192

Paper 18
THE PROBLEM OF PREDICTING RATE OF HEAT
RELEASE IN DIESEL ENGINES
H. C. Grigg* M. H. Syed*

Two simple models for the rate of heat release in diesel engines are described. The factors taken into account
in the models are rate of entrainment of air into the fuel sprays, the rate of turbulent mixing of fuel and air
within the spray, and the chemical kinetics of burning. The models differ in their treatment of the rate of
air entrainment. Comparisons are made with experimental results for a diesel engine running at two speeds
and a variety of turbocharging ratios. The overall agreement with experiment in respect of shape of rate of
heat release diagram is good, with the exception of the naturally aspirated cases where the rate of air entrain-
ment is too low.

INTRODUCTION Fuel injection nozzle hole diameter, m.

THECALCULATION of rate of heat release diagrams is an Activation energy, J/mole.
essential part of engine cycle calculations which are, in Rate of burning of fuel, kg/m3 s.
turn, to become an important part of engine design. The Steric factor, m3/kg s
object of the work reported in this paper is to develop a Discharge coefficient.
relatively simple model to predict rate of heat release Mass flux, kg/s.
diagrams, based upon approximations to the physical Rate of diffusion of air, kg/m3 s.
factors involved. Such a model is more likely to predict Rate of diffusion of fuel, kg/m3 s.
the behaviour of new combustion systems or the behaviour Number of nozzle holes.
of existing combustion systems at new or more extreme Air concentration, kg/m3.
operating conditions than a purely empirical set of Mixed air concentration, kg/m3.
formulae. Fuel concentration, kg/m3.
The present paper reports the progress achieved so far, Mixed fuel concentration, kg/m3.
but further work is required before the model can be Air density ratio.
inserted usefully in engine cycle calculations. Fuel pressure, N/m2.
The model is based on the process of air entrainment Temperature of gases in cylinder, OK.
into fuel sprays, turbulent mixing in fuel sprays, and Time, s.
chemical kinetics. Constants are chosen such that the Injection period, s.
predicted rate of heat release fits experimental data. The Doughnut circulating velocity, m/s.
variations of these constants with engine operating con- Spray tip velocity, mjs.
ditions do not follow a recognizable pattern. The object of Distance penetrated by spray tip, m.
future work will be to increase the realism of the model Distance penetrated by back end of conical plume,
until the constants have a recognizable significance. m.
Cone half angle.
Notation Density of air, kg/m3.
a Doughnut mean radius, m. Density of fuel, kg/m3.
b Doughnut circular cross-section radius, m.
D .VT Diffusivity coefficient, s - l. E X P E R I M E N T A L DATA
The M S . of this paper was received at the Institution on 15th Sep- The experimental data used are diesel engine rate of heat
tember 1969 and accepted for publication on 16thJanuary 1970.
* C.A.V. Ltd, Acton, London, W.3. release diagrams calculated from cylinder pressure records
Proc lnstn Mech Engrs 1969-70 Vol184 Pt 3J

Downloaded from pcp.sagepub.com at The University of Auckland Library on June 4, 2016

 THE PROBLEM OF PREDICTING RATE OF HEAT RELEASE IN DIESEL ENGINES 193

Table 18.1. Engine data and experimental results

Figure Engine Turbo- Overall Conical Conical Doughnut Doughnut Experi- Conical Doughnut
reference speed, charging airlfuel plume steric plume xeric factor diffusivity mental plume fuel fuel
rev/min ratio mass ratio factor,
m3/kgs
diffusivity
coefficient, ?p2s coefficient,
mZ/s
fuel
utilization
utilization utilization
OK112 m2/s
____ - _____
18.3,18.11, 1600 23.0 4.3 x 106 6.7 x 104 6.0 X 10' 0.3521 0.71 0.46 0.71
18.13
18.4 1604 22.7 1.9 x 106 6.7 7.5 x 105 0.2951 0.65 0.61 0.65
18.5 1602 23.0 8.0 x lo5 5.8 7.0 x 105 0.0862 0.65 0.74 0.65
18.6,18.12, 1606 25.8 4.0 x 105 3.4 2.5 x lo5 0.1044 0.84 0.87 0.83
18.14
18.7, 18.15 1202 22.7 2.5 X lo6 1.0 x 102 2.4 x loG 0.2805 0.65 0.48 0.65
18.8 1211 22.2 1.35 x 10' 5.0 12.0 x 105 0,0882 0.67 0.64 0.67
18.9 1205 22.8 4.0 X lo5 5.5 5.0 x 105 0.0871 0.55 0.82 0.55
18.10,18.16j 1209 23.1 5.0 x 105 3.3 5.0 x 105 0.0650 0.83 0.85 0.84

by methods outlined in references (I)* and (2). The engine Cylinder bore: 125 mm
used to furnish the data was a turbocharged Dorman Instantaneous compression ratio : variable (peak value,
6LBT which had a bore of 125 mm and stroke of 130 mm. 13.6).
T h e engine conditions used are given in Table 18.1, Inlet manifold air conditions : variable
together with figure references for some corresponding Valumetric efficienty: variable
results. Nozzle hole size: 0.34 mm
Number of nozzles: 4
FACTORS AFFECTING RATE OF HEAT Mean fuel injection pressure: variable
RELEASE FOR THE MODEL Period of injection: variable
T h e following factors, which are discussed separately, are
considered in constructing the model : It is assumed that the swirl and squish have no effect on
the mixing rate, apart from turning the spray to prevent it
(1) ignition delay; hitting the wall of the combustion chamber.
(2) engine dimensions including, for example, nozzle
hole diameter and number of holes; Period of injection
(3) period of injection;
The periods of injection used are taken from the experi-
(4) rate of entrainment of air into the fuel spray;
ments with which comparisons are being made. The
(5) rate of turbulent mixing of fuel and air; and
injection pressure is assumed to be constant during these
(6) rate of reaction of fuel and air based on chemical
periods; this is sufficiently accurate for the purposes of the
kinetic theory.
model. The injection pressure used in the calculations is
No account is taken of the process of droplet evaporation, that necessary to obtain the measured fuel delivery from
and the significance of this factor in relation to the others the nozzle hole configuration.
is currently under investigation. The evaporation effect
is, for the present, absorbed into the other constants Rate of entrainment of air into fuel sprays
describing the model. The shearing action of the fuel spray on the surrounding
Equation numbers refer to equations in the appendices. air causes air to be entrained into the spray without
appreciable loss of total momentum in the fully developed
Ignition delay flow region a few hole diameters downstream of the nozzle.
The ignition delays used are taken from the experiments The rate of entrainment may be calculated [see equations
with which comparisons are being made. (18.1)-(18.9)] by use of the Schweitzer formula for fuel
spray penetration (3) (4) if the entrained air is assumed to
Engine dimensions have the same velocity as the fuel droplets. Modifications
The engine and fuel injection equipment dimensions taken are made to allow for air density changes arising from the
into account at this stage are: piston motion during the development of the fuel spray
(see Appendix 18.1). This is done by assuming that the
Speed: 1200 and 1600 rev/min fuel spray expands radially about its axis to maintain the
Piston stroke: 130 mm pressure inside the spray equal to the pressure outside the
* References are given in Appendix 18.6. spray with no transfer of mass across the spray boundary.
Proc lnstn Mech Engrs 1969-70 Vol 184 Pt 3J

Downloaded from pcp.sagepub.com at The University of Auckland Library on June 4, 2016

194 H. C. GRIGG AND M. H. SYED

Conical plume model Doughnut model

Fig. 18.1. Diagrammatic representation of the t w o air entrainment models

This is consistent with the conservation of momentum in concentrations on the one hand, and mean concentrations
the developed spray. of unmixed air and fuel corresponding to the weak and
The fuel spray was considered in two different forms rich mixture regions. Therefore, for the purposes of this
(see Fig. 18.1); a set of conical plumes issuing one from investigation it is assumed that the regions of weak mixture
each nozzle hole (which is the configuration that Schweitzer consist solely of air, and the regions of rich mixture consist
observed), and an expanding doughnut of spray fed from solely of fuel vapour. The concentrations of fuel or air, or
a central nozzle. mixed fuel and air, are calculated by dividing the total
I n the conical plume model the fuel spray is allowed to mass of the component in the spray by the total instan-
develop unimpeded as in an infinite atmosphere in a conical taneous volume of the spray. The process of turbulent
form [equations (18.2)-(18.5)]. This is justified by recog- mixing reduces the separate air and fuel vapour concen-
nizing that the air motions in an engine cylinder turn the trations and raises the mixed air and fuel concentrations.
spray so that it maintains a conical form but with a bent The turbulent mixing of fuel vapour and air within the
axis. It is found that the air entrained with such a model body of the fuel spray is represented by a diffusion process,
does not exceed the air available in the cylinder until the but in order to simplify the treatment the mean rate of
combustion process is nearly over, hence no difficulties diffusion per unit volume of the spray is equated to the
arise in this connection. concentration of the component (air or fuel vapour) which
After the end of injection the last fuel injected is taken is diffusing, multiplied by a diffusivity coefficient [see
to behave according to the Schweitzer formula [equations equations (18.10) and (18.11)]. T o obtain an agreement
(18.4) and (18.5)], and the spray therefore continues as an with experiment it was found necessary to vary the
expanding truncated cone. diffusivity coefficient as the spray tip velocity, which
The cone angle of the conical plumes is a function of air necessitates the assumption that the burning jet is
density and can be derived from the Schweitzer formula geometrically similar at different times.
[equation (18.3)]. It should be noted that equations (18.10) and (18.11)
I n the doughnut model it is assumed that the doughnut utilize concentrations rather than concentration gradients,
cross-sectional area and circulating velocity relative to the since characteristic distances for gradients cannot be
surrounding air vary to conserve momentum, and that the determined in a simple treatment of the subject. Conse-
rate of entrainment of air, for unit length of doughnut, is quently the diffusivity coefficient has different units from
the same as for a fuel spray of similar cross-sectional area molecular diffusion constants and its value is therefore
and relative velocity to the surrounding air, and having a dependent on the geometry of the spray.
fixed cone angle of 15" [equations (18.6)-(18.9)]. Turbulent mixing is assumed to begin immediately the
The doughnut cross-sectional area is zero at the begin- fuel spray starts, which ensures a premixed flame at the
ning of injection and increases as it is fed with fuel and end of the ignition delay period.
momentum from the central nozzle. At the end of injection
no further momentum is added, but the cross-sectional Rate of burning of fuel based on chemical kinetics
area continues to expand by entraining air (equation The chemical rate of burning of fuel is calculated by an
(18.8)]. Arrhenius type formula [equation (18.12)] (5) in which the
specific rate of burning is proportional to the product of
Rate of turbulent mixing of fuel and air the mean concentrations of turbulently mixed fuel and
Initially, the air entrained over the fuel spray boundary is air, and is a function of temperature. The temperature is
not fully mixed. There are regions of weak mixture and calculated from the pressure and volume in the cylinder
rich mixture within the spray, and for this reason a dis- using the internal energy data in reference (6) and the total
tinction has been made between mixed fuel and air heat released.

Proc lnstn Mech Engrs 1969-70 Vol184 Pt 3J

Downloaded from pcp.sagepub.com at The University of Auckland Library on June 4, 2016

 THE PROBLEM OF PREDICTING RATE OF HEAT RELEASE IN DIESEL ENGINES 195

Fig. 18.2. Typical heat release diagram

The use of mean concentration of mixed fuel and air During the ignition delay period air is entrained into the
implies that the jet is geometrically similar from time to spray and is mixed, to an extent dependent upon engine
time. Since chemical kinetics is a controlling factor only conditions, with fuel. On ignition, owing to the low
over a short period, i.e. during the first peak, this impli- temperature, this air-fuel mixture burns comparatively
cation may not have a serious effect on the theory. slowly under the control of chemical kinetics to form the
first peak of the heat release diagram. The rate of burning
THE MODEL falls as the mixture is being used up, then rises again to
The factors affecting the rate of heat release (see above) the second peak as more fuel is injected and more air is
are used to form a model with the sequence of events entrained. At all times subsequent to the end of the first
illustrated in Fig. 18.2. peak the temperature is high and the chemical kinetic rate

200 21 0

Experiment.
---- Conical plume model (full model).
x Conical plume model (air entrainment only).
Turbocharging ratio, 0.836. Engine speed, 1600 revlmin.
Fig. 18.3. Comparison between experiment and conical plume model for rate of heat release

Proc lnstn Mech Engrs 1969-70 Vol 184 Pt 3J

Downloaded from pcp.sagepub.com at The University of Auckland Library on June 4, 2016

196 H. C. GRIGG AND M. H. SYED

of burning is too high to be a controlling factor, leaving the plume model, and the full heat release calculation for the
control to the rate of entrainment of air and rate of tur- conical plume model. All three are expressed as calorific
bulent mixing. The height of the second peak is largely values. These results cover four manifold densities at each
controlled by the rate of entrainment, but subsequently of two speeds.
the diffusivity becomes low due to its dependence on spray
tip velocity and the control exercised by the turbulent
mixing, producing the long-drawn-out tail to the rate of Rates of entrainment
heat release diagram. The curves of rate of entrainment in Figs 18.3-18.10 may
Finally, as the concentrations fall, the rate of burning is be regarded (since the sprays were fuel-rich up to the end
once again controlled by the chemical kinetics. of injection) as the heat release that would have been
The calculations were carried out numerically by obtained if there had been no ignition delay and the rate of
forward integration on a digital computer at time intervals turbulent mixing and chemical kinetic burning had been
of half a crank-angle degree. infinitely fast.
The rates of entrainment for the doughnut model are
RESULTS similar in most cases to the rates of entrainment for the
Figs 18.3-18.10 show the comparison between the experi- conical plume model; therefore, only two comparisons are
mental rate of heat release diagrams, the calculated rates given in Figs 18.11 and 18.12, but further critical analysis
of entrainment of air into the sprays for the conical is expected to differentiate between them.

Experiment.
- - - - Conical plume model (full model).
x Conical plume model (air entrainment only).
Turbocharging ratio, 1.26. Engine speed, 1604 rev/min.
Fig. 18.4. Comparison between experiment and conical plume model for rate of heat release

190 200 210

CA
Experiment.
- - - - Conical plume model (full model).
x Conical plume model (air entrainment only).
Turbocharging ratio, 1-72. Engine speed, 1602 rev/min.
Fig. 18.5. Comparison between experiment and conical plume model for rate of heat release

160 170 180 oCA 190 200

Experiment.
---- Conical plume model (full model).
x Conical plume model (air entrainment only).
Turbocharging ratio, 2.1. Engine speed, 1606 rev/min.
Fig. 18.6. Comparison between experiment and conical plume model for rate of heat release

Proc lnstn Mech Engrs 1969-70 Vol 184 Pt 3J

Downloaded from pcp.sagepub.com at The University of Auckland Library on June 4, 2016

 \
THE PROBLEM OF PREDICTING RATE OF HEAT RELEASE I N DIESEL ENGINES 197

I I

I
1
I

i n j e c t ion
- I
I8O OCA
L0
*xx xxxx

Expcriment.
_ _ - -Conical plume model (full model).
x Conical plume model (air entrainment only).
Turbocharging ratio, 0.89. Engine speed, 1202 revjmin.
Fig. 18.7. Comparison between experiment and conical plume model for rate of heat release

- __

- 1 I i! i I----i
OCA 190

Experiment.
____ Conical plume model (full model).
x Conical plume model (air entrainment only).
Turbocharging ratio, 1-28. Engine speed 1211 rev/min.
Fig. 18.8. Comparison between experiment and conical plume:model for rate of heat release

I - - - -

End o f

.
160 170 180 OCA 190 200 210

Experiment.
---- Conical plume model (full model).
x Conical plume model (air entrainment only).
Turbocharging ratio, 1.73. Engine speed, 1205 rev/min.
Fig. 18.9. Comparison between experiment and conical plume model for rate of heat release

Proc lnstn Mech Engrs 1969-70 Vol 184 Pt 3J

Downloaded from pcp.sagepub.com at The University of Auckland Library on June 4, 2016

198 H. C. GRIGG AND M. H. SYED

Experiment.
- - - - Conical plume model (full model).
x Conical plume model (air entrainment only).
Turbocharging ratio, 1.81. Engine speed, 1209 rev/min.
Fig. 18.10. Comparison between experiment and conical plume model for rate of heat release

Experiment.
x Conical plume model (air entrainment only).
o Doughnut model (air entrainment only).
Turbocharging ratio, 0.836. Engine speed, 1600 rev/min.
Fig. 18.11. Comparison between experiment and both air entrainment models for rate of heat release

Experiment.
xConical plume model (air entrainment only).
o Doughnut model (air entrainment only).
Turbocharging ratio, 2.1. Engine speed, 1606 revimin.
Fig. 18.12. Comparison between experiment and both air entrainment models for rate of heat release

Proc lnstn Mech Engrs 1969-70 Vol 184 Pt 3J

Downloaded from pcp.sagepub.com at The University of Auckland Library on June 4, 2016

 THE PROBLEM O F PREDICTING RATE OF HEAT RELEASE IN DIESEL ENGINES 199

Experiment.
x Conical plume model (air entrainment only).
Turbocharging ratio, 0.836. Engine speed, 1600 rev/min.
Fig. 18.13. Comparison between experimental cumulative heat release and air entrainment
(conical plume model) for a naturally aspirated case showing lack o f air entrainment

Experiment.
x Conical plume model (air entrainment only).
Turbocharging ratio, 2.1. Engine speed, 1606 revlmin.
Fig. 18.14. Comparison between experimental cumulative heat release and air entrainment
(conical plume model) f o r a turbocharged case showing t h a t sufficient air is entrained

Experiment.
x Conical plume model (air entrainment only).
Turbocharging ratio, 0.89. Engine speed, 1202 rev/min.
Fig. 18.15. Comparison between experimental cumulative heat release and air entrainment
(conical plume model) f o r a naturally aspirated case showing lack o f air entrainment

Proc lnstn Mech Engrs 1969-70 Vol184 P t 3J

Downloaded from pcp.sagepub.com at The University of Auckland Library on June 4, 2016

200 H. C . GRIGG AND M. H. SYED

Experiment.
x
Conical plume model (air entrainment only).
Turbocharging ratio, 1.81. Engine speed, 1209 rev/min.
Fig. 18.16. Comparison between experimental cumulative heat release and air entrainment
(conical plume model) for a turbocharged case showing that sufficient air is entrained

The rate of entrainment of air in the naturally aspirated the sole controlling factor at this point since, in the highly
cases is too slow to account for the heat released during the turbocharged cases at least, there is still some air left over
period of premixed burning but is fast enough in the turbo- from the period of the first peak which would also be
charged cases. Examples of this are shown by the cumula- burning.
tive curves in Figs 18.13-18.16. This discrepancy in the
naturally aspirated cases has not been explained. The full model
The rate of entrainment curves in Figs 18.5, 18.6, Figs 18.3-18.10 show the comparison between the experi-
18.9, and 18.10 are cut off at the point where the mean mental heat release diagrams and the calculated results
fuel/air ratio has fallen to stoichiometric fuel/air ratio. for the conical plume model only.
Turbulent mixing in the spray then becomes the control- Table 18.1 gives the values of the parameters used to
ling factor to produce the tail to the experimental diagram. obtain a fit for both the conical plume and doughnut forms
T h e cut-off for the other cases is off the diagram. of the model. The steric factor is the multiplying term in
I n general, the experimental second peak corresponds the Arrhenius chemical kinetic burning equation. I n the
with the peak of the entrainment diagrams, but this cannot same equation an activation energy of 1.25 x lo8 J/mole
be taken to mean that the rate of entrainment is necessarily (30000 cal/mole) was used throughout. The value of
adjusted diffusivity coefficient that has been tabulated is
typical, i.e. it is the value of diffusivity coefficient at the
end of injection with the units adjusted to those of a
diffusion coefficient by using a characteristic length and
velocity.
In the conical plume version of the model the character-
istic length is the spray length at end of injection, and the
characteristic velocity is the spray tip velocity at the end of
injection. The corresponding quantities for the doughnut
version are the cross-sectional diameter and the circulating
velocity at the end of injection. This adjustment is
described in Appendix 18.5. Other sets of typical values
for the diffusivity coefficient led to similar sets of relative
values between cases.
The values of diffusivity coefficient and steric factor
were found by adjusting the width and height of the first
peak. The remainder of the diagrams are the results
0.6 08 10 12 14 16 1.8 20 obtained with no further adjustment to the two constants.
T U R B O CHARGING RATIO In the naturally aspirated cases, where the air entrain-
Conical plume model. ment was too small, the best fit was obtained for first peak
_ _ - -Doughnut model. height and tail shape.
Fig. 18.17. Steric factor versus turbocharging ratio Referring to the figures, the calculated curves are of the
for conical plume and doughnut models same general shape as the experimental curves. On balance

Proc lnstn Mech Engrs 1969-70 Vol 184 Pt 3J

Downloaded from pcp.sagepub.com at The University of Auckland Library on June 4, 2016

 THE PROBLEM OF PREDICTING RATE OF HEAT RELEASE I N DIESEL ENGINES 201

the conical plume model was closer to the experiment than Schweitzer formula for fuel spray penetration against
the doughnut model (this is not illustrated in the figures). time :
This observation does not necessarily imply the superi- 32-7d,p1/’t
ority of the conical plume model owing to the doubt thrown x’ =
n
. (18.1) . .
upon the comparison by the lack of fuel utilization in the
naturally aspirated cases.
No account has been taken of heat loss in the calcula- Rate of entrainment for conical plume model
tions, but it usually results in an overall lowering of the Assuming momentum continuity, equation (18.1) gives
diagram by 10 per cent. The generally low values for the for the mass rate of air entrainment for air density constant
experimental fuel utilization (Table 18.1) suggest that the with time:
diagrams are in error. I n general the calculated fuel
utilizations are similar except for the conical plume
3
PA
A? = - tan2 8-
at
ax3
. . . (18.2)
naturally aspirated cases where the initial air entrainment
is low and there is a permanent loss of fuel utilization. The where 20 is the cone angle and
doughnut naturally aspirated results suffer from the same
defect, but subsequent excessive heat release compensates
tan2 6 = 0 . 7 5 ~ n . . (18.3)
for this. The high value for the conical plume model in the The air density, pa, should strictly be modified to take
1205 rev/min, 1.73 T / C case is due to an excessively high account of the fact that the plume contains fuel. However,
second peak. The experimental value of the mass of fuel the air/fuel mass ratio at times soon after the start of
burnt has been obtained from the integral of the experi- injection is relatively large, and p a is a sufficiently good
mental heat release diagram. approximation to what is required.
The calculated amount of fuel burnt is obtained by I n order to take into account the variation of air density
integrating the heat release diagrams up to the point at with time due to piston motion, it is assumed that the
which the rate of heat release falls below 0.5 Chu/lb conical plumes expand or contract to maintain the overall
air/”CA. pressure in the combustion chamber uniform, and that the
T h e values used for the steric factors (Table 18.1) fall mass rate of entrainment of air remains the same as it
with increasing turbocharge ratio in much the same would have been had the expansion or contraction not
manner for both speeds (see Fig. 18.17). The values of the taken place.
adjusted diffusivity coefficients show little relationship Equation (18.2) has to be modified after the end of
with turbocharging ratio, but are reasonably consistent injection since the conical plumes become truncated at
between the two speeds. The high values for the naturally their nozzle end. The new form of the equation is then:
aspirated conditions are probably associated with the PA a
failure to obtain sufficient air entrainment in these two A$ = - tan2 8 - ( x 3 - x E 3 )
3 at
. (18.4)
instances.
where
CONCLUSIONS
32.7dnp1’2(t- t E )
For rate of heat release in a diesel engine, this paper has XE2 =
n
. . (18.5)
described two models which are based upon air entrain-
ment in fuel sprays, turbulent air-fuel mixing, and chemi- The term xE can be interpreted as the ‘penetrations’ of the
cal kinetics. truncated ends of the conical plumes near the nozzle. The
The degree of agreement between theory and experiment partial differentials are evaluated from equations (18.1)
is satisfactory, particularly since the first peak was used to or (18.5).
fix the values of the two adjustable constants. The
subsequent degree of agreement in the second peak and Rate of entrainment f o r doughnut model
tail indicates that further work on the present lines is
worthwhile. The cross-sectional radius, by of the doughnut (see Fig.
18.1) is assumed to be zero at the start of injection and then
ACKNOWLEDGEMENTS to expand, and the circulating velocity is assumed to vary
The authors wish to thank the Directors of C.A.V. so that the circulating momentum flux is kept constant and
Limited for supporting the investigation and for per- the momentum is conserved.
mission to publish this paper. They also wish to thank Hence, at all times during and after injection the
Mr B. E. Knight, Chief Research Engineer, for initiating circulating momentum flux is constant, i.e.
and encouraging the investigation. 7rb2paV2= constant . . (18.6)
APPENDIX 18.1 During injection the total momentum in the doughnut
RATE O F E N T R A I N M E N T OF AIR I N T O FUEL SPRAY
must equal the cumulative momentum from the nozzle or
The rate of entrainment of air into the fuel spray, in both 2P
pFm- dNzk- t = 2.rra.rrb2p,V . (18.7)
forms of the model, is based on an approximation to the 4 PP

Proc Instn Mech Engrs 1969-70 Vol 184 Pt 3J

14
Downloaded from pcp.sagepub.com at The University of Auckland Library on June 4, 2016
202 H. C. GRIGG AND M. H. SYED

After the end of injection equation (18.7) takes the form: The temperatures are calculated from the cumulative
77 2P heat released, the thermodynamic properties of the gases
pFm- dN2k- tI = 2rrarrb2pAV . (18.8) (6), and the mass of gas in the whole cylinder.
PF
where tr is the period of injection.
Equations (18.6), (18.7), and (18.8) determine the A P P E N D I X 18.4
behaviour of the doughnut apart from an undetermined EQUATIONS OF CONTINUITY
constant [equation (18.6)]. This constant is determined by Equations of continuity for the gases and fuel in the un-
assuming that the doughnut entrains air at the same rate mixed and mixed states are set up. The continuity of the
as that part of a conical fuel spray, with a 15” cone angle, overall air in the cylinder is ignored, since it is not fully
which has the diameter 26 and velocity V . With this entrained until near the end of the combustion cycle.
condition:
A P P E N D I X 18.5
ADJUSTED D I F F U S I V I T Y C O N S T A N T S
The remarks regarding p A in the previous section also Referring to equations (18.10) and (18.11) the term D. V ,
hold for this model. has units of [T-l]. It is required for the purposes of
comparison between cases that this be expressed in the
A P P E N D I X 18.2 normal units of a diffusion coefficient, that is [L2T-1].
RATE O F T U R B U L E N T M I X I N G O F F U E L A N D A I R .
I n order to achieve this, D V , must be multiplied by the
The rate of mixing of air and fuel is described in terms of square of a characteristic length and V , must be a
the rate of transfer of air and fuel from an unmixed to a characteristic velocity.
mixed state. T h e rates of transfer per unit volume of the
conical plumes or doughnut are: A P P E N D I X 18.6
& f A = D.VTNA . . (18.10) REFERENCES
(I) LYN,W-T. ‘Calculations of the effect of rate of heat release
AXF = D.V,Np . . (18.11). on the shape of cylinder pressure diagram and cycle efficiency’,
where D. V , is the diffusivity constant, proportional to Proc. Auto. Div. Instn mech. Engrs 1960-61 (No.1).
spray tip velocity. (2) AUSTEN,A. E. W. and LYN,W-T. ‘Relation between fuel
injection and heat release in a direct injection engine and the
nature of the combustion processes’, Proc. Auto. Div. Instn
A P P E N D I X 18.3 mech. Engrs 1960-61 o\To. 1).
RATE O F B U R N I N G O F F U E L (3) SCHWEITZER, P. H. ‘Oil sprays’, Auto. Engrs 1938, 61.
T h e rate of burning of fuel is described by an Arrhenius (4) KNIGHT,B. E. ‘Similarity considerations in assessing diesel
type formula for the specific burning rate: engine fuel spray requirements’, Proc. Instn mech. Engrs
1965-66 180 (Pt 3N), 10.
(5) CHAPMAN, S . and COWLING, T. G. The mathematical theory
of non-uniform gases (Cambridge).
J. M. ‘Chart for the investigation of thermo-
xexp (- 8-31x 1 0 3)~ (18.12) (6) GILCHRIST,
dynamic cycles in internal combustion engines and turbines’,
Proc. Instn mech. Engrs 1948 159.

Proc lnstn Mech Engrs 1969-70 Vol184 Pt 3J

Downloaded from pcp.sagepub.com at The University of Auckland Library on June 4, 2016