Mathematical modelling and fuzzy based evaluation of a

specific engine-CVT transmission combination

Zsolt Farkas1, Attila Piros1, Tibor Bercsey1, István J. Jóri1

1
Budapest University of Technology and Economics,
H-1111 Budapest, Bertalan Lajos u. 3. MG, [email protected]

Abstract:
This paper introduces a new evaluation method of a specific part of the power
train. In this case study was done a fuzzy evaluation using a virtual
combination of a commercial high performance tractor engine and power split
hydrostatic CVT (Continuously Variable Transmission) transmission published.
In the first part of the study the mathematical model of this power train is
described. The engine consumption is calculated from a 3-dimensional
characteristic diagram. This diagram is interpolated of the power measurement
data. The transmission efficiency is calculated from a special mathematical
model. The object of this study was the evaluation of the power train in some
specific phases. Mostly used tractor speed ranges were generated by different
engine speed and transmission ratio combinations. In each case the engine
consumption and the transmission efficiency were calculated based on this
mathematical model. These results were evaluated with the method of corrected
fuzzy mean. All of these calculations were performed in MATLAB. This study
gives a design guide for the future processes of the engine-transmission
matching as a result.

1 Introduction
The tractor is one of the most important machine in the agricultural mechanization
and its transmission system is a key component representing about 30% of the total
tractor first cost. Stepped transmissions have been used since the beginning. Since
late 1995, when the high-powered (194 kW) Fendt Favorit 926 equipped with a
”Vario” transmission was launched, power machines with CVT (Continuously
Variable Transmission) have appeared on the palette of all major tractor
manufacturing companies as well. The CVT unit in the transmission of these power
machines is not included directly in the power flow, but in one of the power branches
after a split. By modifying the ratio gear of a properly selected CVT unit, the
transmission can be geared neutral and the direction of progress of the vehicle can be
changed. The transmission structure produced so is called power split IVT (Infinitely
Variable Transmission). Infinitely variable drives with automatic controls are a new
area of power train design.
This paper introduces a fuzzy based evaluation of an engine – CVT combination.
The evaluation is based on the investigation of engine of Case IH 195 Puma tractor
and Steyr S-Matic transmission.
2 The tested power train elements
The Case IH 195 Puma tractor is built with high performance engine to deliver a
reliable source of power. High pressure common rail diesel technology combined
with electronic fuel management ensures cost-effective operation for all farming
applications. To ensure both power and efficiency, the tractors are equipped with a
four-valve, turbocharged, intercoolered, low-emission diesel engine.
Mobile PTO (Power take-off) brake-test provided the engine characteristics of
vehicle. We used Sigma 5 Dynamometer (Serial No.:022346). During the test there
were four parameters recorded: engine speed [1/min], loading torque [Nm], specific
fuel consumption of the engine [g/kWh] and the throttle position [%] (Figure 1.).
These measurement data were integrated in the mathematical model of the evaluation.

Figure 1. Engine characteristic

The S-Matic transmission (Figure 2.) was tested by us [1,2], is working on the
basis of a hydrostatic-mechanical power splitting. The continuously variable
transmission modulation is accomplished by the variable displacement pump, which
is back to back assembled to the fixed displacement motor. The torque of the input
shaft drives on the ring gear of the summarising planetary unit (P1-P2) one hand and
the pump of a variable axial piston pump other hand.
By variation of the swash plate angle on the piston pump, the rotation speed and
rotation direction of the piston motor can be varied continuously. The gear on the
output shaft of the piston motor is meshing with a gear connected to the sun wheel of
the summarising planetary gear unit (P1-P2). The summarising planetary gear unit is
a five shaft planetary gear set with two input shafts and three output shafts. If the sun
gear is rotating with the some speed value then the ring gear but the opposite
direction the rotation speed of the planetary carrier will be zero. This geared neutral
condition enables the system to hold the vehicle at standstill even on slopes without
activating the brakes. If the speed of the sun gear is zero, the power is transmitted
purely mechanically. Under this condition, the transmission is working in the best
efficiency area.

Figure 2. The virtual prototype model of the S-Matic transmission [1]

The Steyr S-Matic gear system operates with four mechanical transmission ranges
(P3, P4) in the whole tractor speed range. Shifting from one transmission range to the
next one is done at synchronous speed by clutches (K1, K2, K3, K4). To minimize
frictional losses, all clutches are designed as dog clutches. This continuously variable
transmission ratio and this shifting procedure provide there is no torque interruption
and smooth vehicle acceleration from standstill to maximum speed.
The output shaft of the four mechanical transmission range is connected to a
forward / reverse planetary gear (P5). During standstill of the vehicle the forward
(KV) and reverse dog clutches (KR) can be locked simultaneously, which gives the
possibility of keeping the torque flow and geared neutral function active even when
shifting from forward to reverse and reverse to forward.
The transmission efficiency is a highly important factor in the evaluation
procedure therefore it is necessary to include in the mathematical model. The
measuring of a efficiency of the transmission unit can be done a specific test
generally but in this tractor brake test was no opportunity to do it. Therefore the
concerning data were referenced to the direct measurement of the transmission
(Figure3.) [3]. The transmission efficiency is function of vehicle speed. The influence
of the hydrostatic unit to the efficiency graph is significant at the middle of every gear
range, where the hydrostatic power portion transferred is zero. At these operating
points, the power is transferred purely mechanically.
Figure 3. S-Matic transmission efficiency as a function of vehicle speed [3]

3 Developed mathematical model

The evaluation of the engine - CVT system is based on a mathematical model which
generates data of engine speed and transmission ratio pairs. Using these input data the
model calculates the vehicle speed, transmission efficiency and the specific fuel
consumption of the engine. In the current state of the evaluation there are some
limitations. A general purpose tractor can operate with a relative wide speed range. In
the practice this speed range in forward direction is between 0.5 km/h and 40 km/h.
Concerning to the distribution of the speed values on agricultural operations the
evaluated vehicle speed was set to 8 km/h in the procedure [4] (Figure 4.).

Figure 4. Vehicle speed distribution on agricultural operations

During the evaluation all of the possible engine speed and transmission ratio pairs
were generated with the given resolution. The range of the speed and ratio values was
determined with specific limits. These limits are calculated to cover the vehicle speed
range between 4 km/h and 14 km/h.
The transmission efficiency calculation is based on the extension of the efficiency
measurement data with a specific engine speed (2300 rpm) into a broader range of the
engine speeds. The limits of this extension were synchronized with the previous
boundary condition of the vehicle speed. The interpolation was executed in two steps
to decrease the effect of the mechanical gear shifts (Figure 5.). The specific fuel
consumption of the engine was calculated with a selected load case (300 Nm loading
torque).

Figure 5. Interpolated efficiency of the transmission

4 Evaluation of the calculated properties

The evaluation procedure can be separated into three stages. In the first stage the
input data field was generated with a given resolution. The engine speed values were
generated between 1800 rpm and 2600 rpm by 40 rpm steps. The transmission ratio
values located between 2 and 4 by 0.08 resolutions. The resolution can be specified in
any fine step values and it is limited only the available computational capacity. The
second stage of the evaluation includes the calculation the vehicle parameters. These
parameters are the vehicle speed, the efficiency of the transmission and the fuel
consumption of the engine. In the last stage the fuzzy based evaluation and the
visualization of the results can be found.
The evaluation is based on the fuzzy logic [5]. Fuzzy logic handles the ambiguous
cases of parameter evaluation. This way is close to human thinking because by
answering some simple questions the concerning fuzzy membership functions can be
set up. A typical fuzzy membership function (U(HA)) results on its output a number
between 0 and 1. Zero means that the studied value (Ey) is totally outside the set or
range. 1 marks that the value (Ex) is totally inside the range. Any other value (Ez)
between 0 and 1 shows the ratio of the membership. The conventional Boolean logic
of the sets is displayed in Figure 6. The fuzzy implementation is displayed in Figure
7.
Figure 6. The conventional boolean logic of the sets

Figure 7. The fuzzy logic of the sets

The fuzzy evaluation of the vehicle speed is based on the following fuzzy
membership function (Figure 8.a). The evaluated seed value is rated to optimal if it is
equal to 8 km/h. The deviance from this optimal value is rated with lower fuzzy
values up to zero. If the calculated speed is less than 4 km/h or more than 12 km/h the
concerning fuzzy value become to zero therefore these speed ranges are not
acceptable for agricultural operation.
The efficiency of the transmission is the function of the engine speed and the
vehicle speed. It means that the efficiency is not only dependent from the engine
speed and the acceptable efficiency level is varied by the vehicle speed. Renius
determined an objective function for the value of total efficiency [6] (Figure 9.). The
efficiency limit is calculated based on this diagram and the concerning fuzzy
membership function is set up with this value (Figure 8.b).

a, Vehicle speed b, Transmission efficiency c, Engine specific consumption

Figure 8. The fuzzy membership functions
Figure 9. Objective function for total transmission efficiency by Renius [5]
The specific consumption of the tested diesel engine is in the between 340 g/kWh and
420 g/kWh. Thus the membership function for the evaluation of the consumption is
accommodated to these limits (Figure 8.c).
In the last stage of evaluation procedure there are three parameters to rate. These
are the calculated vehicle speed, the transmission efficiency and the specific
consumption. The concerning fuzzy membership functions provide three different
fuzzy values which are have to be transforming back into a single value. This is the
defuzzyfication stage of the evaluation and this single value characterize the overall
quality of the evaluated state of the power train. There are many methods to perform
this defuzzyfication, but in this paper the Corrected Fuzzy Mean (CFM) was used [7].
The uVAR i fuzzy values are defined with the concerning Ui fuzzy membership
functions:

u VARi = Ui (PVARi ) (1)

Introducing the corrected fuzzy mean (RFZ) there is a good opportunity to

summarize as many results of the fuzzy membership function as it is required to
evaluate the specific state. This value is much simpler to calculate than the traditional
results of the fuzzy inference systems.

u VAR1 ⋅ u VAR2 ⋅ … ⋅ u VARn

R FZ = n

∑u
i =1
VARi (2)

These CFM (RFZ ) values indicates well the quality of a given state of the power
train specified by the engine speed and transmission ratio value pair. With calculation
of these CFM values the different states can be comparable.
Reviewing the results of the evaluation there are some important consequences can
be drawn. As the Figure 10. shows at the engine speed-transmission ratio plane, the
best vehicle speed values are concentrated around the 8 km/h value.
Figure 10. Velocity distribution, the best (around 8 km/h) values are highlighted
Concerning to the transmission efficiency the higher values are better (Figure 11.).

Figure 11. Efficiency distribution, the best (higher) values are highlighted
In case of the engine consumption the lower values are better (Figure 12.).
Figure 12. Efficiency distribution, the best (lower) values are highlighted
Analysing the quality values calculated with the Corrected Fuzzy Mean (CFM)
method there is quite conspicuous that the best quality values are not only distributed
on the common best value area of the previous diagrams. There are outlying values
highlighted at the engine speed-transmission ratio plane. It means that a good and
acceptable power train state can be combined with many kind of engine speed,
transmission ratio value pairs (Figure 13.).

Figure 13. Quality distribution, the best (higher) values are highlighted
5 Conclusions
This paper introduced a mathematical model of the high performance tractor power
train. The two main components of this power train are the common rail diesel engine
and the connected infinitely variable transmission (IVT). This model is based on test
measurements and other referenced results concerning to the IVT efficiency. The
computed power train parameters were evaluated with fuzzy evaluation. The final
results were performed by the method of the Corrected Fuzzy Mean (CFM). The
results highlighted that the desired vehicle speed can be combined with many engine
speed and transmission ratio pairs with different parameter values. The qualities of
these solutions are very variable and the maximization of the quality requires careful
selection the engine speed and transmission ratio. This mathematical simulation
enables to determine the optimal values for the integrated engine and transmission
electric control. The maximization of the efficiency of the engine-transmission
collaboration depends on the good control values. This mathematical model can
handle any engine-transmission combinations but it also has more improvement
opportunities. One possible improvement can be the handling of other operation
speed, like transport speed, not only one specific speed. Another improvement would
be the extension of the input parameters from the engine speed and transmission ratio
pairs. Taking into consideration the change of load torque will increase the accuracy
of the mathematical model, therefore the efficiency of the transmission and the
specific consumption of the engine are also depending on the load torque.

Acknowledgements

The authors wish to acknowledge the technical assistance of the Hungarian Institute
of Agricultural Engineering. Furthermore thanks for the support of Hungarian
Scientific Research Found (OTKA K62875).

References

1. Zs. Farkas, I. J. Jóri: Virtual prototype of power split infinitely variable tractor transmission
Synergy and Technical Development (Synergy2009) Gödöllő, 2009
2. Zs. Farkas, N.Q. Hung, S. Tóth, Gy. Kerényi: Modelling of the multiple-range infinitely
variable transmission, Gépészet 2008 Budapest, 29-30.May 2008. G-2008-K-05, ISBN
978-963-420-947-8
3. Heinz Aitzetmüller: Steyr S-Matic – The Future CVT System, Seoul 2000 FISITA World
Automotive Congress, June 12-15, 2000, Seoul, Korea
4. Henning J. Meyer: Durchschnittliche zeitliche Verteilung der Nennfahrgeschwindigkeiten
für Traktoren, Vorlesungfolien Traktoren, TU Berlin
5. Zadeh, L. A: Fuzzy Sets, Information and Control, vol. 8, pp: 338-353, ELSEVIER, 1965,
ISSN 0890-5401
6. Renius, K.Th.: Trends in Tractor Design with Particular Reference to Europe. Journal of
Agricultural Engineering Research 57 1994
7. A. Piros: Fuzzy evaluation in configurable product design, ICDES 2010, Tokyo, JAPAN,
2010. Nov. 17-19.