Get documents about "EXPERIMENTAL AND NUMERICAL ANALYSIS OF FUEL FLOW IN THE
Document Sample
Journal of Kones. Combustion Engines, VoIB, No 1-2,2001
EXPERIMENTAL AND NUMERICAL ANALYSIS OF FUEL FLOW IN
THE DIESEL ENGINE INJECTION NOZZLE
Martin Volmajer, Breda Kegl, Ph.D.
Research assistant Assistant Professor
martin. volmajer@uni-mh.si hreda.kegl@uni-mh.si
Phone: +38622207738 Phone: +38622207732
Fax: +38622207990 Fax: +38622207990
University ofMaribor, Faculty of mechanical engineering,
Smetanova 17, S/-2000 Marihor, Slovenia
Abstract. The geometry of the diesel fuel injection nozzle and fuel flow characteristics in the nozzle
significantly affects the processes of fuel atomisation, combustion and formation of pollutants
emissions in a diesel engine.
In this paper numerical and experimental results of the nozzle fuel flow analysis for a four-hole
injection nozzle Bosch DLLA 148 S 311376 are presented.
To describe the nozzle fuel flow, a three-dimensional computational fluid dynamics (CFD) model is
employed. The CFD package FIRE (AVL Graz) is used for computation. The results represent the fuel
flow characteristics for steady state flow conditions at different needle opening. For this purpose
several three-dimensional models representing different needle lifts are made.
The experimental results are obtained by measuring the fuel flow coefficients at steady state
conditions on the nozzle flow tester made in Engine research laboratory (ERL) at the Faculty of
mechanical engineering in Maribor. The fuel injection pump is driven by an electric motor, the
pressure control valve regulates the pressure at 100 bar and the calibration fluid is injected through the
nozzle into the measuring Plexiglas cylinder.
The fuel flow coefficients obtained from the experimental results at steady flow conditions in the
nozzle are compared with the results of the CFD analysis.
1. Introduction
A modern compression ignition engine should meet ecological and economical requirements.
It should have high performance-fuel consumption ratio, low maintenance costs and it should
enable operation under prescribed emission regulations. As the process of combustion and
further the production of pollutants and noise emissions is mainly controlled by the process of
fuel injection, a lot of effort is put into the development of new and improvement of existing
diesel fuel injection systems.
Considering the fuel injection process in the in-line fuel injection system (FIS), the pump
characteristic is mainly affected by changing the cam profile and the plunger geometry. The
above mentioned pump characteristics and the geometry of the injection nozzle influence the
injection characteristics and further the fuel spray characteristics. The injection and the fuel
spray characteristics connected with the combustion chamber geometry control the
combustion and pollutant formation processes. Therefore the engine operation characteristics
could be improved by improving one of the above mentioned fuel injection systems- or
engine-parts. Even though in recent years many researches have been made considering the
fuel injection systems, as well as their application in variety of the motor vehicles, the fuel
injection nozzle still presents quite an unresearched area.
The part of the nozzle downstream of needle seat is almost the same whether the nozzle is
made for a common rail injection system or unit injector system or even for the in-line FIS.
9
Therefore the results of the researches on one type of the fuel injection system can be applied
on almost every system.
By the nozzle flow characteristics analysis, the flow coefficient, nevertheless its simplicity,
represent very important parameter of the injection nozzle characteristics. The higher the
value of flow coefficient is, the larger fuel quantity is injected per time, which yields higher
outflow velocity (suppose no changes in outflow cross-section) and by this better atomisation.
In the last few years, analysis using computational fluid dynamics (CFD) programs became
important, since this is a relative easy way to analyse the fuel injection process on newly
designed or re-designed fuel injection systems. By using CFD programs a lot of manufacture
and experimental work could be spared, since there is no need to produce every tested variant.
In this way many variants could be analysed in a relatively short period of time.
2. Theoretical backgrounds
Flow coefficient
FLOW COEFFICIENT is defined as ratio between the measured or real (Vreal) and theoretical
(Vth) volume flow injected through the nozzle. According to Bernoulli equation, the
theoretical outflow velocity can be derived from the pressure difference (L1p) and fuel density
(p): J.L = V~eal = V real (eq.I)
-t-dp
•
v,. A
d p
Ad represents the sum of the nozzle hole cross-section area. The Density of the fluid is derived
from equation 2 [I]:
P20 - (0.71- 0.001345· (P20 - 850)). (r - 293) ,
P = .:.-..::c::...--'-- _ ____.:.
--'::.......::::'--_~_'__
(eq.2)
1- p·Ka
D
where P20 represents the density of the fluid at 20 C, T is the temperature and p is the
pressure, while K; is a factor dependent on p, T and P20.
Theoretical-empirical model for determination of flow coefficient
Hardenberg presented in [2] a theoretical-empirical model for calculation of the nozzle flow
coefficients. The flow coefficient is defined as a function of internal nozzle cross-sections
surfaces and the nozzle-hole flow coefficient value (/.ld), which was defined from the series of
the flow coefficient measurements at high needle lifts.
1
(eq.3)
u= (1)2 (Ad A d)2'
l f.ld l
+ A R - As
where A R represent the cross section surface at the needle seat and As in the SAC volume.
3. Injection nozzle design
Analysis were made on the four-hole diesel engine injection nnozzle (Bosch DLLA 148 S
311376) with dimensions presented on figure I and in table I.
Table 1: Nozzle dimensions
IO
Figure 1. Nozzle dimensions
4. Experimental apparatus
The fuel flow coefficient-measuring device is made at the ERL. The testing device is used for
measuring the fuel flow coefficients at the steady state conditions and the procedure is in
commonly known as the Bosch procedure for measuring flow coefficients.
The hydraulic scheme of the testing device is presented on figure 2.
MANOMETER
0- 160bar
2,0 ,
Measuring
1,5 plexiglass
1
1.0
0,5
OUTLET
PIPE
MICROFILTER
Figure 2. Hydraulic scheme ofthe nozzle flow coefficient testing device
The testing device consist of the low and high pressure pump, each driven by an electric
motor, pressure chamber for reduction of pressure waves, pressure regulation valve, three way
electromagnet valve for changing the direction of the flow into the measuring Plexiglas valve
11
or directly into the fuel tank. The nozzle holder enables the measuring of different nozzles.
The needle lift is fixed by using a micrometer.
Uncertainty of measurement
UNCERTAINTY OF MEASUREMENT is the parameter that defines the borders within
which the real value of the measured quantity lies with some probability [3]. The uncertainty
of measurement is divided into two series [3][4]: Type A (UA), which is defined by the
statistical methods (repetition of measurements, regression...) and type B (UB) representing the
producer information about accuracy/precision of some measuring instruments, experiences,
etc. The total uncertainty of measurement (u) is defined as:
u(x)=~u~(x)+u~(x) (eqA)
Where for the directly measured values u is defined as square root of sum of the individual
measurement uncertainties:
U(x)= I.u~j(x)+ I.u~j(x) (eq.5)
j=1 j=1
and for indirectly measured values, the uncertainty is dependent on the number of m mutually
independent directly measured values (xi;i=l,m). The total uncertainty of indirectly measured
quantities u(y) is then:
u(y)= ~(~ r u'(x,)
On figure 3 the uncertainties of the flow coefficient measurement are presented. The
(eq.6)
uncertainty of flow coefficient measurement with respect to the calculation is indirectly
dependent on the real volume flow definition uncertainty and the theoretical volume flow
definition uncertainty. The real volume flow is indirectly derived from the volume and time
measurement, while the theoretical volume is derived from the pressure difference
measurement, hole diameter dimensions with certain production tolerance that is considered
as the uncertainty and density, calculated from the measured temperature and pressure
difference.
DVreaJ DVth
Real volume flow..lPeasurament .uncertainty Theor. volume,]ow.measyre, uncertainty
.-;: .-;: ~
~ T
DVm Dt D.6.p Dd Dp
Volume measurement Time defining Press. dif. Hole Density
un~rta~ty uncertainty measure. diameter measure.
- ......
I
unpertaintv tolerance uncertainty
UVm.A II UVm.B ~
~
<l
~=
..at ut ~
D.6.p
= =
Figure 3. Uncertainty offlow coefficient measurement
12
Measuring process
The volume of the fluid injected at predefined time is measured at constant pressure drop for
different needle lifts. For the needs of the density calculation by using the equation 2, also the
pressure and the temperature are measured.
The calculation of the measured value of flow coefficient with corresponding uncertainties
were calculated for every needle lift from ten time repeated measurement of volume of the
fluid, pressure drop and temperature.
s. Numerical analysis
CFD-CODE - Numerical analyses were taken by using the CFD program FIRE (ver 6.2b),
which is commonly used in automotive industry. The FIRE program uses a finite volume
method for the simulation of fluid flows. The flow variables are calculated by solving the
conservation equations (mass, momentum, turbulence, passive scalar, energy) for general,
non-orthogonal, boundary-fitted co-ordinate systems. The FIRE analysis program uses an
iterative process to approximate the solutions for these equations for each cell. When this is
complete, another iterative process calculates and reconciles the effects that flow variables in
each cell exert upon neighbouring cells as a function of time[5].
Computational model
To analyse the flow characteristics of the in-nozzle flow five different nozzle models,
representing nozzle lifts from 0.05 to 0.35 mm, were made. Since some analysis [6][7] shown,
that the pressure drop in nozzle is significant only in the area of the needle seat, sac chamber
and nozzle holes, the meshes were modelled only for the above mentioned parts. Some further
simplifications considering the use of one half mesh model of the nozzle were made
according to the results of previously made analysis [8], which indicated no significant
difference between the results using either a real model or an one half model of the nozzle.
The mesh models at maximum needle lift of 0.35 mm with relevant number of mesh nodes
and elements for all five models are presented on figure 4.
Inlet: p= 100 bar
h=O.05 mm
No.cells:
39648
h=O.lmm
No.cells:
37632
h=O.2mm
No.cells:
I Outlet: p=l bar I 39648
Hole #1
Figure 4. Mesh model ofnozzle at the needle lift 010.35 mm with specified boundary conditions an
number ofcells for 5 nozzle mesh models representing different needle lifts
13
Initial and boundary conditions
According to steady state analysing conditions, pressure boundary conditions at the in- and
outlet are specified. The fluid used for analysis is the calibration fluid according to ISO
4113[9], with the temperature of 313 K, the density 825 kg/nr' and kinematic viscosity of 2.6
mmvs. K-£ turbulence model is employed. Since maximal velocities are much smaller than
the speed of sound, the fluid is supposed to be uncompressible.
6. Results
Results of the measurements
The flow coefficients derived from the measurements of the volume flow rate; pressure
differences and temperatures are presented in table 2, where the measurement uncertainties
are presented too.
Table 2: Results of the flow coefficient measurements with uncertainties
0.21354 ± 0.0150
0.38933 ± 0.0157
0.51722 ± 0.0165
0.57686 ±0.0169
0.60664 ± 0.0172
0.61799 ±0.0173
0.62380 ± 0.0173
Results of the numerical analysis
The numerical analyses were made mainly for the comparison with the measured values of
the flow coefficients. On the other side, the CFD analysis enables a more complex analysis of
the processes in the nozzle by representing the values of every calculated variable in every
cell of the nozzle model. At the presented analysis, velocity fields and pressure distribution in
the nozzle at different needle lifts, presented on figure 5 and 6, were of the main interest.
At small needle lifts, the main pressure drop occurs at the needle seat, which yields lower
pressures and further lower velocities in the SAC chamber and nozzle holes. At higher needle
lifts the main pressure drop occurs at the SAC chamber and in the nozzle holes or at the
outlets, which results in the higher outflow velocities.
Non-uniform velocity profiles appear at the nozzle holes, especially at the nozzle hole with
larger inclination angle, which should be connected with the sharp edges at the nozzle hole.
Table 3 present the calculated values of the flow coefficients derived from the CFD analysis.
Table 3: Flow coefficents from CFD analysis
14
\ 1\'
j
"
Figure 5. Calculatedpressure distribution/or needle lifts: 0.05, 0.1, 0.2, 0.3 and 0.35 mm
Figure 6. Calculated velocity vectors in nozzle at needle lifts: 0.05, 0.1, 0.2, 0.3 and 0.35 mm
Results of the empirical model
The results from the theoretical-empirical model of Hardenberg are presented in table 3. The
value of )1d is defined according to statement in [10]. The value of )1d for applied nozzle is
0.6694.
7. Comparison of the results
The difference between the results is not significant (Figure 7). Larger difference between the
results of measurements and results of the theoretical-empirical model respectively CFD
analysis appears only at the needle lift of 0.05 mm, with deviation of almost 30% in both
cases. At the other needle lifts deviations lie between 8 and 12 % for CFD and 3 to 7 % for
the theoretical-empirical model. The presented difference in case of CFD analysis is
reasonable, since calculations were made on the simplified model, where the pressure
difference between the inlet and outlet were the same as the difference at the measurement,
not taking into account the simplifications made.
On the other side, the larger difference at the needle lift of 0.05 mm could also be the results
of some uncertainties of the measurement that are not considered in the above presented
equations. One of them is probably looseness of the connection between the needle and the
micrometer used for setting the needle lifts.
15
0,7
0,6
,====;~~~~;;r
t-
0,5 +-------7IIIJ;Ir:-------------1
~ Measurement
0,4 +-----A~------------__i
::1. .... CFD
0,3 +---~~-----------__i
~Empirical
0,2 +--~~---~---------__j
0,1 +-~"------------.---__j
01lF--.....,.-----,----,...------,-----,---,-----i
° 0,05 0,1 0,15 0,2 0,25 0,3 0,35
b [mm]
Figure 7. Comparison ofthe calculated and measured results ofthe flow coefficient values
8. Conclusion
From the presented results the following conclusions could be made. Flow coefficient testing
device constructed at the ERL yields sufficiently precision, with reasonable uncertainties of
the measurement.
To refine the precision of the measurement, by defining the exact value of the pressure
difference, the pressure downstream of the nozzle should be measured, or the nozzle position
should be changed so, that the fluid would be injected directly into the measuring Plexiglas.
For the same purpose, Plexiglas cylinder with high ovalness should be replaced with the
glasslPlexiglas cylinder with proper circle cross-section.
With some small changes, the presented testing device also enables the measurement of the
flow coefficient separately for each nozzle hole, which brings better comparison with the
results of CFD analysis when the simplified models, introducing only one hole, are applied.
9. References
[1] Fomin Ju.Ja., G.V.Nikonov, V.G. Ivanovski, Toplivna aparatura dizeljev,
Masinostroenie, Moskva 1982, p.82-83
[2] H.Hardenberg, Die Nadelhubabhaengigkeit der Durchflussbeiwerte von Lochduesen fuer
Direkteinspritzdieselmotom, MTZ 46( 1985) 4,143-146
[3] WECC-19:Guidelines for the Expression of the Uncertainty of Measurement in
Calibrations, WECC, 1990
[4] USM-SA: Merilna negotovost, A13, Slovenian Standards and Metrology Institute, 1996
[5] FIRE Version 6.2b User Manual, AVL List GmbH, Graz
[6] y'Oishi et.al., A Computational Study into the Effects of the Injection Nozzle Inclination
Angle on the Flow Characteristics in the Nozzle Holes, SAE paper 920580
[7] K.Melcher, J.Chomiak, Experimentelle Untersuchung der Stroemung durch
Dieseleinspritzduesen in stationaer betriebenen Grossmodel, Bosch Techn.Berichte
5(1976)4
[8] M.Volmajer, Numericna in eksperimentalna analiza tokovnih karaktersitik vbrizgalne
sobe dizelskega motorja, Master thesis, Faculty of Mechanical Engineering Maribor,
Maribor 2001
[9] ISO-4113: 1988, Road Vehicles- Calibration Fluid Diesel Injection Equipment
[10] B.Kegl, Inverzna racunska pot za dolocitev konstrukcijskih parametrov vbrizgalnega
sistema dizelskega motorja, doktorska disertacija, Tehniska fakulteta Maribor, Maribor,
1992
16
Related docs
