combustion mechanism of a D.I. die- Basic Engine
sel engine, such as, exhaust emissions, noise and
vibration. For these problems, air motion in a cyl- The basic engine used in this study is a single-
inder is a controlling factor, because it greatly in- cylinder four-stroke D.I. diesel engine. The basic
fluences the entire combustion process, including specifications of this engine are shown in Table I.
liquid fuel vaporization, mixing of fuel and air, and
flame development. Many studies have been car- LDV System
ded out on air motion in the cylinder of a direct
injection diesel engine under motored conditions, Figure 1 shows a LDV system and a modified
using a hot wire anemometer 1-5 or a laser doppler engine. The principal modification of the engine was
velocimeter (LDV). 6-8 Under fired conditions, installing an elongated cylinder liner and a con-
however, swirling gas flow has not been fully ana- necting rod in the basic engine, and fitting glass
lyzed. windows to the elongated cylinder liner and the
The aim of this paper is to present the results of piston with a slit for a laser beam path. These mod-
studies on swirling air flow in a D.I. diesel engine ifications make it possible to measure the air flow
under motored conditions using LDV, and on velocity in the whole cylinder space, including a
swirling gas flow under fired conditions using high- cavity formed in the top of the piston.
speed combustion photography. Histories of local The LDV was operated in a two-beam mode, and
swirl ratios, defined as the ratio of the local tan- a forward scattering system was used for the optical
gential velocity of the swirling flow to an engine arrangement to obtain a high intensity signal burst.
revolving speed, were obtained under both condi- Velocity was determined using a frequency tracker.
tions. A comparison of both histories of local swirl An ensemble average of air flow velocity was taken

201
202 AUTOMOTIVE ENGINE COMBUSTION

TABLE I Inlet V a / v e ~
Engine specifications
Bore I 10 mm
Stroke 125 mm
Clearance Volume 99 cm3
Swept Volume 1187 cm3
Compression Ratio 13
Closing Timing of inlet Valve
Opening Timincj of Exhaust Valve 148clecj.
-133deg.ATDC
ATDC

for 10 cycles to obtain a history of mean velocity

for 1 cycle.
Figure 2 shows measurement positions in a cyl-
inder. Tangential and axial components of the swirl-
ing flow were measured at these positions. 'o'' vo'v"
The engine was driven by an electric motor at a
speed of 16.7 s -1 (1000 rpm), and was not super-
charged.
N : T
High-Speed Combustion Photography , t I I
The engine was also modified to form a light path
in order to take high-speed combustion photo-
graphs, as shown in Fig. 3. The elongated cylinder
ioll [ I II
liner and piston with a transparent piston crown were
installed in the engine. A cylindrical cavity was used,
instead of the toroidal cavity in an actual engine,
to prevent the image from being distorted. As shown
in Fig. 4, the range of 84 percent diameter of the
;~yjJ ,~'i
"Pist~
, ~72
;~ol
I ~ jr"
_I

cylinder bore could be observed.

The engine was operated under the following
conditions: engine speed, 16.7 s -1 (1000 rpm);
amount of fuel injected, 100 mg/cycle (equivalence
ratio d~ = 0.806); boost pressure, 0.064 MPa; and FIG. 2. Measurement positions.
intake air temperature, 373 K. The intake air was
heated in order to obtain the same ignition delay
as in an actual engine. The fuel injector used was ector
a 4-hole nozzle with a hole diameter of 0.38 mm. He<ld
The fuel was a diesel fuel with a density of 0.835
g/cm 3 and a cetane number of 55. ~1 Cavity

ent
Crown
fl
d Piston
S..... gP(lr ..... ~ H ~ ~ ............ lider Liner
;i !i!:[T ,'~ -- Cy[iRd,rHeod Lamp
\~].r Photomultiptier I Lens
re Mirror
~r er ~ ~ vnSz= M....

.2
/ Lens~ Trocker
BeamSpiller J ~ Window
Wind~ I - ~
roroidal Cavity/ _.]
Elongated Cyhnder Liner~
ElongatedConnectinqRod~/-~ Data Recorder ,d Camera

Flc. 3. Modified engine for high-speed combus-

FIG. 1. LDV system and modified engine. tion photography.
 MEASUREMENTS OF SWIRLING FLOW IN DIESEL ENGINE 203

Outside CovityRod~ . . . . .... jCylider LinerWoll Results and Discussion

Inside CovityRodius ~Visible Field
[/
'-~'~. CovityWoli Swirling Flow under Motored Condition
Figure 5 shows the history of tangential and axial
velocity profiles in the horizontal and vertical sec-
tions of a cylinder. Swirling flow is attenuated in
the vicinity of the cylinder wall and can not be re-
~-. ust Volvo
garded as a perfect solid swirl. The tangential ve-
Inlet Volve "Fuel Injector locity at 0 = 90 deg. is much smaller than at 0 =
Spray Axis
- 9 0 deg. The center of the swirling flow almost
FIG. 4. Visible field. coincides with the cylinder axis. The axial velocity

= 10mm

25 mm

5 0 mm

0~10m/s
= -90 deg. 0 deg. 90 deg.
(TDC)
FIG. 5. Histories of a velocity profile in a cylinder.
 204 AUTOMOTIVE ENGINE COMBUSTION

as -~- OJs/03 e (1)

where o~,; local angular velocity of swirling flow

toe; engine revolving speed

Figure 6 shows the histories of R, at the positions

which were selected in the four sections as follows:
i/I
tY (a) 10 mm below the cylinder head, (b) equally dis-
4 tant from the cylinder head and piston head (a me-
dian section), (c) 10 mm above the piston head, and
O
(d) 9 mm below the piston head (in a cavity). The
~3 sections (b), (c) and (d) float with the piston move-
n- ment. Rs at R = 28 mm is larger than that at R
= 48 mm in any section. Rs after the TDC is at-
tenuated as compared with before the TDC. How-
"- 2
ever, R~ seems to take a maximum and a minimum
U3 value symmetrically with respect to the TDC.

Observation of Combustion Process

0 Combustion photographs were taken under the
-160 -120 -80 -40 0 40 80 120 160 conditions of dynamic injection timings 0inj = --28,
(TDC) --13 and - 5 deg. Figure 7 shows diagrams of cyl-
Crank Angle 0 deg. inder pressure and rates of heat release for each
FIG. 6. Histories of a local swirl ratio in a cylin- injection timing. As the injection timing is ad-
der. (floating positions with piston movement) vanced, the heat spike due to premixed combustion
increases.
Figure 8 shows combustion photographs for the
profile is not uniform in any section, and the ve- above injection timings. During the ignition delay
locity is smaller in the section of the cylinder head period, an injected fuel spray impinges on the cav-
side. ity wall and moves along it. Vaporized fuel is dis-
The local swirl ratio Rs was obtained with Eq. tributed by the swirling flow, and the shape of a
(1). flame, immediately after ignition, implies the dis-

10 ,1.0

o8 0.8~
13.

o- 6
t.
o.6
3
~4
_~j~. jdCV'dO -.j rv

0.2 ~
c- I

~,0 o ~
Fuel Injection Fuel Inject,on
r--
Fuel Injection
0
I I I I I L J [ I ] I I I I I I
-30-20-10 0 10 20 30 -30-20-10 0 10 20 30 --30-20-10 0 10 2~0 30 40
(TDC) (TDC) (TDC)
Crank Angle e deg. Crank Angte @ deg. CrQnk Angle 8 deg.
(A) 0in j = -28 deg. (B) 0. = -13 deg. (C) 0 . . = -5 deg.
inj xnj
FIG. 7. Cylinder pressure diagrams and rates of heat release.
 MEASUREMENTS OF SWIRLING FLOW IN DIESEL ENGINE 205

0 = -14 deg. -13 deg. -12 deg. -ii deg. -i0 deg.

(A) O . . = -28 deg.

in 3

0 = -4 deg. -3 deg. -2 deg. -i deg. 0 deg.

(TDC)
(B) 0 . . = -13 deg.
inj

0 = 7 deg. 8 deg. 9 deg. i0 deg. ii deg.

(C) 0in j = -5 deg.

FIG. 8. Combustion photographs.

tribution of vaporized fuel at ignition timing. After ~oe; engine revolving speed
the ignition occurs, the flame spreads rapidly through
the whole combustion chamber. When the injec- Figure 9 shows the comparison between swirl ra-
tion timing is advanced and the heat spike due to tios inside a cavity radius measured by LDV and
pre-mixed combustion increases, the rate of flame combustion photography. In the case of combustion
spreading increases. After flame propagation ceases, photography, a very small amount of fuel (about 10
the flame is driven and rotated by the swirling flow. mg/cycle, equivalence ratio, ~b = 0.081) was in-
An analysis of swirling flow under fired conditions jected and burned. Under this condition, the in-
was made after flame propagation almost finished. crease in cylinder pressure and gas temperature was
very small, and the effect of combustion on the flow
Swirling Flow under Fired Conditions can be regarded as negligible. The history curve of
Rsy obtained by combustion photography is similar
The local swirl ratio of a flame R~f was obtained to that of Rs at the position 10 mm below the cyl-
with Eq. (2). inder head obtained by LDV. R~yis a little smaller
than Rs around the TDC. This difference seems to
I~ss = O~ssl~,~ (2) arise from the differences of the piston cavity shape
and the intake air condition. As a result, it can be
where 0Osf; local angular velocity of a flame considered that the swirl ratio obtained by flame
206 AUTOMOTIVE ENGINE COMBUSTION

0inj = - 5deg. ~,
6
[] :Rs,by LDV
6-
, J /-13deg. o

, \
s'
.o..

0 r:l

03

2 -•:Rsf,by Combustion P h o t o g r a p h y . '2 9 e "9 ~/ e"~.A..~

~-
Outside Cavity Radius
03 \ Inside Cavity Radius ] ~ 1
03 9 9inj = - 5deg.
1 9 - 13 deg.
9 -28 deg.
I I I 1 0 I I 1 I
-% 6
(TDC)
20 40 6o 8o -2o 0
(T D C)
20 40 60 80
Crank Angle 0 deg. Crank Angle 0 deg.
F~c. 9. Comparisons of local swirl ratios obtained FIG. 10. Histories of local swirl ratios of a flame.
by LDV and combustion photography.

movement almost agrees with the swirl ratio of gas

in a cylinder. Thus, the history of Rd for the com-
bustion of a small amount of fuel is recognized as
a standard for R~f inside the cavity radius for var-
ious injection timings. 6 -
--• \
~ ~ /
@inj = - 5 deg.
- 13 deg.
Figure 10 shows the histories of the local swirl
ratio of a flame under the conditions of 0in j ---- --28,
-13 and - 5 deg. Rsf were measured inside and
outside the cavity radius (see Fig. 4). A sudden de- '5 ~, ~. / ~ " Inside Cavity Radius
crease of Rd inside the cavity radius seems to be
delayed when the injection timing is retarded. Fig-
ure 11 shows the comparison between the histories 0
of Rd inside the cavity radius for various injection
timings and for the combustion of a small amount
of fuel. Rsf inside the cavity radius for each injec- "63
tion timing is larger than for the combustion of a 0

small amount of fuel. Swirling flow seems to be ac-

~2- Rsf under Combustion of
celerated by the effect of combustion, especially by
Small Amount of Fuel
the effect of pre-mixed combustion in the early stage
of diesel combustion. The process of acceleration of
a swirling flow due to combustion has been ex- 1
plained as follows.9 During an ignition delay pe-
riod, the mixture of vaporized fuel and air is dis-
tributed along the cavity wall, and air exists inside 0 I I I I
the mixture. When ignition occurs, the mixture -20 0 ZO 40 60 80
burns and expands quickly. The air, which was ro-
(TDC)
tating inside the mixture before ignition, is pushed
Crank Angle e deg
further into the region inside the cavity by the ex- FIG. 11. Effect of combustion on histories of local
pansion of the burning mixture along the cavity wall. swirl ratios of a flame.
 MEASUREMENTS OF SWIRLING FLOW IN DIESEL ENGINE 207

If the angular momentum of the air inside the Instn. Mech. Engrs., Vol. 188, 21/74, pp. 269-
burning mixture is assumed to be conserved, the 280, 1974.
rotational velocity of this air is accelerated. 2. BRANDL, F., REVERENCIC,I., CARTELLIERI,W. AND
DENT, J. C., "'Turbulent Air Flow in the Com-
bustion Bowl of a D.I. Diesel Engine and Its
Conclusions Effect on Engine Performance," SAE Paper No.
A swirling flow in a cylinder of a direct injection 790040, 1979.
diesel engine was investigated under motored and 3. WILLIAMS, T. J. AND TINDAL, M. J., "Gas Flow
fired conditions. The air flow velocity in the mo- Studies in Direct Injection Diesel Engines with
tored engine was measured by LDV. Local swirl Re-Entrant Combustion Chambers," SAE Paper
ratios, which were defined as the ratio of the local No. 800027, 1980.
tangential velocity of the swirling flow to an engine 4. TANABE, S., HAMAMOTO, Y. AND OHIGASHI, S.,
revolving speed, were obtained. The combustion "'Swirl in a Four-Stroke Cycle Engine Cylinder,"
process in the fired engine was observed by high- Bull. JSME, Vol. 21, No. 152, pp. 287-294, 1978.
speed photography. Local swirl ratios of a flame were 5. TINDAL, M. J., WILLIAMS,T. J. AND ALDOORY,M.,
measured by tracing the flame motion on the com- "The Effect of Inlet Port Design on Cylinder Gas
bustion photographs. Motion in Direct Injection Diesel Engines,"
The local swirl ratios at various positions in a cyl- Flows in Internal Combustion Engines, Pro-
inder show different histories under the motored ceedings the Winter Annual Meeting of ASME,
and fired conditions. The value is larger inside a pp. 101-111, 1982.
cavity radius than outside it. Therefore, the swirl- 6. MONAGHAM,M. L., AND PEa'rlVEN, H. F., "'Air
ing flow in a cylinder can be considered to consist Motion and Its Effect on Diesel Performance and
of two regions, that is, the regions inside and out- Emissions," SAE Paper No. 810255, 1981.
side the cavity radius. The swirl motion inside the 7. BALL, W. F., PETrIFER, H. F. AND WATERHOUSE,
cavity radius seems to be accelerated by the effect C. N., "'Laser Doppler Velocimeter Measure-
of combustion, especially by the effect of pre-mixed ments of Turbulence in a Direct-Injection Die-
combustion. These experimental results can be uti- sel Combustion Chamber," Instn. Mech. Engrs.,
lized to verify the results of mathematical modeling C52/83, pp. 163-173, 1983.
for the swirling flow in a cylinder. 8. MURAKAMI,A., ARAI, M., HIROYASU,H. AND KA-
JIYAMA,K., "Air Motion in a D.I. Diesel Engine
Acknowledgment (Analysis of Flow in a Cylinder and a Cavity by
Laser Doppler Velocimeter)," Proceedings of the
This research has been supported by a grant from Spring Annual Meeting of the Japan Society of
Komatsu Ltd. The authors wish to thank M. Shiga Automotive Engineers (in Japanese), pp. 135-140
for his assistance in carrying out the experiment. (Paper No. 841025), 1984.
The authors also would like to thank Dr. M. Arai 9. GYAKUSHI,N., TAKAMOTO,Y. AND TERAI, J.: "Ef-
for stimulating and helpful suggestions and discus- fect of Swirl on Combustion in Diesel Engine
sions. (lst. Report, Swirl Speed in Motored and Fired
Conditions)," Transactions of the Japan Society
REFERENCES of Mechanical Engineers (in Japanese), Vol. 50,
No. 451, B, pp. 770-779, 1984.
1. DENT, J. C. AND DERHAM, J. A., "Air Motion in
a Four-Stroke Direct Injection Diesel Engine,"

COMMENTS

R. R. Maly, Daimler-Benz AG, W. Germany. In of the air flow are shown in Fig. 5. A suitable change
your graphs there was no indication of squish flow. the optical arrangement would make it possible to
Was it possible to separate the squish flow from swirl measure the radial component of the air flow. In
with your technique and what was actually plotted? the combustion photography, only the tangential
component of the gas flow was measured. The squish
Authors" Reply. In this study, LDV measurement flow is a problem that should not be ignored, and
was not made on squish velocity, that is, the radial is the subject for future study.
component of the air flow in a cylinder. The mea- The data plotted in Figs. 6, 9, 10 and 11 are for
sured results of the tangential and axial components the local swirl ratio that is defined as a ratio of local
208 AUTOMOTIVE E N G I N E COMBUSTION

tangential velocity of the air flow (rad/s), that were Authors" Reply. In this study, we did not observe
measured by LDV and combustion photography, to the wall quenching of a flame in a cavity and a cyl-
an engine speed (rad/s). inder, and did not take any data on heat transfer
between a wall and unburned gases.
Swirl ratio and turbulence were measured in a
cylinder during a whole engine cycle (induction,
compression, expansion and exhaust strokes). Al-
S. Durrani, University of Maryland, USA. W h e n though the measurement of the air flow velocity at
you sprayed the fuel along the chamber wall did the inlet port was not made, the jet velocity aspir-
you obtain any data on wall quenching effects on ated through the clearance between a valve and a
heat transfer to the unburned gas? Also, were any valve seat was m e a s u r e d , and its effect on the
measurements made on swirl ratio or turbulence at swirling flow in a cylinder was analyzed. (See, SAE
the inlet part? paper No. 840518 by Nishida et al.)