Energy 49 (2013) 71e85

Contents lists available at SciVerse ScienceDirect

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

Analysis of vehicle exhaust waste heat recovery potential using a Rankine cycle
Antnio Domingues a, Helder Santos b, Mrio Costa a, *
a
Mechanical Engineering Department, Instituto Superior Tcnico, Technical University of Lisbon, Avenida Rovisco Pais, 1049-001 Lisbon, Portugal
b
School of Technology and Management, Polytechnic Institute of Leiria, Morro do Lena e Alto Vieiro Apt. 4163, 2411-901 Leiria, Portugal

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

Article history: This study evaluates the vehicle exhaust WHR (waste heat recovery) potential using a RC (Rankine cycle).
Received 19 January 2012 To this end, both a RC thermodynamic model and a heat exchanger model have been developed. Both
Received in revised form models use as input, experimental data obtained from a vehicle tested on a chassis dynamometer. The
11 October 2012
thermodynamic analysis was performed for water, R123 and R245fa and revealed the advantage of using
Accepted 1 November 2012
Available online 1 December 2012
water as the working uid in applications of thermal recovery from exhaust gases of vehicles equipped
with a spark-ignition engine. Moreover, the heat exchanger effectiveness for the organic working uids
R123 and R245fa is higher than that for the water and, consequently, they can also be considered
Keywords:
Waste heat recovery
appropriate for use in vehicle WHR applications through RCs when the exhaust gas temperatures are
Rankine cycle relatively low. For an ideal heat exchanger, the simulations revealed increases in the internal combustion
Working uid engine thermal and vehicle mechanical efciencies of 1.4%e3.52% and 10.16%e15.95%, respectively, while
Thermodynamic efciency for a shell and tube heat exchanger, the simulations showed an increase of 0.85%e1.2% in the thermal
Heat exchanger efciency and an increase of 2.64%e6.96% in the mechanical efciency for an evaporating pressure of
2 MPa. The results conrm the advantages of using the thermal energy contained in the vehicle exhaust
gases through RCs. Furthermore, the present analysis demonstrates that improved evaporator designs
and appropriate expander devices allowing for higher evaporating pressures are required to obtain the
maximum WHR potential from vehicle RC systems.
2012 Elsevier Ltd. All rights reserved.

1. Introduction
ICEs (Internal combustion engines) are the major source of
motive power in the world, and this is expected to continue for
some decades. Greenhouse effects and depleted petroleum
supplies are crucial issues that the developed worlds economies
are facing. Because of this, governments in industrialized countries
have introduced strict regulations for ICE emissions and fuel
economy standards. In the last two decades, manufacturers have
improved signicantly ICE efciencies by applying a number of new
technologies [1]. In recognition of the need to further reduce
vehicle exhaust pollutant emissions (CO, NOx, hydrocarbons and
particulate matter) and, more recently, also CO2 emissions, there
has been a lot of interest in the development of cleaner and more
efcient vehicle powertrains [2].
In ICEs only about 1/3 of the fuel combustion energy is con-
verted into useful work to drive the vehicle and its accessory loads.
The remainder is engine waste heat dissipated by the engine
exhaust system, coolant system, and convection as well as radiation
* Corresponding author. Tel.: 351 218417186.
E-mail address: mcosta@ist.utl.pt (M. Costa).
from the engine block [3]. Nearly 40% of the heat energy is wasted
with the engine exhaust gases [4]. If the waste heat of an ICE can be
recovered, the engine efciency will be improved [3]. Furthermore,
global warming will be decreased [5].
To increase the ICEs thermal efciency and to reduce CO2
emissions, different WHR (waste heat recovery) techniques were
recently proposed [1e10]. Among the existing WHR techniques, the
most important are the ETC (electrical turbo-compounding), the
MTC (mechanical turbo-compounding), the TIGERS (turbo-gener-
ator integrated gas energy recovery system), the TEG (thermo-
electric generator) and the RC (Rankine cycle) [6,7].
A number of studies [e.g., [1e3, 5,6]] demonstrate that the RC or
the ORC (organic Rankine cycle) has a high WHR potential in
automotive applications. The RC is based on the steam generation
in a secondary circuit, which represents an indirect method of
WHR. This technique has advantages compared with the so-called
direct WHR techniques (e.g., ETC, MTC and TIGERS) that use
a power turbine tted to the vehicle exhaust, which has a much
higher impact on the engine pumping losses. In addition, a RC
allows for high waste energy utilization and it is cheaper than other
WHR techniques such as thermo-electric generators [5].
The choice of the working uid to be used in the RC depends on
a number of factors, namely, thermodynamic, environmental,

0360-5442/$ e see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.energy.2012.11.001
72 A. Domingues et al. / Energy 49 (2013) 71e85

Nomenclature Subscripts
o initial
A area [m2] 1,2,3,4 Rankine cycle process
A/F air/fuel ratio amb ambient
b long side of a rectangular cross section [m] c thermodynamic cycle
cp heat capacity [kJ kg1 K1] cond condensation
d diameter [m] crit critical
Dh hydraulic diameter [m] e effective
E_ exergy ow rate [kW] evap evaporating
f Darcy friction factor exp expansion
F imposed load [N] ext external
h specic enthalpy [kJ kg1]; heat transfer coefcient f working uid
[W m2 K1] g exhaust gases
I_ exergy destruction rate [kW] h hydraulic
k thermal conductivity coefcient [W m1 K1] i internal
L evaporator tube length [m] in inlet
LHV low heating value [MJ kg1] m material
m _ mass ow rate [kg s1] out outlet
N engine speed [rpm] p pump
Nt tubes number pp pinch-point
Nu Nusselt number pump pumping
p pressure [Pa] s isentropic
P vehicle effective power [kW] t turbine
Pr Prandtl number w wall
Q_ heat rate [kW]
Rd fouling factors [m2 K W1] Superscripts
Re Reynolds number m viscosity ratio exponent
T temperature [K]
U overall heat transfer coefcient [W m2 K1] Abbreviations
v specic volume [m3 kg1] BMEP break mean effective pressure
V vehicle speed [km h1] EGR exhaust gas recirculation
W _ power [kW] ETC electrical turbo-compounding
HCFC hydrochlorouorocarbons
Greek symbols HFC hydrouorocarbons
a* aspect ratio of rectangular ducts, ratio of a small to ICE internal combustion engine
large side length MTC mechanical turbo-compounding
b surface area density [m2 m3] NTU number of transfer units
d distance between tubes [m] ORC organic Rankine cycle
Dp pressure drop [Pa] PPTD pinch-point temperature difference
heat exchanger effectiveness RC Rankine cycle
h efciency TEG thermo-electric generator
m dynamic viscosity [N s m2] TIGERS turbo-generator integrated gas energy recovery
r density [kg m3] system
TWC three way catalyst
WHR waste heat recovery

safety, process-related and economic issues. In particular, when
implementing such a system on a moving vehicle with live occu-
pants, the choice must consider worse case scenarios like leakages
or crashes. On that event the uid must be harmless to the vehicle
occupants. For vehicle applications, the low ammability level is
a major concern. Hence, alcohols and hydrocarbons are arguably
not the best candidates, in spite of their good thermodynamic
efciencies. Instead, refrigerants, already used in air conditioning
systems, are usually good candidates. The refrigerants (e.g., R245fa)
are widely used in ORC applications because of their good heat
transfer properties, excellent thermal stability and low viscosity.
They are generally safe (non-ammable) and compatible with most
materials. Under typical low temperature ambient conditions they
do not freeze, which is a major concern with water. However, the
current generation of refrigerants called HFCs (hydrourocarbons)
has a high global warming potential (100-year time horizon) e
950e1070 for R245fae, which means that their use could be
limited or banned in the near future.
Various working uids have been proposed for RC WHR appli-
cations. Saleh et al. [11] investigated 31 pure component working
uids for low-temperature ORCs. Lai et al. [12] investigated several
pure working uids (alkanes, aromates and linear siloxanes) for
high-temperature ORCs. Yamamoto et al. [13] evaluated the
optimum operating conditions of an ORC comparing HCFC
(hydrochlorouorocarbon)-123 and water as working uids. Ring-
ler et al. [14] examined two kinds of RCs for WHR from gasoline ICEs
using water and ethanol as working uids. Chammas and Clodic
[15] compared the performance of water, R245ca and isopentane as
RC working uids in hybrid vehicles.
Since RCs generate additional power without requiring extra
fuel, both the specic fuel consumption and the pollutant emissions
of the vehicle are reduced [8,9]. The performance analysis of
 A. Domingues et al. / Energy 49 (2013) 71e85 73

various system congurations based on RCs is promising. According
with Boretti [6] improvements in fuel economy up to 17% may be
possible, which can serve as a reference basis for the assessment of
the current potential of this technology, whose major downfalls are
the increase of weight, the packaging complexity, the transient
operation and the costs. Of course, the implementation of a RC on
a vehicle requires detailed investigations of all these issues.
To recover the exhaust waste heat, the RC needs to utilize a heat
exchanger to extract energy from the exhaust gases. A heat
exchanger used in such an application has to be able to provide an
adequate surface area in order to achieve high exchange efciency,
while using a small-size and lightweight arrangement. Furthermore,
the pressure drop has to be minimized to avoid excessive pumping
losses that will have a negative impact on the ICE efciency [16].
Mavidrou et al. [16] examined the exhaust gas heat exchanger design
problem, focusing on the usage of different heat exchanger cong-
urations and different types of heat transfer surfaces.
This article presents an analysis of a combined RC (power cycle)
and heat exchanger for vehicle WHR applications, which, to the
best of our knowledge, is currently lacking in the literature.
Furthermore, the present study provides a systematic analysis for
three working uids (water, R123 and R245fa). Initially, this work
evaluates experimentally the exhaust thermal energy contained in
a vehicle equipped with a spark-ignition ICE. Subsequently, the data
obtained are used as input in two developed models: a RC ther-
modynamic model, which includes both energy and exergy anal-
yses, and a heat exchanger model. Among other features, the RC
thermodynamic model allows assessing the RC efciency and the
net power as function of the evaporating pressure of the three
working uids. The heat transfer model permits to perform heat
exchanger sizing calculations and to assess the different heat
exchanger efciencies and pressure drop. Finally, both the ther-
modynamic RC and the heat exchanger models were used together
to evaluate the vehicle exhaust WHR potential using different RCs.
The evaporator and the expander are the most critical compo-
nents of a RC system. The present study considers available
components (evaporator and expander) that allow building a short
term RC prototype for WHR in vehicle applications. Based on the
measured values in the chassis dynamometer experiments and on
the simulation results, the potential power output of the proposed
RC prototype is assessed and compared for different vehicle oper-
ating conditions.
2. Experimental approach
This section describes the experimental approach followed to
gather the data used as input in the RC thermodynamic and heat
exchanger models. It also quanties the exhaust waste heat as
a function of the vehicle operating conditions.
Chassis dynamometer measurements were carried out on
a vehicle equipped with a 2.8 L VR6 spark-ignition engine in order
to measure the exhaust gases mass ow rate and temperature for
several steady state operating conditions (i.e., after engine warm-
up). For each engine speed (2000, 3000 and 4000 rpm) tests were
made for various loads. Table 1 presents the vehicle test conditions
considered in this work. In the table, N is the imposed engine
speed, F is the imposed load, BMEP is the break mean effective
pressure, V is the vehicle speed, Pe is the vehicle effective power (or
brake power) measured at the chassis dynamometer, m _ g is the
vehicle exhaust mass ow rate, Tg,in is the exhaust gases tempera-
ture measured downstream of the TWC (three way catalyst), and
Q_ available is the heat available in the exhaust gases. The test condi-
tions originated exhaust gases mass ow rates and temperatures
ranging from 12.8 g/s to 59.7 g/s and 730.9 K to 1052.3 K.
All measurements were obtained after both the engine and the
TWC reached their steady states. To validate the steady states the
coolant temperature was controlled to 95  C  10  C. In addition,
controlled temperatures had to remain stable within 5  C for, at
least, 1 min. In the present study, the recorded measurements were
always the average of the readings over a period of time of, at least,
2 min.
In order to evaluate the repeatability of the torque and engine
speed measurements, six tests were performed for each steady
state operating condition. In the torque measurements, the
combined uncertainty ranged from 2.01% to 2.97% and no
relationship between the uncertainty and the torque magnitude
was identied. In the engine speed measurements, it was found
that the combined uncertainty decreases monotonically as the
engine speed increases. Consequently, the maximum uncertainty
occurred at idle (2.24%) and the minimum uncertainty at
4000 rpm (1.41%). The BMEP depends on the torque measure-
ment. As a result, giving that the displacement volume is known,
the experimental uncertainties associated with the BMEP calcula-
tions are those of the torque measurements.
In the present study the exhaust gases mass ow rate was
calculated based on the inlet air mass ow rate and stoichiometric
air fuel ratio. The inlet air mass ow rate was measured using the
engine air mass ow rate sensor. Repeatability tests yielded
maximum uncertainties of 3.2%. The uncertainties associated
with the temperature of the exhaust gases were evaluated to be
within 10 K of the mean value.
The exhaust gases properties have been calculated using the
equations shown in Table 2, which were derived using the software
Refprop 9.0 [17]. The composition (mass fractions) of the exhaust
gases was assumed to be 20.4% CO2, 7.8% H2O and 71.8% N2 (minor
components have been neglected).
In the present study, the heat available in the exhaust gases was
calculated through the following equation:

Table 1
Vehicle test conditions.

Operating condition N [rpm] F [N] BMEP [bar] V [km h1] Pe [kW] _ g [g s1]
m Tg,in [K] Q_ available [kW]
1 2000 0 0 0 0 12.8 730.9 6.23
2 2000 500 0.91 31.7 4.26 17.0 790.0 9.47
3 2000 1000 1.75 30.1 8.18 21.0 829.7 12.71
4 2000 1500 2.35 26.6 10.96 23.9 850.7 15.07
5 2000 2000 2.78 23.5 12.96 25.9 868.2 16.89
6 3000 0 0 0 0 17.3 807.3 10.00
7 3000 500 0.98 50.2 6.88 25.8 897.9 17.77
8 3000 1000 1.95 49.7 13.67 31.5 939.6 23.34
9 3000 1500 2.85 48.2 19.97 37.9 968.7 29.47
10 3000 2000 3.77 47.2 26.39 43.0 989.8 34.59
11 4000 0 0 0 0 25.4 869.4 16.60
12 4000 1000 1.98 67.0 18.45 43.0 1001.8 35.25
13 4000 2000 3.98 67.0 37.17 59.7 1052.3 52.81
74 A. Domingues et al. / Energy 49 (2013) 71e85

Table 2
Equations used to calculate the exhaust gas properties.a

Specic heat capacity [J kg1 K1] cpg 956:0 0:3386$Tg  2:476  105 $Tg2
Dynamic viscosity [N s m2] mg 106  3:807 4:731  102 $Tg  9:945  106 $Tg2
Prandtl number Pr 0:774 1:387  104 $Tg 1:863  107 $Tg2 7:695  1011 $Tg3
Thermal conductivity [W m1 K1] kg 103  (4.643 6.493  102$Tg)
Density [kg m3] rg 1:665  2:404  103 $Tg 1:121  106 $Tg2
a
The equations are valid for 400  Tg  1200 K.


Q_ available m
_ g $cpg Tg;in  Tamb (1) The heat absorbed from the exhaust gases by the working uid
in the heat exchanger is given by:
where m _ g is the exhaust gases mass ow rate, Tg,in is the exhaust
Q_ in m
_ f h3  h2 (3)
gases temperature before the RC heat exchanger (i.e., the exhaust
gases temperature after the TWC) and Tamb is the ambient The turbine power is calculated by:
temperature, taken equal to 25  C in the present study. Table 1
includes the values of Q_ available for all the operating conditions. _ t m
W _ f h3  h4 (4)
It is seen that the available exhaust waste heat changes signi-
cantly from low loads and speeds to high loads and speeds. and the heat rejected from the condenser is given by:
Furthermore, regardless of the operating condition, the available
waste heat at the vehicle exhaust is higher than the vehicle Q_ out m
_ f h4  h1 (5)
effective power.
The RC efciency can be dened as the net power produced
referred to the heat received at the heat exchanger as follows:
3. RC thermodynamic model
_ t W
W _ p
The RC is a vapor power cycle used in numerous applications to
hc _
(6)
Q in
generate electrical power. Fig. 1 shows a schematic of a simple RC. It
is composed by four main components: a pump, a heat exchanger, Exergy is a useful concept for evaluating the performance of
a turbine/generator and a condenser. The ideal thermodynamic various energy systems. Exergy is the maximum amount of work
cycle includes the following processes: an isentropic compression that can be done by a process as it approaches the thermodynamic
process in a pump (1e2), an isobaric heat transfer process in a heat equilibrium with its surroundings by a sequence of reversible
exchanger (2e3), an isentropic expansion process through processes [5,18,19]. The exergy of a subsystem is a measure of its
a turbine (or other expansion machine) (3e4), and an isobaric heat distance from equilibrium and, thus, it can classify the energy
transfer process in a condenser (4e1). quality of the subsystem. Exergy destruction rate labels the loss of
The pump supplies the working uid to the heat exchanger, exergy during the process [5]. The exergy destruction is due to
where the working uid is heated and vaporized, removing heat irreversibilities occurring inside the system or the components of
from the exhaust gases. The working uid leaves the heat the system, a control mass for the system or, as in this work,
exchanger in saturated or superheated state. The high enthalpy a control volume for each component and it can be caused by
vapor is then expanded in the expander (usually a turbine), which internal or external factors [20]. As in the previous studies
is coupled to a generator that deliveries the RC power output. After [e.g., [20e22]], in this work, the contributions of the internal and
the expander, the working uid enters the condenser where it external irreversibilities are not recognized separately, being
condensates. calculated as a whole.
The mathematical model of the simple RC uses the thermody- The exergy destruction rate for each process in the cycle
namic energy conservation equations [18]. The model considers (evaporation, expansion, condensation and pumping) can be
a steady state operation with negligible kinetic and potential expressed as follows:
energy effects. Taken these considerations into account, the pump " !#
power is given by: _Ievap T Tg;out
amb m_ f s3  s2 m
_ g cpg ln (7a)
Tg;in
_ p m
W _ f h2  h1 (2)

I_exp Tamb m
_ f s4  s3 (7b)

 
Qin h  h4
(3) I_conddischarged Tamb m
_ f s1  s4  1 (7c)
Tamb
Heat exchanger

Turbine Wt I_pump Tamb m

_ f s2  s1 (7d)
(2) Note that at the condenser the exergy destruction rate expresses
Wp Pump the sum of the irreversibility of the condenser and the exergy
Condenser discharged with the cooling air that ows across the condenser.
(1) (4)
Apart from the simple RC, there are other RC congurations that
Qout permit to increase the recovered thermal energy. For example, the
thermal efciency of a RC can be augmented by adding a preheater
entry>?tlb>

Fig. 1. Schematic of a simple RC. or a regenerator [7]. Mago et al. [21] presented an analysis of

Table 3
Main thermophysical properties of the working uids studied (water, R123 and
R245fa).
Working Category pcr Tcr [K] Normal boiling
uid [MPa] temperature
[ C]
Water e 22.06 647.1 100
R123 HCFC 3.66 456.68 27.8
R245fa HFC 3.64 427.15 14.9
regenerative ORCs using dry uids, where they demonstrated that
regenerative ORCs not only have higher rst and second law ef-
ciencies than basic ORCs, but they also have lower irreversibilities
and lower heat requirements to produce the same power. Recently,
Wang et al. [22] studied the characteristics of a dual loop ORC
system which recovers the waste heat from the exhaust and the
coolant of a spark-ignition ICE. The HT (high temperature) loop
recovers the exhaust waste heat using R245fa as the working uid
and the LT (low temperature) loop recovers the coolant waste heat
and the residual heat from the HT loop using R134a as the working
uid. The results revealed that with this conguration the net
power of the LT loop is higher than that of the HT loop.
However, RCs with regeneration, preheating processes or dual
loop systems require more piping (increased pressure losses) and
more complex devices [7]. Furthermore, compactness is an essen-
tial issue in RCs for vehicle applications. Therefore, the present
study analyzes only the simple RC for different working uids.
The working uids can be classied into three categories
according to the slope of the saturation vapor curve in a Tes
diagram. Dry uids have positive slopes, wet uids have negative
slopes, and isentropic uids have innitely large slopes. A disad-
vantage of using water (or other wet uid) as the working uid is
the need to superheat the steam to prevent condensation during
the expansion, a problem that results directly from the water
thermophysical properties. The occurrence of condensation is
a problem because it can lead to erosion of the turbine blades [7].
The choice of appropriate working uids can circumvent the
superheating requirement, especially in applications that require
operation at low temperatures, or that have a heat source with
relatively low temperature. In these situations, an organic working
uid offers advantages since, in many cases, superheating is
unnecessary. These cycles are called ORCs. In the case of HCFCs
and HFCs the saturation vapor line is almost vertical. This implies
that, without initial superheat, end points lie in the dry vapor
region [23].
Exhaust gases
2 3
Heat exchanger
Pump
1
Condenser
entry>?tlb>
Fig. 2. Schematic of a typical RC waste heat recovery system from ICE exhaust gases.
A. Domingues et al. / Energy 49 (2013) 71e85 75
Fluid selection is one of the most important contributors to the
overall RC performance. There are several general criteria that the
working uids should ideally satisfy. Stability, non-fouling, non-
pcond, Slope of the corrosiveness, non-toxicity and non-ammability are a few prefer-
saturation
T 323 K able physical and chemical characteristics [18]. Several organic uids
[bar] vapor line
for use in RCs have been proposed in the literature [11e15,20e
0.123 Negative
22,24e27]. Among them, the organic uids R123 and R245fa
2.125 Positive
4.012 Positive appear to be the most promising ones for the operating conditions
used in this work, mainly due to its non-ammable behavior and
thermodynamic performance. Therefore, the organic uids R123 and
R245fa have been selected for the present study. In addition, water
has been also considered as working uid in this work because it is
non-toxic and due to its abundant availability. Table 3 presents the
main thermophysical properties of the working uids studied in this
work (water, R123 and R245fa).
Fig. 2 shows a schematic of a typical RC waste heat recovery
system from ICE exhaust gases. Considering the operational
conditions of a simple RC assembled on the ICE exhaust system of
a vehicle, the present model constraints are as follows:
(i) evaporating pressure varying between condensation pressure,
pcond, and critical pressure, pcrit;
(ii) dry expansion for all uids;
(iii) isentropic expander efciency, ht 0.7 [7];
(iv) isentropic pump efciency, hp 0.8 [7];
(v) negligible pressure losses in the heat exchangers and pipes;
(vi) for both organic uids (R123 and R245fa) the condensation
temperature is Tcond 323 K, which corresponds to the
condensation pressures given in Table 3; for water, the
condensation temperature is Tcond 373 K, which corresponds
to a condensation pressure of 1 bar [28];
(vii) the superheating temperature is set as the minimum to
guarantee a dry expansion.
A sensitivity assessment was performed for the values of the
expander and pump isentropic efciencies, water evaporating and
condensation pressures and organic uids evaporating and
condensation temperatures. Table 4 summarizes the results ob-
tained. It is seen that the expander efciency affects signicantly
the RC net power output. Moreover, it is interesting to note that the
RC net power output is more sensitive to the condensation condi-
tions of the working uids than to the evaporating conditions.
Fig. 3 shows Tes process diagrams for water, R123 and R245fa.
The solid line in each gure represents the engine exhaust gases
temperature for the vehicle operating condition 9 (see Table 1).
Note that Tg,out depends on the RC heat exchanger characteristics. In
Air
TWC
Generator
Turbine
4
Air + fuel
76 A. Domingues et al. / Energy 49 (2013) 71e85

Table 4 the present study, we used a Tg,out 200  C in all RC thermody-

Sensitivity assessment of parameters used in the RC thermodynamic model. namics calculations performed. It is seen that only in the case of
Parameter Variation considered Resulting variation water the temperature difference between the exhaust gases and
in the parameter in the RC net power the working uid at the beginning of the evaporation process (the
output (%) so-called pinch-point temperature difference e PPTD) may be
Expander efciency, ht 10% 16 problematic.
10% 0.1
Fig. 4 shows a schematic of the T Q_ diagram used for the pinch-
Pump efciency, hp
Water evaporating pressure, 0.1 bar 1
pevap point analysis in the RC heat exchanger. The PPTD is the smallest
Water condensation pressure, 0.1 bar H4 temperature difference in the RC heat exchanger, establishing the
pcond maximum allowable evaporation pressure and, thus, limiting the

Organic uids evaporating 10 C 5 RC efciency. It is necessary to guarantee a minimum PPTD of 30  C
temperature, Tevap
Organic uids condensation 10  C H9
[8], which is particularly challenging at low loads and low engine
temperature, Tcond speeds.
In the present study, a RC thermodynamic model, which includes
both energy and exergy analyses, was developed to evaluate the
vehicle exhaust WHR potential using a RC. The required thermo-
dynamic and transport properties for the water and the organic
working uids were calculated with the aid of the Refprop 9.0 [17].
a 1000 Fig. 5 shows the algorithm for the RC thermodynamic model, which
Ideal was implemented in Matlab. The thermodynamic properties of the
900 Water Real
Exhaust temperature
working uids studied were calculated for each evaporating pres-
800 sure from the condensing pressure to the critical pressure of the
organic working uids (3.64 MPa for R245fa and 3.66 MPa for R123).
700 In the case of the water the calculation was performed up to 4.4 MPa,
which corresponds to a working uid temperature close to that of
T (K)

600
the exhaust gases at the heat exchanger outlet.
500
4. Heat exchanger model
400

300 The heat exchanger (evaporator) is an essential component in

vehicle WHR applications. The following characteristics are desir-
200 able in an heat exchanger for such applications:
0 1000 2000 3000 4000 5000 6000 7000 8000 9000
b 800
i) high heat exchanger effectiveness;
R123 ii) low pressure drop trough the heat exchanger, which mini-
700 mizes the negative impact of the exhaust back pressure on
the ICE;
600 iii) compactness.
T (K)

500 Improvements in the heat exchanger effectiveness can be ob-

tained by increasing the heat transfer area or the heat transfer
400

Tg,in
T (K)

300

200
600 800 1000 1200 1400 1600 1800
c 800
as
R245fa tg
aus
h
700 Ex
T3
Tg,pp
600
Tpp Working fluid
Tg,out
T (K)

500 T2x T3x

400

300
T2

200
600 800 1000 1200 1400 1600 1800 2000 Q (kW)
s (J/kgK)
Fig. 4. Schematic of the T Q_ diagram used for the pinch-point analysis in the RC heat
entry>?tlb
>

entry>?tlb>

Fig. 3. Tes process diagrams. a) Water, b) R123 and c) R245fa. exchanger.

 A. Domingues et al. / Energy 49 (2013) 71e85 77

Working fluid,
Input variables
Tcond, pcrit

Thermodynamic
conditions

Database Calculate For each

interface efficiency and evaporating
Refprop 9.0 thermophysics pressure
properties

Flow balance in
evaporator

For each
experimental
condiction
Calculate power
output and
working fluid flow

Show results

entry>?tlb>

Fig. 5. Algorithm for the RC thermodynamic model.

coefcient from the exhaust gases side. In a tubular heat exchanger, little affected by the fouling resistances [16]. The fouling resistance
the heat transfer area is usually increased by increasing the number on the gas side is higher than that on the uid side. The particulate
of tubes and/or by using ns inside the tubes. It is well known that matter present in the exhaust gases is the principle responsible for
the exhaust gases heat transfer coefcient is much smaller than the the fouling resistance on the gas side. Note that the present study
RC working uid heat transfer coefcient. As a result, nned examines an evaporator placed after a TWC (see Fig. 2) where the
surfaces are usually placed on the exhaust gases side to increase the particulate matter is rather low. Moreover, we found that consid-
heat transfer area [16]. ering values of the fouling resistances ten times higher than those
A literature survey [e.g., [16, 29]] revealed that shell and tube above, the RC net power output decreases by 6.7%. If the values of
heat exchangers are the most appropriate to be used as evaporator the fouling resistances are not considered the RC net power output
in RCs for vehicle exhaust applications. The heat exchanger model increases by 1.1%.
developed in the present study allows accessing the thermal and The heat transfer coefcient h is calculated as follows:
hydraulic characteristics of various duct geometries (square, rect-
angular and circular) as function of the number of tubes of the heat Nud
exchanger. h kg (9)
Dh
The present heat exchanger model is based on the -NTU
(effectiveness-number of transfer units) method. The overall heat For Red < 2100, the Nusselt number is calculated through the
transfer coefcient U is calculated according to the following Sieder and Tate correlation [30]:
equation [30]:
  1=3
D
1 Nud 1:86 Red $Pr$ h m=mw m ; Red < 2100 (10)
U (8) L
dext dext Rd;i dext 1
lndext =di Rd;ext
di hg di 2km hf For Red > 2100, the Nusselt number is evaluated with the aid of
the Gnielinski correlation [30]:
In the present study the tube was assumed to be made of
aluminum, with km 225 W m1 K1 [30]. For this type of appli-   2=3 
cations, it is extremely difcult to nd accurate data for the fouling f =8Red 1000Pr Dh
Nud 1=2
 1 m=mw m ;
resistances (Rd) appearing in eq. (8), but typical values have been 112:7f =8 Pr 1
2=3 L (11)
assumed to partially account for the fouling effects; specically, Red >2100
Rd,ext 8.8  105 m2 K W1 for all areas in contact with the
working uid and Rd, i 1.8  104 m2 K W1 for all areas in contact In eqs. (10) and (11), m is the uid viscosity at the bulk uid
with the exhaust gases [31]. The overall heat transfer coefcient U is temperature, mw is the uid viscosity at the heat transfer boundary
78 A. Domingues et al. / Energy 49 (2013) 71e85
surface temperature, m 0.14 for Re < 8000 and m 0.25 for
Re > 8000. In eq. (11), f is a logarithmic function of the Reynolds
number:
f 0:79$lnRed  1:642 (12)
The heat exchanger effectiveness, , is calculated from:
h A
m_ ggcpg
1  eNTU 1  e (13)
considering only the exhaust gases ow. Note that this relation is
valid only for condensers and evaporators.
The surface area density, b, for square, rectangular and circular
cross ow geometries is calculated through the following equations
[32]:
4b
b 2
(14a)
b d


2 a* 1 b
b
 (14b)
ba* d b d
2pb
b p 2
(14c)
3b d
The pressure drop through shell and tube heat exchangers
consisting of square, rectangular and circular cross ow geometries
is calculated as follows [32]:
_ 0 2
4fLm=A
Dp Nt (15)
2rg Dh
The heat exchanger considered in the present study is a shell
and tube counter ow type, with the hot uid (exhaust gases) in
tubes and the cold uid (RC working uid) in the shell. In order to
study the effect of the important thermal and hydraulic charac-
teristics of the heat exchanger as a function of the number of the
tubes for different cross ow geometries, rst we had to dene the
main dimensions of the heat exchanger. Hussain and Brigham [29]
demonstrated that the heat exchanger effectiveness decreases for
larger shell diameters, which depend on the cross ow area of the
heat exchanger. On the other hand, it is well known that the
exhaust gases pressure drop decreases for larger cross ow areas of
the heat exchanger tubes. Considering this trade-off and the limited
space available for the installation of the heat exchanger, the cross
ow area of the heat exchanger tubes in the present study was set
equal to the vehicle exhaust duct cross ow area. The vehicle
exhaust duct under study has a cross ow area equal to
A0 2.561  103 m2. To calculate the cross ow area of each tube,
A0 was divided by the number of tubes in the evaporator. In regard
to the heat exchanger length, it was assumed tube lengths of 0.5 m
owing to the dimensions of the vehicle under study. In this case we
have also performed a sensitivity analysis by considering variations
in the heat exchanger length of 0.1 m. The calculations indicated
variations in the RC net power output of 15%, revealing the
signicant impact of this parameter in the present study.
Fig. 6 presents the thermal and hydraulic characteristics (Re, Nu,
h, , b and Dp) of the exhaust gases as a function of the number of
tubes in the evaporator for different cross ow geometries. This
gure was constructed based on the exhaust gases mass ow rate
and temperature of operating condition 9. It is seen that both the Re
and Nu decrease when the number of tubes increases, see Fig. 6a
and b. The tubes hydraulic diameter also decreases as the number
of tubes increases and, as a result, the heat transfer coefcient
increases, as seen in Fig. 6c. Fig. 6d reveals that the evaporator
effectiveness increases as the number of tubes increases. This is
because both the heat transfer coefcient and the surface area
density increase as the number of tubes increases, see Fig. 6c and e.
Fig. 6 also reveals that the pressure drop increases as the
number of tubes increases and that both the heat exchanger
effectiveness and the pressure drop are higher for the rectangular
section, followed by the square section and by the circular cross
section. It is known that low pressure drops are associated with low
heat transfer coefcients. In the present application, the heat
exchanger is to be installed in the exhaust duct of the vehicle. In
such an application, the minimization of the exhaust gases back
pressure on the ICE is of critical importance. The heat exchanger
with circular tubes provides the lowest exhaust gases back pressure
(see Fig. 6f). Furthermore, this is the simplest geometry for
construction, which allows minimizing the heat exchanger costs. As
a result, in spite of the heat exchanger effectiveness penalty, the
conguration with circular tubes was selected for this study.
As already pointed out, Fig. 6f indicates that the exhaust gases
pressure drop increases markedly with the number of tubes, while
Fig. 6d shows that the increase in the heat exchanger effectiveness
for a large number of tubes (40e100) is comparatively small. These
results establish the maximum for the number of tubes in the heat
exchanger: 43 tubes in the present study. Considering the cross
ow area and the hexagonal shell and tube heat exchanger
arrangement, 43 tubes correspond to a tube diameter of 10 mm,
with a distance between the tubes of 4 mm, and a shell inside
radius of 94 mm.
The heat exchanger was divided into three zones for modeling
purposes, as shown in Fig. 7, namely, a preheating zone, an evap-
orating zone and a superheating zone. These zones were consid-
ered as individual heat exchangers with the appropriate boundary
conditions for temperature and mass ow rate. The amounts of
exchanged heat for preheating, evaporating and superheating of
the RC working uid were estimated from the heat transfer rela-
tions and the -NTU method [29,31]. The -NTU method was
applied to the three zones of the heat exchanger. The enthalpy ratio
was used as a rst approximation for each zone.
Fig. 8 shows the algorithm for the heat exchanger model, which
was also implemented in Matlab. As input data, this model uses the
results of the RC thermodynamic model, discussed earlier. The
required thermodynamic and transport properties for the exhaust
gases were calculated with the aid of the Refprop 9.0 [17]. A
parametric iterative method was employed for selecting the
optimum working uid and to obtain the working uid mass ow
rate for the investigated heat exchangers. The calculations were
performed for different evaporating pressures and working uids
until reaching convergence between the working uid mass ows
in the various sub-components of the heat exchanger.
5. Results and discussion
5.1. RC thermodynamic analysis
This section presents a combined rst and second law analysis.
The main objective of this section is to assess the RC efciency, h,
the turbine outlet/inlet expansion ratio, v4/v3, the working uid
mass ow rate, m _ net , as function of the
_ f , and the RC net power, W
evaporating pressure for the working uids studied. To this end, an
adiabatic heat exchanger was considered so that the maximum
vehicle exhaust WHR potential, assuming a Tg,out 200  C (see
section 3) was assessed. The thermodynamic analysis reported
below has been performed for the exhaust conditions corre-
sponding to the vehicle operating conditions 3, 9 and 13 (see
Table 1), which represent low, intermediate and high engine speed
 A. Domingues et al. / Energy 49 (2013) 71e85 79

30000 70
a b Square
25000 60 Rectangular
Circular
50
20000
40

Nu
Re
15000
30
10000
20
5000 10

0 0
10 20 30 40 50 60 70 80 90 100 10 20 30 40 50 60 70 80 90 100
120 100
c 90 d
110
80
h (W m-2 K-1)

100 70

(%)
60
90
50
80 40
30
70
20
60 10
10 20 30 40 50 60 70 80 90 100 10 20 30 40 50 60 70 80 90 100
300 1.5
e f
250
(x 105 Pa)

1
(m2 m-3)

200

150
0.5

100

50 0
10 20 30 40 50 60 70 80 90 100 10 20 30 40 50 60 70 80 90 100
Nt Nt
entry>?tlb>

Fig. 6. Thermal and hydraulic characteristics of the exhaust gases as a function of the number of tubes in the evaporator for different cross ow geometries. a) Re, b) Nu, c) h, d) , e)
b, f) Dp.

and load steady state vehicle operating conditions, respectively.
This allows the assessment of the performance of the RC for
different exhaust mass ow rates and temperatures.
Fig. 9 shows the RC efciency as a function of the evaporating
pressure for the working uids studied. The gure reveals that the
RC efciency (rst law efciency) is very sensitive to the evapo-
rating pressure (or evaporating temperature) and working uid. In
regard to the working uid, the RC efciency at the evaporating
pressure of, for example, 2 MPa is 14.29% for water, 12.03% for R123
and 9.53% for R245fa.
Fig. 10 shows the turbine outlet/inlet expansion ratio (v4/v3) as
a function of the evaporating pressure for the working uids
Tg,out Tg,pp
Preheating Evaporating
zone zone
T2 T2x
entry>?tlb>
studied. The working uid R245fa presents the lower turbine
outlet/inlet expansion ratio (v4/v3), regardless of the evaporating
pressure. This is mainly due to the higher R245fa condenser pres-
sure (4.012 bar) as compared to that of the water (1 bar). The
turbine outlet/inlet expansion ratio (v4/v3) is an important
parameter as it indicates how much the uid volume increases
through the expansion process. It should be noted that the
expansion ratio (v4/v3) can change signicantly according with the
characteristics of the working uid. The expansion ratio is also very
important for the expander selection. When the expansion ratio
(v4/v3) is smaller than 50, expansion efciencies higher than 0.8 can
be achieved using a single stage axial turbine as expander [7].
Figs. 11 and 12 show the working uid mass ow rate and the RC
net power output as a function of the evaporating pressure for the
Tg,pp2 Tg,in
working uids studied for operating conditions 3, 9 and 13. For
a given operating condition, Fig. 11 reveals that the working uid
Superheating
zone R123 requires the highest mass ow rate, being the mass ow rate
required by the working uid R245fa slightly lower. It can also be
T3x T3 seen in Fig. 11 that, for a given operating condition, the water mass
ow rate is one order of magnitude lower than the organic uids
mass ow rate as a consequence of the highest enthalpy increase in

Fig. 7. The three zones of the heat exchanger considered for modeling purposes.
80 A. Domingues et al. / Energy 49 (2013) 71e85

Results of the Define

thermodynamic geometrical
model characteristics

Initial conditions:
temperature and
lenght
for each sub-component

Interface with Exhaust gas

Refprop 9.0 characteristics
Repeat
calculations
for each
sub-component

-NUT method

Calculate
flow balance
for each
sub-component Run
until
convergence
Calculate
exhaust gases
temperature and
sub-component
lenght

Calculate
heat exchanger
power output
and e fficiency
entry>?tlb>

Fig. 8. Algorithm for the heat exchanger model.

the evaporation process. This indicates that the pump power Furthermore, the selected pump needs to be equipped with
required in the ORC is much higher than the pump power required in a frequency converter to control its rotation speed and, hence, the
the water RC. Thus the pump should be carefully selected to mini- evaporating pressure and the mass ow rate of the RC. For a given
mize the power costs under high evaporating pressures. operating condition, Fig. 12 reveals that the RC net power output is

30
0.2
Water
0.18 R123
25
R245fa
0.16
0.14 20
0.12
v4 / v3

0.1 15

0.08
10
0.06
0.04 Water
R123 5
0.02 R245fa

0 0
0 1 2 3 4 5 0 1 2 3 4 5
Evaporating pressure (MPa) Evaporating pressure (MPa)
entry>?tlb> entry>?tlb
>

Fig. 9. RC efciency as a function of the evaporating pressure for the working uids Fig. 10. Turbine outlet/inlet expansion ratio (v4/v3) as a function of the evaporating
studied. pressure for the working uids studied.
 A. Domingues et al. / Energy 49 (2013) 71e85 81

0.25 main cause for the lower RC efciencies for the organic working
uids as compared to water.
As pointed out earlier, Mago et al. [21] presented an analysis of
0.2 the performance of different ORC congurations using dry organic
Working fluid flow (kg/s)

Op. condition 13
0.15 Op. condition 9
Op. condition 3
0.1
0.05
0
0 1 2 3
Evaporating pressure (MPa)
entry>?tlb>
Fig. 11. Working uid mass ow rate as a function of the evaporating pressure for the
working uids studied.
higher for water and lower for R245fa, regardless of the evaporating
pressure.
Fig. 11 shows that the working uid mass ow rate increases for
the engine operating condition 13 as compared to the operating
conditions 9 and 3 due to the increase in the waste heat ow. Fig. 12
reveals that when the waste heat ow increases (from the oper-
ating condition 3 to the operating condition 13), the RC net power
output increases signicantly.
It is important to point out that the condensation pressures
(and temperatures) for the organic uids and water are different. As
mentioned in Section 3, we have chosen a condensation tempera-
ture of 323 K for the organic uids (R123 and R245fa) and 373 K for
water. Despite the higher water condensation temperature, Fig. 12
reveals that water yields higher RC net power output than the
organic uids. The higher water condensation temperature also
allows the use of smaller size condensers as compared to the
organic uids. Note that when the temperature difference between
the condenser and the ambient conditions decreases the condenser
size has to be increased. This is a critical issue when using organic
uids in vehicle exhaust WHR applications that needs to be care-
fully considered in future ORC studies.
The higher temperature difference between the exhaust gases
and the working uid in the heat exchanger for R123 and R245fa
(see Fig. 3) increases the evaporator irreversibilities that are the
uids (R113, R123 and R245ca). These authors showed that super-
Water heating organic uids increase the irreversibilities and decrease the
R123 second law efciency. Note that in the present study the evapo-
R245fa
rating pressure of the organic uids varied between the conden-
sation pressure and the critical pressure (see Table 3). In order to
reduce the evaporator irreversibilities at the evaporator output the
organic uids operated at saturated conditions (see Fig. 3); closer to
the critical pressure the superheating temperature was set as the
minimum to guarantee a dry expansion.
Fig. 13 shows the RC evaporator exergy destruction rate as
4 5 a function of the evaporating pressure for the working uids
studied for operating condition 9. High evaporating pressures
reduce the temperature difference in the evaporator, so the exergy
destruction rate decreases. Increasing the evaporating pressure is
a good method to avoid second law losses. High evaporating
pressures also increase the RC rst law efciency, as shown in Fig. 9.
As a result, increasing the evaporating pressure is a better method
to increase the RC performance than increasing the superheating
temperature.
Previous studies [e.g., [2]] demonstrate that binary mixture
working uids can diminish signicantly the evaporator entropy
generation (exergy destruction) because these uids can reduce the
temperature difference in the evaporator. Research on binary
mixture working uids is still very limited and more work is
necessary to obtain a better understanding of their inuence on the
performance of RCs.
The present results (not shown here) revealed also that the
entropy generation rate in the condenser is signicantly lower than
that in the evaporator mainly because of the smaller temperature
differences that occur in the condenser. The entropy generation rate
in the expander and pump is related to the isentropic efciency of
these devices. Our results indicated that the expander entropy
generation rate is much higher than that in the pump, but
considerably lower than those in the evaporator and condenser.
Consistently with previous studies [e.g., [20]], the present analysis
demonstrates that the evaporator makes the biggest contribution
to the overall entropy generation rate in the RC system.
To complement the present second law analysis, the evaporator
exergy efciency was also determined. To this end, the dead state
was specied as Tamb 25  C and pamb 1 atm.

8 11
Op. condition 13 Water
Op. condition 9 R123 Water, op. condition 9
7 10 R123, op. condition 9
Op. condition 3 R245fa
Exergy destruction rate (kW)

R245fa, op. condition 9

6 9
Net power output (kW)

5 8

4 7

3 6

2 5

1 4

0 3
0 1 2 3 4 5 0 1 2 3 4 5
Evaporating pressure (MPa) Evaporating pressure (MPa)
entry>?tlb> entry>?tlb>

Fig. 12. RC net power output as a function of the evaporating pressure for the working Fig. 13. RC evaporator exergy destruction rate as a function of the evaporating pressure
uids studied. for the working uids studied.
82 A. Domingues et al. / Energy 49 (2013) 71e85

The exergy ow rate of the exhaust gases (assumed to be an Table 5

ideal gas) entering and leaving the RC can be calculated as: Main characteristics of evaporators 1 and 2.

Characteristic Evaporator 1 Evaporator 2

     

 Tg;j pj
E_ j m
Number of tubes, Nt 43 38
_ g cpg Tg;j  Tamb  cpg ln  Rg ln (16)
Tamb pamb Tubes diameter, di (m) 0.01 0.006
Distance between tubes, d (m)a 0.004 0.002
The exergy ow rate of the working uid at any point of the RC Tubes length, L (m) 0.5 0.7
can be determined as: Thickness, e (mm) 0.5
External diameter, dext (m) 0.11

E_ i m
a
_ f hi  hamb  Tamb si  samb  Tubes in an equidistant hexagonal arrangement.
(17)

The evaporator exergy efciency can be expressed as:

5.2. Heat exchanger analysis

E_ E_  E_ This section extends the RC thermodynamic analysis presented

hexergy;evap _ useful _ 3 _ 2
Eavailable Eg;in  Eg;out
where E_ useful and E_ available are the actual exergy used and the
theoretically available exergy at the evaporator, E_ 2 and E_ 3 are the
exergy ow rates of the working uid entering and leaving
the evaporator, respectively, and E_ g;in and E_ g;out are the exergy ow
rates of the exhaust gases entering and leaving the evaporator,
respectively.
Fig. 14 depicts the variation of the RC evaporator exergy ef-
ciency as a function of the evaporating pressure for the working
uids studied for operating condition 9. The results presented in
this gure agree well with the results presented in Fig. 13, since
a decrease in the evaporator irreversibilities represents an incre-
ment in the evaporator exergy efciency. Fig. 14 also reveals that
the evaporator exergy efciency is rather modest regardless of the
working uid. Nevertheless it is important to observe that the
evaporator exergy efciency is about two times higher for water as
compared to the organic uids. The results from the combined rst
and second law analysis show that water yields higher thermody-
namic efciency as compared to the organic uids.
The thermodynamic analysis demonstrates that the water is an
adequate working uid for use in an exhaust heat recovery system
with a RC for the following reasons: as compared to the organic
uids, water (i) yields higher thermodynamic efciency and net
power output; (ii) requires lower quantities of working uid in the
RC (less weight); (iii) condenses easily at atmospheric pressure; (iv)
has lower price and higher abundance; and (v) presents no envi-
ronmental risks. It should be stressed, however, that the use of
water as the working uid in cold regions may be problematic. In
case of frost, water expands at freezing and this can destroy the
equipment in a single cold night.
(18)
in the previous section, by considering also the heat exchanger
model. It examines two shell and tube counter ow type heat
exchangers with the exhaust gases passing through the tubes,
referred hereafter as evaporator 1 and evaporator 2. Evaporator 1
maintains the cross ow area of the vehicle exhaust duct under
study with tube diameters of 1 cm. Specic heat exchangers for
WHR applications are currently not available, so a shell and tube
counter ow type EGR (exhaust gas recirculation) cooler from
a MAN diesel truck was selected as evaporator 2. A similar approach
was considered in previous studies [e.g., [27]]. Table 5 summarizes
the main characteristics of evaporators 1 and 2. The main objective
of this section is to assess the inuence of the heat exchanger on the
RC net power output, W _ net , as a function of the evaporating pres-
sure for the working uids studied.
Fig. 15 shows the working uid mass ow rate as a function of
the evaporating pressure for the working uids studied using
evaporator 1 for the operating conditions 3, 9 and 13. It is seen that
the working uid mass ow rate decreases (approximately 50%)
compared to the working uid mass ow rate obtained for the
optimal heat transfer conditions, as presented in Fig. 11, regardless
of the vehicle operating condition. This is mainly due to the high
thermal resistance of evaporator 1. Fig. 15 reveals an increase in the
organic uids mass ow rate for evaporating pressures higher than
2.25 MPa for the operating condition 13, which is due to the ow
regime transition from laminar to turbulent in the heat exchanger.
It should be stressed that this occurs only for operating condition
13, which represents a high engine speed and load steady state
vehicle operating condition.
Fig. 16 shows the heat exchanger effectiveness as a function of
the evaporating pressure for the working uids studied using
evaporator 1 for the operating conditions 3, 9 and 13. This gure

0.45 0.14
Op. condition 13 Water
0.4 0.12 Op. condition 9 R123
Evaporator exergy efficiency ()

Op. condition 3 R245fa

0.35
Working fluid flow (kg/s)

0.1
0.3
0.08
0.25
0.06
0.2
0.04
0.15
Water, op. condition 9
R123, op. condition 9 0.02
0.1
R245fa, op. condition 9
0.05 0
0 1 2 3 4 5 0.5 1 1.5 2 2.5 3 3.5
Evaporating pressure (MPa) Evaporating pressure (MPa)
entry>?tlb> entry>?tlb
>

Fig. 14. RC evaporator exergy efciency as a function of the evaporating pressure for Fig. 15. Working uid mass ow rate as a function of the evaporating pressure for the
the working uids studied. working uids studied using evaporator 1.
 A. Domingues et al. / Energy 49 (2013) 71e85 83

100 effectiveness for water. A comparison of these results with those in

Op. condition 13 Water Fig. 12 reveals that the tubular heat exchanger causes a signicantly
90
Op. condition 9 R123 decrease in the net power extracted from the RC using water as the
Op. condition 3 R245fa
80 working uid. These results demonstrate that the low net power
70 produced by the RC when using water as the working uid is
mainly due to the low heat exchanger effectiveness (less than 40%).
60
As the vehicle load and speed increase (operating condition 13) the
(%)

50 heat exchanger effectiveness tends to be similar for the three

40
30
20
10
0 rates, as seen in Fig. 15.
0.5 1 1.5 2 2.5
Evaporating pressure (MPa)
entry>?tlb>
Fig. 16. Heat exchanger effectiveness as a function of the evaporating pressure for the
working uids studied using evaporator 1.
reveals that the heat exchanger effectiveness for water is lower
than those for the organic working uids R123 and R245fa,
regardless of the vehicle operating condition. Furthermore, it is
seen that the heat exchanger effectiveness for water decreases
monotonically as the evaporating pressure increases. Fig. 16 also
reveals that the heat exchanger effectiveness for water is similar for
both operating conditions 9 and 13, but decreases signicantly for
the operating condition 3. As for the organic uids, Fig. 16 reveals
yet that the heat exchanger effectiveness decreases from operating
condition 3 to operating condition 13. This is because the organic
uids allow using lower temperature heat sources and yield higher
heat exchanger effectivenesss when the temperatures of the
exhaust gases are relatively low.
It should be noted that when evaporator 1 is considered in the
simulations, the temperature of the exhaust gases decreases only
by about 200  C. In this case the value of Tg,out was well above that
(200  C) considered in the RC thermodynamic analysis in the
previous section.
Fig. 17 shows the RC net power output as a function of the
evaporating pressure for the working uids studied using evapo-
rator 1 for the operating conditions 3, 9 and 13. For the operating
conditions 3 and 9, the gure reveals that the RC net power output
is higher for the two organic uids than for water for evaporating
pressures >1 MPa. This is due to the lower heat exchanger
working uids considered, see Fig. 16. Consistently, for the oper-
ating condition 13, Fig. 17 reveals that for evaporating pressures
<2.25 MPa the RC net power output is (slightly) higher for water
than for the organic uids. In contrast, for evaporating pressures
>2.25 MPa the RC net power output is higher for the organic uids
than for water owing to the increase in the organic uids mass ow
3 3.5
Table 6 shows the heat exchanger characteristics for evaporators
1 and 2 for the operating condition 13 (Table 1). It is seen that the
overall heat transfer coefcient is higher for evaporator 2 than for
evaporator 1, mainly due to the reduced cross ow area of the heat
exchanger (2.5 times lower as compared to the normal cross ow
area of the existing exhaust duct in the vehicle), which increases
the ow velocity of the exhaust gases.
In regard to the heat transfer characteristics of the working
uids, Table 6 demonstrates that the overall heat transfer coef-
cient is higher for water followed by the organic uids R245fa and
R123. Note that Table 6 examines the operating condition 13,
which corresponds to a high vehicle load and high engine speed
(see Table 1). This means that the poor water evaporator effec-
tiveness observed for the operating condition 3 (see Fig. 16) is
reversed for the operating condition 13. Hence, it can be
concluded that water is the best working uid for use in vehicle
WHR applications through RCs when the exhaust gases tempera-
tures are relatively high.
5.3. Inuence of the RC on the ICE thermal and vehicle mechanical
efciencies
This section evaluates the inuence of the RC on the ICE thermal
and vehicle mechanical efciencies. The main objective of this
section is to perform a systematic assessment of the RC vehicle
exhaust WHR potential.
The results from the RC thermodynamic and heat exchanger
models were used to calculate improvements in efciencies. The
ICE thermal efciency was calculated through the following
equation:

3 W_ net W _ net
hth (19)
Op. condition 13 Water _ fuel $LHV
m A
Op. condition 9 R123 _ air = $LHV
m
2.5 F
Op. condition 3 R245fa
Net power output (kW)

where LHV (low heating value) was taken equal to 44 MJ/kg [33].
2 The vehicle mechanical efciency was dened as the ratio of the
useful work produced by the RC, W _ net , and the effective power
1.5 produced by the ICE, Pe:

1 Table 6
Heat exchanger characteristics for evaporators 1 and 2 (operating condition 13, see
Table 1).
0.5
Evaporator Working Thermal Area [m2] Overall heat
uid resistance, transfer coefcient,
0 1/UA [K W1] U [W m2 K1]
0.5 1 1.5 2 2.5 3 3.5
1 Water 0.027 0.675 55.50
Evaporating pressure (MPa)
R123 0.039 38.04
entry>?tlb>

R245fa 0.034 43.69

Fig. 17. RC net power output as a function of the evaporating pressure for the working
2 Water 0.024 0.501 75.17
uids studied using evaporator 1.
84 A. Domingues et al. / Energy 49 (2013) 71e85

_ net
W The RC3 allows assessing the thermal and mechanical efcien-
hm (20) cies of a short term RC prototype. The results found demonstrate
Pe
that future RC prototypes require improved evaporator designs and
The calculations were performed for three operating conditions; expander devices allowing for higher evaporating pressures.
specically, operating conditions 3, 9 and 13 in Table 1. In addition, The operation of the evaporator at higher pressure allows
three distinct RCs were evaluated: increasing the RC rst and second law efciencies and permits
reducing the temperature differences across the heat exchanger
i) RC1 that considers an adiabatic heat exchanger and an surfaces, which minimize the lm boiling effect in which rates of
evaporating pressure of 2 MPa; heat transfer fall sharply [35].
ii) RC2 that considers evaporator 1 (see Table 5) and an evapo- The open literature [e.g., [7e9, 15]] generally analysis the
rating pressure of 2 MPa; maximum WHR potential of a RC tted to the vehicle exhaust. In
iii) RC3 that considers evaporator 2 (see Table 5) and an evapo- the present study the maximum WHR potential was evaluated
rating pressure of 0.52 MPa. In this case a commercial avail- through the RC1. The ICE thermal efciencies obtained in this study
able expander (Green Turbine) that works only with water for the RC1 are in agreement with those presented by Yamada and
as a working uid is considered. Mohamad [9], who reported increases in thermal efciencies of
2.9%e3.7%, as compared with 1.4%e3.52% in this study (see
Note that the RC3 case evaluates the performance of a RC Table 7). In regard to the increase in vehicle mechanical efciency
prototype system that uses existing components: evaporator 2 and (fuel economy), Vaja and Gambarotta [7] reported values around
the Green Turbine, which is currently, to the best of our knowl- 12%, as compared with 10.16%e15.95% in this study (see Table 7).
edge, the most appropriate expander for a RC vehicle application. Srinivasan et al. [8] examined the exhaust waste heat recovery
The main characteristics of the Green Turbine are inlet steam potential of a high efciency, low emissions, dual-fuel, low
maximum pressure of 5.2 bar, inlet steam maximum temperature temperature combustion engine using an ORC. Their results
of 200  C, power of 2.5 kW, weight of 7 kg, length of 25 cm and showed that vehicle mechanical efciency improved by an average
diameter of 19 cm [34]. The reduced mass and dimensions makes of 7% for all injection timings and loads with hot EGR. Finally,
this turbine suitable for vehicle exhaust WHR applications. Chammas and Clodic [15] reported vehicle mechanical efciency of
The control of the RC in a vehicle application is particularly 24%, but for higher evaporating pressures, with the thermal energy
complex due to the (often) transient regime of the heat source. As a being recovered also from the vehicle cooling system.
result, optimizing the RC control is crucial to maximize the
performance of the system. For instance, to maintain the evapo-
6. Conclusions
rator pressure at 5 bar and to guarantee an inlet turbine tempera-
ture constant at 200  C it is needed to control the water ow rate
The vehicle exhaust WHR potential using a RC has been evalu-
through the evaporator by varying the pump speed.
ated with the aid of both a RC thermodynamic model and a heat
Table 7 presents the increase in the ICE thermal and vehicle
exchanger model. Both models used as input experimental data
mechanical efciencies for the three RCs studied. Water is the
obtained in a vehicle tested on a chassis dynamometer. The ther-
working uid with the greatest potential to be used in WHR from
modynamic analysis was performed for water, R123 and R245fa
exhaust gases in vehicles through RC1 as compared to the organic
and revealed the advantage of using water as the working uid in
working uids R123 and R245fa. Specically, at an evaporating
applications of thermal recovery from exhaust gases of vehicles
pressure of 2 MPa, for operating condition 9, the increase in the
equipped with a spark-ignition engine. Moreover, the heat
vehicle mechanical efciency using the RC1 is 15.95%, 13.43% and
exchanger effectiveness for the organic working uids R123 and
10.64% for water, R123 and R245fa, respectively.
R245fa is higher than that for the water and, consequently, they can
In the case of RC2, Table 7 shows that the organic uid R123 is
also be considered appropriate for use in vehicle WHR applications
more suitable to be used in WHR from exhaust gases in vehicles
through RCs when the exhaust gas temperatures are relatively low.
through RCs, especially at lower exhaust gases temperatures.
For an ideal heat exchanger the simulations revealed increases
However, for higher temperatures and higher exhaust gas ows,
in ICE thermal efciency and vehicle mechanical efciency of 1.4%e
RC2 presents higher thermal and mechanical efciencies if water is
3.52% and 10.16%e15.95%, respectively, while for a shell and tube
used as the working uid.
heat exchanger, the simulations showed an increase of 0.85%e1.2%
Table 7 reveals that the RC3 presents thermal and mechanical
in the thermal efciency and an increase of 2.64%e6.96% in the
efciencies of 0.3%e0.85% and 2.17%e3.87%, respectively. As
mechanical efciency for an evaporating pressure of 2 MPa.
compared with the RC1 (see Table 7) these values are rather
However, it is important to note that the thermal and mechanical
modest. However, note that the RC3 accounts for the practical
efciencies can be enhanced with the increase in the evaporating
constrains introduced by the existing RC components.
pressure of the working uid.
The present analysis conrms that RCs have high potential for
Table 7 vehicle exhaust waste heat recovery. However, improved evapo-
Increase in the ICE thermal and vehicle mechanical efciencies for the three RCs rator designs and appropriate expander devices allowing for higher
studied. evaporating pressures are required to obtain the maximum WHR
RC Working Increase of thermal Increase of mechanical potential from vehicle RC systems. Considering increasing fuel
uid efciency [%] efciency [%] prices and environmental issues, this technology will permit to
(3) (9) (13) (3) (9) (13) achieve further reductions in engine specic fuel consumption and
1 Water 2.11 2.98 3.52 15.24 15.95 15.94 CO2 specic emissions.
R123 1.77 2.51 2.97 12.83 13.43 13.43
R245fa 1.40 1.99 2.35 10.16 10.64 10.63
2 Water 0.36 0.96 1.20 2.64 5.14 5.41
References
R123 0.96 1.15 1.15 6.96 6.15 5.23
R245fa 0.85 1.03 1.06 6.18 5.53 4.79 [1] Katsanos CO, Hountalas DT, Pariotis EG. Thermodynamics analysis of
3 Water 0.30 0.72 0.85 2.17 3.83 3.87 a Rankine cycle applied on a diesel truck engine using steam and organic
medium. Energy Conversion and Management 2012;60:68e76.
 A. Domingues et al. / Energy 49 (2013) 71e85
[2] Wang T, Zhang Y, Peng Z, Shu G. A review of researches on thermal exhaust
heat recovery with Rankine cycle. Renewable and Sustainable Energy Reviews
2011;15:2862e71.
[3] He M, Zhang X, Zeng K, Gao K. A combined thermodynamic cycle used for
waste heat recovery of internal combustion engine. Energy 2011;36:6821e9.
[4] Yu C, Chau KT. Thermoelectric automotive waste heat energy recovery using
maximum power point tracking. Energy Conversion and Management 2009;
50:1506e12.
[5] Wang EH, Zhang HG, Fan BY, Ouyang MG, Zhao Y, Mu QH. Study of working
uid selection of organic Rankine cycle (ORC) for engine waste heat recovery.
Energy 2011;36:3406e18.
[6] Boretti A. Stoichiometric H2ICE with water injection and exhaust and coolant
heat recovery through organic Rankine cycles. International Journal of
Hydrogen Energy 2011;36:12591e600.
[7] Vaja I, Gambarotta A. Internal combustion engine (ICE) bottoming with
organic Rankine cycles (ORCs). Energy 2010;35:1084e93.
[8] Srinivasan KK, Mago PJ, Krishnan SR. Analysis of exhaust waste heat recovery
from a dual fuel low temperature combustion engine using an organic
Rankine cycle. Energy 2010;35:2387e99.
[9] Yamada N, Mohamad MNA. Efciency of hydrogen internal combustion
engine combined with open steam Rankine cycle recovering water and waste
heat. International Journal of Hydrogen Energy 2010;35:1430e42.
[10] Miller EW, Hendricks TJ, Peterson RB. Modeling energy recovery using ther-
moelectric conversion integrated with an organic Rankine bottoming cycle.
Journal of Electronic Materials 2009;38:1206e13.
[11] Saleh B, Koglbauer G, Wendland M, Fischer J. Working uids for low-
temperature organic Rankine cycles. Energy 2007;32:1210e21.
[12] Lai N, Wendland M, Fischer J. Working uids for high-temperature organic
Rankine cycles. Energy 2011;36:199e211.
[13] Yamamoto T, Furuhata T, Arai N, Mori K. Design and testing of the organic
Rankine cycle. Energy 2001;26:239e51.
[14] Ringler J, Serfert M, Guyotot V, Hubner W. Rankine cycle for waste heat
recovery of IC engines. SAE paper 2009-01-0174; 2009.
[15] Chammas RE, Clodic D. Combined cycle for hybrid vehicles. SAE paper 2005-
01-1171; 2005.
[16] Mavridou S, Mavropoulos GC, Bouris D, Hountalas DT, Bergeles G. Compara-
tive design study of a diesel exhaust gas heat exchanger for truck applications
with conventional and state of the art heat transfer enhancements. Applied
Thermal Engineering 2010;30:935e47.
[17] Lemmon EW, McLinden MO, Huber ML. NIST reference uid thermodynamic
and transport properties. REFPROP version 9.0. Boulder: National Institute of
Standard Technology; 2010.
[18] Moran MJ, Shappiro HN. Fundamentals of engineering thermodynamics. New
York: John Wiley and Sons; 2005.
85
[19] Cengel YA, Boles MA. Thermodynamics e an engineering approach. 6th ed.
London: McGraw-Hill; 2008.
[20] Wei D, Lu X, Lu Z, Gu J. Performance analysis and optimization of organic
Rankine cycle (ORC) for waste heat recovery. Energy Conversion and
Management 2007;48:1113e9.
[21] Mago PJ, Chamra LM, Srinivasan K, Somayaji C. An examination of regenera-
tive organic Rankine cycles using dry uids. Applied Thermal Engineering
2008;28:998e1007.
[22] Wang EH, Zhang HG, Zhao Y, Fan BY, Wu YT, Mu QH. Performance analysis of
a novel system combining a dual loop organic Rankine cycle (ORC) with
a gasoline engine. Energy 2012;43:385e95.
[23] Hettiarachchi HDM, Golubovic M, Worek WM, Ikegami Y. Optimum design
criteria for an organic Rankine cycle using low-temperature geothermal heat
sources. Energy 2007;32:1698e706.
[24] Maizza V, Maizza A. Unconventional working uids in organic Rankine-cycles for
waste energy recovery systems. Applied Thermal Engineering 2001;21:381e90.
[25] Liu BT, Chien KH, Wang CC. Effect of working uids on organic Rankine cycle
for waste heat recovery. Energy 2004;29:1207e17.
[26] Drescher U, Brggemann D. Fluid selection for the organic Rankine cycle (ORC) in
biomass power and heat plants. Applied Thermal Engineering 2007;27:223e8.
[27] Hung TC, Wang SK, Kuo CH, Pei BS, Tsai KF. A study of organic working uids
on system efciency of an ORC using low-grade energy sources. Energy 2010;
35:1403e11.
[28] Liming F, Wenzhi G, Hao Q, Bixian X. Heat recovery from internal combustion
engine with Rankine cycle; 2010. In: Proceedings of Power and Energy
Engineering Conference; March 28e31, 2010. p. 1e4. Chengdu, China.
[29] Hussain Q, Brigham D. Organic Rankine cycle for light duty passenger vehicles.
In: Directions in Engine-Efciency and Emissions Research (DEER) Confer-
ence; 2011. Presentation avalaible from: http://www1.eere.energy.gov/
vehiclesandfuels/pdfs/deer_2011/wednesday/presentations/deer11_hussain.
pdf; October 3e6, 2011. Detroit, Michigan, USA.
[30] Incropera FP, DeWitt DP, Bergman TL, Lavine AS. Fundamentals of heat and
mass transfer. New York: John Wiley & Sons; 2007.
[31] Kakac S, Liu H. Heat exchangers, selection, rating and thermal design. New
York: CRC Press; 2002.
[32] Shah RK, Dusan PS. Fundamentals of heat exchanger design. New York: John
Wiley and Sons; 2003.
[33] Graboski MS, McCormick RL. Combustion of fat and vegetable oil derived
fuels in diesel engines. Progress in Energy and Combustion Science 1998;
24:125e64.
[34] Green TurbineTM. Available from: www.greenturbine.eu [accessed on 31.03.2012].
[35] Stobart RK. An availability approach to thermal energy recovery in vehicles.
Proceedings of the Institution of the Mechanical Engineers Part D 2007;221:
1107e24.