See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/259334422

Simulation of the Transient Behavior of a 125 CC Scooter with a 1D Commercial

Code

Conference Paper · September 2004

DOI: 10.4271/2004-32-0047

CITATIONS READS
2 204

5 authors, including:

Marco Antonelli
Università di Pisa
63 PUBLICATIONS   331 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Marco Antonelli on 17 December 2013.

The user has requested enhancement of the downloaded file.

 Paper Number 04SETC-29

Simulation of the transient behavior of a 125 cc scooter with a

1D commercial code.
M. Antonelli, L. Nuti, M. Provenzale
Energy Department, University of Pisa

P. Bruno, L. Matteucci
Piaggio Company

Copyright © 2004 SAE International

ABSTRACT integrated simulations, that can take in account the

behavior of both engine and vehicle. By this way, we can
The use of CFD codes is well established in the introduce modifications that are not able to improve in
development of reciprocating engines, owing to the time general sense the curves of torque and power of the
and money saving that can be achieved. The analysis of engine, but are effective in order to fit the engine to the
an engine behavior is usually a quasi-steady type. In vehicle. Such a kind of simulations
fact, with the use of 1D and 3D codes, it is possible to can be carried on by the use of proper CFD codes. One
predict the torque and power output of the engine before of these codes is Wave 5.0, from Ricardo Software. By
it is tested on a bench [1]. However, such an analysis the use of this code, a fluid dynamic, mono dimensional
may result to be not sufficient to predict the effective model of the engine can be built, in association with the
improvements that could be obtained on the transient model of the transmission and the vehicle.
behavior when modifying the intake and exhaust system
of the engine. The aim of this paper is to find a ENGINE AND VEHICLE USED
methodology that is able to quickly optimize the really
important parameters of the engine-transmission For this work was used a four stroke, single cylinder
matching. engine with a displacement of 125 cm3, two intake
In this study, several predictions of acceleration tests of valves and two for exhaust.
a commercial scooter were carried out through the use This engine is connected with a CVT transmission,
of a 1D CFD code. In a close cooperation with Ricardo which operates in combination with a centrifugal clutch.
Software, we implemented a model that simulates the
CVT transmission, that is usually employed in scooters,
taking in consideration the behavior of both the CVT
itself and the centrifugal clutch.
The results of the simulations were compared with the
experimental measures in order to validate the CFD
model, showing a very good agreement. Then, some
modifications were introduced on the intake system of
the engine to improve the response on the acceleration
tests in terms of elapsed time and final speed.

INTRODUCTION

Usually, the standard procedure used when developing

a new engine is the prediction of the curves of torque
and power. A particular attention is to be taken on
optimize such curves in the mostly used range of
rotational speed. The risk of a standard development
engine procedure is wasting time in the optimization of
Figure 1 - Scheme of the transmission
some parameters that result to be ineffective in
improving the desired performances of the entire
vehicle. An useful help may come from
MODEL OF THE ENGINE

COSTRUCTION OF THE MODEL – A model was built

using the standard library of objects available in Wave.
The intake and exhaust system were modeled as
successions of ducts and junctions. The cylinder is
modeled as a variable volume capacity.
The air box and junctions were modeled as zero-
dimensional, constant volume capacities, The model of
the engine is shown in figure 2.

Figure 4 - Torque curve of the original engine

Figure 2 - Original model of the engine

The valves were modeled using the valve lift of the real
engine, introducing the discharge coefficient as a
function of the lift itself, measured through the use of a
fluid dynamic test bench [5]. Atmospheric temperature
and pressure were imposed as boundary conditions for
the elements numbered 301 and 391.

Figure 5 - Engine speed vs time simulation

MODEL OF THE TRASMISSION

The two pre-defined schemes [2] available in Wave

weren’t suitable for describing the CVT of the scooter.
In effects, none of the two models available provides a
CVT gear and the clutch is always before the gear itself.
Figure 3 - Lift law of the intake and exhaust valves
The temperature of the engine walls (head, cylinder
liner, intake and exhaust ducts) were set as dependent
from the engine speed, accordingly to the experimental
data available. Also the combustion modeling provides a
combustion duration dependent from the engine speed.
The duration of the combustion was set on the basis of
experimental data. The carburetor of the engine is
modeled as a duct with an injector. The fuel/air ratio
varies with the engine speed. The fuel is introduced as a
mixture of liquid and vapor, to take in account the effect
of evaporation on the cylinder scavenging.

Figure 6 - Manual transmission model

 Figure 9 - Clutch engagement vs rpm

In the following figure can be seen the shift strategy of

the CVT, where 1 denotes the shorter gear ratio and 2
Figure 7 - Automatic transmission model the longer one.

With a close interaction with the Ricardo Support, it was

possible to create a model with the proper arrangement
of gear, clutch and final transmission. By hand editing, it
was possible to obtain a model which can be
represented in the figure number 9.
Since the CVT varies his gear ratio in function of its
rotational speed, a velocity-dependent shift strategy was
build to match the CVT behaviour.
The engagement of the clutch is a function of the
secondary shaft speed, according to the law reported in
the figure number 8. Figure 10- CVT shift strategy vs rpm

The demand from the driver was simulated by a sudden

increase in throttle position from the position
corresponding to “0” to the position “1” in about 0,5 s.
The demand profile is reported in the following figure.
The value of “1” was not imposed at time “0” in order to
give the necessary numerical stability to the calculus.

Figure 8 - The new model of transmission with CVT

gear

Figure 11 - Demand profile as a function of time

Validation of the model – The model was validated
combining the model of the engine with the model of the
transmission and with the model of the vehicle.
In the following figure is reported the speed of the
scooter in function of the elapsed time, both calculated
and measured.

Figure 12 - Comparison between measured and

calculated speed
The error in the prediction of the performance results to Figure 14 - Difference between steady and unsteady
be about 8% after 2.6 seconds of time simulation, but measured torque
decreases dramatically after just one second to the
value of about 2% and is maintained lower than 1% after At the lowest rpms, the on-road measured torque is
6 seconds of simulation, as showed in the following always lower than the steady-measured one, while at
graph. the highest the engine shows almost the same behavior.
The highlighted point on the graph of figure 15
corresponds to the time at which the throttle is fully
opened (demand = 1).

ATTEMPTS TO IMPROVE THE PERFORMANCE

After the validation of this combined model, it was taken

in consideration to improve the performance on
acceleration test through the variation of the length of
the intake duct, named in the model with the number
“707”. It can be noted that such modification provides a
lower torque at the lowest rpms, while a considerable
gain can be reached at the highest, as stated by many
authors [3], [4], [5], [6], [7].
Figure 13 - Simulation error vs time

COMPARISON BETWEEN STEADY AND ON-

ROAD TEST

On the basis of the results obtained with the first

simulation, several attempts were carried out in order to
improve the transient behavior of the scooter. As a
comparison parameters, the time elapsed to reach 100,
200 and 400 m were considered.
Since in the most part of time during acceleration test
the engine runs at high speed, we attempted to improve
the torque of the engine at the higher rpms.
The unsteady behavior of the engine during the
acceleration test provides a difference between the
torque measured in these conditions and the torque Figure 15 - Torque - Comparison between original
measured at the bench, as shown in the following figure. and shorter intake
This could be related to the difference in gas
temperature between steady and unsteady test, which is
more evident at the lowest rpms.
 This modification provides a faster acceleration in the
first 50 meters, but the result on the 400m is almost the
same than with a simple shorter intake duct.
Here some doubts arise about the effectiveness of such
a system, which is relatively complicated. In the
following graph the gain using variable geometry can be
seen respect to the original engine.

Figure 16 - Performance comparison between

original and shorter intake
Despite this, to avoid excessive difficulties starting on a
climb, the use of a resonator [3], [4], [8] was
hypothesized, in order to compensate the lack of torque
at the lowest rpms. The resonator is activated through a
system of valves, which is actuated in function of the
number of revolutions per minute of the engine, as
provided by the following figure. In this particular case,
the valves of the resonator switch at about 5500 rpms. Figure 19 - Speed gain using resonator
Another matter of interest is the sensitivity to the
carburetor diameter. The carburetor was modeled as a
constant diameter duct, with a diameter giving the same
wave effects of the real carburetor. This method was
necessary, owing to the three dimensional phenomena
happening in such component. It was found that
reducing the diameter of the throat of the Venturi leads
to a strengthening of the amplitude of the waves, thus
enhancing the torque curve at the higher rpms, as can
be seen in the following graph.
Figure 17 - Engine with the resonator
The use of the resonator enhances the torque at the
highest rpms, without sensible losses at the lowest ones.
In other words, the presence of the resonator shifts the
torque curve towards higher rpms, with an improvement
in the peak value.

Figure 20 - Pressure waves in the intake duct

Moreover, the comparison between the two torque
curves are shown, in figure 22. The gain in torque is very
evident from 7000 rpm since to the maximum rpm, with
a decrease at the lowest speeds.

Figure 18 - Comparison between original intake and

resonator
 The mono-dimensional code used showed a good
agreement, with an error in the zone of main interest
lower than 3%, with the experimental data used for the
validation. At the same time, the code proved to be very
fast in evaluating the effectiveness of the solutions
proposed. As a conclusion, it may be said that it
constitutes a really useful aid for the engineers.

ACKNOWLEDGMENTS

[1] G.P. Blair, D. O. Mackey, M. C. Ashe, G. F.

Chatfield, “Exhaust Tuning on a Four Stroke Engine;
Experimentation and Simulation”, SAE Paper 2001 01
Figure 21 - Variations of torque with a lower throat
1897/4218
diameter
Owing to the approximations in the modeling of the [2] Wave Input Blocks Manual, Ricardo Software.
carburetor, this results means that a deeper study on the
carburetor has to be done, because of the sensitivity of [3] G. Ferrari, “Motori a Combustione Interna'', ed. Il
the simulation to the equivalent diameter of the Capitello, Torino 1992
carburetor. The improvement of performances using a
lower throat diameter has to be proven, both with [4] M Badami, F. Millo, G. Giaffreda, “Experimental and
experimental and three-dimensional simulations. Computational Analysis of a High Performance Four-
In the following graph are reported the speed gains Stroke Motorcycle Engine Equipped with a Variable
using, respectively, a shorter intake and a shorter intake Geometry Exhaust System”, SAE Paper 2002-01-0001
together with a resonator.
[5] J. B. Heywood, “Internal Combustion Engines
Fundamentals'', Mc Graw Hill, 1988

[6] G. P. Blair, ”Design and Simulation of Four Stroke

Engines”, edizioni SAE, 1999

[7] R.S. Benson, ”The Thermodynamics and Gas

Dinamics of Internal Combustion Engines”, 1982

[8] F. J. Laimböck, R. S. Kirchberger, “Development of

a 150 cc, 4-Valve CVT Engine for Future Emission and
Noise Limits”, SAE Paper 982052, 1998

Figure 22 - Speed gain using different techniques

CONCLUSION

In this work we presented the study of the whole system

engine-scooter. Firstly, the torque curve of the engine
was predicted and validated through the comparison
with the experimental one. After the prediction of the
torque in steady conditions, a transient simulation of an
acceleration test was done and compared with the
telemetric measures. Finally, some analysis were done
to improve the performance on such test using some
different techniques. This work suggest to perform other
studies on the behavior of the carburetor owing to the
three-dimensional effects that cannot be evaluated with
a mono-dimensional code.
CONTACT

Marco Antonelli
Mechanical Engineer, PhD, MSAE
Tel. 050 2217133
Address: Via Diotisalvi, 2
56126 PISA
e-mail [email protected]

Luca Nuti
Mechanical Engineer, PhD, MSAE
Tel. 050 2217133
Address: Via Diotisalvi, 2
56126 PISA
e-mail [email protected]

Michele Provenzale
Mechanical Engineer, PhD, MSAE
Tel. 050 2217132
Address: Via Diotisalvi, 2
56126 PISA
e-mail [email protected]

Pellegrino Bruno
Mechanical Engineer, PhD
R&D and Operations 2 wheels
ENGINE DESIGN
Piaggio & C. S.p.a.
Tel. 0587 272324
Address: V.le R. Piaggio, 23
56025 Pontedera (PISA)
e-mail [email protected]

Luigi Matteucci
Mechanical Engineer, PhD
R&D and Operations 2 wheels
ENGINE DESIGN
Piaggio & C. S.p.a.
Tel. 0587 272324
Address: V.le R. Piaggio, 23
56025 Pontedera (PISA)
e-mail [email protected]

View publication stats