See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/287202368

Mapping the Design Space of a Recuperated, Recompression, Precompression

Supercritical Carbon Dioxide Power Cycle with Intercooling, Improved
Regeneration, and Reheat

Conference Paper · September 2014

CITATIONS READS
3 637

2 authors:

Andy Schroder Mark G. Turner

University of Cincinnati
6 PUBLICATIONS   22 CITATIONS   
155 PUBLICATIONS   1,051 CITATIONS   
SEE PROFILE
SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Solar Energy View project

Design of a boundary layer ingested fan stage View project

All content following this page was uploaded by Andy Schroder on 17 December 2015.

The user has requested enhancement of the downloaded file.

 Mapping the Design Space of a Recuperated, Recompression,
Precompression Supercritical Carbon Dioxide Power Cycle with
Intercooling, Improved Regeneration, and Reheat
Andy Schroder∗ Mark Turner†
University of Cincinnati, Cincinnati, OH, 45221, U.S.A.

Abstract

A real fluid heat engine power cycle analysis code has been developed for analyzing the zero dimensional
performance of a recuperated, recompression, precompression supercritical carbon dioxide power cycle with
intercooling, improved regeneration, and reheat. Variation in fluid properties within the heat exchangers
is taken into account by discretizing zero dimensional heat exchangers. Variation in performance with
respect to design heat exchanger pressure drops, precompressor pressure ratio, main compressor pressure
ratio, recompression mass fraction, main compressor inlet pressure, and low temperature recuperator mass
fraction have been explored throughout a range of each design parameter. Real turbomachinery efficiencies
are implemented and the sensitivity of the cycle performance and the optimal design parameters is explored.
Sensitivity of the cycle performance and optimal design parameters is also studied with respect to the minimum
heat rejection temperature and the maximum heat addition temperature. An interactive web based tool has
been developed for analyzing the performance of the entire design space in greater detail as well as plotting
temperature difference within the heat exchangers.

Nomenclature heat exchanger’s cooled fluid, J/(kg ∗ K)

cp,Heated specific heat at constant pressure of a
∆T temperature difference between a heat heat exchanger’s heated fluid, J/(kg ∗ K)
exchanger’s cooled and heated sides d heat exchanger pressure drop coefficient,
(TCooled − THeated ), K P a/K
ṁCooled mass flow rate of a heat exchanger’s h total enthalpy, J/kg
cooled fluid, kg/s hi,Cooled total enthalpy of a heat exchanger’s cooled
ṁHeated mass flow rate of a heat exchanger’s fluid, at the heat exchanger cooled side
heated fluid, kg/s inlet, J/kg
η turbomachine isentropic efficiency hi,Heated total enthalpy of a heat exchanger’s heated
φ fraction of desired heat transferred fluid, at the heat exchanger heated side
cp specific heat at constant pressure, J/(kg∗ inlet, J/kg
K) hi total enthalpy at a turbomachine inlet,
CCooled specific heat at constant pressure of a J/kg
heat exchanger’s cooled fluid, based on ho,Cooled total enthalpy of a heat exchanger’s cooled
the mass flow rate of the heat exchanger’s fluid, at the heat exchanger cooled side
cooled fluid, J/(kg ∗ K) outlet, J/kg
CHeated specific heat at constant pressure of a ho,Heated total enthalpy of a heat exchanger’s heated
heat exchanger’s heated fluid, based on fluid, at the heat exchanger heated side
the mass flow rate of the heat exchanger’s outlet, J/kg
cooled fluid, J/(kg ∗ K) ho,max,Heated maximum possible total enthalpy of a
cp,Cooled specific heat at constant pressure of a heat exchanger’s heated fluid, at the heat
∗
Graduate Student, Corresponding Author,
exchanger heated side outlet, J/kg
Email: [email protected] ho,min,Cooled minimum possible total enthalpy of a
†
Associate Professor of Aerospace Engineering heat exchanger’s cooled fluid, at the heat

1
 exchanger cooled side outlet, J/kg 1 Introduction
ho total enthalpy at a turbomachine outlet,
J/kg Closed loop Brayton cycles utilizing Supercritical CO2
hp,Cooled total enthalpy of a heat exchanger’s cooled (S-CO2 ) as their working fluid have gained interest
fluid, at the heat exchanger pinch point, in recent years for electrical power generation due to
J/kg potentially high real cycle thermal efficiencies. The
hp,Heated total enthalpy of a heat exchanger’s heated high efficiencies of closed loop Supercritical CO2 Bray-
fluid, at the heat exchanger pinch point, ton cycles may be possible because operation with the
J/kg compression phase near the critical point results in
p total pressure, P a a cycle that possesses favorable qualities of both the
pi turbomachine inlet total pressure, P a closed loop water Rankine Cycle and the traditional
po turbomachine outlet total pressure, P a open loop air Brayton Cycle. Design of such cycles
pi,Cooled total pressure of a heat exchanger’s cooled requires complex analysis to consider completely real
fluid, at the heat exchanger cooled side fluid property variations, which are a function of both
inlet, P a temperature and pressure.
pi,Heated total pressure of a heat exchanger’s heated The S-CO2 Brayton cycle features low compres-
fluid, at the heat exchanger heated side sion work (low back work ratio) when compared to
inlet, P a a traditional open loop air Brayton cycle. The lower
pp,Cooled total pressure of a heat exchanger’s cooled back work ratio results in a decreased sensitivity of
fluid, at the heat exchanger pinch point, compressor isentropic efficiency (and the isentropic
Pa efficiency of the turbine that drives the compressor)
pp,Heated total pressure of a heat exchanger’s heated on the cycle efficiency. Non-condensing cycles have
fluid, at the heat exchanger pinch point, a narrow heat addition and heat rejection tempera-
Pa tures that does not require evaporative cooling, but
P Rc compressor pressure ratio, po /pi still approximates a Carnot cycle better than an open
P Rt turbine pressure ratio, pi /po loop Brayton cycle. Because the S-CO2 Brayton cycle
s entropy, J/kg does not have to reject latent heat of vaporization at
si turbomachine inlet entropy, J/kg a constant temperature, more recuperation is possible
T total temperature, K and the cycle is more appropriate for dry cooling in hot
Ti turbomachine inlet total temperature, K climates than traditional closed loop water Rankine cy-
To turbomachine outlet total temperature, cles because more of the heat can be rejected at higher
K temperatures. Proposed S-CO2 cycles are typically
Tp heat exchanger pinch temperature, K recuperated cycles which also results in a much lower
TCooled heat exchanger cooled side temperature, pressure ratio than an unrecuperated cycle. Unrecu-
K perated cycles need higher pressure ratios in order to
THeated heat exchanger heated side temperature, be efficient because if the pressure ratio is too low, too
K much energy will not be extracted by the turbine(s)
Ti,Cooled total temperature of a heat exchanger’s and will be wasted. Recuperated cycles can have lower
cooled fluid, at the heat exchanger cooled pressure ratios because they are able to recover energy
side inlet, K that is not extracted by the turbine(s) and transfer it
Ti,Heated total temperature of a heat exchanger’s back to the high pressure side of the cycle. CO2 is a
heated fluid, at the heat exchanger heated gas of choice because it is cheap, inert, non-toxic, and
side inlet, K its critical temperature of 304K (31◦ C) is near ambient
To,Cooled total temperature of a heat exchanger’s temperature, ∼294K (21◦ C). In addition, the S-CO2
cooled fluid, at the heat exchanger cooled Brayton cycle features very high power densities due to
side outlet, K the high fluid density that occurs because of the high
To,Heated total temperature of a heat exchanger’s pressures throughout the cycle and the high molecular
heated fluid, at the heat exchanger heated weight of CO2 when compared to water or air.
side outlet, K Possible applications for S-CO2 engines include
W turbine work, J/kg base load terrestrial electrical power generation, ma-
rine, aviation, and spacecraft electrical power genera-
tion. A S-CO2 engine could be configured as a bottom-

2
 55: Cp vs. Temperature plot: carbon dioxide

15.0
8.4 MPa

7.4 MPa
10.0
9.4 MPa
Cp (kJ/kg-K)

6.4 MPa

10.4 MPa

11.4 MPa
5.00
12.4 MPa
5.4 MPa

20.4 MPa

2.4 MPa

1.4 MPa
0.000
300. 400.

Temperature (K)

Figure 1: Specific Heat (cp ) vs Temperature at various Pressures for Carbon Dioxide. The blue line indicates
the critical pressure of 7.4MPa.

ing cycle using waste heat from a traditional open loop operation subjects the turbomachinery to very different
gas turbine (traditional Brayton cycle) or as a primary inlet conditions. This presents an additional difficulty
cycle with nuclear and solar energy heat sources. It’s in developing appropriate technologies that can operate
also possible that a S-CO2 could serve as a primary cy- efficiently and stall free throughout a wider operating
cle with the combustion of fossil fuels such as coal and range and utilize the lower heat rejection tempera-
natural gas as a heat source, although the benefit in tures possible with variations in ambient air tempera-
this configuration has not yet been explored in depth. ture with time of day, season, and geographic location.
Although there are many potential advantages of These lower heat rejection temperatures could result
S-CO2 Brayton cycles, design, development, and test- in a higher cycle efficiency. These challenges are par-
ing of the appropriate turbomachinery proves to be a ticularly strong in the main compressor which operates
very challenging task. The critical pressure of carbon near, or at the critical point of CO2 .
dioxide is 7.4 MPa. The high pressures required for The high pressures also present increased structural
operation near the critical point results in increased loading and seal leakage issues, which are even more
structural loading of components, as well as very high challenging due to the high operating speeds. Nonlinear
working fluid densities. High working fluid densities re- specific heat mismatch between the high and low pres-
sult in significantly smaller turbomachinery that must sure sides of the cycle causes limitations exchanging
be operated at higher speeds than most familiar turbo- heat between high and low pressure sides, particularly
machinery and prohibit efficient low power, low speed, at lower temperatures and increased complexity in
low cost prototypes to be developed. These high speed modeling and optimizing the cycle layout. The closed
requirements inhibit ease of testing of small turbo- loop design presents additional system complexities.
machinery and small S-CO2 Brayton cycles. Strong Because of all of these design challenges, it is impor-
property gradients near the critical point present addi- tant to establish a well directed development process
tional design challenges due to the variation in fluid in order to have a successful and efficient maturity of
properties within the turbomachinery components. Fig- the components and system. This work investigates
ure 1 illustrates the non-linearity of the specific heat of the impact of system layout, component efficiency,
carbon dioxide near the critical point at different tem- and operating conditions on sizing of the components
peratures and pressures. Figure 1 was created using in the system and the overall system efficiency. Due
REFPROP’s Graphical User Interface [1]. Off design to the highly variable fluid properties the sensitivity

3
of component efficiencies will change at different op- The United States Department of Energy began sub-
erating conditions. As a result, as cycle layout and sidizing the development of engines for concentrating
component sizes change, the sensitivity of the compo- solar power applications in mid 2012[11, 12, 13].
nent efficiencies will also change. Understanding these
relationships is important during design and testing
because it helps impact the overall direction of the
3 Methodology
development process. For example, a development 3.1 Cycle Layout
engine’s layout, components, and overall performance
may be radically different from what is targeted for The layout for the most general cycle considered is
a production engine. One may choose to design an shown in Figures 2-5 and a summary of the state points
engine with a lower overall performance in order focus is shown in Table 1. Figure 2 shows the thermody-
on the design and testing of a particular component namic states on a Temperature Entropy diagram (T −s)
that could be installed in a completely different size for an example cycle configuration. The contours in
production engine. One may also design and build com- Figure 2 are colored by the specific heat of the fluid of
ponents of lower efficiency and different performance all states within a certain range. Black lines indicate
than a production engine, just to test an overall cycle the thermodynamic states throughout the cycle and
layout. the numbered points indicate key component inlet and
outlet states. Colored lines indicate constant pressure
lines starting at each of these numbered state points,
2 Prior Work and the corresponding pressures are indicated in the
legend. Because of the low pressure drop within many
The earliest reference to a supercritical carbon dioxide
components, it is difficult to distinguish the difference
power cycle is that of a patent by Sulzer in 1948[2].
in many of these constant pressure lines without mag-
Among other efforts within the 1900s, studies con-
nifying the figure dramatically. Figure 3 is similar to 2,
ducted by Angelino[3, 4] and Feher[5] in the 1960s
except the horizontal axis is pressure. The fluid type
were significant contributes to the field. Vaclav Dostal
for each region is also labeled. The liquid vapor dome
revived interest in supercritical carbon dioxide power
is collapsed to a single line in Temperature Pressure
cycles with the publication of his doctoral dissertation
(T − p) diagrams. Figure 4 is a schematic which shows
in 2004[6]. Dostal reviewed and compared a number
the main component types in the proposed system,
of cycles and layouts, and primarily analyzed a simple
which include heat exchangers, turbomachines, shafts,
recuperated S-CO2 cycle with reheat and intercooling
tanks, piping, and a generator.
and a simple recompression S-CO2 cycle in his dis-
sertation. He explored heat exchanger volumes and Cycle Efficiency: 51.94%
Line widths scaled by mass fraction.
pressure drops in the simple and recompression cycles. 1,000 Constant
5,000
Pressure 8
6
Dostal researched application specifics, economic anal- Lines
4,580

cp , Specific Heat at Constant Pressure [J/(kg*K)]

10.03MPa
ysis, plant layouts, and control schemes for use of the 900 10.00MPa
25.21MPa
25.16MPa 4,160
recompression cycle with nuclear reactors. 800 25.16MPa
24.93MPa
5 7 3,740
14.65MPa
Sandia National Laboratories has developed two su- 14.61MPa 9
Temperature [K]

700 5.84MPa
5.60MPa
3,320
percritical CO2 test rigs with their contractor, Barber-
Critical Temperature: 304.13K
600 Critical Pressure: 7.377MPa 2,900
Nichols. Their efforts were initially motivated by nu-
2,480
clear power applications. Sandia’s two test rigs have in- 500 4

cluded both a simple S-CO2 cycle and a recompression 2,060

23 10
400 1,640
S-CO2 cycle. They have successfully achieved startup 11
15
14
of both a main compressor/turbine and recompressor/- 300 12 1,220
1 13
turbine loop. Their rigs have incorporated turbine 1,000 1,500 2,000 2,500 3,000 800
alternator/generator compressor assemblies which has Entropy [J/(kg)]
limited their operating speeds to that of the maximum
speed of the alternator/generator[7, 8]. Bechtel Marine Figure 2: Temperature Entropy Diagram for the Pro-
Propulsion Corporation has also been constructing and posed System Layout
operating a similar test rig to that of Sandia’s[9].
Echogen Power Systems has been developing an en- The main compressor is the compressor with the
gine for waste heat recovery applications since 2007[10]. minimum entropy at the inlet (point 1). It is possible
that a high recompression fraction will be used and the

4
 State Point Component Inlet Component Outlet
1 Main Compressor Cooler
2 Low Temperature Recuperators, High Pressure Main Compressor
Side
3 Medium Temperature Recuperator High Low Temperature Recuperators, High Pressure
Pressure Side Side
4 High Temperature Recuperator High Pressure Medium Temperature Recuperator High
Side Pressure Side, Recompressor
5 Heater High Pressure Side High Temperature Recuperator High Pressure
Side
6 High Pressure Turbines Heater High Pressure Side
7 Reheater High Pressure Side High Pressure Turbines
8 Power Turbine Reheater High Pressure Side
9 High Temperature Recuperator Low Pressure Power Turbine
Side
10 Medium Temperature Recuperator Low Pressure High Temperature Recuperator Low Pressure
Side Side
11 Low Temperature Recuperator Total Fraction Medium Temperature Recuperator Low Pressure
Low Pressure Side Side
12 Cooler Low Pressure Side Low Temperature Recuperator Total Fraction
Low Pressure Side
13 Precompressor Cooler Low Pressure Side
14 Low Temperature Recuperator Main Fraction Precompressor
Low Pressure Side, Recompressor
15 Cooler Low Pressure Side Low Temperature Recuperator Main Fraction
Low Pressure Side

Table 1: Summary of State Points

Cycle Efficiency: 51.94%

Line widths scaled by mass fraction.
5,000
1,000 8
6
4,580
cp , Specific Heat at Constant Pressure [J/(kg*K)]

900
Gas 4,160
800 9 7 5
3,740
Temperature [K]

700 3,320

600 2,900
Supercritical Fluid 2,480
500 10 4

11 2,060
32
400 12 1,640
15
14
300 Vapor 1 1,220
13
Liquid
0 5 10 15 20 25 30 800
Pressure [MPa]
Figure 4: Proposed Supercritical Carbon Dioxide
Figure 3: Temperature Pressure Diagram for the Pro- Power Cycle Layout. There are four rotating shafts,
posed System Layout three compressors, four turbines, and up to four recu-
perators.

5
 Cycle Efficiency: 47.31% flow near the critical point on the high pressure side,
Line widths scaled by mass fraction.
1,000 Constant
Pressure 8 making the heat capacity (not specific heat capacity)
6
Lines
900
10.06MPa
10.00MPa
lower. Although the specific enthalpy change (but not
20.47MPa
20.39MPa necessarily the enthalpy change) in the recompressor is
20.39MPa
800 20.19MPa
5
greater than that of the main compressor, making the
8.24MPa 7
8.18MPa
Temperature [K]

700 2.75MPa
2.52MPa
9 recompressor’s efficiency (and the efficiency of the tur-
Critical Temperature: 304.13K bine that powers the recompressor) potentially play a
600 Critical Pressure: 7.377MPa
4
more significant effect on the back work ratio, the ben-
500 3
10 efit gained by better heat capacity matching coupled
400 2
14 11 with a recompression fraction that is not too high can
15 12
result in overall greater cycle efficiencies by employing
300
1 13 a flow split and recompression.
1,000 1,500 2,000 2,500 3,000 3,500 4,000 The cycle has three small, high speed turbines
Entropy [J/(kg)]
which are used to drive the precompressor, recompres-
Figure 5: Example System Layout Showing Two Low sor, and main compressor. Each compressor turbine
Temperature Recuperators - Temperature Entropy Di- pair is on a different shaft. After the high pressure
agram fluid is heated by the recuperators (points 2-5) and ex-
ternal heat source (points 5-6), the flow is split (point
6) and enters these three turbines. The flow is split
main compressor will actually not have the bulk of the rather than each turbine operating in series in order to
mass flow, but the name will continue to be used. The reduce the mass flow rate through each turbine so that
precompressor (points 13-14) is used to compress the the turbine will have a larger pressure drop and can
working fluid and can allow for additional recuperation operate at a lower speed better matched with the com-
(improved regeneration), helping with the specific heat pressor it is powering. In order to simplify the design
mismatch between the high and low pressure sides. space, the present study does not explore the relative
The use of the precompressor in addition to the main mass flow rates through each turbine and assumes all
compressor (points 1-2) also allows for more efficient three turbines have the same isentropic efficiency. Af-
compression overall since the compression portion of ter expanding through the turbines (points 6-7), the
the cycle occurs within a lower temperature range flow is recombined (point 7) and reheated by the exter-
(better approximating a Carnot cycle) and there is nal heat source (points 7-8). After being reheated the
additional heat rejection (intercooling). fluid then enters a power turbine (points 8-9) which
There is a flow split at the exit of the precompressor delivers shaft power to the engine’s external load (an
(point 14) and some mass flow enters a recompressor electrical generator), and is expected to operate at a
and the remaining mass flow passes through an addi- much lower speed than the 3 smaller turbines. This
tional recuperator, heat rejection heat exchangers and is believed to be the first formal proposal of such a
main compressor. The flow then recombines at the exit configuration, and was done so because of the extreme
of the recompressor (point 4). The fraction of the total speed difference expected between the power turbine
mass flow rate that enters the recompressor is called and compressors such that it is not ideal to have a
the recompression fraction. The purpose of the flow single turbine driving the compressors and electrical
split is because of the specific heat mismatch between generator or alternator. If all are on a single shaft, the
the high and low pressure sides. When the specific design of the turbomachinery can’t be optimized very
heat mismatch is too high, much of the low pressure well. The other advantage is that the small turbines
heat cannot be recuperated. In this case, there is addi- and compressors could be placed in containment ves-
tional entropy created in the heat exchanger because sels, eliminating the need for high speed, high pressure
the temperature difference between the high and low rotating seals. The only needed rotating seal would be
pressure fluid streams must be greater in order for heat on the power turbine, which would likely be a larger,
to be able to transfer and because more external heat slower turbomachine operating at a lower pressure,
has to be added to the cycle at lower temperatures. which would result in a more manageable seal design.
Rather than operate with heat exchangers with a high The test rigs designed and operated by Sandia National
specific heat mismatch and large entropy generation, Laboratories were described to have challenges and lim-
some fraction of the fluid is split (the recompression itations by utilizing a starter motor/alternator on each
fraction) and recompressed. This results in less mass shaft, due to speed limitations of the motors and high

6
pressure, high speed rotating seals[7, 8]. In order for FORTRAN functions[1] and a forked version of the
this configuration to work with no starter motors con- python-refprop module[18]. A hybrid approach was
nected to the three smaller turbines, a pressurized tank used to access fluid properties. For commonly accessed
would be required to do a blow down startup procedure properties, property data was populated in advance
in order for the engine to reach its operating speed. using REFPROP and stored into RAM for use with an
It is also possible that a positive displacement pump interpolation function. Less commonly used properties
may be able to be used in place of a tank in order to are accessed directly from REFPROP as needed. This
temporarily provide flow to the smaller turbines. hybrid approach allowed for an increase in run speed
There is a high temperature recuperator (points 4-5 because the interpolation function call was relatively
and 9-10), medium temperature recuperator (points resource expensive, but more efficient at obtaining
3-4 and 10-11), and two low temperature recuperators multiple values simultaneously. Although REFPROP
(points 2-3, 11-12, and 14-15). The low temperature supports many fluids, only carbon dioxide is currently
recuperators have an additional flow split on the high implemented in the present cycle code and all figures
pressure side (at point 2) directing flow to the two shown in the present work are results using carbon
different heat exchangers and then recombining (point dioxide only.
3). When modeling, this more general configuration The fluid property functions were setup to accept
was assumed and the mass fraction of this additional temperature and pressure (T, p), temperature and en-
flow split was explored. The mass fraction on the tropy (T, s), enthalpy and entropy (h, s), pressure and
high pressure side that was explored was called the entropy (p, s), or enthalpy and pressure (h, p) as inputs.
‘‘Low Temperature Recuperator Main Fraction High From these inputs, temperature, pressure, enthalpy,
Pressure Component Mass Fraction’’ and represented entropy, density, specific heat at constant pressure, spe-
the mass fraction that was split at point 2 and then cific heat at constant volume, speed of sound, dynamic
went through the heat exchanger which was cooling viscosity, thermal conductivity, and compressibility
the main mass fraction flow. The low temperature factor could be obtained. Because only two inputs
recuperator cooling the total mass fraction received were required the code is not yet setup in a general
the complement of the ‘‘Low Temperature Recuper- enough way that it can fully function in regions where
ator Main Fraction High Pressure Component Mass mixtures of liquids and vapor can coexist in equilib-
Fraction’’ on the high pressure side. Figure 5 shows an rium (within the liquid vapor dome). The code assumes
example of a configuration where two low temperature all fluids are either supercritical fluids or gases. As
recuperators are in use. In some cases, one or both of a result, cycles where condensing may occur are not
the low temperature recuperators will not exist, and currently studied. Although liquids should work with
in some cases the medium temperature recuperator the present fluid property lookup technique, they were
also will not exist. Figure 2 is an example configura- not considered in order to avoid the chance of a design
tion where only one low temperature recuperator is in configuration where some fluid did condense.
use. The one low temperature recuperator that is in The most complicated portion of the code is the real
use has a much smaller temperature difference on the fluid heat exchanger functions. These functions will
high pressure side than the low pressure side due to be discussed in more detail in Section 3.3. Although
a higher specific heat on the high pressure side. Also considerably simpler, functions have also been created
included in Figure 4 are heat rejection heat exchangers. to model the turbomachines. The isentropic efficiency
Details related to pressure drop and specific heat of for a compressor was defined to be:
the ambient pressure side of the heat exchanger was
not considered in the present analysis and the power ho,ideal − hi
η= (1)
required to pump the coolant was considered to be low. ho − hi
where the compressor outlet enthalpy (ho ) can be deter-
3.2 Computer Code Overview mined from a known compressor isentropic efficiency
A Supercritical Carbon Dioxide Power Cycle analy- (η), pressure ratio (P Rc ), and inlet temperature (Ti )
sis code was created from scratch using Python [14], and pressure (pi ). The compressor inlet enthalpy (hi )
NumPy [15], SciPy [16], and matplotlib [17]. Vari- and entropy (si ) are evaluated using the fluid property
able fluid properties are utilized (i.e. h = h(T, p), functions using the known temperature and pressure
cp = cp (T, p), and s = s(T, p)) throughout the code. as inputs. The ideal compressor outlet has the same
Fluid property data was obtained using REFPROP entropy as the compressor inlet. The outlet pressure

7
for both the ideal and actual compressor outlet is: With a guessed outlet pressure (po ) and a known ideal
outlet entropy (so ), a guessed ideal turbine outlet en-
po = pi ∗ P Rc (2) thalpy (ho,ideal ) could be found using the fluid property
Using a known pressure and entropy, the ideal com- functions. With a guessed ideal turbine outlet enthalpy
pressor outlet enthalpy (ho,ideal ) can be found using the known, a guessed actual turbine outlet enthalpy could
fluid property functions. With all of these parameters be calculated as
known, the actual compressor outlet enthalpy can be ho = hi − η ∗ (hi − ho,ideal ) (8)
found using the relationship:
and the guessed turbine work could then be calculated
(ho,ideal − hi )
ho = + hi (3) W = hi − ho (9)
η
With a known compressor outlet enthalpy and pressure, With a guessed turbine outlet pressure and en-
the compressor outlet temperature (To ) could be found thalpy the fluid property functions could be used to
using the fluid property functions. determine the turbine outlet temperature. The initial
A similar procedure was used for the turbines. The guess for the power turbine outlet pressure was based
isentropic efficiency of a turbine was defined to be: upon no pressure loss in the heat exchangers. After
a power turbine outlet temperature was guessed, a
hi − ho
η= (4) guessed size of the heat exchangers and pressures drop
hi − ho,ideal
could be made. With a new guessed heat exchanger
The turbine inlet enthalpy (hi ) and entropy (si ) were pressure drop, a new power turbine outlet pressure
known in terms of the inlet (Ti ) temperature and pres- guess could be made. The process repeated until the
sure (pi ). The isentropic efficiency (η) was also defined. newly guessed pressure drops stopped changing with
The ideal turbine outlet entropy (so,ideal ) was the same subsequent iterations. If the change in pressure from
as the inlet entropy (si ). For the high pressure turbines, iteration to iteration was less than 0.0001% the pres-
the actual outlet enthalpy (ho ) was defined based on sure was considered to be converged. If the change in
the energy (W ) required to drive the compressors used pressure from iteration to iteration would not decrease
in the cycle: below .0001%, the convergence criteria was relaxed to
ho = hi − W (5) 0.3%. Once a converged turbine outlet pressure was
established, the guessed turbine outlet conditions and
With a known actual outlet enthalpy and inlet en-
turbine work were assumed to be the actual values.
thalpy, an ideal outlet enthalpy (ho,ideal ) could be
If the relaxed convergence criteria could not be met,
calculated
W an exception was raised. If the process converged to
ho,ideal = hi − (6) pressure ratio less than 1, an exception was raised be-
η
Using the ideal outlet enthalpy (ho,ideal ) and entropy cause the engine would not be able to start because
(so,ideal ), the turbine outlet pressure could be deter- more work would be required to drive the compressors
mined using the fluid property functions. The ideal and overcome the heat exchanger pressure drops than
turbine outlet pressure was the same as the actual what was being produced.
turbine outlet pressure. Using a known turbine outlet This iterative process just described was not man-
enthalpy and pressure, the turbine outlet temperature aged by a turbine specific function, but rather, was
and entropy could be found using the fluid property managed by a much larger overall function that laid
functions. out the entire cycle using the heat exchanger functions
The power turbine was solved in a slightly different and turbomachinery functions. The main cycle layout
way. Rather than matching a work output, the power function also has the capability to iteratively solve for
turbine needed to match the pressure ratio so that the the other unknown pressures within the cycle, the pre-
power turbine outlet pressure equaled precompressor compressor inlet pressure and the recompressor outlet
inlet pressure plus the high, medium, and low tempera- pressure. The main cycle layout function coordinated
ture recuperator pressure drops. The iterative process the mass splits in the cycle and ensured all of the com-
worked by guessing a pressure ratio in the power tur- ponent inlet and outlet conditions were in agreement.
bine and then calculating the turbine work output. As was mentioned previously, pumping power for the
The guessed turbine outlet pressure is: ambient pressure side of the heaters and coolers are
assumed to be low and was not considered in the com-
po = pi /P Rt (7) puter code. The heat source currently modeled is that

8
of a constant heat flux (i.e. solar) or a highly regener- multiple places and not necessarily at the ends of the
ated combustion system (heater efficiency is assumed heat exchanger. Understanding the performance of
to be 100%) where no heat is wasted in the exhaust heat exchanges operating with fluids near the critical
flow. With these assumption of low pressure drop, very point is very important in accurately predicting the
simple functions were created for the heaters and cool- performance of real fluid power cycles.
ers and were called by the main cycle layout function. A one dimensional real fluid counterflow heat ex-
The main cycle layout function also totaled all of the changer solver has been developed as part of the Su-
energy flows and determined a cycle efficiency. percritical Carbon Dioxide Power Cycle analysis code.
A set of design exploration functions were created This one dimensional solver takes into account variable
to generate a permutation list and run all permutations fluid specific heats as well as different mass fractions
of the main cycle layout function in parallel. Another between the high and low pressure side. The primary
set of functions was developed to plot the results of the purpose of this solver is to understand the impact of
design exploration, the layout of the cycle, as well as the variable specific heat on heat transfer. Forced
the temperature variation and fluid properties within convection is assumed to be the limiting case where
the heat exchangers. Finally, a set of functions was convection is very high and the temperature difference
developed to display and interact with all of this data within the fluid boundary layers and heat exchanger
using a web server and web browser [19]. solid walls is negligible and the primary cause of the
temperature difference within the heat exchanger is
3.3 Real Fluid Heat Exchangers due to the variable specific heat mismatch between
the two working fluids. Conduction along the length
Most heat exchangers operate with fluids with nearly of the heat exchanger is also assumed to be negligible.
constant specific heats and/or are changing phase. The case of a heat exchanger with very high forced
Counterflow heat exchangers where both fluids have convection is believed to be a reasonable assumption in
constant and similar specific heats have a constant and the case of Supercritical Carbon Dioxide Power Cycles
equal slope for the temperature on both the heated because the very high pressures result in very high
and cooled side. Counterflow heat exchangers where fluid densities that typically lead to very small heat
the two fluids have constant but dissimilar specific exchanger passages with high convection. These very
heats are characterized by constant but unequal sloped small passages are feasible because pure carbon dioxide
fluid temperature. Counterflow heat exchangers where is a fairly inert fluid which will result in minimal scal-
the cooled side is condensed and the heated side has ing of small passages and because carbon dioxide has
a constant specific heat are characterized by a con- a relatively low viscosity. Additionally, the manufac-
stant temperature on the cooled side and a constant turability of heat exchangers with very small passages
sloped temperature on the heated side. Counterflow is becoming a reality through the use of advanced
heat exchangers where the heated side is vaporized and manufacturing techniques.
the cooled side has a constant specific heat are charac- The pressure drop in the heat exchanger is defined
terized by a constant temperature on the heated side as a linear function of the temperature drop, but this
and a constant sloped temperature on the cooled side. pressure drop is not computed based on any assumed
For all of these cases, the location of minimum tem- geometry. The slope of this linear function is defined
perature difference between the high and low pressure by a coefficient, d. The definition of a pressure drop
sides is at an end of the heat exchanger. For the case as a function of temperature allows one to explore the
where the specific heats are constant and similar, the impact of the pressure drop in the heat exchanger on
temperature difference is constant throughout the heat the overall power cycle performance that would be re-
exchanger and therefore the location of minimum tem- quired to achieve this limiting case of high convection.
perature difference also occurs throughout the entire This aids the heat exchanger designer by providing a
heat exchanger. The location of minimum temperature reference on how important it is to minimize pressure
difference is sometimes referred to as the pinch point. loss in order to achieve the very high forced convec-
Heat exchangers operating with fluids near the tion coefficients. The assumption that the pressure
critical point possess wildly nonlinear and dissimilar drop is a linear function of the temperature drop was
specific heats which dramatically complicates perfor- used because it is assumed that the length of the heat
mance analysis. The slope of the fluid temperature exchanger will be related to the temperature drop in
is nonlinear and the minimum temperature difference the heat exchanger and the longer the length of the
between the heated and cooled fluids may occur in heat exchanger, the larger the pressure drop. This

9
assumption could be improved in many ways, but was can be represented by
utilized because of its simplicity and because low sys-
tem component pressure drops are anticipated. (hp,Heated − hi,Heated ) ∗ ṁHeated
The specific heat was evaluated for 200 discrete = (hp,Cooled − ho,Cooled ) ∗ ṁCooled (12)
temperatures in the heat exchanger, on both the high
and low pressure sides. The ratio of these specific and
heats between the high and low pressure side was then
calculated at each temperature. Information about (ho,Heated − hp,Heated ) ∗ ṁHeated
the relative specific heat between the high and low = (hi,Cooled − hp,Cooled ) ∗ ṁCooled (13)
pressure side was used to help accelerate the solution
process. Key information included the average specific where ho,Heated and ho,Cooled are the enthalpies at the
heat ratio, temperatures where the specific heat ratio outlet of the heated and cold side of the heat exchanger.
was equal to 1, the average slope of the specific heat ṁHeated and ṁCooled are the mass flow rates on the
ratios, as well as the average concavity of the specific heated and cooled sides. Equations 12 and 15 could
heat ratios. The logic which utilized these qualities of then be solved for to find the outlet enthalpies
the specific heat ratios aimed to predict the tempera-
ture (Tp ) where pinching occurred: at the high or low (hp,Heated + hi,Heated ) ∗ ṁHeated
ho,Cooled = hp,Cooled −
temperature of the heat exchanger, everywhere in the ṁCooled
heat exchanger, or in the middle of the heat exchanger (14)
(which could be more than one temperature). and
Knowing the location of a pinch point was desired (hi,Cooled − hp,Cooled ) ∗ ṁCooled
because it eliminated one unknown from the problem ho,Heated = + hp,Heated
ṁHeated
since both the high and low pressure fluid streams (15)
were at approximately the same temperature (based The outlet temperatures on the heated (To,Heated ) and
on the assumption of infinitely high convection). With cooled (To,Cooled ) sides could then be found using the
a known temperature where both fluid streams were known outlet enthalpies and the outlet pressures using
at the same temperature, the problem reduced to a the fluid property functions. If the pinch point was at
simple single control volume energy balance with one an endpoint, a similar, but simpler procedure was used
unknown temperature for cases where the pinching because there were less unknowns.
occurred at an inlet or outlet, or two simple control Once all heat exchanger outlet temperatures were
volumes with two unknown temperatures for cases calculated, the solution was verified by using the known
where the pinching occurred in the middle of the heat inlet and outlet temperatures on the low temperature
exchanger. side of the heat exchanger. For each discrete tem-
For the more general case where there were two con- perature on the low pressure side, a temperature was
trol volumes, the pressures at the pinch point (pp,Cooled calculated on the high pressure side using an energy
and pp,Heated ) were defined as balance of a simple control volume, using previously
calculated temperatures on the low temperature side
pp,Cooled = pi,Cooled − (Ti,Cooled − Tp ) ∗ d (10) of the control volume. The resulting temperatures on
the high and low pressure side were then compared to
and
check for any negative temperature differences, which
pp,Heated = pi,Heated − (Tp − Ti,Heated ) ∗ d (11) would indicate a failure of the logic described above to
identify the correct heat exchanger pinch point. The
where pi,Cooled and Ti,Cooled are the inlet pressure on newly computed high temperature outlet and inlet tem-
the cooled side and pi,Heated and Ti,Heated are the inlet peratures were also compared to those of the original
pressure on the heated side. With a known temper- one or two control volume technique. For most sce-
ature and pressure at the pinch point, the enthalpy narios, the logic proved to be successful. Because of
on both the heated (hp,Heated ) and cooled (hp,Cooled ) some errors, which are believed to be due to the fluid
side could be found using the fluid property functions. property database as well as the current discretization
The enthalpy at the inlet on the heated (hi,Heated ) and technique, an error of up to 1.3K was accepted. If the
cooled (hi,Cooled ) sides of the heat exchanger could be negative temperature difference was more than 1.3K,
found using the fluid property functions. The energy a trial and error method was utilized in an attempt
balance for the two control volumes described above to identify the correct pinch point location, and the

10
above process was repeated until an error of less than exchanger. Two specific heats are presented for each
1.3K was achieved. If an error less than 1.3 K was fluid stream. cp,Heated and cp,Cooled are based upon the
not achieved, the heat exchanger function returned an fluid stream’s mass flow, and CHeated and CCooled are
error. based upon mass flow of the cooled fluid stream and
As was mentioned above, the pressure drops in are defined in Equations 16 and 17.
the heat exchangers were defined to be a function of
the temperature changes in the heat exchangers. The ṁHeated
CHeated = cp,Heated ∗ (16)
technique that was just described for finding the heat ṁCooled
exchanger outlet temperatures assumed that the out-
let pressures were known. Because the temperature ṁCooled
changes in the heat exchanger were initially unknown, CCooled = cp,Cooled ∗ = cp,Cooled (17)
ṁCooled
an iterative process was required in order to determine
appropriate pressure drops. An initial guess for the CHeated and CCooled are presented so that heat capac-
heat exchanger pressure drops was based upon the low ity can be compared for heat exchangers with different
temperature inlet/outlet temperatures being equal and mass flow rates on the heated and cooled side, which
the high temperature inlet/outlet temperatures being can happen in cycles with recompression. It is impor-
equal. Temperature changes within the heat exchang- tant to note that cp,Cooled and CCooled have the same
ers were found using the guessed pressure drops based values and as a result, the curves are overlapping. One
upon the initially guessed temperature change. With can see a highly non-linear and dissimilar cp for both
new guessed pressure drops based on calculated heat the heated and cooled sides of the heat exchanger. In
exchanger temperature changes, the heat exchanger the example configuration shown, C is still dissimi-
temperature changes were evaluated again using the lar between the heated and cooled sides of the heat
process described above. The entire process was re- exchanger.
peated until the temperature changes and pressure An important metric for evaluating heat exchanger
drops in the heat exchangers stopped changing. If the performance is the heat exchanger’s effectiveness (),
change in inlet/outlet pressures from iteration to itera- which is defined as
tion was less than 0.0001%, the solution was considered
(ho,Heated − hi,Heated ) ∗ ṁHeated
to be converged. If a change in inlet/outlet pressure of =
min((ho,max,Heated − hi,Heated ) ∗ ṁHeated ,
less than 0.0001% could not be achieved, the conver-
gence criteria was relaxed to a change in inlet/outlet (hi,Cooled − ho,min,Cooled ) ∗ ṁCooled )
pressure of 0.3%. If this relaxation was insufficient, (hi,Cooled − ho,Cooled ) ∗ ṁCooled
= (18)
the heat exchanger function raised an exception. min((ho,max,Heated − hi,Heated ) ∗ ṁHeated ,
It is important to make it clear that no geometry (hi,Cooled − ho,min,Cooled ) ∗ ṁCooled )
is assumed in this heat exchanger solver and that the
purpose of the solver is to aid cycle and heat exchanger where ho,max,Heated is the maximum enthalpy the heated
designers in understanding the thermodynamic limit of side fluid could have if the heated side outlet could
heat exchanger performance due to the variable specific reach the temperature of the cooled side inlet (Ti,Cooled )
heat mismatch. Because no geometry is assumed, no and ho,min,Cooled is the minimum enthalpy the cooled
length dimension is used when plotting the results, but side fluid could have if the cooled side outlet could reach
rather, results are presented as a function of the tem- the temperature of the heated side inlet (Ti,Heated ).
perature of the cooled fluid stream. The temperature of Heat exchanger effectiveness helps one assess how much
the cooled fluid stream is related to the length dimen- of the heat that can be transferred is transferred. For
sion, but no specific relationship is presently assumed. the present study, the heat exchanger effectiveness was
This technique also has not yet been adapted to handle assumed to be 100% based on the assumption of very
heat exchangers where a fluid is boiling or condensing. high convection and all heat that can be transferred
A separate web based interface was also created for is assumed to be transferred. An important metric to
this heat exchanger solver, which plots specific heats, also consider is the maximum amount of heat that can
temperature, and temperature difference for the heated transferred compared to the maximum one may want
and cooled fluid streams in a heat exchanger [19]. to transfer from either the heated or cooled fluids. In
Figure 6 shows an example plot created using the the case of supercritical carbon dioxide power cycle
web based interface. The left plot shows the specific design, one always wants to transfer as much heat as
heats of both the heated and cooled sides of the heat possible, but because of the specific heat mismatch,

11
 Cooled Side Inlet: Temperature=450.0K, Pressure=8.5MPa, Mass Fraction=1.00
Heated Side Inlet: Temperature=305.0K, Pressure=18.5MPa, Mass Fraction=0.60
Pressure Drop=5000 Pa/K, Inlet Pressure Ratio=2.2, φ=0.60
4000 20 2.0
cp,Cooled ∆T
cp,Heated 0
3500 CCooled CHeated/CCooled
CHeated 1
15
3000 1.5

Heat Capacity Ratio, CHeated/CCooled

cp, [J/(kg*K)] and C, [J/(kgCooled*K)]

2500

∆T = TCooled−THeated, [K]
10
2000 1.0

1500
5
1000 0.5

500
0

0
300 320 340 360 380 400 420 440 300 320 340 360 380 400 420 440 0.0
Temperature, Cooled Side, [K] Temperature, Cooled Side, [K]

Figure 6: Sample Heat Exchanger Solution

Cooled Side Inlet: Temperature=350.0K, Pressure=1.0MPa, Mass Fraction=1.00

Heated Side Inlet: Temperature=305.0K, Pressure=1.0MPa, Mass Fraction=1.00
Pressure Drop=0 Pa/K, Inlet Pressure Ratio=1.0, φ=1.00
1000 10 2.0
cp,Cooled ∆T
cp,Heated 0
CCooled CHeated/CCooled
8 1
800 CHeated
1.5
Heat Capacity Ratio, CHeated/CCooled
cp, [J/(kg*K)] and C, [J/(kgCooled*K)]

6
∆T = TCooled−THeated, [K]

600

4 1.0

400
2
0.5
200
0

0
300 310 320 330 340 350 2
300 310 320 330 340 3500.0
Temperature, Cooled Side, [K] Temperature, Cooled Side, [K]

Figure 7: Sample heat exchanger solution with nearly constant and similar specific heats

12
 Cooled Side Inlet: Temperature=450.0K, Pressure=1.0MPa, Mass Fraction=1.00
Heated Side Inlet: Temperature=305.0K, Pressure=1.0MPa, Mass Fraction=0.60
Pressure Drop=5000 Pa/K, Inlet Pressure Ratio=1.0, φ=0.63
1000 60 1.0
cp,Cooled ∆T
cp,Heated 0
CCooled CHeated/CCooled
50 1
800 CHeated 0.8

Heat Capacity Ratio, CHeated/CCooled

cp, [J/(kg*K)] and C, [J/(kgCooled*K)]

40

∆T = TCooled−THeated, [K]
600 0.6
30

400 0.4
20

200 10 0.2

0
0
300 320 340 360 380 400 420 440 300 320 340 360 380 400 420 440 0.0
Temperature, Cooled Side, [K] Temperature, Cooled Side, [K]

Figure 8: Sample heat exchanger solution with nearly constant but dissimilar heat capacities

Cooled Side Inlet: Temperature=350.0K, Pressure=25.0MPa, Mass Fraction=1.00

Heated Side Inlet: Temperature=310.0K, Pressure=5.0MPa, Mass Fraction=1.65
Pressure Drop=0 Pa/K, Inlet Pressure Ratio=0.2, φ=1.00
3000 10 2.0
cp,Cooled ∆T
cp,Heated 0
CCooled CHeated/CCooled
2500 8 1
CHeated
1.5
Heat Capacity Ratio, CHeated/CCooled
cp, [J/(kg*K)] and C, [J/(kgCooled*K)]

2000 6
∆T = TCooled−THeated, [K]

1500 4 1.0

1000 2
0.5
500 0

0
310 315 320 325 330 335 340 345 350 2
310 315 320 325 330 335 340 345 3500.0
Temperature, Cooled Side, [K] Temperature, Cooled Side, [K]

Figure 9: Sample heat exchanger solution where the heat exchanger is pinched at both ends

13
 Cooled Side Inlet: Temperature=450.0K, Pressure=5.0MPa, Mass Fraction=1.00
Heated Side Inlet: Temperature=305.0K, Pressure=25.0MPa, Mass Fraction=0.56
Pressure Drop=0 Pa/K, Inlet Pressure Ratio=5.0, φ=0.96
3000 10 2.0
cp,Cooled ∆T
cp,Heated 0
CCooled CHeated/CCooled
2500 8 1
CHeated
1.5

Heat Capacity Ratio, CHeated/CCooled

cp, [J/(kg*K)] and C, [J/(kgCooled*K)]

2000 6

∆T = TCooled−THeated, [K]
1500 4 1.0

1000 2
0.5
500 0

0
300 320 340 360 380 400 420 440 2
300 320 340 360 380 400 420 440 0.0
Temperature, Cooled Side, [K] Temperature, Cooled Side, [K]

Figure 10: Sample heat exchanger solution where the heat exchanger is pinched in the middle and at the high
temperature end

Cooled Side Inlet: Temperature=450.0K, Pressure=5.0MPa, Mass Fraction=1.00

Heated Side Inlet: Temperature=305.0K, Pressure=25.0MPa, Mass Fraction=1.00
Pressure Drop=5000 Pa/K, Inlet Pressure Ratio=5.0, φ=0.55
3000 80 3.0
cp,Cooled ∆T
cp,Heated 0
70 CHeated/CCooled
2500
CCooled 2.5
CHeated 1
60
Heat Capacity Ratio, CHeated/CCooled
cp, [J/(kg*K)] and C, [J/(kgCooled*K)]

2000 2.0
50
∆T = TCooled−THeated, [K]

1500 40 1.5

30
1000 1.0
20

500 10 0.5

0
0
300 320 340 360 380 400 420 440 300 320 340 360 380 400 420 440 0.0
Temperature, Cooled Side, [K] Temperature, Cooled Side, [K]

Figure 11: Sample heat exchanger solution where the heat exchanger is pinched only at the low temperature end

14
there isn’t always enough heat available to be trans- maximum amount of heat that can be transferred is.
ferred on the cooled side and there isn’t always enough Figure 7 is an example well below the critical pressure,
heat that can be accepted on the heated side. No term so it is not applicable to supercritical carbon dioxide
is known to identify this metric, so a new name will be power cycles, however, it is an important case in under-
given as the ‘‘Fraction of Desired Heat Transferred’’ standing the impact of specific heats on heat exchanger
and will be associated with the symbol φ and defined performance.
as Figure 8 shows an example heat exchanger solution
(ho,Heated − hi,Heated ) ∗ ṁHeated where the specific heats are nearly constant (the same
φ= pressure as the case in Figure 7), but the mass fractions
max((ho,max,Heated − hi,Heated ) ∗ ṁHeated ,
are different so the heat capacities between the heated
(hi,Cooled − ho,min,Cooled ) ∗ ṁCooled )
and cooled sides are different. In this case the heat
(hi,Cooled − ho,Cooled ) ∗ ṁCooled
= (19) capacity on the heated side is always lower than the
max((ho,max,Heated − hi,Heated ) ∗ ṁHeated , cooled side. Within the operating temperature range
(hi,Cooled − ho,min,Cooled ) ∗ ṁCooled ) defined by the heat exchanger heated and cooled side
This new metric can be used to assess how well the inlets, there will not be enough heat capacity on the
specific heats are matched between the heated and the heated side in order to accept all the heat from the
cooled sides. In Figure 6 the ‘‘Fraction of Desired Heat cooled side and the cooled side outlet will never reach
Transferred’’ is 0.60, which means that 60% of the heat the inlet temperature of the heated side (the ‘‘Fraction
that one would like to accept on the heated side could of Desired Heat Transferred’’ is 0.63). The temperature
be transferred to the heated side, or 60% of the heat difference varies linearly within the heat exchanger and
that one would like to have available on the cooled the location of minimum temperature difference (the
side to accept on the heated side could be transferred pinch point) is at the high temperature end only.
to the heated side. Figure 9 shows an example heat exchanger solution
In the right plot of Figure 6, the temperature differ- where the location of minimum temperature difference
ence (∆T ) and ratio of specific heats (CHeated /CCooled ) (pinch point) occurs at both ends. In this example the
is presented. Two reference horizontal lines are pre- specific heats are dissimilar and varying. This particu-
sented for the temperature difference and ratio of spe- lar case is also not applicable to supercritical carbon
cific heats at 0 and 1 respectively. The reference line dioxide power cycles because the heated side is a lower
at 0 is presented along with a slightly negative scale in pressure than the cooled side, however, it illustrates
order to assess the potential error in the solution that a vary interesting solution which helps to understand
is described above. The reference line for a specific heat exchangers with nonlinear and dissimilar specific
heat ratio of 1 is presented in order to illustrate clearly heats. In this case the mass fraction was adjusted
how the heated and cooled sides relative specific heat is so that the average heat capacities of the heated and
changing and which fluid stream has a greater specific cooled sides are well matched. The heat capacity ratio
heat. The relative specific heats are wildly changing in within the heat exchanger changes from greater than
this particular solution and the temperature difference one at the low temperature end of the heat exchanger
as a result also changes dramatically within the heat to less than one at the high temperature end of the
exchanger. It’s also important to note in this particu- heat exchanger. The temperature difference within the
lar example that the temperature difference changes heat exchanger is still fairly low away from the ends.
concavity within the heat exchanger, and the tempera- In Figure 10 a heat exchanger solution is shown
ture difference is close to zero at one local minima and where a pinch point exists at both the high temper-
approximately zero at the absolute minima. ature end and in the middle of the heat exchanger.
In addition to the example solution shown in Fig- The temperatures and pressures are representative of
ure 6, there are several other example solutions that a heat exchanger operating in a supercritical carbon
demonstrate important heat exchanger characteristics. dioxide power cycle. The specific heat capacities are
Figure 7 shows a heat exchanger with nearly constant nonlinear and dissimilar. The mass fraction has been
and similar specific heats with similar mass flow rates adjusted so that the average heat capacities are nearly
on both the heated and cooled sides. In this example the same, which causes the solution to have low tem-
the temperature difference between the heated and perature differences throughout the heat exchanger.
cooled sides is constant and minimum throughout the This example points out that the pinch point is not
heat exchanger (it is pinched everywhere). The spe- necessarily the problem with heat exchangers in su-
cific heats of both sides are very well matched and the percritical carbon dioxide cycles. The heat exchanger

15
 has two pinch points but still has a low temperature exchangers.
difference everywhere so nearly all of the heat from
the cooled side is transferred to the heated side (the 3.4 Cycle Simulation
‘‘Fraction of Desired Heat Transferred’’ is 0.96). Pinch
points may be unusual in heat exchangers with highly Inputs of the cycle simulation include maximum tem-
nonlinear and dissimilar specific heats, but they are perature, minimum temperature, compressor pressure
not the problem, just a distinct feature. The location ratios, turbomachinery component efficiencies, heat ex-
of the pinch point and the temperature away from the changer pressure drop, main compressor inlet pressure,
pinch point may help to indicate how good or bad the and mass fraction for flow splits. An iterative proce-
heat capacity match is between the heated and cooled dure was utilized to find the unknown pressure drop
sides. between states 15 and 1 (the Main Fraction Cooler),
A case where the heated side specific heat capacity the ReCompressor Pressure Ratio, and the Power Tur-
(and heat capacity) is always higher on the heated side bine Pressure ratio. Although an input variable, the
than the cooled side is shown in Figure 11. In this case linear pressure drop vs temperature drop constant was
the heat exchanger relative heat capacities is always fixed to the same value for all heat exchangers in the
changing within the heat exchanger, but because the cycle. Individual heat exchangers currently are not
heat capacity is always higher on the heated side, the able to have different pressure drop constants. The
pinch point is only at the low temperature end and cycle code presently does not include the ability to
the temperature difference is very high at the high handle mixing of fluids of different temperatures where
temperature end of the heat exchanger. Because the the flow is recombined. In order to ensure tempera-
heat capacity on the cooled side is lower, the heated tures were equal when the flow was recombined, some
side will never reach the inlet temperature of the cooled small heaters and coolers were added if necessary on
side (the ‘‘Fraction of Desired Heat Transferred’’ is the high and low pressure sides at the inlet or outlet
0.55). This case is an example where the single pinch of the the heat exchanger. It is also important to note
point at the low temperature end clearly indicates that only on-design conditions were studied. The code
that the heat capacity is much higher on the heated is currently not able to perform transient simulations
side than the cooled side. It’s important to reiterate or off-design studies at this time.
however that the pinch point itself is not the problem,
the heat capacity mismatch is the problem. 3.5 Design Exploration
All of these example heat exchanger solutions demon- A design explorer was developed to run the Supercrit-
strate some distinct characteristics of heat exchangers ical Carbon Dioxide Power Cycle analysis code with
with constant, varying, similar, and dissimilar heat many different input combinations. The design ex-
capacities. The heat exchanger performance plays a plorer was developed in such a way that the design
very important part in the performance of a supercrit- space could be explored in parallel by running the Su-
ical carbon dioxide power cycle. It’s also important percritical Carbon Dioxide Power Cycle analysis code
to note that real heat exchangers will actually have on multiple processors simultaneously. The code can
some non-zero minimum temperature difference and use as many processors as are available on a single
this minimum amount is because the heat exchanger machine. Some effort was conducted to run differ-
can’t have an unlimited length and too high of a pres- ent batches of permutations on different machines,
sure drop. The present results are still useful because however, this functionality was abandoned because of
they provide a rapid solution and clearly isolate the the increased complexity of using multiple machines,
required temperature difference due to specific heat and the dramatic increase in the number of processors
mismatch. If the heat exchanger designer can identify available on a single machine recently. A 48 processor
regions where a temperature difference is required due machine was used for the present study.
to specific heat match, then forced convection coeffi- In order to effectively eliminate some components
cient does not necessarily have to be as high in those such as a low temperature recuperator, the precom-
regions of the heat exchanger and the heat exchanger pressor and/or the recompressor from the system, the
designer can harness that temperature difference rather ‘‘Low Temperature Recuperator Main Fraction High
than requiring high convection (which typically results Pressure Component Mass Fraction’’ could be set to
in a higher pressure drop). This approach may re- 0 or 1, the precompressor pressure ratio could be set
sult in more complex heat exchanger geometry, but to a pressure ratio of 1 and the recompression fraction
could result in lower entropy production in the heat

16
could be set to 0 or 1. laws of thermodynamics were not being violated in
It’s important to note that although the results of any solution. If any laws of thermodynamics were
the design exploration process can be used to explore violated or problems achieving a solution, an exception
optimal system layouts, the approach was not used just was raised, the permutation would be aborted, and
for the purpose of identifying a single optimal solution. the efficiency was set to -1 for that permutation as
Although the methodology utilized is fairly in depth, a way of marking it as an unsolvable permutation.
the level of complexity of assessing a true optimal solu- Contour plots have a minimum efficiency of 0.30 (30%)
tion is significant. With such a new type of technology, presented, and everything less than 0.30 is clipped
economic, control, and safety concerns for example are (appears white instead of blue). Any efficiency less
difficult to formally and accurately formulate. By ex- than 0.30 is essentially ignored so that these unsolvable
ploring the design space completely, rather than using permutations are not considered. For Dataset I, 3.7%
a more efficient adaptive optimization algorithm, engi- of the permutations were unsolvable, and for Dataset
neering judgment can be more appropriately applied to II, 5.2% of the permutations were unsolvable.
the results through the use of easy to use visualization Selected results of the design exploration can be
tools. viewed in Figures 13-19 and the entire dataset can
be manipulated and viewed using the web based user
interface that was developed as part of the Supercritical
4 Results Carbon Dioxide Power Cycle analysis code[19]. A
Two separate runs were conducted in order to explore screenshot of this web based user interface can be seen
a broad range of each parameter as well as a refinement in Figure 12. All Figures in this section, as well as
of key parameters. Tables 2 and 3 provide a summary Figures 2, 3, 5, and 6 were created using the web
of the range and resolution of the parameters studied based user interface, and utilized the high resolution
in both runs, as well as the value plotted. Parameters PDF figure download feature. Contour plots in this
where the value plotted is indicated to be ‘‘Optimal’’ section include dots overlayed over the contour. These
mean that the optimal value is plotted based on the dots indicate the data points and all contour values in
plotting constraints of the other parameters (either between the dots have been interpolated.
fixed or optimal). For plots where a variable is on
an axis, the value plotted parameter is not relevant
because all values are plotted. Parameters where ‘‘Op-
timal’’ is indicated use the optimal value based on the
each value of the parameter on the axes. The first
run surveyed 13 parameters and the total number of
permutations was 20,155,392. This dataset is called
‘‘Dataset I’’. Due to the large number of parameters
explored, the range and resolution of each parameter
was limited, and a second exploration was conducted.
The second exploration focused just on 5 parameters,
PreCompressor Pressure Ratio, Main Compressor Pres-
sure Ratio, Recompression Fraction, Low Temperature
Recuperator Main Fraction High Pressure Component
Mass Fraction, and Main Compressor Inlet Pressure,
which resulted in a total of 1,800,000 permutations.
The remaining parameters were fixed to values which
are assumed to be reasonably attainable component
Figure 12: Web Based Graphical User Interface[19]
efficiencies. This dataset is called ‘‘Dataset II’’.
Extensive efforts were taken to ensure that the heat
Figures 13-14d were generated using ‘‘Dataset I’’.
exchanger and main cycle portions of the code were
Figures 13-14b show the sensitivity of the cycle ef-
written as general as possible, however, with a very
ficiency on the minimum and maximum cycle tem-
large design space being explored, it was difficult to
peratures. Within the limited range and resolution
ensure every permutation could be solved successfully.
studied (three high and three low temperatures), the
Many checks were incorporated into the code to ensure
dependency of the cycle efficiency on the minimum

17
 Parameter Minimum Maximum Number of Values Value Plotted
PreCompressor Pressure Ratio 1.0 4.0 6 Optimal
Main Compressor Pressure Ratio 1.1 4.1 6 Optimal
Recompression Fraction 0.000 0.991 4 Optimal
Low Temperature Recuperator Main Fraction 0.001 0.991 4 Optimal
High Pressure Component Mass Fraction
Main Compressor Inlet Pressure 6 MPa 10 MPa 6 Optimal
Maximum Temperature 798K 923K 3 923K
Minimum Temperature 320K 333K 3 320K
Main Compressor Isentropic Efficiency 0.75 1.00 4 0.85
PreCompressor Isentropic Efficiency 0.80 0.95 3 0.875
ReCompressor Isentropic Efficiency 0.80 0.95 3 0.875
Power Turbine Isentropic Efficiency 0.89 0.93 3 0.93
Main/Re/Pre Compressor Turbine Isentropic 0.84 0.89 3 0.89
Efficiency
Heat Exchanger Pressure Drop 500 Pa/K 0 Pa/K 2 500 Pa/K

Table 2: Dataset I - 20,155,392 permutations - All Parameters - Coarse Exploration

Parameter Minimum Maximum Number of Values Value Plotted

PreCompressor Pressure Ratio 1.0 4.0 20 Optimal
Main Compressor Pressure Ratio 1.1 4.1 20 Optimal
Recompression Fraction 0.000 0.991 15 Optimal
Low Temperature Recuperator Main Fraction 0.001 0.991 15 Optimal
High Pressure Component Mass Fraction
Main Compressor Inlet Pressure 6 MPa 10 MPa 20 Optimal
Maximum Temperature 923K 923K 1 923K
Minimum Temperature 320K 320K 1 320K
Main Compressor Isentropic Efficiency 0.85 0.85 1 0.85
PreCompressor Isentropic Efficiency 0.875 0.875 1 0.875
ReCompressor Isentropic Efficiency 0.875 0.875 1 0.875
Power Turbine Isentropic Efficiency 0.93 0.93 1 0.93
Main/Re/Pre Compressor Turbine Isentropic 0.89 0.89 1 0.89
Efficiency
Heat Exchanger Pressure Drop 500 Pa/K 500 Pa/K 1 500 Pa/K

Table 3: Dataset II - 1,800,000 permutations - Fixed Component Efficiencies and Max/Min Temp, Other
Parameters Refined

18
 viewed using the web based interface for ‘‘Dataset I’’,
Maximum Efficiency=50.57%
0.60 which has a lower resolution for the precompressor
332 0.57 and main compressor pressure ratios, also shows two
0.54 local maxima, although at different locations due to
330
0.51 plotting artifacts and interpolation errors in regions
Minimum Temperature [K]

328 0.48 with high gradients. The main take aways from Figure

Cycle Efficiency
0.45 15 are that eliminating one of the compressors (i.e. the
326
0.42 condition where pressure ratio is 1) will result in a
324 0.39
fairly significant reduction in overall cycle efficiency,
and that when both compressors are used, there is a
0.36
322 wide range of combinations of precompressor and main
0.33
compressor pressure ratios.
320800 820 840 860 880 900 920 0.30
Figure 16 shows the optimal recompression fraction
Maximum Temperature [K]
at each precompressor and main compressor pressure
Figure 13: Cycle Efficiency vs Maximum and Minimum ratio. Comparing to Figure 15, the variation of the
Temperature optimal recompression fraction, near the region of high
cycle efficiency, appears to be significant. Because the
cycle code in its current form does not allow for the
and maximum temperatures appears to be fairly linear. precompressor and main compressor inlet temperatures
Figure 14c shows the cycle efficiency’s sensitivity to to be different, as the precompressor pressure ratio or
the main compressor’s isentropic efficiency. There is main compressor pressure ratio approach 1, the recom-
some non-linearity evident using the 4 different main pression fraction has no meaning. Therefor values of
compressor isentropic efficiencies that were considered. recompression fraction at pressure ratios of 1 should
A larger range and resolution may be preferred in order be ignored (for example, the red vertical region on the
to better assess the importance of the main compres- left portion of Figure 16).
sor’s efficiency because it may be desired to use a Figure 17 shows the cycle efficiency vs recompres-
positive displacement device for some demonstration sion fraction. For much of the design space, the optimal
efforts since the main compressor is expected to be recompression fraction is between approximately 0.40
the most difficult turbomachinery design to perfect. and 0.80. It is important to note when comparing
The cycle efficiency vs the recompressor isentropic effi- Figure 17 to Figures 15 and 16 that the same pres-
ciency is shown in Figure 14d. With the limited range sure ratios could be used for multiple recompression
and resolution shown, the cycle efficiency appears to fractions in Figure 17.
be fairly linear with respect to recompressor isentropic Figure 18 shows the variation in optimal Low Tem-
efficiency. perature Recuperator Main Fraction High Pressure
Figures 15-19 were generated using ‘‘Dataset II’’. Component Mass Fraction vs precompressor and main
The parameters that were not fixed were explored at a compressor pressure ratios. For much of the design
fairly high resolution. Figure 15 shows cycle efficiency space, the elimination of one of the low temperature
vs main compressor and precompressor pressure ratios. recuperator would be possible. This is particularly
The maximum cycle efficiency, using the constraints true because the regions of pressure ratios where a
listed in Table 3 in the column ‘‘Value Plotted’’ is second low temperature recuperator would be useful
51.94%. A Carnot cycle with a maximum temperature are pressure ratios where the cycle efficiency is already
of 923K (650◦ C) and a minimum temperature of 320K lower anyway.
(47◦ C) would have an efficiency of 65.33%. This indi- Figure 19 shows that the cycle efficiency increases
cates that the proposed cycle layout may be able to as the main compressor inlet pressure increases, and
achieve efficiencies of 79.5% of a Carnot cycle efficiency. that the design exploration should be expanded to
As is shown in Table 3, there are 20 precompressor higher main compressor inlet pressures in order to
pressure ratios and 20 main compressor pressure ratios locate the optimal main compressor inlet pressure. It
considered. Observing the results from Figure 15 indi- is important to note, however, that the present study
cates an even higher exploration range should actually does not consider the increased design complexities of
be used because there are some noticeable plotting arti- higher system operating pressures. A more rigorous
facts in regions with high gradients. At least two local optimization coupled with costs may locate a different
maxima are evident in Figure 15. The same plot, when optimal main compressor inlet pressure.

19
 Maximum Efficiency=50.57% Maximum Efficiency=50.57%
0.51 0.510

0.50
0.505
0.49

0.500
Cycle Efficiency

Cycle Efficiency
0.48

0.47
0.495

0.46
0.490
0.45

0.44780 800 820 840 860 880 900 920 940 0.485320 322 324 326 328 330 332 334
Maximum Temperature [K] Minimum Temperature [K]

(a) Cycle Efficiency vs Max Temperature (b) Cycle Efficiency vs Min Temperature

Maximum Efficiency=51.34% Maximum Efficiency=51.18%

0.514 0.512

0.512 0.510

0.510 0.508
Cycle Efficiency

Cycle Efficiency
0.508 0.506

0.506 0.504

0.504 0.502

0.502 0.500

0.500
0.75 0.80 0.85 0.90 0.95 1.00 1.05 0.498
0.80 0.82 0.84 0.86 0.88 0.90 0.92 0.94 0.96
Main Compressor Isentropic Efficiency ReCompressor Isentropic Efficiency

(c) Cycle Efficiency vs Main Compressor Efficiency (d) Cycle Efficiency vs ReCompressor Efficiency

Figure 14

1.0
Maximum Efficiency=51.94%
0.60 4.0
4.0 0.9
0.57
3.5 0.8
3.5 0.54
Main Compressor Pressure Ratio

Recompression Fraction at Maximum Efficiency

Main Compressor Pressure Ratio

0.7
0.51 3.0
3.0 0.6
0.48
Cycle Efficiency

0.5
0.45 2.5
2.5 0.4
0.42
2.0 0.3
2.0 0.39
0.2
0.36 1.5
1.5 0.1
0.33
1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.0
1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.30 PreCompressor Pressure Ratio
PreCompressor Pressure Ratio

Figure 16: ReCompression Fraction at Optimal Cycle

Figure 15: Cycle Efficiency vs PreCompressor and
Efficiency vs PreCompressor and Main Compressor
Main Compressor Pressure Ratios
Pressure Ratios

20
 Maximum Efficiency=51.94%
5 Conclusions and Future Work
0.52
Supercritical CO2 Power Cycles have the potential
0.51 for improvements in efficiency over traditional power
cycles or traditional combined cycle power plants, at
0.50
a dramatically reduced system size. However, the
Cycle Efficiency

highly nonlinear fluid properties present significant

challenges in cycle and component design. The present
0.49
work indicates that a cycle efficiency of 51.94% may be
achievable with a maximum heat source temperature
0.48 of 923K (650◦ C) and a minimum coolant temperature
of 320K (47◦ C). At 79.5% of the efficiency of a Carnot
0.470.0 0.2 0.4 0.6 0.8 1.0 cycle, this cycle is approaching the thermodynamic
Recompression Fraction
limits of a heat engine.
A new system layout has been presented which
Figure 17: Cycle Efficiency vs ReCompression Fraction may help to eliminate some of the design challenges
Low Temperature Recuperator Main Fraction High Pressure Component Mass Fraction at Maximum Efficiency with supercritical carbon dioxide engines. The present
0.90 work demonstrates a cycle analysis code that was writ-
4.0
0.81 ten from scratch specifically to explore a wide range
3.5 0.72
of design parameters for the proposed layout. The
Main Compressor Pressure Ratio

web based visualization tool provides a level of insight

0.63
3.0 and interactivity not possible in a traditional static
0.54
document. This visualization and interactivity helps
0.45
2.5 to better direct the design, testing, and commercializa-
0.36 tion plan for supercritical carbon dioxide power cycles,
2.0 0.27 and better establish the benefits of systems of different
0.18 complexities. Although this work is fairly extensive,
1.5 there is still much to be done to improve the over-
0.09
0.00 all picture of the Supercritical Carbon Dioxide Power
1.0 1.5 2.0 2.5 3.0 3.5 4.0
PreCompressor Pressure Ratio Cycle design space.
Limitations of the present code include the lack
Figure 18: Low Temperature Recuperator Main Frac- of ability to model cycles below the critical tempera-
tion High Pressure Component Mass Fraction at Op- ture (condensing cycles) where a discontinuous change
timal Cycle Efficiency vs PreCompressor and Main in phase and two phase region may exist. The code
Compressor Pressure Ratios could also be improved to incorporate more realism
in the real fluid heat exchangers by accounting for an
0.520 Maximum Efficiency=51.94% additional temperature gradient in the fluids’ thermal
boundary layers and the heat exchangers solid walls.
0.515 The very low minimum temperature difference in the
0.510 heat exchangers is expected to slightly over predict the
cycle efficiency and under predict the ideal recompres-
0.505
Cycle Efficiency

sion fraction. The pressure drop formulation in the

0.500 heat exchangers could also be improved and individ-
0.495 ual relationships applied to different heat exchangers.
The code is presently limited such that the precom-
0.490
pressor inlet temperature must be equal to the main
0.485 compressor inlet temperature, and further flexibility
0.480
could be incorporated to allow for this parameter to
0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00
Main Compressor Inlet Pressure [Pa] 1e7 be changed independently. The code could also al-
low for different turbomachinery efficiencies for the
Figure 19: Cycle Efficiency vs Main Compressor Inlet turbines that power the compressors and a selectable
Pressure mass flow split between these turbines that operate

21
in parallel. The design space explored mostly utilizes [4] G Angelino. Carbon dioxide condensation cycles
a uniformly spaced grid for each parameter. Based for power production. Journal of Engineering for
on the present results, a non uniformly spaced grid Gas Turbines and Power, 90(3):287--295, 1968.
could be harnessed and the design space explored in
more detail. The code could be expanded to allow [5] Ernest G Feher. The supercritical thermodynamic
for off-design, as well as transient simulations. Fluid power cycle. Energy conversion, 8(2):85--90, 1968.
specific heat, pressure drop and pumping power of the [6] V. Dostal. A Supercritical Carbon Dioxide Cycle
atmospheric pressure side of the heat exchangers that for Next Generation Nuclear Reactors. PhD thesis,
transfer heat to and from the engine is currently not Massachusetts Institute of Technology, Dept. of
rigorously addressed. The code could also be expanded Nuclear Engineering, 2004.
such that the coupled engine cycle efficiency as well
as utilization of the heat source are explored and opti- [7] Steven A. Wright, Tom M. Conboy, and Gary E.
mized. The present code currently assumes that the Rochau. Break-even Power Transients for two
heat source is completely utilized and no high temper- Simple Recuperated S-CO2 Brayton Cycle Test
ature heat is unused. Understanding the impact of the Configurations. SCO2 Power Cycle Symposium,
heat source on the optimal cycle design and the overall May 2011.
system efficiency is critical in applying such a cycle
with fossil fuels such as coal and natural gas, as well as [8] Steven A Wright, Ross F Radel, Milton E Vernon,
a replacement for a steam turbine in a combined cycle Gary E Rochau, and Paul S Pickard. Operation
power plant. An estimated system cost could also be and Analysis of a Supercritical CO2 Brayton Cycle.
computed and the cycle efficiency plotted vs estimated SANDIA REPORT, SAND2010-0171, Advanced
system cost. Finally, the cycle could also be improved Nuclear Concepts Department, Sandia National
to allow for different pure fluids as well as mixtures Laboratories, Albuquerque, New Mexico, Septem-
of fluids, as well as incorporate CoolProp[20], a free, ber 2010.
open source, modern fluid property database instead [9] Kenneth J Kimball, Kevin D Rahner, Joseph P
of REFPROP. In addition to improving the cycle code, Nehrbauer, and Eric M Clementoni. Supercriti-
the web based graphical user interface can be enhanced cal Carbon Dioxide Brayton Cycle Development
to allow for plotting dependent variables vs dependent Overview. In ASME Turbo Expo 2013: Turbine
variables, computing and plotting of derivatives, higher Technical Conference and Exposition. American
speed rendering of the thermodynamic state diagrams, Society of Mechanical Engineers, June 2013.
overlaying contour lines of multiple variables on the
same plot, multiple axis on the line plots, among other [10] Echogen Power Systems, LLC. Our Story.
possible features. http://www.echogen.com/about/our-story/
Accessed: February 22nd , 2014.

[11] Energy Department Announces New Concen-

trating Solar Power Technology Investments
to American Industry, Universities. http:
//energy.gov/articles/energy-department-
References announces-new-concentrating-solar-power-
technology-investments-american.
[1] National Institute of Standards and Standard Ref-
erence Data Program Technology. NIST Standard [12] SunShot CSP R&D 2012, June 12, 2012.
Reference Database 23: Reference Fluid Thermo- http://www1.eere.energy.gov/solar/pdfs/
dynamic and Transport Properties-REFPROP. csp_rd_awardees_2012.pdf.

[2] Sulzer. Verfahren zur Erzeugung von Arbeit aus [13] Concentrating Solar Power SunShot Research
Warme, 1948. Swiss Patent 269 599. and Development. http://energy.gov/
eere/sunshot/concentrating-solar-power-
[3] G Angelino. Perspectives for the liquid phase sunshot-research-and-development.
compression gas turbine. Journal of Engineering
for Gas Turbines and Power, 89(2):229--236, 1967. [14] Python. https://www.python.org/.

[15] NumPy. http://www.numpy.org/.

22
[16] SciPy. http://scipy.org/.

[17] matplotlib. http://matplotlib.org/.

[18] python-refprop. https://github.com/

AndySchroder/python-refprop.

[19] Andy Schroder. Supercritical Carbon Diox-

ide Power Cycles. http://andyschroder.com/
CO2Cycle/.

[20] Ian Bell. CoolProp: Fluid properties for the

masses. http://coolprop.sourceforge.net/.

23
View publication stats