Accepted Manuscript

Research Paper

Thermodynamic analysis of recuperative gas turbines and aero engines

C. Salpingidou, Z. Vlahostergios, D. Misirlis, S. Donnerhack, M. Flouros, A.

Goulas, K. Yakinthos

PII: S1359-4311(16)34294-6
DOI: http://dx.doi.org/10.1016/j.applthermaleng.2017.05.169
Reference: ATE 10484

To appear in: Applied Thermal Engineering

Received Date: 21 December 2016

Revised Date: 2 April 2017
Accepted Date: 28 May 2017

Please cite this article as: C. Salpingidou, Z. Vlahostergios, D. Misirlis, S. Donnerhack, M. Flouros, A. Goulas, K.
Yakinthos, Thermodynamic analysis of recuperative gas turbines and aero engines, Applied Thermal Engineering
(2017), doi: http://dx.doi.org/10.1016/j.applthermaleng.2017.05.169

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
 Thermodynamic analysis of recuperative gas turbines and aero engines

C.Salpingidou1, Z. Vlahostergios1, D. Misirlis3, S. Donnerhack2, M. Flouros2, A. Goulas1, K.Yakinthos1

1
Laboratory of Fluid Mechanics & Turbomachinery, Department of Mechanical Engineering, Aristotle University of
Thessaloniki, Thessaloniki, 54124
2
MTU Aero Engines AG, Dachauer Strasse 665, Munich, Germany
3
Technological Educational Institute (TEI) of Central Macedonia, Serres, Greece

Abstract
In the current work, the thermodynamic cycle of a conventional recuperative aero engine, in which a heat
exchanger is placed after the power turbine, is compared with the thermodynamic cycles of two non-
conventional recuperative aero engine configurations. For each configuration, different heat exchanger designs
were used, all having the same core arrangement as the heat exchanger in the conventional recuperation aero
engine which was designed by MTU aero engines AG and has been initially used in the first concept of the
Intercooled Recuperative Aero engine of MTU. The core of the heat exchangers is specially designed to
enhance heat transfer and minimize pressure losses when used as a recuperator in aero engines. Regarding the
non-conventional cycle configurations, the first one is referred to as alternative recuperative cycle, where a
heat exchanger is placed between the high pressure and the power turbine, while the second one is referred to as
staged heat recovery where two heat exchangers are employed, one between the high and power turbines and
the second one at the exhaust, downstream the power turbine. The comparison is based on the efficiencies and
the thrust specific fuel consumption of each thermodynamic cycle. The performance characteristics of the heat
exchangers were defined from previous experimental measurements and computational fluid dynamics. For all
the examined configurations, the aero engine geometrical constrains were taken into consideration, especially for
the alternative recuperative cycle. The results of the study showed that the alternative recuperative and the
staged heat recovery cycles were more efficient than the conventional recuperative cycle for a specific range of
pressure ratios and heat exchangers characteristics. These cycles combined with appropriate geometrical
adaptations and with advanced, temperature resistant ceramics, alloys and other materials have the potential to
further optimize the waste heat management exploitation in aero engines.

Keywords: gas turbines, aero engine, heat exchanger effectiveness, recuperation, staged heat recovery

1. Introduction
Fuel consumption and increased pollutants emissions of gas turbines are important factors that an engineer
should take into account for both environmental and economic reasons, especially when issues regarding global
warming and fuel supply independency arise. For this reasons the design of high efficient gas turbines, with low
fuel consumption, is of great interest and practical importance. Currently, a typical gas turbine operates with a
~40% cycle efficiency, depending always on the compressor overall pressure ratio (OPR). As a result, a ~60% of
the fuel energy is still left unexploited on the gas turbine exhaust gas and is discarded to the environment as
thermal energy. For the exploitation of this discarded thermal energy, the integration of heat exchangers
preheating the compressor discharge air in gas turbines, can be of significant value.
The use of heat exchanger based recuperative gas turbine configurations seems to be the most promising
solution which is becoming continuously more feasible by the improvement of material properties operating in
high temperature environments (e.g. Inconel alloys). For this reason, many researches are now focusing on
improving the design of heat exchangers in order to achieve high heat transfer rates and further improve the
performance of the cycle, [1]. The installation of heat exchangers can enhance gas turbines performance in order
to be competitive to major power production cycle rivals, such as Rankine steam cycles, especially for ground-
based applications, where weight and available space constraints present weak limitations. Due to the importance
of waste heat recovery for energy efficiency optimization various works have been focused on the assessment of
the performance of heat exchangers tubes, Kukulka and Smith [2] or the integration and assessment of the
performance of heat exchangers of various types thremodynamic cycles, Kilkovsk and Jegla [3] or applications,
Klemes and Varbanov [4] .
Regarding gas turbines, the conventional integration of heat exchangers is typically performed with the
installation of a heat exchanger system right after the last turbine in order to exploit the hot-gas high thermal
energy content. As heat is transferred from the hot-gas to the compressor discharge air, the latter enters the
combustion chamber with higher enthalpy content and thus, the cycle fuel demand is reduced leading to
increased cycle thermal efficiency and subsequent pollutants emission. The selection of this position for the heat
exchanger is mainly based on the available space for heat exchangers integration and the overall relative
simplicity for the installation. However, apart from this conventional approach, various researchers, have
investigated adaptations in the conventional recuperative cycle by altering the positioning or/and the number of
heat exchangers in the gas turbine applications.
More specifically, the first widely known efforts in this direction are presented again in the United States in
2002, by Dellenback [5], where an alternative recuperative (AR) gas turbine configuration was presented, with
the heat exchanger being placed between the high pressure and power turbines (inter-turbine heat exchanger). In
this concept, the heat exchangers preheated the compressor discharge air with high-temperature hot-gas, before
the latter was fully expanded across the power turbine. As a result, improved cycle efficiency in relation to the
Brayton and the conventional recuperation cycle was observed for a noticeable range of pressure ratios.
Following the work of Dellenback [5], Cai and Jiang [6] performed a further detailed analysis of the AR cycle
focusing on the effect of irreversibility factors (expressed mainly by the relative pressure losses), in the cycle
efficiency and specific work. Cai and Jiang [6] also tried to clarify some discrepancies between the conclusions
of their work and the work of Dellenback [5], which were mainly the outcome of the selection of some not
completely optimum assumptions in the work of Dellenback [5]. However, it must be mentioned that in the
analysis of Cai and Jiang [6] the recuperator effect on the thermodynamic cycles was included through the use of
a new proposed criterion, named as the average heat transfer temperature difference (AHTTD) between the hot
gas and the cold air in the recuperative process, which is presented in Cai [7], and not by the widely used
recuperator effectiveness criterion. Thus, the results and conclusions of the work of Cai and Jiang [6] are not
directly comparable to usual recuperative gas turbine cycle analyses. Following the analysis of Cai and Jiang [6],
Dellenback [8] in 2006 extended his previous work by using more appropriate assumptions and detailed models.
The analysis showed that AR cycle is capable of higher efficiencies in relation to conventional recuperation but
for a much more narrow range of operational conditions than the one presented by himself in 2002.
Furthermore, in 2006 Dellenback [8] proposed an additional configuration that combines both conventional
(CR) and AR and is referred as Staged Heat Recovery (SHR). In SHR configuration two heat exchangers are
used, the first one (primary) is placed between the high pressure and power turbines, at the same position as in
the AR configuration, and the second one (secondary) at the gas turbine exhaust similarly to the CR cycle. At his
analysis Dellenback [8], showed that there is a small benefit for improvement of cycle efficiency when SHR
configuration is applied. However, the analysis of the effect of the SHR configuration on the cycle efficiency for
a wider range of pressure ratios in relation to heat exchanger effectiveness selection and induced relative
pressure losses of both the primary and secondary heat exchangers is yet to be performed. Even though in
previous analyses, the effect of pressure losses was discussed up to a point, the combined effect of pressure
losses in relation to heat exchanger effectiveness was not discussed in detail.
Such an effort is presented in the present work through a detailed conceptual analysis of these three types of
recuperative thermodynamic cycles (i.e. CR, AR and SHR) for a wide range of the overall pressure ratio taking
also into consideration the combined effect of heat exchanger effectiveness and pressure losses. The main target
of the first part of this paper is to proceed to a detailed comparison of the performance of the different
recuperative thermodynamic cycles but only on a conceptual level and basis in order to identify the most
promising trends and configuration setups. This approach was also followed in the previously mentioned related
works. As a first step, a Brayton cycle is examined for a wide range of pressure ratios in order to set the proper
reference state for comparison. The next step is focused on the examination of the three recuperative cycles and
their impact on cycle efficiency, in relation to heat exchanger effectiveness and pressure losses. For each
recuperative cycle the range of pressure ratio in which the cycle thermal efficiency takes the highest values is
identified. Moreover, SHR and its operational range are examined and discussed in more detail, since SHR
presents the highest degree of operational flexibility, in order to assess the combined effect of both heat
exchangers in the gas turbine efficiency. At the present work, the recuperator effect on the thermodynamic cycle
performance analysis was always included through the use of the widely used recuperator effectiveness criterion
in an effort to perform the comparison of the cycles performance in a more clear and comparable way.
The last part of the present paper is focused on the implementation feasibility of the AR and the SHR
concepts in a recuperative aero engine. For the analysis, the intercooled recuperative aero engine (IRA engine)
concept, invented and developed by MTU Aero Engines AG, is used as the reference engine concept [9] in
which the AR and SHR concepts were implemented in appropriate IRA engine derivative configurations. The
IRA engine reference concept has been in detail investigated in an important number of research projects (i.e.
AEROHEX, NEWAC, LEMCOTEC projects), focused primarily on the implementation of waste heat recovery
recuperation in aero engines applications in an attempt to meet the Advisory Council for Aeronautics Research
in Europe (ACARE) stringent environmental targets for CO2 and NOx reduction.
In this concept a state of the art tubular heat exchanger, invented and developed by MTU Aero Engines AG,
is installed at the exhaust nozzle, downstream the low pressure turbine in a conventional recuperation set-up.
This heat exchanger consists of specially designed and profiled elliptic tubes placed in a 4/3/4 staggered
arrangement. More details about the engine concept can be found in the work of Boggia and Rd [10] and
regarding the heat exchanger design and thermo mechanical performance in the work of Schnenborn et al. [11].
Furthermore, a detailed assessment of intercooled recuperated cores was assessed in detail in the work of
Kyprianidis et al. [12.] which was later followed by an optimization study in the work of Xu et al. [13].
However, all these works were focused in analyses and optimization efforts for conventional recuperation cycles
only without any investigation of alternative recuperative cycles. The present work is attempting to fill this gap.
One of the main issues regarding the implementation of AR and SHR recuperative concepts on IRA-engine
derivative configurations is the limited space between the LPT and the IPT, a limitation not existing for ground
based gas turbine applications. For this reason, in both AR and SHR concepts, heat exchangers suitable to be
installed between the turbines were designed. Having always as reference the MTU tubular heat exchanger
design, various heat exchangers were designed taking into account different axial lengths and tube core
arrangement. For the calculation of the effectiveness of each heat exchanger data from already existing CFD
computations [14], combined with -NTU method, [15] were used. Regarding, the pressure losses, data from
CFD computations and experimental measurements as well were used. In order to achieve proper comparison
between the concepts, the thermal efficiency and the thrust specific fuel consumption (TSFC) were used as
comparative variables. Due to the presented high temperature values the selection of possible recuperator
materials was also taken into consideration. Additional details about this approach can be found in Salpingidou
et al. [16].

2. Development of the thermodynamic model

The first step of the analysis is the development of the computational thermodynamic models for each
thermodynamic cycle. The models were implemented in the following software tools: CAPE-OPEN/COCO [17],
CyclePad [18] and GasTurb11 [19]. For comparison and validation reasons, their results were compared initially
with the ones presented in the work of Dellenback [8], by following also the same conditions and modeling
assumptions.

Figure 1. Brayton cycle designed in CAPE-OPEN/COCO.

The first cycle under investigation is a Brayton cycle without recuperation and is presented in fig.1. It
consists of one compressor and two turbines. The first turbine (High Pressure Turbine-HPT) is used to drive the
compressor, eq.1. The power turbine (PT) produces the net work output which is given by eq.2:

. . . .
W compressor W HPT mc p12 (T2 T1) mc p12 (T4 T3 ) (1)
. .
W net mc p 45 (T5 T4 ) (2)
. .
where c p is the specific heat capacity, W the work, T the temperature and m mass flow

At the next steps, the thermodynamic models of the CR, AR and SHR cycles were developed. In the CR
configuration, a heat exchanger (HEX) is placed downstream the power turbine in order to exploit the thermal
energy of the exhaust gas, whereas in the AR cycle heat is extracted from the gas before it enters the power
turbine. As a result, in the alternative configuration, the gas has reduced enthalpy in the entrance of the power
turbine, a side-effect which has a strong impact on the produced cycle net work and will be further discussed in
next sections. Figures 2 and 3 present the thermodynamic models of the CR and AR configurations respectively.

Figure 2. Conventional recuperative cycle designed in CAPE OPEN/COCO.

 Figure 3. Alternative recuperative cycle designed in CAPE OPEN/COCO.

The last configuration that is modelled is the SHR design. Under this configuration, two HEXs are installed
in the aero engine, the first one (primary) between the high pressure turbine and power turbine and the second
one (secondary) at the exhaust, downstream the power turbine. Exploitation of energy occurs in two stages and
the produced work of the power turbine is less due to the reduced enthalpy in position 12, as shown in fig.4.

Figure 4. Staged heat recovery recuperative cycle designed in CAPE OPEN/COCO.

In order to evaluate and compare the performance of the recuperative cycles, the thermal efficiency of each
cycle is calculated using eq.3:

net work output

cycle (3)
heat input
Where the net work output is the work produced by the power turbine and in case of a Brayton cycle is given
by eq.4. The heat input in the combustion chamber (following the numbering of fig.1) is equal to:

. .
Qin mc p 23 (T3 T2 ) (4)
.
where Q in is the heat input
Another important parameter which strongly affects the thermodynamic cycle efficiency is the HEX
effectiveness () which is defined in eq. 5:
.
Q achieved
. (5)
Q max
. .
where Q achieved is the achieved heat transfer and Q max is the maximum theoretically possible heat transfer in the
HEX.
Moreover, in order to have an 1-1 comparison with the results and analysis of Dellenback [8], no blade
cooling was included in the conceptual analysis (a blade cooling model was included in the engine analysis in
the next sections) and the main parameters of the cycle such as polytropic efficiencies of the turbomachineries,
pressure losses etc, were kept the same as in Dellenbacks analysis and they are summarized in Table1.
Regarding the specific heat capacity of air (cp), the latter was provided as a function of temperature with its
values given by properties tables (CAPE-OPEN [17], CyclePad [20] and GasTurb11 [21]). Since the fluid after
the combustion process is not pure air but a mixture of air and fuel the thermodynamic properties of this mixture
had to be taken under consideration. For this reason the generic fuel of Gasturb 11[19] which is a mixture of
86.08 mass% of carbon and 13.92 mass% hydrogen is used and implemented in CAPE-OPEN. The low heating
value of the fuel is 43.1MJ/kg. Therefore, downstream the combustion chamber the thermodynamic properties
are calculated for the air and fuel (added during the combustion process) mixture.
 Table 1. Main parameters of the thermodynamic models.
Parameter Value
Turbine Inlet Temperature -TIT 1500C
HEX effectiveness- 0.3-0.9
P/P HEX inner flow 2%
P/P HEX outer flow 2%
P/P combustion chamber 3%
Compressor polytropic efficiency - c 90%
Turbine polytropic efficiency t 87%
Inlet conditions 1atm, 15C

3. Comparison of thermodynamic cycles

The first step of the analysis is focused on the selection of the appropriate software and the reproduction of
the results of Dellenback [7] which are used as the reference case for the investigation. Figure 5 shows the
comparative results from all thermodynamic softwares, for Dellenbacks reference case. An AR, CR and a
Brayton cycle having HEX effectiveness equal to 0.7, pressure losses through HEX at 2% for each side and
TIT=1500C are considered as the reference case. More relevant information for the reference case is shown in
Table 1.
As it can be seen in fig.5, the results of the three used software are in close agreement in relation to the
results of Dellenback [8]. Even though in general a good agreement is observed, for pressure ratio values of less
than OPR~22, the results of the alternative recuperative cycle deviate significantly from the results presented in
Dellenback [8], as indicated by the blue area in fig.5. From this pressure ratio and below, the cycle efficiency
begins to drop and conventional recuperation has superior performance to the alternative cycle.
Regarding the software selection, the CyclePad and CAPE-OPEN/COCO software tools can facilitate the
design and implementation of user-designed gas turbine architectures. Nevertheless, GasTurb11 does not provide
this capability and thus, only specific architectures can be calculated e.g. conventional recuperation and Brayton
cycles. Since, CAPE-OPEN/COCO was the most flexible software and also capable for the calculation of the
thermodynamic cycles for all recuperation configurations under investigation, it was chosen for the further
analysis of all cases. CAPE-OPEN/COCO was also suitable for the implementation of appropriate literature-
based turbine blade cooling models which were later included in the analysis for aero engine applications.

Figure 5. Efficiency of conventional, alternative and Brayton cycle as a function of the pressure ratio. Comparison between
Dellenback [8], GasTurb11, CyclePad and CAPE-OPEN/COCO.

Figure 6 shows the thermal efficiency of each configuration as a function of pressures ratio and HEX
effectiveness. The first important conclusion that can be extracted is that HEX effectiveness increase is
beneficial for both recuperative cycles, since as HEX effectiveness increases, the thermal efficiency of the cycles
also rises. It is important to underline that the HEX effectiveness increase has a more intense impact on the
thermodynamic cycle efficiency, for low pressure ratio values.
Regarding the AR cycle, it is concluded from fig.6 that pressure ratio increase leads to increased thermal
efficiency independently of HEX effectiveness. Nevertheless, as OPR increases, the thermal efficiency of the CR
increases up to a certain point (depending on effectiveness value) after which it starts to drop. More specifically,
by examining fig.6, in case of high effectiveness (=0.9) the thermal efficiency of CR constantly drops as OPR
increases, whereas for =0.5 and =0.7 the thermal efficiency becomes higher as OPR increases but after a
certain OPR value, it starts to drop. The higher the HEX effectiveness is, the lower the OPR value at which the
decrease of the efficiency starts.
Comparative results of the cycles show that CR cycle is preferable than the AR cycle for low pressure ratio
values, while the AR cycle is preferable to be used at high OPR. One of the most significant aspects is the
pressure ratio at which the AR cycle starts to outperform which is actually the value at which thermal efficiency
of AR and CR cycles are equal and for further increased OPR the AR cycle efficiency is higher than CR. This
value of the thermal efficiency depends on the HEX effectiveness and it shifts to higher values as the
effectiveness of the HEX is reduced. These OPR values are marked with grey lines in fig.6 and they are also
shown in table1.

Figure 6. Efficiency of all cycles as a function of pressure ratio and varying HEX effectiveness.

Table 2. Pressure ratio at which the AR cycle starts to outperform for various HEX effectiveness
OPR value at which the AR
HEX effectiveness
cycle starts to outperform
0.5 23
0.7 21
0.9 17

At the next stage of the analysis, the performance of the SHR cycle presented in Dellenback [7] is
investigated in detail. For a proper comparison in relation to the CR and AR configurations, the cycle inlet
conditions (i.e. maximum temperature, compressor and turbine efficiencies and pressure losses percentage per
HEX and combustion chamber) were kept the same as in the reference case. Regarding the HEXs effectiveness,
a combination of values ranging from 0.3 up to 0.9 for both HEXs were included in the investigation in relation
to different pressure ratio values.
 Figure 7. Efficiency of all cycles as a function of pressure ratio (grey area indicates the range of SHR efficiency
values for different 1 and 2 combinations)
Comparative results of the SHR cycle performance are presented in fig.7 in relation to the performance of the
CR, AR and Brayton cycles. The SHR efficiency values which are plotted correspond to various combinations of
the HEX effectivenesses, 1 and 2. The maximum SHR efficiency values (indicated with thick grey line in fig.7)
are achieved for different combinations of the primary and secondary HEX effectiveness, as shown in Table 2.

Table 3. SHR maximum cycle efficiency in relation to pressure ratio and corresponding 1 and 2 combinations

OPR 1 2 Cycle Efficiency %

5.75 0.3 0.9 56.27

10 0.3 0.9 54.70
15 0.3 0.9 52.25
20 0.3 0.9 49.86
30 0.9 0.3 48.23
40 0.9 0.3 47.30

For OPR~20 the SHR performance is maximized for a combination of 1=0.3 and 2=0.9, which is
corresponding to a setup more closely adapted to the conventional cycle (low effectiveness value for the primary
HEX and high value for the secondary one). On the other hand, for OPR>30 the SHR performance is maximized
for a combination of 1=0.9 and 2=0.3, which corresponds to a setup more closely adapted to the AR cycle (high
effectiveness value for the primary HEX and low value for the secondary one). Additionally, the selection of the
effectiveness for both the primary and secondary HEX is of crucial importance since it can lead to a cycle
efficiency variation of more than 20% in absolute values, almost independently of the pressure ratio.
Moreover, another important aspect that should be discussed in detail, is the impact of the HEXs
effectiveness sequence. The interchange of the primary and secondary HEX effectiveness has an impact on the
cycle efficiency, which is significant for low pressure ratio cases especially when the difference between the
HEX effectiveness values is high, as presented in Table 3. As it can be seen, the impact of the interchange for
different pressure ratio values leads to a cycle efficiency reduction.
 Table 4. Effect of interchange of HEX effectiveness on cycle efficiency
case1-case2 case1-case2 case1-case2
1 2
at OPR~ 5 at OPR~ 15 at OPR~ 40
Case 1 0.9 0.3
-12.1% -3.5% 4.5%
Case 2 0.3 0.9
Case 1 0.8 0.4
-9.0% -3.2% 2.7%
Case 2 0.4 0.8
Case 1 0.4 0.3
-1.5% -0.6% 0.7%
Case 2 0.3 0.4

The influence of the interchange is particularly intense for low pressure ratios and great differences between
primary and secondary HEX effectivenesses, i.e. 1=0.9, 2 =0.3. The interchange of 1 and 2 leads to a high
decrease of the cycle efficiency by 12.1% at OPR=5.75 and to significant increase at OPR=40. However, the
impact is not so strong when the effectivenesses 1 and 2 are similar, i.e. 1=0.4, 2 =0.3; the change of the
efficiency does not exceed 1.5%.

4. Investigation of the achieved work

Figure 8 describes the net work production of each cycle as a function of the pressure ratio. it can be seen,
the effect of pressure ratio and recuperative cycle configuration strongly affects the achieved net work. The
achieved works of the recuperative cycles are lower than the one of the simple Brayton cycle due to the effect of
the imposed pressure losses of the heat exchanger. These pressure losses reduce the turbine inlet pressure level
and also increase the turbine outlet pressure level. Thus, they limit the expansion degree of the turbine and they
are reducing the turbine achieved work in relation to the simple Brayton cycle. However, due to the achieved
preheating of the compressor discharge air in the recuperator less fuel is now required and thus the efficiency of
the recuperative cycles is higher than the one of the simple Brayton cycle.

Figure 8. Achieved work of all cycles as a function of pressure ratio.

More specifically, regarding the CR cycle, the deviation of the achieved work in relation to the Brayton cycle
is only affected by the HEX pressure losses. On the other hand, the AR cycle presents a much higher work
reduction in relation to the Brayton cycle. In comparison with the CR cycle for the same effectiveness value, the
achieved work of the AR cycle is significantly reduced. In figures 9a, 9b, 9c and 9d the Temperature-Entropy
diagrams of each cycle are presented. he area enclosed by the curves of the diagram represents the net work
production. The increased work reduction of the AR in relation to the Brayton cycle, is the outcome of the
combined effect of both the HEX pressure losses and the HEX effectiveness, with the effect of the latter being
the dominant contributor to the achieved work reduction as it can be seen in fig.9a.
The HEX effectiveness has a strong impact on the achieved cycle work of the AR cycle, as presented in
fig.9c, while it has no effect on the CR cycle work, fig.9b. Regarding the AR cycle, the higher the HEX
effectiveness is, the lower the enthalpy value at the power turbine inlet will be (as seen in fig.10). As a result, the
degree of the power turbine expansion is reduced when the HEX effectiveness value is increased. Thus, the
achieved work of the AR cycle takes its lowest values when =0.9 and its highest values for =0.5. For this
reason, for the CR cycle, different values of HEX effectiveness do not have any effect on the achieved work
since the power turbine expansion process and the turbine inlet enthalpy remains unaffected.
The SHR cycle presents a hybrid behavior in relation to the CR and the AR cycles. As a result, for OPR<20,
for which the CR cycle dominates as it has a higher efficiency, the achieved SHR work is lower than the CR
cycle work, yet higher than the work of the AR cycle. For this pressure ratio, the optimum HEX effectiveness
combination (in terms of cycle efficiency, as presented in Table 1) is more closely adapted to the CR design, by
having a low value for 1 and a high value for 2.
In addition, for OPR>~30 the optimum HEX effectiveness combination for the SHR is more closely adapted
to the AR design, by having a high value for 1 and a low value for 2. As a result the achieved SHR work is
significantly reduced, taking values even lower than the ones of the AR cycle. Finally, for 20<OPR<30 the SHR
work presents an intermediate behavior.

Figure 9a. Temperature-Entropy diagram of all cycles

Figure 9b. Temperature-Entropy diagram of CR.

 Figure 9c. Temperature-Entropy diagram of AR.

Figure 9d. Temperature-Entropy diagram of SHR.

Figure 10. Temperature-Entropy diagram of alternative configuration.

 5. Implementation on aero-engines
5.1. Thermodynamic model of the engine
The last part of the paper is focused on the implementation feasibility of the AR and SHR cycles on aero
engines. In order to proceed with the analysis, the MTU Intercooled Recuperative Aero engine (IRA) is used as
the reference case. The required operational and geometrical data of the engine, e.g. turbomachinery efficiencies
and cruise conditions, were taken into account for the development of the aero engine cycle model. A conceptual
architecture of IRA is presented in fig.11(left). As it can be seen, the IRA engine is a three spool geared turbofan
conventional recuperative engine in which the heat exchangers are placed downstream the low pressure turbine,
following a conventional recuperation setup. An intercooler is placed between the intermediate pressure turbine
(IPC) and the high pressure compressor (HPC) in order to reduce the compression work of the HPC. A matrix of
the originally designed by MTU HEX is shown in fig.11(right). Its core consists of special profiled elliptic tubes,
placed in 4/3/4 staggered arrangement, which provide low aerodynamic losses and enhance heat transfer. The
cold air from the compressor flows inside the tubes, while the hot gas flows on the external side of the HEX, as
shown in fig.11(right). For the proper selection of the IRA engine thermodynamic conditions, feedback from
MTU Aero Engines AG was provided for all necessary components. Thus, more appropriate values than the ones
presented in Table 1, which were corresponding to the values presented in Dellenback [8] and were mainly used
for validation and comparison reasons, were used for compressor and turbine efficiencies in order to better
reflect the current state of the technology.

Figure 11. IRA engine conceptual design (left). MTU HEX (right)

For the analytical study of the performance of the cycles, a thermodynamic cycle model of the IRA engine
was developed in CAPE-OPEN/COCO software which was then validated in relation to an IRA engine
thermodynamic model of GasTurb11. Since the results were in close agreement, as an average deviation of less
than 0.5% was presented, and since the CAPE-OPEN/COCO model was completely customizable and could
easily be adapted to model the AR and SHR concepts on IRA engine derivative configurations (something which
GasTurb cannot accommodate), it was decided to proceed the analysis only with CAPE-OPEN/COCO models.
Regarding the turbine blade cooing model, the latter was based on the suggestions of Young and Wilcock
[22],[23].
Thus, in order to study the feasibility and the performance of AR and SHR cycles, appropriate
thermodynamic models of each cycle were designed in IRA engine derivative configuration. The first step in the
analysis, was the selection of the HEX installation position, taking into strong consideration the IRA engine
general dimensions and by using data from GasTurb11 reference engines. Regarding the AR concept a HEX was
placed between the IPT and LPT, while for the SHR two HEXs were installed, one between IPT and LPT
(primary HEX) and the other at the exhaust nozzle (secondary) HEX.
In the current study, taking into account the limited available space between the IPT and LPT, various heat
exchanger cores with different axial lengths and tube core arrangements (3/2/3, 4/3/4, 5/4/5, 6/5/6 etc.) were
designed and assessed, having always the MTU tubular heat exchanger core design as reference. However, due
to the limited available space and since HEX effectiveness directly HEX available area only heat exchanger
placed in a limited space region in the aero engine, of low effectiveness values were feasible which also
presented high pressure losses due to the increased velocities of the inner and outer flow. The design of the heat
exchanger for a 4/3/4 arrangement is shown in fig.14 and is based on a new annular tubes concept named as
STraight AnnulaR Thermal RECuperator (STARTREC concept), developed by MTU and LFMT in the
framework of LEMCOTEC program. Details about this concept can be found in Misirlis et al [24]. In order to
ensure that the HEX performance characteristics would correspond to the cycle conditions, each HEX was
designed by taking into consideration the thermodynamic conditions at the certain position at which the HEX
would be installed. Furthermore, for each HEX core arrangement under investigation the overall heat transfer
coefficient was derived by using dedicated correlations for the elliptic tubes from previous CFD calculations and
experimental tests as presented in Yakinthos et al. [14], combined with the -NTU method, [15], for cross flow
heat exchangers. Since this process requires the a-priori knowledge of the HEX outlet conditions in order to
derive the values of heat transfer coefficients at the proper reference temperature, the calculation was performed
iteratively.
The pressure losses were calculated using previously derived correlations through CFD computations and
experiments which describe the macroscopic HEX performance and were used for the development of a HEX
heat transfer and pressure losses porosity model, as presented in [14] and [25]. Equations 6, 7 and 8 describe the
pressure losses general formulation.

2
DP a 0 a1v U b0 b1v b1v U

2
(6)
L L
l U 2
f DPstatic / (7)
D2

f C1 ReC2 (8)

where U is the flow velocity, v is the kinematic viscosity, is the dynamic viscosity, is the density, L is the
HEX thickness, D is the tube hydraulic diameter, l is the tubes length, a0, a1, b0, b1, b2 the viscous and inertial
pressure loss coefficients and f the friction coefficient which is a function of Reynolds number, Re. All the above
parameters such as flow velocity or kinematic viscosity correspond to the condition in the certain installation
position of the HEX inside the engine, in order to ensure that the HEX can work properly with the assumed
conditions for the cycle calculation.
Regarding the SHR cycle, the HEX which is placed at the exhaust nozzle is also another new annular tubes
design named as CORN (COnical Recuperative Nozzle) developed by LFMT and MTU, figure 14. Details about
this concept can also be found in Misirlis et al. [24].

Fig. 12. The STARTREC recuperator concept. Fig. 13. The CORN recuperator concept.

Another important issue that must be considered is the proper material selection for the HEXs. Starting with
the CR cycle, the inlet temperature for hot inflow of the HEX is relatively low, less than 700C, since the air has
already been expanded through the turbines. Therefore, the following materials, nickel-chromium alloys such as
Inconel alloy 625, Inconel 617, Haynes 214 or Haynes 230, are appropriate HEX material candidates for air
temperature less than 700C. Regarding, the SHR and AR cycles, the HEX which is placed between the turbines
deals with relatively high temperatures since the hot inflow is the IPT discharge gas. Therefore, for the HEX for
the inter turbine recuperation, ceramic materials, advanced carbon and silicon carbide composites or super alloys
can be considered since they provide temperature resistant behaviour, as mentioned in discussed in McDonald
[26].

5.2. Results
For the AR cycle nine different HEX designs were investigated. The main difference between each design
was the tube core arrangement of the HEX between IPT and LPT, and as a result the axial length of the HEX.
The shortest HEX had a 3/2/3 tube core arrangement while the longest had a 10/9/10 arrangement. The different
core arrangements have a strong impact on the size of the HEX and consequently on the effectiveness and the
pressure losses of the HEX. A schematic illustration of the AR engine is shown in fig.15. Regarding the SHR
cycle, nine different HEX designs, based on various tube core arrangements of the HEX, were also studied. A
schematic illustration of the SHR aero engine is shown in fig.16.
 Figure 14. Illustration of AR cycle. Figure 15. Illustration of SHR cycle.
For the proper evaluation of the recuperative thermodynamic cycles two performance parameters, which are
strongly related with the pollutants emissions, are calculated: the thermal efficiency and the TSFC. By definition
the thermal efficiency is the increase of the kinetic energy divided by the added fuel energy, eq. 9, while the
definition of TSFC is given by eq.10.

th (9)
heat added
.
m (10)
TSFC
Thrust
The results of the conventional recuperative aero engine are used as reference case, therefore the results of
the AR and SHR for eight different tube core arrangements are presented as relative differences in figs.17 and
18.

Tubes core arrangement

Figure 16. Results of AR cycle for various tubes core arrangement. Comparison of TSFC and thermal efficiency
values of AR cycle in relation CR.
 Tubes core arrangement

Figure 17. Results of the SHR cycle for various tubes core arrangement. Comparison of TSFC and thermal efficiency
values of SHR IR cycle in relation to CR cycle.

The AR cycle presents increased TSFC (SFC +13%, +14%) and reduced thermal efficiency (th -11%, -
12%) in relation to the CR cycle. This outcome is a direct reflection of the negative impact of the limited space
between the turbines and hence the available space for a larger HEX installation. Regarding the SHR cycle,
fig.18 shows that for some arrangements the cycle presents enhanced performance in comparison to the CR
cycle. More specifically, for the 4/3/4 and 5/4/5 tubes core arrangement both the thermal efficiency and the
TSFC are improved by ~0.6% and the SHR cycle outperforms the CR cycle. Therefore, the SHR configuration is
a promising concept taking into consideration that even with stringent geometrical constraints, regarding the
available space for installation of the HEX between the turbines, the TSFC and thermal efficiency can be
improved. This improvement can be further increased if additional adaptations of the available space between
the turbines can be implemented by providing more geometrical space for the HEX installation in more flexible
aero engine designs which could incorporate heat exchangers of increased HEX effectiveness and reduced
pressure losses.

6. Conclusions

At the first part of the present work a detailed conceptual analysis of three types of recuperative
thermodynamic cycles for aero engine applications, referred to as conventional recuperation (CR), alternative
recuperation (AR) and staged heat recovery (SHR), for a wide overall pressure ratio range, taking also into
consideration the combined effect of heat exchanger effectiveness and pressure losses is performed. The
following conclusions can be made:
For low pressure ratio values the thermal efficiency of the CR cycle is higher than the one of the AR cycle.
However, for higher pressure ratios, the AR cycle outperforms.
The pressure ratio value at which the AR cycle performance becomes better than the one of the CR cycle
shifts to higher values as the effectiveness of the heat exchanger is reduced.
For low pressure ratios the SHR performance is maximized for low effectiveness value for the primary and
high value for the secondary heat exchanger, a combination which is closer to the CR cycle.
For high pressure ratios the performance of the SHR cycle is maximized for high effectiveness value for the
primary and low value for the secondary heat exchanger, a combination which is closer to the AR cycle.
The interchange of the primary and secondary heat exchanger effectiveness for the SHR cycle affects
significantly the cycle efficiency, especially when the difference between the heat exchanger effectivenesses
values is high.
The achieved work of all cycles presents a work reduction in relation to the Brayton cycle, which is linked to
the heat exchanger pressure losses, effectiveness and positioning in the cycle.
The heat exchanger effectiveness value affects significantly the AR cycle net work since it indirectly
determines the degree of expansion in the power turbine, while it has no effect on the CR cycle net work; the
higher the heat exchanger effectiveness is, the lower the enthalpy value at the power turbine inlet and the
degree of power turbine expansion will be.
The net work of the SHR cycle (for maximum efficiency values) is strongly affected by the overall pressure
ratio and the selection of the effectiveness of the primary and secondary heat exchanger.
 For low pressure ratios, the achieved net work of the SHR is lower than the net work of the CR cycle and
higher than the one of the AR cycle. The optimum heat exchanger effectivenesses combination corresponds
to a configuration more closely adapted to the CR design, i.e. low value for primary heat exchanger
effectiveness and high value for the secondary. For high pressure ratio values, the optimum heat exchanger
effectivenesses combination is more closely related to the AR cycle .

Regarding the implementation feasibility of the AR and the SHR on appropriate derivatives of the an aero
engine which is similar to the geometry of the Intercooled Recuperated aero engine concept (IRA), which was
developed by MTU, the following conclusions can be drawn:
Due to the limited space between the turbines, only low values of inter-turbine HEX effectiveness can be
achievable.
Due to the low inter-turbine HEX effectiveness and increased pressure losses, due to limited available flow
area, the AR cycle has relatively high TSFC and relatively low thermal efficiency in relation to the
conventional recuperation cycle.
The results of the SHR cycle are promising, since for some specific tubes core arrangement of the inter-
turbine HEX, the SHR cycle has better performance than the CR by ~0.6%.
The results can be further optimized if more flexible engine geometrical constraints are taken into account,
since they would result to the possibility of using inter-turbine heat exchangers of enhanced effectiveness and
reduced pressure losses.
The analysis and results of the present work indicates the importance of introducing and assessing new
recuperative thermodynamic cycle for aero engines in order to further enhance their benefits regarding fuel
consumption and pollutants emissions reduction. The latter can be further improved when more properly adapted
and more easily mountable heat exchangers can be integrated in the aero engine configurations since currently
the stringent geometrical constraints limit the applicability of heat exchangers, especially when the latter are
placed in relatively unconventional positions. As a result, a wider degree of aero engine design flexibility can
lead to more beneficial design space and maximize the benefits of recuperation technology in aero engines.
Acknowledgements

This work is financially supported by the E.U. under the LEMCOTEC Low Emissions Core-Engine
Technologies, a Collaborative Project co-funded by the European Commission within the Seventh Framework
Programme (2007-2013) under the Grant Agreement n 283216.
Christina Salpingidou would like to thank Alexander S. Onassis Public Benefit Foundation for the
scholarship.

References

[1] P. Stehlk, Z. Jegla, and B. Kilkovsk, Possibilities of intensifying heat transfer through finned surfaces
in heat exchangers for high temperature applications, Applied Thermal Engineering, vol. 70, no. 2, pp.
12831287, 2014.
[2] D.J.Kukulka, R.,Smith, Comparison of heat exchanger designs using vipertex 1eht enhanced heat
transfer tubes, Chemical Engineering Transactions, 52, 115-120, 2016. DOI:10.3303/CET1652020
[3] B.Kilkovsk, Z.Jegla, Preliminary design and analysis of regenerative heat exchanger, Chemical
Engineering Transactions, 52, 655-660, 2016. DOI:10.3303/CET1652110
[4] J.J.Klemes, P. S.Varbanov, Heat integration including heat exchangers, combined heat and power, heat
pumps, separation processes and process control, Applied Thermal Engineering, vol. 43, 16, 2012.
[5] P. A. Dellenback, Improved Gas Turbine Efficiency Through Alternative Regenerator Configuration,
Journal of Engineering for Gas Turbines and Power, vol. 124, no. 3, p. 441, 2002.
[6] R. Cai and L. Jiang, Analysis of the recuperative gas turbine cycle with a recuperator located between
turbines, Applied Thermal Engineering, vol. 26, no. 1, pp. 8996, 2006.
[7] R. Cai, A new analysis of recuperative gas turbine cycles, in Proceedings of the Institution of
Mechanical Engineers, 1998, vol. 212, no. 4, pp. 289296.
[8] P. A. Dellenback, A Reassessment of the Alternative Regeneration Cycle, Journal of Engineering for
Gas Turbines and Power, vol. 128, no. 4, p. 783, 2006.
[9] A. Goulas, S. Donnerhack, M. Flouros, D. Misirlis, Z. Vlahostergios, and K. Yakinthos,
Thermodynamics Cycle Analysis, Pressure Loss, and Heat Transfer Assessment of a Recuperative
System for Aero-Engines, Journal of Engineering for Gas Turbines and Power, vol. 137, no. 4, Apr.
2015.
[10] S. Boggia and K. Rd, Intercooled recuperated aero engine, Technical report, Advanced Project
Design, MTU Aero Engines, 2004.
[11] H. Schonenborn, E. Ebert, B. Simon, and P. Storm, Thermomechanical Design of a Heat Exchanger for
a Recuperative Aero Engine, in Proceedings of ASME Turbo Expo 2004, Power for Land, Sea and Air,
2004, pp. 17.
[12] K.G. Kyprianidis, T.Grnstedt, S.O.T. Ogaji, P. Pilidis, R. Singh, Assessment of future aero-engine
designs with intercooled and intercooled recuperated cores, J. Eng. Gas Turbines Power 133, 011701,
2011.
[13] L. Xu, K. Kypianidis, T. Grnstedt, Optimization study of an intercooled recuperated aero-engine,
J. Propuls. Power, 29, 424432, 2013.
[14] K. Yakinthos, D. Misirlis, Z. Vlahostergios, M. Flouros, S. Donnerhack, and A. Goulas, Best Strategies
for the Development of a Holistic Porosity Model of a Heat Exchanger for Aero Engine Applications,
in Proceedings of ASME Turbo Expo 2015: Power for Land, Sea and Air, 2015.
[15] W. Kays and A. London, Compact Heat Exchangers, 3rd ed. New York: McGraw-Hill Company, Inc.,
1984.
[16] C. Salpingidou, D. Misirlis, Z. Vlahostergios, S. Donnerhack, M. Flouros, A. Goulas, and K. Yakinthos,
Investigation of the Performance of Different Recuperative Cycles for Gas Turbines/aero Engine
Applications, 2016, vol. 52, no. October, pp. 511516.
[17] COCO - the CAPE-OPEN to CAPE-OPEN simulator. [Online]. Available:
http://www.cocosimulator.org/.
[18] C. Wu, Thermodynamics and Heat Powered Cycles: a Cognitive Engineering Approach. New York:
Nova Science Publishers, 2006.
[19] GasTurb. [Online]. Available: http://www.gasturb.de/.
[20] J. K. Joseph H. Keenan, Gas Tables. New York: John Wiley & Sons, 1948.
[21] Sanford Gordon and Bonnie J. McBride, Computer Program for Calculation of Complex Chemical
Equilibrium Compositions and Applications I. Analysis. NASA Reference Publication 1311.
[22] J. B. Young and R. C. Wilcock, Modelling the Air-Cooled Gas Turbine: Part 1 General
 Thermodynamics, ASME Turbo Expo 2001: Power for Land, Sea, and Air, vol. 124, no. April 2002, p.
V002T04A020--V002T04A020, 2002.
[23] J. B. Young and R. C. Wilcock, Modeling the Air-Cooled Gas Turbine: Part 2Coolant Flows and
Losses, Journal of Turbomachinery, vol. 124, no. 2, p. 214, 2002.
[24] D. Misirlis, Z. Vlahostergios, M. Flouros, C. Salpingidou, S. Donnerhack, A. Goulas, and K. Yakinthos,
Intercooled Recuperated Aero Engine: development and optimization of innovative heat exchanger
concepts, in 2nd ECATS Conference, At Athens , Greece, 2016.
[25] C. Salpingidou, D. Misirlis, Z. Vlahostergios, M. Flouros, S. Donnerhack, and K. Yakinthos, Numerical
modeling of heat exchangers in gas turbines using CFD computations and thermodynamic cycle analysis
tools, CHEMICAL ENGINEERING TRANSACTIONS, AIDIC, vol. 52, pp. 12, 2016.
[26] C. F. McDonald and C. Rodgers, Ceramic Recuperator and Turbine: The Key to Achieving a 40 Percent
Efficient Microturbine, in Proceedings of ASME Turbo Expo 2005, 2005, pp. 963971.
 Thermodynamic models of three recuperative configurations were developed.
Heat exchanger design and engine geometrical constraints affect cycle performance.
Further optimization potential was identified for these cycles.