Energy Conversion and Management 58 (2012) 8493

Contents lists available at SciVerse ScienceDirect

Energy Conversion and Management

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

The thermodynamic characteristics of high efciency, internal-combustion engines

Jerald A. Caton
Texas A&M University, USA

a r t i c l e

i n f o

Article history:
Received 2 December 2011
Received in revised form 3 January 2012
Accepted 4 January 2012
Available online 9 February 2012
Keywords:
Internal combustion engines
Efciency
Emissions
Thermodynamics

a b s t r a c t
Recent advancements have demonstrated new combustion modes for internal combustion engines that
exhibit low nitric oxide emissions and high thermal efciencies. These new combustion modes involve
various combinations of stratication, lean mixtures, high levels of EGR, multiple injections, variable
valve timings, two fuels, and other such features. Although the exact combination of these features that
provides the best design is not yet clear, the results (low emissions with high efciencies) are of major
interest.
The current work is directed at determining some of the fundamental thermodynamic reasons for the
relatively high efciencies and to quantify these factors. Both the rst and second laws are used in this
assessment. An automotive engine (5.7 l) which included some of the features mentioned above (e.g.,
high compression ratios, lean mixtures, and high EGR) was evaluated using a thermodynamic cycle simulation. These features were examined for a moderate load (bmep = 900 kPa), moderate speed (2000 rpm)
condition. By the use of lean operation, high EGR levels, high compression ratio and other features, the
net indicated thermal efciency increased from 37.0% to 53.9%. These increases are explained in a
step-by-step fashion. The major reasons for these improvements include the higher compression ratio
and the dilute charge (lean mixture, high EGR). The dilute charge resulted in lower temperatures which
in turn resulted in lower heat loss. In addition, the lower temperatures resulted in higher ratios of the
specic heats which account for a more effective conversion of thermal energy to work. Other thermodynamic features are described.
2012 Elsevier Ltd. All rights reserved.

1. Introduction and brief literature review

Recent engine research and development activities have been
documented [15] which have demonstrated the use of novel combustion techniques to achieve low nitric oxide emissions with concurrent high thermal efciencies. These activities have used a
variety of novel combustion technologies which have included various versions of low temperature premixed combustion. These
combustion modes are often called homogeneous charge compression ignition (HCCI), pre-mixed charge compression ignition (PCCI),
and so forth. These developments have included higher compression ratios, lean mixtures, high EGR, multiple fuel injections, two
different fuels, variable valve timing, and high inlet pressures
(boost). The design that provides the most effective operation is
not yet known, but these types of features are clearly important.
As mentioned, much work has been completed on the topic of
advanced combustion technologies for IC engines. These technologies have focused on novel combustion. The following will provide
a brief summary of some of the previous work in this area. More
complete reviews of these technologies are provided in [4].
Tel.: +1 979 845 4705.
E-mail address: jcaton@tamu.edu
0196-8904/$ - see front matter 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enconman.2012.01.005

Hanson et al. [1] provided results from a heavy-duty,

direct-injection compression-ignition engine using gasoline with
various injection strategies. Their work centered on 1300 rpm
operation at two different loads: imep of 11.5 and 6.5 bar. Using
41% and 30% EGR, for the high and moderate loads, respectively,
they reported low emissions with about 50% net indicated thermal
efciency.
Kokjohn et al. [2] continued the work of Hanson et al. [1], but
with a dual-fueled HCCI and PCCI concept that resulted in a net
indicated thermal efciency of 50%. They used in-cylinder blends
of gasoline and diesel to extend the range of operation of the HCCI
type combustion. They were able to better match the desired ignition and combustion properties to specic (but changing) fuel
blends. They demonstrated successful operation for several operating conditions. One such case was a 1300 rpm, 11 bar imepnet condition that resulted in negligible emissions and a 50% net indicated
thermal efciency. This was achieved using an overall lean mixture
(u = 0.77), a high level of EGR (45.5%) and an inlet pressure of
200 kPa. Combustion was stable and repeatable.
Wilhelmsson et al. [3] presented work that was based on an HCCI
engine with supercharging using both natural gas and n-heptane.
Although they stated that natural gas appeared to be an ill-suited
fuel for pure HCCI operation, they were able to adopt an operating

85

J.A. Caton / Energy Conversion and Management 58 (2012) 8493

strategy to minimize emissions while attaining a relatively high efciency. They reported net indicated thermal efciencies of 45% and
50% for two cases.
Some investigators have shown that a combination of lean mixtures and high EGR levels works better than either alone. For
example, Tabata et al. [5] have shown that for a moderate load
and speed condition, the highest efciencies were obtained with
a lean mixture (u  0.75) and an EGR level of about 30%.
As an example of more recent work, Eichmeier et al. [6] reported on the use of a small amount of pilot diesel to ignite a highly
diluted gasolineair mixture. They found successful operation for a
heavily boosted engine, and reported a net indicated specic fuel
consumption of about 190 g/kW h (44.6% net indicated efciency)
for one operating condition with a 10 bar imepnet.
The above work suggests that high efciencies are possible for
certain combustion modes. These combustion modes are often attained with some combination of lean mixtures, high EGR levels,
high boost, and high compression ratio. These combustion modes
often result in rapid combustion. With this in mind, the purpose
of the current work is to complete a systematic assessment of engine operation with high levels of EGR, lean mixtures, high compression ratios, and rapid combustion. By considering each
feature in a step-by-step fashion, the impact of each feature will
be quantied. The resulting thermal efciencies will be reported,
and the reasons for the high efciencies will be determined.
2. Engine cycle simulation description
The cycle simulation used in this work has been described in
detail elsewhere (e.g., [7,8]). This simulation is largely based on thermodynamic formulations, and is a complete representation of the
four-stroke cycle including the intake, compression, combustion,
expansion and exhaust processes. The simulation uses detailed thermodynamic gas properties including equilibrium composition for
the burned gases. The combustion process is based on a mass fraction burn relation from Wiebe [9].
The cylinder heat transfer is an important feature of these
engines, and is still not well understood [10]. For this work, two
correlations are used for the convective cylinder heat transfer.
For the conventional conguration (case 1), the correlation from
Hohenberg [11] is used. For the subsequent congurations (cases
25), the correlation from Chang et al. [12] is used. The correlation
proposed by Chang et al. [12] is for one version of a low temperature combustion engine. Case 5 represents a conguration that
possesses the characteristics of a low temperature combustion engine. The other cases (24) are close to a low temperature combustion engine, and the Chang et al. [12] was considered
appropriate. In addition, the use of the Chang et al. correlation
would provide a less confusing comparison. Further comments
on the choice of a heat transfer correlation are provided elsewhere
[10].
The thermodynamic system is the cylinder contents (see Fig. 1).
The engine is in steady-state such that the thermodynamic state at
the beginning of each cycle (two crank-shaft revolutions) is equivalent to the state at the end of the cycle. For the compression,
expansion and exhaust processes, the cylinder contents are spatially homogeneous and occupy one zone. For the intake process,
two zones (each spatially homogeneous) are used. One zone consists of the fresh charge and the other zone consists of the residual
gases.
For the combustion processes, three zones (each spatially
homogeneous) are used. The three zones are: the unburned zone,
the adiabatic core burned zone, and the boundary layer burned
zone. The adiabatic core and boundary layer zones together comprise the burned zone. The total heat transfer is divided in an
appropriate fashion between the unburned and burned zone. The

INTAKE

EXHAUST

Adiabatic
Zone

Qu

Qb
u

Boundary
Layer

Fig. 1. A schematic of the engine cylinder depicting the three zones during
combustion.

heat transfer from the burned zone is assigned in total to the

boundary zone [7,8].
The thermodynamic properties (including pressure and temperature) vary only with time (crank angle) and are spatially uniform
in each zone. The instantaneous composition is obtained from generally accepted algorithms [13] and the species are assumed to
obey the ideal gas equation of state. The instantaneous thermodynamic properties are computed from established formulations [13]
based on the appropriate compositions.
The ow rates are determined from quasi-steady, one-dimensional
ow equations (corrected by an empirical discharge coefcient), and
the intake and exhaust manifolds are assumed to be innite plenums
containing gases at constant temperature and pressure. The instantaneous valve lift is approximated with a sinusoidal shape based on
valve timings and maximum valve lift. The fuel is completely vaporized in the intake ports and mixed with the in-coming air. The combustion efciency is 100% (i.e., no unburnt fuel1). The blow-by is
assumed zero.
Appropriate sub-models are available for EGR [14]. These submodels are based on exhaust gas compositions dictated by the
stoichiometry and include species such as carbon dioxide, water,
nitrogen, oxygen and so forth. For this work, the EGR was considered cooled to the inlet temperature.
The majority of the above assumptions and approximations
have been validated and used in a number of previous simulations.
Previous work [13,15,16] has demonstrated the success of these
types of simulations for duplicating experimental results. In addition to this previous work, examples of more recent work using
similar thermodynamic engine cycle simulations has been published. Bayraktar and Durgun [17] have used a quasi-dimensional
spark ignition engine simulation to study the effects of LPG on engine combustion and performance. Giakoumis [18] has described
the use of a transient, diesel engine simulation to determine the
cylinder wall insulation effects for both rst and second law
considerations. Finally, in 2010, Khalilarya and Javadzadeh [19] describe the development and use of a comprehensive spark ignition
engine simulation to study the details of heat loss within combustion chamber walls.

1
The use of the Wiebe function for the mass fraction burned results in a small
amount of unburned fuel at the end of the combustion period. For example, for values
of the constants a = 5.0 and m = 2.0, this small amount is 0.67% of the original fuel
mass.

86

J.A. Caton / Energy Conversion and Management 58 (2012) 8493

2.1. Energy equations

As a result of the thermodynamic analysis, governing differential equations are obtained for the gas temperatures, the cylinder
pressure, the volumes, and the masses. The instantaneous cylinder
conditions (temperatures, pressure, volumes, masses, and thermodynamic properties) as a function of crank angle are obtained by
the simultaneous numerical integration of the various differential
equations.
To complete the required input information, the boundary conditions for the inlet (temperature and pressure) and for the exhaust
(pressure) are specied. To begin a particular engine cycle calculation, several parameters are not known. The initial amount of exhaust gases (residual), and the initial cylinder gas temperature
and pressure must be assumed. The complete calculation is repeated until the nal values agree with the initial values. Depending on the initial values and the specied tolerance, this procedure
usually nds convergence within about three (3) complete cycles.
2.2. Friction
The engine friction includes the mechanical friction and the
pumping work. The mechanical friction, in turn, includes the rubbing friction (such as from the crank, pistons, and valve train)
and the work associated with the auxiliaries (such as oil and water
pump, and alternator). Algorithms for each of these items were
published by Sandoval and Heywood [20], and these are used here
exactly as presented.
2.3. Exergy parameters

where b is the specic exergy (or exergy for closed systems), u, v,

and s are the specic internal energy, the specic volume, and the
specic entropy, respectively, uo, vo and so are the specic internal
energy, specic volume and specic entropy for the restricted dead
state, respectively, and po and To are the pressure and temperature
of the dead state, respectively. The dead state is dened as the conditions of the environment at a temperature of To and a pressure of
po.
For the ow periods (open system), the ow exergy (or exergy
for ows), bf , is given by:

bf h  ho  T o s  so

where h is the specic enthalpy, ho and so are the specic enthalpy

and specic entropy of the restricted dead state, respectively, and s
is the specic entropy of the owing matter. For ows out of the
system, the owing matter is the cylinder contents, and for ows
into the system, the owing matter must be specied.
Exergy is not a conserved property, and hence, may be destroyed by irreversibilities such as heat transfer through a nite
temperature difference, combustion, friction, or mixing processes.
Between any end states, therefore, the change in the exergy may
be related to the relevant processes:

DB Bend  Bstart
DB Bin  Bout BQ  BW  Bdest

Bfuel DGT o ;po

3. Engine and operating conditions

The engine selected for this study is an automotive, 5.7 l, V8
conguration with a bore and stroke of 101.6 and 88.4 mm, respectively. Table 1 lists the engine specications as used in the present
work. For the Wiebe combustion parameters, the following values
were used as recommended by Heywood [13]: m = 2.0 and a = 5.0.
A moderate load, moderate speed operating condition was examined. Table 2 lists examples of the values of other parameters
(and how obtained) which were needed in this work. Table 3 lists
some of the parameters for case 1.
4. Results and discussion

The exergy of the fuel and the cylinder gases is described by a

number of parameters. For completeness, these parameters are
briey discussed next. Full details are available elsewhere [8].
Once the thermodynamic properties are known for a given set
of conditions, the determination of exergy is fairly straightforward.
In this development, the kinetic and potential energies are neglected (and can be shown to be negligible). At all times, for each
zone:

b u  uo  po v  v o  T o s  so

where DB is the change of the total system exergy for a process, Bend
is the total exergy at the end of the period, Bstart is the total exergy at
the start of the period, Bin is the total exergy transferred into the
system accompanying ow into the system, Bout is the exergy transferred out of the system accompanying ow out of the system, BQ is
the exergy transferred accompanying the heat transfer, BW is the
exergy transfer due to work, and Bdest is the exergy which is destroyed by irreversible processes. This relation may be used to
ascertain the destruction of exergy by solving Eq. (3) to nd Bdest.
For determining the engine efciency, and for completing the
energy and exergy balances, values are needed for the energy
and exergy of the fuel. For the fuel, the lower heating value (LHV)
evaluated for a constant pressure process is used. By denition,
the fuel exergy Bfuel is given by the Gibbs free energy

First, the methodology and constraints of the study are described, and then the results of the study are presented.
4.1. Methodology and constraints
The main idea of this study is to complete a systematic assessment of the various features that have been employed in previous
investigations that resulted in low emissions and high efciencies.
By considering each feature in a step-by-step fashion, the impact of
each feature can be quantied.
The order that the features were added was arbitrary. For the
four features considered (CR, EGR, u and hb), 24 different arrangements are possible. One sequence (the 4th) was selected as representative, but other sequences will be examined. Table 4 is a
description of the cases for the 4th sequence. Obviously, the nal

Table 1
Engine specications.
Item
Number of cylinders
Bore (mm)
Stroke (mm)
Crank rad/con rod ratio
Compression ratio
Inlet valves
Diameter (mm)
Max Lift (mm)
Opens (CA aTDC)
Closes (CA aTDC)
Exhaust valves
Diameter (mm)
Max lift (mm)
Opens (CA aTDC)
Closes (CA aTDC)
Valve overlap (deg)

Value
8
101.6
88.4
0.305
8.0
50.8
10.0
357
136
39.6
10.0
116
371
14

87

J.A. Caton / Energy Conversion and Management 58 (2012) 8493

Table 2
Common engine and fuel input parameters.

Table 3
Case 1 engine and fuel input parameters.
Item

Value used

Case 1: bmep = 900 kPa, CR = 8.0, EGR = 0%, MBT timing

Equivalence ratio
1.0
Engine speed (rpm)
2000
Frictional mep (kPa)
81.3
Inlet pressure (kPa)
92.1
Exhaust pressure (kPa)
105.0
Start of combustion (bTDC)
26.0
Combustion duration (CA)
60
Cylinder wall temp (K)
450

How obtained
Input
Input
From algorithm [20]
Input
Input
Determined for MBT
Input
Input

Table 4
Description of cases: 4th sequence.
Case

Description

1
2
3
4
5

CR = 8; hb = 60 CA; / = 1.0; EGR = 0%; Twall = 450 K; MBT timing

CR = 16; hb = 60 CA; / = 1.0; EGR = 0%; Twall = 450 K; MBT timing
CR = 16; hb = 30 CA; / = 1.0; EGR = 0%; Twall = 450 K; MBT timing
CR = 16; hb = 30 CA; / = 0.7; EGR = 0%; Twall = 450 K; MBT timing
CR = 16; hb = 30 CA; / = 0.7; EGR = 45%; Twall = 450 K; MBT timing

(base)
(CR)
(hb)
(/)
(EGR)

5. Results
Fig. 2 shows the gross indicated, net indicated and brake thermal
efciencies for this operating condition for each case as described in
Table 4. The efciency increases from case 1 through case 5. The net
indicated efciency values are, sequentially: 37.0%, 46.9%, 48.1%,
52.1%, and 53.9%. The brake efciency values are, sequentially:
33.9%, 42.6%, 43.8%, 47.2%, and 47.8%. The greatest efciency

45

Net Indicated
Efficiency

Mechanical
Friction

Brake
Efficiency

40

35

30

Case
Fig. 2. Indicated and brake thermal efciencies for each case for the 4th sequence.

Table 5
Examples of different sequences.

1
2
3
4
5
a

result (comparison of case 1 and case 5) was the same for each sequence. The individual contributions to the efciency gains were
slightly different, however, depending on the order that the features were added. This is quantied below.
The following results are based on a set of assumptions and
approximations which did not consider a few engine characteristics. For example, combustion stability, cycle-to-cycle variations,
or other combustion issues are not included. In addition, knock,
preignition, or other abnormal combustion phenomena are not
considered.
Some items are not known. For example, the exhaust pressure
will be a function of whether a turbocharger or supercharger is
used. For this work, a simple schedule was assumed for the exhaust
pressure. The exhaust pressure was assumed equal to the greater
of 105 kPa or the inlet pressure plus 10 kPa. The sensitivity of this
assumption is examined below.
As a representative fuel, only isooctane is examined. Although
in practice other fuels have been used, the implication of these
other fuels to the thermodynamics is modest [21].
Again, the focus of this work is not to describe the combustion
in any detail. Rather, the focus is to understand the generic thermodynamics based on successful combustion. Of course, successful
combustion is a challenge for some conditions, but this aspect is
beyond the scope of the current work.

Gross Indicated
Efficiency

50

EGR = 45%

Computed
For isooctane [13]
Input
For isooctane [13]
For isooctane [13]

= 0.7

5.733
15.13
319.3 K
44,400
45,670

Pumping
Losses

= 30o

Displaced volume (dm3)

AFstoich
Inlet (airfuel) temperature
Fuel LHV (kJ/kg)
Fuel exergy (kJ/kg)

bmep = 900 kPa

2000 rpm
55
MBT Timing
th
4 Sequence

How obtained

CR = 16

Value used

BASE

Item

Indicated and Brake Efficiency (%)

60

2nd
sequence

3rd
sequence

4th
sequence

5th
sequence

13th
sequence

Base
CR
EGR
/
hb (HEa)

Base
hb
EGR
CR
/ (HE)

Base
CR
hb
/
EGR (HE)

Base
EGR
hb
/
CR (HE)

Base
EGR
/
hb
CR (HE)

HE = high efciency conditions.

increases were achieved for the increase of compression ratio, the

decrease of equivalence ratio, and the increase of EGR. More modest
efciency increases were due to the shorter burn duration. Detail
reasons for these changes are given below.
Some of the gains suggested by the increases of the net indicated thermal efciency are somewhat mitigated in terms of the
brake efciencies. This is due to the increased cylinder gas pressures which increase the mechanical friction component. The higher cylinder pressures are largely due to the higher boost which is
necessary to obtain the same load.
As mentioned above, the features could be added in 24 different
arrangements or sequences. Table 5 is a list of ve (5) of these sequences. Fig. 3 shows the corresponding results for the efciencies
as each feature was added for four (4) different sequences. As
would be expected, the improvement that a particular feature provides depends on the starting condition. This is reected in these
results. Of course the total change (comparing case 1 and case 5)
is the same for all sequences.
Fig. 4 shows the net indicated efciency improvement as a relative percentage contributed by the addition of each feature from
the 24 different sequences. The relative percentage is the absolute
percentage gain divided by the starting efciency percentage. For
example, if a feature increased the net indicated thermal efciency
from 38.0% to 41.5%, this would result in a relative percentage gain
of 9.2% (3.5/38.0). The top of the bars in Fig. 4 represents the average value of the increase for all 24 sequences for each specic feature. The short horizontal lines represent the maximum and
minimum increases for the specic feature. The stars represent
the actual relative increase of each feature for the 4th sequence.
No sequence provided an exact agreement with the average for
each feature. The 4th sequence was one of several sequences that
provided representative results.

88

J.A. Caton / Energy Conversion and Management 58 (2012) 8493

(a)

(b)

Brake
Efficiency

40

35

30

Case

Case

(c)

(d)

CR = 16

CR = 16

Gross Indicated
Efficiency
Net Indicated
Efficiency

= 30o

= 0.7

= 30

EGR = 45%

35

Brake
Efficiency

45

Mechanical
Friction

Gross Indicated
Efficiency

Pumping
Losses

50

= 0.7

Net Indicated
Efficiency

Mechanical
Friction

55

bmep = 900 kPa

2000 rpm
MBT Timing
13th Sequence

EGR = 45%

Pumping
Losses

Indicated and Brake Efficiency (%)

60

BASE

Indicated and Brake Efficiency (%)

Case

bmep = 900 kPa

2000 rpm
MBT Timing
5th Sequence

40

30

Case

50

45

Brake
Efficiency

35

30

60

55

Net Indicated
Efficiency

40

= 0.7

Gross Indicated
Efficiency

CR = 16

30o

30

b=

EGR = 45%

35

= 0.7

CR = 16

40

45

EGR = 45%

Brake
Efficiency

Net Indicated
Efficiency

Mechanical
Friction

50

BASE

45

Mechanical
Friction

Pumping
Losses

= 30o

50

55

Gross Indicated
Efficiency

bmep = 900 kPa

2000 rpm
MBT Timing
3rd Sequence

BASE

Pumping
Losses

Indicated and Brake Efficiency (%)

55

60
bmep = 900 kPa
2000 rpm
MBT Timing
nd
2 Sequence

BASE

Indicated and Brake Efficiency (%)

60

Fig. 3. Indicated and brake thermal efciencies for four different sequences.

These results suggest that the increase of the compression ratio

from 8.0 to 16.0 is the most effective of the four parameter changes
for improving the thermodynamics (i.e., increasing the thermal
efciencies). Operating more lean (from stoichiometric to an
equivalence ratio of 0.7) and using a high level of EGR (from 0%
to 45%) provided similar improvements in the thermodynamics.
In general, the shorter burn duration (from 60 CA to 30 CA) resulted in the smallest improvement.
The remainder of this work will only consider the 4th sequence
as generally representative of the results.
Although no direct verication of the computed results with
experimental data is available, a comparison with similar data is
useful. Table 6 is a comparison of the results (case 5) from this
study with results from experiments by Kokjohn et al. [2]. The engine used by Kokjohn et al. was different than the current work.
Kokjohn et al. reported on a dual-fueled HCCI and PCCI concept
that resulted in a net indicated thermal efciency of 50%. They
used in-cylinder blends of gasoline and diesel to extend the range
of operation of the HCCI type combustion. They demonstrated successful operation for several operating conditions. One such case

was a 1300 rpm, 11 bar imepnet condition that resulted in negligible

emissions and a 50% net indicated thermal efciency. This was
achieved using an overall lean mixture (u = 0.77), a high level of
EGR (45.5%) and an inlet pressure of 200 kPa. Combustion was stable and repeatable.
As shown in Table 6, the conditions are similar and the results
are reasonably close. The computations resulted in a slightly higher
net indicated efciency, and this is most probably due to some idealizations of the model (such as zero blow-by and 100% combustion efciency), and differences between the modeled and the
actual heat transfer. The peak pressure and the nitric oxide emissions, however, were in good agreement. Although this comparison
was not exact, the reasonable agreement of the results between
this study (case 5) and from Kokjohn et al. [2] provides some condence that the overall thermodynamics of the LTC engine is captured with the features outlined above.
Although an exact comparison to the experimental work may
be desired, due to a lack of complete experimental information,
such a comparison was not possible. As is often the case, many major and minor parameters which are needed to complete an exact

89

J.A. Caton / Energy Conversion and Management 58 (2012) 8493

15000

30

4th sequence

25

bmep = 900 kPa

2000 rpm
"MBT" Timing
4th Sequence

Case = 5

Avg
10000

20
Min

Pressure (kPa)

Net Indicated Efficiency Improvement (%)

Max

bmep = 900 kPa

2000 rpm
MBT Timing

15

Case = 4

Case = 3
Case = 2
5000

10

Case = 1

CR

Phi

EGR
EGR

ThetaB
b

Feature
Fig. 4. Relative increase of the net indicated thermal efciency as features were
added in different sequences. The top of the bars represents the average value of the
increase for all 24 sequences. The short horizontal lines represent the maximum
and minimum increases for the specic feature. The stars represent the actual
relative increase of each feature for the 4th sequence.

Table 6
Comparisons to results from [2].
Item

Ref. [2]

This work (Case 5)

Bore/stroke (mm)
Fuels
Inlet pressure (kPa)
Geometric CR
EGR (%)
Equivalence ratio
Speed (rpm)

137/165
Gasoline/diesel
200
16.1
45.5
0.77
1300

102/88
Isooctane
170
16
45
0.7
2000

Results
Imepnet (kPa)
Net ind efciency (%)
Peak pressure (MPa)
Nitric oxide (g/kW h)

1100
50
12
0.01

1015
53.9
12.0
0.003

comparison are not supplied, or in some cases, not even known. For
example, the engine was likely not instrumented for cylinder wall
temperatures. Furthermore, for purposes of this work, only an
approximate comparison was needed to demonstrate the nature
of the results.
Fig. 5 shows the instantaneous cylinder pressures as functions
of crank angle for each case for the 4th sequence. For the higher
compression ratio (case 1 to case 2), higher pressures are observed
as expected. For the shorter burn duration (case 2 to case 3), the
maximum pressures are slightly higher. For the lean operation
(case 3 to case 4), the pressures are higher due to the higher inlet
pressure necessary to achieve the constant load. For the increase of
EGR (case 4 to case 5), the pressures increase most signicantly
due to the higher inlet pressure necessary to achieve the constant
load. As the cases increase from 1 to 5, the maximum cylinder pressure continues to increase. For these cases, the maximum pressure
occurs between about 7.5aTDC (case 5) and 13.8aTDC (case 1).
For this engine load, the maximum cylinder pressure increased

-90

-60

-30

30

60

90

CRANK ANGLE
Fig. 5. Cylinder pressure as a function of crank angle for all ve (5) cases for the 4th
sequence.

from about 4.3 MPa (case 1) to about 12.0 MPa (case 5). These
higher cylinder pressures associated with case 5 are, of course,
important considerations for these types of lean mixture engines.
For example, the engine design needs to be sufciently robust,
and the durability of the engine needs to be veried.
Fig. 6 shows the instantaneous one-zone gas temperature as
functions of crank angle for each case for the 4th sequence. The
one-zone gas temperature is the energy averaged temperature
whenever there is more than one zone. The maximum one-zone
cylinder gas temperature occurs between about 12.2 aTDC (case
5) and 24.8 aTDC (case 1). The temperature increases slightly from
case 1 to case 2 (increase of CR), and increases again from case 2 to
case 3 (short burn duration). The temperature decreases signicantly from case 3 to case 4 (lean), and further from case 4 to case
5 (increase of EGR). The overall decrease of temperature is a key
outcome of the selection of the engine parameters to achieve the
low temperature combustion. This feature also is shown in subsequent gures of the exhaust temperatures.
An average combustion temperature can be dened as the timeaverage of the one-zone temperature during the combustion
process. These average combustion temperatures are, of course,
related to the instantaneous temperatures in Fig. 6. In general,
the average combustion temperatures ranged between about
1306 K (case 5) and 1856 K (case 3).
Another useful average is the mass-average temperature of the
adiabatic zone from start of combustion to 90 aTDC. The average
adiabatic zone temperature ranged from about 1820 K (case 5) to
2426 K (case 1). The relatively low temperatures of case 5 result
in essentially zero nitric oxide formation (see below).
A related observation concerning engine thermodynamics may
be mentioned at this point. Achieving the high thermal efciencies
with low gas temperatures appears to be in conict with results
from the Carnot cycle. For the Carnot cycle, it is well known that
higher temperatures (from the high temperature source) result in
higher efciencies. This is because the Carnot cycle is based on heat
engine concepts. An internal-combustion engine, however, is not a
heat engine. As demonstrated by these results (and explained more

90

J.A. Caton / Energy Conversion and Management 58 (2012) 8493

60

2500

bmep = 900 kPa

2000 rpm
"MBT" Timing
th
4 Sequence

Case = 3

bmep = 900 kPa

2000 rpm
MBT Timing
4th Sequence

55
Exhaust
Energy (%)

50

Case = 1

45

Case = 2

Energy (% of fuel energy)

One-Zone, Average Gas Temperature (K)

3000

2000
Case = 4
1500
Case = 5
1000

40
35
30
25

Relative
Heat Transfer (%)

20

Shorter
Burn
Duration

1
15
Hig
he
rC
R

10

500

0
-90

-60

-30

30

60

35

90

40

2
This previous study [22] is different from the current study. In addition, to the
different heat transfer correlations, parameters such as compression ratio, EGR and
wall temperature were different. Furthermore, the current paper has examined the
signicance of the order that the features were applied and has reported the
relative increase of the efciencies for the different sequences.

50

55

60

Fig. 7. Relative heat transfer and exhaust energy as functions of the net indicated
efciency for the 4th sequence.

1400
High
CR

bmep = 900 kPa

2000 rpm
MBT Timing
4th Sequence

1300

Average Exhaust Temperature (K)

fully below), for an IC engine, low gas temperatures are favorable

due to lower heat losses, and higher conversion of thermal energy
to work (higher ratio of specic heats).
Fig. 7 shows the relative heat transfer and the exhaust energy as
percentages of the fuel energy as functions of the net indicated
thermal efciency for each case for the 4th sequence. This gure
is representative of the next several gures. Typically the net indicated thermal efciency is not used as the independent variable.
This is done to better illustrate the effects of the various features
as the cases progress from case 1 to case 5. Since the net indicated
thermal efciency continually increased as the cases moved from
case 1 to case 5, this serves as a convenient way to illustrate the
results.
The results in Fig. 7 show that the relative heat transfer decreases for the higher compression ratio (case 1 to case 2), and then
remains relatively unchanged for the subsequent cases. For case 5
(the high efciency conguration), the relative heat transfer was
5% of the fuel energy. In a previous study2, Caton [22] reported higher heat transfer for similar conditions for the same engine. This previous study was based on a heat transfer coefcient from Woschni
[23] which provides higher heat transfer than the Chang et al. [12]
correlation. The heat transfer from [22] using the Woschni correlation ranged between about 18% and 25%. Using the Hohenberg [11]
and Chang et al. [12] correlations for this work appears to be more
in-line with the expected heat transfer levels. The implications of
the cylinder heat transfer are discussed in more detail elsewhere [10].
The exhaust energy remains nearly constant for the higher compression ratio (case 1 to case 2), and then gradually decreases for
the addition of the remaining features. As described below, the
decreasing exhaust energy is largely due to the increasing work

45

High
EGR

Net Indicated Efficiency (%)

CRANK ANGLE
Fig. 6. One-zone, average cylinder gas temperature as functions of crank angle for
the 4th sequence.

Lean
Burn

1200

Leaner
Short
BD

1100
4
1000
High
EGR

900

800

700

600
35

40

45

50

55

60

Net Indicated Efficiency (%)

Fig. 8. The average exhaust gas temperature as functions of the net indicated
efciency for the 4th sequence.

which increases due to the more favorable thermodynamics

(higher ratio of specic heats and lower temperatures).
Fig. 8 shows the average exhaust gas temperature as functions
of the net indicated thermal efciency for each case for the 4th sequence. This average exhaust gas temperature is the temperature
that represents the energy of the exhaust gases (enthalpy averaged
temperature). For the increase of compression ratio (case 1 to case
2), the exhaust temperature remains nearly unchanged. For cases

91

J.A. Caton / Energy Conversion and Management 58 (2012) 8493

200

1.35

bmep = 900 kPa

2000 rpm
MBT Timing
th
4 Sequence

5
5
1.30

150

High
EGR

Pressure (kPa)

Average Gamma (expansion stroke)

bmep = 900 kPa

2000 rpm
MBT Timing
4th Sequence

1.25

Leaner
1

High
EGR

pexh
100

40

45

50

Short
BD

Leaner

Higher
CR

2
1.20
35

pin

High
CR

55

60

Net Indicated Efficiency (%)

Fig. 9. The average ratio of specic heats (gamma) for the expansion stroke for the
ve cases of the 4th sequence.

from 2 to 5, however, the exhaust gas temperature continues to

decrease. In general, this reects the higher work production and
the lower cylinder gas temperatures. The nal exhaust gas temperature for case 5 was 717 K which is relatively low. Such low
exhaust temperatures may be problematic with respect to the
use of turbocharging.
Fig. 9 shows the average ratio of specic heats (gamma) during
the expansion stroke as functions of the net indicated thermal efciency for each case for the 4th sequence. The average value of the
ratio of specic heats decreases slightly for the higher compression
ratio (case 1 to case 2). For the remaining cases, however, gamma
increases as the cases progress from 2 to 5 due largely to decreasing temperatures during the expansion stroke. The increase of the
ratio of the specic heats during the expansion stroke results in
increases of the thermal efciency, and is one of the dominant
reasons that explains the improvement in the efciency as the
cases increase from 2 to 5.
Since several features change as the cases move from 1 to 5, it is
difcult to isolate the effect of the increasing ratio of specic heats
on thermal efciency. As an alternative, the simple Otto cycle can
be examined. Since the simple Otto cycle is an adiabatic cycle, the
effects of the ratio of specic heats on efciency are more direct.
For example, an increase of this ratio from 1.30 to 1.31 for a compression ratio of eight results in about a 1% absolute increase of the
thermal efciency (say, from 46.4% to 47.5%). If this improvement
is indicative of actual cycles, then the increase of the ratio of specic heats shown in Fig. 9 may be responsible for about one half
of the efciency gains. This is, however, only a speculation at this
point.
Fig. 10 shows the inlet and exhaust pressures as functions of the
net indicated thermal efciency for each case for the 4th sequence.
The inlet pressure was determined so as to obtain the constant
load of a bmep of 900 kPa. For cases 14, the inlet pressure is below
100 kPa (throttled), and the exhaust pressure was 105 kPa. For case
5, the inlet pressure was 169.5 kPa to attain the required load. This
higher inlet pressure would be obtained with turbocharging or
supercharging. For case 5, the exhaust pressure was 179.5 kPa or
10 kPa greater than the inlet pressure.

Shorter
Burn
50
35

40

45

50

55

60

Net Indicated Efficiency (%)

Fig. 10. Inlet and exhaust manifold pressures as functions of the net indicated
efciency for the 4th sequence.

The sensitivity of the results concerning the assumption of the

exhaust pressure was examined. This was completed by assuming
two additional conditions for case 5: an exhaust pressure of 5 kPa
and 15 kPa greater than the inlet pressure. For these two conditions, the brake thermal efciency changed by factors of 0.995
and 1.005 compared to the base values. Again, the computations
were for a constant bmep of 900 kPa so as the exhaust pressure
changed the inlet pressure needed to change as well.
Fig. 11a shows the exhaust nitric oxide emissions as functions
of the net indicated thermal efciency for each case for the 4th
sequence. The nitric oxide concentrations change little for the increase of compression ratio (case 1 to case 2) and for the decrease
of the burn duration (case 2 to case 3). Fig. 11b shows the effect of
equivalence ratio on nitric oxide concentrations for conditions
related to cases 3 and 4. These results are for a compression ratio
of 16, a burn duration of 30 CA and 0% EGR. As the equivalence
decreases from stoichiometric, the nitric oxide concentration rst
increases, reaches a maximum for an equivalence ratio of about
0.8, and decreases for leaner mixtures. The nitric oxide change with
decreasing equivalence ratio reects the balance of decreasing
temperatures and increasing oxygen concentrations. Fig. 11a
shows that the nitric oxide concentrations increase for the lean
condition (case 3 to case 4). This is consistent with the trends just
discussed and indicates the importance of the increase of the
oxygen concentration relative to the modest decrease of the gas
temperatures.
Fig. 11a shows that for the addition of the nal feature (high
EGR; case 4 to case 5) that the nitric oxide concentrations are about
zero. The nitric oxide emissions decrease to near zero levels for
case 5 due to the low combustion temperatures. This, of course,
was the expected outcome and one of the motivations for the
low temperature combustion technologies. The details of the nitric
oxide computations are provided elsewhere [24] the methodology includes the instantaneous contribution from all zones
throughout the combustion and expansion periods.
Fig. 12 shows the exergy destruction during the combustion
process for each case for the 4th sequence. The exergy destruction

92

J.A. Caton / Energy Conversion and Management 58 (2012) 8493

operating conditions at or near the optimum values for maximizing the thermal efciencies.

30

bmep = 900 kPa

2000 rpm
MBT Timing
4th Sequence

6. Summary and conclusions

25

= 0.7

= 30o

EGR = 45%

15

CR = 16

20

BASE

Exergy Destruction During Combustion

(% fuel exergy)

Fig. 11. (a) Nitric oxide exhaust concentrations as a function of the net indicated efciency for the 4th sequence and (b) the nitric oxide exhaust concentration as a function of
equivalence ratio for conditions related to cases 3 and 4.

Case
Fig. 12. Exergy destruction during the combustion process for each case for the 4th
sequence.

for the base case is about 20% which is typical for automotive engines operating at these conditions [25]. For increases in the compression ratio, the exergy destruction decreases slightly due to the
pressure increases (which has a greater impact than the effects of
the temperature decreases) [26]. For decreases in the burn duration (case 2 to case 3), the exergy destruction decreases slightly
due to the increases of the combustion temperatures (see Fig. 6).
The exergy destruction increases as the mixture is more lean (case
3 to case 4) and as EGR is used (case 4 to case 5) this is due to the
lower combustion temperatures (see Fig. 6) for both these changes.
The effect of combustion temperatures on exergy destruction is
discussed in several references (e.g., [27,28]).
The sensitivity of the values selected for the various features
(e.g., 45% EGR) on efciency were examined for similar conditions
elsewhere [22]. In general, the values selected have provided

This investigation has presented results from both a rst law

and a second law perspective to better understand the thermodynamics of advanced high efciency engines. By systematically
examining the individual contributions of compression ratio, lean
mixtures, fast burn and high EGR, the impact of these features on
the thermodynamic efciency was determined. An automotive engine at 2000 rpm and with a bmep of 900 kPa was examined.
 The nal case (5) resulted in a net indicated efciency of 53.9%
and a brake efciency of 47.8% with negligible nitric oxide emissions. To obtain the bmep of 900 kPa, the inlet pressure needed
to be about 170 kPa for case 5. This will require a turbocharger
or supercharger. Since the exhaust gas energy is signicantly
decreased (exhaust temperature was 717 K), the use of a turbocharger may be problematic.
 Increasing the compression ratio from 8.0 to 16.0 provided one
of the more effective ways to increase the efciencies for these
conditions.
 Decreasing the equivalence ratio from 1.0 to 0.7, and increasing
the exhaust gas recirculation (EGR) to 45% provided signicant
efciency gains largely due to the signicant decrease of the relative heat transfer and the increase of the ratio of specic heats
during the expansion stroke.
 Decreasing the burn duration from 60 CA to 30 CA increased the
net indicated efciency only a modest amount. At least part of the
reason for the modest increase was due to the use of MBT timing.
 Some of the gains suggested by the increases of the net indicated thermal efciency are somewhat mitigated in terms of
the brake efciencies. This is due to the increased cylinder gas
pressures (that are necessary to obtain the same load) which
increase the mechanical friction component.

J.A. Caton / Energy Conversion and Management 58 (2012) 8493

 The exergy destruction during the combustion process

decreases and increases as the various features are considered
due primarily to the increase and decrease of the combustion
temperatures. Overall, as an example, exergy destruction was
about 20.4% and about 24.0% for cases 1 and 5, respectively.
The efciencies were higher for case 5 compared to case 1 even
though the exergy destruction during combustion was higher
for case 5.
 The lower combustion temperatures for the nal conguration
(case 5) result in negligible nitric oxides.

Acknowledgments
This work was stimulated in part by interactions with engineers
and scientists at Oak Ridge National Laboratory. The contents of
this paper, however, do not necessarily reect the opinions or
views of any of these individuals.
References
[1] Hanson R, Splitter D, Reitz R. Operating a heavy-duty direct-injection
compression-ignition engine with gasoline for low emissions. Society of
Automotive Engineers. SAE Paper No. 2009-01-1442; 2009.
[2] Kokjohn SL, Hanson RM, Splitter DA, Reitz RD. Experiments and modeling of
dual-fuel HCCI and PCCI combustion using in-cylinder fuel blending. Society of
Automotive Engineers, SAE Paper No. 2009-01-2647; 2009.
[3] Wilhelmsson C, Tunestal P, Johansson B. Operation strategy of a dual fuel HCCI
engine with VGT. Society of Automotive Engineers. SAE Paper No. 2007-011855; 2007.
[4] Annual Progress Report, FY 2010. Vehicle technologies program, DOE,
advanced combustion engine technologies, energy efciency and renewable
energy; December 2010.
[5] Tabata M, Yamamoto T, Fukube T. Improving NOx and fuel economy for
mixture injected SI engine with EGR. Society of Automotive Engineers. SAE
Paper No. 950684; 1995.
[6] Eichmeier J, Wagner U, Spicher U. Controlling gasoline low temperature
combustion by diesel micro pilot injection. In: Proceedings of the 2011 fall
technical conference of the ASME internal combustion division. Paper No.
ICEF201160042, Morgantown, WV, 0205 October 2011.
[7] Caton JA. A multiple-zone cycle simulation for spark-ignition engines:
thermodynamic details. In: Wong VW, editor. Large-bore engines, fuel
effects, homogeneous charge compression ignition, engine performance and
simulation, ICE-Vol. 372, Proceedings of the 2001 fall technical conference,
vol. 2. Wong VW editor. The ASME Internal Combustion Engine Division,
American Society of Mechanical Engineers, Paper No. 2001-ICE-412, Argonne,
IL; 2326 September 2001. p. 4158.
[8] Caton JA. A cycle simulation including the second law of thermodynamics for a
spark-ignition engine: implications of the use of multiple-zones for
combustion. Transaction of the society automotive engineers journal of
engines, vol. 1113, Paper No. 2002-01-0007, September 2003. p. 28199.
[9] Wiebe JJ. Brennverlauf und Kreisprozess von Verbrennungsmotoren. VEBVerlag Technik Berlin; 1970.

93

[10] Caton JA. Comparisons of global heat transfer correlations for conventional and
high efciency reciprocating engines. In: Proceedings of the 2011 fall technical
conference of the ASME internal combustion division, Paper No. ICEF201160017, Morgantown, WV, 0205 October 2011.
[11] Hohenberg GF. Advanced approaches for heat transfer calculations. Society of
Automotive Engineers, SAE Paper No. 790825; 1979.
[12] Chang J, Guralp O, Filipi Z, Assanis D. New heat transfer correlation for an hcci
engine derived from measurements of instantaneous surface heat ux. Society
of Automotive Engineers. SAE Paper No. 2004-01-2996; 2004.
[13] Heywood JB. Internal combustion engine fundamentals. New York
(NY): McGraw-Hill book Company; 1988.
[14] Caton JA. Utilizing a cycle simulation to examine the use of EGR for a sparkignition engine including the second law of thermodynamics. In: Proceedings
of the 2006 fall conference of the ASME internal combustion engine division,
Sacramento, CA; 58 November 2006.
[15] Heywood JB, Higgins JM, Watts PA, Tabaczynski RJ. Development and use of a
cycle simulation to predict SI engine efciency and NOx emissions. Society of
Automotive Engineers. SAE Paper No. 790291; 1979.
[16] Watts PA, Heywood JB. Simulation studies of the effects of turbocharging and
reduced heat transfer on spark-ignition engine operation. Society of
Automotive Engineers. SAE Paper No. 800289; 1980.
[17] Bayraktar H, Durgun O. Investigating the effects of LPG on spark ignition
engine
combustion
and
performance.
Energy
Convers
Manag
2005;46:231733.
[18] Giakoumis EG. Cylinder wall insulation effects on the rst- and second-law
balances of a turbocharged diesel engine operating under transient load
conditions. Energy Convers Manag 2007;48:292533.
[19] Khalilarya S, Javadzadeh M. Developing of a new comprehensive spark ignition
engines code for heat loss analysis within combustion chamber walls. Therm
Sci 2010;14(4):101225.
[20] Sandoval D, Heywood JB. An improved friction model for spark-ignition
engines. Society of Automotive Engineers. SAE Paper No. 2003-01-0725; 2003.
[21] Caton JA. Implications of fuel selection for an si engine: results from the rst
and second laws of thermodynamics. Fuel 2010;89:315766.
[22] Caton JA. An assessment of the thermodynamics associated with highefciency engines. In: Proceedings of the 2010 fall technical conference of
the ASME internal combustion division, Paper No. ICEF2010-35037, San
Antonio, TX; 1214 September 2010.
[23] Woschni G. A universally applicable equation for the instantaneous heat
transfer coefcient in the internal combustion engine. SAE Transactions, SAE
Paper No. 670931, vol. 76; 1967. p. 306583.
[24] Caton JA. Detailed results for nitric oxide emissions as determined from a
multiple-zone cycle simulation for a spark-ignition engine. In: Wong VW,
editor. ICE, proceedings of the 2002 fall technical conference, vol. 39. The
ASME Internal Combustion Engine Division, American Society of Mechanical
Engineers, Paper No ICEF-2002-491, New Orleans, LA; 811 September 2002.
p. 13148.
[25] Caton JA. Exergy destruction during the combustion process as functions of
operating and design parameters for a spark-ignition engine. Int J Energy Res,
in press. doi:10.1002/er.1807.
[26] Caton JA. The effects of compression ratio and expansion ratio on engine
performance including the second law of thermodynamics: results from a
cycle simulation. In: Proceedings of the 2007 fall conference of the ASME
internal combustion engine division, Charleston, SC, 1417 October 2007.
[27] Dunbar WR, Lior N. Sources of combustion irreversibility. Combust Sci Technol
1994;103:4161.
[28] Caton JA. On the destruction of availability (exergy) due to combustion
processes with specic application to internal-combustion engines. Energy
2000;25:1097117.