Hybrid Solar Thermal Power Plants Jorge Augusto Maciel Milhomem Rodrigues Thesis to obtain the Master of Science Degree in Mechanical Engineering Supervisors: Prof. Aires José Pinto dos Santos Prof. Viriato Sérgio de Almeida Semião Examination Committee Chairperson: Prof. Edgar Caetano Fernandes Supervisor: Prof. Aires José Pinto dos Santos Member of the Committee: Prof. Mário Manuel Gonçalves da Costa June 2019
80
Embed
Hybrid Solar Thermal Power Plants · Hybrid Solar Thermal Power Plants Jorge Augusto Maciel Milhomem Rodrigues Thesis to obtain the Master of Science Degree in Mechanical Engineering
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Hybrid Solar Thermal Power Plants
Jorge Augusto Maciel Milhomem Rodrigues
Thesis to obtain the Master of Science Degree in
Mechanical Engineering
Supervisors: Prof. Aires José Pinto dos SantosProf. Viriato Sérgio de Almeida Semião
Examination Committee
Chairperson: Prof. Edgar Caetano FernandesSupervisor: Prof. Aires José Pinto dos Santos
Member of the Committee: Prof. Mário Manuel Gonçalves da Costa
June 2019
ii
To my family.
iii
iv
Acknowledgments
I want to express my very great appreciation to my supervisors, Professor Aires Santos and Professor
Viriato Semiao, for their support throughout the whole process of developing this work. This thesis would
not have been the same without their consistent openness to help and interest in the work. I feel very
fortunate to have had them as my supervisors.
I would also like to thank my friends Francisco Vieira, Joao Ferreira, Joao Formiga, Luıs Fialho, Nuno
Miguel Guerreiro, Pedro Santana, and Shu Kaiyue for their friendship. I cannot let unmentioned that I
am deeply grateful to Francisco for all the academic support through the years in IST, always having my
back in the different academic challenges that we went through.
To my parents and to my brothers, of course, nothing of this would be possible without you. Thank
you for everything.
v
vi
Resumo
A energia solar termica, especialmente quando obtida atraves da implementacao das tecnologias de con-
centracao de energia solar (CSP), representa uma forma de energia renovavel cada vez mais atrativa.
Um dos fatores determinantes no desenvolvimento destas tecnologias e a sua integracao com ciclos de
potencia de forma eficiente. De modo a que as tecnologias CSP se tornem mais populares e competitivas
no mercado de producao de energia eletrica, e importante desenvolver modelos capazes de apresentar
elevadas eficiencias.
O presente trabalho foca-se na analise termodinamica de tres ciclos combinados com regeneracao
no ciclo a gas aqui propostos, nos quais caracterısticas distintivas como a presenca de uma camara de
combustao em serie juntamente com um concentrador solar (CSP) e divisao de caudais apos a turbina
a gas sao estudadas. O calor e fornecido por concentracao de energia solar nos tres modelos, sendo que
em dois deles, uma camara de combustao serve de suporte para o efeito. As analises termodinamicas
sao efetuadas para diferentes fluidos, nomeadamente, CO2, ar, N2, He e H2 para o ciclo de Brayton, e
R-245fa, R-141b, Cyclohexane, n-Pentane e agua para o ciclo de Rankine, resultando num total de 25
combinacoes. As simulacoes sao executadas por um programa desenvolvido em MATLAB.
Os resultados obtidos mostram que a adicao de uma valvula divisora de caudais como unica variavel
entre dois modelos contribui para um melhor aproveitamento da entalpia do fluido apos a sua saıda da
turbina a gas, resultando, em ultima instancia, em eficiencias globais superiores.
Palavras-chave: Concentracao de energia solar, termodinamica, hibridizacao de ciclos de
potencia.
vii
viii
Abstract
Solar thermal energy, especially concentrated solar power (CSP), represents an increasingly attractive
renewable energy source. One of the key factors that determines the development of this technology is its
integration with efficient power cycles. Thus, as to make CSP an attractive alternative to the conventional
thermodynamic cycles present in most of today’s thermal power plants, it is of high relevance to develop
models that provide attractive global efficiencies and cost effective CO2 mitigation.
The present work focuses on the energetic study of three combined Brayton-Rankine cycles hereby
proposed, in which features as having a combustion chamber in series with a solar receiver, and a stream
splitter subsequent to the gas turbine were studied and compared. Heat is provided by concentrated
solar power, having two of the three models to be hybridised with solar and combustion energy. The
thermodynamic analyses are performed for selected working fluids including CO2, air, N2, He and H2 for
the topping cycles, and R-245fa, R-141B, Cyclohexane, n-Pentane, and Water for the bottoming cycle,
resulting in 25 fluid combinations. A MATLAB program was developed to perform the simulations.
The results show that the CO2 and R-245fa fluid pair provides the highest cycle efficiency curves. In
addition, stream splitting proves to be advantageous, compared to other layouts found in literature.
Keywords: Concentrated solar power, thermodynamics, hybrid power cycles.
5.16 Model 3: global cycle efficiency assuming T5 = 825 K, a regenerator heat transfer area
equal to 6 m2 and a mass flow ratio r = 0.5. . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.17 Brayton cycle’s regenerator effectiveness values for models 1, 2, and 3, assuming a heat
transfer area of 6 m2 and, for models 2 and 3, a mass flow ratio r = 0.5. . . . . . . . . . . 56
xvi
Nomenclature
Greek symbols
α Radiative absorptivity.
ε Radiative emissivity, effectiveness.
η Efficiency.
ρ Fluid density.
σ Stefan–Boltzmann constant.
ref Reference condition.
Roman symbols
A Surface of a control volume, heat transfer area.
C Solar concentration ratio, heat capacity.
cp Specific heat at constant pressure.
g Gravitational acceleration.
G0 Standard solar irradiation.
h Specific enthalpy.
m Mass flow.
Ma Mach number.
p Pressure.
pR Compressor pressure ratio.
pr Relative pressure.
Q Energy transfer by heat across the boundary of a given control volume.
R Specific gas constant.
s0 Specific entropy.
xvii
T Temperature.
U Heat transfer coefficien.
u Specific internal energy.
V Volume of a control volume, and, when vectorized, velocity of the fluid.
v Control volume velocity, specific volume
W Energy transfer by work across the boundary of a given control volume.
z Vertical coordinate of a given control volume.
Subscripts
b Bottoming cycle.
b, cycle Bottoming cycle.
b, pinch Pinch point at the bottoming cycle side of the heat recovery steam generator.
c Cold side of a heat exchanger..
c Combustion chamber.
e Out direction.
h Hot side of a heat exchanger.
i In direction.
p Pump.
r Ratio.
reg Regenerator.
sol Solar receiver.
t Turbine, topping cycle when appearing in mt.
t, cycle Topping cycle.
t, pinch Pinch point at the topping cycle side of the heat recovery steam generator.
Superscripts
˙ Time derivative.
→ Vector.
xviii
Glossary
CSP Concentrated Solar Power.
HRSG Heat Recovery Steam Generator.
xix
xx
Chapter 1
Introduction
1.1 Motivation
Climate change and global warming are major concerns in the present days. The increasing world
population, accompanied by the growth of energy demand makes the development of sustainable and
clean energy technologies of primary importance in order to reduce emissions of carbon dioxide and other
greenhouse gases. Furthermore, fossil fuels are not unlimited and not easily accessible for most countries,
making them to depend on foreign supplies.
The world still relies heavily on fossil fuels in the present days to meet its growing energy needs,
resulting in the continuous increase of carbon emissions. According to U.S. Energy Information Admin-
istration (EIA) [1], world energy consumption through 2040 will increase, on average, for all fuels other
than coal.
One energy source that is vastly abundant in the nature is the sun, namely, the solar radiation that
arrives to the earth surface. For instance, if 1% of the global deserts area were used to harness solar
energy, that would be sufficient to produce the entire annual primary energy consumption of humankind as
electric power, already taking into consideration the conversion factor to useful energy form of electricity
[2]. Of all the electricity generating systems, power plants are the predominant ones, by means of fluid
expansion in a steam turbine, in most of the cases.
The majority of world-wide heating sources for the purpose of generating electricity is still based on
fossil fuels.
1.2 State of the Art
The global demand for electrical energy is increasing with no signs of slowing down. Alongside with
the electricity consumption, comes the growing amount of pollutant gas emissions, resulting mostly from
the use of fossil resources in combustion processes of the actual working power plants. Thus, there is a
significant incentive to use renewable energy sources to produce electricity. Solar energy is being seen
more and more as a promising energy source to achieve better results in both the environmental and
1
financial aspects. A promising technology is concentrated solar power (CSP), which often offers the most
reliable and dispatchable power generation from sun irradiation [3].
1.2.1 CSP Concept
Concentrated solar power (also known as concentrating solar thermal energy, CST) technologies make use
of the entire solar irradiation spectrum by using mirrors or lenses to concentrate the sun’s rays, providing
a source of heat to generate electricity. Thus, CSP technologies are an alternative to fossil fuels or nuclear
reactions regarding energy production.
1.2.2 CSP Technologies
CSP technology is already implemented in different parts of the world, as shown in Fig. 1.1. Spain is the
world leader in regards to CSP, accounting for more than 2300 MW of power capacity. As also shown
in the map, almost half of the worldwide capacity correspond to under construction or in development
CSP projects. Such demonstrates that the technology is fresh and still increasing in development and
implementation, making it relevant.
Figure 1.1 CSP projects around the world [4]
There is currently a great variety of CSP technologies, but the four most developed and promising
ones [5], are overviewed in the following paragraphs.
Parabolic Trough Collectors
A Parabolic Trough Collector, PTC, is a single-axis linear-focus solar collector, composed of a parabolic-
trough-shaped concentrator that reflects the direct solar radiation onto a receiver or absorber tube located
in the focal line of the parabola. Fig. 1.2 shows an example of a PTC system.
The PTC technology is the most common and mature one among the others. According to Wang et
al. [6], 90% of the actual CSP plants use it. Parabolic troughs can deliver useful thermal energy up
to 400◦ C. However, research on new working fluids promise higher temperatures close to 500◦ C in the
2
midterm [7].
Steel structure
Parabolic trough refector
Absorber tube
l
Figure 1.2 A Parabolic Trough Collector [7].
Linear Fresnel Reflectors
A single-axis Linear Fresnel Reflector system, LFR, is composed of long row reflectors that together
reflect direct solar radiation to an elevated linear tower receiver running parallel to the reflector rotational
tracking axis. A schematic representation of this system can be seen in Fig. 1.3.
The LFR technology is less commercially mature than PTC. However, Linear Fresnel Reflector solar
systems have been considered as a cost effective choice for solar thermal power generations due the their
simple structure when compared to the troughs. On the other hand, LFR systems provide lower thermal
performances per aperture area compared to that of the PTC systems, creating a need in the market to
improve the optical efficiency of the LFR systems [8].
Figure 1.3 Schematic representation of a basic LFR system. Each individual reflector tracks the sun byturning about a horizontal axis normal to the page [9].
Solar Power Towers
Also know as Central Towers (CT), Solar Power Tower systems consist of three main components: he-
liostat field, solar collector, and power-block island. Direct solar radiation is reflected and concentrated
3
by a heliostat field - individual mirror assembly with a two axis solar tracking system - onto a receiver
placed at the top of a tower, as it can be visualized in Fig. 1.4. The direct solar radiation is concentrated
in the receiver reaching a high irradiation flux, which is then converted to thermal energy, heating the
working fluid [10].
Figure 1.4 An example of a SPT system in USA [11].
Parabolic Dish Collectors
The Parabolic Dish Concentrator, PDC, system consists of a base support, a concave dish frame, a re-
flecting sheet, a conversion unit, and a sun-tracking system, as illustrated in Fig. 1.5. The tracking
system is allowed to rotate around the horizontal and vertical axis, having the optical axis of the con-
centrator always pointing towards the sun. The direct solar radiation is then reflected and focused onto
the solar receiver. The receiver captures the high temperature thermal energy into a fluid that is either
the working fluid for a receiver-mounted engine cycle, or is used to transport the energy to ground-based
processes. In the case of a receiver-mounted engine (Stirling engine is often used), a directly coupled
generator converts mechanical energy into electricity.
PDC systems are the less implemented in the group of four technologies. According to Pelay et al.
[12], as of 2017, there were no plants in activity using the PDC technology. Nonetheless, it represented
7–9% of plants under construction and planned, making it a non-negligible technology in the future.
Furthermore, the highest conversion efficiency (>30%) among the actual CSP technologies is reached
with a parabolic dish-type point receiver using a Stirling engine [6]. PDC systems have high optical
efficiencies and low start-up losses leading to highly efficient solar energy engines. This gives parabolic
dishes potential to eventually become one of the least expensive forms of renewable energy.
4
Figure 1.5 Illustration of a single PDC system. The bowl shaped concentrator reflects the sun rays ontothe power conversion unit placed centred in the axis [13].
1.2.3 Solar Thermal Power Plants
In order to generate electricity, the concentrated irradiation is converted to heat, driving a heat engine
(in most cases, a steam turbine [14]) connected to a power cycle.
It is expected that this predominance of CSP implemented with steam turbines continue due to its
relatively low costs and high reliability [15]. However, steam power cycles have a major drawback that
is the large amount of water consumption, given, especially, by the need of having cooling water in the
condenser [14]. Since the most promising locations for solar power are located in arid or desert locations
[16], reducing water consumptions is an important aspect in future projects. Gas turbine power cycles
allow for higher working temperatures, resulting in higher global efficiencies. Water consumption is also
considerably lower [14].
Two configurations for the integration of a combustion chamber and a solar heater in a Brayton cycle
are presented by Nathan et al. [15] (Fig. 1.6). The authors of the mentioned work state that having
the solar heater placed before the compressor increases the required work for compression, decreasing
the global efficiency of the cycle. Spelling, in his PhD Thesis ([14]), presents two configurations for the
heat components - parallel and serial positioning - as shown in Fig. 1.7. As stated by the author, the
parallel scheme is a poor choice thermodynamically, since the inlet temperature to the turbine is below
the maximum value achievable in the combustion chamber. Also, the greater the degree of utilization
of solar energy in the parallel scheme, the lower the temperature entering the turbine. This makes the
serial hybridisation a better choice as it can provide a final inlet temperature to the turbine that is equal
to the combustor outlet temperature. Furthermore, the final temperature delivered to the turbine in the
serial scheme is independent of the degree of solar integration.
5
Figure 1.6 Basic options to hybridise a Brayton cycle - having the solar energy source placed before thecompressor (Qsol,LP ) or after (Qsol,HP ) [15].
(a) Parallel configuration
(b) In series configuration
Figure 1.7 Hybridisation schemes for a gas turbine [14].
Dunham and Lipinski [17] made a thermodynamic energy analysis of the global theoretical efficiencies
and power generation of a Brayton and combined Brayton-Rankine cycles with solar energy as the heat
input, considering a gas microturbine. The analysis was performed for multiple selected working fluids.
With a nominal solar heat input of 150 kW and a 50 kWe gas microturbine, the single Brayton cycle
reached a global efficiency of 15.31% and the combined cycle provided a 21.06% global efficiency.
Among the different literature investigated to pursue a study in CSP power plants, this was the
work that offered the most complete data information to further the authors’ study, and, ultimately, to
contribute to the scientific knowledge on this topic. In addition, the models proposed by the authors do
not contain features that present potential performance improvements and relevance to be studied, such
6
as a stream bifurcation or a combustion chamber. These will be present in the thermodynamic models
proposed in this thesis.
Since this work was the starting point of the present thesis, it will be addressed with more detail in
section 2.2.
1.3 Objectives
The present work consists in proposing three different models of solar heated power cycles, having Dunham
and Lipinski’s work [17] as a starting point. Those are three combined Brayton-Rankine cycles with
regeneration in the Brayton cycle. The first model contains a combustion chamber subsequent to the solar
receiver. The second contains the solar receiver as the only source of heat energy and a stream bifurcation
after the gas turbine to preserve the fluid enthalpy before it enters the Brayton cycle regenerator. The
last model presents both the distinctive feature of the formers - combustor chamber and a stream splitter.
The main objectives of the present work are to:
• Model thermodynamically the three cycles, studying and comparing in detail their performances;
• Study the influence that the selected working fluids have on the cycle performances and their
behaviour in the different configurations and conditions;
• Improve the global performance of the new designed cycles when compared to those developed by
Dunham and Lipinski [17].
1.4 Thesis Outline
This thesis is organized in six chapters. Chapter 1 includes an overview of the present topic in the
context of nowadays society needs. Two sections are dedicated to presenting the state of the art and the
objectives of the present work. Chapter 2 begins with a theoretical introduction, explaining the energy
rate equations that are employed in the implementation of the developed algorithm. A summary of the
article at study is presented subsequently. It is then described how the thermodynamic properties of the
fluids are obtained. The third chapter describes the mathematical modelling of the present work’s models.
In the fourth chapter, it is explained the restrictions that are implemented on the models implementation,
and the computer implementation of the methods explained in the preceding chapter. Chapter 5 presents
a discussion of the results obtained from all the proposed model simulations. Chapter 6, the last one,
gives an overview of the conclusions achieved from this thesis and outlines some relevant future work
points regarding the present work subject.
7
8
Chapter 2
Background
2.1 Theoretical Introduction
According to Boettener et al. [18], the energy rate balance equation for open systems, in a given control
volume, is given by
Q− W=d
dt
[∫∫∫ (u+
v2
2+ gz
)ρdV
]+
∫∫ (h+
v2
2+ gz
)ρ(~V · ~n)dA (2.1)
where Q and W are, respectively, the net rate of energy transfer by heat and work across the boundary of
the control volume. u is the specific internal energy, v2/2 is the specific kinetic energy, gz is the specific
potential energy, ρ is the fluid density, V is the volume of the control volume, h is the specific enthalpy,
(~V ·~n) is the dot product between the fluid velocity and the normal vector at a given point of the surface
area, A, of the control volume. A schematic representation of the control volume of an open system is
shown in Fig. 2.1.
Considering steady state, the time derivate is null,
d
dt
[∫∫∫ (u+
v2
2+ gz
)ρdV
]= 0
Dashed line definesthe control volume boundary
Inlet i
mi
meControlvolume
zezi
ui +Vi
2___2
+ gzi
ue +Ve
2___2
+ gze
Exit e
Q
W
Figure 2.1 Schematic representation of the control volume used to develop Eq. (2.1) [18].
9
Having kinetic and potential energy variation to be negligible, the following underlined terms are
neglected, ∫∫ (h+
v2
2+ gz
)ρ(~V · ~n)dA→
∫∫h ρ(~V · ~n)dA
Eq. (2.1) is then simplified to
Q− W =
∫∫h ρ(~V · ~n)dA (2.2)
The control volume can be assumed to have a series of one dimensional locations through which mass
enters, i, and exits, e, the system. In this way, the surface integral in Eq. (2.2) can be expressed as the
summation of outlet minus inlet rates of the transported variables,
∫∫h ρ(~V · ~n)dA =
∑e
me he −∑i
mi hi
At steady-state the mass flow is constant and assuming uniform enthalpy at the inlet and exit sections
one finally gets
Q− W = m
(∑e
he −∑i
hi
)(2.3)
Steam and Gas Turbine
Turbines are widely used in thermodynamic cycles to generate power. The power is generated as
superheated steam or gas enters the turbine and expands to lower pressures while passing through a set
of blades attached to a shaft free to rotate.
Selecting the control volume as that enclosing the turbine (Fig. 2.2), the kinetic and potential energy
variations between inlet and outlet are normally negligible.
cvW·
1
2CV
Figure 2.2 Schematic representation of the turbine control volume. Adapted from [18].
Considering that the turbine is adiabatic, i.e., no heat is exchanged with the surroundings (Q = 0),
the power generated is calculated from
Wt = m(h1 − h2)
where 1 and 2 stand, respectively, for the exit and inlet sections.
Compressor and Pump
Contrary to the turbines case, compressors and pumps exploit work provided by a driving motor in
order to, typically by changing the working fluid state, increase its pressure.
10
Having the control volume to enclose the compressor/pump and neglecting the kinetic and potential
energy variation, as well as the heat transfers with the surroundings, the power can be obtained from
Wp = m(h1 − h2)
Furthermore, assuming an ideal pump (i.e. without internal irreversibilities) and an incompressible
flow (i.e. a Mach number Ma << 0.3), the work per unit mass can be expressed as
Wp
m= v(p1 − p2) (2.4)
With the two previous equations one concludes that, for an ideal pump, the change in enthalpy is
equal to h2s− h1 = v(p2− p1), where h2s is the final enthalpy for an isentropic evolution. The real value
h2 is related to h2s through the pump’s isentropic efficiency.
Heat Exchanger
The heat rate between the control volume and the surroundings is negligible, when compared to the
heat transferred from the hot side to the cold one, making Q = 0. No power is generated nor consumed,
so W = 0. As it is implicit from the name, there is an exchange of heat that takes place, usually between
two fluid streams, so, the following expression represents the general energy balance,
0 =∑e
me he −∑i
mi hi
If two different mass flows pass through the hot and cold sides, as schematized in Fig. 2.3, the energy
rate balance takes the following form,
0 = m1(h2 − h1) + m3(h4 − h3)
In the case where the mass flows passing through both sides present the same rates, the energy rate
balance equation simplifies to
h1 − h2 = h4 − h3
CV
Hot side
Cold side
1
3 4
Figure 2.3 Counterflow heat exchanger schematic illustration. Adapted from [18].
A regenerator and an heat recovery steam generator (HRSG) are employed in the models proposed
in this work. Further explanation of these heat exchangers is presented in chapter 3.
11
2.2 Previous Work
Dunham and Lipinski published their work [17] in 2013 about a regenerative Brayton and combined
Brayton-Rankine cycles fuelled by solar energy alone. In that paper, concentration ratios characteristic
of single-axis concentrator collectors, such as LFR and PTC were examined. The authors examined the
cycles efficiencies and power performances for different working fluids in both the topping and bottoming
cycle. This section reviews their work.
The schematic presentations of the thermodynamic cycles are presented in figures 2.4 and 2.5.
The minimum system pressure for all Brayton cycles was set to the atmospheric pressure (1 bar). The
maximum Brayton cycle pressure, at the compressor exit, was initially varied so that the optimal pressure
ratio that provides the maximum cycle efficiency was found. The minimum temperature, T1, was assumed
to be 308 K and the maximum temperature, T4, was obtained assuming a receiver solar concentration
ratio of C = 70. Pressure drops in relation to the pressure at the entrance of the solar receiver and heat
exchanger were assumed to be 5% and 2%, respectively. The efficiency of the compressor, regenerator
and turbine were 0.796, 0.87 and 0.858, respectively. The power input from the solar receiver was taken
as 150 kW.
Figure 2.4 Brayton cycle scheme [17].
At the bottoming cycle, in the combined cycle model, the temperature at the inlet of the pump was
assumed to be 308 K, and the pressure was set to be the saturation pressure at that temperature or the
atmospheric pressure of 1 bar, whichever was greater. The steam turbine inlet temperature was fixed at
400 K. The pinch point temperature difference at the heat exchanger is fixed to 10 K. The maximum
pressure of the Rankine cycle was made to vary up to 20 bar. The isentropic efficiencies of the pump and
the turbine were 0.6 and 0.68, respectively. Table 2.1 presents the considered parameters concisely.
In order to estimate the optimal operating temperature of the solar receiver, Trec, Dunham and
Lipinski employed the following equation to obtain the theoretical efficiency of converting solar heat into
mechanical work,
η=
(αCG0 − εσ(Treceiver)4 − Uconv(Treceiver − Tamb)
CG0
)︸ ︷︷ ︸
Receiver efficiency
(1− Tamb
Treceiver
)︸ ︷︷ ︸
Carnot efficiency
(2.5)
The different receiver parameters were assumed as follows,
13
• Receiver tubing surface with α = 0.9 and ε = 0.2;
• Overall heat transfer coefficient, Uconv = 10 W/m2K;
• Standard solar irradiation, G0 = 1 Wm−2;
• Atmospheric temperature, Tamb = 303 K;
• Concentration ratio, C = 50, 70, and 100;
• σ is the Stefan-Boltzmann constant (5.67036713×108 Wm−2K−4)
The efficiency was then plotted as shown in Fig. 2.6. The peak efficiency in the curve correspondent
to C = 70 occurred at Treceiver = 831 K. This value was obtained by differentiating Eq. (2.5) with respect
to Treceiver and subsequent equalization to zero. The working fluid was assumed to reach a temperature
of 50 K lower than of the receiver. Thus, T4 was considered to be 781 K in the analysis.
Figure 2.6 Efficiency of a receiver-Carnot heat engine system as a function of receiver temperature forC = 50, 70, and 100 [17].
The Brayton cycle efficiency is obtained from the calculation of the ratio between the cycle power and
the heat input,
ηcycle =Wcycle
Qin
(2.6)
The combined cycle efficiency is calculated as follows,
ηcycle =Wt,cycle + Wb,cycle
Qin
(2.7)
Varying the pressure ratio of the compressor, pR, from 1 to 10, different Brayton cycle efficiency
curves were obtained as shown in Fig. 2.7. The optimal points, as well as the state temperatures, net
cycle power output, and the cycle efficiency for each fluid configuration are presented in Table 2.2. CO2
outperformed the other fluids, providing the highest cycle efficiency of 0.1531.
14
Figure 2.7 Brayton cycle efficiency versus cycle pressure ratio [17]
Regarding the combined Brayton-Rankine cycle, the same analysis was made by obtaining different
values of the global cycle efficiency as a function of the pressure ratio. The peak efficiency and power
results by fluid pair - topping and bottoming fluids combination - are presented in bar charts in Fig. 2.8
and Fig. 2.9. Table 2.3 shows those corresponding specific values. The preponderance of CO2 is notable,
being the fluid that allows for the highest efficiencies and cycle powers.
Figure 2.8 Top cycle efficiency for each configuration of fluids in the combined cycle [17].
Table 2.2 Brayton cycle peak efficiencies and correspondent values of pressure ratio, volume flow, tem-perature at the different states and power generated. Values from [17].
Each of the terms above presented are calculated using the energy mass balance at each cycle com-
ponent at hand,
Wgas turbine = m1(h5 − h6)
Wcompressor = m1(h1 − h2)/0.9797
Wsteam turbine = mb(h14 − h15)
Wpump = mb(h12 − h13)
Having the power output and input to the combined cycle, one can now calculate the global cycle
efficiency,
ηcycle =Wcycle
Qin
30
Chapter 4
Computational Modelling
The proposed models presented in the previous chapter are simulated in a MATLAB program developed
for this thesis. This chapter describes the restrictions that are embed in the MATLAB code and explains
the algorithm implemented in the simulations.
4.1 Restrictions
The implementation of restrictions is crucial to guarantee the reliability of the results in a real life context.
This section describes those restrictions, being organized in two subsections: thermodynamic properties
restrictions, which are applicable across all three models; model restrictions, having some nuances that
vary with the used model.
4.1.1 Thermodynamic Properties Restrictions
Each working fluid in the topping and bottoming cycle of the models are constrained to operate in a
certain limited temperature range. In the case of the bottoming cycle, CoolProp has these thermodynamic
restrictions incorporated, preventing the program to keep running when the range values are not verified.
In the Brayton cycle, however, restrictions have to be implemented given that there is no such library.
These are incorporated by programming in the code so that situations such as fluid condensation do not
occur.
As explained more thoroughly in the next chapter, temperatures at the exit of the Brayton cycle
regenerator and at the exit of the HRSG (both cases at pressures close to 1 bar) can attain excessively
low temperatures, which brings to the necessity to have temperature restrictions. Table 4.1 shows the
minimum temperatures for each gas at which the program is allowed to run. Except for CO2, those
correspond to the triple point temperatures, and since they are used by Dunham and Lipinksi [17], the
same was done in this work. In the CO2 case, however, because its triple point temperature is too high
comparing to the others (216.6 K), its restriction is pushed further down to 195 K in order to increase
the span of results (the increase is not extreme, but it is significant). This is the chosen value because,
at the critical states as the one mentioned above, the gas is at around 1 bar, which, as shown in Fig. 4.1,
31
CO2 can operate on. The point (-78.5◦ C/194.5 K, 1 atm/1.01 bar) in the diagram is used as a reference
point for this decision.
Table 4.1 Minimum temperatures for the Brayton cycle fluids.
Fluid Lower limit temperature
CO2 195.0 K
Air 60.0 K
N2 63.151 K
He 2.1768 K
H2 13.957 K
Figure 4.1 Phase diagram of CO2 [23].
4.1.2 Model Restrictions
As mentioned before, there is also a set of model restrictions that have to be implemented into the
developed program. Several restrictions are implemented with the purpose of rejecting solutions that
result from physics incoherences occurring in intermediate steps and others aiming for the models to
comply with the desired design conditions.
Starting with the physical meaning, the models are restricted to operate in a set of conditions that,
for instance, avoid a component to exhibit thermal energy rate input (e.g. solar receiver) greater than
that leaving the equipment (in this case, the condition h4 > h3 has to be verified). These constraints
have to be introduced into the program code since the software does not have any physics considerations
or sensibility incorporated. This type of restriction is applied to the following components: regenerator,
solar receiver, combustion chamber, and HRSG, avoiding negative energy rates and negative mass flow
rates.
Regarding the models intended design characteristics, a few more restrictions are implemented.
32
Heat dissipation for the fluid to reach initial conditions
It is desired that no other source of energy is added to the cycle to get the Brayton cycle fluid to
achieve the initial cycle conditions (308 K, 1 bar) at the compressor inlet. In this way, the fluid at states
(8)/(10)/(11) in models 1/2/3 is restricted to have temperatures equal or greater than 308 K.
Brayton cycle regenerator
In models 2 and 3, the efficiency is calculated from the NTU relations (as described in chapter 3).
As in the recurrently mentioned work [17], the effectiveness is fixed at 0.87 (which one can percept as
relatively high value in the heat exchangers domain), in the mentioned models the regenerator is designed
to provide maximum effectiveness values at around 0.87.
Combustion chamber
The purpose of adding a combustion chamber in models 1 and 3 is to have it as a support to the
solar receiver, operating together so that the cycles generate higher electric powers at higher efficiencies.
Accordingly, in order to avoid a combustion chamber providing the fluids with disproportionally higher
values of heat than the solar receiver, the models are restricted to run by the following restriction:
Qc < Qs.
4.2 Numerical Model
In this section it is described the algorithm used do implement the models described in Chapter 2.
In order to study the performance of the models proposed in the present work, the pressure ratio in
the Brayton cycle (pR = p2/p1) and the temperature at the exit of the combustion chamber (T5, in the
cases of model 1 and model 3) constitute the first variable inputs of the system. The ratio, r, between the
mass flow rates resulting from the stream splitting in models 2 and 3 is assumed to have a given value.
The main algorithm employed is presented in the form of flowchart schematized in Fig. 4.2.
The program, developed in MATLAB, follows the following steps:
1. Start by selecting the model chosen by the user that is firstly simulated.
2. Select the gas used in the Brayton cycle. Depending on the selected gas, different thermodynamic
property data bases are loaded and become ready to be called in. Restrictions are also considered
accordingly.
3. Select the Rankine cycle fluid. Different fluids result in different thermodynamic properties, and in
different pressure values at the inlet and outlet of the pump (as mentioned in section 2.3).
4. Read the input variables. The values to be iterated by the computer code correspond to the entries
in the chosen vectors of pressure ratio, ~p2, and temperature at the exit of the combustor, ~T5. Given
increment values, ∆p and ∆T , are taken such that: ~p2(i + 1) = ~p2(i) + ∆p and ~T5(i + 1) =
~T5(i) + ∆T5. These values are then specified and explained in chapter 5. The iterations are made
till the last entries of pressure and temperature, np and nT , respectively, are reached.
33
5. Evaluate the model that was selected and decides whether it starts an iterative process by varying
T5 or not. As aforesaid, this process is not applied to model 2 given the absence of a combustion
chamber.
6. Assume the first entry value of ~T5, if applied, and, subsequently, the first value of ~p2.
7. For each value of p2 and T5, Brayton and Rankine cycles’ outputs are calculated.
8. Perform post-processing of data so that a comprehensive analysis of the systems output variation
with the variable inputs (e.g., plots, bar charts) is possible.
After getting the results from these iterations, further simulations are made in order to study the
models behaviour with other variables, such as the mass flow ratio that governs the stream splitter after
the gas turbine in models 2 and 3 and the regenerator design - heat transfer area - (models 2 and 3).
The method to accomplish the results of these studies is identical to the one presented in the previously
mentioned Flowchart.
Regarding the fixed parameters, the same values as those used by Dunham and Lipinksi [17] article
are taken. In terms of components efficiencies, the gas and steam turbines have an isentropic efficiency
of 0.858 and 0.68, respectively. The isentropic efficiency values taken for the compressor and the pump
are 0.796 and 0.6, respectively. The solar receiver operates at a solar concentration ratio of C = 70. For
the heat recovery steam generator, the temperature difference at the pinch point is assumed to be 10 K
until varied in upcoming simulations presented in the next chapter.
34
Start
Select Brayton cycle
fluid
Select Rankine cycle
fluid
Input vectors:
2, 5
Select Model
Evaluate
Model
Model 1 or Model 3
T
T
5 assumes the first
entry value of T5
p2 assumes the first
entry value of p2
Model 2
Calculate Brayton
cycle outputs
Calculate Rankine
cycle outputs
Calculate Global
cycle outputs
p2 + p
Post Processing
T5 + T
p2 p2 n
p2 = p2 n
T5 T5 n
5 = 5
p
T T n
T
p
T
p
Figure 4.2 Flowchart representation of the main employed algorithm.
35
36
Chapter 5
Results and Discussion
5.1 Comparison with Reference Results
As the models previously proposed and described are coded into a computer program developed from
scratch in MATLAB, a comparison with the ones obtained in Dunham and Lipinski’s [17] work is per-
formed.
Fig. 5.1 displays the global efficiency curves obtained in the present work and the ones obtained in
the previously mentioned work. One can observe that the results match, validating the program and the
thermodynamic properties data base.
pR
1 2 3 4 5 6 7 8 9 10
ηcycle
0
0.05
0.1
0.15
0.2
CO2
Air
N2
He
H2
(a) MATLAB (b) Dunham and Lipinski’s [17] obtained results
Figure 5.1 Brayton cycle results comparison between the results from MATLAB in the present workand those obtained in the work of Dunham and Lipinski [17].
In regards to the combined cycle, Table 5.1 shows the results from the MATLAB simulations. Com-
paring to the results obtained in the previous work (section 2.2, Table 2.3), these do not distance with
any considerable magnitude from the ones previously obtained, have the highest deviation to be on the
fluid pair He & Cyclohexane, presenting a difference of ηprevious − ηMATLAB = 0.0234. With respect to
the results per fluid pair compared to the others within each set of results, the same tendency is verified.
37
The differences of the absolute values can be explained from the distinct thermodynamic properties data
bases (recalling section 2.3, in the present work the CoolProp’s library was used).
Table 5.1 Peak global efficiency values obtained from the simulation of the article’s combined cycle.
Figure 5.3 Model 1: influence of temperature T5 on the global efficiency for each fluid in the Rankinecycle and CO2 in the Brayton cycle.
41
R-245fa R-141b Cyclohexane n-Pentane Water
ηcycle
0
0.05
0.1
0.15
0.2
0.25
0.3
825 K
850 K
875 K
900 K
(a) CO2
R-245fa R-141b Cyclohexane n-Pentane Water
ηcycle
0
0.05
0.1
0.15
0.2
0.25
0.3
825 K
850 K
875 K
900 K
(b) Air
R-245fa R-141b Cyclohexane n-Pentane Water
ηcycle
0
0.05
0.1
0.15
0.2
0.25
0.3
825 K
850 K
875 K
900 K
(c) N2
R-245fa R-141b Cyclohexane n-Pentane Water
ηcycle
0
0.05
0.1
0.15
0.2
0.25
0.3
825 K
850 K
875 K
900 K
(d) He
R-245fa R-141b Cyclohexane n-Pentane Water
ηcycle
0
0.05
0.1
0.15
0.2
0.25
0.3
825 K
850 K
875 K
900 K
(e) H2
Figure 5.4 Model 1: highest cycle efficiency values for each fluid combination at temperatures T5 equalto 825 K, 850 K, 875 K, and 900 K.
42
R-245fa R-141b Cyclohexane n-Pentane Water
Pow
erOutput[kW
]
0
10
20
30
40
50
60
70
80825 K
850 K
875 K
900 K
(a) CO2
R-245fa R-141b Cyclohexane n-Pentane Water
Pow
erOutput[kW
]
0
10
20
30
40
50
60
70
80825 K
850 K
875 K
900 K
(b) Air
R-245fa R-141b Cyclohexane n-Pentane Water
Pow
erOutput[kW
]
0
10
20
30
40
50
60
70
80825 K
850 K
875 K
900 K
(c) N2
R-245fa R-141b Cyclohexane n-Pentane Water
Pow
erOutput[kW
]
0
10
20
30
40
50
60
70
80825 K
850 K
875 K
900 K
(d) He
R-245fa R-141b Cyclohexane n-Pentane Water
Pow
erOutput[kW
]
0
10
20
30
40
50
60
70
80825 K
850 K
875 K
900 K
(e) H2
Figure 5.5 Model 1: highest cycle power output values for each fluid combination at temperatures T5
equal to 825 K, 850 K, 875 K, and 900 K.
43
The results displayed in figures 5.3, 5.4, and 5.4 lead to the following conclusions:
• The maximum power output value at a highest global cycle efficiency point (76.32 kW) occurs with
the fluid pair CO2 and n-Pentane at a temperature T5 = 875 K, which also corresponds to the
highest global efficiency point (0.255);
• Power outputs and global efficiencies tend to rise with the increase of temperature T5, for 825 K
< T5 < 875 K. For temperatures T5 higher than 875 K, this tendency is not always verified due to
the imposed restrictions to the combined Brayton-Rankine cycle.
• CO2 consistently provides the highest global cycle efficiencies, excluding the cases at which no
results complying with the restrictions are obtained.
5.2.2 Model 2
The second model novelty, compared with those in the reviewed literature, is the addition of a stream
splitter after the gas turbine, maintaining solar energy as the only source of energy. Since there is no
combustion chamber added, the temperature T4 of the gas entering the gas turbine is calculated from an
energy balance to the solar receiver.
The range of values of the pressure ratio varies between 1 and 10, as this span covers the results of
interest to the study.
The regenerator effectiveness, which depends on the mass flow rates and temperatures on its cold and
hot side, was calculated from the effectiveness - NTU relations, as explained in 3.2.
An iterative process consisting in varying values of mass flow ratio, r = m8/m1, and heat transfer area
of the regenerator led to the decision of choosing 0.5 and 6 m2 as their values, respectively, to perform
the subsequent simulations. Considering the appropriate restrictions presented in section 4.1, the global
efficiency is plotted against the pressure ratio in Fig. 5.6.
Analysing the results, one can notice that all the efficiency curves maintain the trend of reaching
a maximum value at a certain pressure ratio value. However, comparing to model 1, the global cycle
efficiency values fall more abruptly after reaching their peaks. Another important difference, comparing
figures 5.6 (model 2) and 5.3 (model 1), is the fact that, in the latter case, possible operation points
spread around maximum efficiency, except for T5 = 875 K and T5 = 900 K, whereas in the former, they
are always located beyond that point, i.e., the global cycle efficiency never reaches its maximum value.
This happens mainly due to the gas temperature restrictions, explained in section 4.1.1. In order to get
an insight on how these restrictions influence the possible results, Fig. 5.7 shows the results of global
efficiency with (left) and without (right) temperature restrictions, using CO2 in the Brayton cycle.
One might wonder why temperature constraints have such a larger impact on the results of model 2
(and model 3, as well) than on those of model 1.
As explained before, the stream splitter has the advantage of supplying fluid to the HRSG at a
larger enthalpy compared to model 1. However, it can be too large to meet the system’s requirements.
Recalling the pinch point equation presented in section 3.1 (eq. (3.12)), that can be rearranged yielding
h8 − h10 = (h7 − h10) − A × (h7 − ht,pinch), where A = (h11 − h10)/(h11 − hb,pinch) > 1. By fixing the
44
enthalpies of the Rankine cycle and the pinch point difference, A becomes a constant, as well as ht,pinch.
From the above equation, one concludes that h8 < h10 whenever h7 > (ht,pinch × A − h10)/(A − 1).
In other words, there is a threshold value for the enthalpy at the inlet of the HSRG hot side, above
which the enthalpy at the hot side outlet is smaller than the enthalpy at the cold side inlet, which is
obviously physically impossible. An equivalent analysis can be applied to model 2, replacing states 7, 8,
10, 11 by 6, 7, 12, 13 (likewise for model 3, replacing 10, 11 by 13, 14). Models 2 and 3 provide higher
temperatures at the inlet of the HRSG hot side. In fact, those values may be so high that it results in
too low temperatures at its outlet, making it difficult to comply with the thermodynamic constraints
mentioned previously. This is the main reason why models 2 and 3 have their results considerably more
restricted than those of model 1, and also explains the huge difference in Fig. 5.7 between the results
with and without these constraints.
To allow the system to work with high enthalpy values h6, thus avoiding thermodynamic constraints,
two strategies were implemented:
1. Increase the value of the enthalpy h13 at the steam turbine inlet;
2. Increase the pinch point temperature difference.
These two options can be mathematically justified by the enthalpy threshold value mentioned pre-
viously. Similarly to what happens in model 1, in model 2 the condition h7 < h12 is met whenever
h6 > (ht,pinch×A−h12)/(A− 1), where A = (h13−h12)/(h13−hb,pinch). Therefore, to avoid unrealistic
solutions, the threshold value h6,max = (ht,pinch×A−h12)/(A−1) must be as large as possible. Differenti-
ating h6,max with respect to A, one concludes that h6,max increases when A decreases and, differentiating
A with respect to h13, the conclusion is similar, i.e., A decreases when h13 increases. Therefore, h6,max is
a monotone increasing function of h13, which explains the first strategy. The mathematical explanation
for the second strategy is much more straightforward, since, from the equation of h6,max, it is obvious
that h6,max increases with ht,pinch.
45
pR
0 2 4 6 8 10
ηcycle
0
0.1
0.2
0.3
0.4
0.5
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(a) CO2
pR
0 2 4 6 8 10
ηcycle
0
0.1
0.2
0.3
0.4
0.5
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(b) Air
pR
0 2 4 6 8 10
ηcycle
0
0.1
0.2
0.3
0.4
0.5
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(c) N2
pR
0 2 4 6 8 10
ηcycle
0
0.1
0.2
0.3
0.4
0.5
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(d) He
pR
0 2 4 6 8 10
ηcycle
0
0.1
0.2
0.3
0.4
0.5
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(e) H2
Figure 5.6 Model 2: global cycle efficiency versus pressure ratio with r = 0.5 and a heat transfer areaof the regenerator equal to 6 m2
46
pR
0 2 4 6 8 10
ηcycle
0
0.1
0.2
0.3
0.4
0.5
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(a) Gas temperatures restrictions applied
pR
0 2 4 6 8 10
ηcycle
0
0.1
0.2
0.3
0.4
0.5
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(b) No gas temperatures restrictions applied.
Figure 5.7 Model 2: influence of the gas temperatures restrictions on the possible results using CO2 inthe Brayton cycle for all Rankine fluids.
Increase of T13
The temperature T13 was raised from 400 K to 427 K, given that 427 K is the lowest temperature of
all the critical points (which corresponds to R-245fa, according to CoolProp [20]). The increase of h13
also implies an increase in the steam turbine expansion, i.e., in the net power output of the Rankine cycle.
Figures 5.8 and 5.9 show the global efficiency and power outputs obtained from the implementation of
the first strategy. As the behaviour is similar for every gas, only the results for CO2 are presented. The
rise of the temperature at the steam turbine inlet increases the range of physically acceptable results, as
well as the global cycle efficiencies, but not to the extent desired, since in Fig. 5.8 the possible solutions
are still far from the point of maximum efficiency.
The results for the power outputs are shown in Fig. 5.10. The most striking feature in figures 5.9 and
5.10 is the increase in global efficiency and net power output for the pair of fluids CO2 and R-245fa.
pR
0 2 4 6 8 10
ηcycle
0
0.1
0.2
0.3
0.4
0.5
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(a) Global efficciency curves for T13 = 400 K
pR
0 2 4 6 8 10
ηcycle
0
0.1
0.2
0.3
0.4
0.5
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(b) Global efficciency curves for T13 = 427 K
Figure 5.8 Model 2: global efficiency curves for temperature T13 = 400 K and T13 = 427 K at the inletof the steam turbine, using CO2 in the Brayton cycle.
47
CO2
Air N2
He H2
ηcycle
0
0.05
0.1
0.15
0.2
0.25
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(a) 400K
CO2
Air N2
He H2
ηcycle
0
0.05
0.1
0.15
0.2
0.25
R245FA
R141b
Cyclohexane
nPentane
Water
(b) 427K
Figure 5.9 Model 2: global maximum efficiency values for temperatures T13 = 400 K and T13 = 427 Kat the inlet of the steam turbine.
CO2
Air N2
He H2
Pow
erOutput[kW
]
0
5
10
15
20
25
30
R-245fa R-141b Cyclohexane n-Pentane Water
(a) 400K
CO2
Air N2
He H2
Pow
erOutput[kW
]
0
5
10
15
20
25
30
R-245fa R-141b Cyclohexane n-Pentane Water
(b) 427K
Figure 5.10 Model 2: net power at maximum global efficiency points for temperatures T13 = 400 K andT13 = 427 K at the inlet of the steam turbine.
In order to compare the results obtained with the ones from the work of Dunham and Lipinski [17],
tables 5.4 and 5.5 show the maximum global efficiency values of model 2 at temperatures at the steam
turbine’s inlet 400 and 427 K, respectively. For the pair of fluids CO2 and R-245fa, mentioned previously,
the increase in global efficiency equals 41.4 %.
Table 5.4 Model 2: maximum global efficiency values for T13 = 400 K.
In the second strategy, the value T13 = 400 K was kept constant and the pinch point temperature
difference was raised above 10 K. In Fig. 5.11, results are presented for the CO2 and R-245fa, Air and
R-141b fluid pairs, having each plot different curves corresponding to different pinch point temperature
differences. These two fluid pairs were chosen randomly to demonstrate that the same tendency is verified,
regardless of the working fluids.
As observed, the increase of the pinch point temperature difference decreases the global efficiency val-
ues, while also broadening the span of acceptable results. However, for the former fluid pair (which is the
most affected by the restrictions implemented), even at the lowest efficiency curves (highest ∆Tpinch point),
the global cycle efficiency is higher than the value obtained with a pinch point temperature difference
equal to 10 K, due to thermodynamic constraints. As it is not realistic to have ∆Tpinch point values as
high as 100 K, and as it would lower every other configuration’s global efficiency, its value should not
differ much from the value of 10 K mentioned in the work of Dunham and Lipinski [17].
pR
0 2 4 6 8 10
ηcycle
0
0.1
0.2
0.3
0.4
0.5
∆Tpinch=10 K
∆Tpinch=25 K
∆Tpinch=50 K
∆Tpinch=100 K
(a) CO2 & R-245fa
pR
0 2 4 6 8 10
ηcycle
0
0.1
0.2
0.3
0.4
0.5
∆Tpinch=10 K
∆Tpinch=25 K
∆Tpinch=50 K
∆Tpinch=100 K
(b) Air & R-141b
Figure 5.11 Model 2: global efficiency curves for different values of pinch point temperature differencewith T13 = 400 K.
Regenerator Heat Transfer Area
Other specification of the model cycles that influences its performance is the regenerator design and,
more specifically, its heat transfer area. Thus, the behaviour of the global efficiency curves when the heat
transfer area of the regenerator varies is analysed, keeping a mass flow ratio, m8/m1, equal to 0.5, and a
temperature at the inlet of the steam turbine T13 = 400 K. For the sake of conciseness, the fluid pair Air
49
and R-245fa was chosen, since the global efficiency curves show the same behaviour.
Fig. 5.12 shows the global efficiency evolution with the variation of the heat transfer area. The
larger the regenerator heat transfer area, the greater the global cycle efficiency, since the regenerator
effectiveness increases with its area of heat transfer (section 3.2). Furthermore, the pressure ratio at
which the curves reach their maximum level tends to decrease with the area increase. However, the range
of physically acceptable results does not improve. It is also worth noticing that no acceptable results are
obtained for areas of 8 m2 or higher.
pR
0 2 4 6 8 10
cycle
0
0.1
0.2
0.3
0.4
0.5
1m2
2m2
4m2
6m2
8m2
Figure 5.12 Model 2: regenerator heat transfer area influence on global cycle efficiency, for the fluid pairAir and R-245fa, assuming T13 = 400 K, ∆Tpinch point = 10 K and a mass flow ratio r = m8/m1 = 0.5.
Mass Flow Ratio
The mass flow ratio r = m8/m1 also affects the regenerator effectiveness. In the results presented in
Fig. 5.13, the regenerator heat transfer area was kept constant and equal to 6 m2.
The results show that the increase of the mass flow ratio implies a reduction of global efficiencies,
which can be explained by the following reasons:
– From the NTU relations (3.2), the increase of m8/m1 decreases the regenerator effectiveness and,
therefore, the Brayton cycle efficiency;
– The mass flow through the HRSG (m6 = m1(1 − r)) decreases, which results in a decrease of the
bottoming cycle mass flow, (mb = m6(h6 − h7)/(h13 − h12)) and, consequently, of the Rankine
power output and efficiency.
It is noteworthy that, with the increase of the mass flow ratio, the pressure ratios at which global
efficiencies reach their maximum values tend to be relocated to higher values, and that the curves inter-
section with the horizontal axis occurs for lower pR values. Regarding the range of acceptable results,
(solid lines) a tendency towards a broader spectrum can be observed when the mass flow ratio increases,
despite the lower efficiencies.
50
pR
0 2 4 6 8 10
ηcycle
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.20
0.35
0.50
0.65
0.80
Figure 5.13 Model 2: mass flow ratio r = m8/m1 influence on global cycle efficiency, for the fluid pairAir and R-245fa, assuming a regenerator heat transfer area equal to 6 m2, T13=400 K and ∆pinch point =10 K.
Nevertheless, the maximum global efficiency obtained in the work of Dunham and Lipinski [17]
(0.2106) could to not be reached, which can be explained by the following reasons:
• In the mentioned authors work, the regenerator effectiveness is considered constant and equal to
0.87, whilst in the present model it depends on different parameters such as the heat transfer area,
mass flow rates and specific heats, the maximum value being 0.87. Such disparities prevent a
possible and fair comparison between those cycles to be obtained. This issue will be addressed in
the next section;
• This model forces the Brayton cycle fluids to go to more extreme conditions due to its own config-
uration by having the stream bifurcation, which implies more severe restrictions in the operating
conditions.
On the other hand, however, efficiencies for several fluid combinations were raised, compared to
Dunham and Lipinski’s work [17]. Those correspond to the combinations: Air - R-141b, He - R-141b, He
- Cyclohexane, He - n-Pentane, H2 - R-141b, H2 - Cyclohexane, and H2 - n-Pentane.
5.2.3 Model 3
This model, compared to the previous one, brings the novelty of having a combustion chamber placed
after the solar receiver.
In a first approach, it is expected that this model provides the highest global efficiency values, com-
pared to the other two, due to the implementation of a combustion chamber, which increases the expansion
in the gas turbine. However, the analysis is not so straightforward.
Fig. 5.14 presents the global cycle efficiency results for the metallurgical temperature limit of the gas
turbine, T5 = 1173 K, assuming mass flow ratio of r = m9/m1 = 0.5, and a regenerator heat transfer
area equal to 6 m2. The two last parameters were chosen in order to make a comparison with the results
51
of model 2. The pR values range from 1 to 40 for this first simulation and are bounded to 20 thereafter,
since there are no relevant results above that value.
As expected, the efficiencies are overall higher than the ones obtained with model 2, which can
be observed comparing figures 5.14 and 5.6, and no acceptable results are obtained. However, when
compared to the peak efficiencies in model 1, using the same temperature T5 = 1173 K (Fig. 5.2), the
global efficiencies of model 3 are considerably lower. Such outcome is mainly due to the fact that in
model 1 (as well as in the work of Dunham and Lipinski [17]) the regenerator effectiveness was considered
constant and equal to 0.87, whereas in models 2 and 3 this parameter was calculated using the NTU
relations, imposing a maximum value of 0.87. To make a comparison on equal terms, the same procedure
should be applied to model 1. Efficiency curves are presented in Fig. 5.15 (model 1) and Fig. 5.16 (model
3), assuming a temperature at the inlet of the gas turbine T5 = 825 K and a regenerator heat transfer
area of 6 m2. The peak efficiency values are presented in tables 5.6 and 5.6 .
52
pR
5 10 15 20 25 30 35 40
ηcycle
0
0.1
0.2
0.3
0.4
0.5
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(a) CO2
pR
5 10 15 20 25 30 35 40
ηcycle
0
0.1
0.2
0.3
0.4
0.5
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(b) Air
pR
5 10 15 20 25 30 35 40
ηcycle
0
0.1
0.2
0.3
0.4
0.5
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(c) N2
pR
5 10 15 20 25 30 35 40
ηcycle
0
0.1
0.2
0.3
0.4
0.5
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(d) He
pR
5 10 15 20 25 30 35 40
ηcycle
0
0.1
0.2
0.3
0.4
0.5
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(e) H2
Figure 5.14 Model 3: global cycle efficiency assuming T5 = 1173 K, a regenerator heat transfer areaequal to 6 m2 and a mass flow ratio r = 0.5.
53
pR
0 5 10 15 20
ηcycle
0
0.05
0.1
0.15
0.2
0.25
0.3
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(a) CO2
pR
0 5 10 15 20
ηcycle
0
0.05
0.1
0.15
0.2
0.25
0.3
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(b) Air
pR
0 5 10 15 20
ηcycle
0
0.05
0.1
0.15
0.2
0.25
0.3
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(c) N2
pR
0 5 10 15 20
ηcycle
0
0.05
0.1
0.15
0.2
0.25
0.3
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(d) He
pR
0 5 10 15 20
ηcycle
0
0.05
0.1
0.15
0.2
0.25
0.3
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(e) H2
Figure 5.15 Model 1: global cycle efficiency assuming T5 = 825 K and a regenerator heat transfer areaequal to 6 m2.
54
pR
0 5 10 15 20
ηcycle
0
0.05
0.1
0.15
0.2
0.25
0.3
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(a) CO2
pR
0 5 10 15 20
ηcycle
0
0.05
0.1
0.15
0.2
0.25
0.3
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(b) Air
pR
0 5 10 15 20
ηcycle
0
0.05
0.1
0.15
0.2
0.25
0.3
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(c) N2
pR
0 5 10 15 20
ηcycle
0
0.05
0.1
0.15
0.2
0.25
0.3
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(d) He
pR
0 5 10 15 20
ηcycle
0
0.05
0.1
0.15
0.2
0.25
0.3
R-245fa
R-141b
Cyclohexane
n-Pentane
Water
(e) H2
Figure 5.16 Model 3: global cycle efficiency assuming T5 = 825 K, a regenerator heat transfer areaequal to 6 m2 and a mass flow ratio r = 0.5.
55
Table 5.6 Model 1 (with the regenerator effectiveness calculated from the NTU relations): peak globalefficiency values for each fluid combination.
The results show that model 3 outperforms model 1. With the exception of the pair CO2 and R-245fa,
every fluid combination in model 3 yields higher peak global efficiencies than in the case of model 1. It
is worth pointing that no acceptable results are obtained in model 1 for He.
To understand the influence of the stream splitter on the regenerator effectiveness, Fig. 5.17 presents
the effectiveness values for the three models, under the same design conditions, imposing a maximum
effectiveness value of 0.87. The effectiveness for the first model is considerably lower. Since there is
no stream splitter in model 1, the larger mass flow on the regenerator hot side is the reason for this
behaviour.
pR
0 5 10 15 20
ηreg
0
0.2
0.4
0.6
0.8
0.87
1
CO2
Air
N2
He
(a) Model 1
pR
0 2 4 6 8 10
ηreg
0
0.2
0.4
0.6
0.8
0.87
1
CO2
Air
N2
He
H2
(b) Model 2
pR
0 5 10 15 20
ηreg
0
0.2
0.4
0.6
0.8
0.87
1
CO2
Air
N2
He
H2
(c) Model 3
Figure 5.17 Brayton cycle’s regenerator effectiveness values for models 1, 2, and 3, assuming a heattransfer area of 6 m2 and, for models 2 and 3, a mass flow ratio r = 0.5.
These results, reconciled with the global efficiencies results, show the degree of influence that the
regenerator effectiveness has in the cycles performances and how the stream splitting provides better
results.
56
Chapter 6
Conclusions and Future Work
6.1 Conclusions
The main goal of this work was to study the performance of three combined Brayton-Rankine cycles with
solar thermal power generation and, if possible, improve the results of the work of Dunham and Lipinski
[17].
The main conclusions are:
1. The fluid pair CO2 and R-245fa provides the highest global cycle efficiency curves for all the models.
In the case of model 2 and 3, however, due to the stream splitting, this pair is the most affected
in terms of thermodynamic restrictions, resulting in the lowest global efficiencies, despite having
better maximum efficiencies.
2. In model 1, with its regenerator effectiveness calculated from the NTU relations, the fluid pair CO2
and n-Pentane stands out as the best pair, providing an efficiency value of 0.1649. The fluid pair
H2 and R-141b provides the best cycle performance for model 2, leading to a global efficiency of
0.1848. In regards to model 3, the best fluid pair is CO2 and R-141b, providing an efficiency value
of 0.2045.
3. With the model operating with the same Brayton cycle regenerator design conditions, as in the
work of Dunham and Lipinski [17] (εreg = 0.87), having a combustion chamber supporting the solar
receiver in the generation of heat (with Qc < Qsol) provides higher global efficiencies than with
solar alone. This also supports the idea that the combination of both types of energy can feed a
power cycle more efficiently.
4. With εreg being obtained from the NTU-effectiveness relations, models 2 and 3 outperform model
1, which is similar to the cycle studied in the work of Dunham and Lipinski [17]. Such results allow
us to conclude that the implementation of the stream splitter is beneficial.
57
6.2 Future Work
As mentioned in chapter 1, technologies for the exploitation of solar energy are developing at a significant
pace due to the present environmental and economical context. Therefore, research in this field is very
relevant and some suggestions can be made regarding future work:
1. In addition to the energy analysispresented in this work, an exergetic analysis should also be done,
in order to identify the components of the cycle with major exergy destructions.
2. A further analysis on a wider range of fluids.
3. Finally, as in any engineering study, an economic analysis is paramount. As the goal is to maxi-
mize gains and minimize costs, these conflicting objectives can be studied using a multi-objective
optimization, which examines the required trade-off between increased electricity production and
additional investment cost. This analysis should also take into account CO2 emissions, which will
be increasingly penalized by taxes and regulations, in the near future.
58
Bibliography
[1] L. Capuano. International Energy Outlook 2018. U. S. Energy Information Administration, 2018.
[2] G. Knies. A brief overview on global energy, water and carrying capacity perspectives. Clean Power
from Deserts - The DESERTEC Concept for Energy, Water and Climate Security, 2009.
[3] E. Pihl, J. Spelling, and F. Johnsson. Thermo-Economic Optimization of Hybridization Options
for Solar Retrofitting of Combined-Cycle Power Plants. Journal of Solar Energy Engineering,