Energy 68 (2014) 877e885

Contents lists available at ScienceDirect

Energy
journal homepage: www.elsevier.com/locate/energy

Dynamic heat transfer model for temperature drop analysis and heat
exchange system design of the air-powered engine system
Qiyue Xu, Maolin Cai, Yan Shi*
School of Automation Science and Electrical Engineering, Beijing University of Aeronautics and Astronautics, Beijing 100191, China

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 19 November 2013
Received in revised form
18 February 2014
Accepted 27 February 2014
Available online 21 March 2014

In the operation process of an air-powered engine (APE) system, temperature drops happening in
critical locations can inuence the engines performance negatively, and even lead to the ice blocking
problem. To predict temperature drops during the operation, rstly, the thermodynamic model of
the APE and a calculation method for equivalent air temperatures at intake and exhaust ports are
described. The cooling mechanism of the pressure-reducing process is analyzed. Then a simplied
calculation model of the throttling effect for dynamic temperature analysis is proposed. Furthermore, a complete dynamic model of the APE system is established, by considering models mentioned
above and models of the pressure tank and the supply pipeline as well. The models feasibility on the
temperature drop analysis is veried by comparing with corresponding experiments. Simulation of a
practical APE system is carried out. Under specic parameter settings, temperature drops of critical
locations in the system are predicted. On this basis, the supply system of compressed air is modied
and a principle structure of the heat exchange system for the APE system is proposed. The
analysis results in this paper can provide a theoretical support for the design of the heat exchange
system.

2014 Elsevier Ltd. All rights reserved.

Keywords:
Air-powered engine
Temperature drop
Throttling effect
Dynamic model
Heat exchange

1. Introduction
Under the background of energy crisis and environmental issue,
it has become increasingly important to nd environmentally
friendly energies to replace fossil fuels [1,2]. The compressed air, as
a kind of medium, is considered to have superior energy storage
density to other media like batteries. It can be easily obtained from
the power generation process of renewable energies, such as solar
energy, wind energy and tidal energy [3e6]. The air-powered engine (APE) is driven by compressed air. As a lightweight, nonpolluting and safe power device, it can be applied to transportation, general aviation and small-scale power generation in the
future. Especially, serving as engines of motor vehicles, the APE
discharges breathable air at low temperature and low pressure, so
the operation of the APE can achieve zero carbon emissions, which
is signicant for air pollution abatement in some industrial country,
especially in China.

* Corresponding author. Tel.: 86 15810834177.

E-mail address: yesoyou@163.com (Y. Shi).
http://dx.doi.org/10.1016/j.energy.2014.02.102
0360-5442/
2014 Elsevier Ltd. All rights reserved.

In the past 10 years, MDI, a French company of air-powered

vehicles, developed a series of APEs [7,8]. In China, some universities and companies have also been conducting researches and
designs on kinds of APEs [9e13]. But generally limited by problems of low working efciency and low temperature, the APE is
still in developing stage. Low temperature due to the airs
expansion and the local throttling can lead to poorer power performance of the compressed air, and can even result in ice
blockings at critical locations of the APE system which is a problem that cannot be ignored. Liu Hao and Chih Yung Huang both
mentioned that there was a certain temperature decrease in the
cylinder of the APE and higher inlet pressures resulted in lower
temperatures [14,15]. Zhai Xin reported that increasing the inlet
temperature is an effective way to enhance the engines power
and efciency, and Zhai also studied the optimization of the APEs
inlet tube for the best heat transfer performance [16]. But few
studies established complete heat transfer models of the APE
system, as well as methods to calculate the air temperature drops
at critical locations, especially the high pressure reducing valve
and the exhaust port of the APE.
Several studies provided references for the temperature analysis
of the compressed air system. The JouleeThomson coefcient was

878

Q. Xu et al. / Energy 68 (2014) 877e885

used to calculate the temperature changes of the throttling [17e

20]. The real gas effect on temperature was considered in the discharging process of high pressure vessels [21,22]. But most of these
methods need to solve transcendental equations which have large
calculation complexity, thus they are not suitable for modeling of
dynamic pneumatic systems.
This paper focuses on the temperature drop analysis of the APE
system. Dynamic air temperature calculation methods of locations
including the APEs ports and the throttling pressure reducer are
proposed. Then the complete dynamic heat transfer model of the
APE system can be established for temperature analyses and the
design of the heat transfer system. Results of this paper are helpful
to predict temperature drops in the practical APE system, to prevent the ice blocking problem and to improve the performance of
the APE system.
2. Introduction of the APE system
2.1. Working principle of the APE
For a piston-type APE, the compressed air expands in the cylinder, pushing the piston to output shaft power. Its operation is
shown in Fig. 1: in the suction power stroke the compressed air
enters the cylinder through the intake valve, driving the piston
downward. Then the intake valve closes after a specic crank
angle while the compressed air expands to push the piston down
and output work. When the piston is near the bottom dead center
the exhaust valve opens so that the air with residual pressure
discharges under the impetus of the piston. After the piston
moves back to the top dead center, the APE completes a work
cycle.
2.2. Locations of the temperature drop
This paper only concerns the piston-type APE. However, for any
kind of APE, the compressed air goes through it will cool due to the
expansion power process.
At room temperature, the real gases except hydrogen, helium
and neon cool upon expansion by the JouleeThomson (JeT) effect.
Therefore when the compressed air discharges from a container or
is forced through a pressure-reducing valve it has temperature

drops. In the APE system, apart from the cylinder of the APE, the
high pressure tank and the pressure reducer are typical locations
where the air temperatures are predicted to decrease. Though the
compressed air experiences dry processing, it cannot be guaranteed
100% without water. Thus when the temperature of air is lower
than the dew point, the locations mentioned above are considerably possible to have ice blockings. In the following sections,
thermodynamic models of these locations are established for
analysis.
3. Thermodynamic model of the APE and calculation method
of the equivalent air temperatures
3.1. Thermodynamic model of the APE
The model based on the crank angle was built at rst. In the
modeling process, the following assumptions are made.
(1) The compressed air is ideal, which means specic heat u and
specic enthalpy h are only related to the temperature.
(2) The air in the cylinder is uniform during the thermodynamic
process.
(3) The air ow in and out of the cylinder is considered as quasisteady, one-dimension and isentropic.
(4) The airs kinetic energy and potential energy are ignored.
(5) There is no leak during the working cycle.

3.1.1. Energy equation

The energy equation can be given in the form of temperature
differential equation:


dT
1
dV
ct Ah fTa  T h1 G1 h2 G2  p

 uG
df
mCv
df

(1)

where G1 dm1/df, G2 dm2/df, G dm/df, m is the mass of the

air in the cylinder, Cv is the constant volume specic heat, ct is the
heat transfer coefcient, Ah(f), Ta are the total heat transfer area
and the temperature of the internal walls, respectively, here Ta is
assumed to be equal to the room temperature, T is the temperature of the air in the cylinder, h1, h2 are the specic enthalpies of
the air ow in and out of the cylinder, respectively, m1, m2 are the
mass of the air ow in and out, respectively, V is the instant volume of the cylinder, u is the specic internal energy and f is the
crank angle.
The change rate of the volume is described by:

p
lsin fcos f 7
dV
6
D2 S4sin f q5
df
8
2
1  l sin2 f

(2)

where D is the diameter of the cylinder, S is the stroke of the piston,

l is the crank ratio.
The total heat transfer area is:

Ah

Fig. 1. Working cycle of the APE.

D2


q
1
2
DS 1  cosf
1  1  l sin2 f

(3)

3.1.2. Continuity equation

The intake and exhaust airow rate of the APE can be calculated
as follows:

Q. Xu et al. / Energy 68 (2014) 877e885

Gi

v
0 1 u
u
k 1

u
1 @ A t k
2
k  1;
A f pi
u
Rg Ti k 1

8
>
>
>
>
>
>
>
>
>
<

po

pi

v
>
>
0 1 u
>
" 2  k 1 #
u
>
>
>
1 @ A u
2k 1
po k
po
>
t
k
>
>

;
: u A f pi k  1 R T
p
p
g i

 k
2
k1
k1
(4)

 k

po
2
k1
>
k1
pi

where u is the angular speed of the crank, k is the adiabatic

exponent of the air, Rg is the gas constant, pi, Ti are the upstream
pressure and temperature, respectively, po is the downstream
pressure, A(f) is the effective sectional area of the intake or exhaust
valve:

Af pSv fDv sin2av Sf=2cosav

(5)

where Dv is the valve diameter, av is the valve contact angel, Sv(f) is

the motion of the valve lift which is expressed as:

" 
Sv f i H

f fa
fo

3

f fa
 15H
fo

4

f fa
6H
fo

879

5 #
(6)

where i is the drive ratio of the rocker, H is the cam lift, fo is the cam
rise angle, fa is the advance angle. The motion of the valve fall can
be described in the symmetrical form.
3.1.3. State equation
Ideal air meets the equation of the state:

pV mRg T

(7)

especially the latter are important locations in the temperature

analysis.
However, the temperature of the air going through the ports is
time varying. For the intake port, when the intake valve is open
thus the port can be regarded as a part of the cylinder, the air
temperature at the port changes with the temperature in the cylinder. When the valve is closed, the air temperature at the port can
be set equal to the upstream temperature Tin in the inlet pipe. As to
the exhaust port, the air temperature at the port is closer to the
room temperature when the exhaust valve is closed. And in the
exhaust process, due to the airow from the cylinder, the air temperature is regarded as equal to the temperature in the cylinder.
In addition, the intake and exhaust ports are both very short and
the airow quickly in the operation. Thus, for the convenience of
analysis, it is assumed that the inner walls of the ports have no heat
transfer with the air.
The state of the valves can be judged exactly by monitoring the
mass of the air in the cylinder. Under certain parameter settings,
Fig. 2 shows curves of the cylinder temperature and mass simulated
during a period, and angles (f1, f2, f3 and f4) which mark the
intake and exhaust phases. Thus in a relatively short period of time,
the equivalent air temperature at the intake port can be expressed
by:

where Rg is the gas constant of air [J (kg K)1].

Z f2

3.1.4. Torque equation

In addition, the dynamic model of the APE is based on the differential equations of the crankshaft which relate the crank angle
with the time:

8
df
>
>
>
< dt u
>
Mp Ms  Mn  Mi  Me
du
>
>

:
dt
ICr

Tine

f1

Tdf

Z f1
0

Z
Tin df

360

f2

Tin df

360

(9)

where dm/df > 0, if f (f1, f2).

And the equivalent air temperature at the exhaust port is:

(8)

where u is the angular velocity, t is the time, Mp is the driving

torque generated by the expansion of the compressed air, Ms is an
instant stating torque, Mn is a tunable load torque, Mi (i 1, 2) is the
real time torque caused by the reaction force of the intake or
exhaust valve, Me is the torque due to the friction loss and the
mechanical efciency and ICr is the total moment of inertia of
rotating parts on the crankshaft. These torques analytical expressions are not provided since they have no effect on the heat transfer
system.
3.2. Equivalent air temperatures at intake and exhaust ports
Due to the throttling of the valve and the air expansion in the
cylinder, the air temperatures of the intake and exhaust ports are
predicted to be low. And blocking of the ports seriously affects the
performance of the APE. Hence intake and exhaust ports of the APE

Fig. 2. Curves of the cylinder temperature and mass simulated during a period.

880

Q. Xu et al. / Energy 68 (2014) 877e885

Z f4

Tdf

f3

Texe

Z f3
0

Z
Ta df

360

f4

Ta df
(10)

360

where dm/df > 0, if f (f3, f4).

4. Simplied model of throttling effect for dynamic

temperature analysis
In high pressure pneumatic systems, temperature effects
generated by the throttling cannot be ignored. The following
analysis shows the greater the pressure difference is, the more
signicant the temperature change becomes. The pressurereducing valve is an important component of the APE system,
and it is also a main location of low temperature due to the
throttling. Because the state of the source air is in changing during
APEs operation, the air pressure and temperature before the
reducing valve is not constant. So it is necessary to establish a
dynamic model of the pressure reducer for temperature drop
analysis.
The temperature change of the throttling process is explained by
the JeT effect:
The throttling process proceeds very quickly so that the process
can be considered adiabatic. On this basis the enthalpy of a gas
remains constant before and after the throttling:

hv1 hv2

(11)

where hv1 and hv2 denote the molar enthalpies (J mol1) of the gas
before and after the throttling, respectively.
The molar enthalpy of a real gas can be described in terms of the
gass state parameter as:

"

dhv Cp dTv Vm  Tv

vVm
vTv

 #
dpv

(12)

where Cp is the molar heat capacity at constant pressure

[J (mol K)1], Tv, Vm and pv are the temperature, molar volume
(m3 mol1) and pressure of the gas, respectively.
In the isenthalpic process dhv 0, therefore it can be obtained
from Eq. (12):

"

dTv



1
vVm
 Vm dpv
Tv
Cp
vTv p

(13)

Then the JeT coefcient mj is dened in terms of the partial

derivative of Tv with respect to pv at constant enthalpy:

mj

vTv
vpv

"

"

1
vVm
Tv
Cp
vTv

##


 Vm

(14)

It can be derived from the equation of the state of ideal gas:

vVm
vTv

R
Vm

pv
Tv

!
a
pv 2 Vm  b RTv
Vm

(16)

where a is a measure of the attraction between the particles

(Pa m6 mol2), b is a correction term considering the gas molecular
volume (m3 mol1). For air, values of a and b are 0.1361 and
0.0000367, respectively. Then the temperature change of the
compressed air due to the throttling can be calculated based on Eqs.
(11)e(16). It should be noted that the temperature of a real gas may
either increase or decrease, depending not only on the type of the
gas but also on the initial state before expansion. This phenomenon
can be reected by the sign of mj. For high pressure air at room
temperature mj is always positive during the expansion process
[23]. Since vpv < 0, thus vTv must be negative.
However, numerical methods of solving transcendental equation like Eq. (14) are complex and time-consuming. Researchers
such as Luo and Yuan [21,24] tried to calculate the JeT coefcient
based on the compressibility factor but faced similar problems. And
in dynamic pneumatic systems, calculation considering the throttling effect is even impracticable. In this paper, a simplied calculation method of the throttling effect for the dynamic temperature
analysis is proposed.
On both sides of Eq. (16), solving the partial derivative of Vm
with respect to Tv at constant pressure yields:



3
vVm
RVm

3  aV 2ab
vTv p
pv Vm
m

(17)

Substituting Eq. (17) into Eq. (14) yields:

mj

3 2aV 2  3abV
1 pv bVm
m
m
$
3  aV 2ab
Cp
pv Vm
m

(18)

To avoid integral calculations, equivalent values of pv and Vm are

used to replace their real time values during the throttling process.
The equivalent pressure is the algebraic average of the pressures
before (pv1) and after (pv2) throttling:

pv

pv1 pv2
2

(19)

where pv2 is also equal to the set pressure of the reducer pvs.
The equivalent molar volume is calculated by substituting Eq.
(19) into Eq. (16):

V m f Tv1 ; pv

(20)

where Tv1 is the air temperature before throttling. Using computational tools, the only expression of real value of Vm can be obtained by solving Eq. (16). But the expression is too long to be
described here.
Substituting Eqs. (19) and (20) into Eq. (18) yields the equivalent
average JeT coefcient mj . Then the air temperature after throttling
is described in form of the following expression:

(15)
Tv2 Tv1  mj pv1  pv2
1

where R is the gas constant [J (mol K) ], so that for an ideal gas

mj 0. Its temperature remains constant during the throttling
process. But the ideal gas state equation ignores the molecular
potential energy, which has an inconvenient impact on the
temperature change during the rapid throttling process of a
real gas of high pressure. The real gas Van der Waals state
equation is:

(21)

To compensate the errors caused by the differential calculation

and the equivalent treatments, the coefcient mj is amended as
amj b and a quadratic term g(pv1  pv2)2 is added to the
expression. In addition, a correction term d(pv1  pv2)/pv1pv2 is
added considering the inuence of the initial air pressure on
the temperature drop. Eventually, the air temperature after throttling is:

Q. Xu et al. / Energy 68 (2014) 877e885

Table 1
Values of the empirical coefcients.

b  105/K Pa1

g  1014/K Pa2

d  106/K Pa

3.869

4.9654

1.6

1.3



p pv2
Tv2 Tv1 pv1 pv2 amj b  gpv1 pv2 2 d v1
pv1 pv2

(22)

881



dTs
1
dQs
dms
dms

hs
 us
ms Cv df
df
df
df

(23)

where ms, hs, us, Ts are the mass, the specic enthalpy, the specic
internal energy and the temperature of the air in the tank,
respectively. The heat absorbed by the air in the tank Qs includes
two parts which are Qs1 and Qs2. Qs1 is from the natural convective
heat transfer with the inner wall of the tank:

dQs1 cs As Tws  Ts

(24)

where a, b, g and d are empirical coefcients. They can be obtained

and veried through the nonlinear tting of a large amount of
related data.
The following comparison shows that it can be applied in a wide
range of pressure from 30 MPa to 1 MPa. The data in related articles
[21e26] are used. Here, corresponding values of a, b, g and d are
shown in Table 1.
Table 2 shows typical values of the throttling temperature drop
calculated by SeReK equation and PeR equation in Ref. [21]. The
dynamic calculation method used in this paper is denoted by DTE
(dynamic calculation method of throttling cooling effect) and its
corresponding results are also shown in Table 2.
As it can be seen, most of the results calculated by DTE method
are within the range of the above two methods results. Therefore
the DTE method is reliable in accuracy in the pressure range from
30 MPa to 1 MPa compared to other methods of throttling effect.
And the working pressure of the practical APEs air tank is within
this range. In addition, due to the reduced integral operation the
DTEs calculating speed is superior to other methods, thus it is
suitable for simulation of dynamic models of the APE system.

where ls is the thermal conductivity of air, Hs is the height of the

tank. The Nusselt number is calculated by:

5. Heat transfer models of the air tank and the pipe

where g is the gravitational acceleration, na is the kinematic viscosity of air, Tm is the reference temperature which can be dened
as the average of Tws and Ts. Values of the empirical coefcients can
refer to reference [27]:

The air tank and the pipe are also basic components of the APE
system. Unlike the throttling process, temperature changes in these
models are relatively slow, and the convective heat transfer is
considered in the temperature differential equation. So the air is
considered to be ideal to simplify the calculation. The ideal gas state
equations and continuity equations of the air tank and the pipe are
similar as in Section 3.

The temperature differential equation deduced from the energy

equation is expressed by:

Table 2
Typical values of the throttling temperature drops calculated by different methods
when Tv1 298 K.
Pv2/MPa
22

Pv1 30 MPa
SePeK
PeR
DTE

5.9
6.9
6.3
5.4
6.6
8.9

la
Hs

Nus

(25)

Nus Cs Grs Pr n

(26)

where Pr is the Prandtl number of air, the Grashof number is

dened as:

Grs

gjTws  Ts jHs3
Tm n2a

8
>
< Cs 0:59; n 0:25;
Cs 0:0292; n 0:39;
>
: C 0:11; n 0:333;
s

(27)

Grs 0; 3  109

Grs 3  109 ; 2  1010

Grs 2  1010 ; N

(28)

In addition, the radiation heat transfer of the tanks inner wall

Qs2 is calculated by:

(29)

where is the radiation emissivity, s is the StefaneBoltzmann

constant.

5.2. Model of the pipe

The temperature differential equation of the pipe is:

15

Pv1 12 MPa
SePeK
PeR
DTE
Pv1 20 MPa
SePeK
PeR
DTE

cs



4
dQs2 As s Tws
 Ts4

5.1. Model of the air tank

Method

where As is the heat transfer area, Tws is the temperature of the

tanks inner wall which is assumed to be equal to the
room temperature. The coefcient of natural convective heat
transfer is:

13.7
16.2
19.3

10

3.2
3.8
3.3

13.2
15.4
15.3

17.9
21.1
22.5

23.1
27.3
31.2

14.1
16.4
14.5

25.1
29.3
25.6

30.4
35.6
31.2

36.2
42.6
37.2

22.5
26.4
28.3

34.4
40.5
39.1

40.2
47.4
46.8

46.5
55.1
48.9



dTp
dQp
dmpe
dmpa
dmp
1

hpe
hpa
 up
mp Cv df
df
df
df
df

(30)

Denitions of the parameters in Eq. (30) can refer to those in

Eqs. (1) and (23). The forced convective heat transfer coefcient of
the pipe is calculated by:

cp

la
Dp

Nup

(31)

where Dp is the inner diameter of the pipe, and the Nusselt

number is:

882

Q. Xu et al. / Energy 68 (2014) 877e885

Table 3
Major structural parameters of the verication prototype.
Parameter

Value

Bore  stroke
Cylinders initial volume
Crank ratio
Intake advance angle
Intake duration angle
Exhaust advance angle
Exhaust duration angle

0.052  0.050 m
0.00003 m3
0.263
10
85
10
120

0:5
0:4
Nup 0:023R0:8
Tr Twp
e Pr

(32)

where Twp is the temperature of the pipes inner wall which is

assumed to be equal to the room temperature, Tr is the reference
temperature which can be dened as the average of Twp and Tp, the
Reynolds number is calculated by:

Re

D p up

na

(33)

here up is the ow velocity of the compressed air in the pipe.

The models mentioned above were modularized with all control
variables parameterized, so that the dynamic simulation model of
the APE system can be established by connecting these models.

6. Experimental study on the APE system

To verify the feasibility of the APE, a prototype was modied
from a single cylinder piston-type internal combustion(IC) engine
by transforming its valve system. Major structural parameters of
the prototype are shown in Table 3.
The temperature change during the operation of a high pressure
system is very fast, thus, it is not easy to measure the temperatures
accurately. For ease of measurement and analysis of the data, at
present, a low pressure air system with highest supply pressure of
1 MPa was set up for the verication prototype. Its schematic
structure can refer to Fig. 3. The compressed air discharges from an
air tank (0.6/1.3S-II, Haikong vessel Co. LTD, Qingdao, China), goes
through the pipeline and a pressure reducer (AC40B-04G-TV, SMC
Corporation, Tokyo, Japan), and then supplies to the APE. And a
dedicated test bench for the APE was designed and built to measure
the APEs operating parameters and performances, including cylinder pressure, rotate speed, power and torque of the APE as well as
temperatures of the air at intake and exhaust ports, as shown in
Fig. 4. Associate with this article, the air temperatures were
monitored by installing two PT100 temperature transmitters (with
range of measurement 50  C to 50  C) close to the intake and
exhaust ports. And a pressure transducer (JYB-KO-MAG1, COLLIHIGH, Beijing, China) was installed at the inlet pipe of the APE to

Fig. 4. Phtograph of the verication APE system and its test bench.

measure inlet pressure. In the experiments, each time the air tank
was charged by a compressor (UD11A-10C, United OSD, Shanghai,
China) until the air pressure reached 1 MPa. Then the compressor
was turned off and the switch valve of the tank was opened to
supply compressed air to the APE. Several tests were done at
different inlet pressures. And the systems operating parameters
are shown in Table 4.
Meanwhile, using the models mentioned above, a dynamic
simulation model of this APE system was built to verify its accuracy
on temperature drops analysis. Its principle diagram is shown in
Fig. 3. Parameters of the model are set as in Tables 3 and 4. The
simulation system was running until the air pressure in the tank
was lower than the set pressure of the reducer.
Although it is not related to the temperature analysis, to briey
illustrate the validity of the APEs thermodynamic model, the
average powers simulated when the set speed is 500r/min are
compared with the measured data, as can be seen in Table 5. With
the simulated powers as references, the average deviation is about
4.1% which is an acceptable error.
During running processes of the dynamic system model, air
temperatures in the tank (Ts), at the export of the pipeline (Tp) and
at the export of the reducer (Tv) were monitored. Fig. 5 shows the
simulated temperature curves when the set pressure of the reducer
is 0.7 MPa. It should be noted that the actual values of Tp and Tv are
uctuant. For convenience of observation, they are averaged. As the
air pressure in the tank drops, the air temperature in the tank
continuously decreases in the discharging process, but with small
amplitude. Due to the heat exchange with the pipeline, the air
temperature before the reducer approximately restores to the room

Table 4
Operating parameters of the test bench.
Parameter

Value

Volume of the tank

Initial pressure in the tank
Length and diameter of the pipeline
Inlet pressure
Room temperature
Set rotate speed

600 L
1.0 MPa
4  F0.01 m
0.4e0.9 MPa
309 K
500 r min1

Table 5
Simulated powers and measured powers at different inlet pressures.

Fig. 3. Principle diagram of the dynamic simulation model of the verication APE
system.

Inlet pressure/MPa

0.4

0.5

0.6

0.7

0.8

0.9

Power simulated/W
Power measured/W
Deviation/W

106
97
9

163
160
3

219
205
14

275
265
10

331
323
8

387
380
7

Q. Xu et al. / Energy 68 (2014) 877e885

883

Fig. 7. Schematic diagram of a practical APE system.

Fig. 5. Curves of ps Ts, Tp and Tv during the running process of the dynamic model
when the inlet pressure is 0.7 MPa.

temperature. Although the pressure difference is relatively small,

the throttling cooling effect of the pressure reducer still can be
observed. And as the upstream pressure in the tank drops, the
temperature decrease amplitude reduces.
Then the equivalent air temperatures at intake and exhaust
ports were calculated when the simulation time is around 120 s, as
can be seen in Fig. 6 The intake and exhaust air temperatures of the
experiments were also measured at about 2 min after the start of
each test, and are shown in the gure for comparison.
As can be seen, after running the same period of time, the
measured intake air temperatures have a satisfying agreement with
the calculated values, as the maximum difference between them is
1.1 K. While the exhaust air temperatures are consistently lower
than the measured data. The average temperature deviation is
5.25 K. That is caused by the deviation between the sensor
mounting position and the actual calculating position of the

exhaust port. But the temperature differences standard deviation is

about 1.2 K, it shows the temperature change trend of the simulated data is consistent with the measured ones. Therefore, the
dynamic model has a good reference value in estimating the temperature drops of the real APE system, and the calculation method
of equivalent air temperatures at intake and exhaust ports is useful.
From the results it also can be concluded that, as the inlet
pressure rises, the cooling effect of the throttling process reduces,
thus, the temperature drop amplitude before the APEs air intake
decreases. But the expansion cooling effect in the cylinder increases, the exhaust air temperature gradually gets close to 0  C. In
the practical high pressure APE system, even if the intake air
temperature is close to room temperature, high inlet pressures will
lead to extremely low exhaust temperatures. So heat exchanger of
the intake and exhaust system of the APE is an important direction
to prevent the ice blocking.

7. Simulation for temperature analysis of the practical APE

system
Serving as the engine of vehicles, the practical APE has its inlet
pressure higher than 1 MPa so that it can output enough power.
And sufcient compressed air is needed to meet the demand of a
certain mileage. Thus the practical APE systems air source is a high
pressure system, as shown in Fig. 7. Besides the high pressure tanks
group, in order to stabilize the intake ow, a low pressure buffer
tank and a reducer is added before the inlet of the APE. Design
parameters of the practical APE system including the corresponding APEs major structures are shown in Table 6. A dynamic system
model was built according to the principle in Fig. 7. Air pipes before
the high pressure reducer and after the low pressure reducer are

Table 6
Design parameters of the practical APE system.

Fig. 6. Intake and exhaust air temperatures at different pressures when the running
time is about 2 min.

Parameter

Value

Volume of the tanks group

Initial pressure in the tank
Length and diameter of the pipeline
Volume of the buffer tank
Set pressure of the buffer tank
Inlet pressure
Room temperature
Bore  Stroke
Cylinders initial volume
Crank ratio
Intake advance angle
Intake duration angle
Exhaust advance angle
Exhaust duration angle
Set rotate speed

100 L
30 MPa
1  F0.01 m
70 L
5.0 MPa
2.0 MPa
293 K
0.085  0.088 m
0.00008 m3
0.3165
10
60
30
190
1000 r min1

884

Q. Xu et al. / Energy 68 (2014) 877e885

ignored because they are relatively short, then the pipelines in the
system model are simplied as one pipe model between the high
pressure reducer and the buffer tank.
In the simulation process, the room temperature was set as
293 K. Thus, temperatures of all inner walls in the system model
were ideally assumed to be equal to 293 K. The APE model was
running until the pressure in the buffer tank was lower than 5 MPa.
In the case of rotate speed 1000 r min1, the engines output power
reaches 5 kW. During the operation, air temperatures in the high
pressure tank (Tsh), at the export of the high pressure reducer (Tvh),
at the export of the buffer tank (Tsl) and at the export of the low
pressure reducer (Tvl) were monitored and are shown in Fig. 8.
As can be seen, Tsh drops dramatically, at the end of the operation its lowest value reaches 258 K. Due to the large pressure difference, the instant temperature drop after the reducer is severe,
especially in the initial stage of the operation. Though the temperature recovers slowly after 150 s, the highest value of Tvh is
lower than 257 K. Sustained low temperature makes the reducers
outlet very easy to have ice blocking. Heat exchanges with the
pipeline and the buffer tank is effective so that Tsl is higher than
280 K at most of the time. But after the throttling of the low
pressure reducer, Tvl which is equal to the inlet temperature drops
to lower than 270 K. At about 300 s, the inlet air temperature is
265.8 K. Then according to Eqs. (9) and (10), the calculated equivalent air temperatures at intake and exhaust ports of the APE are
268.8 K and 225.6 K, respectively. It shows the very bad temperature condition at the exhaust port. Furthermore, the pneumatic
power of the air Pa [28] is calculated by the following equation:

Fig. 9. Structure diagram of the heat exchange system for the APE system. (1) Tanks
group; (2) supply pipeline; (3) medium pipeline; (4, 10) heat exchange pipe; (5) pump;
(6, 12) transmissions; (7) low pressure reducer; (8) buffer tank; (9) cylinder head; (11)
energy recovery device; and (13) radiator assembly.

low temperature of the inlet air also affects the APEs power output
performance.
8. Heat exchange system design for the APE system

here min, Tin and pin are the mass, temperature and pressure of the
air in the inlet pipe, respectively. Since the inlet air temperature is
10% lower than the room temperature, in the case of same mass
ow and pressure, the inlet compressed air loses 10% available
energy due to the temperature drop.
Through the analyses above it can be concluded: low temperature locations in the system include the outlet of the high pressure
tanks, outlets of the pressure reducers and the exhaust port of the
APE. In a vehicle mounted APE system, any of these locations having
ice blocking can cause abnormal operation of the engine. And the

A heat transfer system for the practical APE system is designed

with its structure diagram shown in Fig. 9. On the basis of the
original system shown in Fig. 7, a set of heat exchange pipe (4) with
total length 4 m and diameter 0.006 m is added between the tanks
group (1) and the high pressure reducer (11). Size of the pipe after
the reducer (10) is changed into the same size as pipe (4). Positions
of the buffer tank (8) and the low pressure reducer (7) are swapped.
Compared with the position order in Fig. 7, this measure can make
the temperature difference between the inlet air and the inner wall
of the buffer tank (8) bigger, resulting in better heat transfer effect
in the tank. For comparison, other parameters remain unchanged as
shown in Table 6. Then the modied model was simulated. Air
temperature curves in the high pressure tanks (1) (Tsh), at the
export of the pipe (4) (Tph), at the export of the pipe (10) (Tpl) and at
the export of the buffer tank (8) (Tsl) are shown in Fig. 10. As can be
seen, these improvements can make the air temperature in the
supply pipelines (2) more close to the room temperature. When all
the temperatures are higher than 273 K, no ice block will happen

Fig. 8. Curves of Tsh, Tvh, Tsl, and Tvl during the running process of the dynamic model.

Fig. 10. Curves of Tsh, Tph, Tpl, and Tsl during the running process of the dynamic model.

Pa

dmin
p
Rg Tin ln in
dt
pa

(34)

Q. Xu et al. / Energy 68 (2014) 877e885

before the inlet of the APE. When the inlet air temperature is about
282 K, the equivalent temperatures at intake and exhaust ports
raise to 285 K and 230 K, respectively.
For piston-type engine, the intake and exhaust system located
on the cylinder head (9), thus a heat exchange chamber is placed in
the cylinder head (9) so that a heat transfer medium can ow in it
to prevent excessive cooling. The heat transfer medium is stored in
an engine radiator assembly (13) which is composed of a radiator
tank and a fan. The medium of low temperature absorbs heat from
the atmosphere through the radiator and the temperature rises.
Then the liquid is extracted by a pump (5) out of the tank. Through
the medium pipelines (3), it ows through the cylinder head (9),
heat exchange pipes (4), (10) and returns to the radiator tank (13).
In addition, the high pressure reducer is a major cause of energy
loses in the pneumatic system. In the simulation of the high pressure APE system, the average air ow through the reducer is
0.0425 kg/s. Calculated by Eq. (34), the pneumatic power loses due
to the reducing process is approximately 6.3 kW which is even
higher than output power of the APE. Therefore the high pressure
reducer is modied into a pressure reducing and energy recovery
device (11). It is functionally combined by a high pressure rotary
motor and a reducer. When the APE is not running, it plays the role
of a pressure reducer. When there is mass ow in the pipeline, part
of the energy can be recycled by the rotary motor to drive the pump
(5) and the radiator fan (13) through transmissions (6) and (12).
Thus the heat exchange system can be running automatically
during operation of the engine without additional energy
consumption.
9. Conclusions
The thermodynamic model of the APE and a calculation method
for equivalent air temperatures at intake and exhaust ports were
described. A simplied calculation method of throttling effect for
the dynamic temperature analysis was proposed. Models of the air
tank and the pipe were considered. Then the dynamic heat transfer
model of the APE system was established to predict temperature
drops in the operation process, to prevent the ice blocking problem
and improve the performance of the APE system. Conclusions are
summarized as follows:
(1) The simplied calculation method of throttling effect for the
dynamic temperature analysis is proved to be effective in a
pressure range from 30 MPa to 1 MPa.
(2) The dynamic system models feasibility on the temperature
drop analysis is veried by comparing its simulated results
with the measured data in corresponding experiments.
(3) The outlet of the high pressure tanks, outlets of the pressure
reducers and the exhaust port of the APE are low temperature locations during the operation of the system. Heat
exchanging with the buffer tank and pipelines is effective to
recover temperatures. Simulation results show that the heat
exchange system can maintain the air temperatures in the
supply system above 0  C, preventing the ice blocking
problem before the inlet of the APE. In addition, the energy
loss due to the high pressure-reducing process can be used to
drive the heat exchange system.
The analyses of this paper can provide a theoretical basis for the
design of the heat exchange system for the APE system. The

885

proposed simplied calculation method of throttling effect can also

be applied to temperature analysis of other high pressure gas
systems.
Acknowledgements
The research work presented in this paper is nancially supported by grant 51375028, 51205008 of the National Natural Science Foundation of China.
References
[1] Veziroglu TN, Sahin S. 21st centurys energy: hydrogen energy system. Energy
Convers Manag 2008;49:1820e31.
[2] Li Y, Chen H, Zhang X, Tan C, Ding Y. Renewable energy carriers: hydrogen or
liquid air/nitrogen. Appl Therm Eng 2010;30:1985e90.
[3] Ibrahim H, Ilinca A, Perron J. Energy storage systems e characteristics and
comparisons. Renew Sustain Energy Rev 2008;12:1221e50.
[4] Lund H, Salgi G. The role of compressed air energy storage (CAES) in future
sustainable energy systems. Energy Convers Manag 2009;50:1172e9.
[5] Foley A, Daz Lobera I. Impacts of compressed air energy storage plant on an
electricity market with a large renewable energy portfolio. Energy 2013;57:
85e94.
[6] Drury E, Denholm P, Sioshansi R. The value of compressed air energy storage
in energy and reserve markets. Energy 2011;36:4959e73.
[7] Motor Development International (MDI) Home Page. Available online: http://
www.mdi.lu/english/index.php [accessed 08.03.13].
[8] Motor Development International (MDI). Web Page. AIRPod. Available online:
http://www.mdi.lu/english/airpod.php [accessed 27.08.12].
[9] Zuo CJ, Qian YJ, An D, Ouyang MG, Yang FY. Experimental study on airpowered engine. Chin J Mech Eng 2007;43:93e7.
[10] He W, Wu Y, Ma C, Ma G. Research of air powered engine system using twostage single screw expander. Chin J Mech Eng 2010;46:139e43.
[11] Wang Y, Liu XY, Zhou Y. Design and performance analysis of a triangle rotary
piston air-powered engine. Mech Sci Technol Aerosp Eng 2009;28:950e4.
[12] Nie XH, Yu XL, Hu JQ, Chen PL. Inuence of inlet and outlet valves open timing
on pneumatic engine and their optimal design. J Eng Des 2009;16:16e20.
[13] Xu QY, Cai ML, Yu QH, Fu XH. Multi-parameter and multi-objective optimization method for air powered engine. J Beijing Univ Aeronaut Astronaut
2013;39:803e7.
[14] Liu H. Research on air powered vehicle engine. PhD thesis. China: Zhejiang
University; 2004.
[15] Huang CY, Hu CK, Yu CJ, Sun CK. Experimental investigation on the performance of a compressed-air driven piston engine. Energies 2013;6:1731e45.
[16] Zhai X. Research on hybrid of compressed-air and fuel. PhD thesis. China:
Zhejiang University; 2005.
[17] Cherkez R. Theoretical studies on the efciency of air conditioner based on
permeable thermoelectric converter. Appl Therm Eng 2012;38:7e13.
[18] Gong M, Wu J, Cheng Q, Sun Z, Liu J, Hu Q. Development of a 186 C cryogenic preservation chamber based on a dual mixed-gases JouleeThomson
refrigeration cycle. Appl Therm Eng 2012;36:188e92.
[19] Maytal BZ, Pfotenhauer JM. The JouleeThomson effect, its inversion and other
expansions. In: Miniature JouleeThomson cryocooling. New York: Springer;
2013. pp. 39e72.
[20] Luo Y, Wang X. Exergy analysis on throttle reduction efciency based on real
gas equations. Energy 2010;35:181e7.
[21] Luo YX. A study of the real gas effects on high pressure pneumatics and the
basic theoretical and experimental researches of pressure reduction system.
PhD thesis. China: Zhejiang University; 2011.
[22] Luo Y, Wang X, Ge Y. Real gas effects on charging and discharging processes of
high pressure pneumatics. Chin J Mech Eng 2013;26:61e8.
[23] Yin Y, Shen L, Fu J, Dai Y. Outlet temperature characteristics of pneumatic
pressure-reducing valve for hydrogen vehicles. Chin J Constr Mach 2009;7:
383e7.
[24] Yuan WM, He S, Yuan ZM, Liu C, Zhang HL. Numerical methods of using BWRS
equation of state in calculating compressibility factor of natural Gas. Pipeline
Tech Equip 2009;3:004.
[25] Liu X, Mei N, Ding S, Zhao J. Experimental investigation of JouleeThomson
effect for ammonia solution throttle process. J Therm Sci Technol 2009;4:015.
[26] Mao W, Zhang L. A method study for calculating JouleeThompson coefcient.
Spec Oil Gas Reservoir 2002;5:013.
[27] Yang S, Tao W. Heat Transfer. Beijing: Higher Education Press; 1998.
[28] Cai ML, Kawashima K, Kagawa T. Power assessment of owing compressed
air. J Fluids Eng 2006;128:402e5.