Relating gas turbine performance to combined

cycle efficiency
Simon Frick

Thesis for the Degree of Master of Science

Thesis advisors: Professor Magnus Genrup and Dr. Klas Jonshagen

This thesis for the degree of Master of Science in Engineering has been conducted
at the Division of Thermal Power Engineering, Department of Energy Sciences,
Lunds Tekniska Högskola (LTH) – Lund University (LU) and at Siemens Indus-
trial Turbomachinery AB (SIT AB).
Supervisor at SIT AB: Dr. Klas Jonshagen
Supervisor at LU-LTH: Professor Magnus Genrup
Examiner at LU-LTH: Associate Professor Marcus Thern
 ”The two fundamental laws of thermodynamics are, of course, insufficient to
determine the course of events in a physical system. They tell us that certain
things cannot happen, but they do not tell us what does happen.”

- Alfred J. Lotka, (1922)

Abstract
keywords: combined cycle efficiency, second law efficiency, exergy, HRSG, ir-
reversibility, heat engine, thermodynamics, T-q diagram.

The gas turbine has been around for over a century, providing power for a
variety of applications. The efficiency, i.e. the amount of power produced per
kilogram of fuel provided, has increased steadily over the years and is today
greater than 44 percent in a state of the art gas turbine. For electricity pro-
duction the efficiency can be increased further by combining the gas turbine
with a steam turbine. The energy in the hot exhaust gases can be used to boil
water into steam, which can then be used to drive a steam turbine producing
additional electricity. In a combined cycle the efficiency can reach 63-64 percent
with current technology. These power plants are expensive to operate primarily
due to the fuel prices, which heavily drives the need for even higher efficiency.
This is why companies operating older power plants often consider upgrading
their components.
When an upgrade is considered there is often an information gap between
the OEM (original equipment manufacturer) and the buyer, in this case the
plant owner. If the plant is not delivered as a turnkey by a single OEM, parts
and components are purchased from different OEMs specialized on the specific
equipment, leading to a situation where the OEM have limited knowledge about
the environment in which their equipment operates. Commonly the solution to
this is to create models and apply extensive heat and work balance equations
to calculate the impact on the plant following an upgrade of a component. This
is a time consuming and complicated task which also makes it unnecessarily
expensive since an expert has to be involved to perform the calculations. In this
report a method of predicting the combined cycle efficiency change is presented.
It is a compact formulation which has the potential to speed up the process
considerably. This could provide a tool for the OEM to quickly be able to
provide an answer to the customer, regarding the potential efficiency increase.

4
Contents

1 Introduction 9
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.3 Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.4 Previous work and literature study . . . . . . . . . . . . . . . . . 12
1.5 Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.5.1 IPSEPro . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.5.2 Newton-Raphson method . . . . . . . . . . . . . . . . . . 13

2 Theory 14
2.1 Heat engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2 Carnot cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2.1 Reversible isothermal expansion . . . . . . . . . . . . . . 15
2.2.2 Reversible adiabatic expansion . . . . . . . . . . . . . . . 15
2.2.3 Reversible isothermal compression . . . . . . . . . . . . . 15
2.2.4 Reversible adiabatic compression . . . . . . . . . . . . . . 15
2.3 First and second laws of thermodynamics . . . . . . . . . . . . . 16
2.3.1 First law . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.3.2 Second law . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.4 Brayton cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.5 Rankine cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.6 Combined cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.7 Second law efficiency . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.7.1 Exergy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.8 Carnot efficiency vs thermal efficiency . . . . . . . . . . . . . . . 20
2.9 Logarithmic mean temperature . . . . . . . . . . . . . . . . . . . 22
2.10 Exergy balance of the combined cycle . . . . . . . . . . . . . . . 24
2.10.1 Exergy input . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.10.2 Exergy out . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.10.3 HRSG irreversibilities . . . . . . . . . . . . . . . . . . . . 25
2.10.4 ST irreversibilities . . . . . . . . . . . . . . . . . . . . . . 25
2.10.5 Pumping irreversibilities . . . . . . . . . . . . . . . . . . . 25
2.10.6 Exergy balance . . . . . . . . . . . . . . . . . . . . . . . . 26

5
3 HRSG theory 27
3.1 HRSG pressure levels . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.2 T-Q diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.3 Stack temperature . . . . . . . . . . . . . . . . . . . . . . . . . . 28

4 Methodology 30
4.1 Model description . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.1.1 1PNRH . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.1.2 2PNRH . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.1.3 3PNRH . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.1.4 Reheat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.2 Exergy balance evaluation . . . . . . . . . . . . . . . . . . . . . . 32
4.3 Irreversibility change with exhaust temperature . . . . . . . . . . 33
4.3.1 Why use the second law efficiency? . . . . . . . . . . . . . 34

5 Derivation of correlation 35
5.1 Derivation of combined cycle efficiency equation . . . . . . . . . . 35
5.2 Evaluation of combined cycle efficiency equation . . . . . . . . . 36
5.2.1 Gas turbine exhaust temperature . . . . . . . . . . . . . . 36
5.2.2 Gas turbine exhaust mass flow . . . . . . . . . . . . . . . 37
5.2.3 Specific heat . . . . . . . . . . . . . . . . . . . . . . . . . 37
5.3 Derivation of correction factors . . . . . . . . . . . . . . . . . . . 38
5.3.1 Correction factors . . . . . . . . . . . . . . . . . . . . . . 39
5.4 Calculation procedure . . . . . . . . . . . . . . . . . . . . . . . . 40
5.5 Spray cooling of live steam . . . . . . . . . . . . . . . . . . . . . 41

6 Validation 42

7 Discussion and analysis 44

7.1 Correction factors . . . . . . . . . . . . . . . . . . . . . . . . . . 45
7.2 Stack temperature decrease - first law . . . . . . . . . . . . . . . 45
7.3 Steam mean temperature increase - second law . . . . . . . . . . 46
7.4 Different number of pressure levels . . . . . . . . . . . . . . . . . 46
7.5 HRSG configuration . . . . . . . . . . . . . . . . . . . . . . . . . 46
7.6 Sources of error . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

8 Conclusions 48

9 Future work 49
9.1 Software implementation . . . . . . . . . . . . . . . . . . . . . . . 49
9.2 Improvement of correction factors . . . . . . . . . . . . . . . . . . 49
9.3 HRSG optimization . . . . . . . . . . . . . . . . . . . . . . . . . 49

6
Nomenclature

1P N RH Single pressure no reheat

2P N RH Two pressure no reheat
3P N RH Three pressure no reheat

T̄ Mean temperature
ṁ Mass flow
η2nd Second law efficiency
ηBC Bottoming cycle efficiency

ηcc Combined cycle efficiency

ηC Carnot efficiency
ηHRSG HRSG efficiency

ηSC Steam cycle efficiency

ηth,rev Internally reversible efficiency
ηth Thermal efficiency
BC Bottoming Cycle

BC Bottoming cycle
cp Specific heat at constant pressure
CCP P Combined cycle power plant

Cond Condenser
CV Control volume
CW Cooling water
E Exergy

7
exh Exhaust
f Fuel
GT Gas turbine
H Enthalpy

HP High pressure
HRSG Heat recovery steam generator
I Irreversibility

K Kelvin
LHV Lower heating value
LP Low pressure
N Ox Nitrogen Oxides

OEM Original Equipment Manufacturer

p Pressure
Q Heat

R Gas constant
S Entropy
SIT Siemens Industrial Turbomachinery
ST Steam turbine

stm Steam
T − q Temperature heat transfer diagram
TH Temperature High

TL Temperature Low
V Volume
W Work

8
1. Introduction

Finspång in Östergötland is known for its long history of industrial develop-

ment and production. For more than three centuries iron canons were produced
here and delivered abroad and to the Swedish armed forces. When the canon
production closed down in the early 20th century, the Ljungström brothers,
Birger and Fredrik, saw the possibilities of the available factories and favorable
location. They decided to locate their newly founded production company Sven-
ska Turbinfabriks AB Ljungström (STAL) in Finspång. In the year 1913 they
started building steam turbines in the empty workshops [1]. Their new radial
steam turbine construction were both smaller and more efficient than any other
turbine on the market at the time. Over 100 years later, a considerable number
of power producing machines have been sold worldwide. Since the start, the
company has changed many times and in 2003 Siemens took over as new own-
ers and founded Siemens Industrial Turbimachinery AB (SIT). As of today the
products being developed and sold in Finspång consist of a number of medium
sized gas turbines for industrial and electrical power production.

1.1 Background
The use of a single gas turbine (GT) for electrical power production leaves a
lot of potential energy unused, as a lot of heat is lost in the hot exhaust gases.
To take care of this loss it is common to use a steam turbine (ST) driven by
steam produced by the heat from the gas turbine. The combination of one or
more gas turbines with one or more steam turbines is known as a combined
cycle power plant (CCPP). The steam is produced in a heat recovery steam
generator (HRSG), basically consisting of a number of heat exchangers with
water on the cold side and exhaust gas on the hot side. Depending primarily on
the exhaust mass flow, exhaust temperature and the optimization of the HRSG
a specific amount of energy can be recovered from the hot exhaust gases. Since
no extra fuel is burned, i.e. the heat input to the combined cycle is the same
as for the simple cycle GT, the over all thermal efficiency ηcc can be increased
substantially.
In order to maximize the useful work produced per unit fuel supplied, it is
desirable to use the highest possible heat addition temperature and the lowest
possible heat rejection temperature. This is easily concluded from Carnot’s

9
Figure 1.1: Carnot efficiency as a function of TH , TL is kept constant at 273 K

theorem. If the temperature at which heat is supplied is approaching infinity,

and or the temperature at which heat is rejected approaches zero the Carnot
efficiency approaches unity, as seen in figures 1.1 and 1.2. The heat rejection
temperature is directly proportional to the Carnot efficiency, i.e. the incremental
decrease of the temperature always results in the same increase in efficiency.
This is not the case with the heat addition temperature, the gain in efficiency
levels out at a certain temperature, and the further increase of the temperature
will not result in a significant increase in efficiency.
In reality, where the machines operates, there are numerous limitations that
dictates the possibilities in choosing these temperatures. On the cold side it is
usually the plant location that determines the availability of heat sink i.e. a
river or sea close to the power plant. It is not economically viable to lower the
temperature readily available at the site, therefore there is not much that can
be done about the heat rejection temperature. On the hot side the primary
obstacles deterring an increase in temperature is limits relating to material
characteristics. Under operation the gas turbine rotor is exposed to extreme
conditions. The available materials needed to withstand the environment are
limited. Most metal alloys melting points are well below the temperatures that
a state of the art gas turbine first rotor blades are exposed to. With better
materials the temperature can be increased for better cycle efficiency, however
nitrogen oxide (NOx) formation is strongly depending on temperature which
effectively limits the temperature due to emission regulations.
For a given location with a given temperature in the heat sink, the only
way of reducing the environmental impact and cost is to increase the thermal
efficiency of the plant.[2] The gas turbine exhaust gases are always warmer than
the ambient temperature. Theoretically it is then possible to drive a heat engine
using the temperature difference between the hot exhaust gases and the cold
heat sink. By utilizing the otherwise rejected heat and producing more useful
work without increasing the fuel input the thermal efficiency of the plant is

10
Figure 1.2: Carnot efficiency as a function of TL . TH is kept constant at 1523
K

increased. The Rankine cycle is well suited to be combined with the Brayton
cycle using the exhaust gases as a source for the heat needed to generate steam.
CCPP can reach an overall thermal efficiency of above 63% in state of the art
power plants today.

1.2 Objectives
The objective of this project is to develop a method that can be used in a first
attempt to adequately and quickly determine the change in combined cycle ef-
ficiency of a power generating plant, when variation in key plant parameters
occurs. The objectives is also to present an explanation of the underlying mech-
anisms at work in the processes using fundamental thermodynamics with focus
on the application of the second law efficiency approach.

1.3 Approach
The approach taken in this project is to relate the thermal efficiency to the
Carnot efficiency i.e. using the second law efficiency. This makes it possi-
ble to produce a compact formulation that can be used to predict the change
in combined cycle efficiency, based on knowledge about the combined cycle effi-
ciency prior to the change in gas turbine performance. The second law approach
evaluates the different components performance compared to their theoretical
maximum, thus revealing the true potential of the component. [3] For a specific
power plant incorporating a combined cycle the total efficiency is commonly
well known to the plant owner. However it is not straight forward to predict
how the efficiency will vary if the GT performance is changed, e.g. if an upgrade

11
of the GT is considered. To determine this there is usually a need for extensive
and time consuming heat balance calculations. In this project a simpler and
more compact way to reach the same goal is presented and evaluated. The
second law efficiency greatly depends on the temperatures in the cycle. As the
exhaust temperature increases it will lead to an increase in both Carnot effi-
ciency and thermal efficiency, and the fact that the second law efficiency is the
ratio between these two efficiencies advocates the thesis that it is possible to
assume a constant second law efficiency with satisfying accuracy. The approach
is evaluated when a change in exhaust temperature and or mass flow occurs.
Four correction factors are introduced into the equation to reduce the errors in
the predictions. The combined cycle efficiency is dependent on a large number
of parameters. Depending on the specific situation the knowledge of the power
plant may differ, e.g. when an upgrade of GT performance is proposed by an
original equipment manufacturer (OEM) to a site owner. The OEM may have
limited information about the site data, but would like to be able to predict
the overall efficiency increase said GT upgrade would result in. It is in such a
situation the approach proposed in this report could be considered.

1.4 Previous work and literature study

Previous work closely related to the subject have been carried out by Gülen. In
reference [3] a second law approach is used to evaluate the efficiency of a Rankine
bottoming cycle of a combined power plant. In the exergy analysis a control
volume is placed around the bottoming cycle (BC) and the exergy supplied to
the system is simply the exergy in the exhaust gases from the topping cycle. The
exergy related to heat transfer out of the RBC is split between the condenser,
HRSG and the fuel preheater. The exergy destruction inside the control volume
is quantified in the HRSG, in the steam turbines and in the condenser. There
are also minor parts of exergy destruction related to electrical and mechanical
losses in the generator, shafts and bearings. Gülen derived suitable expressions
for the component irreversibilities in the bottoming cycle, these equations are
applied herein to clarify how different cycle parameters will affect the combined
cycle efficiency.
El Masri [4] concluded that the second law based approach to analyze com-
bined cycle power production can be useful for pinpointing and quantifying
losses when optimizing a new plant.

1.5 Tools
1.5.1 IPSEPro
The program that is used in this report is IPSEPro-PSE from SimTech Simu-
lation Technology GmbH. It is a heat and mass balance calculating and sim-
ulating program that quickly solves equations using matrix categorization and

12
then applying the Newton-Raphson method root-finding algorithm. The Pro-
cess Simulation Environment (PSE) is used together with the model developing
kit IPSEPro-MDK to build a representation model of a two pressure no re-
heat (2PNRH) combined cycle. This is done by starting from standard models
of each component and then recode the components so that every component
have the required variables e.g. irreversibilities or ambient temperature. This
model is then used to test the sensitivity of different components when varying
important input data. The results are then plotted together.

1.5.2 Newton-Raphson method

The working principle of the Newton-Raphson method is calculating the root
to a function on the form f (x) = 0 by providing an initial estimation xn , and
then comparing the function value to a tangent line to find a better estimation
xn+1 . The process is repeated until the difference between the estimations is
below a given tolerance i.e. the system has converged.

g(x) = f 0 (xn )(x − xn ) + f (xn ) (1.1)

f (xn )
xn+1 = xn − (1.2)
f 0 (xn )
The estimation xn+1 , found by intersecting the function g(x) and the x-axis,
is generally closer to the root than xn . An example of this can be seen in figure
1.3.

Figure 1.3: Representation of Newton-Raphson method

13
2. Theory

2.1 Heat engine

The thermodynamic definition of a heat engine is a device that can convert
thermal energy into mechanical energy which can be used to produce mechanical
work. The heat engine has been around since the antiques [5], but the concept
was first used on a wider front as a power source driving useful mechanical
loads in the era of the industrial revolution. The heat engine is often confused
with the theoretical cycle it operates according to, however the heat engine is a
physical machine and a thermodynamic cycle is the theory behind it.
The process of an arbitrary heat engine is driven by taking thermal energy
from a heat source and use it to heat up a working substance. The working sub-
stance can then be lowered down to a lower temperature level while mechanical
energy is extracted in the process. The state of the working substance is then
changed by transferring heat to a heat sink. It is the difference in state of the
working substance that is exploited to create a net work output from the engine.
During the process there are losses of energy to the surroundings because of
heat transfer through the system walls and also internally in the engine because
of friction. Naturally these irreversibilities lowers the efficiency of the engine.
There are a number of successful applications of different heat engines. The
Otto engine and the Diesel engine are just two of the most well-known machines
often used for mechanical drive of cars and ships. They both differ considerably
from the most common power generating engines, but all heat engines share the
aforementioned characteristics i.e. they receive heat from a high temperature
source, they convert part of the heat to work and they reject the waste heat to
a sink and they are all cycle operated.

2.2 Carnot cycle

The theoretical upper limit of a heat engine was first defined by the French en-
gineer and physicist Sadi Carnot in 1824 [6]. Carnot stated that any heat engine
operating between the same two temperatures can never achieve an efficiency
higher than the Carnot efficiency, (eq. 2.1) i.e. the Carnot efficiency dictates the
limit of possible work produced by providing the engine with a certain amount
of heat. Carnot’s principle became the foundation on which the second law

14
of thermodynamics were built upon. It can be said that the Carnot engine is
a physically equivalent of the second law. The Carnot Cycle is a theoretical
version of an internally reversible heat engine.
TL
ηCarnot = ηth,rev = 1 − (2.1)
TH
The closed cycle consists of four different steps. When these four steps have
all occurred the working medium will have returned to its original state, ergo
the cycle is complete.

2.2.1 Reversible isothermal expansion

The working medium is first expanded at a constant temperature TH thereby
doing work on the surroundings. When a gas is expanding the temperature
drops an increment dT , but since the process is very slow (quasi-static) the heat
provided by the energy source will have enough time to heat up the working
medium equally dT degrees, so that the temperature is kept constant. The total
amount of heat transfered to the gas during this step is QH , i.e. the energy input
from the heat source.

2.2.2 Reversible adiabatic expansion

At this point the energy input is removed, and is not in direct contact with the
working medium any more. The process is adiabatic i.e. no heat is transfered
out of the system. The working medium continues to do work on the surrounding
by expanding, thereby slowly lowering the temperature to TL .

2.2.3 Reversible isothermal compression

At the third step the working medium is connected to a heat sink with the
temperature TL . An external force is applied to the system, compressing the
gas so that the volume decreases and the pressure increases. The heat generated
in the compression step is transfered to the heat sink so that the temperature
is kept constant at TL .

2.2.4 Reversible adiabatic compression

In the final step the working medium is no longer connected to the heat sink. The
temperature increase from the continuing compression will then adiabatically
rise the temperature back to TH . The working medium have now returned to
the first state and the cycle can be repeated.

15
2.3 First and second laws of thermodynamics
2.3.1 First law
Originally the first law of thermodynamics was empirically developed over many
years of practice. It was not officially stated until the year 1850 when Ger-
man physicist and mathematician Rudolf Julius Emanuel Clausius restated Sadi
Carnot’s principle, the Carnot cycle [7]. His statement is known as the ”ther-
modynamic approach” to the nature of the relationship between heat and work.

”In a thermodynamic process involving a closed system, the increment in

the internal energy is equal to the difference between the heat accumulated by
the system and the work done by it.”

This can be expressed as the difference in internal energy of a closed sys-

tem which is always equal to the heat provided to the system minus the work
produced by the same system.

∆U = Q − W (2.2)
I can also be interpreted in the way that the total energy of an isolated
system is constant. [6] It is possible to change its form by transforming the
energy from one form into another, but it is not possible to create or destroy
energy.

2.3.2 Second law

”There have been nearly as many formulations of the second law as there have
been discussions of it.”
Philosopher and physicist P.W. Bridgman, (1941)

The first law of thermodynamics is dealing with the conservation of energy

for a system. The second law puts further limitations on the system by intro-
ducing a new state variable, entropy, and stating that the entropy of a closed
system is over time always either the same or increasing. The only case when
the entropy is constant is for a system undergoing a reversible process. How-
ever, such a process is physically impossible to achieve since irreversibilities are
always present in the system. To define the absolute entropy for a system in a
certain state the third law of thermodynamics is needed. However the absolute
value is not as interesting as the change in entropy. The change in entropy for
a system going from one state to another is defined in eq. 2.3.
Z 2
δQ
∆Ssys = S2 − S1 = + Sgen (2.3)
1 T
The Sgen term is the entropy generated because of system irreversibilities.
This term is zero in the reversible case. Key is that the entropy increase is
always related to a reference temperature.

16
 I
δQ
≤0 (2.4)
T
The relation famously known as Clausius inequality was first developed by
Clausius in the 1850s. Clausius proposed that ”A transformation whose only
final result is to transfer heat from a body at a given temperature to a body at
a higher temperature is impossible.”

2.4 Brayton cycle

The Brayton cycle has its name from American mechanical engineer George
Brayton [8] and is the theoretical model that describes a constant pressure heat
engine such as a gas turbine. A real gas turbine must be run as an open cycle
because of the exhaust gas composition, however the model is a closed cycle
process reusing the exhaust gases in the compressor intake. This enables closed
cycle calculations and analysis considering the engine as a closed system.

Figure 2.1: Representation of the Brayton cycle working principle [9]

Since there are many applications using the Brayton cycle a lot of research
and development have been carried out, resulting in a number of different ma-
chines using different layout to make the cycle better for a specific task. Inter-
rupting the expansion in the turbine and reheating the working medium to a
higher temperature will increase the work output from the cycle. This is because
of the fact that the work required to compress or expand the working medium is
proportional to the specific volume of the working medium. Since the working
medium can be considered an ideal gas, and the reheating is carried out under
constant pressure, following the equation of state (eq. 2.5) it is easily shown

17
that an increase in temperature will increase the specific volume of the working
medium.

pv = RT (2.5)
Reheating is also favorable in the sense that the maximum temperature of
the cycle can be kept at a level at which the nitrogen oxide (NOx) emissions are
complying with the regulations. While at the same time extracting more work
from the cycle.
Intercooling of the working medium between the compression stages will
have similar effect on the cycle performance, following the same logic. How-
ever, intercooling and reheating will not improve the thermal efficiency of the
cycle since intercooling will reduce the temperature at which heat is added and
reheating will increase the temperature at which heat is rejected. Following
equation (2.1) one can see that the Carnot efficiency will decrease. For single
cycle use reheat and intercooling is always used together with regeneration, i.e.
the exhaust gas heat is used to increase the temperature of the air prior to
compression. Lowering the need for heat input to the system.

2.5 Rankine cycle

The most common steam cycle using water steam as working medium is modeled
by the Rankine cycle. The cycle consist of four steps similar to the Carnot
cycle. However the Rankine cycle is a ”real” cycle in the sense that it takes into
consideration the real characteristics of the working medium.

Figure 2.2: Rankine cycle working principle

In the first step the condensed water is pumped up to a desired pressure.

This occurs with the water in fluid form. Following the same logic as in the
section for the Brayton cycle, it is very favorable to increase the pressure with
a low specific volume of the working medium. As seen in (eq.2.6 ) the specific
work required to increase the pressure of the fluid is directly proportional to the

18
specific volume and pressure difference. In the Rankine cycle all the compression
occurs when the working medium is in liquid form, this is why the compression
work is small.

wpump,in = v(P2 − P1 ) (2.6)

The water is then heated under constant pressure by adding heat from a
heat source. The heat source may consist of a boiler burning a suitable fuel,
or of a heat exchanger using the heat from a hot medium. At a temperature
dictated by the pressure the water will reach its saturation temperature and
begin to evaporate. When the all of the working medium is fully evaporated
the steam is usually superheated to a even higher temperature by continuing to
add heat from the heat source. The steam is then expanded in a steam turbine
and mechanical work can be extracted from the cycle by letting the pressure
and temperature drop. The fluid exiting the steam turbine is usually in the
mixed zone, with high enough steam fraction to prevent corrosion on the last
turbine stage. The mixture is then condensed back to liquid at constant pressure
transferring heat to a cold heat sink and the cycle is completed. The available
heat sink greatly affects the potential efficiency of the cycle. It may consist of
a nearby lake or an air cooled cooling tower.

2.6 Combined cycle

As the name suggests the combined cycle consists of a combination of two ther-
modynamic cycles, the Brayton cycle for the gas turbine system and the Rankine
cycle for the steam cycle system. As previously mentioned the reason for com-
bining the cycles is that the Brayton cycle exhausts gas at a temperature often
above 500◦ C, which makes it possible to drive the steam cycle using the energy
in the exhaust gas. The reference to the Brayton cycle as the ”topping cycle”
and the Rankine cycle as the ”bottoming cycle” is due to the cycles relative
positioning on the temperature entropy diagram, where the Brayton cycle is
situated above the Rankine cycle.
The combined cycle efficiency is the total work output divided by the heat
input of the topping cycle. This can be derived into eq. 2.7, where it is apparent
that an increase in GT efficiency not always leads to an increase in combined
cycle efficiency.

ηCC = ηGT + ηBC (1 − ηGT ) (2.7)

ηBC = ηHRSG ηSC (2.8)

19
 Figure 2.3: Combined cycle

2.7 Second law efficiency

2.7.1 Exergy
When considering a heat engine in a specific location with a defined heat source
and heat sink, the difference between total energy and useful energy is the
lost energy or waste energy of the system. The theoretical upper limit of the
useful energy that can be produced by the heat engine is the exergy of said
system. This is illustrated in figure 2.4. If all the processes in the system are
reversible, then the exergy equals the actual work produced by the system. This
is impossible in a real application, since there are always irreversibilities in the
system. But the concept gives an indication of how well the system is performing
in its environment, and the room left for improvements.
Wnet,out QL
ηth = =1− (2.9)
Qin QH
ηth
η2nd = (2.10)
ηth,rev

2.8 Carnot efficiency vs thermal efficiency

Applying an exergy balance on a Carnot cycle results in
in
EC = WC (2.11)

20
Figure 2.4: Graphic representation of total energy input to the bottoming cycle

Figure 2.5: Graphic representation of the exergy loss distribution, internal irre-
versibilities

Since
Z
in TL
EC = (1 − ( ))dQ (2.12)
TH
in
EC = ηC Q (2.13)
and there is no thermal exergy leaving the cycle since the temperature at
the exit is at the low reference temperature.
For the irreversible ”real” Brayton cycle the exergy balance equation yields
X
in Out CV
EB = WB + EB + IB (2.14)
Which makes the difference between eqs. (2.11) and (2.14)
X
in in
EC − EB = WC − WB − [E Out + I CV ]B (2.15)

21
 Leading to
Q Q
X
WB = WC − Iin − IOut − I CV ]B (2.16)
Where
Z 
Q in in TL TL 
Iin = EC − EB = − dQB (2.17)
T TH
and
Z 
Q Out TL 
IOut = EB = 1− dQA (2.18)
T
Q Q
The irreversibilities IOut and Iin are related to the temperature addition
to and rejection from the Brayton cycle taking place at a temperature differ-
ence rather than at a constant high temperature as is the case n the Carnot
cycle. And due to heat rejection to the atmosphere at a temperature above the
reference temperature. Thus the Brayton cycle efficiency is always lower than
the corresponding Carnot efficiency following equation 2.19. [10] An analogous
analysis can be performed both for the Rankine bottoming cycle as well as the
whole combined cycle.

(I Q + IOut
Q CV
P
WC WB ) IB
ηC − ηB = − = in + (2.19)
QB QB QB QB

2.9 Logarithmic mean temperature

Throughout this project all mean temperatures used are defined using the log-
arithmic mean. The first law gives the energy balance, i.e. the incremental flow
of heat and work of a process is balanced by the internal energy.

du = dq − dw (2.20)
According to the second law, for a reversible process the entropy generation
is equal to the incremental heat divided by the temperature.
dq
ds ≥ (2.21)
T

T ds ≥ dq (2.22)

dw = pdv (2.23)
Definition of enthalpy
h = u + pv (2.24)
Combining equation 2.20 to 2.24 gives the second law equation, or the Gibbs
equation.

22
 T ds − pdv = d(h − pv) (2.25)

T ds − pdv = dh − pdv − vdp (2.26)

dh = T ds + vdp (2.27)
Derivation of T̄ PN
i=1 ∆hi
T̄ = (2.28)
∆soverall
h3 − h2
⇒ T̄ = (2.29)
s3 − s2
From the Gibbs equation of entropy (2.25) one can derive the expression for
the incremental difference in entropy.
dh dT
ds = = cp (2.30)
T T
Where cp is the specific heat at constant pressure. The volume times pressure
difference part is zero for a theoretically isobaric heat addition. Integration of
eq. 2.30 gives the overall difference in entropy.

∆s = s3 − s2 (2.31)

Z 3
1
∆s = cp dT (2.32)
2 T
 
T3
∆s = cp ln (2.33)
T2
Inserted in Equation 2.29 gives the T̄ , logarithmic mean temperature de-
pending only on the two temperatures T3 and T2 .

cp (T3 − T2 )
T̄ =   (2.34)
cp ln TT32

T3 − T2
⇒ T̄ =   (2.35)
ln TT32

23
2.10 Exergy balance of the combined cycle
2.10.1 Exergy input
X T0

0 = ṁGT · (eexh − estck ) ± 1− · Q̇i − ẆBC − I˙ (2.36)
i
T̄i

When applying this exergy balance to the model created in IPSE the dom-
inating exergy input to the system is the mass flow related exergy leaving the
GT exhaust.
Ėin = ṁexh · eexh (2.37)
Where

V2
eexh = (hexh − h0 ) − T0 (sexh − s0 ) + + gz (2.38)
2
The subindex 0 refers to the dead state. The kinetic and potential exergy is
disregarded as it is negligible. This exergy input is known from GT calculations.
The other exergy input consists of the pumping work in the system. Since
the pumps are driven by electrical motors and not by steam turbines using
steam produced in the system, the total energy input to the pumps equals the
exergy input. However this exergy is subtracted form the generator output when
determining the cycle net output, and it is therefore unnecessary to take any
further measures to quantify the pumping exergy input.

2.10.2 Exergy out

The exergy leaving the BC is the net work output from the generator, the exergy
related to heat transfer out of the condenser and exergy related to mass transfer
out of the stack. The largest part is the net work output from the generator.
The stack exergy is strongly depending on the stack temperature and exhaust
mass flow.
 
Ėstck = ṁexh · (hstck − h0 ) − T0 (sstck − s0 ) (2.39)

The heat that is transfered away from the BC in the condenser could have
been used to drive a heat engine and produce work. This work equals the exergy
transfered, and is determined according to the equation below.
 
T0
Ėcond = Q̇cond · 1 − (2.40)
T̄CW
Where the cooling water mean temperature, T̄CW , is defined using the same
method as for the heat exchanger mean temperatures. The heat rejected from
the condenser is large in any steam cycle. However the exergy lost due to heat
transfer out of the bottoming cycle is not very large. This is because of the
relatively low temperature difference between the cooling mass flow entering
the condenser and leaving the condenser.

24
2.10.3 HRSG irreversibilities
For the complete HRSG the expression for the irreversibilities related to heat
transfer between the cold steam mass flow and the hot exhaust gases can be
expressed as eq. 2.44. In the exergy balance the heat supplied to the HRSG is
expressed as the available part of the heat output from the hot exhaust gases,
eq.2.42. Using the definition of exergy destruction in rate form the exergy
destruction in each of the heat exchangers in the HRSG can be determined.

Q̇
Ėdest = Ṡgen · T0 = · T0 (2.41)
T
 
TL
Q̇ = Q̇in · 1 − (2.42)
TH
As the temperature varies in the different parts of the HRSG an expression
for the mean steam and exhaust temperature is used to approximate the inlet
and outlet conditions of each heat exchanger.
The mean temperature is defined as:
h2 − h1
T̄ = (2.43)
s2 − s1
 
T0 T̄stm
I˙HRSG = Q̇in · 1− (2.44)
T̄stm T̄exh
This approach can be adapted to the individual heat exchanger in the model
for a better prediction of the total exergy destruction rate in the HRSG.

2.10.4 ST irreversibilities
The exergy destruction in the steam turbines is calculated using standard defi-
nitions. This is an easy procedure when the steam mass flow and entropy before
and after each turbine is known. The entropy difference is determined using the
ST isentropic efficiency. When another pressure level is introduced in the model
the calculation gets more extensive as more states needs to be calculated, but
the procedure is the same. When all the irreversibilities are calculated they are
added together as the total ST irreversibility.

I˙ST = ṁ · T0 · (sout − sin ) (2.45)

2.10.5 Pumping irreversibilities

Since the pumping work in the bottoming cycle is low and the pumps in the
evaporators are not increasing the pressure but are only circulating the fluid,
the entropy generation related to pumping is also low. And since the exergy
destruction is strongly linked to the entropy generation, it to is low.

25
 I˙pump = ṁ · T0 · (sout − sin ) (2.46)
The pumps in the model have an isentropic efficiency of 0.9.

2.10.6 Exergy balance

Looking back to the exergy balance equation 2.36 When the exergy destruction
and rejection is subtracted from the exergy input the remaining exergy matches
the net work output very well. For the 2PNRH there is a 0.3 percent difference
which is small considering that the temperatures used in the exergy destruction
model are mean temperatures.

26
3. HRSG theory

3.1 HRSG pressure levels

The heat recovery steam generator is an integral part of the combined cycle. The
optimization of the HRSG is heavily affecting the overall plant efficiency. The
increased performance of the HRSG that comes from adding another pressure
level originates from the better possibility to transfer heat to the steam from
the exhaust gases and by that lower the stack temperature and increase the
steam mean temperature. [11] By increasing the number of pressure levels the
influence of variation in exhaust temperature decreases. This is because of the
relative shift of LP economizer pinch point to the left in the T-Q diagram.

3.2 T-Q diagram

To visualize the HRSG performance it is common to use a diagram with energy
transfer on the horizontal axis and temperature on the vertical axis, known as
a T-Q diagram.

Figure 3.1: Schematic representation of a single pressure HRSG

27
 The GT exhaust is represented by the upper line in the diagram and the mass
flow is from hot to cold as the heat energy transfer to the colder steam mass flow
occurs. The steam mass flow is represented by the lower line. When the water
is evaporated the temperature of the fluid is constant, and the temperature of
the exhaust gas is decreasing, this leads to an increasing temperature difference
between the two lines. Theoretically this temperature difference could be used
to drive a Carnot engine thus increasing the energy produced by the system
and thereby also increasing the efficiency of the plant. In the evaporator the
difference in temperature between the cold side of the steam and the cold side
of the exhaust gases is known as the pinch point, illustrated in figure 3.1. This
is an important design parameter when optimizing the HRSG. If there were
no temperature difference between the two lines in the T-Q diagram then the
HRSG would operate at its theoretical maximum efficiency. However there
are physical limitations that makes constructing such a HRSG impossible in
reality. The most obvious reason for this being the evaporation of steam under
constant temperature, effectively making the lines diverge. A common way to
counter this effect is to use several pressure levels in the HRSG. By splitting the
steam into different parts generated at different pressures it is possible to reduce
the mass flow in each evaporator, which makes the constant temperature line
shorter. By increasing the number of pressure levels the efficiency of the HRSG
is increased, however each extra pressure level comes with its own evaporator
and often economizer and super heater as well, leading to a substantial cost
increase with each extra level. Combined with the diminishing effect on the
gain in efficiency by exceeding three pressure levels, HRSG:s with two or three
pressure levels is commonly used.
Another physical limit is the fact that to reduce the pinch points the heat
transfer area must be increased. Thus leading to increased cost both due to
more material being used and also due to larger HRSG footprint.

3.3 Stack temperature

The stack temperature is of vital importance in the thermodynamic evaluation
of the HRSG and thus the combined cycle. It basically dictates the portion of
the heat provided in the exhaust gases that is wasted to the surroundings. In a
HRSG the stack temperature is either determined by the pinch point in the LP
economizer [12] or by reaching the minimum allowed stack temperature due to
condensation of water from the exhaust gases. Damage to the economizer tubes
due to condensation is generally prevented by keeping the temperature above
a certain limit. However, this limit will not be active at full load nor in most
of the operational window unless the fuel contains large amount of sulfur. If
this is the case, the stack temperature will be lowered by an increase in exhaust
temperature. In the T-q diagram 3.2 the line representing the exhaust gases
will pivot around the fixed economizer pinch point.
As shown by the graph the stack temperature is reduced following an increase
in exhaust temperature. From the figure it is also clear that the decrease in stack

28
 Figure 3.2: Fixed economizer pinch point

temperature is always less than the increase in exhaust gas temperature. Since
the stack temperature is a result of the LP evaporator cold side temperature it
is the position of this point in the T-q diagram that dictates this difference.
When a second pressure level is introduced in the HRSG the optimum LP
boiler pressure level is being pushed down. This lower LP pressure leads to
the pinch point in the boiler being pushed to the left in the T-q diagram. In a
three pressure HRSG the decrease in stack temperature following an increase in
exhaust temperature would relatively be even smaller, since the LP pressure is
generally lower for a three pressure plant and hence the evaporation temperature
is lower.
Figure 3.2 presents a good visualization on the increase in steam mean tem-
perature when the exhaust temperature increases. It can be seen that when an
increase in exhaust temperature part of the LP steam is transferred to the HP
boiler, i.e. the HP boiler line in the T-q diagram is longer. Since the HP boiler
operates on a higher temperature this mass flow transfer leads to an increase in
overall steam mean temperature of the steam cycle, explaining the increasing
BC efficiency. This is since the increased steam mean temperature effectively
reduces the irreversibilities in the cycle, thus increasing the efficiency.

29
4. Methodology

In this section the methodology of the project is presented in order to describe

what has been done to accomplish the objectives. The models created in IPSE
are described in detail along with the assumptions and decisions made during
the creation process. The correlation between second law efficiency and com-
bined cycle efficiency is derived. The component irreversibilities dependence
on changes in exhaust temperature and mass flow is evaluated. The constant
second law efficiency approach is introduced, derived and improved upon by
developing the correction factors.

4.1 Model description

In this project numerous thermodynamic calculations were carried out based on
heat and work balance equations. This would not have been possible without
specific models of the systems that were examined. IPSEpro provides the pos-
sibility to build highly customizable models, including the possibility to rebuild
the different components, alter the governing equations and to program desired
features during the project. The resulting data from the cycle calculations could
then be exported to produce graphs and tables that can be presented.
Since the result of this project is going to be used to evaluate power plants
using different number of pressure levels in their bottoming cycles the equations
and approximations will have to be altered to fit the number of pressure levels
in the specific power plant. When the number of pressure levels increase it gets
harder to get good approximations. This is mainly because of the mass flow
being split into different streams in the HRSG and ST. Because of this, the
combined cycle efficiency is first determined for the single pressure bottoming
cycle, and then developed to work for two and three pressure levels.

4.1.1 1PNRH
The simplest model used in this project is the single pressure no reheat (1PNRH)
bottoming cycle. The combined cycle model used is a 1x1 configuration, i.e there
is only one gas turbine delivering heat to the steam cycle. This configuration
will stay the same throughout the project. In this model the HRSG consists of
three heat exchangers. One economizer, one evaporator and one super heater.

30
Figure 4.1: The 1PNRH model built in IPSEpro

31
 There are three pumps in the model, one to circulate the evaporator, one to
pump the condensate to the deaerator and one to increase the pressure of the
deaerated water before it enters the economizer. As previously mentioned the
pump work is very small compared to the generator output. This means that
the total pumping irreversibilities is small as well.
Although only one pressure level exists in the model, there are two steam
turbines. This is because of the bleed mass flow needed to the deaerator.
In this simple model no pressure losses are taken into consideration. The
HRSG is considered adiabatic, i.e no heat transfer to the surroundings are mod-
eled.

4.1.2 2PNRH
The two pressure no reheat cycle built in IPSE is similar to the one pressure
cycle in every way except for the additional low pressure boiler and added high
pressure economizer. The model has also been extended to include fuel preheat-
ing, extracting a small water mass flow after the high pressure economizer. With
this model it is possible to capture the irreversibility decrease in the components
since steam mass flow is transfered between the pressure levels. The model is
also expanded to include a temperature controller on the super heater to be able
to control the temperature of the hot steam entering the steam turbine. This is
done by spraying water from the HP economizer stream to the stream leaving
the HP boiler.

4.1.3 3PNRH
The three pressure model was built in IPSE for the purpose of calculating the
correction factors applicable to a combined cycle incorporating three pressure
levels.

4.1.4 Reheat
Reheat means heating up the steam leaving the first steam turbine to the same
temperature as before the first steam turbine, thereby increasing the mean steam
temperature of the BC and by that also increasing the efficiency. To accomplish
this another heat exchanger is introduced into the HRSG.

4.2 Exergy balance evaluation

To understand the exergy transfer in the combined cycle it is important to dif-
ferentiate between exergy destruction internally in the system and exergy losses
related to energy being transfered out of the system. This differentiation is
illustrated in figure 2.5. There are internal irreversibilities causing exergy de-
struction in the heat exchangers, condenser, steam turbines and pumps. These
components must be evaluated respectively and then added together. There is

32
also a miscellaneous loss to take care of small losses related to pipe friction and
valves.The model created in IPSEpro is used for finding an appropriate expres-
sion for the temperature or temperatures at which the exergy destruction in the
system occurs and how this temperature varies when the exhaust temperature
and mass flow is altered. It is therefore important to evaluate the exact exergy
destruction in the model, in order to know if the approximation is sufficiently
accurate. The model is evaluated using second law analysis of the different
parts. An exergy balance equation is used and the irreversibility rate from each
part of the system is summed up to the total cycle irreversibility.

4.3 Irreversibility change with exhaust temper-

ature
The bottoming cycle irreversibilities depends on the GT exhaust temperature.
This can be seen in the equations in the presented theory. The mean steam
temperature increases which leads to decreased irreversibility. Following an
increase in exhaust temperature the different parts of the bottoming cycle will
be individually changed, however the common trend is that an increased GT
exhaust temperature will result in lower bottoming cycle irreversibilities. The
graphs below represent the irreversibility of the major parts of the bottoming
cycle for a combined cycle with dual pressure HRSG.

Figure 4.2: Exhaust temperature in [◦ C] on the x-axis and irreversibility in [kW]

on the y-axis

33
 It is apparent that the pumping irreversibilities are to small to be paid much
attention in the analysis of the overall performance. It is also clear that the
HRSG irreversibility change is largest, strongly linked to the mass flow being
shifted from the LP to the HP boiler.

4.3.1 Why use the second law efficiency?

In figure 4.3 the second law efficiency, Carnot efficiency and bottoming cycle
efficiency is plotted as a function of exhaust temperature. The Carnot efficiency
increases since the hot temperature is the exhaust temperature, simply following
equation 2.1. The bottoming cycle efficiency is also increased as the exhaust
temperature increases. This is because of the higher heat addition temperature
in the bottoming cycle caused by the larger fraction of steam lifted to the high
pressure boiler. The second law efficiency is the ratio between the bottoming
cycle efficiency and the Carnot efficiency eq. 2.10. Since both the bottoming
cycle efficiency and Carnot efficiency increases the second law efficiency will
remain close to constant. This correlation is something that can be used when
relating the GT performance to the combined cycle efficiency.
However, due to material limitations in the steam turbine the superheater
temperature reaches its limit when the exhaust temperature rises, this leads to
the need for spray cooling to comply with this limit. The spray mass flow is
extracted after the HP economizer, at around 330 ◦ C and is mixed with the
superheated steam. This inevitably leads to a reduced second law efficiency
since the BC efficiency does not increase as much as the BC Carnot potential.
This can be seen in figure 4.3.

Figure 4.3: Efficiency variation with exhaust temperature

34
5. Derivation of correlation

5.1 Derivation of combined cycle efficiency equa-

tion
The combined cycle efficiency is the ratio of power generated per heat input.
PGT + PBC
ηCC = (5.1)
Qin
The only heat input to the system is the fuel entering the GT combustion
chamber. This is quantified as the mass flow of fuel multiplied by the lower
heating value LHV.

Qin = ṁf uel · LHV (5.2)

The power produced by the bottoming cycle can be related to the heat input
and the steam cycle efficiency. The heat input to the bottoming cycle equals
the heat leaving the topping cycle.

PBC = ṁGT · cp (Texh − Tstck ) · ηBC (5.3)

The thermal efficiency of the bottoming cycle can be expressed using the
second law efficiency and the Carnot efficiency.
 
TL
ηBC = η2nd 1 − (5.4)
TH

Combining the above equations one can rewrite the combined cycle effi-
ciency as follows. Where the Carnot heat engine represents the bottoming cycle
operating between the gas turbine exhaust temperature and the condenser tem-
perature.
 
Tcond
PGT + ṁGT · cp (Texh − Tstck ) · η2nd · 1 − Texh
ηCC = (5.5)
ṁf · LHV

35
5.2 Evaluation of combined cycle efficiency equa-
tion
Based on GT calculations most of the parameters in equation 5.5 is known. The
GT power, mass flow and exhaust temperature is known from these calculations.
The condenser temperature is treated as the ambient temperature plus a con-
stant depending on the cooling possibilities i.e. water cooling or air cooling.
The specific heat is found from gas tables since the composition is known from
GT calculations. The unknown parameters are the stack temperature and the
second law efficiency, these are the parameters that need to be treated sepa-
rately to get a good approximation. This is explained in the following sections.
The specific heat also needs some commenting.

5.2.1 Gas turbine exhaust temperature

Figure 5.1: Variation in GT exhaust temperature

In figure 5.1 the prediction of the combined cycle efficiency using equation
5.5 is presented when the exhaust temperature is varying in the range between
545 and 565 ◦ C. This interval is chosen since a 20 degree temperature increase is
considered reasonable to expect when an upgrade of the GT is implemented. The
reference point is where the exhaust temperature is 545, the second law efficiency
is calculated for that loadpoint and is then kept constant to illustrate how
this impacts the combined cycle efficiency. Information about the calculation
procedure can be found in section 5.4. The gray line represents equation 5.5 and
the black line is plotted with values from the heat and work balance calculations
in which the second law efficiency is free to vary for each point calculated. It
can be concluded that correlation can predict the combined cycle efficiency with
an accuracy in the order of 0.04 percentage points i.e. 0.02 percentage points
per 10 degrees increased exhaust temperature.

36
5.2.2 Gas turbine exhaust mass flow

Figure 5.2: Variation in GT exhaust mass flow

In figure 5.2 the variation on the x-axis is the increase in exhaust mass flow.
The gray line is equation 5.5 with the constant second law efficiency and the
black line is the heat and work balance calculation. Similarly to the exhaust
temperature variation the equation follows the calculation data with a slight
under-prediction for larger increases in exhaust mass flow. It is evident that
there is room for an increased accuracy in the prediction from the equation,
which leads to the introduction of the correction factors for the second law
efficiency and stack temperature.

5.2.3 Specific heat

The specific heat in the exhaust gas is depending on the gas composition, pri-
marily on the carbon dioxide content. This is in turn depending on the GT fuel
to air ratio. In equation 5.5 the specific heat will be unknown due to its depen-
dence on the stack temperature. When applying the equation the specific heat
is given by a gas table where the inputs will be the exhaust pressure, temper-
ature and composition. To get a good estimation a mean specific heat is used.
The mean value is found by estimating the stack temperature as the corrected
constant stack temperature (presented in section 5.3.1). This estimation leads
to an error, however it is small because the stack temperature will have limited
effect on the value from the gas table as seen in figure 5.3. Furthermore, the
stack temperature will not vary much compared to the exhaust temperature.

37
 Figure 5.3: Specific heat sensitivity to stack temperature prediction.

A sensitivity analysis was carried out to verify this. The corrected stack
temperature was increased by 10 percent to determine the impact on the spe-
cific heat, the result from a 10 percent over-prediction of the corrected stack
temperature was less than 0.0008 decrease of specific heat as can be seen in
figure 5.3. The same number for an under-prediction of the corrected stack
temperature by 10 percent. This is with good margin to the actual corrected
stack temperature.

5.3 Derivation of correction factors

To attain a working correction factor it is imperative to find out how the second
law efficiency and stack temperature deviates from their original values, and to
separately allocate the change of these parameters to the change in GT mass
flow and exhaust temperature. This has to be done carefully since all parame-
ters are strongly depending on each other. This is well described by Gülen in
ref. [3]. The method applied in this project is to use the IPSE model and the
model development kit to state an equation that represents the combined cycle
efficiency using only varying parameters, i.e. no constants or approximations.
Another equation is then stated in which the second law efficiency is kept con-
stant. The first correction factor is inserted in the equation. Then a variation
of the GT parameter for said correction factor is made in the IPSE model. The
values of both the exact equation and the equation with the constant second
law efficiency and correction factor are collected in a vector. The value of the
correction factor will then vary with the GT parameter to correct the efficiency.
The values are plotted against the GT parameter and since it is linear, both for
the exhaust gas temperature and the mass flow, a linear equation can be created
from the values. The interesting part is the coefficient in front of the variable,
because it is a measure of how much the parameter influences the efficiency.
The sensitivity of the second law efficiency correction factors was tested by

38
changing the steam cycle parameters. The dominating parameter affecting the
correction factors was found to be the number of HRSG pressure levels, this
has been pointed out earlier in the report. The change in pinch points for the
evaporators made no significant impact on the correction factors at all.

5.3.1 Correction factors

The correction factors are derived using the IPSE-Pro models of the system.
There are four factors in the equation, one each for correction of the second law
efficiency when exhaust mass flow and temperature is changed, and one each
for correction of the stack temperature when the same change is made. All
four correction factors are determined using the same method. Since both the
second law efficiency and the stack temperature varies close to linearly when the
exhaust parameters are changed, they are a dimensionless number that follows
a linear equation with the change in the exhaust parameter as a variable.

xη∆ṁ = 1 − 0.00009176 · ∆mGT (5.6)

This correction factor is introduced to compensate for the change in second law
efficiency when the exhaust mass flow is changed.

xη∆T = 1 − 0.0003285 · ∆T (5.7)

This correction factor is introduced to compensate for the change in second law
efficiency when the exhaust temperature is changed. These factors will correct
the decrease in second law efficiency when a change of GT mass flow and or tem-
perature occurs. It can be seen that the decrease is rather small. The coefficient
basically says that to obtain a good combined cycle efficiency following a change
in exhaust temperature and mass flow it is necessary to lower the second law
efficiency by 0.0003285 times the change in exhaust temperature and 0.00009176
times the change in exhaust mass flow. This again agrees with the thesis that
the second law efficiency is essentially independent of these parameters. Due to
the steam turbine temperature limit the live steam is usually sprayed down to a
set temperature by bypassing some of the HP boiler mass flow to the HP super
heater. If the change in exhaust temperature were to decrease, this spray mass
flow would be reduced down to a certain point after which it would be shut of
completely. At this point the correction factor for exhaust temperature effect
on second law efficiency is no longer needed.

xT ∆ṁ = 1 + 0.00037427 · ∆mGT (5.8)

This correction factor is introduced to compensate for the change in stack tem-
perature when the exhaust mass flow is changed. In figure 3.2 an increase in
exhaust temperature can be seen, and how this affects the stack temperature
and second law efficiency.

xT ∆T = 1 − 0.00022203 · ∆T (5.9)

39
This correction factor is introduced to compensate for the change in stack tem-
perature when the exhaust temperature is changed. These factors will take care
of the change in stack temperature when a change in GT mass flow and or tem-
perature occurs. All four factors are inserted in equation 5.5. Which leads to
the final form of the equation. 5.10.

 
Tcond
PGT + ṁGT · cp (Texh − xT ∆T · xT ∆ṁ · Tstck ) · xη∆T · xη∆ṁ · η2nd · 1 − Texh
ηCC =
ṁf · LHV
(5.10)
The divergence of the lines in fig. 5.1 depends on the second law efficiency
being kept constant. As stated before, the second law efficiency is nearly con-
stant in the equation. The correction factor is introduced in order to get an
even better prediction of the combined cycle efficiency. The correction factors
presented in this part of the report is suitable for a dual pressure HRSG, in table
the correction factors for other systems can be found. The stack temperature
and both correction factors related to it is the corrected stack temperature.

Tstck,corr = xT ∆T · xT ∆ṁ · Tstck (5.11)

The second law efficiency with the related correction factors is the corrected
second law efficiency.

η2nd,corr = xT ∆ṁ · xη∆ṁ · η2nd (5.12)

5.4 Calculation procedure

Starting from equation 5.5 the data from a power plant where the combined
cycle efficiency is known is inserted. If the stack temperature is not known
an estimate is made based primarily on the number of pressure levels in the
HRSG. It is suitable to hard-code this estimation into the software used when
implementing this method, making it simple for the user since there is no need
for experience to make the estimation. As a rule of thumb a good estimation
would be 110 ◦ C for a 1 pressure HRSG, 95 ◦ C for a 2 pressure HRSG and 80 ◦ C
for a 3 pressure HRSG. The specific heat is found from gas tables, this is done
using the known exhaust gas composition, exhaust pressure and temperature.
The mean value is used for the specific heat, i.e. the value of the specific heat is
calculated both for the exhaust temperature and the stack temperature and then
averaged. The equation can then be solved for the second law efficiency. When
this is done equation 5.10 is applied with the upgraded GT performance data
along with the correction factors. The correction factors are calculated based on
the difference in exhaust gas temperature and exhaust mass flow, all according
to their respective equation presented previously in the report. Both the stack
temperature and the second law efficiency is kept constant and is adjusted by
the correction factors. This results in a new combined cycle efficiency which

40
can be compared to the old value to determine the difference. No iteration is
needed.

5.5 Spray cooling of live steam

In cases where the exhaust temperature decreases after the upgrade it is favor-
able to leave the correction of the second law efficiency with regard to exhaust
temperature out of the equation.

Figure 5.4: Spray mass flow influence on correction factor for second law effi-
ciency

In figure 5.4 the value of the correction factor is presented as the exhaust
temperature decreases. When the exhaust temperature decreases it will even-
tually pass the point where spray cooling is no longer needed since the steam
temperature is below the steam turbine limit without cooling. This is why
there is a change in inclination. If the correction factor is not removed from the
equation it will lead to an over-prediction, which is clearly seen in the figure.
Following the assumption that the spray mass flow is designed to be low at the
design point in the HRSG, it is likely that the spray mass flow reaches zero after
a small decrease in exhaust temperature. This leads to the suggestion that said
correction factor should be removed for any decrease in exhaust temperature,
even if this is not necessarily true for all HRSG designs.

41
6. Validation

The equation was validated against data from calculations of an upgrade project
of a Siemens combined cycle consisting of two gas turbines and one steam tur-
bine. The prediction of the combined cycle efficiency difference when compared
to the data was found to be in the low hundredth of a percent, which makes for
a close enough approximation for a first quick calculation. To put the results
in context the approach proposed in this report was compared to the approach
of keeping the BC thermal efficiency constant in the prediction of combined cy-
cle efficiency change. The thermal efficiency was calculated from eq. 2.7 using
the same data from Siemens upgrade project. This showed that the prediction
from the second law approach is closer to the actual efficiency change by over
30 percentage points. By that the objectives can be considered to be achieved.
To perform a further statistically reliable validation, more site data would be
needed.
Validation with Siemens gas turbine fleet was carried out. Four different
scenarios were examined using existing GT models. An increase in turbine inlet
temperature by 25 and 50 degrees, an increase of compressor mass flow by 5
percent and an increase in diffuser recovery by 10 percent. The prediction result
is presented in the table below, as deviation from the combined cycle efficiency
calculated by the model.

In the cases where the exhaust temperature decreases the correction factor

42
for the second law efficiency related to exhaust temperature difference was not
included in the equation, as explained in section 5.5. The error from predicting
the combined cycle efficiency without the correction factors are presented. It
can be seen that the correction factors makes the prediction considerably more
precise in cases where the exhaust temperature changes.
The comparison in precision in the table is from equation 2.7 where the
thermal efficiency is kept constant to predict the combined cycle efficiency after
an upgrade. It can be seen that the second law approach is far more precise in
the prediction.

43
7. Discussion and analysis

The results from the project i.e. the derived combined cycle efficiency equation
and especially the understanding of what affects its precision implies that it is
indeed possible to implement the approach when predicting the combined cycle
efficiency change.
The ambition of this project was to find a way to quickly predict the com-
bined cycle efficiency increase. With focus on the simplicity of the method in
order to increase the speed of the calculation process and to make it easier to
perform an estimation on the combined cycle efficiency increase without the
need for extensive experience from heat and work balance calculations. Ini-
tially the potential approach of using exergy destruction calculations to predict
the combined cycle efficiency was evaluated. However, it was concluded that
such an approach would not effectively reduce the complexity of the calcula-
tions, i.e. it presented no substantial reduction of equations and heat and work
balance calculations needed to obtain the result compared to the conventional
procedure. Thus the exergy balance approach was rejected in favor of the more
straightforward equation derived in this report.
When the first equation (eq. 5.5) describing the combined cycle efficiency
prediction was derived it was evaluated by comparing the predictions of the
equation with heat and work balance calculations. The heat and work balance
models described in section 4.1 were created to this purpose. Different param-
eters in the model was changed to find out which parameters made the largest
impact on the prediction. It was concluded that when the GT exhaust temper-
ature and mass flow was changed, it made the prediction less accurate when the
stack temperature and second law efficiency was kept constant. The sensitivity
of the estimation of GT exhaust gas mean specific heat was examined since
the stack temperature after the upgrade is unknown and this affects the mean
specific heat in the equation. With knowledge about this sensitivity the use of
the corrected stack temperature in the gas table when estimating the specific
heat is the best approach, considering generality.
One challenge was to find a good balance between applicability and gener-
ality. Based on the fact that a large number of parameters affect the combined
cycle efficiency it is extremely difficult to produce a compact formulation that
can be used for the many different bottoming cycles that are in operation world
wide. Every assumption made in the model that is used to produce the correc-
tion factors inevitably leads to sacrifices in the applicability of the equation and

44
since the end result, i.e. the prediction of the combined cycle efficiency, has to
be fairly close to the correct value in order to be at all useful it makes the whole
project rather delicate.

7.1 Correction factors

When a change in GT performance is considered e.g. an upgrade of the GT
is proposed which will result in a change of the parameters involving the bot-
toming cycle, there are limitations in the system that dictates the maximum
possibility of changing certain key parameters. That is, to use the combined cy-
cle efficiency equation for a prediction in efficiency change, the equation needs to
be altered further since the input parameters are not individually independent
of each other, this leads to the need for the correction factors to compensate
for both stack temperature and second law efficiency being kept constant. The
correction factors will compensate the slight alterations in second law efficiency
and stack temperature when a change in exhaust mass flow or temperature is
made. To begin with, two correction factors were introduced into the equation,
one for the stack temperature and one for the second law efficiency. It was
then concluded that the change in exhaust temperature and mass flow affects
the stack temperature in opposite ways, an increase in exhaust mass flow will
increase the stack temperature and an increase in exhaust temperature will de-
crease the stack temperature. This leads to the need for another correction
factor for the stack temperature. Since the prediction must be applicable to
different GT upgrades, and since there is no general connection between the
change in exhaust parameters there has to be a separate correction factor for
the change in exhaust temperature and mass flow to correct the constant second
law efficiency. This led to the final expression for the equation used to predict
the efficiency change, eq. 5.10.

7.2 Stack temperature decrease - first law

The increase in exhaust temperature will result in a decrease of stack tem-
perature. This will increase the combined cycle efficiency, which is easy to
understand applying the first law of thermodynamics. The heat captured by
the HRSG increases, leading to an increased power output from the ST without
increasing the fuel consumption of the machine. This can be seen in figure 3.2,
the total heat input into the HRSG increases both because of the exhaust tem-
perature being higher and the stack temperature being lower. The fuel input to
the combined cycle is unchanged hence the efficiency increases.

45
7.3 Steam mean temperature increase - second
law
From a second law point of view the increase in combined cycle efficiency comes
from the increased steam mean temperature, i.e. the increased exhaust temper-
ature results in a larger HP mass flow, effectively shifting the LP boiler steam
to the HP level. This effect is also illustrated in fig. 3.2. The HP boiler operates
on a higher temperature compared to the LP boiler, thus increasing the mean
steam temperature. The higher mean steam temperature effectively reduces the
irreversibilities of the system and consequently reduces the total internal losses,
leading to an increased overall efficiency of the combined cycle.
Regarding the reasoning behind the use of a second law approach, this is
primarily because it presents a powerful tool when comparing different alterna-
tives of components in a power plant. While the results are always in agreement
with the first law of thermodynamics, the major advantage of using the second
law approach is that it provides a good representation of which components to
target when improving the power plant. Furthermore the second law approach
also contributes to understanding the mechanisms of the bottoming cycle and
how the components reacts to changes in the cycle, since it comprehensively
quantifies the irreversibilities. Furthermore, the equations describing the irre-
versibilities presented in the theory that are derived by Gülen in ref [3] can
individually or added together be used to understand how the change in various
parameters will affect the total internal losses in the systems.

7.4 Different number of pressure levels

The need for a set of correction factors derived for the correct number of pres-
sure levels is clear when following the reasoning about applicability. The stack
temperature does not vary in the same way for a dual pressure HRSG as for a
triple pressure one, as can be seen in figure 3.2. This is the main reason behind
the need for different correction factors when considering steam generators with
different number of pressure levels.

7.5 HRSG configuration

The correction factors are developed using a model with a defined HRSG con-
figuration. When applying the equation 5.10 to a power plant this assumes that
the HRSG configuration is similar to the one created in the model. If the HRSG
of the plant differs from the modeled one that could lead to a source of error in
the prediction. The material limitations of the steam turbine leads to the need
for spray cooling of the HP super heated steam to control the temperature. As
previously mentioned this affects the second law efficiency since the increase is
hindered due to the temperature control. This aspect is something that need
to be considered by the user.

46
7.6 Sources of error
The correction factors are straight line equations, simply to keep them compact.
This ultimately leads to an introduction of error since the model calculated
values marginally deviates from this straight line. This is once again a balance
between compactness and precision of the equation. To capture the deviations
from the straight line equation would require the introduction of a significantly
higher order polynomial equation.
There are several controllers in the model used to create the correction fac-
tors. The HP SH spray being of major importance since it introduces nonlinear-
ities into the system. There is also a bleed steam controller to the deaerator. If
the change in GT performance is changed passed the limits in these controllers
the system responds nonlinearly making the linear correction factors less precise
leading to a decrease in prediction accuracy.

47
8. Conclusions

With knowledge about the site prior to an upgrade i.e. known stack temperature
and combined cycle efficiency the equation can predict the efficiency change
after the upgrade within a couple of hundredth of a percent following an exhaust
temperature increase by 15 ◦ C. This is without the need for any iteration. When
the correction factors were updated to fit the pressure levels of the validation
site the prediction error was reduced by more then half, down to just above one
hundredth of a percent on the combined cycle efficiency.

48
9. Future work

9.1 Software implementation

The scope of this project is limited to the presentation of a method of predicting
the change in the combined cycle efficiency for a power generating plant. To
make use of the proposed method it will have to be implemented in a suitable
software.

9.2 Improvement of correction factors

Improvement of the correction factors to better predict efficiency change for a
specific HRSG configuration. This is preferable to consider in the case that
more information exists about the specific site targeted.

9.3 HRSG optimization

The results presented in this report could also be of use in the process of de-
signing a new HRSG and efficiently optimizing it for combined cycle efficiency
of the plant.

49
Bibliography

[1] Siemens AG. Historia. www.sit-ab.se/01_historia.html.Accessed:

2018-03-23., 2018.

[2] A.M.Y. Razak. 2 - thermodynamics of gas turbine cycles. In A.M.Y. Razak,

editor, Industrial Gas Turbines, pages 13 – 59. Woodhead Publishing, 2007.
[3] S Can Gülen and Raub W. Smith. Second law efficiency of the rankine
bottoming cycle of a combined cycle power plant. 132, 01 2010.

[4] M. A. El-Masri. On thermodynamics of gas-turbine cycles: Part 1 — second

law analysis of combined cycles. Journal of Engineering for Gas Turbines
and Power, 10 1985.
[5] Bunch B Hellemans A. The Timetables of Science: A Chronology of the
Most Important People and Events in the History of Science. 1991.

[6] Boles M Cengel Y. Thermodynamics: an engineering approach, 7th edn.

2010.
[7] C. Truesdell. The Tragicomical History of Thermodynamics, 1822–1854.
1980.

[8] C Borth. Masters of mass production. Bobbs-Merrill Co., 1945.

[9] H.I.H. Saravanamuttoo, G.F.C. Rogers, and H. Cohen. Gas Turbine The-
ory. 50 years. Prentice Hall, 2001.
[10] Horlock J.H. Advanced gas turbine cycles. 2003.

[11] Shoaib Khanmohammadi and Ahmad Reza Azimian. Exergoeconomic eval-

uation of a two-pressure level fired combined-cycle power plant. Journal of
Energy Engineering, 141(3):1 – 13, 2015.
[12] M. A. El-Masri. Design of gas turbine combined cycle and cogeneration
systems - theory practice and optimization. Seminar notes, Thermoflow,
Inc., Sadbury, MA, 2007.

50