Energy Conversion and Management 101 (2015) 470–480

Contents lists available at ScienceDirect

Energy Conversion and Management

journal homepage: www.elsevier.com/locate/enconman

Supercharging system behavior for high altitude operation of an aircraft

2-stroke Diesel engine
Antonio Paolo Carlucci ⇑, Antonio Ficarella, Domenico Laforgia, Alessandro Renna
University of Salento, Department of Engineering for Innovation, Via per Monteroni, 73100 Lecce, Italy

a r t i c l e i n f o a b s t r a c t

Article history: Different studies on both 2- and 4-stroke engines have shown how the choice of different supercharging
Received 19 March 2015 architectures can influence engine performance. Among them, architectures coupling one turbocharger
Accepted 3 June 2015 with a mechanical compressor or two turbochargers are found to be the most performing in terms of
engine output power and efficiency. However, defining the best supercharging architecture for aircraft
2-stroke engines is a quite complex task because the supercharging system as well as the ambient con-
Keywords: ditions influence the engine performance/efficiency. This is due to the close interaction between super-
Engine supercharging
charging, trapping, scavenging and combustion processes. The aim of the present work is the comparison
Supercharging architecture
2-stroke Diesel engine
between different architectures (single turbocharger, double turbocharger, single turbocharger combined
Trapping efficiency with a mechanical compressor, single turbocharger with an electrically-assisted turbocharger, with inter-
Scavenging efficiency cooler or aftercooler) designed to supercharge an aircraft 2-stroke Diesel engine for general aviation and
General aviation unmanned aerial vehicles characterized by a very high altitude operation and long fuel distance. A 1D
UAV model of the engine purposely designed has been used to compare the performance of the different
supercharging systems in terms of power, fuel consumption, and their effect on trapping and scavenging
efficiency at different altitudes.
The analysis shows that the engine target power is reached by a 2 turbochargers architecture; in this
way, in fact, the cylinder filling, and consequently the engine performance, are maximized. Moreover, it is
shown that the performance of a 2 turbochargers architecture performance can be further improved con-
necting electrically and not mechanically the low pressure compressor and turbine (electrically-assisted
turbocharger). From an energetic point of view, this system has also proved to be particularly convenient
at high engine speed and load, because it is possible to extract power from the electric turbocharger with-
out a penalty on specific fuel consumption.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction
One of the crucial points of the use of a reciprocating internal
combustion engine for aircraft propulsion is the performance
decline with the increase in altitude. This is due to the negative
effect of air density reduction on cylinder filling process. The solu-
tion to this problem is the engine supercharging [1]. To super-
charge a 2- and 4-stroke engines a turbocharger (TC), and/or a
mechanical compressor, can be used: the first involves an air com-
pressor fed by a turbine driven by exhaust gases; the second one
consists of a compressor mechanically coupled to the engine
crankshaft [1].
⇑ Corresponding author at: Systems for Energy and Environment, University of
Salento, Dept. Engineering for Innovation (CREA), Via per Monteroni, 73100 Lecce,
Italy. Tel.: +39 0832 297751, mobile: +39 340 6925653; fax: +39 0832 297777.
E-mail address: paolo.carlucci@unisalento.it (A.P. Carlucci).
In 2-stroke engines, the cylinder filling and emptying phases are
driven by the difference between inlet and outlet pressure.
Therefore, the supercharging system of a 2-stroke engine is respon-
sible for the air density increase and cylinder scavenging. Due to
the strict interaction among supercharging, scavenging and com-
bustion, 2-stroke engines are very sensitive to the variation of
engine working parameters and ambient conditions with altitude
[2,3].
Several studies on both 2- and 4-stroke engines have shown
how the supercharging architecture can influence the engine per-
formance. Pohorelsky et al. [4] simulated numerically different
supercharging systems, characterized by positive displacement or
dynamic compressors mechanically coupled to the engine before
or after a TC with Waste-Gate (WG) or variable nozzle turbine to
ensure power and scavenging targets of 2-stroke Diesel engines.
They proved that the best configuration is to place the positive dis-
placement compressor upstream the TC with WG. Systems with

http://dx.doi.org/10.1016/j.enconman.2015.06.009
0196-8904/Ó 2015 Elsevier Ltd. All rights reserved.
 A.P. Carlucci et al. / Energy Conversion and Management 101 (2015) 470–480 471

Nomenclature

AFRcmb trapped air/fuel ratio m_C compressor mass flow

AP air purity m_T turbine mass flow
ATDC after top dead center n engine speed
BSFC brake specific fuel consumption p_amb International Standard Atmospheric pressure
cpC mean specific heat at constant pressure during com- p2/p1 compressor pressure ratio
pression phase p4/p3 turbine expansion ratio
cpT mean specific heat at constant pressure during expan- Pb engine brake power
sion phase R propeller radius
CP propeller power coefficient SE scavenging efficiency
CAD crank angle degree T_amb International Standard Atmospheric temperature
DR delivery ratio T1 compressor inlet temperature
EAT electrically assisted turbocharger T3 turbine inlet temperature
eCOM electrical compressor tcPLUS difference between the power produced by the turbine
eTURB electrical turbine and the power required by the compressor
HP high pressure TC Turbocharger
LP low pressure TE trapping efficiency
mas mass of fresh charge entering the cylinder during the UAV unmanned aerial vehicle
scavenging phase V aftercooler volume
msc mass of fresh charge short-circuiting the cylinder during VGT variable geometry turbocharger
the scavenging phase (fraction of mas) WG waste gate valve
mcp mass of combustion products exiting the cylinder dur- Dp aftercooler pressure drop
ing the scavenging phase e aftercooler efficiency
mtr total mass trapped into the cylinder after the scaveng- qa ambient air density
ing phase (with valves and ports closed) gsC compressor isentropic efficiency
mtas mass of intake fresh charge trapped into the cylinder gsT turbine isentropic efficiency
mex mass of exhaust gases not discharged from the cylinder gm turbocharger mechanical efficiency
(fraction of mcp)
mar mass of residual air from the previous cycle not having
participated to the combustion

electric compressor and Electrically Assisted Turbocharger (EAT)
required excessive electric power.
Knoll [5] modeled numerically and compared different super-
charging architectures for a 2-stroke engine prototype: a mechan-
ical compressor, a TC coupled with a crankcase scavenging pump,
and a TC combined with an external mechanical compressor. The
last one proved to be the best option, thanks to the compression
power reduction, and beneficial effect of the exhaust backpressure
on trapping efficiency.
Mattarelli [6] analyzed two different architectures for a
4-stroke Diesel engine. The first one comprises two TCs; the second
one comprises a TC and a mechanical positive displacement com-
pressor. The analysis revealed that, using a mechanical compressor
and a Variable Geometry Turbocharger (VGT), the engine perfor-
mance at full-load are improved together with the fuel consump-
tion and pollutants control [7].
Chen et al. [8] studied the influence of two-stage turbocharging
system parameters on the engine performance. It was found that
engine efficiency and performance are strongly influenced by high
pressure (HP) and Low Pressure (LP) compressors efficiency, by
intercooler efficiency and cooling water temperature. In
off-design operation, the turbocharging system shall be designed
to ensure the two compressors to operate with a similar value of
the compressor efficiency. A similar analysis with similar results
has been presented by Liu et al. [9], in which a matching method
for two-stage turbocharging system has been proposed. Shan
et al. [10] proposed a simulation method for the design and analy-
sis of an aerial turbocharging system, composed by two TCs, two
intercoolers, and two turbine by-pass valves. It was concluded that,
operating the two by-pass valves correctly, it is possible to meet
the design objectives at different altitudes, and to operate the com-
pressors to minimize the compression work.
Divekar et al. [11] simulated the performance of a supercharg-
ing scheme decoupling the intake-boost system from the exhaust
recovery system using electric supercharging and
turbo-generation with two separate electric machines and a dedi-
cated energy storage buffer. In this way, the system transient
response is faster if compared to a conventional turbocharged sys-
tem, and a reduction in fuel consumption over driving cycles char-
acterized by frequent transient operation is obtained. These
conclusions are confirmed by Katrasnik et al. [12].
Terdich et al. [13] showed that the hybridization of the TC is the
most effective way to recover the transient response of a TC; VGT,
moreover, decreases the fuel consumption. Mamalis et al. [14]
showed that the system with two TCs leads to the best extension
of the operational range of a 4-stroke Homogeneous Combustion
Compression Ignition engine, i.e. an engine in which well-mixed
fuel and oxidizer are compressed to the point of auto-ignition.
The aim of the present work is to compare different architec-
tures – single turbocharger, double turbocharger, single tur-
bocharger combined with a mechanical compressor, single
turbocharger with an EAT, with intercooler or aftercooler – to
supercharge an aircraft 2-stroke Diesel engine for general aviation
and Unmanned Aerial Vehicles (UAV). This engine is characterized
by a very high altitude operation and long fuel endurance. To carry
out tests, a 1D model of the engine has been realized with AVL
BOOST software. The performance of the different supercharging
systems were compared to determine the most performing system
in terms of power and specific fuel consumption. Moreover, the
results have been validated analyzing the cylinder scavenging
and filling processes. Lastly, the effect of the EAT compression ratio
on the power produced or required by the EAT at part load has
been also analyzed to define the best control strategies of fuel
consumption.
472 A.P. Carlucci et al. / Energy Conversion and Management 101 (2015) 470–480

Table 1 4.E+03
Intake conditions and required performance per engine bank with altitude variation. geometric flow area
Altitude p_amb T_amb Power (% referred to effecve flow area
take-off sea level) per 3.E+03
bank at 2000 rpm
[m] [kft] [bar] [°C] [%]

[mm2]
0 (sea level) 0 1.01 15 100
3050 10 0.69 6 100 2.E+03
5180 17 0.53 18 89
10,670 35 0.23 54 54

1.E+03

2. Structure of the engine model

0.E+00
100 120 140 160 180
The required engine targets, in terms of output power referred CAD ATDC [°]
to sea-level power for take-off are reported in Table 1 together
with related International Standard Atmospheric pressure Fig. 1. Total geometric and effective inlet area per cylinder as a function of CAD
(p_amb) and temperature (T_amb). In its initial configuration the during the descending piston stroke.
engine was a Diesel 2-stroke boxer engine composed by six cylin-
ders arranged in two banks, each one fed by a TC, an aftercooler
3.E+03
and an intake plenum. The scavenging system is ‘‘Uniflow’’ (inlet geometric flow area
ports and exhaust valves). The main specifications of the engine effecve flow area
are reported in Table 2. The inlet ports and exhaust valves opening
and closing angles are reported in Crank Angle Degrees After Top 2.E+03
Dead Center (CAD ATDC).
The software used for modeling the engine is AVL BOOST
[mm2]

v2011.2. This software solves 1D time dependent equations pre- 2.E+03

dicting the mass flux in intake and exhaust systems; 0D approach
is used to calculate the combustion development, in-cylinder pres-
sure, the thermal exchanges and the indicated work into the cylin- 8.E+02
ders. Pressure losses, due to three-dimensional effects in particular
zones of the engine like inlet ports and exhaust valves, were con-
sidered through appropriate discharge coefficients [15], defined
0.E+00
as the ratio between effective and geometric flow areas [1]. In 70.0 90.0 110.0 130.0 150.0
Figs. 1 and 2 the value of both these areas, at inlet and exhaust,
CAD ATDC [°]
are respectively shown as a function of CAD.
In 2-stroke engines, the scavenging process is crucial in deter- Fig. 2. Total geometric and effective exhaust area per cylinder as a function of CAD
mining the engine performance. Referring to Fig. 3(a) and (b), the during the opening phase.
quality of the scavenging process can be quantified through the
Scavenging Efficiency (SE) and Trapping Efficiency (TE) [16,17]:

mtas 
SE ¼  ð1Þ
mtas þ mex þ mar all closed


mtas 
TE ¼ ð2Þ
mas all closed

If mar is negligible, SE is almost equal to the Air Purity (AP),

which is the fraction of the mass of fresh air in the cylinder when
the valves and ports are closed (SE 6 AR):

mtas þ mar 
AP ¼ ð3Þ
mtas þ mex þ mar all closed

The definitions of SE and TE imply, for example, that with a high

boost pressure the cylinder is very effectively scavenged (therefore
SE  1 since mex  0), whereas it is not efficiently scavenged when

Table 2 Fig. 3. Scavenging process (a) and following compression stroke (b) with ports and
Engine main specifications. valves closed.

Cycle 2-stroke Diesel Uniflow

Bore/Stroke 1
Compression ratio 17.2:1
Injection system Common rail too much fresh air has been pumped into the cylinder compared to
Propeller/engine speed ‘‘n’’ 2000 rpm the mass actually trapped (mtas  mas ) TE  1). The choice of the
Inlet ports opening/closing 115 CAD ATDC/245 CAD ATDC supercharging layout must therefore take into account its behavior
Exhaust valves opening/closing 80 CAD ATDC/250 CAD ATDC in terms of SE and TE.
 A.P. Carlucci et al. / Energy Conversion and Management 101 (2015) 470–480 473

(4 cylinders)

(3 plenum)

(2 aercooler)

(5 turbine) (1 compressor)

Fig. 5. Scheme of one bank of the prototype used during the experiments.

Fig. 4. Scavenging model for Uniflow system compared with perfect mixing and
perfect displacement models.
The scavenging models define SE as a function of the Delivery
Ratio (DR), which is the ratio between mas and the air mass, at
ambient conditions, ideally required to fill the engine displacement
(equal to the product between the ambient density and the cylin-
der displacement). In this work, a scavenging model, correspond-
ing to Eq. (4) and shown in Fig. 4, has been implemented. This
model is halfway between the ideal models of perfect mixing and
perfect displacement, as suggested in literature [16] on the basis
of experimental tests on 2-stroke Uniflow Diesel engines:
SE ¼ 1  expð4:35 þ 1:36DR  0:22DR2 Þ
In order to describe the combustion development, the Vibe
function was used [1,15] thanks to its easiness. The Vibe function
models the heat release rate into the cylinder as a function of
CAD based only on the start of combustion angle, combustion
angular duration and a shape parameter.
Regarding the supercharging system, each component can be
simulated either in simplified or full mode. The simplified calcula-
tion is based on the power balance and requires as input the
desired value for the compressor pressure ratio p2/p1; referring to
turbochargers (compressor mechanically coupled to turbine), the
power balance leads to the following equation:
"  k1 #
p k 1 1
_ T  gsT  cpT  T 3  1  4
m _C
m  c T
p3 gm gsC pC 1
" k1 # V
p2 k
  1  tcPLUS ¼ 0
p1
tcPLUS, defined as the difference between the power produced by
(4) through the inlet ports. After the combustion process, the
exhaust gases are discharged in the exhaust ducts and then at
the turbine inlet.
In order to run the model, the values for the fuel mass flow rate
(14 g/s, measured by means of a AVL 733S fuel balance) and pres-
sure ratio of the compressor (1.85, obtained averaging over time
the instantaneous pressure measured downstream and upstream
the compressor by means of two AVL piezoresistive sensors type
LP11DA) have been set equal to those measured during the exper-
iments on a real engine prototype performed at Sea Level. Engine
output power was measured by means of a Borghi & Saveri
eddy-current dynamometer, while the air mass flow was estimated
by means of a calibrated diaphragm.
According to real data, the characteristic settings of the model
ð4Þ are reported in Table 3.
Moreover, the engine friction mean effective pressure – exper-
imentally set during motored conditions on a range of ‘‘n’’ – varies
almost linearly with ‘‘n’’ and is equal to 2.6 bar when n = 2000 rpm.
The engine model, and the Vibe model of the combustion phase,
have been calibrated with experimental data acquired on the real
engine prototype, operating at n = 2000 rpm and full load. In detail,
the angular start of combustion was set equal to 13 CAD ATDC,
the combustion angular duration was set equal to 80 CAD while
the shape parameter was fixed equal to 1.5. In Table 4, the relative
Table 3
Characteristic settings of the model.
Aftercooler
19.2 L
e 0.7
ð5Þ
Dp Varying with the square of the volumetric flow rate
and equal to 0.2 bar in operating conditions used for
the model calibration

the turbine and the power required by the compressor, takes into Intake ports
account the possible production or demand of mechanical/electric Number per cylinder 14
Maximum cylinder 3700 mm2
power.
overall area
The full mode, instead, is based on the dynamic balance of tur- CA variation See Fig. 1
bomachineries, and therefore on the characteristic maps of turbine Exhaust valves
and compressor. Number per cylinder 2
Maximum lift 12 mm
CA variation See Fig. 2
3. Settings and calibration of the engine model for the
simulation of the different architectures Turbocharger
Model Simplified
gsC 0.7 (according to compressor and turbine
In Fig. 5, the model of the engine in its initial configuration is characteristic maps)
shown. In particular, a compressor (1) driven by the turbine (5) gsT 0.65 (according to compressor and turbine
delivers the compressed air in the aftercooler (2) and then in the characteristic maps)
gm 0.98 (as suggested in [15])
plenum (3); from the plenum, the air enters the three cylinders
474 A.P. Carlucci et al. / Energy Conversion and Management 101 (2015) 470–480

error of the computed with respect to the measured values is Table 5

reported. The model was then used as basis for calculating the out- List of the supercharging architectures and targets of the simulations.

put power and specific fuel consumption of the engine when Architecture Computational Goals
supercharged with different supercharging systems at different settings
altitudes. Once assumed the symmetry and independence of the TC (Fig. 5) One bank, 1 TC and AFRcmb equal to Calculation of the
two banks, the analysis was done on a single bank to reduce the one aftercooler per 20; TC in engine
computing duty. bank simplified mode; performance once
n = 2000 rpm; fixed the
Table 5 summarizes the supercharging architectures and the simulations at 0, supercharging
targets of the simulations. During the simulations, the trapped 3050, 5180, conditions
Air/Fuel Ratio (AFRcmb) was always imposed equal to 20 if not para- 10,670 m altitude
metrized (like in TTCHybrid system, please refer to Table 5). This TTC-1, One bank, 2 TCs AFRcmb equal to Evaluation of the
choice was made because the amount of fuel burning in the engine TTC-2 and one aftercooler 20; TC in supercharging
depends on the air trapped in the cylinder, and consequently on (Fig. 7) (TTC-1) or simplified mode; conditions (mass
intercooler (TTC-2) n = 2000 rpm; flow and
the supercharging architecture.
per bank simulations at 0, compression ratio)
In all the previous architectures, the presence of the WG valve 3050, 5180, required to
was not considered. This decision was taken because the goal of 10,670 m altitude guarantee the
this analysis was to determine the supercharging system in work- target
ing conditions at maximum efficiency. performance;
comparison
between
architectures with
4. Results and discussion aftercooler (TTC-1)
and intercooler
4.1. Performance with one TC per bank (TC model) (TTC-2)
TCC, CTC One bank, 1 TC, 1 AFRcmb equal to Evaluation of the
The scheme with one TC and one aftercooler per bank (Fig. 5) (Fig. 10) mechanical 20; TC in supercharging
has been simulated setting the model as specified in Table 5. In compressor on LP simplified mode; conditions (mass
(CTC) or HP (TCC) n = 2000 rpm; flow and
Fig. 6a and following, Pb is reported in % of Pb target at sea level
side and one simulations at 0, compression ratio)
for different compressor pressure ratios and different operating aftercooler per 3050, 5180, required to
altitudes; in the same figure, the Pb targets for the different alti- bank 10,670 m altitude guarantee the
tudes are reported. In Fig. 6b and following, the engine relative target
BSFC are also reported in % of a target value calculated as average performance;
comparison
of the target values at different altitudes. It can be concluded that between
TC supercharging system does not guarantee the achievement of Pb architectures with
target operating with acceptable compressor ratios. In fact, while the mechanical
the target at sea level and 3050 m could be reached with a com- compressor on LP
(CTC) and HP (TCC)
pressor pressure ratio respectively equal to 3 and 4.2, the target
side with respect
at 5180 and 10,670 m would require a compressor pressure ratio to the TC
higher than 4.5; this value cannot be currently reached by a single
TTCHybrid One bank, 1 HP TC HP TC in complete Performance
compressor. It is therefore necessary to study alternative super- (Fig. 13) and one aftercooler mode with a evaluation for
charging architectures. per bank, 1 LP commercial map; different ‘‘n’’ (1500
compressor and 1 simulation at 0, and 2000 rpm) and
LP turbine 3050, 5180, altitudes; maps
4.2. Performance with two turbochargers (TTC model) connected to the 10,670 m altitude; setting for the EAT
engine in order to pressure ratio of and fuel mass flow
represent an EAT electric rate management
Two architectures with two turbochargers have been compared
system compressor in order to
with different position of the cooler: as aftercooler (TTC-1 model, (eCOM) and guarantee the Pb
see Fig. 7a); as intercooler (TTC-2 model, see Fig. 7b). AFRcmb target for all the
Considering the limitations in weight and bulk, the possibility parametrized operating
of having a second cooler per bank was not contemplated. conditions
Scheme TTC-1 (a), theoretically identical to TC scheme, has been
analyzed to highlight that the two compressors would work with
compression ratios acceptable for the current state of the art in
order to reach the engine output targets.
the HP compressor pressure ratio is slightly lower than the LP com-
The model was set as specified in Table 5, while Fig. 7 also
pressor pressure ratio. As reported by Liu et al. [9], this is also the
reports the compressor pressure ratios used for simulations.
working condition corresponding to the lowest compression work.
These values not only ensure the achievement of the Pb target at
In Fig. 8, the behavior of the two architectures is compared.
different altitudes; they also correspond to the working condition
The scheme with the intercooler (TTC-2) leads to a longer life of
with the lowest BSFC. In particular, this condition is obtained when
the second compressor, which works at a lower temperature. The
scheme with aftercooler (TTC-1), on the contrary, leads to better
Table 4 engine efficiency. In fact Fig. 8a shows that BSFC is lower with
Relative error of the computed with respect to the measured values. TTC-1 architecture than TTC-2. Since TTC-1 is also characterized
Value Relative error [%] by a higher TE (see Fig. 8b), the final result is that the compressed
air mass flow rate in TTC-1 is lower than in TTC-2. Additionally,
Engine brake power (Pb) 1.8
Brake specific fuel consumption (BSFC) +1.9 using TTC-1 requires lower overall compression ratio (Fig. 8c),
Air mass flow 10.0 and therefore a lower supercharging power in order to reach the
Turbine pressure ratio +4.2 same Pb target.
 A.P. Carlucci et al. / Energy Conversion and Management 101 (2015) 470–480 475

120 125
(a) (b)
100 100

80 75

ΔBSFC [%]
%Pb@sea level

60 50

40 25

20 0
1.5 2.0 2.5 3.0 3.5 4.0 4.5
0 - 25
1.5 2.0 2.5 3.0 3.5 4.0 4.5 Compressor Pressure Ratio [-]
Compressor Pressure Ratio [-]
TC-0 TC-3050 TC-5180 TC-10670 TC-0 TC-3050 TC-5180 TC-10670
target 0-3050 m target 5180 m target 10670 m

Fig. 6. Engine brake power Pb (a) and brake specific fuel consumption BSFC (b) at different altitudes and compression ratios with TC architecture.

(4 cylinders) (4 cylinders)
(a) (b)
(3 plenum) (3 plenum)

(2 aftercooler)

(5’ –HP turbine) (5’ –HP turbine)

(1’ –HP compressor) (1’ –HP compressor)

(2 intercooler)
(5 –LP turbine) (5 –LP turbine)
(1 –LP compressor) (1 –LP compressor)

altitude Low pressure – high pressure altitude Low pressure – high pressure

0 [m] 1.75 1.70 0 [m] 2.00 1.90

3050 [m] 2.20 1.95 3050 [m] 2.50 2.30
5180 [m] 2.30 2.20 5180 [m] 2.80 2.50
10670 [m] 3.00 2.30 10670 [m] 3.20 3.00

Fig. 7. TTC-1 (a) e TTC-2, (b) schemes. In the table, the values of low pressure and high pressure compressors ratio used during simulations.

The reason of this behavior lies in the architecture of the com-
pression system; in particular, once the compression ratios and
efficiency and the cooler efficiency are fixed, the architecture with
intercooler causes a higher outlet (plenum) temperature; this, in
turn, leads to less air mass trapped into the cylinders, to a lower
amount of fuel burned and lower output power. Therefore, to
increase the air mass trapped, the outlet pressure must be
increased, this leading to the higher compression ratios values
for TTC-2 architecture scheme reported in Fig. 7. This is expected
to cause the difference between the pressure at cylinder inlet
and exhaust to increase. In Fig. 9, the difference between the pres-
sure at cylinder inlet and exhaust is compared during the scaveng-
ing phase with TTC-1 and TTC-2 schemes for the lowest and
highest operating altitude. As expected, this difference is higher
with TTC-2 with beneficial effects on scavenging efficiency or AP
(Fig. 8d) but detrimental effects on TE. The overall effect is there-
fore a better behavior of TTC-1 scheme, becoming more evident
increasing the altitude. This scheme has been chosen for further
considerations.
4.3. Performance with a mechanical compressor and a TC per bank
To obtain a lower bulk and weight configuration a supercharg-
ing system with a TC, a mechanical compressor on the LP side (CTC,
Fig. 10a) or on the HP side (TCC, Fig. 10b) and an aftercooler per
bank has been analyzed.
The model was set as specified in Table 5. In Fig. 10, the values
of the compression ratio set for the TC and the mechanical com-
pressor are reported. Similarly to TTC simulations, these values
not only guarantee the attainment of the Pb target at different alti-
tudes, they also correspond to the working condition with the low-
est BSFC. In Fig. 11, the behavior of the two systems is compared.
Still in Fig. 11, the results obtained with TTC-1 scheme are reported
for comparison purposes.
As well known, the power required by the mechanical compres-
sor is provided by the engine crankshaft. Having the power to drive
the mechanical compressor, the indicated power must be higher
than the target value. Therefore, a higher quantity of air and fuel
must be trapped and injected into the cylinder. Consequently, a
476 A.P. Carlucci et al. / Energy Conversion and Management 101 (2015) 470–480

(a) 15 (b) 0.8

10
BSFC [%] 0.7

TE [ ]
5
0.6
0
TTC 1
TTC 1
TTC 2 TTC 2
5 0.5
0 3000 6000 9000 12000 0 3000 6000 9000 12000
altitude [m] altitude [m]

(c) 11
(d) 0.94
Overall Compressor
Pressure Ratio [ ]

9
0.93
7

AP [ ]
0.92
5
0.91
3 TTC 1
TTC 1
TTC 2 TTC 2
1 0.90
0 3000 6000 9000 12000 0 3000 6000 9000 12000

altitude [m] altitude [m]

Fig. 8. Comparison between performance related to TTC-1 e TTC-2 schemes.

turbine determines a lower backpressure at the cylinder exit with

0.3
0 TTC 1
0 TTC 2
respect to the double turbine TTC scheme, thus penalizing even
0.15 10670 TTC 1 further the TE with a consequent short-circuiting of the fresh
10670 TTC 2
charge. Based on the previous analysis, it can be concluded that
pinlet pexhaust [bar]

0 using a mechanical compressor leads to higher engine BSFC; there-

fore, although TCC and CTC might be characterized by a lower
0.15 weight, bulk and intake and exhaust system complexity, TTC-1
scheme was selected for further energy utilization improvements.
0.3

0.45 IPO IPC

0.6
80 100 120 140 160
CAD ATDC
Fig. 9. Difference between the pressure at cylinder inlet and exhaust during the
scavenging phase; comparison between TTC-1 and TTC-2 architectures for the
lowest and highest operating altitude.
worsening of the BSFC is to be expected when employing the
mechanical compressor compared to the TTC architecture. This is
evident in Fig. 11a, in which it is also highlighted that the lowest
specific fuel consumption between TCC and CTC schemes is
obtained with the mechanical compressor positioned on the LP
side. In these conditions, indeed, the air enters the mechanical
compressor at a lower intake temperature, which means that less
compression power is required. The need to trap a higher amount
of air is achieved increasing the pressure in the plenum as visible in
Fig. 11b where the mean values of the pressure in the plenum for
TCC, CTC and TCC-1 architectures are plotted. As a consequence, TE
worsens (Fig. 11c) while the AP increases with TCC (Fig. 11d). This
behavior can be further explained analyzing the difference
between the pressure at cylinder inlet and exhaust for the three
schemes (Fig. 12); as expected, this difference is highest for TCC
and lowest for TCC-1, during all the scavenging phase.
Lastly, it can be observed that, with TCC and CTC schemes, only
one expansion stage is used; this determines an unacceptable
operation of the turbine, that has to work with expansion ratios
as high as 6 at highest altitudes (10,670 m). Besides, a single
4.4. Performance with hybrid double stage per bank (TTCHybrid)
BDC
180 200 220 240 The last architecture analyzed is the TTCHybrid. The TTCHybrid
scheme was composed by two supercharging levels, with a classic
TC on the HP side and a compressor (named eCOM) and a turbine
(eTURB) on the LP side, connected electrically by means of a
motor/generator system (EAT configuration). This system was orig-
inally proposed either to increase the power transfer to the com-
pressor to achieve a higher boost pressure and to accelerate the
TC in response to an acceleration requirement [18]. This system
is proposed in this study because it allows to uncouple the speed
of the eCOM from that of the eTURB and also allows to manage
the energy fluxes. In fact, with tcPLUS > 0 it would be possible to
store energy in a battery pack able to give it back in case
tcPLUS < 0. However, this system causes a penalty in terms of
weight and bulk of the supercharging system. In this case the pres-
ence of the WG valve was not considered to minimize the engine
fuel consumption, and also because, with this scheme, the tur-
bochargers speed can be controlled acting on the motor/generator
system.
In Fig. 13, the model of the TTCHybrid scheme is shown; the
mechanical coupling between the eCOM and the eTURB in AVL
BOOST software can be realized with the thermal engine but not
with an electric motor/generator; therefore, the output power of
the engine, calculated by the code, takes into account for the
tcPLUS. The net power produced by the engine only will be equal
to the difference between the power calculated with the model
and the value of tcPLUS.
In order to evaluate the engine performance and the energy
fluxes of the LP supercharging system (eCOM + eTURB), several
 A.P. Carlucci et al. / Energy Conversion and Management 101 (2015) 470–480 477

(4 cylinders) (4 cylinders)

(3 plenum) (3 plenum)

(2 aftercooler) (2 aftercooler)

(1 –LP compressor)

(1’ –HP compressor)

(5 –LP turbine) (5 –LP turbine) (1’ –HP compressor)
(1 –LP compressor)

(a) (b)
altitude Turbocharger – mechanical compr. altitude Turbocharger – mechanical compr.

0 [m] 2.60 1.50 0 [m] 2.30 1.60

3050 [m] 3.60 1.40 3050 [m] 3.00 1.70
5180 [m] 4.3 1.70 5180 [m] 3.50 1.70
10670 [m] 5.00 1.40 10670 [m] 4.00 2.20

Fig. 10. TCC (a) e CTC (b) schemes. In the table, the values of turbocharger and mechanical compressor ratio used during simulations.

(a) 40
TTC 1
(b) 5
TTC 1
30 TCC 4 TCC
CTC CTC
BSFC [%]

20
pplenum [bar]

10 2

0 1

10 0
0 3000 6000 9000 12000 0 3000 6000 9000 12000
altitude [m] altitude [m]

(c) 0.80 (d) 0.97

0.95
0.70
TTC 1
AP [ ]
TE [ ]

TCC 0.93
CTC
0.60
0.91 TTC 1
TCC
CTC
0.50 0.89
0 3000 6000 9000 12000 0 3000 6000 9000 12000
altitude [m] altitude [m]

Fig. 11. Performance comparison among TTC-1, CTC e TCC schemes.

simulations have been run varying the load at different ‘‘n’’ (1500
and 2000 rpm) and altitudes. The power is that required by the
propeller directly coupled to the engine crankshaft, whose varia-
tion with ‘‘n’’ is described by the Renard law:
PðnÞ ¼ C P qa n3 R5
Knowing the values of the power required at the different alti-
tudes at 2000 rpm (target values), it is possible to know the power
required at 1500 rpm as:
 n 3
PðnÞ ¼ P n¼2000  ð7Þ
2000
Concerning the HP TC, the full mode has been used, as previ-
ously specified in Table 5. This choice was made since the aim of
these simulations was to describe the system behavior at part
ð6Þ loads, where the turbomachineries cannot be assumed working
at design point. The full mode has been implemented based on
commercial maps for the turbomachineries, and chosen consider-
ing the mass flow rates and compression ratios calculated with
478 A.P. Carlucci et al. / Energy Conversion and Management 101 (2015) 470–480

0.9
0.75 0 TTC 1
0 CTC
0.6 0 TTC
pinlet pexhaust [bar]

0.45
0.3
0.15
0
0.15
0.3
IPO IPC
0.45
BDC
0.6
80 100 120 140 160 180 200 220 240
CAD ATDC

Fig. 12. Difference between the pressure at cylinder inlet and exhaust during the
scavenging phase; comparison between TTC-1, CTC and TCC architectures at sea
level.

(4 cylinders)

(3 plenum)

(2 aercooler)

(6 –mech.
conn. between
LP TC and
engine shas) tcPLUS<0 (1’ – HP compressor)
tcPLUS>0

(5’ – HP turbine)
(1 – LP compressor)

(5 – LP turbine)

Fig. 13. TTCHybrid scheme. tcPLUS > 0 if the power produced by the turbine
exceeds the power required by the compressor; tcPLUS < 0 if the power required by
the compressor exceeds the power produced by the turbine.

TTC-1 model; in particular, the maps of the selected TC are Fig. 14. Compressor and turbine characteristic maps of the chosen commercial
reported in Fig. 14. In this way, no assumptions were required con- turbocharger.
cerning the value of the pressure ratio of the HP compressor.
Moreover, a TC mechanical efficiency of 0.98 for the HP TC has
been set [15].
As for the LP side of the supercharging system, given the double
electric connection, a mechanical efficiency equal to 0.97 has been
% Pb@sealevel
used both on the compressor and on the turbine sides, resulting in
a 0.95 overall mechanical efficiency. During the simulations, the
pressure ratio of the eCOM was parameterized in the range 1–2. 160
A further limit on the maximum power required by the eCOM 120
equal to 10 kW has been imposed for design constraints of the
electric system. Finally, in order to vary the engine load, AFRcmb 80
has been varied in the range 20–35. 40
In this way, a 3D surface has been obtained for each value of
altitude and ‘‘n’’, fitting the power delivered by the engine for each
value of both AFRcmb and eCOM pressure ratio; this surface, inter-
sected with a planar surface at constant power equal to Pb target,
gives the values of PR,eCOM and AFRcmb corresponding to the equi-
librium engine and load (see for example Fig. 15). Among these val-
Fig. 15. 3D surface plotting engine brake power as a function of both AFRcmb and
ues, excluding those related to a power of the eCOM higher than eCOM pressure ratio for n = 2000 rpm at Sea Level (the dotted line is the
10 kW, it is possible to plot the graphs shown in the following intersection between the 3D surface and the Pb = Pb@sea level plane).
 A.P. Carlucci et al. / Energy Conversion and Management 101 (2015) 470–480 479

+53.8 1.5 5 +34.6 2 10

BSFC [%]
PR,eCOM [-]

PR,eCOM [-]
tcPLUS [kW]
BSFC [%]

tcPLUS [kW]
+15.4 1.25 +25.0 1.5 0

-5

-23
0 [m] 1 10670 [m]
-10 +15.4 1 -10
20 23.3 26.6 30 20 24 28 32
AFRcmb [-] AFRcmb [-]

Fig. 16. BSFC, tcPLUS and PR,eCOM at different altitudes for a constant output power at 1500 rpm.

-4.5846 1.394 0.8 +8.84 2 1.5

PR,eCOM [-]
BSFC [%]
BSFC [%]

tcPLUS [kW]
tcPLUS [kW]

tcPLUS [kW]
-4.5838 1.385 0.65 +8.78 1.9 0.5

0 [m] 10670 [m]

-4.5829 1.376 0.5 +8.72 1.8 -0.5
20.10 20.16 20.22 20.30 20 20.47 20.93 21.4
AFRcmb [-] AFRcmb [-]

Fig. 17. BSFC, tcPLUS and PR,eCOM at different altitudes for a constant output power at 2000 rpm.

34 conditions, the eCOM works with a compression ratio lower than

0 1.3. However, these points are characterized by higher values of
26 fuel consumption (see Fig. 16). Consequently, the engine could be
3050
managed by either minimizing the fuel consumption (with no
5180
BSFC [%]

18 energy production or demand by the EAT) or producing energy

10670
10
2 The operation at 2000 rpm corresponds to the cruise of the air-
6 almost equal to 20 and PR,eCOM about 1.4, except at 10,670 [m]
TCC CTC TTC 2 TTC 1 TC
supercharging architecture
Fig. 18. Comparison of the relative best BSFC for the different architectures
considered in this paper, all simulated at engine target output conditions for the
different altitudes.
figures, where, for each altitude, ‘‘n’’ and AFRcmb, the PR,eCOM
required to obtain the Pb target is shown.
The BSFC and tcPLUS have been also estimated running the
engine model for the particular values of Pb, AFRcmb and eCOM
pressure ratio. For this architecture, the BSFC was calculated con-
sidering both the power output of the crankshaft (always positive)
and the tcPLUS (positive or negative). For example, referring to
Fig. 16, at sea level, to obtain the target power with an AFRcmb
equal to 23.3, a PR_eCOM equal to 1.15 will be required. This will
also determine a BSFC penalty of about 3%. At the same time, a
power surplus of the EAT equal to 0.2 kW will characterize the
whole system. With this tool it is possible to define a control strat-
egy (energy storage/request) of the EAT, based on the power sur-
plus tcPLUS and on the pressure ratio of the eCOM.
from the TC with a penalty on fuel consumption.
4.4.2. Results at 2000 rpm
craft. It shows a limited operating range at Pb target with AFRcmb
TTChybrid with PR,eCOM between 1.8 and 2 (see Fig. 17). BSFC is anyway rel-
atively low and results show values substantially equal to TTC-1
scheme. These operating points allow to produce energy from
the EAT as well; therefore, it is convenient to manage the system
in these conditions in order to work with the maximum possible
overall power output, being the fuel consumption basically con-
stant. The TTCHybrid scheme allows, therefore, to manage the LP
supercharging system, being able to act directly on the working
point of the compressor thanks to the decoupling of the compres-
sor and turbine shafts; moreover, the possibility of storage in the
batteries allows to optimally manage the energy fluxes varying
the load during flight cycles, avoiding the energy wastes induced
by other control logics, like the WG. However, compared to
TTC-1 case, even if characterized by a lower fuel consumption at
2000 rpm, this system is more complex and characterized by
higher weight and bulk. Summarizing, Fig. 18 compares the rela-
tive best BSFC for the different architectures considered, all simu-
lated at engine target output conditions for the different altitudes.
Data referring to TC configuration at 5180 m and 10,670 m are not
reported since, as previously observed, the target power cannot be
reached with a compressor pressure ratio lower than 4.5.

4.4.1. Results at 1500 rpm 5. Conclusions

The simulations at 1500 rpm show the possibility to produce
energy from the TC up to 3 kW with AFRcmb in the range 20–23 The aim of the present work is to compare different architec-
and only at the highest operating altitudes; under such working tures – single TC, double TC, single TC combined with a mechanical
480 A.P. Carlucci et al. / Energy Conversion and Management 101 (2015) 470–480
compressor, EAT, with intercooler or aftercooler. This engine must
be characterized by a very high altitude operation and long fuel
distance. Therefore, a 1D model of the engine has been realized
with AVL BOOST software. The performance of the above men-
tioned supercharging systems have been compared to determine
the most performing one in terms of power and specific fuel con-
sumption to supercharge an aircraft 2-stroke Diesel engine for gen-
eral aviation and UAV. Moreover, the effect of the EAT compression
ratio on the power produced or required by the EAT at part load
has been also analyzed to define best performing the control
strategies in terms of fuel consumption.
The analysis with the two turbochargers system showed that it
was possible to reach the target power for each altitude. Moreover,
positioning the air cooler after the two compression stages reduces
the specific fuel consumption. The two turbochargers architecture
was shown to be more effective even if compared with architec-
tures in which a single TC was used together with a mechanical
compressor.
Conclusively, a hybrid supercharging system has been analyzed.
It comprised a classic HP TC and a LP compressor and turbine elec-
trically connected. The analysis revealed that the hybrid super-
charging architecture is characterized by a fuel consumption
lower than that of the two turbochargers architecture and compa-
rable with single TC architecture; compared to this last one, the
hybrid architecture is also able to reach the power targets for all
operation altitudes. Though more complex and high in weight
and bulk, this scheme allows to extract power from the EAT with-
out a penalty on fuel consumption.
References
[1] Heywood J. Internal combustion engine fundamentals. McGraw-Hill; 1988.
[2] Ravi M, Marathe A. Effect of inlet and exhaust pressures on the scavenging
characteristics of a carbureted uniflow scavenged engine. SAE Paper 920840;
1992.
[3] Ravi M, Marathe A. Effect of port sizes and timings on the scavenging
characteristics of a uniflow scavenged engine. SAE Paper 920782; 1992.
[4] Pohorelsky L, Brynych P, Macek J. Air system conception for a downsized two-
stroke diesel engine. SAE Paper 2012-01-0831; 2012.
[5] Knoll R. AVL two-stroke diesel engine. SAE Paper 981038; 1998.
[6] Mattarelli E. Comparison among different 2-stage supercharging systems for
HSDI diesel engines. SAE Paper 2009-24-0072; 2009.
[7] Mattarelli E, Rinaldini CA, Mazza A, Oliva M. Development of a 2-stage
supercharging system for a HSDI diesel engine. SAE Paper 2009-01-2757;
2009.
[8] Chen J, Zhuge W, Zheng X, Zhang Y. Investigation of influence of two-stage
turbocharging system on engine performance using a pre-design model. In:
Proceedings of ASME Turbo Expo 2013: turbine technical conference and
exposition GT2013, San Antonio, Texas, USA; June 3–7, 2013.
[9] Liu Y, Zhuge W, Zhang Y, Zhang S, Zhang J, Huo X. A matching method for two-
stage turbocharging system. In: Proceedings of ASME Turbo Expo 2014:
turbine technical conference and exposition GT2014, Dusseldorf, Germany;
June 16–20, 2014.
[10] Shan P, Zhou Y, Zhu D. Mathematical model of two-stage turbocharging
gasoline engine propeller propulsion system and analysis of its flying
characteristic. In: Proceedings of ASME Turbo Expo 2014: turbine technical
conference and exposition GT2014, Dusseldorf, Germany; June 16–20, 2014.
[11] Divekar PS, Ayalew B, Prucka R. Coordinated Electric supercharging and turbo-
generation for a diesel engine. SAE Paper 2010-10-1228; 2010.
[12] Katrasnik T, Medica V, Trenc F. Analysis of the dynamic response improvement
of a turbocharged diesel engine driven alternating current generating set.
Energy Convers Manage 2005;46:2838–55.
[13] Terdich N, Romagnoli A, Martinez-Botas RF, Pesiridis A. Mild hybridization via
electrification of the air system: electrically assisted and variable geometry
turbocharging impact on an off-road diesel engine. In: Proceedings of ASME
Turbo Expo 2013: turbine technical conference and exposition GT2013, San
Antonio, Texas, USA; June 3–7, 2013.
[14] Mamalis S, Nair V, Andruskiewicz P, Assanis D, Babajimopoulos A. Comparison
of different boosting strategies for homogeneous charge compression ignition
engines – a modeling study. SAE Paper 2010-01-0571; 2010.
[15] Avl BOOST, Users Guide for AVL BOOST – AVL List GmbH; 2008.
[16] Blair GP. Design and simulation of two-stroke engines. Warrendale: SAE
International; 1996.
[17] Heywood JB, Sher E. The two-stroke cycle engine: its development, operation
and design. USA: Taylor and Francis; 1999.
[18] Baines NC. Fundamentals of turbocharging. USA: Concepts NREC; 2005.