THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT.
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by Leonid Tartakovsky, Tuesday, October 16, 2012 03:48:23 AM
Simulation of Wankel Engine Performance Using
Commercial Software for Piston Engines
Leonid Tartakovsky, Vladimir Baibikov, Marcel Gutman and Mark Veinblat
Technion Israel Inst. of Technology
Jonathan Reif
Elbit Systems Ltd
ABSTRACT
At present the market of Wankel engines is limited to some
special applications. This fact explains absence of
commercial software products specially developed for this
engine simulation and prediction of its performance.
Conversely, there are available and widely used software
products for simulation of reciprocating-piston engines
performance. Some attempts are known in using this software
for prediction of Wankel engine performance. This paper
details an approach used in these attempts. Main differences
between both types of engines are summarized and principles
of a virtual reciprocating-piston engine compilation are
developed. A method of virtual blowing was developed for
assessment of discharge coefficients for intake and exhaust
ports. Comparison of simulation results with the measured
performance of two UAV Wankel engines showed sufficient
accuracy of the suggested approach.
INTRODUCTION
Nowadays there are many commercial software products
dealing with simulation of reciprocating-piston (RP) engines
performance. The most popular of them are the BOOST and
FIRE software of AVL (Austria), WAVE of Ricardo (UK),
GT-SUITE of GTI (USA) etc. In contrast, there is no
commercially available software specially developed for
simulating Wankel engine performance. Some attempts of
creating Wankel engine simulation software were made at the
period between seventies and beginning of nineties
[1,2,3,4,5,6,7,8,9,10,11,12,13]. However, none of them was
appropriately developed to contemporary commercial level.
2012-32-0098
20129098
Published
10/23/2012
Copyright © 2012 SAE International
doi:10.4271/2012-32-0098
Most probably, this results from the absence of significant
demand for such software since the present market of Wankel
engines is very limited. The UAV application may be
mentioned which utilizes such well known benefits of
Wankel engines compared to RP counterparts as higher
power-to-weight ratio, compactness and lower noise/vibration
levels. Taking this into account, a prediction of Wankel
engine performance using commercially available software
for RP engine might be of great interest.
Unfortunately, direct use of RP engine simulation software is
impossible for performance modeling of Wankel engines.
Because of that two different approaches may be applied:
• Development of specific software which is usually based on
detailed experimental data for some definite Wankel engine
[1,2,3,4,5,6,7,8,9,10,11,12,13]. Taking into account the very
limited available experimental database on Wankel engines,
such software is usually calibrated for only one engine type
and its applicability for other engines may be problematic. Of
course, development of new software requires very
significant investment of time and resources.
• Use of available RP engine software appropriately modified
to account for unique features of a Wankel engine [14,15,16].
In the work [14] the first attempt of such approach was made
when some components of the RP engine code were replaced
for taking into account main peculiarities of a Wankel engine
design and working cycle. The work was fulfilled on the base
of non-commercial RP software without detailed analysis of
RP and Wankel engines similarity criteria. In the paper [15]
the modified commercial AVL BOOST software was used
for simulation of Wankel engine performance where the
simulated Wankel engine was replaced with the virtual 3-
 THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT.
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by Leonid Tartakovsky, Tuesday, October 16, 2012 03:48:23 AM
cylinder, 4-stroke RP engine. The paper does not present
criteria for choice of the virtual RP engine dimensions, heat
transfer and heat release calculations. In the work [16] the
commercial software GT SUITE was used for prediction of
the Wankel combustion chamber pressure as a function of
rotor angle. As in [15], the simulated Wankel engine was
replaced with the virtual 3-cylinder, 4-stroke RP engine.
Some assumptions that do not reflect peculiarities of Wankel
engine geometry and working cycle were made in [15]: heat
release and heat transfer calculation procedures were used
unmodified and did not take into account changes of
combustion and heat transfer processes in Wankel engine
compared with RP counterpart, piston stroke of the virtual RP
engine was taken equal to eccentric offset of the simulated
Wankel engine, etc. These simplifications do not allow
prediction of Wankel engine performance with an acceptable
accuracy.
The approach of [15 and 16] with use of the GT-POWER
software of the GT-SUITE package is discussed and further
developed in this paper. The GT-SUITE is an up-to-date
integrated program for engines simulation and design
analysis. It contains the following components: GT-POWER
- engine simulation for performance and acoustic analysis;
GT-DRIVE - vehicle performance and cycle analysis for fuel
economy and emissions; GT-VTRAIN - valve train
kinematics, dynamics and tribology; GT-FUEL - injection
system pressure and flow analysis; GT-COOL - engine heat
and cooling system analysis; GT-CRANK - crankshaft
dynamics and torsional vibrations analysis. All six
components of GT-SUITE are based on a common set of
multi-physics libraries and thus have a large degree of
commonality among them. The GT-POWER is the 2- and 4-
stroke SI and diesel engine simulation tool widely used
nowadays by engines makers. Among its advantages are ease
of use and tight integration with the rest GT-SUITE
components. Its flexibility and usefulness are further
enhanced by the possibility of integration with STAR-CD,
Fluent, Simulink, MS/EXCEL and other general purpose
software. Due to the large size of its user base, the GT-
POWER has been thoroughly validated by real-life
applications. Use of this software which is focused on RP
engines for simulation of Wankel engine performance has
required development of the algorithm for composing a 3-
cylinder, 4-stroke virtual RP engine that would be
geometrically similar to the considered Wankel engine.
Development of special procedures that would take into
account differences between Wankel and RP engines in
combustion and heat transfer, as well as assessment of
discharge coefficients of intake and exhaust ports, was
required too.
MAIN DIFFERENCES BETWEEN
WANKEL AND PISTON ENGINES
The following main differences between Wankel and piston
engines that affect their performance can be mentioned:
• Difference in patterns of working chamber volume and
surface dependence on the angle of shaft rotation;
• Duration of the Wankel working cycle is 1.5 times longer
compared with a 4-stroke piston engine (in terms of the angle
of shaft rotation). Complete working cycle of a Wankel
engine takes place in each working chamber per one rotor
revolution, or per three shaft revolutions. In other words, one
working cycle in a Wankel engine occurs during 1080
degrees of shaft rotation (or 360 degrees of rotor rotation),
compared to 720 degrees of the crankshaft rotation in a 4-
stroke piston engine;
• “Hot” and “cold” stator zones of the Wankel working
chamber surfaces are separated contrary to a piston engine,
where the same working chamber surfaces are heated and
cooled in-turn. This, in combination with charge rotational
movement together with a working chamber, leads to
differences in the heat transfer conditions;
• Unfavorable shape of the Wankel working chamber (Figure
1) leads to the higher surface-to-volume ratios and larger
relative value of crevice volumes, where flame quenching
takes place. This causes differences in combustion patterns;
• Wankel working chamber has more complicated design of
seals (apexes and rotor side seals). This leads to the
possibility of higher charge leakage values;
• Wankel engine differs fundamentally from piston engine in
its kinematic mechanism. This leads to differences in internal
friction power losses.
Figure1. Unfavorable shape of the working chamber in a
Wankel engine [17].
 THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT.
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by Leonid Tartakovsky, Tuesday, October 16, 2012 03:48:23 AM

SIMILARITY CRITERIA
The method of Wankel engine performance simulation by
using RP engine software which is discussed in this paper
requires compilation of some similarity criteria for these
engines and composing a virtual piston (VP) 4-stroke engine
that would allow simulation of Wankel performance.
The following similarity criteria of Wankel and the VP
engines have been defined:
• Displacement equality;
• Compression ratio equality;
volume ratio for the VP engine should not differ from the
corresponding values of the modeled Wankel engine by more
then 1% at minimal working chamber volume values of these
engines.
A flow chart of the developed algorithm for calculation of the
geometric similarity parameters is presented in the Appendix
B. Results of the computations are shown in Tables 1, 2 and
Figures 4 - 6 for the example of the 802W Wankel engine. Its
main parameters are:
• Naturally aspirated, spark ignition, single-rotor, combined
air-liquid cooling;

• Identical behavior of the working chamber volume • Rated shaft speed - 8,000 rpm;
dependence on angle of shaft rotation;
• Rated brake power - 52 kW;
• Identical behavior of working chamber surface-to-volume
ratio change vs. angle of shaft rotation; • Displacement of each working chamber - 343.9 cc;

• Identical behavior of intake and exhaust ports discharge • Compression ratio - 7.4;
coefficients vs. angle of shaft rotation.
• Eccentricity - 11.6mm;
In addition to these similarity criteria which may be • Generating radius - 71.5mm;
summarized as a geometric similarity, the following
important aspects should be taken into account: • Rotor width - 80mm.

• Instantaneous values of the heat transfer through the As can be seen from Table 2, a maximal deviation from the
combustion chamber walls vs. angle of shaft rotation have to geometric similarity at the top dead center (TDC) does not
be calculated separately for non-firing (“cold”) and firing exceed 0.8% and at the bottom dead center (BDC) - 3.5%. It
(“hot”) parts of the working cycle; is clear that from the viewpoint of combustion and
• Modeling the combustion should take into account performance simulation, high accuracy of the geometric
specificities of the combustion process in Wankel engine. similarity at TDC is much more important then at BDC,
where combustion does not take place.
GEOMETRIC SIMILARITY
ALGORITHM
The main purpose of the algorithm is to ensure that the VP
engine would meet the criteria of geometric similarity listed
in the previous section. It means that the VP engine has to be
of the same displacement, compression ratio, behavior of
working chamber volume and surface-to-volume ratio change
versus shaft angle as the modeled Wankel engine.

Equations characterizing geometric parameters of Wankel

[18] and piston [19] engines are presented in Appendix A.
Designation explanations are presented in Fig.2 (for a
Wankel engine) and in Fig.3 (for a piston engine). Joined
solution of these equations for Wankel and piston engines
allows finding geometric parameters (bore, stroke and
connecting rod length) of a VP engine geometrically similar
to the modeled Wankel engine. Unfortunately, an analytic Figure 2. Designations for Wankel engine parameters.
solution of these equations gives different results for various
angles of shaft rotation. Therefore, achievement of absolute
geometric similarity at each moment of the engines operation
cycle is impossible and approximate iterative solution was
applied with the following criterion of convergence. Values
of the working chamber volume, surface and surface-to-
 THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT.
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by Leonid Tartakovsky, Tuesday, October 16, 2012 03:48:23 AM

Figure 3. Designations for piston engine parameters.

Figure 4. Working chamber volume vs. angle of shaft
Table 1. Parameters of the virtual 4-stroke piston engine. rotation for 802W Wankel and virtual piston engines

Figure 5. Working chamber surface area vs. angle of

shaft rotation for 802W Wankel and virtual piston
engines.

Table 2. Deviations from the geometric similarity of the

VP engine.

Figure 6. Working chamber surface/volume ratio vs.

angle of shaft rotation for 802W Wankel and virtual
piston engines.

As can be seen from Tables 1, 2 and Figures 4,5,6, the

geometrical similarity of the VP and the modeled Wankel
engines is kept closely, especially in the vicinity of TDC.
 THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT.
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by Leonid Tartakovsky, Tuesday, October 16, 2012 03:48:23 AM

DISCHARGE COEFFICIENTS FOR DC values were calculated for each studied position of the
rotor using the isentropic gas flow rate equation for subsonic
INTAKE AND EXHAUST PORTS velocity of outflow [20], where the real process effects were
Values of discharge coefficients (DC) for intake and exhaust taken into account to some extent by using a typical range of
ports are very important because they significantly affect the polytrophic exponent n values:
calculations of gas exchange in the engine working chamber.
It is clear that they greatly influence the accuracy of the
engine performance simulation. Therefore, DC data used as
the inputs in a simulation approach must be found as accurate
(2)
as possible.
Where: Q - gas mixture flow rate; DC - discharge coefficient;
Obviously, an accurate DC assessment would be based on
A - area of the port; n - polytropic exponent; ρo - upstream
experimental measurements, when air is blown through the
real orifice. Unfortunately, these results are not always mixture density; p0 - upstream pressure; p - downstream
available on early stages of engine development. Therefore a pressure.
method of flow simulation through the intake and exhaust
ports was developed. These calculations were performed for
few different positions of the rotor relative to the
corresponding port. Appropriate charts were prepared using
Solid Works software. Examples of this virtual blowing are
presented in Figures 7 (for intake port) and 8 (for exhaust
port).

Gas flow rate values through the port were computed using
the FLUENT software for each studied position of the rotor.
Unburned mixture with air/fuel ratio of 14.5 (typical for SI
Wankel engines) and burned mixture of N2/CO2/H2O vapor
(typical for complete combustion) were used for the
simulation of gas flow through the intake and exhaust port,
respectively.
Figure 7. Simulation scheme for intake port blowing.
Molecular weight of each gas mixture was calculated by
equation:
(1)
where: Ci - mole fraction of the component “i” Mi -
molecular weight of the component “i”.
The data that have been used for the calculation of gas
mixtures molecular weight and calculation results for intake
and exhaust ports are given in Table 3.
Figure 8. Simulation scheme for exhaust port blowing
Input data, such as pressure and temperature in the working
chamber, that are required for the calculation carrying out,
have been assessed using a typical indicatory diagram of a
Wankel engine. Upstream data for the exhaust port and
downstream data for the intake port were spatially averaged
for the combustion chamber volume for each considered
opening area of the corresponding port. Downstream data for
the exhaust port and upstream data for the intake port were
taken as ambient conditions (1 bar and 20 degrees of
Celsius). Since pressure values in the working chamber that
have been assessed based on the typical indicatory diagrams
cannot be sufficiently accurate, the calculations were made
for several values of the pressure in the vicinity of the
calculated point. This allowed assessment of DC sensitivity
to pressure value p in the working chamber. Fig. 9 illustrates
example of calculation results for 50% opening of exhaust
port. As can be seen from this Figure, minimal DC value was
found for the combination of the minimal pressure and
polytrophic exponent values. Accordingly, the maximal DC
value anticipated for the highest values of pressure and n.
Deviation of the computed DC values in the realistic range of
pressures and polytrophic exponents was found to be less
than ±12%. In this work the DC values for each port opening
were assumed to be the average of those found by the
computations for nine combinations of pressure value in the
working chamber and polytropic exponent in the range of
 THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT.
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by Leonid Tartakovsky, Tuesday, October 16, 2012 03:48:23 AM

1.5<p<2.5 bar (exhaust port), 0.7<p<0.9 bar (intake port) and Table 4. DC values calculated for Wankel engine and
1.2<n<1.4. those typical for reciprocating engines

Table 3. Results of molecular weights calculation for

unburned and burned mixtures.

Figure 10. Values of intake port DC at various port

openings.

Figure 9. Results of DC calculation for 50% opening of

exhaust port.

Results of computations that were performed by using the

described approach are shown in Figure 10 (for intake port)
and Figure 11 (for exhaust port). As can be seen from Table
4, maximal values of the DC obtained by using the suggested
approach for the Wankel engine are substantially higher than Figure11. Values of exhaust port DC at various port
those of reciprocating piston engines [21]. Obviously, this openings.
may be explained by design simplicity of Wankel intake/
exhaust ports in comparison with 2-stroke engines and, HEAT TRANSFER COEFFICIENT
especially, with 4-stroke engines where intake/exhaust flows
Obviously, the heat transfer coefficient (HTC) calculated for
are restricted by poppet-valves. Figures 10 and 11 contain DC
the virtual piston engine should reflect the heat transfer
values only for the process of ports opening. The same values
conditions in the simulated Wankel engine. As described
of the discharge coefficients were assumed also for the ports
earlier, the main factors leading to different conditions of heat
closing process. These values were used as an input data for
transfer in Wankel engine compared with a piston counterpart
the engine simulation software.
are:
• Unfavorable shape of Wankel working chamber with
significantly higher surface-to-volume ratio;
 THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT.
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by Leonid Tartakovsky, Tuesday, October 16, 2012 03:48:23 AM

• Non-firing (“cold”) and firing (“hot”) stator zones of

Wankel working chamber surfaces are separated contrary to
piston engine;
• Charge rotational movement together with a working
chamber.

These differences do not allow straightforward application of

the HTC calculation procedure of GT-Power commercial
software [21] based on the Woschni correlation for an engine
without swirl, as described in detail in Section 12.4.3 of [22].
In this approach the first factor was taken into account by
composing geometrically similar virtual piston engine with
the same surface-to-volume ratio of a working chamber, as
described earlier. Regarding the last two factors, it is
impossible to use standard computation routine of RP engine
software for assessment of the heat transfer coefficient.
Therefore, in this work the approach used in [9, 13 and 14]
was applied. This method is based on the classical
relationship between the dimensionless groups nowadays
known as Nusselt, Reynolds and Prandtl numbers that was
introduced for the first time by W. Nusselt [23, 24] for
turbulent flow in pipes or over flat plates:
(3)
where: Nu = hL/k is Nusselt number; Re = vL/v is Reynolds
number; Pr = µcp/k is Prandtl number; K, a, b - constants; h -
heat transfer coefficient; L - characteristic length; k - thermal
conductivity of gas; v - characteristic velocity; v - gas
kinematic viscosity; µ - gas dynamic viscosity; cp - specific
heat at constant pressure.
Values of the constants: C1 = 1; C2 = 0.00324 m/(s·K) were
taken from [14].
Substitution of all the data into the equation (3) and its
solution for h gives HTC values for each angle of the
crankshaft rotation. Fig. 12 presents Wankel HTC behavior
over crankshaft angle of the virtual piston engine operated
under rated power, as calculated by using the described
algorithm. This Figure contains also the HTC values for the
same operation regime of this engine that were computed
using the commercial software procedure. As can be seen, the
HTC values computed by using the new algorithm for a
Wankel engine are larger by 10…15% than those assessed
with the aid of a commercial software procedure for
reciprocating piston engines. The difference is a result of the
peculiarities of a Wankel engine, discussed above.

In [9, 13 and 14] Prandtl number is assumed to be unity and

the flow pattern is assumed to be similar to that of the forced
convection over a flat plate. The latter provides the following
values of the constants in equation (3), [14]: K = 0.087; a =
0.8.

The characteristic length L for Wankel engine was assumed

to be equal to the rotor width, i.e. Lw = H, and for the virtual
piston engine it was assumed to be equal to the cylinder bore,
i.e. Lp = B. The characteristic velocity v was assessed in
different ways for suction/compression and for combustion/
expansion strokes. For the suction/compression stroke (non-
firing):
(4)
where: n - Wankel engine shaft speed; R - generating radius.
The instantaneous characteristic velocity for the combustion/
expansion stroke (firing) was given by:
(5)
Figure 12. Heat transfer coefficient (HTC) patterns for
the virtual piston engine under rated brake power - 52
kW: 1 - computed by using commercial software
procedure; 2 - computed using the algorithm developed
for Wankel engine.
MODEL VALIDATION
The validation of the developed approach was carried out by
comparison of the predicted 802W and 802 Wankel engines
performance with the available experimental data for these
engines. The 802 engine differs from the 802W one by lesser
rotor width (68mm compared with 80 mm). The calculations
 THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT.
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by Leonid Tartakovsky, Tuesday, October 16, 2012 03:48:23 AM

were carried out for the same ambient and cooling conditions
that were observed under the engines tests. For each
calculated regime piston surface temperature of the virtual RP
engine was assumed equal to that one of the simulated
Wankel rotor face, temperatures of cylinder head and
cylinder surfaces were assumed equal to these of the Wankel
housing inner surface on the base of experimental data. HTC
values were computed as described in the previous section.
The engine tests and calculations were carried out over
regimes of full load and propeller curves. The latter was
needed since the simulated Wankel engines are intended for
UAV propulsion, i.e. they are loaded by propeller where the
relationship between the engine load P and speed n is
described by the equation P=kn3.
Figure 15. Wankel engines 802 and 802W: predicted
Figures 13,14,15 show satisfactory correspondence of the (lines) and measured (dots) values of the fuel
calculated and measured values of brake power and fuel consumption over the propeller curve.
consumption at various operation regimes of the 802W and
802 engines. As can be seen from Figures 13, 14, 15, the maximal
discrepancy of the measured and predicted results is about
5% and 3-10% for the brake power and fuel consumption,
respectively. This indicates the acceptability of the model
presented here.

SUMMARY AND CONCLUSIONS

In contrast to piston engines, none of Wankel engine
simulation software was appropriately developed to
contemporary commercial level up to date. At present only
AVL BOOST commercial software has additional block for
Wankel engine performance prediction by replacing it with a
virtual RP engine. Most probably, this results from the
absence of significant demand for such software, since the
market of Wankel engines is very limited. So, a possibility of
Wankel engine performance prediction using commercially
Figure13. Wankel engines 802 and 802W: predicted
available software for reciprocating-piston engine might be of
(lines) and measured (dots) values of the brake power
great interest.
over full load curve.
Since straightforward use of the piston engine simulation
software is impossible, there are different approaches to
simulation of Wankel engine performance. Some of them are
focused on the development of new software for the specific
Wankel engine on the basis of its detailed experimental data.
Another way is based on adaptation of available piston
engine commercial software products. The latter approach is
detailed in this paper. It required development of the Wankel
-to- RP engine geometric similarity algorithm and
compilation of a virtual RP engine, which would allow
assessment of Wankel engine performance. Peculiarities of
heat transfer and combustion processes in a Wankel engine
were taken into account as well.

The developed algorithm ensures meeting the geometric

Figure 14. Wankel engines 802 and 802W: predicted similarity criteria (equality of displacement, compression
(lines) and measured (dots) values of the fuel ratio, working chamber volume and surface-to-volume ratio
consumption over full load curve. vs. crankshaft angle). It allowed finding base dimensions of
the virtual piston engine that can be used for performance
 THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT.
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by Leonid Tartakovsky, Tuesday, October 16, 2012 03:48:23 AM
simulation of a Wankel engine. The applied methodology of
the heat transfer coefficient computing allowed accounting
for such features of Wankel engine as “hot” and “cold” stator
zones separation, unfavorable shape of working chamber and
the charge rotational movement.
For the case when discharge coefficients cannot be
experimentally obtained (the common situation at early
engine development stages), a method of blowing simulation
through the intake and exhaust ports was developed. This
method includes compilation of charts for intake and exhaust
port zones, when each of the charts has different rotor
position relative to the port (i.e. different rate of the port
opening). Unburned mixture with air/fuel ratio of 14.5 and
burned mixture of N2/CO2/vapor H2O were assumed for
blowing through the intake and exhaust port respectively. The
gas flow rate through the port can be computed using any
available CFD software for different positions of the rotor,
with appropriate calculation of discharge coefficients.
Combustion rate and heat release behavior were simulated by
using the conventional approach for piston engines - the
Wiebe equation. Values of the constants in the latter were
appropriately changed based on the available literature data
for Wankel engines.
This approach was demonstrated through the example of the
802 and 802W engines brake power and fuel consumption
simulation at various operation regimes. A comparison of the
predicted Wankel engines performance with the measured
data has shown satisfactory correlation.
The developed simulation approach can be used for
optimization of Wankel engine parameters, such as ignition
advance map, intake and exhaust ports timing and geometry,
air/fuel ratio, as well as assessment of the influence on engine
performance of air filter and exhaust muffler characteristics,
flight altitude, etc.
REFERENCES
1. Danieli, G., Keck, J., and Heywood, J., “Experimental and
Theoretical Analysis of Wankel Engine Performance,” SAE
Technical Paper 780416, 1978, doi: 10.4271/780416.
2. Horsfield, B., “Computer Modeling [Wankel Rotary
Combustion Engine]” Electronics & Computing Monthly
3(6): 56-7, 1983.
3. Weston, K.C., “Computer Simulation of a Wankel Rotary
Engine - Analysis and Graphics”, Proceedings of the 1986
Summer Computer Simulation Conference, USA, July 28-30,
1986.
4. Roberts, J., Norman, T., Ekchian, J., and Heywood, J.,
“Computer Models For Evaluating Premixed and Disc
Wankel Engine Performance,” SAE Technical Paper 860613,
1986, doi: 10.4271/860613.
5. Shih, T., Schock, H., and Ramos, J., “Fuel-Air Mixing and
Combustion in a Two-Dimensional Wankel Engine,” SAE
Technical Paper 870408, 1987, doi: 10.4271/870408.
6. Lee, C. and Schock, H., “Regressed Relations for Forced
Convection Heat Transfer in a Direct Injection Stratified
Charge Rotary Engine,” SAE Technical Paper 880626, 1988,
doi: 10.4271/880626.
7. Ramos, J.I., Shih, T.I-P., Schock, H.J., “Wankel engine
modeling”, Proceedings of the International Centre for Heat
and Mass Transfer: 469-483, 1989.
8. Li, Z., Steinthorsson, E., Shih, T., and Nguyen, H.,
“Modelling and Simulation of Wankel Engine Flow Fields,”
SAE Technical Paper 900029, 1990, doi: 10.4271/900029.
9. Bartrand, T.A., and Willis, E.A., “Performance of a
Supercharged Direct-Injection Stratified-Charge Rotary
Combustion Engine”, NASA technical memorandum 103105,
1990.
10. Kim, K.J., Chae, J.O., and Chung, T.Y., “A Performance
Simulation for Spark Ignition Wankel Rotary Engine,” SAE
Technical Paper 912479, 1991, doi: 10.4271/912479.
11. Raju, M., and Willis, E., “Three Dimensional Analysis
and Modeling of a Wankel Engine,” SAE Technical Paper
910701, 1991, doi: 10.4271/910701.
12. Raju, M., “Heat Transfer and Performance
Characteristics of a Dual-Ignition Wankel Engine,” SAE
Technical Paper 920303, 1992, doi: 10.4271/920303.
13. Bartrand, T.A., and Willis, E.A., “Rotary engine
performance computer program. User's guide”, NASA
Contractor Report 191192, October, 1993.
14. Norman, T.J., “A performance model of a spark ignition
Wankel engine: including the effects of crevice volumes, gas
leakage and heat transfer”, M.S. thesis, Massachusetts
Institute of Technology, Massachusetts, June 1983.
15. Wendekerer, M., Grabovski, L., Pietrikovski, K. and
Magryta, P., “Phenomenological model of a Wankel engine”,
Silniki Spalinowe, 50(3), 2011.
16. Handschuh, R.F. and Owen, A.K., “Analysis of apex seal
friction power loss in rotary engines”, NASA/TM
-2010-216353, September 2010.
17. Poojitganont, T., Izweik, H.T., and Berg, H.P., “The
simulation of Flow Inside the Wankel Combustion
Chamber”, http://www.tsme.org/ME_NETT/ME_NETT20/
article/pdf/tsf/TSF031.pdf, Nov. 2011.
18. Yamamoto, Kenichi, “Rotary engine”, Published by
Toyo Kogyo Co., LTD. 1969.
19. Automotive handbook, 8th edition, Robert Bosch GmbH,
2011.
20. Mechanical engineering handbook, V.2, Moscow, 1961
(in Russian).
 THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT.
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by Leonid Tartakovsky, Tuesday, October 16, 2012 03:48:23 AM

21. http://www.gtisoft.com/applications/
a_Engine_Performance.php
22. Heywood, John B., “Internal combustion engine
fundamentals”, McGraw -Hill Inc., 930 p., 1988.
23. Nusselt, W., “Das Grundgesetz des Wärmeüberganges”,
Gesundheits-Ingenieur, 38 (42): 477-482 and (43): 490-496,
1915.
24. Nusselt, W., “Der Wärmeübergang im Rohr”, Z. VDI,
61, 1917.

CONTACT INFORMATION
Dr. Leonid Tartakovsky
Faculty of Mechanical Engineering
Technion - Israel Institute of Technology
Technion City, Haifa 32000, Israel
tartak@technion.ac.il

ACKNOWLEDGMENTS
The authors are grateful to MAFAT, Israeli Ministry of
Defense for the financial support of this project.

DEFINITIONS/ABBREVIATIONS
BDC - bottom dead center
DC - discharge coefficient
HTC - heat transfer coefficient
RP - reciprocating piston
SI - spark ignition
TDC - top dead center
UAV - unmanned aerial vehicle
VP - virtual piston
 THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT.
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by Leonid Tartakovsky, Tuesday, October 16, 2012 03:48:23 AM

APPENDIX

APPENDIX A
Rotor lobe surface area Ac is a constant value and was calculated by using the modified approximating formula from [14]:

(A1)

Where: Rx - radius of the circular arc approximation

(A2)

β is one half the angle subtended by the approximate circular arc:

 THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT.
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by Leonid Tartakovsky, Tuesday, October 16, 2012 03:48:23 AM

(A3)
m is multiplier considering rotor lobe recess.

Following [14], in the equation (A3) values of R and e must be inserted in inches.

Instantaneous value of the housing surface area as function of the angle of rotor rotation was calculated analytically using the
following equation derived in this work:

(A4)

Instantaneous value of the side plate surface area consists of two parts - the first is the constant one Acompr (corresponds to
compression volume) and the second is the variable one Adispl (corresponds to displacement volume):

(A5)

Values of these parts were found using the following equations [16]:

(A6)

Where φmax = 3e/R.

(A7)

APPENDIX B
 THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT.
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by Leonid Tartakovsky, Tuesday, October 16, 2012 03:48:23 AM

Fig. B1. Flow chart of virtual piston-to-Wankel engines geometric similarity algorithm. For designations see Fig.B2 and a list
below.

Fig. B2. Designations of Wankel engine geometry elements.

Symbols designation
Ac - rotor lobe area;
As - area between rotor lobe arc and chord XY;
Vcr - volume of rotor recession;
Δα - angle step of calculations;
 THIS DOCUMENT IS PROTECTED BY U.S. AND INTERNATIONAL COPYRIGHT.
It may not be reproduced, stored in a retrieval system, distributed or transmitted, in whole or in part, in any form or by any means.
Downloaded from SAE International by Leonid Tartakovsky, Tuesday, October 16, 2012 03:48:23 AM

n - number of calculation steps;

k - multiplier for conrod length iteration;
α(i) - instantaneous value of rotor rotation angle;
Ai - instantaneous value of area between arch Li and chord XY;
Li - instantaneous value of trochoidal arch;
Aw - instantaneous value of one working chamber surface area of Wankel engine;
Vw - instantaneous value of one working chamber volume of Wankel;
Vcw - compression volume of Wankel;
Vw - maximal value of one working chamber volume of Wankel;
Vhw - displacement of Wankel one working chamber;
B - virtual piston engine cylinder bore;
R - crank radius;
h(i) - instantaneous value of distance between piston and TDC;
Ap(i) - instantaneous value of piston engine working chamber surface area;
Vhp(i) - instantaneous value of piston engine volume that corresponds to h(i);
Vcp - compression volume of piston engine;
Vp(i) - instantaneous value of piston engine working chamber volume.

The Engineering Meetings Board has approved this paper for publication. It has Positions and opinions advanced in this paper are those of the author(s) and not
successfully completed SAE's peer review process under the supervision of the session necessarily those of SAE. The author is solely responsible for the content of the paper.
organizer. This process requires a minimum of three (3) reviews by industry experts. SAE Customer Service:
Tel: 877-606-7323 (inside USA and Canada)
All rights reserved. No part of this publication may be reproduced, stored in a
Tel: 724-776-4970 (outside USA)
retrieval system, or transmitted, in any form or by any means, electronic, mechanical, Fax: 724-776-0790
photocopying, recording, or otherwise, without the prior written permission of SAE. Email: CustomerService@sae.org
ISSN 0148-7191 SAE Web Address: http://www.sae.org
Printed in USA