J. ugric. Engng Res.

(1990) 47, 165-176

A Universal Mathematical Model of Diesel

Engine Performance
H. D. HARRIS;* F. PEARCE?

l Rimik Pty Ltd, Toowoomba, Queensland,Australia

t University College of Southern Queensland, Toowoomba, Queensland,Australia
(Received 12 September 1989; accepted in revised form 28 January 1990)

At any throttle setting, and as torque is increased, the speed of a diesel engine will
decrease from its value at zero torque. The difference between the speed at zero torque and
the current speed is the engine speed depression, which increases until the fuel pump is at full
stroke and delivering “full fuel”. If torque increases further, the engine then operates along
the locus of these points where the fuel delivery is a maximum, referred to as the “full fuel
line”.
A mathematical model of the performance of a governed diesel engine has been developed
using these concepts. This model expresses torque, power and fuel consumption as universal
functions of engine speed at zero torque and the depression of engine speed below the engine
speed at zero torque. Each parameter is described by a function involving nine coefficients,
which are established by p.t.o. testing of a particular engine. The form of the functions allows
them to meet the boundary conditions of engine operation. The model has been validated by
applying it to three engines, where it gives accurate predictions of torque, power and fuel
consumption. It is also shown that the full fuel line can be accurately represented by
expressing the engine speed at its intersection with a constant throttle (or governor) line as a
polynomial function of the zero torque (or idle) speed represented by that governor line. The
boundaries of engine operation are then this full fuel line, the maximum and minimum
governor lines, and the zero torque axis.
This model of engine performance is in a form which enables its use in the modelling of
tractor performance and the indirect measurement of engine performance.

1. Introduction
A number of mathematical models of tractor engine performance have been proposed
in recent years as part of the continuing interest in optimizing tractor performance and in
the modelling of farming operations. These models have typically been based on a least
squares fit of a polynomial expression, involving some of the parameters engine speed,
throttle position, duration of injector needle lift, and exhaust gas temperature. They have
been able to predict torque or power and possibly fuel consumption, but they have all
been difficult to apply generally because they have been based on a large number of
measurements, or have used parameters which are specific to a particular installation.
There has also been no assurance that the same form of expression would be applicable to
any other engine, and they have suffered from the basic deficiency that they have not met
the elementary boundary conditions of engine operation, namely that zero power is produced
at zero engine speed, that zero torque is produced at any idle speed, and that fuel is
consumed at any idle speed even when output torque is zero.
The only model which has attempted to address any one of these boundary conditions
is that of Jahns,’ and the present work used the Jahns model as a starting point for the
development of an engine performance model which is universally applicable, meets the
necessary boundary conditions, is based on a minimum data set, properly defines the full
165
 166 MATHEMATICAL MODEL OF DIESEL ENGINE PERFORMANCE

Notation

a,, b,, c,, A, hm un, vn coefficients defined in the text

1
aij, bij, cij coefficients of the power, fuel consumption and torque
models, respectively
dij coefficients of a fuel consumption model proposed by Jahns’
F fuel consumption, ml/s
G mass of fuel injected/cycle, kg
N engine speed, rev/min
N engine speed at the intersection of the full fuel line and a
governor line
D injector needle lift duration, ms
P power, kW
R engine speed depression below Z( = Z - N), rev/min
Rf engine speed depression at the intersection of the full fuel line
and a governor line
S position of governor control shaft, counts
T torque, Nm
Z engine speed at zero torque (idle speed), rev/min
AT exhaust gas temperature difference above ambient, “C

fuel line, uses engine parameters which are easily measured and are not specific to a
particular installation, and is accurate. This model also allows the calculation of contours
of constant power, constant fuel consumption and constant torque.

2. Previous proposals
Previous workers have used a variety of engine parameters to represent engine power,
torque and fuel consumption. For consistency, the expressions that they have derived
have been rewritten in the variables used here.
Koertner et al.* used engine speed(N), idle speed(Z) and control shaft position(S) to
correlate measurements of power(P) and fuel consumption(F). P and F were found to be
functions of S, N, Z, and S x N.
Chancellor and Smith’ used a rotary potentiometer to measure throttle position, and
expressed engine torque(T) as
T=d+e. s+f. N+g. a,
where 6 = 0 or -1, depending on the direction in which the potentiometer is moving, and
Q, is the potentiometer reading. The speed N is divided into five ranges, and d, e, f and g
are constants which depend on the range within which the engine speed falls.
They also noted that torque may be a function of the engine speed depression R, but
did not pursue the possible correlation. Their expression for T is not universal in form
because it depends on the installation and it does not meet the boundary conditions. Fuel
consumption was not predicted.
Schrock ef aL4 established correlations for power(P) and fuel consumption(F) using
engine speed(N) and control shaft position(S). Their expressions were of the form
P = a, + a,N + a,S + a3N2 + a,S. N
F = b. + b, N + b2P + b3N2b,S2 + b$ . N
H. D. HARRIS; F. PEARCE 167

S is defined as an encoder count, with no definition of the zero position, so it is a variable

which depends on the particular installation. The expression for P does not meet the
necessary boundary conditions, because it is not zero at any idle speed or when the speed
is zero. The full fuel line was defined by an expression for maximum S in terms of P and
N.
McKiernan et ~1.~ investigated the use of exhaust gas temperature (AT) and injector
needle lift duration(D) for predicting power and fuel consumption. While they found that
AT responded too slowly to changed conditions and did not yield accurate correlations,
the correlations based on D were accurate. The expressions derived were
T=a,+a,N+a,D +a,N. D +a4N@5
and
F=b,+b,N+b,D+b,N. D
The torque relationship does not yield T = 0 at D = 0, as would be expected. Similarly, F
is not equal to zero at D = 0.
de Souza and Milanez’ considered three methods for indirectly predicting the torque of
diesel engines, expressing torque as functions of fuel consumption, exhaust gas tempera-
ture and engine control lever position measured by engine idle speed. The functions used
were based on a general form involving the particular variable and N, and using ten
coefficients. The expressions were simplified by adding terms until the accuracy of the
correlation improved by a negligible amount.
Writing G = mass of fuel injected per cycle, they found
T = a, + a,G + a,G2 + a,N + a4N2
This relationship does not yield T = 0 at either G = 0 or N = 2.
For the correlation based on exhaust gas temperature, they found
T = b. + b,AT + b2AT2 + b3NAT + b,N2
The expected boundary conditions here would be T = 0 at either AT = 0 or N = 2, and
these are not met.
The correlation based on engine idle speed (2) was in two parts, for either the full fuel
line or for the governed range. This is the only correlation method among those so far
described which deals explicitly with operation along the full fuel line. de Souza and
Milanez6 gave
T = do + d,N + d2N2 + d3N3
on the full fuel line, and
T = h, + h, N + h2Z + h3Z2
in the governed range. Again the boundary condition that T = 0 at N = Z is not satisfied.
The only author among those surveyed who attempted to represent engine performance
using an expression based on physical reasoning and which met the necessary boundary
conditions, was Jahns.’ He proposed that there should be an expression of universal form
with coefficients specific to each type of engine, which are calculable from tests performed
on these engines. He then modelled fuel consumption with an equation whose structure
was determined by the boundary condition that fuel consumption > 0 when T = 0, and by
the assumption that a polynomial provides a sufficient representation. His equation was
2 3
F = d,,N + d,,N + d,3N
+ (dzlN + d22N2 + d23N3)T
+ (d3,N + d,,N’ + d33N3)T2
168 MATHEMATICAL MODEL OF DIESEL ENGINE PERFORMANCE

The nine coefficients, dij, involved in this expression are found from the values of engine
speed(N) and torque(T) at nine points evenly distributed on the performance map.
Jahns also represented the full fuel line and the full throttle line, respectively by
T = u, + u,N + u,N* + u4N3 + u5N6
and
T=vI+v2N+v3N2
The coefficients Ui and Vi can be found from eight further test points on these boundary
lines. The expression for the full throttle line does not meet the boundary condition that
T = 0 at full idle speed.
It can be seen from the above discussion that previous work has been deficient in the
prediction of either torque or fuel consumption, that the expressions proposed have not
been universally applicable and, with the exception of the work of Jahns, have not been
able to meet the necessary boundary conditions.

3. Objectives
The main objective of this work was to develop an engine model which predicts fuel
consumption, torque and engine power using simply measured independent variables.
Furthermore, it was required that the model developed should be universally
applicable, incorporating coefficients specific to each engine type. The amount of data
required to establish these coefficients was to be minimized so that the work involved in
the p.t.o. testing of tractors would be reduced.
Jahns’ suggested that because tractor performance would vary with operation time and
engine maintenance, and between nominally identical engines of the same type, an
accuracy of f5% was adequate for this type of modelling. The same accuracy criterion
was set here.

4. Tbe model
The torque-speed performance space of a governed diesel engine is seen as being
bounded by the zero torque line (speed axis), the full throttle governor line, and full fuel
line and a minimum throttle governor line. As shown in Fig. 1, the full throttle line and

Full fuel line

Full
0
throttle
line

Fig. 1. Definitions of engine parameters in T - N and R - Z space

H. D. HARRIS; F. PEARCE 169

the full fuel line intersect at the rated point. It is assumed that engines of the same type
will have the same full fuel line extending from the point of maximum torque to beyond
the rated point. At the same time, engines of the same type may vary in that the full
throttle line may be located either to the right or left of the given line.
It is also assumed that any governor line may be defined by the value of engine speed at
zero torque (or the idle speed) which we denote by 2. The full throttle idle speed is
therefore Z,, and there is a minimum throttle idle speed which we may denote Zmin.
Preliminary work using data from intensive p.t.o. testing of a number of tractors showed
that along any governor line characterized by 2 the torque and fuel consumption are
simple functions of the depression of engine speed below Z, denoted R. This suggested
that a model should be sought which used Z and R as its independent variables,
particularly as the simply measured engine speed N is related to Z and R by the
relationship N = Z - R. It was also found that Z is a function of the throttle position,
which can be measured by simple potentiometric arrangements.
Fig. I also shows how the torque-speed space has now been replaced by the R - Z
space. The full throttle governor line has become the line Z = Z,,, the full fuel line has
become the line R =f(Z), and the minimum throttle line has become Z = Zmin. The
engine speed axis is now R = 0. It is proposed that engine power torque and fuel
consumption should be defined in this space as functions of Z and R.

4.1. Definition of model functions

Following the approach taken by Jahns,’ who structured his relationship for fuel
consumption on physical reasoning, we have written
Fuel consumption F= (bllZ + b,2Z2 + b13Z3)
+ (b2,Z + b2*Z2 + &Z”) . R
+ (b3,Z + &*Z* + b33Z3). R* ml/s (1)
= b, + b2 . R + b3 . R* ml/s at constant Z (2)
This expression meets the boundary condition that fuel consumption = f (Z) at R = 0, i.e.
fuel is used to overcome internal losses when output torque is zero. It also shows a
quadratic variation of F with R at constant Z, which was found to be an adequate
representation in our preliminary work.
A similar form was then used in proposing an expression for power, namely
Power P = (a,,R + a,*R* + alJR3)
+ (a*1 R + u2*R2 + u,,R3) . Z
+ (~31 R + ~32R* + a33R3) . Z* kW (3)
=u,+u,.Z+u,.Z*kWatconstantR (4)
This relationship meets the boundary condition
Power = 0 at R = 0 (i.e. at zero torque).
Torque can then be calculated knowing the engine speed, N.
An alternative approach is to use a similar expression for torque, namely
Torque T= (c,,R + c,*R* + c1,R3)
+ (CZAR + ~22R* + c23R3). Z
+ (~31 R + ~32R* + c33R3) . Z* Nm (5)
= c, + c2 . Z + c3 . Z* kW at constant R (6)
170 MATHEMATICAL MODEL OF DIESEL ENGINE PERFORMANCE

This relationship meets the necessary boundary condition

Torque=0 at R =0
The coefficients Uij, bij and Cij can be evaluated from nine points taken in a standard p.t.o.
test, or by using a multilinear fit technique they can be adjusted to yield the best fit to a
larger number of points. These points preferably include the values at R = 0 and at the
intersection of the governor (constant 2) lines with the full fuel line, and should be evenly
distributed across the operating range of the engine. Five governor lines with five points
on each is a suitable compromise and yields an accurate fit when used to evaluate the nine
coefficients for each variable. Typical values of the correlation coefficient, r*, for the fit of
these expressions to measured values of fuel consumption, power and torque are in the
range of 0.99 to 0.999.

4.2. Model boundaries

The boundaries of the model are simply defined in R - Z space (as shown in Fig. 1) by
z=z,, and Z = Zmin
and
R=O and R=f(Z),
which is the transformed full fuel line.
It has been found convenient to define the full fuel line by
Nf = do + d,Z + d2Z2 + d,Z3 (7)
Nr is the value of N at the intersection with the full fuel line of the governor line defined
by Z. This cubic representation yields very accurate fits to the data from a number of
tractors, with values of r* greater than O-99. It can also be used to limit the value of R
appropriate to a given Z by writing
Rf=Z-Nf
If R > Rf, then the engine is operating along the full fuel line.

4.3. Calculation of contours

It is possible to use Eqns (l-6) to calculate lines of constant fuel consumption, power
and torque in the R - Z space, which can then be simply transformed to the F - N,
T - N or P - N space by using N = Z - R.
From Eqn (2), for a given Z, we can derive an expression for the line of constant fuel
consumption F,
R = (-b2 f db; - 4b3(b, - F))/2b, (8)
where bl, b2 and b3 are all functions of Z.
From Eqn (4), for a given R, we can derive an expression for the line of constant power
p,
z = (-a* f l/a: - 4a3(q - P))/2u, (9)
where a,, a2 and a3 are all functions of R.
From Eqn (6), for a given R, we can derive an expression for the line of constant
torque T,
Z = (-c2 f Vc: - 4c3(c1 - T))/2c3 (10)
where cl, c2 and c3 are all functions of R.
H. D. HARRIS; F. PEARCE 171

5. Validation of the model

One of the requirements of this model was that it should be universally applicable
across all tractor engines, with only the coefficients being changed. Accordingly, evidence
is reported here of the application of Eqns (l), (3), (5) and (7) to three different engines.
Data were measured using a standard brake type p.t.o. dynamometer, which had been
modified to allow accurate measurement of torque and speed using a computer controlled
data acquisition system. Fuel flow rate was measured using “Micro-Oval” positive
displacement flow transducers, which had been modified to give five pulses per revolution
of the rotors as compared with the one pulse per revolution of the standard transducers.
A transducer was placed in each of the main and the return flows, and the nett flow was’
taken as their difference.
At each throttle setting, the value of 2 was measured by disconnecting the p.t.o. and
measuring the engine speed with a hand held tachometer. The p.t.o. was then engaged
and four more data sets were acquired at that throttle setting until the full fuel line was
reached.
The data were then processed using a commercial software package to create the
functions defined in Eqns (l), (3) and (5), and calculate the coefficients of the multilinear
fits, Uij, bij and ci,. These coefficients were then used to calculate values of fuel
consumption, power and torque for comparison with the measured values.
Data for 2 and Nr were also processed to yield the coefficients di of the polynomial fit,
Eqn (7). This polynomial was used to limit the values of N used in the calculation of the
comparisons.
Data from three tractors have been processed in this way, and the results have been
plotted to show how accurately the models will fit the data, and how the models may be
used to predict fuel consumption and torque throughout their range of applicability.

5.1. Ford 7W5

These coefficients were used to calculate the torque-speed and the fuel consumption-
speed curves shown in Fig. 2, together with the data points used to calculate the
coefficients. Values of torque derived using N and the power calculated via the power
coefficients are practically identical with those calculated from the torque coefficients, and
so they have not been plotted separately. Values for the goodness of fit are shown in
Table 2.
Table 1
Coefficients of the model for the Ford TWS

Coefficients (calculated from 25 points)

Power aI1 -0.870 aI2 2.93 x 1O-3 aI3 2.76 x 1O-6

azI 8.22 x 10m4 az2 - 1.46 x lo-” a23 -6.67 x lOmy
a3( -4.39 x 10-n -5.61 x 10-l” a33 3.49 x lo-”
Fuel consumption b,, 5.63 x 10-j ::I 8.25 x lo-’ b 13 4.76 x 10-l’
b 2, 1.73 x 1O-5 b,? - 1.16 x lo-’ b 4.43 x lo- I2
b 31 -3.06 x lo-’ b,, 4.34 x 10-l’ b: -1.66 x lo-‘J
Torque CII -2.77 Cl2 0.012 Cl3 1.65 x lo-’
C2I 3.95 x 10-3 C22 -9.52 x W6 C2.3 -2.69 x lo-’
C3l -5.57 x 10-7 C32 -6.54 x lo-*’ c33 1.29 x 10-l’
-5.75
Full fuel CO -70.5 Cl 1.54 C2 -2.71 x lo-’
C3 5.28 x lo-”

Range of Z 1680 to 2600 rev/min

172 MATHEMATICAL MODEL OF DIESEL ENGINE PERFORMANCE

9r

a 5r
E
2
El
u 4
I=5 3-

2-

l-

0
1200
Engine speed. revlmin

Fig. 2. Comparison of torque and fuel consumption models for Ford 7’WS (25 data points). 0, data
points on which the torque model is based; - - -, predictions of the torque model; 0, data points on
which the fuel consumption model b based; -, predictions of the fuel consumption model

It may be seen that the model fits the measured points very well, and that torque and
fuel consumption curves for values of 2 (such as 2470 and 1815) between those used for
deriving the coefficients appear to be plausible. In developing this approach it was found
to be necessary to include data points as near as possible to the intersection of the
governor line and the full fuel line in order for the predicted constant 2 line (or governor
line) to actually reach the full fuel line. It is also necessary to include a governor line at
each end of the speed range because predictions outside this range can be in error.
When the coefficients are calculated from 15 points rather than 25, at 2 = 2601, 2154
and 1683, the predicted torques for other values of 2 are not accurate. Where other
measured points are available, at Z of 2319 and 1963, the predicted torque is in error in
the region of the full fuel line, as shown in Fig. 3. It appears from this that data from
more than three governor lines are necessary to properly model engine torque across its
entire range of speed.
In contrast, Fig. 3 also shows that the prediction of fuel consumption using 15 points to
calculate the coefficients yields results comparable with those achieved using 25 points.

5.2. Case 2090

Only four governor lines were measured for this engine, at Z values of 2280, 2080, 1868
and 1642 rev/min. The coefficients have therefore been calculated from 20 data points,
Table 2
Goodness of fit for the Ford TW5

Correlation coefficient Maximum error

Power 0.9919 24 kW
Fuel consumption 0.9995 0.2 ml/s
Torque 0.9936 10 Nm
H. D. HARRIS; F. PEARCE 173

Engine speed,revlmln

Fig. 3. Comparison of torque and fuel consumption models for Ford 7W5 (15 data points). 0, data
pointi on which the torque model is based; A, other torque data points; - - -, predictions of the
torque model; 0, data points on which the fuel consumption model is based; A, other fuel
consumption data points; -, predictions of the fuel consumption model

Table 3
Coefficients of the model for the Case 2O!Mt

Coefficients (based on 20 points)

Power aI1 -1.56 -2.73 x 1O-3 aI3 4.85 x lo-’
*2! 1.39 x 10-j 7.92 x 1O-6 a= -6.45 x lo-*
*3t 1.75 x lo-’ 032 -344 x 10-9 a33 1.99 x 10-l’
Fuel consumption b,, 2.78 x 1O-3 b,, -2.53 x lo-’ b,, 8.18 x 10-10
b 21 -1.62 x 1O-4 b,, 1.76 x lo-’ b, -4.38 x 10-l’
b 31 103 x 1O-6 b,, -1XI2 x lo-’ b,, 2.56 x lo-l3
Torque CII -846 Cl2 1.96 x 1O-2 Cl3 1.66 x to-4
C21 8.71 x 1O-3 CZ2 6.86 x 10-6 CL3 -2.47 x lo-’
C3l -1.62 x 1O-.6 c32 -864 x lo-’ c33 8.02 x lo-”
Full fuel CO 3319 Cl -4.18 C2 2.48 x 1O-3
Cl -3.89 x lo-’

Range of 2 2280 to 1640 rev/min

Table 4
Goodness of fit for the Case 2MO

Correlation coeficient Maximum error

Power 0.9988 2.8 kW
Fuel consumption 0996 0.16 ml/set
Torque 0.9991 14 Nm
174 MATHEMATICAL MODEL OF DIESEL ENGINE PERFORMANCE

Engine speed, rcvlmin

Fig. 4. Comparison of torque and fuel consumption modeLri for Case 2090 (20 data points). Same key
as in Fig. 2

but the predictions of torque and fuel consumption match the measured points very well,
as can be seen in Fig. 4. Calculations have also been made at Z values of 2180, 1990 and
1755 t-pm, and these curves appear to be plausible representations of torque and fuel
consumption.

5.3. John Deere 3140

The torque and fuel consumption curves are shown in Fig. 5. Other data were available
at Z values of 1762 and 1604 r-pm, and these have also been plotted along with the model
Table 5
Coefficients of the model for John Deere 3140

Coefficients (based on 25 points)

Power alI 6.88 aI2 -5.04 x 10-z aI3 4.67 x lo-’

h -6.48 x lo-” 4, 4.29 x 1O-5 a23 -2.42 x lo-”
031 1.51 x 10-6 -8.35 x lo- a33 -6.20 x lo-l3
Fuel consumption b,, 5.91 x 10-4 ab9z -2.59 x lo-’ b,, 1.45 x lo-“’
b 21 5.01 x lo-’ 6;; -4.99 x lo-’ bz3 1.30 x lo-”
b 31 -8.55 x lo-” bxz 8.05 x 10-l’ b,, -1.51 x IO-l4
Torque Cl1 24.66 Cl2 -0.200 Cl3 2.27 x 1o-4
51 -0.233 c22 1.75 x 10-4 c23 - 1.43 x 10-7
C3l 544 x 1o-6 c32 -3.49 x 10-x c33 1.07 x 10-I’
Full fuel CU - 1984 Cl 2.87 C? -7.05 x 1o-4
C3 1dO x lo-’

Range of Z 2563 to 1918 rev/min

H. D. HARRIS: F. PEARCE 175

Table 6
Goodness of tit for Jobn Deere 3140

Correlation coefficient Maximum error

Power 09986 2.0 kW
Fuel consumption 0987 0.2 ml/set
Torque 0.9986 12 Nm

predictions. At 2 = 1762, the torque is predicted adequately, but at Z = 1604 the

predicted torque is not accurate.
Torque and fuel consumption curves have also been calculated for Z values of 2490,
2350, 2160 and 1980 rev/min, and they show that the model can interpolate between the
data points. All these curves intersect the full fuel line at the expected values of torque
and fuel consumption, while the fuel consumption values at zero torque also lie on the
known curve.
When 15 points are used to calculate the coefficients, it is possible to compare the
predictions of the model with the data on the two other governor lines. The 15 points
chosen were on the governor lines for Z values of 2563, 2080 and 1918. Fig. 6 shows the
calculated results for Z = 2415 and 2057. While the values of torque and fuel
consumption are predicted reasonably well, the intersection of these curves with the full
fuel curve is inaccurate. Again, it seems desirable to use more than 15 points in
calculating the coefficients of the model.

6. Conclusions
A mathematical model for the prediction of fuel consumption, torque and power of a
diesel engine has been proposed. This model is based on a universal form involving nine
coefficients for each of these parameters, and improves on previous proposals in that it

400 8

-7

E 3001 c

-5 $
z

\ -4 P
\ :
\ . -3 E
\

I -2 j
- 4
-1

0 I l .1 -\ \ I ‘.-\ I\.- ‘\ , \- .\ ). , .26:O,

1200 1400 1600 1800 2000 2200 2400
Engine speed, revlmin

Fig. 5. Comparison of torque and fuel consumption models for John Deere 3140 (25 data points). 0,
data points on which the torque model is based; A, other torque data points; - - -, predictions of the
torque model; 0, data points on which the fuel consumption model is based; -, predictions of the
fuel consumption model
176 MATHEMATICAL MODEL OF DIESEL ENGINE PERFORMANCE

400 -0

E 300

iI:: ~~~:i:

b ’ -1
\ \ \
\ \ \ \
\ \ \&
0 I I I \_ 11. L
1200 1400 1600 1800 2000 2200 2400 26C%
Engine speed, revlmin
Fig. 6. Comparison of torque and fuel consumption modeLF for John Deere 3140 (15 data points).
Same key as in Fig. 4

meets the necessary boundary conditions of engine performance, is accurate, and can be
applied to any engine. The engine variables used by the model are the engine speed at
zero torque and the rev/min depression below this speed, which are both easily
measured.
The nine coefficients for each parameter for a given engine are preferably calculated
from a minimum of 20 points on four governor lines, using a multilinear fit technique.
The five points on each governor line should include the data at zero torque and the data
at the intersection of the governor line with the full fuel line.
The full fuel line can be represented by a third order polynomial between engine
rev/min at the governor line/full fuel line intersection and the value of engine rev/min at
zero torque for that governor line.

Acknowledgements
This work was supported by way of a grant under the Australian Government’s National Energy
Research, Development and Demonstration Program.

References
’ Jahns, G. A Method of Describing Diesel Engine Performance Maps. ASAE Paper NCR 83-103,
1983, ASAE St Joseph, MI
’ Koertner, R.; B&ford, L. L.; Lane D. E. Tractor Instrumentation for Measuring Fuel and
Energy Requirements. ASAE Paper 75-1547, 1975, ASAE St Joseph, MI
’ ChanceUor, W. J.; Smith, N. E. Tractor Engine Torque Transducer Using Throttle Position and
RPM. ASAE Paper 85-1557, 1985, ASAE St Joseph, MI
4 S&rock, M.; Matteson, D.; Blumanhourst, M.; Thompson, J. G. A Device for Aiding Gear
Selection in Agricultural Tractors. Transactions of the ASAE 1986, 29: 1232-1236
’ McKieruan, M.; Lalk, T. R.; Stout, B. A.; Searcy, S. W. Estimating Diesel Engine Performance
By Indirect Methods. SAE Technical Paper Series 871606, 1987
’ de Souza E. G.; Milanez, L. F. Indirect Evaluation of the Torque of Diesel Engines.
Transactions of the ASAE 1988, 31: 1350-1354