Student Magnus Gule Dynamic process simulation of thermal power plant on offshore oil and gas installations Trondheim, March 11th, 2016 Project work NTNU Norwegian University of Science and Technology Faculty of Engineering Science and Technology Department of Energy and Process Engineering
Welcome message from author
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
Student Magnus Gule
Dynamic process simulation of thermal
power plant on offshore oil and gas
installations
Trondheim, March 11th, 2016
Pro
ject w
ork
NT
NU
Norw
egia
n U
niv
ers
ity
of
Scie
nce a
nd T
echnolo
gy
Faculty o
f E
ngin
eering S
cie
nce a
nd T
echnolo
gy
Depart
ment of E
nerg
y and P
rocess E
ngin
eering
1
2
Preface
This project was written autumn 2015 at NTNU, Department of Energy and Process
Engineering. I would really like to thank my supervisor Lars Olaf Nord for guidance
and motivational boost during my work. I’d also like to thank Rubén Mocholí
Montañés for technical guidance and advising during the modelling and simulation
process.
3
Abstract
Offshore installations represent today one of the largest demands related to electrical
power consumption on the Norwegian continental shelf (NCS). To cover the supply
almost all platforms are powered by one or more simple gas turbines, which emits
approximately 80% of all greenhouse gasses produced in the offshore petroleum
sector. Due to increases in CO2-taxation set by the Norwegian Government,
companies are currently looking for alternatives like the combined cycle to reduce
emissions and increase efficiency on the gas turbines.
During early 1999 and 2000, a set for three combined cycles were installed on the
Norwegian continental shelf, and have later been upgraded by newer heat recovery
steam generators (HRSGs). Due to harsh weather conditions, both weight and
volume limitation, and the need for variable power demand, makes further retrofit
installations of CCs challenging.
There exist few literature studies on transient operation conditions for onshore
combined cycles, and close to none for offshore platforms. The few studies
investigated primary concern triple-pressure onshore combined cycle power plants
(CCPPs) with drum-based HRSGs. Newer offshore HRSGs are based on drum-less
once-through steam cycles is therefore important to study.
A comparative literature study of different HRSG-skids was evaluated and different
regulation techniques for part-load of the steam cycle investigated. Based on this a
dynamic model was built in Dymola based on pre-simulated steady-state data
resembling the Oseberg D combined cycle. The HRSG-module was parameterized
using two-phase counter-current heat exchangers from the Thermopower library and
calibrated to the original data.
The preliminary model shows expected heat transfer behavior, but initially show
large oscillations and maintain unstable startup conditions especially regarding high
pressures. This is explained through rigid boundary conditions for the current model.
Implementation of valve control and pumps to the model is vital to further work and
to understand the real behavior of the combined cycle.
4
Contents
Preface ......................................................................................................................... 1
Abstract ....................................................................................................................... 3
Contents ...................................................................................................................... 4
List of figures and tables .............................................................................................. 7
Acronyms and Abbreviations ....................................................................................... 9
Nomenclature ............................................................................................................. 10
1 Introduction ........................................................................................................ 11
1.1 Background and motivation .......................................................................... 12
1.2 Thesis objective and outline ......................................................................... 13
1.3 Limitations of work....................................................................................... 14
1.4 Further work ................................................................................................ 14
1.5 Literature study ............................................................................................ 15
2 Combined Cycles Power Plants – An introduction ............................................. 15
3 The components of a CCPP ............................................................................... 17
3.1 The gas turbine............................................................................................. 18
3.2 The steam cycle ............................................................................................ 19
3.3 HRSG ........................................................................................................... 20
3.3.1 Finned tube heat exchanger ................................................................... 21
3.3.2 Sections in the HRSG ............................................................................ 23
3.4 Water treatment ........................................................................................... 25
4 Onshore and offshore BCC ................................................................................. 27
4.1 Existing CC platforms .................................................................................. 27
4.2 Eldfisk ........................................................................................................... 28
4.2.1 Replacement of HRSG ........................................................................... 29
4.3 Oseberg D ..................................................................................................... 30
5
4.3.1 Dual OTST replacement in 2010 ............................................................ 32
4.4 Snorre B ....................................................................................................... 32
4.5 Backup power on offshore platforms ............................................................. 32
4.6 Challenges for further CC offshore................................................................ 33
5 HRSG designs and configurations ....................................................................... 35
5.1 Drum based HRSG ....................................................................................... 35
5.2 The influence of pressure levels ..................................................................... 37
5.3 Once Through Steam Generator ................................................................... 39
5.4 Development of compact HRSGs .................................................................. 41
5.5 Circular Steam Generator (CSG) ................................................................. 42
5.6 Differences - Drum and Once Through Boilers (rydd) .................................. 43
6 Control aspects and regulation ........................................................................... 47
6.1 Drum level regulation ................................................................................... 47
6.1.1 Shrink and swell ..................................................................................... 47
6.1 Off-design and partial arc control ................................................................. 49
6.1.1 Part load regulation ............................................................................... 51
6.1.2 Partial arc .............................................................................................. 52
6.2 Bypass stack flow ......................................................................................... 53
6.3 Supplementary firing .................................................................................... 54
6.4 Startup and shutdown of CC ........................................................................ 54
7 Dynamic and Steady-state modelling .................................................................. 55
8 Modelling with Dymola/Modelica ....................................................................... 56
8.1 Working with ThermoPower ........................................................................ 56
8.2 Fundamental equations ................................................................................. 58
8.3 Metal wall model .......................................................................................... 59
8.1 Discretization in FEM .................................................................................. 60
6
8.2 Approach to model and workload ................................................................. 61
8.3 HRSG build-up ............................................................................................. 62
8.4 Challenges of parametrization ....................................................................... 64
8.5 Pressure drop calculations ............................................................................ 66
8.6 Calibrating the HRSG heat transfer ............................................................. 67
9 Evaluation of model ............................................................................................ 70
10 Results of preliminary tests ................................................................................ 72
10.1 Warm start-up ramp: 5 min ...................................................................... 72
10.2 Part-load: GT ramp 80% to 100% and back ............................................. 76
11 Review of work ................................................................................................... 80
12 Conclusion .......................................................................................................... 81
13 Appendix ............................................................................................................ 83
14 References ........................................................................................................... 90
7
List of figures and tables Figure 1: Ts-diagram of the combined cycle. Brayton as top cycle, and Rankine as bottom cycle. ...... 16
Figure 2: Schematic illustration of a dual-pressure CCPP with steam drums. ...................................... 18
Figure 3: General Electric LM2500+G4 gas turbine. [8] ........................................................................ 19
Figure 4: The HRSG units of 660MW land-based combined cycle power plant (CCPP)[9] . ................ 21
Figure 5: Right: High pressure supeheater fin configuration of Oseberg D. Left: Illustration from
frbiz.com [10] .......................................................................................................................................... 22
Figure 6: Schematics of vertical HRSG module. The illustration shown gives the exact number of
pipes, and thermodynamic values during steady-state operating. (Edited figure by source: Thermoflow-
data by Lars O. Nord [14]) ..................................................................................................................... 24
Figure 7: TQ-diagram of single pressure HRSG ..................................................................................... 25
Figure 8: Illustration of the flows inside a deaerator [16] ....................................................................... 26
Figure 9: Skid of the Eldfisk combined cycle plant from 1999 [4] .......................................................... 28
Figure 10: Oseberg D 1999 steam power cycle, drum based HRSG [4] .................................................. 31
Figure 11: Model of OTSG module installed for Oseberg D 2010. Coutesy of Macchi [21] .................... 31
Figure 12: Comparison of weight reduction of drum-based HRSG versus OTSG [24]. Note that wet
weight axis starts at 100 tons. ................................................................................................................ 34
Figure 13: Size comparison between different HRSG skids (Courtesy of HRS [19]) .............................. 35
Figure 14: Courtesy of ISA.org [25] ........................................................................................................ 36
Figure 15: A typical layout of a single-pressure CCPP, with corresponding TQ-diagram [6] ................ 37
Figure 16: Triple pressure level HRSG with reheating. .......................................................................... 38
Figure 17: (right) TQ-diagram of Dual pressure level HRSG. A dual-pressure combined cycle with
reheating (Reheating is the last part on the hot end) [6] ....................................................................... 38
Figure 18: Efficiency of different CCPP configurations without parasitic or step-up. Heat to power
effiency. [6] .............................................................................................................................................. 39
Figure 19: Simplified CC for a once-through steam generation (OTSG) system. Illustration taken from
Jonshagen et.al. [6] ................................................................................................................................. 40
Figure 20: Comparative sizing of traditional onshore drum-based HRSG and newer OTSG. [29] ......... 41
Figure 21: Circular HRSG concept by HRS [19] .................................................................................... 43
Figure 22: Circulations of water inside Vertical OTSG and Drum-based HRSG. (Courtesy of NEM
Group [32]).............................................................................................................................................. 44
Figure 23: Courtesy of Milton R. Beychok [33] ...................................................................................... 44
Figure 24: Drum based skit to the left; OTSG to the right. Courtesy of IST [27] ................................. 45
Figure 25: Measurement of drum level. [25] ........................................................................................... 48
Figure 26: Measurement using gage glass ............................................................................................... 49
Figure 27: Sliding pressure opeation [35] ................................................................................................ 50
Figure 28: Illustraiton of a partial-arc inlet [6] ....................................................................................... 52
Figure 29: Bypass duct with integrated silencer for Vertical HRSG. Source: Courtesy of HRS [19] ..... 53
8
Figure 30: Finite element method (FEM) 1D analysis schematic diagram [42] ..................................... 58
Figure 31: HRSG build-up. Components inside are numbered and described in table. ......................... 63
Figure 32: Definition of different exchange surfaces in the HE2ph model .............................................. 65
Figure 33: Single HE validation with seperate pressure drop module .................................................... 67
Figure 34: HRSG module for calibration ................................................................................................ 69
Figure 35: Preliminary model for simulation. ......................................................................................... 70
Figure 36: Semi-stable model of the Oseberg D plant, including pumps and pressure-control thorugh
valves. ..................................................................................................................................................... 71
Tables:
Table 1: HRSG specification for Oseberg D fins and tubing .................................................................. 22
Table 2: Table showing the details about the components inside the HE2ph ........................................ 62
Table 3: Average HTC values for a HRSG............................................................................................. 68
9
Acronyms and Abbreviations
BCC Bottoming Combined Cycle
CC Combined Cycle
CCGT Combined Cycle Gas Turbine
CCGT Combined-Cycle Gas Turbine
OTSG Once-Through Steam Generator
CCPP Combined Cycle Power Plant
CSG Circular Steam Generation
DAEs Ordinary Differential-Algebraic Equations
FEM Final Element Methods
FG Flue Gas
FPSO Floating Production Storage and Offloading Unit
GHG Greenhouse Gases
GT Gas Turbine
GWP Global Warming Potential
HE Heat Exchanger
HP High pressure
HPB High Pressure Boiler
HPE High Pressure Economizer
HPS High Pressure Superheater
HRSG Heat Recovery Steam Generator
HTC Heat Transfer Coefficient
LTE Low Temperature Economizer
NCS Norwegian Continental Shelf
NCS Norwegian Continental Shelf
PDAE Partial Differential and Algebraic Equations
WHRU Waste Heat Recovery Unit
10
Nomenclature
𝜙𝑖 heat flux enetering pipe across lateral surface [W/m2]
ℎ specific enthalpy [J/kg]
𝜔 wetter perimeter [m]
𝐶𝑓 Fanning friction factor [-]
𝜌 density [kg/m3]
𝑤 massflow [kg/s]
𝐾𝑓 Hydraulic friction coefficient [-]
𝐾𝑓,𝑐 Friction factor correction coefficient [-]
𝑔 acceleration of gravity [m/s]
𝑝 pressure [Pa]
𝐴 area [m2]
𝑡 time [s]
𝑥 1-dimentional direction [m]
𝑇 temperature [K]
𝑄 heat transfer rate
11
1 Introduction
The offshore industry is heavily reliant on flexible and reliant energy production for
their daily operations. Gas turbines (GTs) are generally the running source for both
electric and mechanical power demand offshore, and is responsible for about 27% of
the total Norwegian CO2-emissions [1, 2]. With political motivation to reduce
greenhouse gas emissions (GHGs) and the coherent increase of CO2-taxations in 2013
[3], the Norwegian offshore industry is looking towards reliant alternatives to the
simple gas turbines cycles to power the Norwegian continental shelf (NCS). To reduce
taxation cost, electrification of the NCS is being evaluated together with improved
combined cycle technology to meet the required emission levels. Nevertheless,
combined-cycle gas turbines (CCGTs) has since 1991 only been implemented on a
total of three platforms on the NCS. Challenges regarding offset operation conditions,
flexibility, space and weight requirements and the need for make-up water remains
the primary issues for the implementation of new CCGTs offshore.
This study will have its focus on the transient operations conditions of the bottoming
steam cycle. A study of both drum-type and once-through steam generation (OTSG)
for both operation on offshore and onshore will be investigated. Control aspects and
regulation of the two different steam generation systems are also being discussed.
Based on this study, an appropriate model for an offshore bottoming combined cycle
(BCC) will be built and thus simulated. Limitations to the current heat recovery
units will be presented and evaluated. The focus will be on dynamics of the CC in
transient part-load operation, as well as warm startup.
12
1.1 Background and motivation
The Norwegian oil and gas sector represents today one of the largest demands related
to electrical power consumption in the industry. The political motivation to reduce
greenhouse gas emissions is of high interest both nationally and internationally, which
in 2013 resulted in a doubling of the current CO2-taxation in Norway. While gas
turbines still being the primary source for both electrical and mechanical energy
supply for operation on the NCS, companies currently operating the fields are looking
for alternatives to reduce the cost due to the CO2-taxation. Although there has been
high profiled investments in R&D projects regarding CO2-cleansing and deposition,
the only practical solutions has been improvement of current GT technology. Pål
Kloster [4] claims that the CO2-reduction in existing combined cycle systems offshore
on the NCS represented between 50-92 ktonnes CO2 per platform in 1999. This
corresponds to savings in emission taxation of about 22-40 million NOK per year, and
represents a fuel and emission reduction of about 25% compared to the traditional
usage of single gas turbine cycles. With the increased expenses in emission taxes and
CO2-quotas, projects like COMPACTS from SINTEF, and EFFORT/PETROMAKS
from NTNU have originated to research on potential weight reduction, materials and
space requirements that has been the long-lasting limitations holding back the
implementation of CC in the Norwegian oil industry.
The main concerns about implementing combined cycle technology offshore is
primarily related towards
Rapid changes in both heat and power demand.
Volume and weight requirements for retrofit design on existing platforms
already installed with single-cycle gas turbines.
Availability of make-up water and qualified purification equipment in the
steam cycle.
Lifetime and reliability
Thermos-mechanical fatigue and creep on materials, corrosion and lifetime
reduction of components due to irregular to discontinuous operation demand.
Justification regarding to total investment cost and savings, related to CO2-
taxation policies, which has changed massively since 1991.
13
It is thus important to look at bottoming combined cycles as an alternative to
electrification of large parts of the NCS, which is heavily debated in politics to date.
Platforms which operate remotely from other oil and gas installations would likely
favor the most of CC-technology. The political debate is in largely biased toward
electrification, and hence it is important to include alternatives while taking lifetime
emission analysis into the evaluation.
This is especially important given that 80% of the NCS emission stem directly from
gas turbine operation. As mentioned, a reduction of about 25% in fuel and emissions
represents a substantial amount when taken electrification into account.
Nonetheless, regarding offshore as the main objective of study, onshore power
production is also heavily reliant on dynamic operation with a fluctuating abundance
of renewable energy being produced in Europe. The need to quickly regulate power
production will be of high interest in the foreseeable future.
1.2 Thesis objective and outline
The aim for the study is to investigate and simulate a chosen bottoming combined
cycle applicable for offshore platforms, with emphasis on flexible and transient
operation conditions.
The work to be done can be summarized as:
Literature study on current waste heat recovery units (WHRU) technology,
both drum-type (HRSG) and once-through steam generation systems (OTSG),
onshore as well as offshore.
Investigating control aspects of the steam cycle, and how various components
in the cycle regulate and perform during transient operation. A control
strategy should be included for the steam cycle.
Acquire and use suggested thermodynamic parameters and dimensions based
on already existing BCC-technology offshore and scientific studies.
14
Constructing and selecting an applicable steam cycle for further study based
on operation conditions in offshore environments.
Build a bottom-up model using Dymola/Modelica simulation software.
Simulation will be focused on plant load changes.
Suggest further work and improvements of current model
It should be emphasized that only preliminary results of the model are expected from
the work of this project thesis. Studies of dynamic behavior of a power plant can be
time consuming, and it is expected that the current work will be material for further
study for a potential MSc thesis.
1.3 Limitations of work
Due to the large coverage of literature studies on combined heat cycles, with primary
focus on onshore plants, it is vital that the work is done within well-defined scope
and limits. Some of the of these are made specifically here, and others explained
further on in the report.
Emissions of NOx and CO2 linked to gas turbine offset operation conditions
will not be studied.
Dynamic model build will be based on already existing steady-state acquired
data. Study of transient operations and behavior of the system are the main
objective.
Detailed studies of thermal expansion and contraction of thermal fatigue will
not be included. Neither will start-up procedures, or shut-down procedures of
CC, with regard to scheduled or controlled steps.
1.4 Further work
Simulations including steam-drum dynamic model and additional components
to fulfill the whole power plant system.
Include multiple pressure cycles and drums in simulation.
15
Build a refined and more detailed HRSG-component,
Detailed data validation (offset-data) with steady state simulation in Dymola
toward other software.
More detailed analysis of the behavior of a the HRSG system with
Study actual limitations to HRSG components regarding fatigue, stresses and
equipment limitations.
1.5 Literature study
Prior to the project thesis a course regarding literature search was attended and
recommended by supervisor Lars O. Nord available on NTNU. The course introduced
to detailed library search engines, both available at NTNU and to external databases.
The initial approach came to harvest material regarding “offshore dynamic combined
cycle” literature from the ORIA, the NTNU database. However, none to few articles
were initially found. Many publications on onshore three-pressure stage drum-based
HRSGs was found, most steady-state literature, but also some dynamic behaviors.
Most of the offshore study can be linked to research done here at NTNU, and is
regarded as quite new material in the context of bottoming cycle history.
It was long evaluated that dynamic modelling of solar powered BCC was the closest
to dynamic CC studies that could be found. In December 2015 two major articles
regarding dynamic simulation were released [5], [2]. Also co-supervisor Rubén Mocholí
Montañés have been of great help with references regarding the main theory of the
study. The result is that most material of this study came from references of these
papers, because of the lack of study of dynamic combined cycle systems found by
general search engines.
2 Combined Cycles Power Plants – An introduction
Since the introduction of the industrial revolution back in the 1800s, the world’s
energy demand has quickly grown. The need for electric power grew rapidly, and with
it more efficient power-producing plants. When it became clear that combustion
16
products had an impact on the environment and that fossil fuel resources are finite,
efforts devoted to developing more energy efficient power plants intensified.
The steam boiler (or rather the steam cycle) represents today about 80% of all
land-based electricity generation in the world [6]. Energy sources like coal, oil and gas
fill up the most common and finite sources, but steam production from nuclear,
biomass and solar energy does also represent a large part of the energy-mix.
A various combinations of thermodynamic heat-cycles have through history
been tested with the aim of combining high power, flexibility and low waste heat.
These cycles are referred to as combined cycles, and are typically defined as any
power-producing unit consisting of two or more power-producing cycles. They consist
of a topping cycle and a bottoming cycle, where the waste heat from the topping
cycle constitutes the heat input to the bottoming cycle. The most common combined
cycle consists of a gas turbine (Brayton cycle) as a topping cycle and a steam-
turbine-based cycle (Rankine cycle) as a bottoming cycle. See figure 1. Plants
employing this combination are common, and are normally referred to as combined
cycle power plants, or CCPPs for short.
Figure 1: Ts-diagram of the combined cycle. Brayton as top cycle, and Rankine as bottom cycle.
17
3 The components of a CCPP
Figure 2 shows a simplified schematic of a CCPP and its three main components: a
gas turbine, a steam turbine and a HRSG. The gas turbine operates at high
temperature and pressure where it produces power connected to a generator. The
flue-gas from the gas turbine is relatively hot and contain a large amount of energy.
This energy is fed to the HRSG where the flue-gas is evaporates and superheat water
to a high temperature under high pressure. The high-temperature steam is delivered
to a steam turbine where it expands and produces work. The steam is routed through
the steam turbine, and thereby condensed by the evaporator using cold water, usually
seawater as the cooling fluid. The condensed water is then routed back to the HRSG
for a new cycle of recovering heat, and the process repeats.
The energy in the exhaust gases, which is by far the major loss in the gas turbine, is
thus utilized to produce additional power from a second cycle. Roughly about two
thirds of the power of the plant is produced by the gas turbine.
18
Figure 2: Schematic illustration of a dual-pressure CCPP with steam drums.
3.1 The gas turbine
The gas turbine (GT) is the topping cycle, which is a normal open Brayton cycle
with some special features. A simple-cycle gas turbine, i.e a gas turbine not operated
in a combined cycle, has maximum efficiency at a high pressure ratio. A high
pressure ratio results in a large expansion and hence the temperature of the exhaust
gas is low. The low exhaust gas temperatures means a reduction in stack loss and
therefore a high efficiency. Too high pressure ratio results in the compressor
consuming a large amount of energy in comparison to the energy that can be added
with the fuel without exceeding the maximum turbine inlet temperature.
When a gas turbine is to be designed for a combined-cycle operation, it is no longer
desirable to have a low flue-gas temperature. A low flue-gas temperature would give a
low steam admittance temperature in the bottoming steam cycle, which limits the
efficiency and power output [7]. Therefore, the pressure ratio of the gas turbine
should be lower than for a single-cycle unit to ensure that the bottoming cycle
19
receives energy at a suitable temperature for steam production. On the other hand,
too low a pressure ratio will lead to poor gas turbine efficiency.
In today’s plants, with GT combustor outlet temperatures of 1400-1500C and steam
turbine inlet temperatures of 450-600°C, the pressure ratio should be in the range of
17-22 bar. In a sequentially fired gas turbine, the pressure ratio will be higher, with
the second combustor at the pressure level of an ordinary combine-cycle gas turbine.
Sequentially fired gas turbines are though not much used anymore the 1970 [6].
In some plants, like in the offshore Oseberg D combined cycle, two or more gas
turbines run in parallel to generate higher gas temperatures for higher efficiency. This
is common on the few combined cycle plants introduced on the NCS. This also gives
extra flexibility and better part-load performance [4]. More details on these will be
further discussed in chapter 4.
Figure 3: General Electric LM2500+G4 gas turbine. [8]
3.2 The steam cycle
The bottoming cycle is an ordinary Rankine cycle utilizing water as the heat transfer
fluid. Unlike the topping cycle, the bottoming steam cycle is a closed cycle, i.e. the
working fluid never leaves the system. In the simplest possible Rankine cycle, water is
boiled and superheated in a boiler. The steam is then passed from the boiler to a
steam turbine where it is expanded, producing work. The steam is then condensed in
20
a condenser using an ambient heat sink as seawater or air, depending on
configuration and location of operation. Finally, the water is pumped up to the
operating pressure and returned to the boiler.
In the Rankine cycle the working fluid is pressurized in the liquid state, which is
much less energy consuming than compression in the gaseous state, as in the Brayton
cycle. The pressure ratio is therefore not limited by the work consumption but by the
two-phase region (gas-liquid) at the end of the expansion. To further increase the
pressure of a Rankine cycle, the steam can be reheated part of the way of the
expansion. This allows a higher admission pressure for a given maximum steam
temperature without exceeding the maximum moisture content at the end of the
expansion. If the moisture content is too high at the end of the expansion, the steam
turbine will suffer from erosion, which drastically shortens its lifetime.
When reheating is introduced the next limit on admittance pressure is the physical
blade length in the high-pressure steam turbine. The preheating energy is supplied by
steam extracted from the steam turbine at a number of pressure-levels. Preheating
reduces the amount of fuel required in the boiler. In a combinced cycle this is not
beneficial because it will reduce the recovery of heat from the exhaust.
To control the steam production, a method called variable pressure control is used in
the steam cycle. This will be discussed in chapter 6.
3.3 HRSG
The HRSG is the interface between the top and bottoming cycle. It can be compared
to a big heat exchanger where the heat from the high-energy exhaust is transferred to
the water or steam. The HRSG is the largest component of the combined cycle,
where the tube arrangements represent the heavies one. This is due to the large
surface area needed to recuperate the heat from the exhaust because of the poor heat
transfer properties to the flue gas [4].
21
The HRSG can include various pumps, steam-drums, feedwater tanks, valves, rack
structure (manifold), deaerator, chemical dosing station – depending on configuration
and supplier.
The heat transfer in a HRSG mainly consists of convection, unlike the ordinary steam
boilers on coal plants where radiation contributes to the heat transfer. The simplest
form of HRSG operates at only one pressure level, which means that water only boils
one pressure and circulates in one cycle. To increase the efficiency of the HRSG
additional evaporators working at different pressures can be introduced. It is custom
to use multiple pressure-levels on onshore power plants, but single-level on offshore
cycles. The reason for the increase in efficiency is discussed in chapter 5.2.
Figure 4: The HRSG units of 660MW land-based combined cycle power plant (CCPP)[9] .
3.3.1 Finned tube heat exchanger
Most designs have a stack of vertical or horizontal tubes connected in series, where
the liquid gradually absorb the heat from the fluegas on its way out the stack. The
tubes usually have extended fins for increased surface area towards the flue gas.
22
Figure 5: Right: High pressure supeheater fin configuration of Oseberg D. Left: Illustration from
frbiz.com [10]
HRSG – tube material property sheet
(@ 20°C unless otherwise specified)
Incoloy 800HT
[11]
TP407 [12]
Density 7940kg/m3 7700 kg/m3
Specific Heat Capacity 460 J/kg•°C 460 J/kg K
Thermal conductivity: @ 100°C
@ 500°C
13.0 W/m°C
19.5 W/m°C
23.0 W/mK
25.0 W/mK
Thermoflow mean values @ 260°C 15.6 W/m°C 26.1 W/mK
Table 1: HRSG specification for Oseberg D fins and tubing
The fin material should ideally have a large thermal conductivity to minimize
temperature variations from its base to its tip. The efficiency of the fins are directly
related to Δ𝑇𝑏, the temperature difference, and thus the driving force for the heat
transfer. [13]. The fin efficiency 𝜀𝑓 is dependent on the geometry of the fin
arrangements, and will vary with the external flow pattern and arrangements of the
tube bundle.
𝑄𝑓 = 𝜀𝑓 ⋅ ℎ ⋅ 𝐴𝑐,𝑏 ⋅ 𝜃𝑏
23
𝜃𝑏 = 𝑇(𝑥) − 𝑇0 = Δ𝑇𝑏
3.3.2 Sections in the HRSG
The flue gas entering the duct also need to decelerate to low enough speeds for the
gas to make contact with the heat exchanger. This is why the cross sectional area
increases radically when the flue gas enters the duct of the HRSG.
There are generally three noticeably separate sections in the WHRU/HRSG where
water is sequentially heated all the way from liquid to superheated steam. See figure
5. Each section is represented as one line in the TQ-diagram as seen in figure 7. The
notation used is economizer (ECO), evaporator (EVAP) and superheater (SUPH).
The economizer heats the liquid water up to saturation (𝑇𝑠𝑎𝑡), while the evaporator
transfers the needed evaporation enthalpy for the phase-change (𝑇𝑠𝑎𝑡 = 𝑐𝑜𝑛𝑠𝑡) and
the superheater for heats the fully gaseous steam. Because of heating sequence, each
section of the WHRU must correspond to their respectively needed heat transfer
range, which means the exhaust first comes in contact with the superheater, then the
evaporator, and then the economizer before it leaves the WHRU.
24
Figure 6: Schematics of vertical HRSG module. The illustration shown gives the exact number of
pipes, and thermodynamic values during steady-state operating. (Edited figure by source: Thermoflow-
data by Lars O. Nord [14])
25
Figure 7: TQ-diagram of single pressure HRSG
To date there are several configurations in making the HRSG setup, through different
pressure levels and arrangements of separation of fluid, which will be discussed more
in detail in chapter 5.
3.4 Water treatment
The pressure at the end of the expansion and in the condenser is well below
atmospheric pressure and, therefore, it is inevitable that air will leak into the system.
The condenser pressure lies normally around 0.05 bar with small variations depending
on design. Dissolved oxygen is a major problem because it causes serious corrosion in
the system [6]. Water also combines with dissolved carbon dioxide to form carbonic
acid, which causes further corrosion. It is thus clear that access to purified water is
essential to avoid corrosion and long-term fatigue to the components of the
bottoming cycle.
26
One way to blow off dissolved gasses in the steam cycle is to use a deaerator. The
principle is based on the fact that the solubility of gases in saturated water is almost
zero. The feedwater is heated to saturation by adding steam in a closed tank. The
steam and the gases will rise to the surface where they are cooled, and the gases and
some steam are vented out of the system at the top of the deaerator [15].
From a thermodynamically perspective, it is more beneficial to use energy at a lower
temperature for deaeration. For instance, part of the flow from the economizer outlet
can be flashed and used in the deaerator. To do this, the drum approach temperature
must be higher than normal and all the water must be passed through the unit. The
higher the approach temperature to the deaerator, the more gases can be removed.
The energy consumed in the deaeration process will be supplied from the evaporator,
which results in less low-pressure steam being available in the turbine. However, no
additional deaerator tank is required, which reduces the initial cost and saves space.
Figure 8: Illustration of the flows inside a deaerator [16]
27
4 Onshore and offshore BCC
In the worldwide power industry, the CCPP is a popular power-producing unit that
has a number of desirable features such as high efficiency, low initial cost, relative
short construction time, and small footprint and short start-up time. This has
resulted in a common choice for the land based power industry, and the number of
CCPPs has thus increased around the world, from 5% to almost 20% [6]. The state-
of-the-art efficiency lies very close to reaching 60% on onshore power plants [4].
These desirable features have made the CCPP a common choice for the power
industry, especially in the abundance of gas and oil.
Most offshore installations to date are only powered by simple gas turbine cycles,
with no bottoming cycle implemented, which have varying efficiency of 33-39% at
their optimal design point depending on manufacturer and model. The few platform
having installed BCC offshore has a total plant net efficiency around 50% [17]. This
is far lower than what onshore power plants are capable of, since the limitations to
size and weight on offshore platforms are stricter than onshore, and thus the area to
recuperate heat is less.
4.1 Existing CC platforms
Existing combined cycle implementations goes as back as far to the 1990s when the
Norwegian CO2-taxation law was implemented [18]. The heat recovery units on all of
the current tree platforms using this technology is being quickly replaced and
improved over time, and have gone through multiple improvements.
Oseberg was the first offshore platform to install a CC which was in operation in
1999. Later the same year, the combined cycle on Eldfisk was in operation. The
Snorre B combined cycle was planned to be in operation in 2000 [4]. The WHRU’s at
Eldfisk and Oseberg has later been replaced by more compact designs.
The platform steam cycles range from 10-15MWe, many connected to common GE
LM2500 GT’s.
28
4.2 Eldfisk
Figure 9: Skid of the Eldfisk combined cycle plant from 1999 [4]
The Eldfisk (Phillips) water injection platform has four LM1600 gas turbine driving
water injection pumps and one PGT 25 (LM2500) for gas compression. The main
generator is a LP10 steam turbine with a 5.4 MW Typhoon generator set and two 2.1
MW diesels as back up. The steam for the steam turbine is produced in a triple-inlet
WHRU-SG recovering heat from the PGT25 and two of the LM1600 gas turbines
(figure 9). Since the steam turbine is the sole producer of electricity under normal
operation the steam inlet valve must control the flow to the steam turbine according
to the power demand. The steam production and power demand are not directly
linked, so the steam cycle is designed to produce a minimum of 10% more steam than
normally required. This is to ensure control possibilities at load changes. The surplus
steam will be routed directly to the condenser via the steam turbine bypass valve.
This is done by pressure control. The steam turbine is designed for 10.3 MW
electricity production. There is no steam extraction from the steam turbine. There is,
however, heat recovery in the WHRU-SG for production of fresh water in a seawater
evaporator. This system is totally independent of the steam system.
29
An elegant feature of the Eldfisk water injection concept is the use of injection water
as cooling water for the condenser prior to injection. Thus there is no extra lifting of
seawater for the steam condenser. Reduced fuel consumption on the generator sets
compared to simple cycle gas turbine solution, will be approximately 23 MSm3/year.
This represents reduced CO2-emissions of about 50 000 tonnes/year. The Eldfisk
steam bottoming cycle was put in operation during the 4th quarter of 1999 [4].
4.2.1 Replacement of HRSG
In recent years, the old HRSG unit were replaced with a HRS Circular Steam
Generator units combined with LM2500 and LM1600’s into one single unit. These are
installed with a tangential inlet which reduces pressure drop, distributes exhaust
evenly to the coils, and overcomes the potential problem for GT interaction
resonance. Two units has been installed in a single lift module to replace hot casing
rectangular units that suffered stress induced cracking of the casings and supports.
The two units produce enough steam to generate 10MWe, sufficient for the complete
platform. [19]
Figur 1: Eldfisk Circular Steam Generation (CSG) installation [19]
30
4.3 Oseberg D
The combined cycle on Oseberg D uses two gas turbine packages (PGT25+) to drive
two gas compressors for reinjection. The package consists of a LM2500+ gas
generator and a power turbine designed by Nuovo Pigone. The power turbines are
rated to 30MW each and have 40.3% efficiency at ISO condition. The steam turbine
is located on the neighbor platform Oseberg A, and a 400m steam pipe connects the
HRSG with the steam turbine. The steam turbine is rated to about 19MW with the
gas turbines running on full load, this correspond to total plant efficiency of 50%.
The HRSG is of a double inlet module and recover heat from both gas turbines. The
exhaust has a temperature of 480°C and the steam is produced at one pressure level.
The HRSG package is placed over the two gas turbines and the design is of a vertical
gas flow arrangement with forced circulation. The gas turbines run independently of
the steam cycle, and the steam turbine will produce electricity from whatever steam
is produced in the HRSG. This ensures simple regulation [4].
Another operating possibility is to extract steam from the steam turbine to utilize as
process heat. One by-pass stack and one diverter are fitted to allow simple regulation.
[20]. A process cycle sheet is shown in figure 10:
31
Figure 10: Oseberg D 1999 steam power cycle, drum based HRSG [4]
Figure 11: Model of OTSG module installed for Oseberg D 2010. Coutesy of Macchi [21]
32
4.3.1 Dual OTST replacement in 2010
In 2010 two new OTSG modules from Aibel produced in cooperation of IST [22] and
Macchi [21] was replaced with the old one which has been in operation for less than
ten years. When in operation the OTSG will recover waste heat from two 28 MW
gas turbines.[22] The dimensions are approximately 20 x 20 x 25 meter with a weight
of 700 tons. The new modules is a OTSG that utilizes two gas export compressors to
generate steam to power a steam turbine which will generate 15MWe for the
platform. [23]
4.4 Snorre B
The Snorre B platform uses another combined cycle concept. The combined cycle
produce only electricity and run continuously at 100% load to maximize the efficiency
and cut the payback time. Both the two DR63P (LM2500+) gas turbine and the
LP17 steam turbine package are used for generator drive. The HRSG is of a double
inlet type and has incorporated supplementary firing in case of a gas turbine shut
down. The Snorre B platform export surplus electricity to Snorre TLP. This inter-
platform power distribution makes better flexibility and utilization of the electricity.
At design point and 100% load the gas turbines produce about 30MW each and the
steam cycle produce 17.3MW electricity. There is also a possibility for extraction of
steam with a total energy of 8.0 MW, and then the ST production will be 15.2MW.
[2].
4.5 Backup power on offshore platforms
This three cases show how different the configuration of power generation may be
offshore and how each combined cycle are designed especially for each platform.
Because the platforms are self-sufficient in energy and the energy demand varies, it is
important to have a system which is reliable and easy to regulate. All the combined
cycles have backup capacity and god regulations. Offshore Power Generationn 25 The
Oseberg D platform has a bypass stack with diverters and backup capacity is covered
by gas turbine genset. The Eldfisk platforms ensure reliability by using a dual fueled
33
5.25MW gas turbine in backup and god regulation of steam to the ST. On the Snorre
platform surplus energy is delivered to a neighboring platform and deficit exhaust
energy is covered by supplementary firing.
4.6 Challenges for further CC offshore
In the offshore industry, space and weight requirements are essential when it comes
to placing new equipment on a platform or a FPSO. A steam bottoming cycle needs
to be simple, with low weight and volume. This is one of the main reason why so few
offshore installations have combined cycles installed to date [4].
Other problems is particularly related to very transient and off-design operation
conditions, treating of feed and makeup water, and extraordinary corrosive conditions
on the platforms. This has led to multiple research intensives, especially in the
Norwegian oil industry to expand CC implementation.
COMPACTS goals is to reduce the weight of the steam turbine with its accessories
by up to 50 percent. A large part of the weight reductions will come from reduction
in the framework which currently contribute 50 percent of total weight. Most
research points toward replacing steel components with lighter metals like aluminum
in the framework, or titanium or Inconel in the heat exchangers. [18]
Lars O. Nord et.al. [24] did a study of reducing the weight of steam bottoming cycles
on offshore installations, showing how different components contribute to the weight
of the total cycle. Figure XY shows how especially the removing of bypass stack
reduces the weight considerably.
34
Figure 12: Comparison of weight reduction of drum-based HRSG versus OTSG [24]. Note that wet
weight axis starts at 100 tons.
The trend shows that implementations of compact OTSG and CSG technologies may
open up possibilities for further platforms implementing CC technology. The details
about different HRSGs will be discussed in the next chapter.
35
5 HRSG designs and configurations
Historically, the HRSG design has commonly been associated with the drum-based
horizontal steam cycle for land based CCPPs. Alternative designs like the Once
Through Steam Generator (OTSG) and Circular Steam Generator (CSG), and
variations of the HRSG has been in development since the 1990s. However, many of
the skids can be categorized into vertical and horizontal designs, each having pros
and cons considering configuration and operation conditions. Different models and
skids will be described in this chapter.
Figure 13: Size comparison between different HRSG skids (Courtesy of HRS [19])
5.1 Drum based HRSG
Drum-type circuits typically use natural circulation for horizontal designs and forced
circulation for vertical designs. This is because horizontal evaporator pipes are more
susceptible to backflow so that pumps are required for preventing system instability.
Drum-based HRSG are the most common design worldwide, and connects the
economizer, evaporator and the superheater in an HRSG. There is a drum for every
36
pressure level in the HRSG. The cross-section of the drums are circular, is installed
on the top of horizontal HRSGs, and on the top or side of vertical HRSGs. The
length depends on the size of the plant, and may be up to 15m. The drum is filled
wiith about 50% liquid water where the water at the surface is at the boiling point.
Hot water from the economizer is entering the drum and is distributed in the water
volume. A sketch of a drum is given below in figure 14:
Figure 14: Courtesy of ISA.org [25]
The drum is a pressure vessel with two main functions:
Separating liquid water and saturated steam,
Remove impurities in the feed water
From the drum, water is led to the evaporator. Back from the evaporator a two-
phase mixture is returned to the drum, above the water volume. This two-phase
mixture goes normally though hydro cyclones, where the steam and the liquid water
are separated. The steam leaves at the top of the drum for the superheater, while the
liquid water returns down into the water volume. The steam/liquid water in the
drum is at the saturation state, ensured by heat and mass transfer between the
37
saturated steam in the upper half of the drum, and the liquid water at the boiling
point in the lower half of the drum.
5.2 The influence of pressure levels
Considering the first and second laws of thermodynamics when designing a HRSG
means a compromise between the following:
As much energy as possible should be recovered, i.e, the T-Q diagram should
be extended as far as possible along the x-axis, and
The temperature difference Δ𝑇 (𝑝𝑖𝑛𝑐ℎ), i.e. the area between the lines in the
T-Q diagram, should be minimized.
Single-pressure CCPPs are used when a short start-up time is important, or if the
heat remaining in the flue-gas can be recovered for an external process, such as in
CHP plants. With only one steam drum it will have a relatively small heat storage
capacity and will therefore respond quickly to load control and have a short start-up
time. Single pressure is common in offshore combined cycle plants [6]
When examining the TQ-diagram of the single-pressure CCPP we see that a lot of
heat is not recovered, because the Δ𝑇 between the lines are extended during the
evaporation.
Figure 15: A typical layout of a single-pressure CCPP, with corresponding TQ-diagram [6]
38
Figure 16: Triple pressure level HRSG with reheating.
Figure 17: (right) TQ-diagram of Dual pressure level HRSG. A dual-pressure combined cycle with
reheating (Reheating is the last part on the hot end) [6]
On the other hand, when examining multiple-pressure diagrams like figure 17, we see
that the heat area between the lines are smaller, and thus more heat can be
recuperated at different pressure levels.
39
It is therefore evident that multiple pressure levels lead to increase in efficiency for
CCPPs. However this also lead to higher weight and more components in form of
extra drums and would therefore not be the optimal solution in i.e. offshore
planforms. The initial cost will increase as the piping and other material must be
stronger to withstand the higher pressure. A high pressure results in a low volume
flow entering the high-pressure gas turbine, which is a problem especially for small
CCPPs.
Figure 18: Efficiency of different CCPP configurations without parasitic or step-up. Heat to power
effiency. [6]
5.3 Once Through Steam Generator
The Once-through or Benson type steam generator was invented as early as 1930 and
was then based on supercritical conditions, meaning pressures above 221 bar –
traditionally found in coal-fired power plants operating today [26]. Still, development
and implementation of OTSG in CCGT plants are relatively new and are naturally
operation in subcritical conditions. Siemens have previously tested both vertical and
horizontal OTSGs attaining operation conditions up to 58% efficiency with land-
based development plants. (Tripple pressure).
40
Figure 19: Simplified CC for a once-through steam generation (OTSG) system. Illustration taken from
Jonshagen et.al. [6]
Siemens Power Generation Group, Germany claims some of the features and
advantages over drum-based HRSG are: [26]
No thick-walled drum, which is required for high pressure stages.
15-25 per cent less weight of pressure parts. This has also been described for
offshore installations by Lars Nord [24].
Attractive operating characteristics, with good flexibility and short start-up
times.
Fast cycling due to thin walls and therefore low thermal stresses
Compact lightweight pressure bundle
Other advantages include zero blowout before startup [27], but again rises the
challenge of purified water, which is generally not a problem on drum-based HRSGs.
There are requirements to demineralized feedwater, and additional utilities are thus
needed. There are also strick requirements for using stainless steel on feedwater
41
piping from the polisher to OTSG for cycling plants, which is not that critical in
drums.
Even though, main regulation problems like deposition and carry over are avoided
(explained further in chapter 6), and allows the OTSG to run dry for high
temperatures, given the steel can attain tolerate the thermal stresses.
As pressures and temperatures have become higher, the once through boiler has
become more attractive [28]. Once through designs avoid the need for a steam drum
to separate the steam and water mixture after it leaves the evaporator. High pressure
drums require very thick walled sections, which increases the overall weight of the of
the HRSG, while exerting thermal stresses on drums.
Figure 20: Comparative sizing of traditional onshore drum-based HRSG and newer OTSG. [29]
5.4 Development of compact HRSGs
Early gas turbine waste heat recovery units (WHRU’s) were designed to standards
like the API 560 in the early 1980’s [19], which was a robust and well proven fired
heater standard. These were heavy and bulky designs most adapted onshore power
plants. There was also a requirement to bypass the heat recovery unit for start-up
42
and also to control the amount of heat recovered in the heat exchanger. The bypass
required a pair of exhaust gas dampers to modulate and control the process fluid to a
temperature set point. In addition, a silencer was required between the gas turbine
and the WHRU to reduce noise levels from the GT. This resulted is a set of casings,
ducts, dampers and stacks that had to be assembled on the platform with suitably
designed structures to support them.
During the late 1980’s improvements were made to make the WHRU’s more compact
with the introduction in the North Sea of the integral bypass, this helped by having
the dampers, bypass and heat exchanger in an assembled form, however the silencer
and bypass stacks were still supplied separately.
Efforts to incorporate heat exchanger, bypass, silencer and stack to under one single
assembly intensified during the 1990s. Engineers focused on minimizing the amount
of steel required and to make a small footprint as possible. Support structures were
also a big issue, especially on floating platforms such as FPSO’s where motion added
considerably to the weight of the structures, making the center of gravity a problem.
Wind loading was also a factor because of the traditionally rectangular design on the
stacks. [19]
5.5 Circular Steam Generator (CSG)
This led to the development of circular WHRU’s, but introduced problems with
integrating coiled tubes into the exchanger. According to Wickham [19] circular
design will reduce overall weight by up to 25%. The advantages of circular coils is
that there are no return bends, which finds place in rectangular units where the
water bends 180 degrees to transfer it to the next pass and this adds weight and
pressured drop with no heat transfer benefit. Additionally, the amount of welding per
tube is greatly reduce, which will reduce welding and radiography by 86% and
improve the integrity of the coil.
Another feature is that the bypass could be positioned internally concentric with the
coils and also incorporate a silencer. To facilitate this concept a radial vane damper
was introduced to effectively control gas flow.
43
Figure 21: Circular HRSG concept by HRS [19]
Although developed specifically for offshore the first commercial units of this type
were employed on land based location in Texas on Solar Centaur 40 gas turbines. A
total of 29 units have been installed offshore worldwide by September 2015, including
three CSG modules for Eldfisk in the Norwegian Sea [19].
5.6 Differences - Drum and Once Through Boilers
The HRSG can be vertical or horizontally built. A vertical design give a small
footprint, but the steel structure is more expensive as it has to carry the heat
exchangers as well as additional equipment (i.e drums), depending on framework and
special design considerations. In a horizontal HRSG the exchanger tubes are
vertically placed and more suitable for natural circulating boilers and is the most
common type found in the world to date. The great majority of CCPPs in the world
uses horizontal natural-circulation drum-based HRSGs. Once-trough steam generators
were in 2008 only representing about 48 of a total of almost 2000 large CCPPs
worldwide [30].
Coherently, most literature focus on horizontal natural-circulation drum-based
HRSGs, and majority of these are land based CCPPs. [31]. Thus, there are few to
44
none studies on the dynamic behavior of forced-circulation-compact HRSGs without
drums, and even less on once-through type in any papers.
Figure 22: Circulations of water inside Vertical OTSG and Drum-based HRSG. (Courtesy of NEM
Group [32])
Figure 23: Courtesy of Milton R. Beychok [33]
45
Figure 24: Drum based skit to the left; OTSG to the right. Courtesy of IST [27]
Drum-type circuits typically use natural circulation for horizontal HRSGs and forced
circulation for vertical designs [34]. This is because horizontal evaporation pipes are
more susceptible to backflow so that pumps are required for preventing system
instability. Although HRSGs can be designed with evaporators that function without
the use of circulation pumps, the variable operational range and specific design
criteria can limit the applicability in offshore operation.
The vertical HRSG has horizontally placed tubes and need forced circulations by the
use of pumps. Operational experience shows that combined cycle plants with vertical
HRSGs are cycling tolerant systems, as the design permits the tubes to
expand/contract freely and independently of one another [29]. In contrast, the
evaporator-tubes for horizontal designs are hanging vertically in a more rigid harp
structure. In order to support their own weight, a larger wall thickness must be
selected compared to vertical HRSGs, resulting in higher thermal inertia of the
system. [30]
In contrast to the drum-type boiler, the water content of the system is fully
evaporated in a single passage so that significantly more heat transfer surface is
required. Once-through omits the necessity of a drum and strongly reduces the water
inventory, resulting in less thermal inertia and more flexibility in operation [2].
46
However, with once-trough systems the need for purified feedwater is critical. While
drum-based HRSGs can be installed with deaerator-functionality with the i.e. the
economizer, the OTSG need a separate purifications unit, being chemical dosing
station or an ordinary deaerator.
If deaeration is performed in the low-pressure drum, no separate deaeration unit is
required. To do this, the drum approach temperature must be higher than normal
and all the water must be passed through the unit. The higher the approach
temperature to the deaerator, the more gases can be removed. The energy consumed
in the deaeration process will be supplied from the evaporator, which results in less
low-pressure steam being available in the turbine. However, no additional deaerator
tank is required, which reduces the initial cost and saves space, favorable for offshore
installations.
As mentioned, the thick-walled components (drum in particular) of the HRSG
restrict the permissible start-up gradients of the combined cycle. Accordingly,
preferred application of the once-through design is in the high pressure (HP) stage.
The technology also enables supercritical steam parameters, given that future
developments in gas turbine technology are expected to continue the trend towards
greater exhaust mass flows at higher temperatures.
Regulation is needed on drum-based start-up procedures to level the water inside the
drums, and to hold the pressure within acceptable limits in every cycle. The problem
naturally grows with increased number of pressure-stages in the HRSG.
Superheaters and reheaters in the HRSGs are subject to severe thermomechanical
cyclces due to an increase of heating gradients and of the number of transitional
periods. In addition, steam drums are also stressed because they have great thickness
and many weak points such as down-comers, risers and steam pipes connected to the
main body. Indeed, during transient operation, they are subject to pressure and
temperature variations which induce low-cycle fatigue.
47
6 Control aspects and regulation
6.1 Drum level regulation
The level in the drum must be controlled to the limits specified by the manufacturer.
If the drum level does not stay within these limits, there may be water carryover. If
the level exceeds the limits, boiler water carryover into the superheater or the turbine
may cause damage resulting in extensive maintenance costs or outages of either the
turbine or the boiler. If the level is low, overheating of the water wall tubes may
cause tube ruptures. A rupture or crack most commonly occurs where the tubes
connect to the drum. Damage may be a result of numerous or repeated low drum
level conditions where the water level is below the tube entry into the drum.[25]
It is common with cracked or damaged water tubes as a result of time delayed trips
or operators having a trip bypass button. When the drum level gets too low, the
boiler must have a boiler trip interlock to prevent damage to the tubes and cracks in
the tubes where they connect to the boiler drum. The water tubes may crack or
break where they connect to the drum, or the tubes may rupture resulting in an
explosion. The water tube damage may also result in water leakage and create
problems with the drum level control. The water leakage will affect the drum level
because not all the water going into the drum is producing steam
Poor level control also has an effect on drum pressure control. The feedwater going
into the drum is not as hot as the water in the drum. Adding feedwater too fast will
result in a cooling effect in the boiler drum reducing drum pressure and causing boiler
level shrinkage. This can be demonstrated by pouring tap water into a pan of boiling
water. [25]
6.1.1 Shrink and swell
Shrink and swell must be considered in determining the control strategy of a boiler.
During a rapid increase in load, a severe increase in level may occur. Shrink and swell
is a result of pressure changes in the drum changing water density. During a rapid
increase in load, a severe rise in level may occur because of an increase in volume of
48
the bubbles. This increased volume is the result of a drop in steam pressure from the
load increase and the increase in steam generation from the greater firing rate to
match the load increase (i.e., bubbles expand). If the level in the drum is too high at
this time, it may result in water carryover into the superheater or the turbine. The
firing rate cycle can result in drum pressure cycles. The drum pressure cycles will
cause a change in drum level.
The firing rate change has an effect on drum level, but the most significant cause of
shrink and swell is rapid changes in drum pressure expanding or shrinking the steam
bubbles due to load changes. When there is a decrease in demand, the drum pressure
increases and the firing rate changes, thus reducing the volume of the bubbles (i.e.,
bubbles get smaller). A sudden loss in load could result in high drum pressure causing
shrinkage severe enough to trip the boiler on low level. A boiler trip at high firing
rates creates a furnace implosion. If the implosion is severe enough, the boiler walls
will be damaged due to high vacuum in the furnace.
Figure 25: Measurement of drum level. [25]
Typically, for redundancy, there are three different methods used to measure drum
level. In the "Boiler drums/level measurement" example, the bull's eye technology is a
direct reading level measurement. The differential pressure transmitter represents the
level control measurement, and the probe type sensor is a common method for level
49
alarms and low and high level shutdown. Note the connections in the second
illustration are not realistic. (Figure 25 above)
Figure 26: Measurement using gage glass
The basic indication of the drum water level is commonly shown in a sight gage glass
(bull's eye) connected to the boiler drum. Due to the configuration of the boiler, and
the distance the boiler drum is from the operator, a line-of-sight indication may not
be practical. The gage glass image can be projected with a periscope arrangement of
mirrors. There are a number of methods for visual drum level measurement. Other
methods are a closed circuit television and the use of fiber optics.
The sight glass reading is affected by the temperature/density of the water in the
sight glass. The water in the sight glass is cooler than the water in the boiler drum.
6.1 Off-design and partial arc control
At part load and off-design conditions the exhaust heat energy may change, which
affect the steam production in the HRSG, and consequently the ST. The ST turbine
50
is designed to follow the GT without control of the power output. Most steam cycle’s
s in combined cycle plants use sliding pressure operation down to 50% load. This
ensures good utilization of the exhaust energy and high efficiency. Below 50% load,
the live stream pressure is held constant by a valve at the steam turbine inlet. This
introduces throttling losses, and increasing stack losses [35]. The sliding pressure
operation is illustrated below.
Figure 27: Sliding pressure opeation [35]
At part load the ST have approximately constant volume flow. This implies that the
velocity vectors remain unchanged, hence the efficiency is constant [15]. Stodala law
helps us calculate off-design operation for the steam turbine, when the turbine nozzles
are not choked. [36].
For condensing turbines, where the pressure ratio is low and the ratio of swalling
capacity is almost 1 the simplified Stodola’s law coefficient can be simplified to:
𝐾𝑡 = [𝑚2] =�̇�
√(𝑝𝑖𝑛.𝑑𝑒𝑠 𝜌𝑖𝑛,𝑑𝑒𝑠) ∙ √1 − (𝑝𝑜𝑢𝑡,𝑑𝑒𝑠
𝑝𝑖𝑛,𝑑𝑒𝑠)
2
Where �̇� is the steam mass flow and 𝑝𝑖 and 𝜌𝑖 is respectively the design (or nominal)
pressure and density either in or out of the ST. The subscript “des” refers to the
value computed at design point conditions.
51
6.1.1 Part load regulation
The most common principle for reducing the load of a combined cycle is by reducing
both the airflow inlet of the compressor on the GT and reducing the fuel flow. This
provides a higher part-load efficiency of the combined cycle compared to only choking
the fuel [15]. The airflow reduction is done with VIGV (variable inlet guide vanes)
which can change the inlet angel of the flow into the first stage of the GT
compressor. The combination of regulating the mass flow and fuel flow, makes it
possible to maintain a high TIT (turbine inlet temperature), and consequently high a
exhaust gas temperature.
The VIGV’s may typically reduce the mass flow down to 40% of GT load. At loads
below this level, the TIT is reduced by reduction of fuel only and the efficiency drops
quicker. [37]
Regulating the load also introduces challenges regarding emission of NOx and CO,
but will not be discussed further.
The simplest turbine model is based on a fixed isentropic efficiency. However, the
efficiency is dependent on the turbine load, and should not be considered to be
constant at off-design conditions. The turbine's off-design conditions are predicted
with the correlation proposed by Schobeiri [38]. The correlation expresses the relation
between the isentropic efficiency (ℎ𝑖𝑠) and the dimensionless flow coefficient. The
isentropic efficiency is a function of the rotational speed n and the isentropic enthalpy
drop Δℎ𝑖𝑠.
𝜂𝑖𝑠 = 𝜂𝑖𝑠,𝑑𝑒𝑠 ⋅ √Δℎ𝑖𝑠.𝑑𝑒𝑠
Δℎ𝑖𝑠⋅ (2 −
𝑛
𝑛𝑑𝑒𝑠⋅ √
Δℎ𝑖𝑠.𝑑𝑒𝑠
Δℎ𝑖𝑠)
When the admittance pressure of a turbine is changed, the volume flow passing
through it will naturally be affected. The turbine’s ability to swallow a certain mass
flow is called the turbine capacity, and is a measure of the turbine size or rather the
52
area of the turbine inlet. The turbine capacity is essential for off-design modelling
especially for combined cycles which are operated at sliding pressure.
6.1.2 Partial arc
The steam turbine in a combined cycle has two special features. Firstly, largescale
combined cycles normally generate steam at more than one pressure level, and
therefore the steam turbine is equipped with steam induction points. Large steam
turbines consist of a number of turbine cylinders and the induction point will
normally be located between them. To control the steam production, a method called
variable pressure control is used in the steam cycle. A conventional steam boiler
plant sometimes uses partial-arc control, which has superior part-load efficiency.
Partial-arc control utilizes a fixed pressure in the boiler at all loads, giving high part-
load efficiency.
Figure 28: Illustraiton of a partial-arc inlet [6]
However, a constant pressure in the boiler is undesirable for the heat recovery in the
WHRU. With a fixed evaporation pressure the WHRU becomes stiff, inflexible and
inefficient. The part-load efficiency of the turbine is secondary to efficient heat
recovery. Instead of using a fixed pressure, the combined-cycle bottoming cycle is
controlled by sliding pressure control. Sliding pressure control means that the inlet
control valves are always fully open, the inlet volume flow is constant at all loads,
resulting in a more or less fixed pressure ratio over the first stage of the steam
turbine. Therefore, the temperature gradients are much smaller than those in partial-
arc control. [7]
53
6.2 Bypass stack flow
Bypass stack is normally installed on most high-pressure drum-based HRSGs. The
bypass is needed due to keep thermal stresses in the HRSG components within
allowable limits during startup procedure. This is especially important to high
pressures drums which is built with thick drum walls, and thus the thermal
expansion and contraction in the drums are higher than in low or medium-pressurized
drum cycles. The stresses are generated by the uneven distribution of the metal
temperature inside the drum. Kim et al. [39].
Figure 29: Bypass duct with integrated silencer for Vertical HRSG. Source: Courtesy of HRS [19]
The bypass stack is a simplified way to regulate both temperature, pressure and
massflow through the HRSG at the cost of volume and weight to the whole
installation. A diverter at the bottom of the bypass stack regulates the flow into the
HRSG stack or into the by-pass stack, normally installed vertically.
54
6.3 Supplementary firing
The plant output can be increased by supplementary firing, which means that a duct
burner is placed at the HRSG inlet. The oxygen surplus in the flue-gas is usually
sufficient for combustion. Supplementary firing increases the plant output at the
expense of efficiency. Supplementary firing was more common in early CCPPs
because the gas turbine exhaust was not enough for the steam cycle. It is rare to find
supplementary firing in a modern plant, but in some cases it is used to increase the
flexibility of the power plant.
6.4 Startup and shutdown of CC
The need to quickly shutdown and startup a combined cycle power plant is essential
at modern power plants. Above all, the gas turbine can be started and loaded
quickly. Because its reaction time is short, it is capable of following quick changes
and surges in load. The gas turbine and, more importantly, the steam turbine is also
sensitive to thermal stresses on the turbine blades. Therefore, regulated startup time
is necessary to avoid material stresses and fatigue from expansion and contraction on
the different parts of the turbines. Hot and cold startup can range between 50 to 170
minutes respectively depending on the size of the CCPPs [35]. It is vital that the
temperature gradient at startup is held to reduce the lifetime of the turbine blades,
and shorten the service inspection time for replacing of blades.
Nevertheless, the HRSGs drums and piping system is often limiting factor in a CC-
plant, especially if there’s multiple pressure levels [40]. The steam drum in particular,
are subject to high thermal stress during start-ups, which is generated by the uneven
distribution of the metal temperature.
The cold start-up time of a typical HRSG can range from 45 min to 2 hours or more.
[39]. During start-up all operating parameters including temperature, pressure and
mass flow will increase rapidly. During the base load operation, the steam exiting the
evaporator will have a mild degree of superheat and should enter the superheater in a
completely dry condition, which will enter the superheater in complete dry condition.
55
The steam from the evaporator will not only be saturated but will also contain free
water. This needs to be removed in a separator. [28]
7 Dynamic and Steady-state modelling
Whereas conventional HRSG design is largely based on well-known steady-state
models, detailed modelling and dynamic simulation of the relevant components are
necessary in order to evaluate and optimize HRSG design with respect to fast start-
up and shutdown capability. A discussed, most of current research is done through
steady-state modelling where the assumption of operation point is on design point of
the power plant.
Traditionally steady-state modelling has been the practice for modelling and
designing critical parameters for a power plant. The results from these calculations
are defined through boundary conditions and the desired components included in the
model. This type of modelling does not provide any details about the dynamic events
taking place in the components, but is excellent in giving a more global view of the
performance. The principle is that each component has a set of inlets and outlets at
which the stagnation properties are calculated using thermodynamic equations. Each
component depends on the others, and in order to model the complete cycle the
components should be linked together to form a network representing the whole
power plant.
There are numerous types of software developed to simulate steady-state. Dynamic
modelling on the other hand is considering the time-varying equations and boundary
conditions which the steady-state modelling ignores. This is not to say that the same
equations does not apply to the dynamic modelling, but it includes the time-varying
equations as well as the steady-state equations. When investigating dynamic behavior
of a system we usually simulate a defined model with rigid components and
geometries.
Since the pressure in the HRSG varies depending on which cycle we’re focusing on,
real gas conditions are typically used in calculations for the steam throughout the
56
HRSG. This means that the density and enthalpy deviates from that of ideal gas. For
both drum- and once-through based heat recovery, we have a mix of fluid water and
steam depending on location in the cycle.
Fortunately, the properties of water and steam is perhaps the most well documented
we can find in any research. There is an international organization, IAPWS that
defines the equations, which describes the various properties of water. There exists an
international organization, IAPWS that defines the equations which describes the
various properties for water. The latest data from IAPWS-97 is also included in the
Thermopower package described as IP-97. [41]
8 Modelling with Dymola/Modelica
8.1 Working with ThermoPower
The tool used to evaluate the power plant cycles was Dymola, a proprietary GUI
extension of the open-source based modeling language Modelica, with the package
Thermopower for thermal power and heat conversion systems. The last update of
Thermopower version 3.1 was released in 2011 with moderate, to few updates and
changes [41]. The recommendations to use the software originated from supervisor
Lars Nord and PhD and candidate Rubén Mocholí Montañés, the latter who have
experience using the software for dynamic simulation in organic Rankine cycles.
Unlike most commercial power plant simulation tools, Dymola/Modelica does not
function as a “black-box” software, showing the calculations behind the generated
results in clear detail. The Modelica, and thus the Thermopower library, is open
source, and modular, which means that the inherent properties of every model can be
split into simpler and simpler components, and the direct equations and code behind
the models can be examined. This also gives the user great freedom to modify the
code and customize components to comply with their models, which will be vital for
once-trough systems which is not well-documented nor included in ThermoPower by
default.
57
The adjacent downside of Modelica is that of making quick simplified models, for i.e.
idealized power plant. This depends on whether the actual model has been premade
in the library. If non-existing, the components in the cycle has to be built from
bottom-up in detail. A detailed discussion of this problem is explained in the HRSG-
build-up in chapter 8.3.
This problem is exemplified in that all models coming with the library are drum-
based, with multiple pressure levels, which is most typical design for land based
CCPPs. The library is mainly based on onshore three-pressure stage CCPPs, and
thus the buildup of a once-through model needs to be done from scratch.
Along with extensive work to simulate dynamic behavior of the system – which
indicate that one need results from steady-state conditions prior to the dynamic
calculations – can make the work time-consuming with parameterization of the
model. The initial conditions must also be changed for every part-load case
investigated, which means the whole parametrization of the power-cycles has to be
re-initialized.
An issue with Modelica is that it does not support 1-dimentional partial differential
and algebraic equations (PDAEs), which makes it impossible to model distributed-
parameter processes. Modelica uses finite element methods to approximated models of
such processes, which is described by ordinary differential algebraic equations
(DAEs).
58
8.2 Fundamental equations
Figure 30: Finite element method (FEM) 1D analysis schematic diagram [42]
The equations used by Modelica are based on partial differential equations (PDEs),
which is linearized and converted to ODE’s for solving, using the finite-element
method (FEM). Casella and Schiavo [42] has describe the equations in detail for one-
dimensional fluid flows with heat transfer in pipes. Summarized, the thermodynamic
intensive variables can be derived with respect to one-dimensional length x and time
t, illustrated in figure 30. Within this framework, the dynamic balance equations for
mass, momentum (neglecting the kinetic term), energy (neglecting the diffusion term)
and partial mass can be formulated as follows: [42]
Mass balance equation:
𝐴𝜕𝜌
𝜕𝑡+
𝜕𝑤
𝜕𝑥= 0
Dynamic momentum equation:
1
𝐴
𝜕𝑤
𝜕𝑡+
𝜕𝑝
𝜕𝑥+ 𝜌𝑔
𝑑𝑧
𝑑𝑥+
𝐶𝑓𝜔
2𝜌𝐴3𝑤|𝑤| = 0
Energy balance equation:
59
𝜌𝜕ℎ
𝜕𝑡+
𝑤
𝐴
𝜕ℎ
𝜕𝑥=
𝜕𝑝
𝜕𝑡+
𝜔
𝐴𝜙𝑒
Partial mass equations: (Advection equation)
𝜌𝜕𝜉𝑘
𝜕𝑡+
𝑤
𝐴
𝜕𝜉𝑘
𝜕𝑥= 0, 𝑘 = 1, … , 𝑁𝑠
𝜙𝑖 heat flux enetering pipe across lateral surface [W/m2]
ℎ specific enthalpy [J/kg]
𝜔 wetter perimeter [m]
𝐶𝑓 Fanning friction factor [-]
𝜌 density [kg/m3]
𝑤 massflow [kg/s]
𝜉𝑘 mass fraction of the kth component [-]
𝑁𝑠 number of chemical species in the fluid [-]
8.3 Metal wall model
Dynamic heat transfer through the lateral surfaces, or the wall of the tube is
described by Fourier’s equation. With focus on the heat capacity in the middle of the
tube and neglecting longitudinal heat conduction along the pipe, we get:
𝜌𝑚𝑐𝑚𝐴𝑚�̇�𝑚(𝑥, 𝑡) = 2𝜋𝑟𝑖𝜙𝑖(𝑥, 𝑡) + 2𝜋𝑒𝜙𝑒(𝑥, 𝑡)
where 𝜌𝑚 is metal density, 𝑐𝑚 is the metal specific heat capacity, 𝐴𝑚is the cross-
sectional area, 𝑇𝑚 is the tube temperature in the middle of the wall, 𝑟𝑖 and 𝑟𝑒are the
60
internal and external radius, and 𝜙𝑖 and 𝜙𝑒 are the corresponding heat fluxes across
the tube sufraces; the latter two can be calculated as:
𝜙𝑒 =𝜆
𝑟𝑒 ⋅ ln (𝑟𝑖 + 𝑟𝑒
2𝑟𝑒 )
(𝑇𝑒 − 𝑇𝑚)
𝜙𝑖 =𝜆
𝑟𝑖 ⋅ ln (𝑟𝑖 + 𝑟𝑒
2𝑟𝑖 )
(𝑇𝑖 − 𝑇𝑚)
where 𝜆 is the metal thermal conductivity and 𝑇𝑖 and 𝑇𝑒 are the temperatures of the
internal and external surface, respectively. The heat conduction equation is then
easily discretized by sampling it in N equally spaced nodes. More accurate
approximations of Fourier’s equation can be adopted if a more detailed description of
the heat flows and temperatures is needed (e.g. for thermal stress studies, not
included in this report).
8.4 Discretization in FEM
The approximated solution of the four PDEs - in the previous chapter - can be
obtained through several numerical methods. However, there’s only one method that
allow one to transform a PDE into a set of ordinary differential equations (ODEs) or
differential-algebraic equations (DAEs) with respect to time are suitable to use within
the Modelica framework. Here the focus is on the numerical methods termed Finite
Element Methods (FEM) used.
The FEM is based on the discretization of the solution region into basic elements.
The choice of selecting a proper set of nodes for every calculation is critical to every
customized model. More defined nodes for heat transfer problems will generate more
accurate and fast responding dynamics of the system, but also oscillate faster and
create tendencies of non-converging solutions to the iterations.
61
It will later be shown that the number of nodes will be held low, due to initial
modelling and stabilization of the model.
8.5 Approach to model and workload
Due to less components and a more simplified steam cycle, a once-through model
based on steady-state data from an Oseberg D skid without drums will be used as
initial approach. The steady-state data simulate a “half” Oseberg D cycle, meaning
only one of the two gas turbines in operation. The Dymola model will be iterated and
built up piecewise and validated towards the Thermoflow-data, especially regarding
the HRSG unit with multiple heat exchangers. Integration of more advanced flow-
components (pumps, evaporators, regulators) will be implemented when the HRSG-
model is stable and performs as expected.
The procedure to construct the Dymola model that follows is a result of multiple
conversations with co-supervisor Rubèn M. Montañès, and exchange of experience
with the use of ThermoPower library:
Work to be done:
Acquire steady-state Thermoflow data by Lars Nord (appendix) [14]
Conversion of data to fit HE input units.
HRSG-split into seven different heat exchangers:
Each HE has declared parameters unique for its section and function. Static
input and out conditions for flow and pressure are set.
Parameterize each unique HE. Initialize and stabilize toward steady-state
conditions for pressure, heatflow and temperature.
Combine HEs in series. Stabilize initial conditions.
Integrate all 7 HE to total HRSG. Validate heat transfer, pressure and
temperature.
If stable and working as expected, implement bottoming cycle with simplified
gas turbine, steam turbine and idealized condenser after steam turbine.
62
8.6 HRSG build-up
A piece-wise buildup of the HRSG-module, as previously stated, is the best practice
to validate the steady-state data from Thermoflow for an initial model. Properties for
the different HRSG-sections vary largely depending on whether the water is
evaporating, pre-heating or super-heating inside the tubes. External surface and fin-
configurations also vary between each section and are defined in detail in the
Thermoflow data. [14]
In the context of object-oriented modelling, it is convenient to split the model of a
generic heat exchanger (HE) into several interacting parts, belonging to three
different classes [42]
The model of the fluid flowing within a defined volume,
The model of the metalwalls enclosing the fluid
And the model of the heat transfer between the fluid and the metal, or
between the metal and the outer world.
Each heat-exchanger (HE) used to design the total HRSG is based on the
Thermopower model HE2ph, which is a counter-current multi-phase heat-exchanger
schematically illustrated in figure 31 below.
I II III IV V VI VII
fluidFlow convHT metalTube cC heatFlowDistrib
ution
convHT2N gasFlow
1D fluid flow
model for
water/steam
(finite
volumes, 2-
phase)
1D
convect
ive heat
transfer
.
Cylindrical
metal tube –
1 radial node
and N axial
nodes.
Counter-
current heat
transfer
adaptor for
1D.
Same heat flow
through two
different
surfaces.
1D convective
heat transfer
between two
DHT connectors
with a different
number of nodes.
1D fluid
flow
model
for gas
(finite
volumes)
Table 2: Table showing the details about the components inside the HE2ph
63
Figure 31: HRSG build-up. Components inside are numbered and described in table.
64
8.7 Challenges of parametrization
The Thermoflow data is based on finned heat exchangers, described briefly in chapter
3.3.1. Since the Thermopower library does not contain any finned heat exchangers,
the closest resemblance to our data will be a bare counter current tube two-phase
heat exchanger (HE2ph). Just one out of seven HEs are bare (no finned), and thus
this would be the only exchanger that would be expected to behave close to the
parameters given by the steady-state model.
The HRSG are parameterized with many a lot of manipulated values in order to be
correctly assembled, when compared to the original steady-state Thermoflow data
[14] (appendix) For instance, the external and internal radius of the tubes are
calculated based on the total volume of metal for each meter of tube. (see equation
below). How the amount of metal on the finned tubes should be included in a bare-
tube model has been hard to evaluate. The same goes for the external tube, where
the defined area resembles the bare tube, and not with the finned total area for the
majority of the exchangers. The external tube area used in the simulations are the
prime outside surface from the Thermoflow data, which is assumed to resemble the
bare external area of the tube subtracting area occupied by the fins.
𝑟𝑖𝑛𝑡 =4𝑉𝑓𝑙𝑢𝑖𝑑
𝐴𝑖𝑛𝑡𝑒𝑟𝑛𝑎𝑙,𝑓𝑙𝑢𝑖𝑑⋅
1
2
𝑟𝑒𝑥𝑡 =4(𝑉𝑓𝑙𝑢𝑖𝑑 + 𝑉𝑚𝑒𝑡𝑎𝑙)
𝐴𝑒𝑥𝑡𝑒𝑟𝑛𝑎𝑙,𝑡𝑢𝑏𝑒⋅
1
2
The first approach using the prime outside-surface opened up new problems regarding
wrong external and internal radius when calculated, which is considered to be critical
with regard to mass flow, turbulence and pressure-drop inside the tube. With this in
mind, the external tube surface was defined by the external diameter of the tube. The
difference in numbers are noticeable especially on exchangers with small fin spacing
which occupies larger tube surface.
The reason for this comes clear when examining the parameterization of the HE2ph
model, where the external tube surface is used to calculate the external radius of the
65
metalTube (III), illustrated as a yellow stroke in figure 32. The external tube surface
is in direct contact with an arbitrary defined exchange surface to the gas, through
heatFlowDistributor (IV) meaning that all heat exchange will go through these two
areas.
Figure 32: Definition of different exchange surfaces in the HE2ph model
Thus, defining the external tube surface and the gas exchange surface can be done
separately, which is needed for finned tubes. Recapping from chapter 3.3, the
effective heat transfer of a finned heat exchanger can be regulated through any of the
variables in the equation below [13].
𝑄𝑓 = 𝜀𝑓 ⋅ ℎ ⋅ 𝐴𝑐,𝑏 ⋅ 𝜃𝑏
𝜃𝑏 = 𝑇(𝑥) − 𝑇0 = Δ𝑇𝑏
Thus, the heat transfer coefficient, finned area, or the efficiency can be regulated to
fit the cross-flow data.
66
8.8 Pressure drop calculations
Thermoflow uses ESCOA correlations to calculate the overall pressure-drop though
each heat exchanger [43]. Unfortunate, neither Weiermann nor ESCOA correlations
for pressure drop exist in Dymola, which is suggested to be one of the overall most
accurate correlations done by a study of Næss [44]. In Dymola, the pressure-drop
correlation factor can be set to correspond to the data, calculated by the operation
point of the nominal values. The hydraulic friction coefficient (Kf) is defined by the
operating point, (e.g. nominal values) in Dymola, thus defined by the hardware
pressure drop across the HE. The option to pick either is arbitrary, since the
coefficient will define the pressure drop if calculated, and Kf and Fanning friction
factor (Cf) is correlated.
𝐾𝑓 =Δ𝑃𝑛𝑜𝑚 ⋅ 𝜌𝑛𝑜𝑚
�̇�2⋅ 𝐾𝑓𝑐
𝐶𝑓 = 2𝐾𝑓 ⋅𝐴3
𝜔ℎ𝑦𝑑 ⋅ 𝑙
𝜔ℎ𝑦𝑑 =4 ⋅ 𝐴
𝑃
Pressure calculations alternatives in Dymola:
Friction factor calculated from operating point: the hydraulic friction
coefficient is specified by a nominal operating point (�̇�𝑛𝑜𝑚, Δ𝑝𝑛𝑜𝑚, 𝜌𝑛𝑜𝑚).
Friction factor is computed from the a defined constant value of Fannings
riction factor(𝐶𝑓,𝑛𝑜𝑚)
Fannings friction factor is computed by Colebrook`s equation, assuming
turbulent flow. (Re > 2100)
The last alternatives include Colebrooks equation for turbulent flow, which would
generally be the best option for pressure-drop calculations. However, the use of
Colebrook resulted in inconsistent results and difficulties validating the data.
Nevertheless, because of relatively small pressure-drops and problems initializing the
models with it, an external seperate pressure-drop has been implemented as a pipe-
pressure-drop outside the HE like shown in the figure 33 below. This temporarily
solves the heat transfer problem, but leaves pressure solving for further work.
67
Figure 33: Single HE validation with seperate pressure drop module
8.9 Calibrating the HRSG heat transfer
The equation below shows a the overall heat transfer equation for a tube as function
of heat transfer coefficient for both sides (ℎ𝑖), heat conduction coefficient 𝑘𝑡𝑢𝑏𝑒, and
fouling factors (R), where the latter is ignored due to low values. Each term represent
a resistance to the heat flow.
1
𝑈0 ⋅ 𝐴0= ∑
1
ℎ𝑖 ⋅ 𝐴𝑖
𝑖
𝑛=1
+ln (
𝑑𝑜
𝑑𝑖) ⋅ 𝑑𝑖
2 ⋅ 𝑘𝑡𝑢𝑏𝑒 ⋅ 𝐴𝑖+ ∑ 𝑅𝑓,𝑖
𝑖
𝑛=1
68
In table presented, average heat transfer coefficients for each medium in a typical
HRSG is shown.
Section of
HRSG Flue gas
Water in
economiser
Water in
evaporator HP Steam
Heat transfer
coefficient
(HTC)
(𝑾/𝒎𝟐𝑲)
50 500 2500-10000 1000
Table 3: Average HTC values for a HRSG
Usually, the heat transfer is much higher on the water/steam side, than of the gas
side, but the limiting factor is highest resistance, or the lowest heat transfer, which
will be the gas convection. Therefore, the adjustments in the heat transfer coefficient
on the gas side is of importance when refining the data in Dymola model.
A problem with the HTC in the HE-model is that it is set static, so dependency on
thermodynamic variables like temperature and pressure won’t influence the
coefficient. For the current premade HE-model library of Thermopower, this is the
only option set for the current model. The same is given for thermal conductivity.
The conductivity coefficient through both fins and tube has been averaged for the
numbers given in the data, which is defined at an approximate average temperature
in HRSG for 260°C. The value is normally set to about 20 W/mK in general HRSG
models found in the library, and corresponds good with the numbers calculated of
finned pipes in chapter 3.3.
69
Figure 34: HRSG module for calibration
In general, the HE with the larges impact will influence the system, and calibration
when downwards to the second largest. After some trial and error, some tendencies
and “rules” began to emerge. The impact of these reduces with each number:
1. Adjusting heat exchanger with largest difference in heat transfer.
2. Adjusting down the boiler (evaporator) adds more heat to the superheater
(left) side. The right side (economizer-side) does not get affected much.
3. In general: HEs on the left side is more affected by changes. Right side, few to
none.
HPB1 has the largest area and by far the greatest heat transfer by an order of 5-7x
compared to the other exchangers. The fine-adjustment of this exchanger was critical
to get the others calibrated afterwards without much readjusting the previous ones.
As depicted on figure 34, one important difference to the HRSG model is the removal
of the high-pressure pump between the low temperature economizer (LTE) and high
pressure economizer (HPE0). This is due to the difficulties of initializing the system
when the pump runs outside its pump characteristic.
Thus, the design conditions for the LTE has been set to approximately 18 bar, which
rises the saturation temperature for the water, and thus expensive heat transfer can
easily occur at that would normally be restricted by 𝑇𝑠𝑎𝑡.
The calibration results and error on every heat exchanger is included in the appendix.
70
9 Evaluation of model
Figure 35: Preliminary model for simulation.
It should be highly noted that the preliminary results in the current model is a work
in progress and does not correspond to the steady-state quality of the Thermoflow
model. Also the transient results simulated do not necessarily represent real ramp-
times on the Oseberg CCGT either, and are merely suggested ramp-values evaluated
from conversations with supervisor Lars O. Nord, mixed with numbers found in
literature-studies of Mertens and book of Kelhofer [2], [35].
The lack of control systems in the model gives the HRSG-module more freedom to
respond quickly to changes than on a real CCGT-plant, which have far more control
valves, start-up criteria and regulation parameters. Previous attempts to include
PID-controllers and control valves on earlier models have failed due to the lack of
good initial conditions on the HRSG-module.
Even though developing a control strategy from the beginning of the project work
was suggested, difficulties with initializing even the simplest HE-models have
postponed this to be topic for further work. Controls can be implemented when
model behavior is thoroughly understood, initial conditions properly implemented and
71
stabilized conditions are obtained in accordance with the steady-state model data.
For this reason, no control-strategies have been introduced to the current model and
not evaluated during the course of building the model.
Figure 36: Semi-stable model of the Oseberg D plant, including pumps and pressure-control thorugh
valves.
Implementing pumps have been a thorough problem in completing the model, but a
semi-stable model has been produced (figure 36 above) able to simulate 238 seconds
of start-up. Due to oscillating start-up conditions, mainly from the bad initialization
in the HRSG, the flow in both the HP feed-pump after the LTE (1st economizer) and
the condensing forwarding pump, will operate outside their pump characteristics
defined by the Thermoflow data sheets, and thus stop the iteration while simulating.
Options to force front-flow and introduce check-valves on the pumps have failed due
to negative flow-patterns in the initial 100 seconds of the simulation.
Starting the model with steady-state initialization is also an option, but experience
shows that too many dependencies on other dynamic initializations influence the
start-up negatively. Thus, stable start-up data is needed to initialize a more complex
steam-cycle, which is a topic for further work.
72
10 Results of preliminary tests
One interesting parameter, that is not included in this dataset, is the steam-quality
in each HE during transient operation. To visualize the evaporation-data at transient
conditions would help see where the boiling occurs, and in which heat exchanger.
Furthermore, the results are only preliminary, and does not necessarily give
indications on how the steady-state HRSG would behave.
Recent research shows that an option for disabling the dynamic momentum term in
the HE-models may help stabilize fluctuating pressure oscillations [42]. This is though
most relevant to sonic phenomena, and may not influence the current model
conditions noticeably when looking at the current timescale. Also suggestions of a
numerical stabilization coefficient (𝛼) found in the same paper could help stabilize
the model, which is said to be predefined in current HE-code. It should be
emphasized to investigate these options in further work for these models.
10.1 Warm start-up ramp: 5 min
A start-up from stabilized initial conditions at 42% GT-load, with the rest of the
cycle being at nominal values. The initial ignition point is considered by the GT-
model in Thermopower to be minimum running condition for the gas turbine, still at
chocked conditions. Ramping is linearly up toward 100% load at the span of 5
minutes, giving approximately 2.1MW/min ramp time for the GT. Extended views of
the data can be found the appendix.
73
0
5000
10000
15000
20000
25000
30000
3800 4000 4200 4400 4600 4800 5000 5200 5400
Po
wer
ou
tpu
t [k
W]
time [s]
GT el.power and ST mech power output
GasTurbine,P_el SteamTurbine,Pm
0
2 000
4 000
6 000
8 000
10 000
12 000
14 000
16 000
18 000
3800 4000 4200 4400 4600 4800 5000 5200 5400
Q -
he
at t
ran
sfe
r to
wat
er/
ste
am [
kW]
time [s]
HRSG heat transfer to fluid
HPS3,fluidFlow,Q HPS1,fluidFlow,Q HPS0,fluidFlow,Q HPB1,fluidFlow,Q
HPE3,fluidFlow,Q HPE0,fluidFlow,Q LTE,fluidFlow,Q
74
The most differentiating results lies in the high inlet pressure in the economizers and
throughout the evaporator, with values close to 50 bar. Even though the source of
this has not been investigated thoroughly, it is likely that either the modified 18 bar
nominal value of LTE, in combination with the static boundary condition of the
inletMassflow could be the cause.
The vast pressure-drop down to approximately 18 bar (nominal HRSG-value), is
caused over the pressure-drop module after the HPB1 evaporator. Though designed
to have a nominal pressure-drop of only 1.3 bar, the drop is nearly 25 bar, and it is
not clear if the pressure-drop has any coherence with the boundary condition set on
the inlet, or that the module itself causes it. More investigation of the inlet conditions
are needed to debug these phenomena.
0,0
50,0
100,0
150,0
200,0
250,0
300,0
350,0
400,0
450,0
500,0
3800 4300 4800 5300
tem
pe
ratu
re [
de
g C
]
time [s]
Temp.water/steam HRSG module
Tw_HPS3_in,T
Tw_HPS1_in,T
Tw_HPS0_in,T
Tw_HPB1_in,T
Tw_HPE3_in,T
Tw_HPE0_in,T
Tw_LTE_in,T
SteamTurbine,steamState_in,T
10
15
20
25
30
35
40
45
50
3800 4300 4800 5300
pre
ssu
re [
bar
]
time [s]
Pressure.water/steam inlet HRSG modules
Tw_HPS3_in.flange.p
Tw_HPS1_in.flange.p
Tw_HPS0_in.flange.p
Tw_HPB1_in.flange.p
Tw_HPE3_in.flange.p
Tw_HPE0_in.flange.p
Tw_LTE_in.flange.p
SteamTurbine.steamState_in.p
75
Temperature in through the last superheater increases by almost 250°C from the
ramp-up, while the other exchangers has an overall increase of about 30-50°C.
The heat transfer at 42% load gives the front HPS3 high relative heat transfer,
because of its designed heat transfer properties. Thus the heat transfer at the
backend receives lower temperatures than the superheater absorbs.
As expected, when the load increases, the heat transfer shifts to the boiler, while
reducing the relative amount of heat absorbed by the superheater.
76
10.2 Part-load: GT ramp 80% to 100% and back
The second part-load simulations tests the regulation from 80% GT-load to 100% and
back to 80% over period of 1000 seconds. Both the ramp-up and down are 1MW/min
with a constant design-load over 400 seconds in the middle.
5 000
7 000
9 000
11 000
13 000
15 000
17 000
19 000
21 000
23 000
25 000
27 000
3800 4000 4200 4400 4600 4800 5000 5200 5400
Po
wer
ou
tpu
t [k
W]
time [s]
GT el.power and ST mech power outputPart-load: 80% to 100% and back. Period T = 16.67 min
SteamTurbine,Pm GasTurbine,P_el
77
0
2 000
4 000
6 000
8 000
10 000
12 000
14 000
16 000
18 000
3800 4000 4200 4400 4600 4800 5000 5200 5400
Q -
he
at t
ran
sfe
r to
wat
er/
ste
am [
kW]
time [s]
HRSG heat transfer to fluidPart-load: 80% to 100% and back. Period T = 16.67 min
HPS3,fluidFlow,Q HPS1,fluidFlow,Q HPS0,fluidFlow,Q HPB1,fluidFlow,Q
HPE3,fluidFlow,Q HPE0,fluidFlow,Q LTE,fluidFlow,Q
78
0
50
100
150
200
250
300
350
400
450
500
3800 4000 4200 4400 4600 4800 5000 5200 5400
tem
pe
ratu
re [
de
g C
]
time [s]
Temp.water/steam inlet HRSG modulesPart-load: 80% to 100% and back. Period T = 16.67 min
Tw_HPS3_in,T Tw_HPS1_in,T Tw_HPS0_in,T
Tw_HPB1_in,T Tw_HPE3_in,T Tw_HPE0_in,T
Tw_LTE_in,T SteamTurbine,steamState_in,T
79
10
15
20
25
30
35
40
45
50
3800 4000 4200 4400 4600 4800 5000 5200 5400
pre
ssu
re [
bar
]
time [s]
Pressure.water/steam inlet HRSG modulesPart-load: 80% to 100% and back. Period T = 16.67 min
Tw_HPS3_in.flange.p Tw_HPS1_in.flange.p
Tw_HPS0_in.flange.p Tw_HPB1_in.flange.p
Tw_HPE3_in.flange.p Tw_HPE0_in.flange.p
Tw_LTE_in.flange.p SteamTurbine.steamState_in.p
80
The same phenomena for the shift in relative heat transfer to the HEs can be seen in
the part-load change too. The temperature lag is largest for the high-pressure
exchangers, with over 300 seconds before the temperature starts increasing. The
HPB1 boiler is though quite responsive and follows the transient close to the actual
GT-load.
The steam generator also shows a lag with respect to the power output adjacent with
the gas turbine. It is thus clear that the power generation in the steam turbine will
go on despite the lower load change in the GT.
11 Review of work
The initial approach was to get familiar with the simulation software
Dymola/Modelica. While the main Modelica library contain documentation on all of
its components, the Thermopower library was not documented directly into the
library, but externally through various research papers and publications on its
webpages. The Thermopower package is transparent in its buildup of code and
models, but has few or none pre-made models beside the elementary components that
construct these. This is especially relevant to multicomponent heat exchangers. This
made the learning the modulation of the models ineffective, and manual research into
the source-code itself was the way to investigate the behavior of the various steam
cycle components.
A mistake not clarified early on in the project, was the initial approach to test
existing models in the Thermopower library. Even though Thermopower do contain
some test-models of HRSGs, a thorough compatibility issue with current version of
Dymola resulted all the initial models to fail compilation, due to different variable
dependencies with current Modelica library. This source of this error was made
evident late into the project, and thus a lot of time were used construction models
that would not compile, even the most basic HE. Together with a disjoint sets
documentation behind the library led to the evaluation of using an alternative
software to simulate the work instead.
81
However, the Thermopower models was later made compatible (with help from
Rubèn M. Montañès) with the current version of Dymola, and thus a two-phase HE-
model could be built and tested.
Emphasizing the need of proper steady-state data to model even the simplest heat
exchangers was not considered critical during the startup of the project. Making
dynamic modelling directly is a near to impossible task, unless steady-state
initialization of the models are achieved forehand. Nevertheless, for full-scale
simulation of a CC, steady-state data is vital in order to obtain useful data for a full-
scale transient operation.
12 Conclusion
A preliminary dynamic HRSG model based on Oseberg D steam cycle was developed
in Dymola. The lack of a cross-current heat exchangers with defined fin-area made
the initial models hard to iterate. Static and predefined heat transfer properties and
initialization options made the primary work focusing on debugging and
parameterizing of the models.
It is advised that newer models should be built on dimensionless numbers and
common variables for heat transfer and flow, and thus ease of transfer of parameters
and validation of results between different power simulation software. This is
reasonable since Thermopower allows for the detailed build-up of such models, and
such work should be done in further studies of combined cycle technology.
The results from the simulation show expected behaviors related to heat transfer
during transient operation. However, validation and refinement of the current model
is needed to get more accurate and reliable transient data for the CC. The HRSG-
module shows highest sensitivity to transient operation, and is unfortunately the
most rigid module in the cycle. Properties of a more representative HRSG-module has
been suggested, and focus on stabilization and proper initialization should be the
focus for further work.
From reviewed current research papers, another library called ThermoSysPro is
widely used for both dynamic and steady-state simulations of drum-based natural
82
circulating horizontal HRSG-modules, mostly onshore [45]. The distinct differences
between these libraries have not been examined through this report. They should be
considered reviewed for later work in order to emphasize the differences the two
libraries represent.
83
13 Appendix
Part-load trapes input signal: (1 = 100%)
Name Description Value
amplitude Amplitude of trapezoid 0.20
rising Rising duration of trapezoid [s] 300
width Width duration of trapezoid [s] 400
falling Falling duration of trapezoid [s] 300
period Time for one period [s] 1000
nperiod Number of periods (< 0 means infinite number of
periods)
1
offset Offset of output signal 0.80
startTime Output = offset for time < startTime [s] 4000
84
GasTurbine
Name Value Description
maxPower 25e6 [W]
flueGasNomFlowRate m_fluegas Nominal flue gas flow rate [kg/s]
flueGasMinFlowRate 57.97 Minimum flue gas flow rate [kg/s]
flueGasOffFlowRate flueGasMinFlowRate/100 Flue gas flow rate with GT switched off [kg/s]
fuelNomFlowRate 1.312 Nominal fuel flow rate [kg/s]
fuelIntFlowRate 0.7677 Intermediate fuel flow rate [kg/s]
fuelMinFlowRate 0.4966 Minimum fuel flow rate [kg/s]
fuelOffFlowRate 0.011 Flue gas flow rate with GT switched off [kg/s]
constTempLoad 0.60 Fraction of load from which the temperature is kept constant
intLoad 0.42 Intermediate load for fuel consumption computations
flueGasNomTemp 480 Maximum flue gas temperature [K]
flueGasMinTemp 274.85 Minimum flue gas temperature (zero electrical load) [K]
flueGasOffTemp 90 Flue gas temperature with GT switched off [K]
fuel_LHV 50.047e6 Fuel Lower Heating Value [J/kg]
fuel_HHV 55.533e6 Fuel Higher Heating Value [J/kg]
Steam Turbine Stodala – Nominal/design values.
Type Name Default Description
Boolean explicitIsentropicEnthalpy true Outlet enthalpy computed by
isentropicEnthalpy function
MassFlowRate wnom m_steam Inlet nominal flowrate [kg/s]
Pressure pnom 16.5 Nominal inlet pressure [Pa]
Real eta_mech 0.98 Mechanical efficiency
Boolean allowFlowReversal system.allowFlowReversal = true to allow flow reversal,
false restricts to design direction
Real eta_iso_nom 0.92 Nominal isentropic efficiency
Area Kt �̇�
√(𝑝𝑖𝑛 .𝑛𝑜𝑚𝑖𝑛𝑎𝑙 𝜌𝑖𝑛,𝑛𝑜𝑚𝑖𝑛𝑎𝑙) ∙ √1 − (𝑝𝑜𝑢𝑡,𝑛𝑜𝑚
𝑝𝑖𝑛,𝑛𝑜𝑚)
2
Kt coefficient of Stodola's law
[m2]
Real partialArc_nom 1 Nominal partial arc
Initialisation
MassFlowRate wstart m_steam Mass flow rate start value [kg/s]
Real PRstart 16.5/p_cond Pressure ratio start value
85
HRSG Oseberg D - Default values, steady stateheat transfer to water/steam (Q) [kW]
HPS3 HPS1
Ideal Real Increase Q if + Ideal Real Increase Q if +
3542 2804 20,84 % 339 229 32,45 %
DIFF [kW] -738 DIFF [kW] -110
HPE3 (OTB) HPE0
Ideal Real Increase Q if + Ideal Real Increase Q if +
2094,9 1401 33,12 % 1918,6 1866 2,74 %
DIFF [kW] -693,9 DIFF [kW] -52,6
HPS0 (OTB) HPB1 (OTB)
Ideal Real Increase Q if + Ideal Real Increase Q if +
726,3 795 -9,46 % 16779,5 18249 -8,76 %
DIFF [kW] 68,7 DIFF [kW] 1469,5
LTE CALIBRATION
Ideal Real Increase Q if + % Total ideal Q 25400,3
2481,75 3240 -30,55 % Total real Q 25344
DIFF [kW] 758,25 DIFF ideal/real 56,3
Comment:Default values derived in SI-units and entered into HE2ph Dymola/Modelica model. In total: 7 heat exchangers
Small initial overall heat transfer difference. Locally on every HE, up to 30% increase or decrease
86
HRSG Oseberg D -Corrected values, steady state (5000s)heat transfer to water/steam (Q) [kW]
HPS3 HPS1 HPS0 (OTB)
Ideal Real Δ Ideal Real Δ Ideal Real Δ
3542 3573 -0,88 % 339 341 -0,59 % 726,3 728 -0,23 %
31 2 1,7
Steam In Steam In Steam In
16.94 p 16.78 p 16.99 p 17.7 p 17.08 p 17.71 p
250 T 255.4 T 234.5 T 241.1 T 204.5 T 210.3 T
2916.1 h 2929.9 h 2877.4 h 2890.9 h 2794.7 h 2807.6 h
Steam Out Steam Out Steam Out
16.5 p (B.C) 16.5 p 16.94 p 16.77 p 16.99 p 17.66 p
430.1 T (B.C) 430.1 T 250 T 255.4 T 234.5 T 241.1 T
3319 h (B.C) 3338 h 2916.1 h 2929.9 h 2877.4 h 2890.9 h
Gas In Gas In Gas In
480 T (B.C) 480 T 439.9 T 438.2 T 436 T 434.2 T
Gas Out Gas Out Gas Out
439.9 T 438.2 T 436 T 434.2 T 427.8 T 425.6 T
HPE3 (OTB) HPE0 LTE
Ideal Real Δ Ideal Real Δ Ideal Real Δ
2094,9 2104 -0,43 % 1918,6 1940 -1,12 % 2481,75 2488 -0,25 %
HT to water 9,1 21,4 6,25
Water In Water In Water In
18.2 p 47.5 p 18.44 p 47.61 p 18.77 p (B.C) 47.90 p
152.8 T 152.6 T 101.5 T 100.4 T 33.66 T (B.C) 32.39 T
645.2 h 646.1 h 426.7 h 424.4 h 141 h (B.C) 140.0 h
Water Out Water Out Water Out
17.93 p 47.46 p 18.2 p 47.47 p 1.054 p 47.62 p
206.9 T 207.3 T 152.8 T 152.6 T 101.1 T 100.4 T
883.8 h 886.7 h 645.2 h 646.1 h 423.7 h 424.4 h
Gas In Gas In Gas In
231.9 T 222.7 T 206.9 T 196.6 T 183.8 T 172.5 T
Gas Out Gas Out Gas Out
206.9 T 196.6 T 183.8 T 172.4 T 153.8 T (B.C) 153.8
HPB1 (OTB) Calibration
Ideal Real Δ Total ideal Q 25400,3
16779,5 16804,4 -0,15 % Total real Q 25490,4
24,9
Water In Increase if Q+
17.93 p 47.4638 DIFF ideal/real -90,1
206.9 T 207.347
883.8 h 886707 Comment:Steam Out * B.C. indicates boundry conditions set static for the system.
17.08 p 17.7146 * Inlet water is set to 18.72 bar. HP_Pump integration avoided.
204.5 T 210.322 * p[bar], T[C], M[kg/s], h [kJ/kg]
2794.7 h 2807.6 * Steam Properties: IAPWS-IF97
Gas In
427.8 T 425.599
Gas Out
231.9 T 222.677
87
SI-unit corrected Thermoflow data
HPS3 HPS1 HPS0 (OTB) HPB1 (OTB) HPE3 (OTB) HPE0 LTE UNIT
Fin-tube type Serrated fins Bare Serrated fins Serrated fins Serrated fins Serrated fins Serrated fins
Tube arrangement Staggered Staggered Staggered Staggered Staggered Staggered Staggered
Fin material TP409 N/A TP409 TP409 TP409 TP409 TP409
Tube material Incoloy Incoloy Incoloy Incoloy Incoloy Incoloy Incoloy
Tube outer diameter 0,03175 0,03175 0,03175 0,03175 0,03175 0,0254 0,0254 [m]
Tube outer radius (rext) 0,015875 0,015875 0,015875 0,015875 0,015875 0,0127 0,0127
Tube INNER radius (rint) 0,01397 0,01397 0,01397 0,01397 0,01397 0,011049 0,009652
Tube wall thickness 0,001905 0,001905 0,001905 0,001905 0,001905 0,001651 0,003048 [m]
Fin height 0,009525 0 0,009525 0,009525 0,009525 0,0127 0,0127 [m]
Fin spacing 0,004373 0 0,002229 0,002229 0,002229 0,007557 0,008303 [m]
Fin thickness 0,001 0 0,001 0,001 0,001 0,001 0,001 [m]
Number of fins per meter 186,1 0 309,7 309,7 309,7 116,9 107,5
Serrated fin segment width 0,00397 0 0,00397 0,00397 0,00397 0,00397 0,00397 [m]
# of serrated fin segments 26,63 1 26,63 26,63 26,63 22,11 22,11
Un-serrated height / fin height 0,2 1 0,2 0,2 0,2 0,2 0,2
Longitudinal row pitch 0,07 0,07 0,07 0,07 0,07 0,07 0,07 [m]
Transverse tube pitch 0,07142 0,07288 0,07142 0,07142 0,07142 0,07142 0,07142 [m]
# of tube rows (longitudinal) 6 1 1 15,62 6 5 5
# of rows per pass 3 1 2 2 2 1 1
# of tubes per row (transverse) 28 28 29 28 28 28 28
Tube length 7,127 7,127 7,127 7,127 7,127 7,127 7,127 [m]
Gas path transverse width 2,036 2,036 2,036 2,036 2,036 2,036 2,036 [m]
Gas path frontal area 14,51 14,51 14,51 14,51 14,51 14,51 14,51 [m^2]
HX total outside surface area 660,9 19,86 88,12 2654,1 998,5 393,3 368,2 [m^2]
Maximum gas velocity 22,93 19,58 23,18 20,18 16,48 12,3 11,54 [m/s]
Gas pressure drop 0,004295 0,0002623 0,0005651 0,01326 0,004039 0,001384 0,001209 [bar]
Steam side velocity 27,81 66,72 30,56 0,2969 0,2869 0,868 1,09 [m/s]
Steam side DP from hardware 0,2842 0,8659 0,0537 1,335 0,008 0,1442 0,2856 [bar]
Heat transfer from gas 3555000 341200 729100 16842000 2102700 1925800 2491000 [W]
Heat transfer to steam 3529000 338700 723600 16717000 2087100 1911400 2472500 [W]
HPS3 HPS1 HPS0 (OTB) HPB1 (OTB) HPE3 (OTB) HPE0 LTE
Gas In 480,00 439,90 436,00 427,80 231,90 206,90 183,80 [C]
Gas In 78,40 78,40 78,40 78,40 78,40 78,40 78,40 [kg/s]
Gas Out 439,90 436,00 427,80 231,90 206,90 183,80 153,80 [C]
Gas Out 78,40 78,40 78,40 78,40 78,40 78,40 78,40 [kg/s]
Steam Out 16,50 16,94 16,99 17,08 17,93 18,20 1,05 [bar]
Steam Out 430,10 250,00 234,50 204,50 206,90 152,80 101,10 [C]
Steam Out 3319,00 2916,10 2877,40 2794,70 883,80 645,20 423,70 [kJ/kg]
Steam Out 8,75 8,75 8,75 8,75 8,75 8,75 8,75 [kg/s]
Steam In 16,94 16,99 17,08 17,93 18,20 18,44 1,09 [bar]
Steam In 250,00 234,50 204,50 206,90 152,80 101,50 33,66 [C]
Steam In 2916,10 2877,40 2794,70 883,80 645,20 426,70 141,00 [kJ/kg]
Steam In 8,75 8,75 8,75 8,75 8,75 8,75 8,75 [kg/s]
88
HPS3 HPS1 (BARE!) HPS0 (OTB) HPB1 (OTB) HPE3 (OTB) HPE0 LTE
Inlet Gas side gasNomPressure 1,0430 1,0387 1,0384 1,0379 1,0246 1,0206 1,0192 bar
Inlet Fluid side fluidNomPressure 16,94 16,99 17,08 17,93 18,20 18,44 1,09 bar
exchSurface_G 660,9 19,86 88,12 2654,1 998,5 393,3 368,2 m2
exchSurface_F (including fins) 105,10 17,52 18,14 273,60 105,10 69,27 60,51 m2
extSurfaceTub (only tube) 119,43 19,90 20,62 310,91 119,43 79,62 79,62 m2
extSurfaceTub (prime outer surface) 97,33 19,86 7,13 214,70 80,78 70,38 71,14 m2
gasVol 5,138 0,858 0,850 13,361 5,132 4,567 4,57 m3
fluidVol 0,7341 0,1224 0,1267 1,9111 0,7341 0,3827 0,2920 m3
serratedFinVol 6,74E-03 0,00E+00 1,94E-03 2,92E-02 1,12E-02 4,70E-03 4,33E-03 m3
unserratedFinVol 1,69E-03 0,00E+00 4,84E-04 7,30E-03 2,80E-03 1,18E-03 1,08E-03 m3
tubeInconelVol 2,14E-01 3,56E-02 3,69E-02 5,57E-01 2,14E-01 1,23E-01 2,14E-01 m3
metalVol_total (not used) 0,222 0,036 0,039 0,593 0,228 0,129 0,219 m3
metalVol_only_tube 0,2139 0,0356 0,0369 0,5568 0,2139 0,1229 0,2136 m3
Efficency of HE MULTIPLY with lambda, gammaG/F 0,78 0,08 0,13 0,89 0,68 0,49 0,45 -
heat capactity rhomcm 3597200 3597200 3597200 3597200 3597200 3597200 3597200 J/m3*K
conductivity lambda 20,86 20,86 20,86 20,86 20,86 20,86 20,86 W/m*K
gamma_G 137,30 96,00 85,30 111,90 106,20 130,80 128,00 W/m^2*K
HTC gas side gamma_G_corrected (fin efficiency) 107,53 96,00 11,05 99,23 72,64 63,73 57,48 W/m^2*K
HTC fluid side gamma_F 754,40 1839,50 1086,90 25807,00 3045,00 6761,00 6141,00 W/m^2*K
FFtype_G FFtypes.OpPointFFtypes.OpPointFFtypes.OpPointFFtypes.OpPointFFtypes.OpPointFFtypes.OpPointFFtypes.OpPoint
Hydraulic res. CoefficientKfnom_G - - - - - - - not needed - calculated
P-drop(nom) dpnom_G 0,004295 0,0002623 0,0005651 0,01326 0,004039 0,001384 0,001209 [bar]
density gas inlet rhonom_G (ideal gas law used) 0,474800389 0,499436728 0,502056586 0,507653444 0,695561723 0,728900456 0,764709826 [kg/m3]
Fanning Fric.Factor Cfnom_G - - - - - - - not needed
FFtype_F FFtypes.OpPointFFtypes.OpPointFFtypes.OpPointFFtypes.OpPointFFtypes.OpPointFFtypes.OpPointFFtypes.OpPoint
Kfnom_F - - - - - - -
P-drop (nom) fluid dpnom_F (hardware pdrop) 0,2842 0,8659 0,0537 1,335 0,008 0,1442 0,2856 [bar]
density fluid inlet rhonom_F (use Steam.density_pT(p, T)) Se row 43+44 Se row 43+44 Se row 43+44 Se row 43+44 Se row 43+44 Se row 43+44 Se row 43+44 [kg/m3]
Cfnom_F - - - - - - -
Initialization HPS3 HPS1 (BARE!) HPS0 (OTB) HPB1 (OTB) HPE3 (OTB) HPE0 LTE
T_startbar_G 459,95 437,95 431,9 329,85 219,4 195,35 168,8 [C]
pstart_G gasNomPressuregasNomPressuregasNomPressuregasNomPressuregasNomPressuregasNomPressuregasNomPressure
Tstartbar_M default default default default default default default
pstart_F 16,94 16,99 17,08 17,93 18,20 18,44 1,09
Ssinit FALSE FALSE FALSE FALSE FALSE FALSE FALSE
FluidPhaseStart Twophase Twophase Twophase Twophase Twophase Twophase Twophase
HPS3 HPS1 (BARE!) HPS0 (OTB) HPB1 (OTB) HPE3 (OTB) HPE0 LTE
Pipe pressure drop Given NO dP in HX (linear to w) [Pa/kg*s] 3248,74257 9898,26246 613,8545953 15260,631 91,44947417 1648,376772 3264,746228
Pipe cross-area Pipe cross-section 0,002452464 0,002452464 0,002452464 0,002452464 0,002452464 0,001534108 0,001170697
Fraciont of nominal flow rate at which linear friction equals turbulent frictionwnf 0,05 0,05 0,05 0,05 0,01 0,05 0,04
12. Fin thermal conductivi ty @ 500 F (260 C) [W/m-C] 2,61E+01 2,61E+01 2,61E+01 2,61E+01 2,61E+01 2,61E+01 26,13
13. Fin thermal conductivi ty s lope [W/m-C^2] 7,50E-03 7,50E-03 7,50E-03 7,50E-03 7,50E-03 7,50E-03 0,01
14. Tube thermal conductivi ty @ 500 F (260 C) [W/m-C] 1,56E+01 1,56E+01 1,56E+01 1,56E+01 1,56E+01 1,56E+01 15,58
15. Tube thermal conductivi ty s lope [W/m-C^2] 1,68E-02 1,68E-02 1,68E-02 1,68E-02 1,68E-02 1,68E-02 0,02
16. Pass inlet & exi t DP (0=1 vel . head, 1=180 deg. bend) 0,00E+00 0,00E+00 0,00E+00 0,00E+00 0,00E+00 0,00E+00 0,00
17. Water/steam side fouling factor [m^2-C/W] 8,81E-05 8,81E-05 8,81E-05 8,81E-05 8,81E-05 8,81E-05 8,81E-05
18. Gas side fouling factor [m^2-C/W] 1,76E-04 1,76E-04 1,76E-04 1,76E-04 1,76E-04 1,76E-04 1,76E-04
19. Gas s ide convective h.t.c. adjustment factor 1,00E+00 1,00E+00 1,00E+00 1,00E+00 1,00E+00 1,00E+00 1,00
20. Gas s ide pressure drop correction factor 9,00E-01 9,00E-01 9,00E-01 9,00E-01 9,00E-01 9,00E-01 0,90
21. Water s ide h.t.c. adjustment factor 1,00E+00 1,00E+00 1,00E+00 1,00E+00 1,00E+00 1,00E+00 1,00
89
Outputs: HPS3 HPS1 (BARE!) HPS0 (OTB) HPB1 (OTB) HPE3 (OTB) HPE0 LTE
22. Number of tube rows (longitudinal ) 6,00 1,00 0,51 15,62 5,88 5,00 5,00
23. Number of rows per water s ide flow pass 3,00 1,00 2,00 2,00 2,00 1,00 1,00
24. Number of tubes per row (transverse) 28,00 28,00 29,00 28,00 28,00 28,00 28,00
25. Tube length [m] 7,13 7,13 7,13 7,13 7,13 7,13 7,13
26. Gas path transverse width [m] 2,04 2,04 2,04 2,04 2,04 2,04 2,04
27. Gas path frontal area [m^2] 14,51 14,51 14,51 14,51 14,51 14,51 14,51
28. Min. gas free flow cross section / frontal area 0,50 0,56 0,47 0,46 0,46 0,59 0,60
29. H.T. surface area / min. free flow cross section 15,26 2,43 25,20 25,20 25,20 9,15 8,52
30. Gas s ide Nusselt number Re coefficient 0,23 0,18 0,15 0,20 0,20 0,27 0,27
31. Primary tube surface / tota l heat transfer surface 0,15 1,00 0,08 0,08 0,08 0,18 0,19
32. Gas s ide friction factor Re coefficient 6,72 0,34 10,30 9,13 9,13 5,66 5,48
33. Heat exchanger prime outside surface [m^2] 97,33 19,86 7,13 214,70 80,78 70,38 71,14
34. Heat exchanger total fin area [m^2] 563,60 0,00 80,99 2439,40 917,70 322,90 297,00
35. Heat exchanger total outside area [m^2] 660,90 19,86 88,12 2654,10 998,50 393,30 368,20
36. Radiation beam mean length [m] 0,14 0,15 0,14 0,14 0,14 0,19 0,19
37. Gas mass flux before tube bundles [kg/m^2-s ] 5,40 5,40 5,40 5,40 5,40 5,40 5,40
38. Gas mass flux @ min. free flow cross section [kg/m^2-s ] 10,87 9,57 11,43 11,63 11,63 9,12 9,07
39. Gas face veloci ty [m/s] 11,39 11,05 10,96 9,37 7,66 7,28 6,87
40. Maximum gas veloci ty [m/s] 22,93 19,58 23,18 20,18 16,48 12,30 11,54
41. Gas Reynolds Number 10681,00 10029,00 12169,00 13301,00 14647,00 9782,00 10528,00
42. Gas Prandtl Number 0,71 0,71 0,71 0,72 0,72 0,72 0,72
43. Gas convective Nusselt Number 85,35 62,85 58,67 84,41 89,96 95,12 101,50
44. Gas s ide convective heat transfer coeff. [W/m^2-C] 135,90 92,35 84,49 111,50 106,00 130,60 127,90
45. Gas s ide radiative heat transfer coeff. [W/m^2-C] 1,40 3,65 0,81 0,47 0,18 0,23 0,14
46. Gas side total adjusted h.t.c. [W/m^2-C] 137,30 96,00 85,30 111,90 106,20 130,80 128,00
47. Fin effectiveness 0,71 0,00 0,79 0,74 0,75 0,59 0,59
48. Heat exchanger effectiveness 0,78 0,08 0,13 0,89 0,68 0,49 0,45
49. Effective / tota l external area 0,76 1,00 0,81 0,76 0,77 0,66 0,67
50. Gas pressure drop [mi l l ibar] 4,30 0,26 0,57 13,26 4,04 1,38 1,21
51. Water s ide flow cross section area [m^2] 0,05 0,02 0,04 0,03 0,03 0,01 0,01
HPS3 HPS1 (BARE!) HPS0 (OTB) HPB1 (OTB) HPE3 (OTB) HPE0 LTE
52. Water s ide mean veloci ty [m/s] 27,81 66,72 30,56 0,30 0,29 0,87 1,09
53. Water s ide Reynolds Number 216952,00 809633,00 420715,00 54856,00 47528,00 82396,00 49147,00
54. Water s ide Prandtl Number 0,99 1,08 1,15 0,88 0,98 1,36 2,67
55. Water s ide Nusselt Number 425,90 1256,70 759,20 136,40 125,90 217,70 180,20
56. Water s ide pressure drop from hardware [bar] 0,28 0,87 0,05 1,34 0,01 0,14 0,29
57. Water s ide pressure drop correction factor 1,00 1,00 1,00 1,00 1,00 1,00 1,00
58. Water side heat transfer coefficient [W/m^2-C] 754,40 1839,50 1086,90 25807,00 3045,00 6761,00 6141,00
59. Overall heat transfer coefficient [W/m^2-C] 51,38 87,32 38,84 69,27 56,15 72,27 66,90
60. Estimated minimum tube wal l temperature [C] 343,50 249,70 300,00 211,20 169,00 113,60 57,97
61. Estimated maximum tube wal l temperature [C] 454,70 264,30 320,70 243,00 214,40 160,80 117,80
62. Maximum al lowable tube wal l metal temperature [C] 648,90 648,90 648,90 648,90 648,90 648,90 648,90
63. Estimated maximum fin tip temperature [C] 465,10 N/A 355,40 311,40 220,70 187,50 155,90
64. Recommended maximum fin metal temperature [C] 676,70 N/A 676,70 676,70 676,70 676,70 676,70
65. Estimated maximum al lowable water s ide pressure [bar] 139,00 151,10 149,70 153,80 154,80 175,50 367,30
Thermodynamics:
66. Gas massflow across HX [kg/s ] 78,40 78,40 78,40 78,40 78,40 78,40 78,40
67. Gas temperature entering HX [C] 480,00 439,90 436,00 427,80 231,90 206,90 183,80
68. Gas enthalpy entering HX [kJ/kg] 621,56 576,21 571,86 562,56 347,74 320,92 296,36
69. Gas temperature exi ting HX [C] 439,90 436,00 427,80 231,90 206,90 183,80 153,80
70. Gas enthalpy exi ting HX [kJ/kg] 576,21 571,86 562,56 347,74 320,92 296,36 264,58
71. Heat transfer from gas [kW] 3555,20 341,20 729,10 16842,50 2102,70 1925,80 2491,00
72. Water/steam massflow exi ting HX [kg/s ] 8,75 8,75 8,75 8,75 8,75 8,75 8,75
73. Water/steam pressure entering HX [bar] 16,90 17,00 17,10 17,90 18,20 18,40 1,10
74. Water/steam temperature entering HX [C] 250,00 234,50 204,50 206,90 152,80 101,50 33,70
75. Water/steam enthalpy entering HX [kJ/kg] 2916,08 2877,37 2794,65 883,76 645,19 426,70 141,05
76. Water/steam pressure exi ting HX [bar] 16,50 16,90 17,00 17,10 17,90 18,20 1,10
77. Water/steam temperature exi ting HX [C] 430,10 250,00 234,50 204,50 206,90 152,80 101,10
78. Water/steam enthalpy exi ting HX [kJ/kg] 3319,45 2916,08 2877,37 2794,65 883,76 645,19 423,67
79. Heat transfer to water/steam [kW] 3528,80 338,70 723,60 16717,10 2087,10 1911,40 2472,50
80. Log mean temperature di fference [C] 104,70 195,60 212,20 91,60 37,70 67,20 100,30
90
14 References
[1] SSB, “Utslipp av klimagasser, 1990-2014, endelige tall,” no. SSB, 18.
desember 2015, 2014.
[2] N. Mertens, F. Alobaid, R. Starkloff et al., “Comparative investigation
of drum-type and once-through heat recovery steam generator during
start-up,” Applied Energy, vol. 144, pp. 250-260, 4/15/, 2015.
[3] O. Technology, “Norway to set high carbon tax on oil and gas
production,” Offshore Technology, pp. 1, 17 October 2012, 2012.
[4] P. Kloster, “Energy Optimization on Offshore Installations with
Emphasis on Offshore Combined Cycle Plants,” in Offshore Europe
Conference, Aberdeen, Scotland, 1999.
[5] N. Mertens, F. Alobaid, T. Lanz et al., “Dynamic simulation of a triple-
pressure combined-cycle plant: Hot start-up and shutdown,” Fuel, vol.
167, pp. 135-148, 3/1/, 2016.
[6] K. Jonshagen, “Modern Thermal Power Plants-Aspects on Modelling
and Evaluation,” Doctor Thesis, Department of Energy Sciences, Lund
University, Room M:B, M-building, Ole Römers väg 1, Lund University,
Faculty of Engineering, 2011.
[7] K. Jonshagen, and M. Genrup, “Improved load control for a steam cycle
combined heat and power plant,” Energy, vol. 35, no. 4, pp. 1694-1700,
4//, 2010.
[8] G. Aviation, “product sheet of the LM2500+G4 gas turbine,” 2015.
[9] P. Construction. "Luna Energy Facility - 600 MW Combined Cycle
Power Plant," 29.feb 2016; http://www.pcl.com/Projects-that-
Inspire/Pages/Luna-Energy-Facility-600-MW-Combined-Cycle-Power-
Plant.aspx.
[10] "Carbon Steel Titanium Spiral Finned Tube For Boiler Economizer,"
2016.
[11] S. c. I. A. 800, 2004.
91
[12] N. A. Stainless. "Flat Products Stainless Steel Sheet T409,"
http://www.northamericanstainless.com/wp-
content/uploads/2010/10/Grade-409.pdf.
[13] T. L. Bergman, F. P. Incropera, D. P. DeWitt et al., Fundamentals of
Heat and Mass Transfer: Wiley, 2011.
[14] O. N. Lars, "Thermoflow Output data: Half Oseberg D " O. N. Lars, ed.,
2015.
[15] B. Olav, “Natural Gas Technology, Thermal power generation Autumn
2014,” pp. 176, 2014.02.14, 2014.
[16] Wikipedia, “Daerator sketch,” 2016.
[17] L. O. Nord, and O. Bolland, “Design and off-design simulations of
combined cycles for offshore oil and gas installations,” Applied Thermal
Engineering, vol. 54, no. 1, pp. 85-91, 5/14/, 2013.
[18] M. J. Mazetti, "COMPACTS," COMPACTS,
http://www.sintef.no/en/projects/sintef-energy-research/compacts/, [26.
feb 2016, 2014].
[19] M. Wickham, “Examining Applications Of Offshore Technology To
Onshore Gas Compressor Stations For The Profitable Generation Of
CO2-Free Power,” Heat Recovery Solutions Ltd, 2015.
[20] R. Farmer, "Combined Cycles Offshore, "North Sea platforms are
converting mech drives to
comb cycle operation"," Gas Turbine World, Gas Turbine World, 1998.
[21] M.-S. P. Generation, "MACCHI Complete HRSG Module Offshore
Platform Norway," 2015.
[22] IST, "IST awarded two contracts totalling $10 million," Toronto,
Ontario, IST, 2008.
[23] Aibel, "StatoilHydro awards 58 MUSD Oseberg EPCI contract to Aibel,"
B. Sandal, ed., Aibel, 2008.
[24] L. O. Nord, E. Martelli, and O. Bolland, “Weight and power
optimization of steam bottoming cycle for offshore oil and gas
installations,” Energy, vol. 76, pp. 891-898, 11/1/, 2014.
92
[25] J. Gilman, “Steam drum, Boiler drum level control,” ISA.org, July-
August 2010, 2016.
[26] J. Franke, U. Lenk, R. Taud et al., Advanced Benson HRSG makes a
successful debut, London, ROYAUME-UNI: Global Trade Media, 2000.
[27] T. Landon, “ Once Through Steam Generators - Design, Operation, and
Maintenance Considerations,” in McIlvaine Company Hot Topic Hour,
2013.
[28] F. Starr, “Background to the design of HRSG systems and implications
for CCGT plant cycling.pdf,” OMMI, vol. 2, no. 1, pp. 17, April 2003,
2003.
[29] P. Fontaine, “Cycling Tolerance - Natural Circulation Vertical HRSGs,”
ASME, 16-19, 2003
2003.
[30] E. Jeffs, Generating Power at High Efficiency: Combined Cycle
Technology for Sustainable Energy Production: Elsevier Science, 2008.
[31] F. Alobaid, R. Starkloff, S. Pfeiffer et al., “A comparative study of
different dynamic process simulation codes for combined cycle power
plants – Part B: Start-up procedure,” Fuel, vol. 153, pp. 707-716, 8/1/,
2015.
[32] N. Group, "Vertical OT HRSGs."
[33] M. R. Beychok, Fundamentals of stack gas dispersion: Milton R.
Beychok Irvine, 1994.
[34] H. Klingenhöffer, “Corrosion Books: Stationäre Gasturbinen. By: C.
Lechner, J. Seume - Materials and Corrosion 11/2003,” Materials and
Corrosion, vol. 54, no. 11, pp. 909-910, 2003.
[35] R. Kehlhofer, F. Hannemann, F. Stirnimann et al., Combined-cycle gas
& steam turbine power plants, 3rd ed. ed., Tulsa, Okla: PennWell, 2009.
[36] N. Yoan, Ellipse Law: Miss Press, 2012.
[37] Ø. Flatebø, “Off-design Simulations of Offshore - Combined Cycles,”
Master of Science in Mechanical Engineering, Department of Energy and
93
Process Engineering, Norwegian University of Science and Technology,
2012.
[38] M. T. Schobeiri, Turbomachinery Flow Physics and Dynamic
Performance: Springer Berlin Heidelberg, 2004.
[39] T. S. Kim, D. K. Lee, and S. T. Ro, “Analysis of thermal stress
evolution in the steam drum during start-up of a heat recovery steam
generator,” Applied Thermal Engineering, vol. 20, no. 11, pp. 977-992,
8/1/, 2000.
[40] H. I. H. Saravanamuttoo, Gas Turbine Theory: Pearson Prentice Hall,
2009.
[41] P. d. Milano. "ThermoPower - Open library for thermal power plant
simulation," 10th Des 2015, 2015; http://thermopower.sourceforge.net/.
[42] F. Schiavo, and F. Casella, “Object-oriented modelling and simulation of
heat exchangers with finite element methods,” Mathematical and
Computer Modelling of Dynamical Systems, vol. 13, no. 3, pp. 211-235,
2007/06/01, 2007.
[43] R. Hofmann, F. Frasz, and K. Ponweiser, “Heat transfer and pressure
drop performance comparison of finned-tube bundles in forced
convection,” WSEAS Trans. Heat Mass Transfer, vol. 2, no. 4, pp. 72-
88, 2007.
[44] E. Næss, “Experimental investigation of heat transfer and pressure drop
in serrated-fin tube bundles with staggered tube layouts,” Applied
Thermal Engineering, vol. 30, no. 13, pp. 1531-1537, 9//, 2010.
[45] B. E. Hefni, “Dynamic Modeling of Concentrated Solar Power Plants
with the ThermoSysPro Library (Parabolic Trough Collectors, Fresnel
Reflector and Solar-Hybrid),” Energy Procedia, vol. 49, pp. 1127-1137,
//, 2014.
Related Documents
