sign in sign up
<<
Home , Cheng cycle, Kalina cycle, Working fluid

Advanced Power Cycles with Mixtures as the Working Fluid

Maria Jonsson

Doctoral Thesis Department of Chemical Engineering and Technology, Processes Royal Institute of Technology Stockholm, Sweden, 2003

Advanced Power Cycles with Mixtures as the Working Fluid

Maria Jonsson

Doctoral Thesis Department of Chemical Engineering and Technology, Energy Processes Royal Institute of Technology Stockholm, Sweden, 2003

TRITA-KET R173 ISSN 1104-3466 ISRN KTH/KET/R--173--SE ISBN 91-7283-443-9

Contact information: Royal Institute of Technology Department of Chemical Engineering and Technology, Division of Energy Processes SE-100 44 Stockholm Sweden

Copyright © Maria Jonsson, 2003 All rights reserved

Printed in Sweden Universitetsservice US AB Stockholm, 2003

Advanced Power Cycles with Mixtures as the Working Fluid Maria Jonsson Department of Chemical Engineering and Technology, Energy Processes Royal Institute of Technology, Stockholm, Sweden

Abstract The world demand for electrical power increases continuously, requiring efficient and low-cost methods for power generation. This thesis investigates two advanced power cycles with mixtures as the working fluid: the Kalina cycle, alternatively called the - cycle, and the evaporative turbine cycle. These cycles have the potential of improved performance regarding electrical efficiency, specific power output, specific investment cost and cost of electricity compared with the conventional technology, since the mixture working fluids enable efficient energy recovery. This thesis shows that the ammonia-water cycle has a better thermodynamic performance than the as a bottoming process for natural gas- fired gas and gas-diesel , since the majority of the ammonia-water cycle configurations investigated generated more power than steam cycles. The best ammonia-water cycle produced approximately 40-50 % more power than a single- steam cycle and 20-24 % more power than a dual-pressure steam cycle. The investment cost for an ammonia-water bottoming cycle is probably higher than for a steam cycle; however, the specific investment cost may be lower due to the higher power output. A comparison between combined cycles with ammonia-water bottoming processes and evaporative cycles showed that the ammonia-water cycle could recover the exhaust gas energy of a high pressure ratio gas turbine more efficiently than a part-flow evaporative gas turbine cycle. For a medium pressure ratio gas turbine, the situation was the opposite, except when a complex ammonia- water cycle configuration with reheat was used. An exergy analysis showed that evaporative cycles with part-flow humidification could recover energy as efficiently as, or more efficiently than, full-flow cycles. An economic analysis confirmed that the specific investment cost for part-flow cycles was lower than for full-flow cycles, since part-flow humidification reduces the exchanger area and humidification tower . In addition, the part-flow cycles had lower or similar costs of electricity compared with the full-flow cycles. Compared with combined cycles, the part-flow evaporative cycles had significantly lower total and specific investment costs and lower or almost equal costs of electricity; thus, part-flow evaporative cycles could compete with the combined cycle for mid-size power generation.

Language: English

Keywords: power cycle, mixture working fluid, Kalina cycle, ammonia-water mixture, reciprocating internal combustion , bottoming cycle, gas turbine, evaporative gas turbine, air-water mixture, exergy

List of Appended Papers

This thesis is based on the following papers, referred to by the Roman numerals I- VIII:

I. Jonsson, M., Thorin, E. and Svedberg, G. (1999). Gas Engine Bottoming Cycles with Ammonia-Water Mixtures as Working Fluid. In: Proceedings of the 1999 International Joint Power Generation Conference, Burlingame, California, USA, July 25-28, 1999. PWR-Vol. 34, 55-65. II. Jonsson, M. and Yan, J. (2000). Diesel Engine Bottoming Cycles with Ammonia-Water Mixtures as Working Fluid. In: Proceedings of the 2000 Spring Technical Conference of the ASME Internal Combustion Engine Division, San Antonio, Texas, USA, April 9-12, 2000. ICE-Vol. 34-1, 55-64. ASME Paper No. 2000-ICE-257. III. Jonsson, M. and Yan, J. (2000). Exergy and Pinch Analysis of Diesel Engine Bottoming Cycles with Ammonia-Water Mixtures as Working Fluid. International Journal of Applied , 3(2), 57-71. Adapted from a paper published in: Proceedings of ECOS 2000, Enschede, The Netherlands, July 5-7, 2000. IV. Jonsson, M. and Yan, J. (2001). Ammonia-Water Bottoming Cycles: A Comparison between Gas Engines and Gas Diesel Engines as Prime Movers. Energy, 26(1), 31-44. V. Jonsson, M. and Yan, J. (2001). Gas Turbine with Kalina Bottoming Cycle versus Evaporative Gas Turbine Cycle. In: Proceedings of the 2001 International Joint Power Generation Conference, New Orleans, Louisiana, USA, June 4-7, 2001. ASME Paper No. JPGC2001/PWR-19005. VI. Jonsson, M. and Yan, J. (2002). Exergy Analysis of Part Flow Evaporative Gas Turbine Cycles - Part 1: Introduction and Method and Part 2: Results and Discussion. In: Proceedings of ASME Turbo Expo 2002, Amsterdam, The Netherlands, June 3-6, 2002. ASME Paper Nos. 2002-GT-30125 and GT-2002-30126. VII. Jonsson, M. and Yan, J. (2003). Economic Assessment of Evaporative Gas Turbine Cycles with Optimized Part Flow Humidification Systems. Accepted for publication in: Proceedings of ASME Turbo Expo 2003, Atlanta, Georgia, USA, June 16-19, 2003. ASME Paper No. GT2003-38009. VIII. Jonsson, M. and Yan, J. (2003). Humidified Gas Turbines - A Review of Proposed and Implemented Cycles. Manuscript to be submitted for journal publication.

The papers are appended after the summary.

Table of Contents

1 Outline of the Thesis 1 2 Introduction 3 2.1 Background 3 2.2 Objective of the Thesis 4 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines 5 3.1 Theory and Previous on Ammonia-Water Cycles 5 3.1.1 Properties of Ammonia-Water Mixtures 5 3.1.2 The Principle of the Ammonia-Water Cycle 6 3.1.3 Comparison of the Ammonia-Water and the Rankine Cycle 8 3.1.4 Previous Work on Ammonia-Water Cycles 9 3.1.4.1 Cycle Studies 9 3.1.4.2 Existing Ammonia-Water Cycle Power Plants 12 3.1.4.3 Properties of Ammonia-Water Mixtures 14 3.2 Reciprocating Engines for Power Generation 15 3.2.1 The Principle of Gas and Gas-Diesel Engines 15 3.2.2 Power Plants 16 3.3 Studies of Ammonia-Water Bottoming Cycles for Gas and Gas-Diesel Engines 19 3.3.1 Method 19 3.3.1.1 Assumptions and Fixed Parameters 19 3.3.1.2 Cycle Configurations and Optimization 21 3.3.1.3 Exergy Analysis 23 3.3.2 Ammonia-Water Bottoming Processes for Gas Engines 23 3.3.3 Ammonia-Water Bottoming Processes for Gas-Diesel Engines 26 3.3.3.1 First Law Analysis 26 3.3.3.2 Second Law and Pinch Analyses 28 3.3.4 Comparison between Ammonia-Water Bottoming Processes for Gas and Gas-Diesel Engines 32 3.4 Discussion 33 3.4.1 Method and Assumptions 33 3.4.2 Economical and Technical Aspects 34 3.4.3 Results 35 3.4.4 Suggestions for Future Work 36

4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation 37 4.1 Gas Turbines for Power Generation 37 4.2 Theory and Previous Work on Evaporative Gas Turbines 39 4.2.1 Gas Turbines with Air-Water Mixtures as the Working Fluid 39 4.2.1.1 Water-Injected Gas Turbines 40 4.2.1.2 Steam-Injected Gas Turbines 41 4.2.1.3 Evaporative Gas Turbines 42 4.2.2 Previous Work on Evaporative Gas Turbines 43 4.2.2.1 The Humid Air Turbine 44 4.2.2.2 The EvGT Project and Related Studies 46 4.2.2.3 Part-Flow Humidification 47 4.2.2.4 The Humidification Tower 48 4.2.2.5 Water Recovery and Water and Air Quality 49 4.2.2.6 Properties of Air-Water Mixtures 50 4.3 Studies of Evaporative Gas Turbines with Part- and Full-Flow Humidification 51 4.3.1 Method 51 4.3.2 A Comparison between Combined Cycles with Kalina Bottoming Processes and Evaporative Gas Turbines 52 4.3.2.1 Method and Cycle Configurations 53 4.3.2.2 Results 56 4.3.3 Exergy Analysis of Part-Flow Evaporative Gas Turbines 58 4.3.3.1 Method and Cycle Configurations 58 4.3.3.2 Results 58 4.3.4 Economic Analysis of Evaporative Gas Turbines 61 4.3.4.1 Method and Cycle Configurations 62 4.3.4.2 Results 65 4.4 Discussion 70 4.4.1 Method and Assumptions 70 4.4.1.1 Thermodynamic Analysis 70 4.4.1.2 Economic Analysis 71 4.4.2 Economical and Technical Aspects 72 4.4.3 Results 73 4.4.4 Suggestions for Future Work 75 5 Conclusions 77 6 References 79 7 Nomenclature 97 8 Acknowledgments 101

1 Outline of the Thesis

The thesis is a summary of eight technical papers, which are appended to the thesis. In Chapter 2, the general background to the study, the two power cycles investigated, and the objective of the thesis are presented. In Chapter 3, the ammonia-water cycle is addressed. In Section 3.1, the properties of the ammonia-water working fluid, the principle of the ammonia- water cycle, the reasons for the high efficiency of the ammonia-water cycle compared with the steam Rankine cycle, and some of the previous work on ammonia-water cycles in different applications are discussed. In Section 3.2, reciprocating internal combustion engines for power production are described. The principles of spark-ignition gas engines and compression-ignition gas-diesel engines are explained and some examples of reciprocating engine power plants, where engine is recovered, are given. Section 3.3 presents the calculations and the results from Papers I-IV concerning ammonia-water bottoming cycles when used with reciprocating engines. In Section 3.4, the methods, some technical and economical aspects of the ammonia-water cycle, the results, and some suggestions for further work are discussed. In Chapter 4, the evaporative gas turbine cycle is investigated. In Section 4.1, gas turbines for power generation are described. In Section 4.2, which to some extent is based on Paper VIII, different gas turbine cycles with air-water mixtures as the working fluid are discussed along with their principles and some previous work on evaporative gas turbines. In Section 4.3, the methods and results of the studies on part- and full-flow evaporative gas turbines for power generation presented in Papers V-VII are described. The methods, some technical and economical aspects of the evaporative cycle, the results, and some suggestions for further work are discussed in Section 4.4. The conclusions of this thesis are drawn in Chapter 5. Electrical efficiencies given in this work are based on the lower heating value of the fuel.

1

2 Introduction

2.1 Background

The demand for energy and electricity increases steadily. The International Energy Agency (IEA) predicts that the energy and electricity demands in the world will increase by 1.7 % and 2.4 % per year, respectively, from 2000 to 2030 (IEA, 2002). In 2030, the IEA forecasts that fossil fuels will still account for the largest part of the energy demand and most of the new power generating capacity will be natural gas-fired combined cycles. Hence, for a sustainable use of the available natural resources, efficient, low-cost and low-emission processes for the conversion of fuels into power and heat should be developed. Natural gas is considered the most environmentally friendly of the fossil fuels and it can efficiently be converted into power by reciprocating internal combustion engines1 and gas turbines. Recovering the waste heat in the engine or turbine exhaust gas enhances the efficiency of the power generation process. In this thesis, two innovative power cycles that employs this waste heat for increased power generation and efficiency have been investigated: the Kalina, or ammonia-water, cycle, first proposed by Kalina (1983), and the evaporative gas turbine cycle. The working fluids in these cycles are mixtures, which result in efficient heat recovery and high electrical efficiencies. In Sweden, the present demand for electricity is mainly met by hydro and nuclear power stations, covering approximately one half of the demand each (SEA, 2002). Wind power stations and power generation from biomass, coal and oil contribute with below 10 % to the total yearly electricity generation. Condensing oil-fired power plants and gas turbines are primarily used as peak-load capacity. However, there are plans to phase out the nuclear power plants. In this case, replacement power generating capacity is required and increased power generation in biomass-based, hydro or wind power stations can probably not wholly replace the lost generation capacity. One possible alternative is natural gas-based power production, since natural gas can be imported from countries near Sweden. Both ammonia-water and evaporative cycles could be used for efficient power generation from natural gas, in Sweden and in the world at large, where the natural gas consumption is predicted to double between 2000 and 2030 and the largest part of the increase will be used for power generation (IEA, 2002). The ammonia-water cycle has been studied since 1988 at the Division of Energy Processes, Royal Institute of Technology. Ammonia-water processes when used as bottoming cycles for gas turbines, industrial waste heat power cycles and direct-fired cycles have been investigated in addition to the thermophysical properties of ammonia-water mixtures. The evaporative gas turbine has been

1That is, spark-ignition and compression-ignition engines.

3 Advanced Power Cycles with Mixtures as the Working Fluid

studied at the Division since the beginning of the 1990s. Evaporative cycles fueled with natural gas, gasified biomass, externally-fired with biomass, closed evaporative cycles, water recovery from the exhaust gas, air and water quality issues, the humidification tower and the thermophysical properties of air-water mixtures have been studied. In addition, the Division has participated in the construction of an evaporative gas turbine pilot plant.

2.2 Objective of the Thesis

The objective of the present work was to investigate advanced power cycles with the potential of improved performance compared with today’s technology for small-to-mid-size power generation. The ammonia-water cycle is a further development of the steam Rankine cycle that has a higher efficiency for power generation than the steam cycle for several applications. The advantage of the ammonia-water cycle over the steam cycle is most pronounced for a moderate- sensible heat source with a large temperature drop. The waste heat generated by a reciprocating internal combustion engine has these characteristics; thus, an ammonia-water bottoming process could increase the total electrical efficiency more than a steam bottoming process. However, ammonia-water bottoming cycles for reciprocating engines have not been studied in detail previously. Therefore, ammonia-water bottoming cycles for gas and gas-diesel engines were designed and compared with steam cycles in this thesis. Thermodynamic analyses, including an exergy analysis, and a simple economic analysis were used to explore the advantages of the ammonia-water cycle compared with the steam cycle as an engine bottoming cycle. The total power outputs of the combined cycles varied between 18 MWe and 55 MWe in the studies performed. The evaporative gas turbine is an advanced power cycle with a high specific power output (kJ/kg air) and high electrical efficiency. The evaporative cycle has the potential of less expensive power generation compared with a combined cycle, since the evaporative cycle efficiency is approximately equal to the efficiency of a combined cycle, the specific power output is higher for the evaporative cycle than for the combined cycle, and the cost for the steam turbine of a steam bottoming cycle is eliminated for an evaporative cycle. In this thesis, evaporative cycles were investigated thermodynamically, by energy and exergy analyses, and economically to further evaluate the cycle. The studies performed focused on optimizing the design of the humidification and heat recovery system, with special regard to the amount of air that should be passed through the humidification tower. Part-flow evaporative cycles with steam injection were compared with full-flow evaporative, steam-injected and combined cycles. The power outputs of the investigated cycles varied between 16 MWe and 108 MWe.

4

3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

In the first part of this thesis (Chapter 3), ammonia-water cycles utilized as bottoming cycles for reciprocating internal combustion engines are investigated. In Section 3.1, the properties of the ammonia-water cycle working fluid and the principle of the ammonia-water cycle are addressed along with the characteristics of the ammonia-water cycle that enables more efficient power production compared with a steam Rankine cycle, and previous work in the area. Section 3.2 reports on power production with reciprocating engines. In Section 3.3, the studies of gas and gas-diesel engine bottoming cycles, presented in Papers I-IV, are described. In Section 3.4, the methods and assumptions, economical and technical aspects, the results and the need for further work are discussed.

3.1 Theory and Previous Work on Ammonia-Water Cycles

3.1.1 Properties of Ammonia-Water Mixtures The properties of ammonia and water are suitable for a power cycle working fluid mixture. For example, the substances are soluble in each other and easily separable. Ammonia and water have different boiling ; thus, the ammonia-water mixture evaporates over a large temperature range, which is an advantage for power generation from a sensible heat source. Both substances are inexpensive and extensively used in industry. Ammonia and water have approximately the same molecular weights; hence, a steam turbine can be used in the ammonia-water cycle with only minor alterations (Kalina and Leibowitz, 1987). Table 3.1 gives some properties of ammonia and water.

Table 3.1. Some properties of ammonia and water (Reid et al., 1987).

Ammonia (NH3) Water (H2O) Molecular weight [kg/kmol] 17.0 18.0 Boiling point at 1.013 bar [K] 239.8 373.2 Freezing point at 1.013 bar [K] 195.4 273.2 Critical temperature [K] 405.5 647.3 Critical pressure [bar] 113.5 221.2

5 Advanced Power Cycles with Mixtures as the Working Fluid

vapor vapor

f d e c b azeotropic liquid a Temperature [K]Temperature composition [K]Temperature

x=0 Mass fraction [kg/kg] x=1 x=0 Mass fraction [kg/kg] x=1 Figure 3.1. Temperature-composition diagram for an azeotropic (left) and a non-azeotropic mixture (right) at a constant pressure. The mass fraction of the more volatile component is indicated on the x-axis.

Ammonia and water form a non-azeotropic mixture. The difference between an azeotropic mixture and a non-azeotropic mixture is illustrated in Figure 3.1, which shows temperature-composition diagrams for the two types of mixtures. In the figure, the upper line represents the state of saturated vapor at the dew point temperature, which is where liquid starts to condense from the vapor. The lower line in the diagram is the saturated liquid line, where the mixture is at the bubble point temperature and vapor starts to form from the liquid mixture. Thus, in between the lines, there is a two-phase mixture. For a pure substance, the bubble point and the dew point are identical. Therefore, for the pure compositions in the diagrams (i.e., x=0 and x=1), the saturation lines meet. For an azeotropic mixture, there is one composition, called the azeotropic composition, where the liquid and vapor in equilibrium have the same composition and temperature. For a mixture of this composition, the boiling liquid forms vapor of the same composition and temperature as the liquid and the mixture behaves as a pure substance. For a non- azeotropic mixture, the temperature and composition continuously change during boiling, as illustrated by the points a to f in Figure 3.1. When the mixture starts to boil at the bubble point temperature, given by the point a, a vapor that is richer in the more volatile component is formed with a composition given by the point b. When the mixture continues to boil, the temperature increases and the point c is reached, where the concentrations of the remaining liquid and of the vapor formed are given by the points d and e, respectively. Eventually, the point f is reached, where the mixture is a saturated vapor at the dew point temperature and the concentration of the vapor is the same as the concentration the liquid had at the beginning of the evaporation process.

3.1.2 The Principle of the Ammonia-Water Cycle The simplest possible ammonia-water cycle is shown in Figure 3.2. Compared with a steam Rankine cycle, the exhaust gas heat recovery part of the ammonia-water cycle has the same basic layout, while the distillation-condensation subsystem (DCSS) of the ammonia-water cycle differs to the steam cycle condensation

6 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

turbine ∼ HRVG generator DCSS Working mixture superheater separator Basic mixture NH3-enriched vapor NH3-lean solution evaporator reheater Exhaust gas Charge air Jacket water economizer Lubricating oil Cooling water low- high-pressure pressure condenser condenser

Figure 3.2. The simplest possible ammonia-water cycle, configuration Ia in Papers I-IV, and a legend for Figures 3.2, 3.6, 3.8 and 4.3.

system. The DCSS enables the working fluid ammonia concentration to be varied by efficiently recovering energy internally in the cycle for driving a mixture separation process. The working fluid is evaporated and superheated in the heat recovery vapor generator (HRVG) before expansion through the turbine. The relatively high ammonia concentration of the working fluid in the HRVG is called the working mixture composition. A reheater recovers energy in the turbine outlet stream to drive the separation process in the separator. The flash tank separator produces one stream of ammonia-lean saturated liquid and one stream of ammonia-rich saturated vapor. The ammonia-lean liquid stream is throttled and absorbs the working mixture stream from the turbine before condensation in the low-pressure condenser; the resulting concentration is called the basic mixture composition. A pump increases the pressure of the basic mixture condensate and the stream is split: one of the resulting streams is sent to the separator via the reheater and the other stream is mixed with the ammonia-rich vapor from the separator to restore the working mixture concentration. The two-phase working mixture is condensed in the high-pressure condenser and the pressure of the stream is increased to the maximum cycle pressure before the HRVG. A stream with a high concentration of ammonia, like the turbine outlet stream, cannot be condensed by cooling water of a normal temperature, since the high ammonia concentration would result in a very low condensation temperature at the pressure level in the condenser. An increased turbine back pressure is required to raise the condensation temperature; however, this decreases the cycle power output. The DCSS reduces the high ammonia concentration of the turbine outlet stream by absorption in an ammonia-lean solution and allows turbine expansion to a lower pressure than possible in an ammonia-water mixture Rankine cycle, which

7 Advanced Power Cycles with Mixtures as the Working Fluid

is a Rankine cycle with a fixed concentration ammonia-water mixture as the working fluid.

3.1.3 Comparison of the Ammonia-Water and the Rankine Cycle A non-azeotropic mixture, like an ammonia-water mixture, boils at increasing temperature and condenses at decreasing temperature if the pressure is constant, while a pure substance, like water, boils and condenses at a constant temperature. The different boiling behaviors of the two working fluid types are illustrated in Figure 3.3, which shows the temperature profiles for an ammonia-water cycle HRVG and a steam Rankine cycle HRSG (heat recovery steam generator) for a sensible heat source. The sliding boiling temperature of the ammonia-water mixture enables closer matching of the working fluid temperature profile to the heat source temperature profile than is possible with a pure substance. The exergy destruction of the process depends on the temperature difference for heat exchange. Hence, the exergy destruction in the ammonia-water cycle HRVG is smaller than in the steam cycle HRSG. The advantage of the ammonia-water cycle’s HRVG is most pronounced for sensible heat sources with large temperature drops. In the condenser, the temperature drop is smaller and the advantage of the ammonia-water cycle over the steam cycle is not as pronounced. At atmospheric pressure, pure ammonia boils at -33 °C and pure water at 100 °C; therefore, the working fluid in the ammonia-water cycle starts boiling at a lower temperature than in the steam cycle and the ammonia-water cycle can extract more energy from a heat source. In addition, the ammonia-water cycle can utilize

400 Heat source 350 Steam Rankine cycle

300 Ammonia-water cycle

250

200

150 Temperature [°C] Temperature 100

50

0 01234567 Transferred heat [MW] Figure 3.3. Temperature profiles of the HRVG in an ammonia-water cycle and the HRSG in a steam cycle with a sensible heat source.

8 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

low-temperature energy, for example, in the DCSS, which the steam cycle cannot use. Another characteristic of the ammonia-water cycle that further increases the power output compared with a steam cycle is that the ammonia-water cycle has a larger potential for internal heat recovery due to the DCSS, where low-temperature energy is recovered internally to drive the separation process.

3.1.4 Previous Work on Ammonia-Water Cycles The ammonia-water cycle can generate power from waste heat from gas turbines, reciprocating engines and industrial processes and from geothermal energy. In addition, the cycle can be used in direct-fired power processes based on solid fuels. Previous studies on the ammonia-water cycle have focused on designing efficient cycle configurations, demonstrating the technology and ammonia-water mixture properties. Kalina, the inventor of the ammonia-water cycle investigated in this work, founded the company Exergy to market the ammonia-water cycle technology. The company has devised and patented a large number of Kalina cycle configurations for the above-mentioned applications. In this work, only cycle configurations proposed by Kalina are called Kalina cycles, while cycle configurations designed by the author are called ammonia-water cycles.

3.1.4.1 Cycle Studies Combined cycles with ammonia-water bottoming processes have been extensively studied. The studies show that the ammonia-water cycle can produce more power than the steam Rankine cycle as a gas turbine bottoming cycle. The first publication on the Kalina cycle (Kalina, 1983) included a combined cycle application. Kalina (1984) investigated several commercial combined cycles and found that the combined cycle electrical efficiency increased by up to 20 % with a Kalina instead of a steam bottoming cycle. El-Sayed and Tribus (1985a) investigated combined cycles with Kalina bottoming processes thermodynamically. EPRI (1986a) investigated combined cycles with power outputs of about 6 MWe. The Kalina combined cycle had 5-14 % higher electrical efficiency and 2-5 % lower cost of electricity than the steam combined cycles. The specific investment cost (USD/kWe) of the Kalina combined cycle was approximately 11 % higher than for a single-pressure steam combined cycle and approximately 4 % lower than for an advanced dual-pressure steam cycle. Kalina (1993) investigated similar Kalina cycles and found 20-30 % higher power outputs for the Kalina bottoming cycles compared with dual-pressure and triple-pressure steam cycles. Kalina and Leibowitz (1987) studied combined cycles where the bottoming cycle power output was approximately 200 MWe. A Kalina bottoming cycle had 16-30 % higher power output than a triple-pressure steam bottoming cycle. Kalina et al. (1991) presented a 227 MWe Kalina combined cycle with a net efficiency of 52.8 %. The Kalina cycle generated 16 % more power than a dual-pressure steam bottoming cycle. The investment cost for the Kalina bottoming cycle was almost 20 % higher and the specific investment cost was 2 % higher than for the steam bottoming cycle. The Kalina cycle turbine was less expensive than the steam turbine, since the

9 Advanced Power Cycles with Mixtures as the Working Fluid

volumetric flow rate in the low-pressure part of the ammonia-water turbine was much smaller than in the steam turbine, while the Kalina cycle heat exchangers were more expensive. However, the additional power output of the Kalina cycle gave an economic benefit compared with the steam cycle. Park and Sonntag (1990) calculated a 28 % higher power output, 5 % higher first law efficiency and 15 % higher second law efficiency for a Kalina bottoming cycle for a gas turbine compared with a single-pressure steam bottoming cycle. Marston and Hyre (1995) investigated combined cycles with power outputs of about 240 MWe. Combined cycles with Kalina bottoming processes of different complexity had 1-4 % higher power outputs and efficiencies compared with combined cycles with triple- pressure steam bottoming processes. The efficiency was 51.9 % for the most complex Kalina combined cycle and 49.9 % for the triple-pressure steam combined cycle. Ammonia-water combined cycles for of power (about 30 MWe) and district heating were studied by Olsson et al. (1991). The ammonia-water cycle generated more heat and power than a single-pressure steam cycle and low district heating temperatures were found to be to the advantage of the ammonia-water cycle. General Electric has had a licensing agreement with Exergy for Kalina combined cycles and performed several studies of this application. For example, a 412 MWe Kalina combined cycle with a net efficiency of 59.6 % was compared with a 396 MWe triple-pressure reheat steam combined cycle with an efficiency of 57.3 %. The Kalina combined cycle showed a higher investment cost than the steam combined cycle; however, for high fuel prices (i.e., over 4 USD/GJ) and long operation time (i.e., more than 3,500 hours/year), the Kalina cycle had a net economic benefit due to its higher efficiency (Bjorge et al., 1997). Bjorge et al. (1996) compared 500 MWe combined cycles: a Kalina combined cycle had an efficiency of 58.4 %, while a combined cycle with a triple- pressure reheat steam bottoming process had an efficiency of 55.8 %. Smith et al. (1996) presented a detailed design and operational features for a Kalina combined cycle. Few studies of ammonia-water bottoming cycles for reciprocating engines have been found in literature. In the first Kalina cycle publication (Kalina, 1983), diesel engine ammonia-water bottoming cycles could generate 26-49 % more power than a steam cycle. Kalina cycle configurations using only the diesel engine exhaust gas as a heat source or both the exhaust gas and the jacket water were presented. Exergy (2001) presented a diesel engine Kalina bottoming cycle where the jacket water preheats the ammonia-water mixture before the HRVG, where the exhaust gas is the heat source. Two diesel engine power plants of different sizes were investigated. In the first case, a steam bottoming cycle had a power output of 8 MWe, while a Kalina cycle could generate 13.3 MWe. In the second case, a steam bottoming cycle could generate 1.87 MWe and a Kalina cycle 3.24 MWe. The ammonia-water cycle has also been considered as a bottoming cycle for a molten carbonate fuel cell (Richter and Lobachyov, 1996). Another investigated application of the ammonia-water cycle is recovery of industrial waste heat for power production. In a study by Olsson et al. (1994), it

10 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

was shown that the ammonia-water cycle could generate more power from iron and steel industry waste heat than a steam cycle, especially when the minimum outlet temperature of the heat source was low. Isaksson et al. (1999) investigated power generation from waste heat in a pulp mill. For this application, an ammonia- water cycle generated 2.6 MWe, an generated 2.0 MWe and a steam Rankine cycle generated 0.4 MWe. Bisio (1992) investigated Kalina cycles for power generation from waste heat in ironworks and steelworks. The ammonia-water cycle can use geothermal heat sources for power production. This application was already considered in the first Kalina cycle publication (Kalina, 1983). Desideri and Bidini (1996) compared steam flash cycles, Rankine cycles with ammonia or ammonia-water mixtures as working fluids and Kalina cycles without separators for geothermal applications. Kalina and Leibowitz (1989) and Leibowitz and Markus (1990) presented some of Exergy’s studies on geothermal Kalina cycles. In these studies, it was shown that the Kalina cycle has a higher power output for a specified geothermal heat source compared with organic Rankine cycles and steam flash cycles. The Kalina cycle has a larger heat exchanger area compared with the other cycles; however, this extra cost is offset by the higher power output. Leibowitz and Markus (1990) presented a 40 % lower specific investment cost for a geothermal Kalina cycle compared with an organic Rankine cycle. Ansaldo Energia in Italy has had a licensing agreement with Exergy for geothermal applications of the Kalina cycle. Lazzeri and Bruzzone (1995) compared steam flash cycles with the Kalina cycles designed by Ansaldo Energia for geothermal applications and showed that the Kalina cycles provided over 10 % higher power outputs. Lazzeri (1997) and Lazzeri and Diotti (1998) presented more of Ansaldo Energia’s work in this area. Direct-fired ammonia-water cycles have mainly been studied for coal-fueled applications. Kalina (1989) proposed a coal-fired large-scale Kalina cycle with approximately 15 % higher electrical efficiency than the current reheat steam cycles. Kalina (1991) presented a 100 MWe direct-fired Kalina cycle with almost 20 % higher efficiency and 9 % lower investment cost compared with a steam cycle. Zervos (1992) compared a version of the same Kalina cycle configuration with a steam cycle when both cycles had coal-fired atmospheric fluidized bed combustors. For a 100 MWe power output, the investment cost was approximately 20 % lower and the cost of electricity was 11 % lower for the Kalina cycle compared with the steam cycle. Zervos (1993) compared direct-fired Kalina cycles with different coal-fired power plants: the Kalina cycle and the integrated gasification combined cycle (IGCC) showed the highest efficiencies. Kalina and Pelletier (1996) presented simplified direct-fired Kalina cycles for small-to- medium-scale power generation. Dejfors et al. (1998) found that for biomass- fueled cogeneration of power (23-27 MWe) and district heating, the ammonia- water cycle produced slightly less power than a steam cycle, while for a condensing power plant the ammonia-water cycle had a slightly higher power output (2 %) than a steam cycle. Asea Brown Boveri (ABB) has had a licensing agreement with Exergy for direct-fired applications of the Kalina cycle. Currently, the Japanese

11 Advanced Power Cycles with Mixtures as the Working Fluid

company Ebara holds this license. ABB studied the direct-fired Kalina cycle within the US Department of Energy program for advanced pulverized-coal-fired low- emission systems (LEBS). Some of ABB’s work in this area is described by Regan et al. (1996, 1997). Enick et al. (1997) presented two direct-fired Kalina cycle configurations with low emission boiler systems (LEBS). Rankine cycles with fixed concentration ammonia-water mixtures as the working fluid have been investigated as well. In a study by Olsson et al. (1991), a gas turbine bottoming mixture cycle used for the cogeneration of heat and power generated more heat and power than a steam Rankine cycle. Compared with an ammonia-water cycle, the mixture Rankine cycle produced less heat and power. The advantage of a mixture Rankine cycle over an ammonia-water cycle is its simplicity. The economics of the ammonia-water cycle have been investigated by EPRI (1986a) (combined cycles), Gajewski et al. (1989) (combined cycles), Kalina et al. (1991) (combined cycles), Bjorge et al. (1997) (combined cycles), Kalina and Leibowitz (1989) (geothermal applications), Lazzeri (1997) (geothermal applications), Kalina (1991) (direct-fired applications) and Zervos (1992) (direct- fired applications). In most of these studies, the investment cost for the ammonia- water cycle was higher than for a steam cycle; however, when the ammonia-water cycle generates more power than the steam cycle, a higher investment cost can be tolerated. A major part of the extra investment cost is due to the larger heat exchanger area of the ammonia-water cycle compared with the steam cycle, which is explained by two characteristics of the ammonia-water cycle. Firstly, there are often more heat exchangers in the ammonia-water cycle than in the steam cycle and secondly, the more closely matched temperature profiles in the ammonia- water cycle heat exchangers result in a smaller driving for the heat transfer process, thus a larger heat exchanger area is required. On the other hand, the turbine in the ammonia-water cycle should be smaller and less expensive than the steam cycle turbine. This is explained by the smaller volumetric flow rate in the ammonia-water turbine compared with the steam turbine, since the pressure in the ammonia-water cycle is often higher than in the steam cycle, although the mass flow rate in the ammonia-water turbine is larger, as more working fluid can be evaporated than in the steam cycle. A standard backpressure steam turbine can be used since ammonia and water are similar substances. In the steam cycle, the turbine back pressure is often below the atmospheric pressure, resulting in large final turbine stages and the necessity of a vacuum system. The pressure in the ammonia-water cycle is most often below critical and above atmospheric; hence, the turbine back pressure is higher than the atmospheric pressure and no vacuum system is required.

3.1.4.2 Existing Ammonia-Water Cycle Power Plants In 1992, a Kalina demonstration plant started operation at the US Department of Energy’s Energy Technology Engineering Center in California. This plant, constructed by Exergy, validated the theory of the ammonia-water cycle. At first

12 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

the plant used waste heat at a temperature of approximately 540 °C to generate 3 MWe of power and accumulated 5,200 operation hours. In 1996, a gas turbine replaced the waste heat stream as the Kalina cycle heat source. As a combined cycle, the plant generated 6 MWe and operated for a total of 1,500 hours. The maximum pressure and temperature of the Kalina cycle were 110 bars and 516 °C (Leibowitz and Mirolli, 1997). The design of the cycle is described by Kalina and Leibowitz (1988) and operating experiences are provided by Leibowitz (1993). The plant is no longer operational since the closing of the Energy Technology Engineering Center. A 1.6 MWe geothermal Kalina cycle started operation as a municipal power plant in Húsavík, Iceland, in 2000. The ammonia-water cycle decreases the temperature of the geothermal brine stream from 121 °C to 80 °C and the remaining energy in the stream is used for heating purposes, for example district heating. The plant components, plant operation and safety features are presented by Mirolli et al. (2002). The plant design and economics, compared with an organic Rankine cycles, are described by Leibowitz and Mlcak (1999). Valdimarsson (2002) simulated different power cycles for the conditions in the Húsavík plant. An ammonia-water cycle had approximately 20 % higher power output compared with an organic Rankine cycle and a steam flash cycle. The Icelandic company X-Orka has the license for the Kalina cycle in Europe and markets the cycle for power generation from low-temperature heat sources, for example geothermal energy and waste heat (Húsavík Energy, 2003). There is a Kalina cycle at the Sumimoto Metals Kashima steelworks in Japan (Sumimoto Metals, 2003). The plant, constructed by Ebara, generates 3.1 MWe from 98 °C cooling water (Exergy, 2003). In addition, Ebara has constructed a Kalina cycle demonstration plant that operated in 1999 and 2000 at a refuse incineration plant in Fukuoka, Japan. Here, the Kalina cycle acted as a bottoming cycle to a waste-fired steam cycle. The working fluid in the Kalina cycle was evaporated in the steam cycle condenser and superheated by a fraction of the exhaust gas from the waste incinerator. The Kalina cycle was simplified, for example by removing the separator, to reduce the cost for small-scale power generation. The design power output of the Kalina cycle was 2.6 MWe and the maximum pressure and temperature were 39 bars and 294 °C (Furuya and Mori, 2001a). The demonstration plant was also used for heat exchanger evaluation (Furuya and Mori, 2001b). At the Waseda University in Japan, there is an experimental ammonia-water cycle. The ammonia-water cycle works as a bottoming cycle to a steam cycle that in turn is a bottoming cycle to a gas turbine. The ammonia-water cycle, that can generate 60 kWe, is driven by low-pressure steam from the steam cycle. The system also includes an ammonia-water cycle. The purpose of the plant is to investigate systems for distributed cogeneration of power and hot and cold water. Takeshita et al. (2002) and Amano et al. (2001) presented results from the experimental plant.

13 Advanced Power Cycles with Mixtures as the Working Fluid

3.1.4.3 Properties of Ammonia-Water Mixtures To evaluate ammonia-water cycles, thermophysical properties of the working fluid are required. The term thermophysical properties includes both the thermodynamic properties, for example and , and the transport properties, for example viscosity and thermal conductivity, of the working fluid. To calculate the thermodynamic performance of an ammonia-water cycle, an accurate correlation for the working fluid thermodynamic properties is required. Ammonia-water mixtures have been used in absorption heat pumps and for a long time; however, power generation applications involve higher temperatures and and the correlations available when the Kalina cycle was proposed did not cover this area. In the studies on the ammonia-water cycle found in literature, different thermodynamic property correlations have been used. In the present work, a correlation developed by Stecco and Desideri (1989), based on the work by Ziegler and Trepp (1984) and El-Sayed and Tribus (1985b), has been utilized. In other studies, the correlations developed by El-Sayed and Tribus (1985b), Park (1988), Kouremenos and Rogdakis (1990) and Ibrahim and Klein (1993) have been used. Experimental data to validate the theoretical correlations are scarce. Thorin et al. (1998) listed the available thermodynamic property correlations and compared some of the correlations (i.e., Stecco and Desideri (1989), Ibrahim and Klein (1993), El-Sayed and Tribus (1985b) and Park and Sonntag (1990)). It was found that ammonia-water cycle efficiencies calculated with the different correlations differed only by a few percent. However, properties of saturated liquid and vapor differed more between the correlations. Thorin (2000) compared a new correlation presented by Tillner-Roth and Friend (1998) with the correlations by Stecco and Desideri (1989) and Ibrahim and Klein (1993) regarding saturation properties, and cycle efficiencies. All correlations gave similar results at low pressures and temperatures, while the differences increased at high pressures and temperatures. The largest difference in power cycle efficiencies for different correlations was 4 % corresponding to 1-1.5 percentage points. The correlation by Tillner-Roth and Friend had a more reasonable behavior close to the critical point. For accurate design and sizing of the ammonia-water cycle components, reliable transport properties are needed; however, correlations and experimental data for these are few. Thorin (2001) presented correlations for ammonia-water mixture transport properties found in literature. Different correlations for the thermodynamic properties (i.e., Stecco and Desideri (1989) and Tillner-Roth and Friend (1998)) resulted in a 7 % difference in the total heat exchanger area and area differences of up to 24 % for individual heat exchangers. Different transport property correlations resulted in a 3 % difference in the total heat exchanger area and up to 10 % area differences for individual heat exchangers. The use of ammonia-water mixtures as a power cycle working fluid has raised concerns about possible corrosion and nitriding (i.e., nitrogen reacts with nitride forming elements and forms hard nitrides) of the cycle components and decomposition of ammonia. These aspects were investigated in the Kalina cycle

14 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

demonstration plant at the Energy Technology Engineering Center. Experiments with different materials in the high-temperature parts of the plant did not show any nitriding and corrosion of the cycle components could not be detected (Leibowitz, 1993). Plant operation showed that the rate of ammonia decomposition was very low; the water in the mixture may prevent the thermal decomposition of ammonia. However, copper and copper-based alloys should not be used anywhere in the cycle, since these material are corroded by ammonia. At high temperatures, ammonia may decompose into hydrogen and nitrogen; therefore, Hastelloy (a nickel-based alloy), low-nickel iron alloys and alloys containing cobalt should be avoided in the high-temperature parts of the cycle, since these materials catalyze ammonia decomposition. Otherwise, conventional materials can be used in the major components of the cycle, and carbon steel may be used in the low- temperature parts of the cycle. (Bjorge et al., 1997) Another concern was the safety when operating a power plant with ammonia in the working fluid. Ammonia is a toxic substance; however, the distinct odor ensures that any leakages will be detected before the concentration of ammonia reaches harmful levels. Ammonia is gaseous at atmospheric pressure and lighter than air; hence, it is easy to ventilate. Ammonia is difficult to ignite and has a narrow flammability range. From ammonia synthesis plants, absorption refrigeration processes with ammonia as the working fluid and the use of ammonia as a fertilizer and feedstock there are extensive experiences for the safe use of ammonia. (Bjorge et al., 1997)

3.2 Reciprocating Engines for Power Generation

3.2.1 The Principle of Gas and Gas-Diesel Engines A spark-ignition gas engine works according to the and is fueled with natural gas. The pre-mixed, stoichiometric or lean, air-fuel mixture is supplied to the cylinders before the compression begins. In open chamber engines, the spark plug is located in the combustion chamber. For bores above approximately 0.2 meters and air-to-fuel ratios higher than 1.7 times the stoichiometric air demand, the spark plug cannot generate a sufficient amount of energy to ignite the air-fuel mixture. Therefore, lean-burn engines often have a pre- chamber design, for increased air-to-fuel ratio and engine power output. In the pre-chamber, a spark plug ignites a small amount of pilot gas that ignites the mixture in the cylinder. (Stenhede, 1998a) There are several engine constructions, developed by different engine manufacturers, for fueling diesel engines with natural gas. One example is the gas- diesel engine, or gas injection diesel engine, manufactured by Wärtsilä. This engine works according to a pure diesel combustion principle: natural gas compressed to 350 bars (Ahnger, 1996) is injected directly into the combustion chamber immediately before the end of the compression stroke. The air-fuel mixture is

15 Advanced Power Cycles with Mixtures as the Working Fluid

ignited by a small amount of pilot fuel, corresponding to 3-5 % of the total fuel supplied (Ahnger, 1996), or by glow plugs. Crude oil, diesel oil or heavy fuel oil can be used as pilot fuels. The main fuel for the engine can be either natural gas or the pilot fuel; thus, the fuel flexibility is higher than for a gas engine. The efficiency of a gas-diesel engine depends on the power requirement for the natural gas compression. Diesel engines have roughly the same engine efficiencies as gas engines and large engines have higher efficiencies than small engines. Engine efficiencies range from about 35 % for small engines to over 50 % for the largest low-speed two-stroke diesel engines (Niemi, 1997).

3.2.2 Reciprocating Engine Power Plants Reciprocating engines are employed by utilities and independent power producers for base- and peak-load power generation, for industrial and residential cogeneration of heat and power and as standby power plants, since engines have short startup times. In addition, engines are used for propulsion and mechanical drive. The market for power generating reciprocating engines expanded in the 1990s; the increase is mostly accounted for by small diesel engines for standby applications. The majority of reciprocating engines for power generation have power outputs from 1 MWe to 20 MWe; however, diesel engines with power outputs over 50 MWe are available (MAN B&W, 2003). Small engines (1-3.5 MWe) are typically utilized for standby applications, while engines with power outputs above 10 MWe are mostly employed for continuous power generation (McNeely, 2001). Several engines can be combined in large base-load power plants with outputs of up to 300 MWe. In Sweden, most of the power-producing diesel engines are installed in reserve power plants, although there are some diesel engines for combined power and district heating generation. Gas engines are used for power generation from landfill gas or sludge digestion gas produced in wastewater plants (Anonymous, 1989) and for industrial cogeneration (Niemi, 1997). Most engine power plants are fueled by diesel oil or heavy fuel oil, but the use of gaseous fuels, for example natural gas and landfill gas, is increasing. Solid fuel- fired engines are under development. The investigations of coal as an engine fuel have focused on coal-water slurries (Niemi, 1997; Anonymous, 1989), although dry coal powders have been studied as well (Anonymous, 1989). For biomass, the development work has concentrated on gasification (Rensfelt, 1987) and dry wood powders (Olsson et al., 1998). Nevertheless, biomass can be used for production of other engine fuels: dimethyl ether that can fuel modified diesel engines (Sorenson, 2000), biodiesel that is produced from vegetable oils (Agarwal and Das, 2000) and methanol and ethanol that can be used in both spark-ignition and compression-ignition engines (Winell and Svedberg, 1997). However, these fuels have mostly been considered for transport applications. The amounts of health damaging carbon monoxide, unburned hydrocarbons and particulates generated by a reciprocating engine depend mostly on the combustion efficiency, which is usually high in reciprocating engines; however,

16 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

stoichiometric gas engines may have high levels of carbon monoxide and hydrocarbons. Particulate emissions are very low for gas engines, while oil-fired diesel engines have higher particulate levels (Niemi, 1997). The amount of , which causes soil acidification, depends on the sulfur content of the fuel; natural gas contains very low amounts while heavy fuel oil can contain relatively high levels of sulfur. Nitrogen oxides, which cause soil acidification and photochemical smog, are formed from nitrogen in the combustion air and/or in the fuel. The amount of nitrogen oxides generated from nitrogen in the air is mainly determined by the temperature of the combustion process. The high air-to- fuel ratio in lean-burn gas engines decreases the combustion temperature and the generation of nitrogen oxides is reduced, while stoichiometric gas engines and diesel engines generate higher levels of nitrogen oxides. The power plant efficiency and the fuel determine the specific emission of carbon dioxide, that is the mass of carbon dioxide per generated MJ of electricity. The specific emission of carbon dioxide, which contributes to the global warming, is lower for natural gas than for other fossil fuels, since natural gas has a higher hydrogen-to-carbon ratio than, for example, oil. The waste heat generated by reciprocating engines can be recovered to produce additional power, steam or hot water. One alternative is turbocharging which increases the engine power output by compressing the charge air in a driven by the engine exhaust gas. In modern engines, all of the exhaust gas energy is not required for the turbocharging process. Hence, a part of the exhaust gas stream can be used in a turbine for additional power generation. This so-called turbocompounding process can increase the power output of the engine by up to 5 % (Niemi, 1997). A bottoming cycle, for example a steam or organic Rankine cycle, can use the rest of the exhaust gas energy. Combined cycles with two-stroke low-speed diesel engines and steam bottoming cycles can reach total electrical efficiencies of 54 %, while gas engine combined cycles usually have efficiencies below 47 % (Niemi, 1997). The exhaust gas temperatures of reciprocating engines are rather low; often below 300 °C for large two-stroke low-speed engines and below 350 °C for large four-stroke medium-speed engines, while the exhaust gas temperature of a smaller engine is higher (Niemi, 1997). Diesel engines have lower exhaust gas temperatures than gas engines due to the different combustion principles. Supplementary firing increases the temperature of the exhaust gas, thus raising the power output of a bottoming cycle or the steam or hot water generation in a cogeneration system. The steam cycle is the most common bottoming process for reciprocating engines and some examples of such power plants will be given here. Two slow- speed diesel engines with a steam bottoming cycle started operation in Florida, USA, in 1984 (Osenga, 1984). The bottoming process increased the power output by 6-8 % to a total output of slightly above 40 MWe. In a Macau power plant, there are six two-stroke low-speed diesel engines; the two largest engines, installed in 1996, have power outputs of 50 MWe each (Cordeiro and Jensen, 1996). Turbocompounding and bottoming steam cycles provide a total efficiency of 46 %

17 Advanced Power Cycles with Mixtures as the Working Fluid

and increase the total power output by 5 %, from about 220 MWe to 231 MWe. In Pakistan, a base-load heavy fuel oil-fired power plant with eight diesel engines (120 MWe) and one steam turbine (8.3 MWe) and a net efficiency of 46 % began operation in 1997 (Anonymous, 1998). Applications for combined generation of heat and power can also be found. In Pennsylvania, USA, a base-load plant with three gas-diesel engines (5.4 MWe each) and one steam turbine (1.4 MWe) has a net electrical efficiency of 38.8 % (O’Keefe, 1995); the plant generates hot water for a greenhouse as well (Anonymous, 1992). A heavy fuel oil-fired plant in Vaasa, Finland, that generates power and district heating with two diesel engines (11.3 MWe and 22.6 MWe) and a dual-pressure steam bottoming cycle (5 MWe) started commercial operation in 1998 (Wärtsilä, 2003). When operated for maximum power output, the net power output of the plant is 38 MWe, the district heating output is 12 MWh, the electrical efficiency is 53.6 % and the total efficiency is 70.9 %. When operated for a high heat output, the power output is 33 MWe, the district heating output is 30 MWh, the electrical efficiency is 47.0 % and the total efficiency is 90.0 %. Since the exhaust gas temperature of reciprocating engines is relatively low, organic Rankine bottoming cycles have attracted interest. EPRI (1986b) investigated one steam and two organic Rankine bottoming cycles for diesel engines. Five of these systems were installed in US power plants in the 1970s and 1980s, but only the steam bottoming cycle was offered for sale at the time of EPRI’s report. The French company Bertin et Cie manufactured organic Rankine bottoming cycles for diesel engines in the 1970s and the beginning of the 1980s (Gröndalen, 1987) and one of their systems was installed in a diesel power plant in Corsica (Verneau, 1984). The main competition for natural gas-fueled engines is gas turbines (Niemi, 1997); nevertheless, engines have some advantages over gas turbines for small- scale power generation. The engine has a higher electrical efficiency than the gas turbine for small power outputs, since the engine has a higher combustion temperature, and the part-load efficiency is higher: the engine efficiency is almost unaltered down to approximately 40 % part-load. The engine efficiency is not as dependent on the ambient conditions as the gas turbine efficiency. In addition, the engine can be started and stopped quickly and is therefore suitable for applications with varying power demands. An engine has a higher power-to-heat ratio than a gas turbine for combined heat and power production, which is an advantage if the price of electricity is high. In cogeneration applications, the engine exhaust gas is used for steam generation, while hot water is generated from the engine cooling system. In a gas turbine, all the waste heat is contained in the exhaust gas and the exhaust temperature is often higher than for an engine; hence, gas turbines are better suited for cogeneration of steam than engines. An advantage of the gas turbine is that, for the same engine volume, a turbine can handle larger working fluid flows than a reciprocating engine, since the turbine process is continuous and the engine process is intermittent (Wilson, 1984). Hence, the gas turbine has a high power-to-weight ratio. A reciprocating engine power plant is often of a modular

18 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

construction with several engines. If the power demand varies, a power plant with several power-producing units is more flexible than a plant with one major unit. A spark-ignition gas engine does not require compressed natural gas, which is an advantage compared with the gas turbine and the gas-diesel engine. For power outputs below 50 MWe, engine combined cycle power plants are competitive with gas turbine combined cycles (Niemi, 1997).

3.3 Studies of Ammonia-Water Bottoming Cycles for Gas and Gas-Diesel Engines

In Section 3.3, the studies of ammonia-water bottoming processes for gas and gas- diesel engines presented in Papers I-IV are described.

3.3.1 Method

3.3.1.1 Assumptions and Fixed Parameters The process simulation program IPSEpro developed by SimTech Simulation Technology (IPSEpro) was used for the cycle simulations in Papers I-IV. The thermodynamic properties of the ammonia-water mixture were calculated with a library of subroutines developed by Stecco and Desideri (1989). The gas engines 16V25SG and 18V34SG and the gas-diesel engines 18V32GD and 18V46GD, manufactured by Wärtsilä, were used in the simulations. The first set of numbers of the engine name is the number of cylinders and the second set of numbers is the cylinder in centimeters. The assumptions and fixed parameters used for the engines and the bottoming cycles are shown in Table 3.2; more details can be found in Papers I-IV. The calculations in Paper I were based on five gas engines and one bottoming cycle. The calculations in Papers II and III were based on three gas-diesel engines and one bottoming cycle. The calculations in Paper IV were based on a fuel input of 64-71 MW and the number of engines depended on the engine size, as shown in Table 3.2. The exhaust gas and charge air were assumed to be ideal and the lubricating oil was assumed to be an incompressible liquid. The pressure drops of the waste heat streams were ignored. The maximum pressure in the ammonia-water cycles was set to 115 bars. At pressure levels above this value, the properties of the ammonia-water mixture are not calculated correctly by the library of subroutines used in this work, since the critical pressure of the mixture has been reached. The power required by the engine driven pumps has been included in the engine efficiencies and all calculations were based on the lower heating value of the fuel. Equations for calculating the bottoming cycle power output and the total electrical efficiency of the engines and the bottoming cycle are shown in Papers I, II and IV. The natural gas supplied to the gas-diesel engines must be compressed to 350 bars before

19 Advanced Power Cycles with Mixtures as the Working Fluid

Table 3.2. Input data for the gas engines in Papers I and IV (Stenhede, 1998b), the gas-diesel engines in Papers II-IV (Maunu, 1999) and the bottoming cycles in Papers I-IV. Gas engines Gas-diesel engines 16V25SG×91 18V34SG×51 18V32GD×41 18V46GD×21 P [MW] 28.1 (3.1) 27.5 (5.5) 26.0 (6.5) 31.4 (15.7) Qfuel [MW] 69.4 (7.7) 65.5 (13.1) 64.1 (16.0) 71.2 (35.6) ηel [%] 40.5 42.0 40.7 44.2 meg [kg/s] 51.3 (5.7) 48.0 (9.6) 59.2 (14.8) 62.2 (31.1) teg, in [°C] 410.0 418.0 321.0 330.0 teg, min [°C] no restriction no restriction 130.0 130.0 2 3 2 3 cp, eg [kJ/kg, K] 1.05 , 1.1 1.05 , 1.1 1.07 1.07 Qeg [MW] - - 12.1 (3.0) 13.3 (6.7) mca [kg/s] 49.5 (5.5) 46.5 (9.3) 57.9 (14.5) 60.7 (30.4) tca, in [°C] 139.0 149.0 200.0 200.0 tca, out [°C] 45.0 45.0 50.0 50.0 cp, ca [kJ/kg, K] 1.015 1.015 1.01 1.01 Qca [MW] 4.7 (0.5) 4.9 (1.0) 8.8 (2.2) 9.2 (4.6) mjw [kg/s] 243.0 (27.0) 250.0 (50.0) 240.0 (60.0) 222.0 (111.0) tjw, in [°C] 96.0 91.0 91.0 91.0 tjw, out [°C] 89.2 86.5 85.2 86.4 pjw [bar] 3.0 3.0 4.0 4.0 Qjw [MW] 7.0 (0.8) 4.7 (0.9) 5.9 (1.5) 4.3 (2.2) mlo [kg/s] 116.1 (12.9) 151.0 (30.2) 112.0 (28.0) 150.0 (75.0) tlo, in [°C] 85.1 73.7 75.9 74.0 tlo, out [°C] 74.0 63.0 63.0 63.0 cp, lo [kJ/kg, K] 2.1 2.1 2.1 2.1 Qlo [MW] 2.7 (0.3) 3.4 (0.7) 3.0 (0.8) 3.5 (1.7) Methane content of natural gas (only significant for gas engines) 0.80 Bottoming cycles pmax, cI, cII and cIII [bar] 115 tcooling water [°C] 15 (in), 25 (out) ∆pHRVG, cI, cII and cIII [bar] 5 tdeaerator, R1 and R2 [°C] 105 ∆ptbn to cnd lp, cI, cII and cIII [bar] 0.01⋅pin aturbine, min [kg/kg] 0.90 ∆pHRSG, R1 and R2 [bar] 0.04⋅pmax ηis, turbine 0.80 ∆teg – vapor [°C] 30 ηgen, tbn, 16V25SG and 18V32GD 0.965 ∆teg - liquid [°C] 15 ηgen, tbn, 18V34SG and 18V46GD 0.97 ∆tca – liquid [°C] 10 ηis, pump 0.80 4 ∆tliquid - liquid [°C] 5 ηmech 0.98 1The values are based on nine 16V25SG engines, five 18V34SG engines, four 18V32GD engines and two 18V46GD engines (as in Paper IV). Values for one engine are shown within parentheses. 225-120 °C 3120-450 °C 4The mechanical efficiency is used for the generator, turbine and pumps in the ammonia-water cycle.

20 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

injection into the cylinders; however, the power needed for the compression was not included in the calculations. In Papers III and IV, a reference state of 288.15 K and 1.013 bars was used in the exergy analysis. The first and second law efficiencies of the bottoming cycles in Paper III were calculated as the bottoming cycle power output divided by the energy or exergy contents of the waste heat streams, as explained in the paper. The total amount of energy or exergy available to the bottoming cycle was used in the calculation of these efficiencies, although all cycles could not use all the energy or exergy. In a gas engine, the only fuel is natural gas and the exhaust gas temperature can be reduced to low temperatures without a risk for corrosion. Therefore, to calculate the amount of energy and exergy in the gas engine exhaust gas in Paper IV, a minimum temperature of 50 °C was assumed. In a gas-diesel engine, the pilot fuel oil contains sulfur and the exhaust gas temperature should not go below 130 °C to avoid corrosion.

3.3.1.2 Cycle Configurations and Optimization Ammonia-water cycle configurations found in literature were used as starting points for the simulations. These single-pressure cycles had been designed as gas turbine and diesel engine bottoming cycles, using one or two waste heat streams from the turbine or engine as heat sources. The engines in this work provide four waste heat streams that can be used as heat sources by a bottoming cycle: the exhaust gas, the charge air, the jacket water (i.e., the engine cooling water) and the lubricating oil. The charge air is cooled to increase the power output of the engine. Consequently, the initial ammonia-water cycles were changed in order to use these heat sources efficiently and maximize the power output. To compare the ammonia-water cycles with the conventional technology, both single-pressure and dual-pressure steam Rankine cycles were simulated. The simulated configurations are summarized in Table 3.3. Configurations called “a” employ only the exhaust gas as a heat source, while configurations called “b” can use all the waste heat streams as heat sources. Configurations Ia and Ib are further developments of the

Table 3.3. Bottoming cycle configurations investigated in Papers I-IV. 16V25SG 18V34SG 18V32GD 18V46GD Only exhaust cIa, cIIa cIa, cIIa cIa, cIIa cIa, cIIa Ammonia-water gas (a) cycles All waste heat cIb, cIIb, cIb, cIIb, cIb, cIIb, cIb, cIIb, streams (b) cIII cIII cIII cIII Only exhaust R1a1 R1a1 R1a, R2a R1a, R2a Steam Rankine gas (a) cycles All waste heat R1b2 R1b2 R1b, R2b R1b, R2b streams (b) 1Paper IV 2Paper I

21 Advanced Power Cycles with Mixtures as the Working Fluid

3.5 optimal version

3.4 five confs. four confs. three confs. six confs. 3.3 five configurations

Power output [MW] 3.2

one configuration 3.1 0.78 0.80 0.82 0.84 0.86 0.88 0.90 0.92 0.94

Working mixture ammonia concentration [kg/kg] Figure 3.4. Power output versus the working mixture ammonia concentration for a number of simulated versions of configuration IIb for the engine 18V32GD.

ammonia-water cycle shown in Figure 3.2 (Kalina, 1983). Configurations IIa and IIb are more complex than configurations Ia and Ib with more internal heat recovery (El-Sayed and Tribus, 1985a). Configuration III has two separators, which result in additional ammonia concentration levels in the cycle (Kalina, 1983). This configuration is always allowed to use all waste heat streams as heat sources. For the steam cycles, “1” implies a single-pressure HRSG and “2” a dual-pressure HRSG. The configurations differ to varying degrees for the four engine models investigated, especially the “b”-configurations. To optimize an ammonia-water cycle with a fixed configuration, the ammonia concentration of the working mixture is held constant while the basic mixture concentration is varied until the maximum power output is determined. The basic mixture concentration is varied in this way for several values of the working mixture concentration until the overall maximum power output is found. When designing ammonia-water cycles for a new application, as in the present work, the optimization process is more complicated. Varying the ammonia concentration of the working fluid for a certain initial cycle configuration changes the optimal cycle layout. Thus, the cycle layout must be changed continually during the optimization process. To illustrate this procedure, the optimization process for the configuration IIb for the engine 18V32GD is shown in Figure 3.4. The ammonia concentration of the working fluid and the cycle layout are changed in a step-by- step process until the optimal configuration with the highest possible power output is found.

22 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

The single-pressure steam cycles were optimized by varying the maximum cycle pressure. To optimize the dual-pressure steam cycles, the high-pressure level in the HRSG was held constant while the low-pressure level was varied until the maximum power output was found. This process was repeated for a number of values of the high-pressure level until the highest overall power output could be determined. During the optimization of the steam cycle configurations, some of the heat exchangers were moved to find the most efficient cycle configurations. A single-pressure ammonia-water cycle has the same number of degrees of freedom available for cycle optimization as a dual-pressure steam cycle. The two main ammonia concentrations, the working mixture and the basic mixture compositions, correspond to the two evaporative pressure levels in the steam cycle. The temperature profile of a boiling or condensing ammonia-water mixture is non-isothermal; thus, the minimum temperature difference can be located anywhere in a heat exchanger. When designing an ammonia-water cycle, all heat exchanger temperature profiles must be verified to ensure that the minimum temperature difference allowed is heeded.

3.3.1.3 Exergy Analysis In Papers III, IV and VI, the ammonia-water and evaporative gas turbine cycles were evaluated with exergy analyses, sometimes called a second law analysis. The first law of thermodynamics states that energy cannot be destroyed; it is only converted between different forms. However, the potential for extracting work from different energy sources, for example 1 MJ of electric power stored in a battery or 1 MJ of in steam, is different and this difference cannot be fully evaluated in an energy analysis. Therefore, the thermodynamic property exergy, emanating from the second law of thermodynamics, is useful, since it quantifies how much of the energy can be converted into work. Unlike energy, exergy can be destroyed due to irreversibilities, which occur in all real processes since they are not completely reversible. Irreversibilities are caused by, for example, heat transfer through a finite temperature difference, spontaneous chemical reactions, unrestrained expansion or friction in viscous fluids and between mechanical components. Exergy analysis is a powerful tool for investigating and comparing different cycle configurations, since the exergy destruction and losses in different cycle components can be computed. The exergy destruction is defined as the exergy destroyed due to internal irreversibilities in the cycle components, for example due to heat transfer, while exergy losses (i.e., external irreversibilities) are defined as the exergy lost from the cycle, for example in the cooling water and the stack. The exergy analyses in this work followed the method described by Moran (1989). Details on the calculation methods can be found in Papers III and VI.

3.3.2 Ammonia-Water Bottoming Processes for Gas Engines In the study presented in Paper I, ammonia-water bottoming cycles for gas engines were compared with single-pressure steam cycles. In addition, a simplified

23 Advanced Power Cycles with Mixtures as the Working Fluid

Table 3.4. Results of the gas engine study based on five gas engines and one bottoming cycle. The bottoming cycle power outputs are shown within parentheses. 16V25SG×5 18V34SG×5 1 1 1 1 Ptot ηel, tot xwork , xbas Ptot ηel, tot xwork , xbas [MW] [%] p2 [bar] [MW] [%] p2 [bar] Engines 15.6 40.5 - - 27.5 42.0 - - cIa 18.0 (2.3) 46.6 0.71 0.48 31.6 (4.1) 48.3 0.71 0.48 cIb 18.5 (2.9) 47.9 0.85 0.45 32.4 (4.9) 49.5 0.87 0.48 cIIa 18.0 (2.4) 46.8 0.74 0.49 31.7 (4.2) 48.5 0.74 0.48 cIIb 18.7 (3.1) 48.5 0.88 0.42 32.6 (5.1) 49.7 0.87 0.46 cIII 18.3 (2.7) 47.5 0.74 0.39 32.2 (4.7) 49.2 0.73 0.39 R1b 17.6 (2.0) 45.7 12 - 31.0 (3.5) 47.4 13 - 1ammonia concentration levels [kg/kg] in the ammonia-water cycles 2maximum pressure in the steam cycles

55 16V25SG 50 18V34SG 45

40

35

30

25

20

15

Difference in power output [%] 10

5

0 cIa cIb cIIa cIIb cIII Figure 3.5. Power outputs of the ammonia-water bottoming cycles compared with the power outputs of the steam bottoming cycles in the gas engine study.

economic study was included to evaluate the competitiveness of the ammonia- water cycle compared with the steam cycle. Some simulation results are presented in Table 3.4 and in Figure 3.5, the power outputs of the ammonia-water bottoming cycles are compared with the steam cycle power outputs. The differences between ammonia-water and steam bottoming

24 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

gen tbn ∼ sh sep rht ca1 jw ca2 ca3 eva lo

eco ca4

cnd hp rht

rht

cnd lp

Figure 3.6. The ammonia-water bottoming cycle configuration IIb for the gas engine 16V25SG. A legend can be found in Figure 3.2.

cycles for different engines in Figures 3.5, 3.7 and 3.12 were computed as the difference in power output between the ammonia-water and the steam bottoming cycles, divided by the power output of the steam bottoming cycle. Figure 3.5 shows that the ammonia-water bottoming cycles generated 17-54 % more power than the steam cycles. Configuration IIb for the gas engine 16V25SG, shown in Figure 3.6, had the highest power output compared with a steam cycle. The other cycle configurations can be found in Paper I. The exhaust gas temperature after the HRVG varied between 53 °C and 79 °C for the ammonia-water cycles and was approximately 172 °C for the steam cycles. This shows that the ammonia-water cycles could recover more exhaust gas energy than the steam cycles. The configurations Ib, IIb and III could use almost all of the energy in the charge air, jacket water and lubricating oil streams, whereas the steam cycles could only use a small part of this energy. As there is no limitation on the exhaust gas temperature, exhaust gas could be used both in the HRSG and for preheating the condensate before the deaerator in the steam cycle, in contrast to the steam cycles in Paper I that utilize other waste heat streams than exhaust gas for condensate preheating. The use of only exhaust gas or all waste heat streams as heat sources in the steam cycle has no influence on the power output. Configuration III was originally developed as a diesel engine bottoming cycle. However, the results of this study show that configuration II is a more suitable engine bottoming cycle configuration.

25 Advanced Power Cycles with Mixtures as the Working Fluid

In most studies found in literature, the investment cost was higher for the ammonia-water cycle than for the steam cycle. The ammonia-water cycles in this study generated more power than the steam cycles; hence, a higher investment cost can be tolerated for the same pay-off-time. The larger heat exchanger area of the ammonia-water cycle accounts for the major part of the extra investment cost for an ammonia-water cycle compared with a steam cycle. Therefore, the cost of the extra heat exchanger area was calculated and compared with the extra investment cost that could be allowed for the ammonia-water cycle because of its higher power production. It was assumed that the investment cost per heat exchanger unit area, shown in Paper I, was the same for the ammonia-water and the steam cycles. The calculations showed that the extra heat exchanger area of the ammonia- water cycles accounted for 10-15 % of the extra allowable investment cost. Thus, the higher investment cost of the ammonia-water cycle compared with the steam cycle might be compensated for by the extra amount of power generated by the ammonia-water cycle. The turbine accounts for a large part of the investment cost for a bottoming cycle and this cycle component should be considered when comparing the investment costs for the ammonia-water and the steam cycles. In this study, the mass flow rate through the turbine in the ammonia-water cycles was 55-82 % larger than in the steam cycles, since the ammonia-water cycles could recover more energy. However, the condenser pressure was approximately 0.04 bars in the steam cycles and 1.3-2.5 bars in the ammonia-water cycles, which resulted in 94-97 % lower volumetric flow rates in the low-pressure part of the turbine in the ammonia-water cycles compared with the steam cycles. This indicates that the ammonia-water cycle turbine may be less expensive than the steam cycle turbine.

3.3.3 Ammonia-Water Bottoming Processes for Gas-Diesel Engines

3.3.3.1 First Law Analysis In the study presented in Paper II, ammonia-water bottoming cycles for gas-diesel engines were compared with single-pressure and dual-pressure steam cycles. The cycle layouts differ from the configurations designed for the gas engines in Paper I, since the temperatures and mass flow rates of the waste heat streams differ for the gas and gas-diesel engines, as can be seen in Table 3.2. In Table 3.5, some results of the study are presented. In Figure 3.7, the power outputs of the ammonia-water and the steam cycles are compared. Ammonia-water and steam cycle configurations using only exhaust gas as a heat source (“a”- configurations) have been compared with each other and cycle configurations using all heat sources (“b”-configurations) have been compared with each other. The ammonia-water cycles have been compared with both single-pressure (“1-p”) and dual-pressure (“2-p”) steam cycles. All ammonia-water cycles generated more power than the corresponding steam cycles, except for the configuration Ia when compared with the dual-pressure steam cycle configuration 2a. Configuration IIb

26 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

Table 3.5. Results of the gas-diesel engine study based on three gas- diesel engines and one bottoming cycle. The bottoming cycle power outputs are shown within parentheses. 18V32GD×3 18V46GD×3 1 1 1 1 Ptot ηel, tot xwork , xbas , Ptot ηel, tot xwork , xbas , 2 2 2 2 [MW] [%] phi [bar] plw [bar] [MW] [%] phi [bar] plw [bar] Engines 19.5 40.7 - - 47.2 44.2 - - cIa 21.4 (1.9) 44.6 0.71 0.51 51.5 (4.3)48.2 0.68 0.49 cIb 22.8 (3.2) 47.4 0.92 0.67 54.1 (6.9)50.7 0.93 0.72 cIIa 21.6 (2.0) 44.9 0.72 0.49 51.8 (4.6)48.5 0.70 0.46 cIIb 23.0 (3.5) 47.9 0.89 0.51 54.6 (7.4)51.1 0.88 0.52 cIII 22.7 (3.2) 47.3 0.92 0.67 54.0 (6.8)50.6 0.91 0.67 R1a 21.4 (1.8) 44.5 4.7 - 51.3 (4.1) 48.0 5.1 - R1b 21.9 (2.4) 45.6 2.4 - 52.4 (5.2) 49.1 2.6 - R2a 21.5 (2.0) 44.8 9.3 2.0 51.7 (4.5)48.4 10.3 2.0 R2b 22.3 (2.8) 46.5 8.7 1.7 53.3 (6.2)50.0 9.0 1.7 1ammonia concentration levels [kg/kg] in the ammonia-water cycles 2pressure levels in the steam cycles

50 18V32GD 1-p 45 18V46GD 1-p 40 18V32GD 2-p 35 18V46GD 2-p 30 25 20 15 10 5

Difference in power output [%] 0 -5 -10 cIa cIb cIIa cIIb cIII Figure 3.7. Power outputs of the ammonia-water bottoming cycles compared with the power outputs of the steam bottoming cycles in the gas-diesel engine study.

was the best ammonia-water cycle configuration; it produced 43-47 % more power than a single-pressure steam cycle and 20-24 % more power than a dual-pressure steam cycle. Configuration IIb for the gas-diesel engine 18V32GD is shown in Figure 3.8. The rest of the cycle configurations can be found in Paper II.

27 Advanced Power Cycles with Mixtures as the Working Fluid

gen tbn ∼ sh sep rht jw ca3 lo

eva ca4

rht ca1

rht eco rht

ca2 cnd hp cnd rht lp

Figure 3.8. The ammonia-water bottoming cycle configuration IIb for the gas-diesel engine 18V32GD. A legend can be found in Figure 3.2.

The largest differences in power output between the ammonia-water and the steam cycles occurred in cases using all heat sources. This is illustrated by the difference in electrical efficiency between the “a”-version and “b”-version of a specific configuration: this difference was 2.4-3.0 percentage points for the ammonia-water cycles and 1.0-1.6 percentage points for the steam cycles. In the cases where all heat sources were integrated (“b”-configurations), the ammonia- water cycles could utilize all the useful energy in the streams, except for configuration IIb, which could not use all the energy in the charge air. The steam cycles could only use the exhaust gas and the charge air as heat sources. Increasing the temperature of the condensate from the high-pressure condenser for configurations Ia, Ib and III improved the power output of the cycles. For the configuration III, the highest power output was attained when the condenser before the HRVG was removed.

3.3.3.2 Second Law and Pinch Analyses In the study presented in Paper III, exergy analyses were performed on the bottoming cycle configurations designed in Paper II to evaluate the exergy destruction in the different cycle components. Only the thermomechanical, or physical, exergy was calculated. The bottoming cycles are closed systems and the streams external to the cycles do not undergo any chemical reactions; hence, it is not necessary to calculate the chemical exergy for an exergy analysis. To investigate

28 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

Table 3.6. First and second law efficiencies for the gas-diesel engine bottoming cycles. 18V32GD 18V46GD 18V32GD 18V46GD

ηI [%] ηII [%] ηI [%] ηII [%] ηI [%] ηII [%] ηI [%] ηII [%] cIa 9.3 30.6 10.3 32.9 R1a 8.9 29.3 9.8 31.2 cIb 15.6 51.5 16.4 52.6 R1b 11.4 37.7 12.3 39.4 cIIa 9.8 32.4 10.9 35.0 R2a 9.7 32.1 10.7 34.3 cIIb 16.9 55.7 17.7 56.5 R2b 13.6 44.7 14.6 46.8 cIII 15.4 50.9 16.3 52.0

45 18V32GD 40 18V46GD

35

30

25

20

15 Exergy destruction [%] 10

5

0 cIa cIb cIIa cIIb cIII R1a R1b R2a R2b Figure 3.9. Total component exergy destruction as a percentage of the total exergy available to the bottoming cycle.

how well the heat exchanger networks of the cycles had been designed, pinch analyses in the form of temperature profiles were performed on some of the cycle configurations. The first and second law efficiencies of all configurations are shown in Table 3.6. The table illustrates that the efficiencies of the ammonia-water cycles were higher than for the steam cycles, except for configuration Ia, which had a lower efficiency than the Rankine configuration 2a. A comparison of the second law efficiencies of the ammonia-water and steam cycles shows the same trends as Figure 3.7. The total exergy destruction of all bottoming cycle components due to internal irreversibilities, expressed as a percentage of the exergy available to the bottoming cycle in the four waste heat streams and in the pump work, is shown in Figure 3.9.

29 Advanced Power Cycles with Mixtures as the Working Fluid

0.6 hot streams R2b cold streams R2b )/T [-]

0 0.5 hot streams cIIb cold streams cIIb 0.4

0.3

0.2

0.1 Exergetic temperature scale (T-T scale temperature Exergetic 0.0 0 1020304050 Transferred heat [MW] Figure 3.10. Temperature profiles for the configurations IIb and 2b for the engine 18V32GD with an exergetic temperature scale.

The “b”-configurations that could use a large fraction of the waste heat generated more power than the “a”-configurations, while their exergy destruction was higher. Among the “b”-configurations, configuration IIb produced more power than configurations Ib and III and had a lower exergy destruction. Configuration Ib, for example, could extract slightly more energy from the waste heat streams than configuration IIb; however, configuration IIb was more efficient overall and generated more power than configuration Ib. The sum of the cycle exergetic efficiency in Table 3.6 and the total component exergy destruction in Figure 3.9 for a cycle configuration is not 100 %, since exergy losses in the form of heated cooling water and unused exergy in the waste heat streams were not included in the calculated component exergy destruction. In addition, the cycle exergetic efficiency was calculated as the gross power output divided by the total exergy in the waste heat streams, while the component percentage exergy destruction was calculated as the internal irreversibility in the component divided by the total exergy available to the cycle in the waste heat streams plus the pump work. In Figure 3.10, the temperature profiles of the configurations IIb and 2b for the engine 18V32GD are shown with an exergetic temperature scale. The concept of an exergetic temperature scale was discussed by McIlvried et al. (1998) and in Paper III. The figure shows that more energy is transferred to the ammonia-water cycle than to the steam cycle, as the ammonia-water cycle can use low-temperature heat sources since the ammonia-water mixture starts to boil at a lower temperature than pure water. Since a large part of the gas-diesel engine waste heat was available at low temperatures, this energy could not find efficient use in the steam cycle. Configurations IIb and 2b could utilize all the energy in the exhaust gas. However,

30 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

cIacIb cIIacIIb cIII 0

-5

-10

-15

-20

-25

-30

-35 18V32GD 1-p -40 18V46GD 1-p Difference in exergy destruction [%] exergy destruction in Difference -45 18V32GD 2-p 18V46GD 2-p -50 Figure 3.11. Exergy destruction in the ammonia-water cycle HRVGs compared with the exergy destruction in the steam cycle HRSGs.

the steam cycle could only use some of the energy in the charge air, while the ammonia-water cycle could use most of the charge air energy and all the energy in the jacket water and the lubricating oil. The exergy analysis showed that the HRVG/HRSG was the component for which the exergy destruction differed most between the ammonia-water and the steam cycles. The non-isothermal boiling of a mixture is of most advantage when the temperature drop of the heat source is large, as it is for the exhaust gas stream. The distance between the temperature profiles of the heat source and the working fluid was larger in the steam cycle HRSGs than in the ammonia-water cycle HRVGs, as can be seen in Figure 3.10. A large temperature difference implies a large exergy destruction for the heat transfer process. In Figure 3.11, the HRVG/HRSG exergy destruction for the ammonia-water and the steam cycles has been compared. Ammonia-water cycles that use only the exhaust gas as a heat source have been compared with the corresponding steam cycles, while cycle configurations that use all the heat sources have been compared with each other. The ammonia-water cycles have been compared with both single-pressure (“1-p”) and dual-pressure (“2-p”) steam cycles. A negative value implies that the exergy destruction was smaller in the ammonia-water cycle HRVG than in the steam cycle HRSG. Regarding the exergy destruction in the condensers, there was no clear distinction between the ammonia-water and the steam cycles. The constant temperature condensing of a pure substance was not a disadvantage to the steam cycle when the temperature increase of the cooling water in the condenser was as small as in this study (10 °C).

31 Advanced Power Cycles with Mixtures as the Working Fluid

Configuration IIb for the engine 18V32GD was simplified by removing one of the DCSS reheaters. It was found that the component exergy destruction, the temperature profiles and the power outputs were almost the same for the two versions of configuration IIb. This shows that complex ammonia-water cycles can be simplified and the investment cost reduced without large power output losses.

3.3.4 Comparison between Ammonia-Water Bottoming Processes for Gas and Gas-Diesel Engines In the study presented in Paper IV, the ammonia-water cycles simulated in the first two studies were compared to analyze the differences between bottoming cycles designed for gas and gas-diesel engines. Single-pressure steam cycles were used as the basis for the comparison. The steam cycles used as a comparison to the gas engine ammonia-water cycles were not the same in Paper IV as in Paper I. The layouts of the steam cycles designed for the gas-diesel engines were more realistic than the steam cycle configuration used in Paper I. Therefore, the gas engine steam cycles were redesigned to be comparable to the gas-diesel engine steam cycles. However, the differences in power outputs between the two versions of gas engine steam cycles were very small. Differences between the two engine types result in dissimilar characteristics, for example temperature and mass flow rate, for the engine waste heat sources. These characteristics had a large impact on the bottoming cycle layouts and power outputs. The most important waste heat source was the exhaust gas and its characteristics significantly influenced the bottoming cycle. The gas engine waste

50 16V25SG 45 18V34SG 18V32GD 40 18V46GD 35 30

25

20

15

10 Difference in power output [%] 5

0 cIa cIb cIIa cIIb cIII Figure 3.12. Power outputs of the ammonia-water bottoming cycles compared with the power outputs of single-pressure steam bottoming cycles for the gas and gas-diesel engines.

32 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

heat was mostly in the form of high-temperature exhaust gas; the exhaust gas contained approximately 75 % of the waste heat exergy. For the gas-diesel engines, the greatest part (about 57 %) of the waste heat exergy was in the form of exhaust gas; however, a considerable amount of waste heat was contained in the charge air, which had a lower temperature than the exhaust gas. The exergy content of the waste heat from the gas engines was 32.2 %, whereas the mean value of the exergy content of the waste heat from the two gas-diesel engine models was 30.8 %. Thus, most of the gas engine ammonia-water bottoming cycles generated more power than the gas-diesel engine ammonia-water cycles, when compared with the corresponding steam cycles. In Figure 3.12, the power outputs of the ammonia-water cycles designed for the two types of engines are compared with their respective steam cycles. All gas engine ammonia-water bottoming cycles were compared with steam cycles using only the exhaust gas as a heat source. For the gas-diesel engines, ammonia-water cycles using only exhaust gas as a heat source were compared with the steam cycle 1a, while the ammonia-water cycles using all waste heat sources were compared with the steam cycle 1b. Some of the cycle configurations are shown in Paper IV. When the power outputs of “a”-configurations and “b”-configurations are compared, a clear distinction can be seen between the “a”-configurations for the gas and gas-diesel engines, for which the exhaust gas was the only heat source. The exhaust gas temperature was higher for the gas engines than for the gas-diesel engines. In addition, there was no minimum value for the gas engine exhaust gas temperature, while there was a minimum temperature of 130 °C for the gas-diesel engine exhaust gas due to the risk of corrosion. Therefore, the “a”-ammonia-water cycles had higher power outputs for the gas engines than for the gas-diesel engines, compared with the steam cycles. The same distinction cannot be observed for the “b”-configurations. Configurations utilizing all the waste heat sources were complex and included many heat exchangers, which resulted in varying power outputs for the different cycle configurations and no clear trends could be detected.

3.4 Discussion

In the first part of this thesis, ammonia-water cycles have been investigated as bottoming processes for gas and gas-diesel engines. The ammonia-water cycles were compared with the conventional bottoming cycle used today for reciprocating engines, that is the steam Rankine cycle.

3.4.1 Method and Assumptions The assumptions made for the calculations in this work influenced the results of the analysis. The exhaust gas temperatures after the gas engine bottoming cycle HRVGs were low, in some cases almost as low as 50 °C. This resulted in a large

33 Advanced Power Cycles with Mixtures as the Working Fluid

heat exchanger area in the low-temperature part of the HRVG. Furthermore, there is a risk of condensation and corrosion in the HRVG and the stack; therefore, the low-temperature heat exchangers should be constructed of stainless steel, which is more expensive than carbon steel. The assumption that the gas engine exhaust gas temperature was unrestricted may give the gas engine bottoming cycles an unfair advantage in comparison with the gas-diesel engine bottoming cycles. Pressure drops were included for some streams as percentages of the heat exchanger inlet pressure; however, when designing an ammonia-water cycle in detail, more exact pressure drop calculations are required. Standard values were assigned for the component efficiencies and not values of real components. Since the same values were used for the ammonia-water and steam cycles, it is assumed that the influence of these assumptions on the differences in power output between the ammonia- water and steam cycles is insignificant. Moreover, the results of the ammonia-water cycle analysis are dependent on the property model used for the ammonia-water mixture. Accurate working fluid thermodynamic properties are important for identifying the cycle configuration with the highest power output. In addition, reliable transport properties are required for design of the cycle components. However, the correlations for the thermophysical properties of ammonia-water mixtures at the temperature and pressure levels relevant for power cycles are not experimentally verified for all conditions in an ammonia-water cycle. The steam cycle used as a comparison to the ammonia-water cycle should be of approximately the same complexity as the ammonia-water cycle. If the number of heat exchangers is used as a measure of the cycle complexity, the ammonia-water cycles in the present work are more complex than the steam cycles. If the number of HRVG/HRSG pressure levels is the basis of comparison, the ammonia-water cycles should be compared with single-pressure steam cycles. If instead the number of degrees of freedom available for the optimization of the cycle is used for comparison of the cycles, the ammonia-water cycles should be compared with dual-pressure steam cycles. The possibility to change the ammonia concentration of the working and basic mixtures in the ammonia-water cycle represents two degrees of freedom, corresponding to the two degrees of freedom represented by the two HRSG pressure levels of a dual-pressure steam cycle. In the studies presented in the Papers I-IV, different numbers of engines were used as the basis for the calculations. Reciprocating engine power plants are of a modular construction and more engines can be added for increased plant power output. The electrical efficiencies of the engines are the same disregarding the number of engines. The bottoming cycle efficiency, on the other hand, is dependent on the power plant size and very small bottoming cycles are expensive in relation to their power outputs.

3.4.2 Economical and Technical Aspects The investment cost for the ammonia-water cycle is probably higher than for a steam cycle due to the larger heat exchanger area of the ammonia-water cycle. However, the higher power output may result in a lower specific investment cost

34 3 Ammonia-Water Processes as Bottoming Cycles for Reciprocating Engines

for the ammonia-water cycle. Calculations in Paper I showed that the additional power generated by an ammonia-water cycle compared with a steam cycle could compensate for the higher heat exchanger cost. In Paper III, it was shown that one heat exchanger could be removed from configuration IIb for the gas-diesel engine 18V32GD with only a slight decrease in the cycle power output. Most of the complex ammonia-water cycle configurations designed in this work could probably be simplified by removing some small heat exchangers, causing only minor decreases in cycle power output. The ammonia-water cycle has an economic advantage regarding the vapor turbine. The ammonia-water cycle turbine is smaller and may therefore be less expensive than the steam cycle turbine, since the turbine back pressure is higher in the ammonia-water than in the steam cycle. In addition, the ammonia-water cycle condenser pressure is above atmospheric pressure, while in the steam cycle condenser, the pressure is below atmospheric and a vacuum system is required. When the charge air, jacket water and lubricating oil cannot be cooled sufficiently in the bottoming cycle, additional heat exchangers are required to cool these streams. The cost for these heat exchangers should be included in the steam cycle investment cost. The pressure levels in the steam cycle HRSGs in this work are relatively low, especially in the gas-diesel engine bottoming cycles, which then require steam pipes with large diameters. The ammonia-water cycle plants installed and operated have shown that the cycle works according to the theory. Corrosion, nitriding, ammonia decomposition and ammonia safety have not caused problems in Exergy’s demonstration plant or in other plants, and conventional materials and components can be used, although with some precautions.

3.4.3 Results The studies performed show that, as a reciprocating engine bottoming process, the ammonia-water cycle has a better thermodynamic performance than the steam cycle. All the ammonia-water cycle configurations designed generated more power than the corresponding steam cycles, except for the simplest ammonia-water cycle configuration for the gas-diesel engines compared with a dual-pressure steam cycle. The most efficient ammonia-water bottoming cycle, configuration IIb, generated 44-54 % (Paper I) or 40-50 % (Paper VI) more power than a single-pressure steam cycle for the gas engines and 43-47 % more power than a single-pressure steam cycle and 20-24 % more power than a dual-pressure steam cycle for the gas-diesel engines. In the four studies, the total power outputs of the engines and ammonia- water bottoming cycles varied between 18 MWe and 55 MWe and the total electrical efficiencies varied between 45 % and 51 %. The properties of the ammonia-water mixture working fluid and the ability of the distillation- condensation subsystem to change the working fluid concentration result in the higher power outputs of the ammonia-water cycles compared with the steam cycles. Due to the working fluid characteristics, the ammonia-water cycle can utilize low-temperature heat sources that the steam cycle cannot use, since the ammonia-water mixture starts boiling at a lower temperature than pure water.

35 Advanced Power Cycles with Mixtures as the Working Fluid

The layout and resulting power output of an ammonia-water bottoming cycle is dependent on the characteristics of the waste heat streams produced by the topping cycle. This is illustrated by the differences between the ammonia-water bottoming cycles designed for the gas and gas-diesel engines. The gas engine ammonia-water cycles generated more power than the gas-diesel engine ammonia- water cycles, mostly due to the higher exergy content of the gas engine exhaust gas. When an ammonia-water bottoming cycle was added to a gas engine, the relative power output increase was higher than when added to a gas-diesel engine. Hence, an ammonia-water bottoming cycle is more advantageous for a gas engine than for a gas-diesel engine. The highest total electrical efficiency of the combined engine and bottoming cycles was found for the gas-diesel engine 18V46GD and the ammonia-water configuration IIb; however, this is explained by the high efficiency of the 18V46GD engine.

3.4.4 Suggestions for Future Work To evaluate the real advantage of the ammonia-water cycle compared with the steam cycle, detailed designs and economic studies are required. A thermodynamic analysis merely shows the possible improvements in electrical efficiency and power output of the ammonia-water cycle over the steam cycle. Reliable thermophysical properties, validated by experimental data, for the ammonia-water mixture are needed for an accurate design. The ammonia-water plants constructed thus far have shown that the ammonia-water cycle works according to theory; however, more ammonia-water cycle plants are required in order to completely demonstrate the technology for different applications. The specific investment cost for power plants using new technology is higher than for conventional power plants until a number of power plants have been constructed and the technology is proved.

36

4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

In the second part of this thesis (Chapter 4), evaporative gas turbine cycles are investigated. In Section 4.1, gas turbines for power generation are described. The principles of gas turbine cycles with air-water mixtures as working fluids, especially evaporative gas turbines, are discussed in Section 4.2 along with some previous work concerning the subject. This section contains material from Paper VIII, which reviews different humidified gas turbine cycles. In Section 4.3, three studies about evaporative gas turbines, presented in Papers V-VII, are addressed. In these studies, combined cycles consisting of gas turbines with ammonia-water bottoming processes were compared with evaporative gas turbines and part-flow evaporative cycles were evaluated by exergy and economic analyses. The methods and assumptions, economical and technical aspects of the evaporative cycle, the results and the need for further work are discussed in Section 4.4.

4.1 Gas Turbines for Power Generation

Gas turbines are utilized for base-, intermediate- and peak-load power generation, cogeneration of heat and power, propulsion, and mechanical drive. In recent years, the number of gas turbines for power generation ordered has increased, particularly medium-size (30-60 MWe) aeroderivative gas turbines and large (over 120 MWe) heavy-duty industrial gas turbines (McNeely, 2001). Micro gas turbines (i.e., gas turbines with power outputs about or below 100 kWe) were introduced commercially in the end of the 1990s for distributed power generation and cogeneration on deregulated markets, although the total number of micro gas turbines is still low (Capstone, 2003). For the later part of the 1990s, it was estimated that more than half of the orders for new fossil fuel-fired power plants were based on gas turbines (Luckey, 2000). Moreover, according to a recent study by the International Energy Agency (IEA, 2002), natural gas-fired combined cycles will account for the most of the new power generating capacity. In an open cycle gas turbine, compressed ambient air is used to combust the gaseous or liquid fuel. Expansion of the hot high-pressure combustion gases through a turbine drives the compressor and a generator for electrical power generation. There are gas turbines with a range of sizes: from micro gas turbines to large industrial gas turbines. As an example, the Gas Turbine World Handbook

37 Advanced Power Cycles with Mixtures as the Working Fluid

(GTW, 2001) presents gas turbines with power outputs from 0.2 MWe to 334 MWe and electrical efficiencies from 16 % to 42 %. Most stationary gas turbines are fueled with natural gas, while aircraft gas turbines are fueled with kerosene, a light petroleum distillate (Eastop and McConkey, 1993). Gas turbines which utilize solid fuels as their primary energy source have been studied. Solid fuels can be used in gas turbines either by burning the fuel and using the cleaned combustion gases directly in the gas turbine or by gasifying the fuel and firing the cleaned gases in the gas turbine combustor. Some coal-fueled gas turbine power plants have been constructed with pressurized fluidized bed combustion or gasification processes (Rukes and Taud, 2002). Other alternatives for solid or low-quality fuels are externally-fired or closed cycle gas turbines, where the fuel energy is transferred indirectly to the gas turbine through a heat exchanger that replaces the combustion chamber of an open cycle gas turbine. A few closed cycle gas turbines have been in operation (Bathie, 1996). The predominant pollutants generated by a gas turbine are nitrogen oxides (NOx). To reduce the amount of nitrogen oxides produced, water or steam can be injected in the combustion chamber to decrease the firing temperature (Bathie, 1996; Schorr, 1992). An advantage of this method is that the gas turbine power output is increased due to the increased working fluid mass flow rate. In addition, lean low-NOx burners can reduce the emissions of nitrogen oxides (Bathie, 1996). Both methods can achieve low levels of nitrogen oxides, 25 ppmvd (parts per million by volume, dry, 15 % oxygen) is a common level. Natural gas-fired turbines produce very low levels of sulfur oxides, since this fuel contains very little sulfur. Modern gas turbine combustors have combustion efficiencies of almost 100 %, which result in only small amounts of carbon monoxide and unburned hydrocarbons (Wilson, 1984). Improvement of the gas turbine efficiency and power output can be accomplished in many ways. One mean is increased turbine inlet temperature and pressure ratio; however, this requires advanced materials and larger cooling flows for the gas turbine hot parts. Today, gas turbines with firing temperatures of 1430 °C (i.e., the General Electric 7001H and 9001H) and pressure ratios of 35 (i.e., the Rolls-Royce Trent) are available (GTW, 2001). Gas turbines for stationary power generation have benefited from research and development work on aircraft gas turbines, which have also been converted to power generation applications. Another mean for increased power output is compressor intercooling, which also raises the efficiency for cooled aeroderivative gas turbines, although not for industrial gas turbines (Macchi et al., 1995). Intercooling reduces the compression work and the cooling air temperature, thus enabling higher firing temperatures or lower cooling flow rates. Saidi et al. (2000) reviewed different types of intercoolers. Moreover, the gas turbine efficiency can be augmented by recovering the exhaust gas energy, for example by a recuperator that preheats the compressed air prior to the combustor, thus decreasing the fuel flow rate. Gas turbine recuperators are relatively scarce and have mostly been used for propulsive gas turbines, for which a low specific fuel consumption (kg/kWh) is important (McDonald and Wilson,

38 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

1996). Furthermore, recuperators are required for competitive electrical efficiencies for micro gas turbines and primary surface recuperators have been developed for this application (Lagerström and Xie, 2002; Utriainen, 2001). Intercooling combined with recuperation increases the efficiency and power output more than either method applied alone. One example of an intercooled recuperated gas turbine is the WR-21, which will be used in the UK Royal Navy’s next generation warships. Final testing of the WR-21 was scheduled for 2002 (McCarthy and Scott, 2002; Ashley, 1998). Another alternative for increased efficiency is to generate steam from the exhaust gas energy for reforming the natural gas fuel in a so-called chemically recuperated gas turbine (Carcasci et al., 1998). Alternatively, the steam can be used in a combined cycle. Combined cycles have become common in power plants in the last decades and systems with power outputs from below 10 MWe up to almost 1,000 MWe and electrical efficiencies from 40 % to 60 % are available (GTW, 2001). The smaller combined cycles are most often used for cogeneration. Reheat or sequential combustion, where additional fuel is burned in a second combustor, increases the gas turbine power output and the efficiency of a combined cycle (e.g., the ALSTOM GT 24 and GT 26 gas turbines). In the most efficient commercially available combined cycle, based on a General Electric H technology gas turbine, advanced gas turbine materials and a triple-pressure steam cycle with reheat of the steam by gas turbine cooling result in a net efficiency of 60 % and a power output of 480 MWe (50 Hz) or 400 MWe (60 Hz) (Valenti, 2002). The first system of this type (50 Hz) was scheduled for startup in 2002. Integrated systems with gas turbines and fuel cells have the potential of high efficiencies, about 70 % in the near term and up to 80 % in the long term according to Layne et al. (2000). Another possibility for increased power output and efficiency is water or steam injection in the gas turbine; such cycles are further considered in Section 4.2.

4.2 Theory and Previous Work on Evaporative Gas Turbines

4.2.1 Gas Turbines with Air-Water Mixtures as the Working Fluid Different gas turbine cycles where the working fluid is humidified to increase the specific power output and the electrical efficiency and decrease the specific investment cost have been suggested. Possible additional advantages of humidification are reduced formation of nitrogen oxides, decreased power degradation caused by high ambient temperatures or low ambient pressures (i.e., at high elevations) and improved part-load efficiency. Water or steam injection increases the mass flow rate of working fluid that expands through the turbine. This raises the specific power output, since the compressor work remains constant and the work required to compress the water used for injection or steam generation is small. In addition, if the exhaust gas energy is recovered by

39 Advanced Power Cycles with Mixtures as the Working Fluid

preheating water for injection, generating steam for injection or heating the combustion air, the cycle efficiency is raised.

4.2.1.1 Water-Injected Gas Turbines There are many systems in operation where water is injected at the compressor inlet when the ambient temperature is high, thus decreasing the power output degradation experienced by gas turbines on hot days. The water cools the air and increases its density and the compressor mass flow rate. Systems that saturate the air can increase the power output by 5 % to 10 % and the efficiency by 1.5 % to 2.5 % and decrease the NOx emissions by 10 % for a conventional diffusion combustor (Johnke and Mast, 2002). Overspray systems, where more water than is required for saturation is injected and the remaining water droplets evaporate in the compressor, can increase the power output by 10-20 % and the efficiency by 1.5-3 % and decrease the NOx emissions by 20-40 % (Johnke and Mast, 2002). Water injection between compressor stages can be used for intercooling. One example is the General Electric Sprint system for the LM6000 gas turbine, where water is injected at the low-pressure and high-pressure compressor inlets, thus increasing the power output by more than 8 % at ISO conditions and by more than 32 % at 32 °C ambient temperature (de Biasi, 2000a). Water injection inside the compressor has been proposed, for example in the TopHat cycle (Anonymous, 2001), where the resulting low compressor outlet temperature enables efficient recuperation. Gas turbines with water injection after the compressor combined with recuperation have been suggested. A schematic of such a cycle is shown in Figure 4.1. The injected water increases the mass flow rate of working fluid and

~ ~

air air recuperator superheater

steam

economizer evaporator water

economizer water

Figure 4.1. Schematic recuperated water-injected gas turbine (left) and steam-injected gas turbine (right).

40 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

decreases the temperature of the compressed air, thus augmenting the heat recovery in the recuperator. The water is assumed to be injected into a chamber, to slow down the air and allow the water sufficient time to evaporate. Water-injected cycles for power generation have been studied by Gasparovic and Stapersma (1973), Mori et al. (1983), El-Masri (1988a), Frutschi and Plancherel (1988), Annerwall and Svedberg (1991), Bram and de Ruyck (1997), Horlock (1998) and Aronis and Leithner (2002). Szargut (2002) and de Ruyck et al. (1996) investigated water-injected cycles for cogeneration of power and district heating. Rolls-Royce has presented a cycle based on the Trent gas turbine with water injection for intercooling and aftercooling. The expected advantages of this cycle are an increased electrical efficiency, over 50 % compared with 42 % for the simple cycle, and a lower specific investment cost than for a combined cycle (Stambler, 2002; Wei, 2002).

4.2.1.2 Steam-Injected Gas Turbines In the steam-injected gas turbine, schematically shown in Figure 4.1, steam generated from exhaust gas heat is injected before, in or after the combustor or between turbine stages. Steam-injected cycles became more common in the middle of the 1980s with the introduction of the Cheng and STIG cycles and today there are several hundred gas turbines with steam injection in operation. The development of steam-injected gas turbines was reviewed by Tuzson (1992). There is at least four commercialized steam-injected gas turbine cycles: the Cheng (Cheng and Nelson, 2002), the General Electric STIG (Badeer, 2000), the Ishikawajima- Harima IHI-FLECS (Uji, 1999) and the Mashproekt Aquarius cycles (Lupandin et al., 2001; de Biasi, 2001a). Steam-injected gas turbines with power outputs of 2-51 MWe and efficiencies of 32-44 % are available (GTW, 2001). In these systems, all or a part of the steam generated from the exhaust gas energy is injected in the gas turbine, depending on the compressor surge margin. Retrofit with steam injection can be a low-cost alternative for power augmentation for gas turbines used for peak-load or cogeneration (Cheng and Nelson, 2002). Steam-injected gas turbines are mostly used for small-to-medium-size cogeneration of power and steam in industry (e.g., breweries, food, chemical and pulp and paper industries) or community facilities (e.g., hospitals, universities and district heating), especially in applications where the steam load or the electricity price vary. Such applications will most likely increase in number when more power markets are deregulated. Different steam-injected gas turbine configurations for power generation have been investigated, for example cycles with intercooling (Larson and Williams, 1987), sequential combustion (Fischer et al., 2001), intercooling, recuperation and sequential combustion in different combinations (Annerwall and Svedberg, 1991) and also combined with steam cooling of the gas turbine (De Paepe and Dick, 2000), cycles with a topping steam turbine in which the steam is expanded prior to injection in the gas turbine (Frutschi and Plancherel, 1988; Foster-Pegg, 1989), intercooling and a topping steam turbine (Macchi et al., 1995), and a topping steam turbine combined with sequential combustion (Hofstädter et al., 1998; Rice, 1993a,

41 Advanced Power Cycles with Mixtures as the Working Fluid

b, c). El-Masri (1988b) presented a gas turbine cycle with two recuperators, aftercooling, and possibly intercooling, by water injection and steam injection that was more efficient than steam-injected gas turbines. A modified version of this cycle was investigated by Bolland and Stadaas (1993). Steam-injected gas turbines for cogeneration of power and district heating have been investigated by Frutschi and Wettstein (1991), Krause et al. (1998) and Bartlett et al. (2002).

4.2.1.3 Evaporative Gas Turbines In the evaporative gas turbine, schematically shown in Figure 4.2, the compressed air is aftercooled before entering a humidification tower in which it is humidified by hot water. The humidification increases the flow rate through the expander and, since the compressor work is constant, the specific power output is augmented. In addition, the cycle efficiency is improved, since the energy in the exhaust gas is recovered by humid air in the recuperator and by water in the economizer and the energy from aftercooling is recovered into the cycle. Alternative cycle configurations compared with the one shown in Figure 4.2 are possible, for example with intercooling, without aftercooling and with external cooling water for increased intercooling and aftercooling. In the humidification tower, air and water are brought into direct counter- current contact. In the simultaneous mass and heat transfer process, some of the water evaporates and saturated heated air leaves at the top of the tower. The water evaporates at a temperature determined by the partial pressure of steam in the air and not at the boiling temperature corresponding to the total pressure in the cycle. Thus, the evaporation occurs at a lower temperature than is possible in a conventional boiler, where water evaporates isothermally at the boiling point

~

air recuperator

aftercooler economizer

hum. tower

water Figure 4.2. Schematic evaporative gas turbine.

42 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

corresponding to the total system pressure, and the evaporative cycle can use low- temperature energy for the evaporation of water. In contrast, the water-injected and steam-injected cycles can only use low-temperature energy for preheating water. Moreover, the pinch point limits the heat recovery in the HRSG of the steam-injected cycle. It has been shown that the exergy destruction of the air-water mixing process in the humidification tower is lower than for mixing by direct injection of water or steam; hence, the electrical efficiency of the evaporative cycle is higher than for the water-injected or steam-injected cycles (Macchi et al. 1995; Chiesa et al., 1995). Since the water temperature decreases downwards in the humidification tower, the air meets water of a higher temperature as it ascends in the tower. The humidity ratio of the air increases upwards in the tower, which increases the evaporation temperature and the evaporation process is non- isothermal. Hence, the temperature profiles for air and water are closely matched and the exergy destruction for the humidification process is low. In contrast, the isothermal boiling process in a steam-injected cycle has a larger exergy destruction due to the poor matching of the exhaust gas and water temperature profiles. Independent power producers and distributed generation have become more common due to privatization and deregulation of the power markets. These trends require flexible power plants with high efficiencies for low-to-medium power outputs. Consequently, the number of gas turbines ordered has increased (McNeely, 2001) and the market for small-to-mid-size combined cycles has enlarged (Bärring et al., 2000). The evaporative cycle should be competitive with the combined cycle for these applications. The electrical efficiency of the evaporative cycle is at the same level as for the combined cycle, while the specific investment cost should be lower, since the steam turbine of the steam bottoming cycle is excluded and the specific power output is higher. For smaller power outputs, the steam bottoming cycle accounts for a larger part of the total investment cost for a combined cycle; hence, the evaporative gas turbine should be especially competitive with the combined cycle for small-to-medium power outputs.

4.2.2 Previous Work on Evaporative Gas Turbines Evaporative gas turbines have been investigated for a long time. Some examples are the different cycles patented by Martinka (1940) and Anxionnaz (1952) and the cycle presented by Mori et al. (1983). In the 1980s, the interest in evaporative gas turbines increased and a research program on the humid air turbine (HAT), the evaporative cycle patented by Fluor Daniel (Rao, 1989), was commenced in the USA. In the beginning of the 1990s, another research program on evaporative gas turbines was initiated in Sweden. In this thesis, the general term humidified gas turbine is used for gas turbines with air-water mixtures as the working fluid. The term evaporative gas turbine is used for gas turbine cycles including a humidification tower, while the term HAT cycle is used for cycles of the type patented by Rao (1989).

43 Advanced Power Cycles with Mixtures as the Working Fluid

4.2.2.1 The Humid Air Turbine The US research program on the intercooled, aftercooled and recuperated humid air turbine involved the Electric Power Research Institute (EPRI), consultant companies, gas turbine manufacturers and utility companies. The program investigated different HAT cycle configurations integrated with coal gasification (IGHAT) or fueled with natural gas and an aeroderivative gas turbine for humid air operation. EPRI (1991) presented a report on HAT cycles fueled with gasified coal or natural gas. The IGHAT was predicted to have higher efficiency, lower investment cost and a 15 % lower cost of electricity compared with a combined cycle integrated with coal gasification (IGCC). The IGHAT investment cost was lower since it could recover low-temperature energy in the form of hot water from an inexpensive quench-cooled gasifier, while the IGCC required expensive syngas coolers for generation of steam for the steam cycle. The specific investment cost for the natural gas-fueled HAT was higher than for a combined cycle; however, the higher efficiency of the HAT, about 4 percentage points, resulted in equivalent costs of electricity. For both fuels, the HAT had a better part-load performance than the combined cycle and was less sensitive to high ambient temperatures. Previous work had shown that the optimal HAT would have a high pressure ratio (35±5) and the turbine outlet mass flow rate would be 20-30 % greater than the compressor inlet flow. Therefore, an aeroderivative gas turbine was investigated for operation in the HAT cycle, since such gas turbines have high pressure ratios and can comply better with flow rate variations than industrial gas turbines (Cohn, 1993). Another driving force for using aeroderivative gas turbines in evaporative cycles is that aeroderivative gas turbines are not as suitable as industrial gas turbines for combined cycle applications, since the high pressure ratio results in a low exhaust gas temperature. Therefore, other cycle layouts than combined cycles are required for high electrical efficiencies for power cycles based on aeroderivative gas turbines. EPRI (1993) presented a report on an IGHAT using the proposed gas turbine FT 4000 HAT, based on the Pratt & Whitney aircraft engine PW 4000. The turbine was designed for fueling with syngas and was also modified for natural gas operation. The design calculations showed that the IGHAT had approximately the same efficiency as an IGCC, although the specific investment cost was 11 % lower and the cost of electricity was 8 % lower. With natural gas, the HAT efficiency was 5 % higher than for a combined cycle. However, the HAT specific investment cost was higher than for the combined cycle, resulting in an only slightly lower cost of electricity for the HAT. It was estimated that the development cost for the FT 4000 HAT would be above 200 million USD (1992 USD). Day and Rao (1993) presented results for a redesigned natural gas-fired FT 4000 HAT with a pressure ratio slightly above 40. The HAT cycle had a net power output of 200 MWe and a net efficiency of 55.4 %, while a combined cycle had an output of 201 MWe and an efficiency of 52.6 % for an ambient temperature of 21.5 °C.

44 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

The CHAT (cascaded humidified advanced turbine) was proposed to solve the problem of flow rate mismatch between the compressor and turbine in the HAT. The CHAT can be based on standard gas turbine components, which reduces the cycle development cost. Compared with a combined cycle, the CHAT has higher part-load efficiency, lower investment cost and retains its power output at high ambient temperatures. This intercooled, humidified, recuperated and reheated cycle has two shafts: one power-generation shaft with a low-pressure compressor and expander and one power-balanced shaft with intermediate-pressure and high- pressure and expanders. Compared with a simple cycle gas turbine, the work input to the low-pressure compressor is reduced, since the high-pressure expander pressure ratio is lower than the total compressor pressure ratio on the power-balanced shaft. Hence, the intermediate-pressure and high-pressure compressors supply most of the required compression work (Nakhamkin et al., 1996). Nakhamkin et al. (1998) presented a CHAT with a pressure after the high- pressure compressor of 68 bars, a power output of 316 MWe and a net efficiency of 55.5 %. The specific investment cost was 10-15 % lower than for a combined cycle at ISO conditions and 20-25 % lower at 35 °C. A CHAT demonstration plant based on the Rolls-Royce Allison 501-KB7 gas turbine, with a power output of 12.1 MWe and a net efficiency of 46.4 %, has been proposed (de Biasi, 1999). Further studies of the CHAT have shown higher power outputs and electrical efficiencies than combined cycles based on commercially available gas turbines for mid-size power generation (30-150 MWe). With higher turbine inlet temperatures, CHAT efficiencies over 60 % are anticipated (de Biasi, 2001b). In the recuperated and reheated CASH (compressed air storage with humidification) cycle, compressed air produced at off-peak periods for a low cost is stored in, for example, an underground cavern, for power generation during periods with high electricity prices. There exist two non-humidified compressed air (CAES) plants: a 290 MWe plant in Germany and an intercooled, recuperated and reheated 110 MWe plant in the USA. (Cohn et al., 1999) In the CASH cycle, the stored air is extracted and humidified in a tower by hot water heated in the compressor intercooler and aftercooler and by exhaust gas. Power generated by the turbines drives the compressors via an electric motor. When the power demand is high, none of the power generated by the turbines is used for compression and all the generated power can be delivered to the grid, resulting in a plant with a low specific cost. An additional advantage of the CASH cycle is that the flow mismatch of the HAT is avoided when the compressed air is stored. A CASH cycle with integrated coal gasification (IGCASH) can save 20-40 % in specific investment cost for the peak capacity compared with an IGCC cycle (Wolk and Cohn, 1993). The IGCASH can be integrated with a natural gas-fired CASH cycle in a so-called CASHING (compressed air storage with humidification integrated with natural gas) plant, for decreased investment cost and increased flexibility compared with an IGCASH plant (Cohn, 1993).

45 Advanced Power Cycles with Mixtures as the Working Fluid

In addition, externally coal-fired HAT cycles have been investigated (Robson and Seery, 1998) and for power plants with carbon dioxide recovery, a HAT could have a lower specific investment cost than a combined cycle (Rao and Day, 1996).

4.2.2.2 The EvGT Project and Related Studies A research program on evaporative gas turbines (EvGT) was initiated in Sweden in the beginning of the 1990s. The program involves utility companies, gas turbine manufacturers, research organizations and universities. The objectives of this program are to demonstrate the evaporative gas turbine technology in a pilot plant, investigate the humidification process and the water circuit chemistry, and to propose future plant designs. An evaporative gas turbine pilot plant has been constructed and different small-to-mid-scale cycle configurations and applications have been investigated. Ågren (2000), Rosén (2000), Lindquist (2002) and Bartlett (2002) presented some of the work within the project. All the studies cited in this section were not directly financed by the EvGT project; nevertheless, they are connected to this project. In one study, a mid-size (70-80 MWe) intercooled evaporative cycle with part- or full-flow humidification was investigated for generation of power only or power and district heating (Nilsson, 1996). The evaporative cycles had electrical efficiencies on the same level as the combined cycles, whereas they had higher total efficiencies for cogeneration. The specific investment costs were 15-30 % lower and the electricity and heat production costs were lower for the evaporative cycles compared with the combined cycles. In addition, evaporative cycles for cogeneration of power and district heating were investigated by Rydstrand et al. (2002). A combined cycle was compared with a conventional steam-injected cycle, a part-flow evaporative cycle with steam injection and a saturated steam-injected cycle without a steam superheater. Compared with the combined cycle, the humidified cycles all had similar electrical efficiencies (up to 50 %) and higher total efficiencies. Moreover, the humidified cycles had significantly lower specific investment costs and were economically feasible at lower costs of electricity compared with the combined cycle. Another investigated application for the evaporative cycle is industrial cogeneration of power (1-50 MWe) and steam and/or hot water (Melin and Simonsson, 2001). The best economic performance of the evaporative cycle was found for applications with low demands of electricity and demand of low- temperature energy, for example hospitals and airports. In addition, industries like the food and pulp and paper industries are interesting for the evaporative cycle. If electricity and hot water (50-60 °C) are generated, the evaporative cycle could have an electrical efficiency of 45 % and a total efficiency of over 90 %. In addition, Nilsson et al. (2001) discussed evaporative cycles for industrial cogeneration (a few to about 80 MWe). If there is a demand for low-temperature energy, it is favorable with high humidification rates (i.e., full-flow humidification), while for power production only or cogeneration of process steam, lower humidification rates can be used (i.e., part-flow humidification combined with a boiler for steam

46 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

generation). The steam-injected cycle was deemed to be the strongest competitor to the evaporative cycle, since it is commercialized and can be used in the same applications. Biomass, refuse or waste heat can replace part of the natural gas input to the cycle, and evaporative gas turbines for base-load power generation using such energy sources have been studied (Simonsson et al., 2002). The external energy can heat the water outlet stream from the humidification tower, generate steam for injection in the cycle and superheat the humid air after the tower. Furthermore, externally biomass-fired evaporative cycles for cogeneration of power and district heating have been investigated by Yan et al. (1996). Flue gas condensation recovered the water vapor in the exhaust gas and generated district heating. A natural gas topping combustor increased the turbine inlet temperature. Yan et al. (1995) investigated the same system, without the flue gas condenser, for power generation only. Olsson (1999) studied combined and evaporative cycles for cogeneration of power and district heating from gasified biomass. For a combined cycle with pressurized gasification, the electrical efficiency approached 45 % and the evaporative cycle efficiency was close to this value, while the total efficiency was lower for the evaporative cycle unless for low district heating temperatures. There is only one evaporative gas turbine in operation: a pilot plant in Lund, Sweden, that started operation as a simple cycle in 1997. The plant is based on a Volvo VT600 gas turbine with a simple cycle power output of 600 kWe and an efficiency of 22 %. In the final plant configuration, the compressed air is aftercooled, humidified in a packed bed tower by water heated in the aftercooler and an exhaust gas economizer and recuperated. As an evaporative cycle without an aftercooler, the efficiency was 35 %, while addition of an aftercooler increased the efficiency by about one percentage point (Lindquist, 2002). The flow mismatch between the compressor and turbine is met by bleeding off some of the compressed air. (Lindquist et al., 2000a) Lindquist et al. (2000b) described the operation of the plant with startup strategies and water circuit control. Full power output can be reached almost as fast for the evaporative cycle as for the simple cycle. The pilot plant can be self-sufficient with water by cooling the exhaust gas to 35 °C in the flue gas condenser and the water treatment system can achieve a sufficient feedwater quality (Ågren et al., 2002).

4.2.2.3 Part-Flow Humidification In most of the evaporative cycles studied, the whole compressed air flow is sent to the humidification system; however, cycles with part-flow humidification have been suggested. Within the EvGT project, part-flow evaporative cycles have been investigated and Westermark (1998) proposed a cycle where the part-flow was between 10 % and 30 % of the compressor inlet air mass flow rate. Previously, Nakamura et al. (1985, 1987) had suggested evaporative and water-injected cycles with higher percentages of part-flow. In part-flow cycles, only a fraction of the air is passed through the humidification system; the remaining air bypasses the tower and is mixed with the humidified air before the recuperator or the combustor.

47 Advanced Power Cycles with Mixtures as the Working Fluid

Part-flow humidification reduces the heat exchanger area and tower volume, which decreases the investment cost compared with a cycle with full-flow humidification. In addition, it has been shown that it is unnecessary to cool the whole air flow in the aftercooler and afterward heat it up again in order to achieve an efficient heat recovery; hence, compared with full-flow cycles, the electrical efficiency is retained or increased for part-flow cycles. Aagren et al. presented part-flow evaporative cycles based on an industrial gas turbine (1997a) and an aeroderivative gas turbine (1997b) with a packed bed humidification tower divided in two sections. A two-section tower can provide a lower water outlet temperature than a one-section tower, thus increasing the heat recovery. Part-flow humidification reduced the efficiency slightly for the industrial gas turbine cycle with a one-section tower. With a two-section tower, the efficiency was the same for the part- and full-flow cycles. For the aeroderivative gas turbine, part-flow increased the efficiency for the cycle with a one-section tower and a two- section tower increased the efficiency further. Ågren and Westermark investigated evaporative cycles with part-flow humidification for an aeroderivative gas turbine (2001a) and an industrial gas turbine (2001b). In the cycles, high-temperature energy is used to generate steam that is mixed with the working fluid and the humidification tower recovers low-temperature energy below the boiling point corresponding to the total pressure in the system. For the aeroderivative case, the efficiency had a maximum (52.9 %) for a part-flow corresponding to 12 % of the compressor intake mass flow rate. For the industrial gas turbine, full-flow humidification resulted in the highest electrical efficiency (52.6 %); however, the efficiency was relatively constant down to a part-flow of 50 %. Bartlett and Westermark (2003a, b) found that steam-injected part-flow evaporative cycles with high pressure ratios were better suited for non-intercooled gas turbines than full-flow cycles, since the part-flow cycles required less gas-gas heat exchange and had higher electrical efficiencies and specific power outputs. For intercooled gas turbines, the full-flow evaporative cycles had the highest efficiency, although the part-flow cycles had similar efficiencies for high pressure ratios. An economic comparison showed that part-flow evaporative and steam- injected cycles had the largest potential for mid-size power generation. The full- flow cycles required intercooling and a high-performance recuperator with a lower cost than today’s recuperators in order to be competitive. The non-intercooled humidified gas turbine cycles had lower specific investment costs (15-35 %) and lower costs of electricity (3-12 %) than combined cycles and intercooling should reduce the cost of electricity further.

4.2.2.4 The Humidification Tower The humidification tower is a key component in the evaporative cycle that has not been employed previously in power cycles. Different methods for humidification of air were discussed by Dalili (2003). Accurate modeling of the simultaneous mass and heat transfer in the humidification tower is required for reliable cycle design and cost estimation. Enick et al. (1995), Gallo et al. (1995), Ågren (2000) and

48 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

Lindquist et al. (2002) proposed models for packed bed humidification towers, which have been used in most evaporative cycle simulations, and Dalili and Westermark (2002) performed an experimental investigation of the packed bed humidifier in the Swedish pilot plant. An alternative to the packed bed humidifier is the tube humidifier. Dalili and Westermark (2001, 2003) presented experimental results from a tube humidifier pilot plant, which consists of a tube with rods on the outer surface. Inside the tube, ascending air is brought into counter-current contact with a water film falling downward, while the exhaust gas flows downward on the outside of the tube. Thus, the separate economizer required for a packed bed humidifier is avoided with a tube humidifier. A tube humidifier is favorable for small-size gas turbines due to its effective heat and mass transfer characteristics and compactness.

4.2.2.5 Water Recovery and Water and Air Quality Water consumption and quality have been much discussed for humidified gas turbines, since the cost for water increases the operational cost for the plant and water and its contaminants may cause erosion and corrosion in the cycle. There are some examples of water recovery in humidified gas turbines: the steam-injected gas turbine Aquarius has water recovery (de Biasi, 2000b), there is a Cheng cycle with water recovery in an Italian car-manufacturing factory (Macchi and Poggio, 1994) and the Swedish evaporative gas turbine pilot plant is self-supporting with water (Ågren et al., 2002). In all cases, the recovered water is cleaned from substances that may cause corrosion in the water circuit or gas turbine before re-use in the cycle. Different alternatives for water recovery in steam-injected cycles have been investigated by De Paepe and Dick (1999), Nguyen and den Otter (1994) and Blanco and Ambs (2002), and Cataldi (2001) investigated water recovery in evaporative cycles. Gas turbines are sensitive to contaminants that may enter the cycle with the inlet air, the fuel and, in the case of humidified cycles, the water. For direct water- injected and steam-injected gas turbines, demineralized water is required. Evaporative cycles, on the other hand, can use water of a lower quality, since the water evaporates into the air in the humidification tower, leaving non-volatile contaminants in the liquid phase. Bartlett and Westermark (2001a) presented a model for the flow of alkali salts in an evaporative cycle. Alkali metals and alkaline earth metals may cause corrosion in the turbine and in the water circuit. Substances can be transferred between the air and water streams in the humidification tower and in a direct-contact flue gas condenser. If there are contaminants in the water, only volatile compounds and compounds in entrained droplets can be transferred to the air, therefore a droplet separator should be included after the humidification tower. Bartlett and Westermark (2001b) presented results from experimental investigations in the evaporative gas turbine pilot plant regarding the beneficial effect of different inlet air filters and flue gas condensation on the air and water quality. For evaporative cycles with water recovery, efficient inlet filters reduce the need for water treatment, since a filter greatly decreases the amount of salt particles

49 Advanced Power Cycles with Mixtures as the Working Fluid

entering the cycle. If an is not used, the water in the humidification tower captures most of the particles in the air stream.

4.2.2.6 Properties of Air-Water Mixtures For humidified gas turbines, accurate thermophysical properties (i.e., thermodynamic properties and transport properties) for air-water mixtures are required for reliable simulations of the cycle performance, sizing and design of the cycle components, and cost estimation. The air-water system deviates from ideal behavior, especially at conditions of high pressures and low temperatures. Dalili et al. (2001) reviewed the available thermodynamic property models for air-water mixtures. The models found were limited in temperature, pressure and/or humidity range. The best available model was the Hyland and Wexler model (Hyland and Wexler, 1983), despite a limited temperature range. A comparison between ideal and real gas data showed that the real model predicted a higher saturation humidity at a specific temperature, since water is more volatile in the real model. The cycle efficiency seemed to be only slightly affected by different thermodynamic models. Ji and Yan (2002a) proposed a real gas model, based on a modified Redlich- Kwong , for the thermodynamic properties (i.e., humidity, enthalpy and entropy) of air-water mixtures. The new model was compared with other real gas models: up to 50 bars and 400 K, the models agreed with each other, while the deviations were larger at higher pressures. Ji and Yan (2002b) validated the new model by comparing saturated properties calculated with the model to experimental data available in literature for oxygen-water, nitrogen-water and air- water systems and to other thermodynamic property models. Experimental data for the air-water system are scarce. The comparisons showed that the new model is valid for saturated properties up to 300 °C and 200 atm. Yan et al. (2003) calculated some components in an evaporative cycle with ideal, ideal mixture and real models, among them the Ji and Yan model, for the air-water mixture thermodynamic properties. It was found that the humid air superheater outlet temperature varied up to 18 °C for the different models. Comparisons of the superheated properties calculated with the Ji and Yan model, the ideal model and the ideal mixture model showed that the Ji and Yan model can be used up to 1500 °C and 200 bars. One possible problem in humidified gas turbines is combustion, since the water may cause combustion instability and reduced combustion efficiency resulting in increased emissions of carbon monoxide and unburned hydrocarbons. Day et al. (1999) described experiments and calculations performed to determine the effect of combustion air water content on emissions, stability limits, operation and ignition. The moisture in the air reduced the NOx emissions, while the carbon monoxide emissions were not significantly increased. In the evaporative cycle pilot plant, the NOx emissions were below 10 ppm and the levels of unburned hydrocarbons and carbon monoxide were very low (Lindquist et al., 2000a).

50 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

4.3 Studies of Evaporative Gas Turbines with Part- and Full- Flow Humidification

In Section 4.3, the studies on evaporative gas turbine cycles with part- and full- flow humidification presented in Papers V-VII are described.

4.3.1 Method All evaporative cycles in this work were simulated in GateCycle, developed by GE Enter Software (GateCycle). Since the packed bed humidification tower cannot be modeled in GateCycle with sufficient accuracy, the tower was simulated in Excel using an in-house model, which is described in detail by Ågren (2000). The tower and GateCycle models were connected via the program CycleLink. The thermodynamic properties of air and water were calculated by an ideal mixture model in GateCycle and by the Hyland and Wexler (1983) real gas model in the tower. In the ideal mixture model, all components except water are assumed to be ideal gases, the water properties are calculated with a steam table and mixtures of water and other components are assumed to be ideal. The evaporative cycles in this work employ the industrial gas turbines GTX100 and Cyclone, manufactured by ALSTOM, and the aeroderivative Trent, manufactured by Rolls-Royce. The GTX100 and Trent were utilized in Papers V- VII, while the Cyclone was only used in Paper VII. The Trent and GTX100 models were based on models by Ågren and Westermark (2001a, b) that were mostly derived from GateCycle gas turbine library data. The Cyclone model was based on data from GateCycle and ALSTOM (Nilsson, 2002). The data utilized for the gas turbines and the evaporative cycle simulations in Papers V-VII are displayed in Table 4.1. The data for the gas turbines differ slightly between the studies, since some input data were changed. All electrical efficiencies were based on the lower heating value of the fuel. In Papers V and VII, the fuel was natural gas that has a lower heating value of 47.5 MJ/kg and in Paper VI, the fuel was methane that has a lower heating value of 50.0 MJ/kg. In the simulations of the humidified cycles, the expander cooling flows were adjusted until the first rotor inlet temperature attained the value specified in Table 4.1. The air stream that passes through the humidification system experiences a pressure drop; in the part- flow cycles, this pressure drop was counteracted by a booster fan in Papers V and VI and by an ejector in Paper VII. The water at the humidification tower inlet was subcooled by 10 °C to prevent boiling in the tower. In all humidified cycles investigated, the water in the exhaust gas was recovered by flue gas condensation. In Papers V and VI, the compressor inlet air mass flow rate in the evaporative cycles was kept the same as in the simple cycle gas turbines. Hence, when the working fluid was humidified, the expander size and the power output were increased. In Paper VII, the intake air mass flow rate to the humidified cycles was varied until the expander size was the same as for the simple cycle. The definition

51 Advanced Power Cycles with Mixtures as the Working Fluid

Table 4.1. Input data for the simple cycle gas turbines at ISO conditions and the evaporative cycles in Papers V-VII. Values without footnotes are the same in all three papers and the Cyclone is only used in Paper VII. Trent GTX100 Cyclone P [MW] 49.81, 3, 49.62 41.61, 41.42, 41.53 12.9 1, 3 2 1 2, 3 ηel [%] 40.7 , 40.6 35.9 , 35.8 35.0 pressure ratio 35.0 20.1 16.7 tcombustor outlet [°C] 1336 1358 1278.5 ∆pcombustor [%] 1 3 3 tfirst rotor inlet [°C] 1253 1275 1217 1, 3 2 1, 3 2 meg [kg/s] 158.3 , 158.2 121.3 , 121.2 38.6 1, 2 3 1, 2 3 teg [°C] 433.0 , 433.1 554.5 , 555.0 555.0 peg [bar] 1.038 1.038 1.038 3 3 3 Vexpander outlet [m /s] 314.3 283.2 90.1 ηpolytropic, compressor 0.895 0.901 0.907 1 2, 3 ηpolytropic, expander 0.876 0.853 , 0.852 0.859 ηgenerator 0.983 0.958 0.963 ηcombustor 0.999 0.999 0.99 Wloss, expander [%] 0.5 0.5 0.5 Cycle simulations

ηis, pump 0.85 ∆pcompressor inlet [bar] 0.00625 1, 2 1, 2 (GTX100), 3 1, 2 (Trent) ηis, booster fan 0.70 ∆pgas, hum. system [%] 5 , 3 ηshaft & bearings (booster fan, pumps) 0.90 ∆pwater, hum. system [%] 2 ηel. drive (booster fan, pumps) 0.95 ∆peg [bar] 0.02475 3 ∆tgas – gas [°C] 30 toutlet water FWH [°C] 130 ∆tgas – liquid [°C] 15 tsubcooling tower inlet [°C] 10 3 ∆tmin, humidification tower [°C] 4 HTU [m] 0.24 3 ∆tpinch, boiler [°C] 10 ηthermodynamic, ejector 0.20 1Paper V 2Paper VI 3Paper VII

of constant expander size was that the expander outlet volumetric flow rate should be the same as in the simple cycle.

4.3.2 A Comparison between Combined Cycles with Kalina Bottoming Processes and Evaporative Gas Turbines The Kalina cycle and the evaporative gas turbine cycle are advanced and efficient power cycles with mixtures as the working fluid: ammonia and water in the Kalina cycle and air and water in the evaporative cycle. A mixture enables non-isothermal boiling, which reduces the exergy destruction of the evaporation process. In

52 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

Paper V, combined cycles with ammonia-water bottoming processes were compared with evaporative gas turbines to evaluate the respective merits of the cycles.

4.3.2.1 Method and Cycle Configurations The input data for the gas turbines are shown in Table 4.1 and Table 4.2 shows the data for the ammonia-water bottoming cycles. The author simulated the ammonia- water cycles, while the results for the evaporative cycles were obtained from Ågren and Westermark (2001a, b).

Table 4.2. Input data for the ammonia-water bottoming cycles in Paper V. pmax [bar] 115 tcooling water [°C] 15 (in), 25 (out) ∆pHRVG [bar] 5 aturbine, min [kg/kg] 0.90 ∆ptbn to cnd lp, cI and cIV [bar] 0.05 ηis, turbine 0.85 ∆ptbn to cnd lp, cII [bar] 0.10 ηgen 0.9604 ∆tgas – gas [°C] 30 ηmech, turbine 0.98 ∆tgas – liquid [°C] 15 ηis, pump 0.80 ∆tliquid – liquid [°C] 5 ηmech, pump 0.875

∼

tbn ∼ tbn ∼ tbn ∼

sh1 sh2

eva1 eva2 sep rht eco

cnd cnd hp lp

Figure 4.3. Ammonia-water cycle configuration IV. A legend can be found in Figure 3.2.

53 Advanced Power Cycles with Mixtures as the Working Fluid

Three ammonia-water cycle configurations were utilized as bottoming processes for the GTX100 and Trent gas turbines. Configuration I (Kalina, 1983), shown in Figure 3.2, is the simplest possible ammonia-water cycle. Configuration II (El-Sayed and Tribus, 1985a) is the same as configuration II in Chapter 3. Configuration IV, shown in Figure 4.3, includes three turbines with reheat (sh2) between the first and second turbine stages and recooling with a recuperative evaporator (eva2) between the second and third turbine stages (Kalina and Leibowitz, 1987). The recuperative evaporator improves the matching of the temperature profiles in the HRVG since there is a surplus of high-temperature energy in the superheater section and a deficiency of energy in the evaporator section. This configuration is called III in Paper V, while it is called configuration IV in this thesis to avoid confusion with configuration III in Chapter 3. All ammonia-water cycle configurations are shown in Paper V. The ammonia-water cycles, simulated in IPSEpro, were optimized for the highest possible electrical efficiency by varying the ammonia concentration levels, as described in Section 3.3.1.2, and, to some extent, the pressure levels in configuration IV. The

Cooling G

INTERCOOLER FUEL RECUPERATOR Feed 1−β Flow splitter β HUMID AIR SUPER- HEATER (H-SH)

BOILER

ECONOMIZER (ECO)

AFTERCOOLER (AC)

HUMIDIFI- CATION TOWER

BOOSTER FAN FLUE GAS CONDENSER AIR/FLUE GAS (FGC) STEAM To feed WATER

Figure 4.4. Evaporative cycle based on the GTX100.

54 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

cycle layouts were not changed, thus these cycle configurations could be called Kalina cycles instead of ammonia-water cycles. In the evaporative cycle based on the GTX100, the gas turbine is intercooled, as shown in Figure 4.4. After the compressor, the air stream is split and one part is sent to the humidification tower. The humid air from the tower is superheated by the exhaust gas and mixed with the air that bypassed the tower and the steam generated from the exhaust gas. The mixture is heated in a recuperator before the combustor. Water entering the humidification tower is heated by the compressed air in the intercooler and aftercooler and by exhaust gas in the economizer. In the evaporative cycle based on the Trent, shown in Figure 4.5, the gas turbine is not intercooled. After the air stream split after the compressor, the hot air is used for steam generation and superheating before the aftercooler. The humid air from the tower is mixed with steam generated from the compressed air and the exhaust gas. The exhaust gas this mixture before it is mixed with the dry air that has bypassed the tower. The water used in the tower is heated by the compressed air in the aftercooler and by exhaust gas in the economizer. Ågren and Westermark (2001a, b) varied the percentage of part-flow and the cycles with the highest

Cooling G

β 1−β WATER AIR/FLUE GAS β = fractional FUEL STEAM flow split to humidification

MIXER 1 HUMID AIR SUPER- HEATER (H-SH)

STEAM SUPER- HEATER (S-SH) MIXER 2 BOILER 1

BOILER 2

ECONOMIZER (ECO)

FEEDWATER HEATER (FWH) AFTERCOOLER (AC) HUMIDIFI- CATION TOWER

BOOSTER FAN PUMP1 FLUE GAS CONDENSER (FGC) PUMP2

Figure 4.5. Evaporative cycle based on the Trent.

55 Advanced Power Cycles with Mixtures as the Working Fluid

electrical efficiencies have been utilized for the comparison in this study. For the GTX100 cycles, the case where all of the air is passed through the humidification system had the highest electrical efficiency and for the Trent cycles, the case where 12 % of the compressor inlet mass flow rate of air is passed through the tower had the highest electrical efficiency.

4.3.2.2 Results The results of the study are summarized in Table 4.3. The power outputs of the evaporative cycles are higher than for the combined cycles, since the compressor flow rate is the same in all cycles, while the expander flow rate is higher in the evaporative cycle due to the humidification, as explained in Section 4.3.1. In Figure 4.6, the electrical efficiencies of the combined cycles with ammonia- water bottoming processes have been compared with the electrical efficiencies of the evaporative cycles. For the GTX100, ammonia-water cycle configurations I and II had lower electrical efficiencies than the evaporative cycle, while configuration IV had a higher efficiency. For the Trent, all ammonia-water cycle configurations had higher efficiencies than the evaporative cycle. The main reason for the different behavior of the GTX100 and Trent cycles is the lower exhaust gas temperature of the Trent, resulting from the Trent’s high pressure ratio. The results of the study imply that the ammonia-water cycle can use energy at this temperature level slightly more efficiently than the investigated evaporative cycle configurations. The combined cycle with a configuration IV bottoming process gave 1.3-1.4 percentage points higher efficiency than both evaporative cycles, which shows that this ammonia-water cycle configuration is efficient for both aeroderivative and industrial gas turbines due to the flexibility of the ammonia- water mixture working fluid. The efficiency of configuration IV could be increased by replacing the simple DCSS by a more complex DCSS with a higher degree of internal heat recovery. Since the GTX100 was intercooled and recuperated in the evaporative cycle, the ammonia-water cycle configuration I was simulated as a bottoming cycle for a GTX100 with intercooling and with intercooling and different percentages of recuperation, that is 50 % or 100 % of the gas turbine exhaust gas was used to heat the compressed air before the combustor while the rest of the exhaust gas was sent directly to the bottoming cycle HRVG. The energy from the intercooler was recovered by the bottoming cycle. With only intercooling, the total power output was increased while the electrical efficiency was decreased. Intercooling and 50 % recuperation resulted in a higher efficiency than the case with only intercooling, although the efficiency was lower than for the base case (i.e., a simple cycle gas turbine as topping cycle). With intercooling and 100 % recuperation, the efficiency was almost the same as for the base case, while the fraction of the total power output supplied by the bottoming cycle was reduced. Both the ammonia-water and the evaporative cycle can use low-temperature energy for evaporation. In this study, the ammonia-water cycle could recover more energy than the evaporative cycle. This is shown by the exhaust gas temperature

56 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

Table 4.3. Results from the comparison of combined cycles with ammonia-water bottoming processes to evaporative cycles. GTX100 Trent Ptot [MW] ηel, tot [%] Ptot [MW] ηel, tot [%] Simple cycle gas turbine 41.6 35.9 49.8 40.7 Ammonia-water cI 60.2 (18.7)1 52.0 65.1 (15.3)1 53.2 Ammonia-water cII 60.5 (19.0)1 52.2 65.5 (15.7)1 53.5 Ammonia-water cIV 62.4 (20.9)1 53.9 66.5 (16.7)1 54.3 Evaporative cycle 78.32 52.62 97.93 52.93 1The power outputs of the ammonia-water bottoming cycles are shown within parentheses. 2For the GTX100 evaporative cycle, 100 % of the compressor inlet air mass flow rate is humidified (Ågren and Westermark, 2001b). 3For the Trent evaporative cycle, 12 % of the compressor inlet air mass flow rate is humidified (Ågren and Westermark, 2001a).

1.5 GTX100 Trent

1.0

0.5

0.0 [percentage points]

-0.5 Difference in electrical efficiency electrical in Difference

-1.0 cI cII cIV Figure 4.6. Comparison of the electrical efficiencies of the ammonia-water combined cycles and the evaporative cycles.

after the HRVG, which was 60-70 °C for the ammonia-water configuration I, 70- 80 °C for configuration II and 53-54 °C for configuration IV. For the evaporative cycles, the exhaust gas temperature before the flue gas condenser was 106 °C for the GTX100 cycle and 94 °C for the Trent cycle. The specific investment cost for the evaporative cycle should be lower than for a combined cycle with an ammonia-water bottoming process, since the evaporative cycle avoids the cost for the bottoming cycle turbine. In addition, the specific

57 Advanced Power Cycles with Mixtures as the Working Fluid

power output is higher for the evaporative cycle than for the ammonia-water combined cycle, which should reduce the specific investment cost further. On the other hand, humidifying the gas turbine working fluid requires design changes to the gas turbine, which are possibly expensive and are not necessary when the gas turbine is employed in a combined cycle.

4.3.3 Exergy Analysis of Part-Flow Evaporative Gas Turbines In the two-part Paper VI, an exergy analysis was performed on the evaporative cycles that were studied in Paper V. In the first part of the paper, the methods are described and in the second part of the paper, the results are presented and discussed. The purpose of the exergy analysis was to evaluate cases with different percentages of part-flow, to explore the advantages of part-flow humidification.

4.3.3.1 Method and Cycle Configurations The data used for the simulations of the evaporative cycles are shown in Table 4.1. The evaporative cycle configurations, shown in Figures 4.4 and 4.5, are explained in Section 4.3.2.1. To perform an exergy analysis of a cycle, the exergies of the streams in the cycle must be calculated; hence, the enthalpies, and compositions of the streams are required. The entropy is not calculated in GateCycle, so this property had to be calculated separately. Thus, the cycles were simulated in GateCycle and data for all streams (i.e., temperature, pressure and composition) were transferred to Excel where the enthalpies and entropies were computed. In GateCycle, the enthalpies of all substances except water were calculated with the ideal JANAF table data curves (Chase, 1998). Therefore, polynomial expressions of the JANAF table data (NIST, 2001) were used to compute the enthalpies and entropies of the ideal gases in Excel. These expressions and the coefficients used in them are shown in Paper VI. The properties of water were calculated by the property formulation IAPWS-IF97 (Wagner and Kruse, 1998) in GateCycle and by the formulation IAPWS 95 (SteamTab; IAPWS, 1996) in Excel. The differences between the two steam property formulations are very small. In both GateCycle and Excel, the ideal gases and water were assumed to form an ideal mixture; the calculation of enthalpy, entropy and exergy of ideal mixtures is described in Paper VI. The reference states for the property formulations used differ and the conversion between different reference states is explained in Paper VI. Since both enthalpy calculation methods were based on the same data sources, the differences between the enthalpy values computed in GateCycle and Excel were small.

4.3.3.2 Results The power outputs, fuel exergy inputs and cycle exergetic efficiencies are presented in Table 4.4. The parameter β is the air mass flow rate fraction that is sent to the humidification system after the removal of the expander cooling air. The highest exergetic efficiencies were found for the GTX100 cycle with full-flow humidification, although the exergetic efficiency was relatively constant from full-

58 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

Table 4.4. Results of the exergy analysis of part-flow evaporative cycles. GTX100 Trent β [%] P [MW] Efuel [MW] ηII [%] P [MW] Efuel [MW] ηII [%] 100 77.9 155.7 50.0 107.7 221.1 48.7 80 77.6 155.2 50.0 105.1 214.6 49.0 60 77.1 154.4 49.9 102.4 207.7 49.3 40 76.2 153.0 49.8 99.5 200.6 49.6 20 74.1 150.4 49.3 95.6 192.2 49.7 10 71.9 147.7 48.7 92.5 186.6 49.6 5 69.8 145.3 48.0 89.7 182.1 49.3 2.5 68.0 143.4 47.4 87.3 178.5 48.9

Table 4.5. Exergy destruction as a percentage of the fuel exergy for the cases with the highest cycle exergetic efficiencies.

Edestruction/Efuel [%] GTX100 (β=100 %) Trent (β=20 %) Combustor 25.1 (55.3 %) 25.3 (53.7 %) Expander 7.3 (16.0 %) 7.1 (15.1 %) FGC (water recovery) 5.7 (12.5 %) 5.7 (12.0 %) Compressors, fan, pumps 2.0 (4.5 %) 2.4 (5.1 %) Superheaters, recuperators 2.3 (5.1 %) 1.0 (2.1 %) Mixers 0.5 (1.0 %) 3.1 (6.6 %) Water preheaters 1.3 (3.0 %) 0.7 (1.5 %) Humidification tower 1.0 (2.1 %) 0.2 (0.4 %) 0.2 (0.5 %) 1.6 (3.3 %) Miscellaneous (inlet duct, ) 0.04 (0.1 %) 0.04 (0.1 %) Total exergy destruction 45.4 (100 %) 47.2 (100 %)

flow humidification to 60 % part-flow, and the Trent cycle with 20 % part-flow (i.e., 14 % of the compressor intake air mass flow rate is sent to the humidification tower). Figure 4.7 shows the exergetic efficiencies for all part-flow cases. In Table 4.5, the exergy destruction due to internal irreversibilities in the components is given as a percentage of the fuel exergy for the cases with the highest cycle exergetic efficiencies. For the water recovery system, the exergy lost due to heat transfer to the surroundings was included in the exergy destruction. The sum of the exergetic efficiency and the total exergy destruction as a percentage of the fuel exergy for a cycle is not 100 %, since the net power output was used to calculate the exergetic efficiency of the cycle instead of the gross power output. In addition, the method used to calculate the exergy in the heat transferred from the water recovery system to the surroundings was only an approximation. The largest exergy destruction occurred in the combustor, while the exergy destruction of the heat recovery system was low.

59 Advanced Power Cycles with Mixtures as the Working Fluid

50.5

50.0

49.5

49.0

48.5

48.0 Exergetic efficiency [%] efficiency Exergetic

47.5 GTX100 Trent 47.0 020406080100 β [%] Figure 4.7. Exergetic efficiencies of the evaporative cycles.

The main difference between the gas turbines was the pressure ratio, which is 20 for the GTX100 and 35 for the Trent. A high pressure ratio results in a low exhaust gas temperature; the expander outlet temperature was approximately 465 °C for the Trent cycles and 577 °C for the GTX100 cycles. The exergy flow rate of a stream depends on the mass flow rate, temperature, pressure and composition; hence, the exergy flow rate of the exhaust gas and the potential for heat recovery differed between the GTX100 and the Trent cycles. Another factor influencing the heat recovery is the cycle configuration. Since the Trent was not intercooled and had a high pressure ratio, the compressor outlet temperature was high, while the GTX100 was intercooled and had a medium pressure ratio, thus the compressor outlet temperature was relatively low. The different temperature levels explain the trends in Figure 4.7: in the GTX100 cycles, the compressor outlet temperature was relatively low and further cooling in the aftercooler and recuperation by the high-temperature exhaust gas was beneficial for cases with high percentages of part-flow. In the Trent cycles, on the other hand, the compressor outlet temperature was high and the air used for humidification could not be heated to its initial temperature by the relatively low-temperature exhaust gas and the air that bypassed the tower could not be recuperated; hence, partial flow cases had higher efficiencies than the full-flow case. In Figure 4.8, the difference in exergy flow rate of the exhaust gas between the expander outlet and after the heat recovery system (i.e., before the water recovery system) has been divided by the exergy flow rate at the expander outlet. The computed value can be viewed as an exergetic efficiency of the heat recovery system. The maximum exergetic efficiency for the heat recovery system occurred for the part-flow cases

60 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

77 GTX100 76 Trent

75

74

73 [%] 1 )/E

2 72 -E 1 71 (E 70

69

68

67 0 20406080100 β [%]

Figure 4.8. Difference in exergy flow rate between the expander outlet (E1) and after the heat recovery system (E2) divided by the exergy flow rate at the expander outlet (E1).

that had the highest cycle exergetic efficiencies. This shows the importance of the heat recovery system. Reducing the percentage of part-flow increases the humidification tower water outlet temperature. Thus, the water that recovers energy from the exhaust gas is warmer than in a full-flow case and the heat recovery is impaired. Nevertheless, Figure 4.8 shows that the exergetic efficiency of the heat recovery system was high at low part-flows for the Trent cases and almost constant down to 60 % part-flow for the GTX100 cases. Consequently, part-flow evaporative cycles have as efficient or more efficient heat recovery as full-flow cycles, while the investment cost is lower due to the reduced heat exchanger area and tower volume. Diagrams presenting the variations of the exergetic efficiency with the percentage of part-flow for different components are shown and discussed in more detail in Paper VI.

4.3.4 Economic Analysis of Evaporative Gas Turbines Different humidified gas turbine cycles have been presented and analyzed thermodynamically; however, in order to evaluate the cycles for different applications, economic analyses are required. In Paper VII, steam-injected part- flow and full-flow evaporative cycles were compared thermodynamically and economically with steam-injected gas turbines and combined cycles.

61 Advanced Power Cycles with Mixtures as the Working Fluid

4.3.4.1 Method and Cycle Configurations The input data used for the thermodynamic simulations of the cycles are shown in Table 4.1. The water outlet temperature was 130 °C since the deaerator required this temperature. A deaerator was included to reduce the oxygen content in the water so that the boilers and their economizers could be manufactured in carbon steel. The tower packing was stainless steel Norton Intalox 2T for which the height of a transfer unit was estimated from measurements in the Swedish evaporative gas turbine pilot plant (Dalili and Westermark, 2002). The three evaporative cycle configurations investigated in Paper VII differ from the configurations in Papers V and VI. In Paper VII, none of the cycles are intercooled, since the Trent, GTX100 and Cyclone gas turbines used in the simulations are non-intercooled originally. The same cycle configurations were simulated for all gas turbines. One of the evaporative cycles was a steam-injected part-flow evaporative (PEvGT) cycle with an ejector, shown in Figure 4.9. The ejector uses steam generated from energy in the compressed air and exhaust gas to increase the pressure of the humid air stream that has passed through the humidification system to the compressor outlet pressure. The principle of the ejector is explained in Paper VII. In comparison to the booster fan used for the same purpose in Papers V and VI, the ejector has no moving parts and is less expensive. The steam mass flow rate and pressure required by the ejector were

cooling air

compr. turbine G compr. turbine G fuel fuel

humid air superheater ejector steam superheater after- cooler economizer

DEA boiler humid air superheater feedwater boiler heater

after- feedwater cooler heater economizers hum. tower hum. DEA FGC tower feedwater heater air water feedwater heater air steam water FGC

Figure 4.9. Part-flow evaporative gas turbine (PEvGT, left) and full-flow evaporative gas turbine (FEvGT, right).

62 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

computed by a method, described in Paper VII, where the so-called thermodynamic efficiency of the ejector is utilized. The other two evaporative cycles investigated, the FSEvGT and FEvGT cycles, used full-flow humidification. The FSEvGT cycle has the same layout as the PEvGT cycle, although all the compressed air passes the humidification tower and the ejector is excluded, since it was assumed that the ejector would require excessive amounts of steam to function. The FEvGT cycle, shown in Figure 4.9, has no boilers and all water evaporation occurs in the humidification tower. In addition, steam-injected gas turbines were simulated. In these cycles, the exhaust gas energy is used to generate superheated steam that is added to the compressed air before the combustor. The equipment cost data were estimated from data found in literature and manufacturer data. Cost data from earlier years were updated with inflation indexes, as described in Paper VII. The cost data for the equipment were given in different currencies and the methods used for converting the costs are discussed in Paper VII. When the free on board cost was given for a piece of equipment, the method suggested by Guthrie (1969) was used to calculate the installed cost, as shown in Equation (4.1). In this method, the module factor (FM) accounts for direct and indirect costs due to the installation of the component. Materials other than carbon steel, design, pressure and temperature can increase the free on board cost, although these factors do not necessarily increase the installation cost. In such cases, a correction factor is used to adjust the installation cost. The module and correction factors used in the calculations were based on Guthrie (1969).

= ⋅ + 1 ()−  Installed cost Free on board cost 1 FM 1  (4.1)  correction factor 

Table 4.6. Cost data sources, module factors and correction factors.

Cost data source FM Corr. factor Gas turbine GTX100, Cyclone (Nilsson, 2002); 1.4 1 Trent (GTW, 1999) Hum. tower Oelkrug (1999) - - packing Hum. tower shell Oelkrug (1999) 4.34 2.6-3.3 + 1.31 Heat exchangers Wahlberg (2001) 3.39 1-7 Ejector Ulrich (1984) 2.95 1 Pumps Ulrich (1984) 3.48 1.9-2.9 Water recovery Cataldi (2001), Westermark (2001) - - system (installed cost) Chimney Ulrich (1984) (installed cost) - - 1Additional factor to allow for distributors, connections, fixtures and droplet separator.

63 Advanced Power Cycles with Mixtures as the Working Fluid

Table 4.7. Input data for the heat exchanger sizing and costs. Type Material U [W/m2, °C] Cost [USD/m2] S-SH smooth tubes ss 90 110 H-SH smooth tubes ss 90 110 Boiler smooth tubes cs 105 (eg), 450 55 (compressed air) ECO for tubes w. fins tubes in cs, 700 290 boiler on eg side fins in cs ECO for tubes w. fins tubes in ss, 700 350 hum. tower on eg side fins in cs FWH tubes w. fins tubes in ss, 700 350 on eg side fins in cs AC tubes w. fins tubes in ss, 700 350 on eg side fins in cs

Table 4.8. Indirect costs as fractions of the total equipment cost and assumptions for the cost of electricity. Indirect costs Fraction Land purchase, surveys, site preparations 0.05 Specific services (local) 0.01 Fees in addition to contactors’ fee 0.02 Contingency 0.10 Confidence limit 0.02 Contractors’ fee 0.03 Total indirect costs 0.23 Assumptions for the cost of electricity Interest rate [-] 0.06 Economic life [years] 20 Fuel price [USD/MWh fuel] 10.3 Fixed operation and maintenance cost [% of investment cost] 2 Variable operation and maintenance cost [USD/MWh fuel] 0.83 Annual full-load hours [hours/year] 6,000

The sources of the cost data, the module factors and the correction factors, shown in Table 4.6, are discussed in detail in Paper VII. Table 4.7 presents the overall heat transfer coefficients (Hewitt et al., 1994) and cost data (Wahlberg, 2001) for the heat exchangers. The investment cost for the power plant is the sum of the direct costs (i.e., equipment) and indirect costs (e.g., land purchase and contingency). The indirect costs and the assumptions for the calculation of the cost of electricity are shown in Table 4.8.

64 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

4.3.4.2 Results The power outputs and electrical efficiencies for all cases are shown in Figure 4.10. The cycles with the highest degree of humidification, defined as the mass flow rate of water evaporated in the humidification tower and/or boilers divided by the compressor inlet air mass flow rate, had the highest power outputs. For the Trent, the part-flow cycles had lower power outputs and higher efficiencies than the full- flow cycles, while the situation was the opposite for the GTX100 and Cyclone. The main difference between the GTX100 and Cyclone cycles on the one hand and the Trent cycles on the other hand was the gas turbine pressure ratio, which was 35 for the Trent, 20 for the GTX100 and 17 for the Cyclone. The pressure ratio determines the compressor outlet temperature, which was 579 °C for the Trent, 450 °C for the GTX100 and 408 °C for the Cyclone, and the expander outlet temperature, which was 466 °C in the Trent cases, 590 °C in the GTX100 cases and 486 °C in the Cyclone cases. These temperature levels imply that when the percentage of part-flow was decreased for the Trent evaporative cycle, more of

73.0 52.5 Trent Trent 72.0 52.0 71.0 51.5 70.0

69.0 [%] 51.0 el η P [MW] 68.0 FEvGT 50.5 67.0 STIG FSEvGT 50.0 66.0 PEvGT 65.0 49.5 55.5 50.0 GTX100 55.0 GTX100 49.5 54.5 49.0 54.0

[%] 48.5 el

53.5 η P [MW] 48.0 53.0 52.5 47.5 52.0 47.0 17.0 50.0 Cyclone 16.8 16.6 49.0 Cyclone 16.4

[%] 48.0 16.2 el η P [MW] 16.0 47.0 15.8 15.6 46.0 0 1020304050 0 1020304050 β [%] β [%] Figure 4.10. Power outputs and electrical efficiencies for all cases.

65 Advanced Power Cycles with Mixtures as the Working Fluid

the high-temperature compressed air bypassed the tower and was not cooled in the aftercooler; hence, the electrical efficiency was higher for the part-flow cases than for the full-flow cases. For the GTX100 and Cyclone cases, the situation was the opposite: when the percentage part-flow was decreased, the recuperation of the high-temperature exhaust gas was reduced and more of the relatively cold compressor outlet air bypassed the tower without any recuperation at all. The FSEvGT cycles had slightly higher power outputs and electrical efficiencies than the FEvGT cycles, since the combination of boilers and a humidification tower recovered heat more efficiently than merely a humidification tower. The electrical efficiencies of the steam-injected cycles were low for all gas turbines, since there was no recuperation in these cycles and the evaporator section pinch point in the HRSG restricted the heat recovery from the exhaust gas. The power outputs of the GTX100 and Cyclone steam-injected cycles were relatively higher than the power output of the steam-injected Trent, since a high exhaust gas temperature is beneficial for heat recovery with a boiler.

41000 740 ]

Trent e Cyclone 40000 730 720 39000 710 38000 700 37000 690 FEvGT 680 36000 STIG 670 35000 FSEvGT

Investment cost [kUSD] 660

PEvGT Spec. inv. cost [USD/kW 34000 650 31500 600 ]

GTX100 e GTX100 590 31000 580 30500 570 30000 560 550 29500 540 29000 530 Investment cost [kUSD] 28500 Spec. inv. cost [USD/kW 520 11800 560 Cyclone Trent 555 11600 550 11400 545 540 11200 535 530

11000 Spec. inv. cost 525

Investment cost [kUSD] 10800 520 0 1020304050 0 1020304050 β [%] β [%] Figure 4.11. Investment costs and specific investment costs for all cases.

66 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

The investment costs and specific investment costs are shown in Figure 4.11. Any irregularities in the graphs in Figures 4.11 and 4.12 are due to the factors used in Guthrie’s method (Guthrie, 1969) for calculation of the installed equipment cost. The part-flow cycles had lower specific investment costs than the full-flow cycles for all gas turbines, since the heat exchanger areas and tower volumes were smaller for the part-flow cases, while, at least for the GTX100 and Cyclone cases, the power outputs were higher for the part-flow cases than for the full-flow cases. The FSEvGT cases had higher total and specific investment costs than the FEvGT cases, since the thermodynamic performances were similar for the cycles, while the FSEvGT cycle recovered energy with boilers that were more expensive than the large humidification tower in the FEvGT cycle. The steam-injected cycles had the lowest total and specific investment costs since the cycle configuration was simple. The trends for the costs of electricity for all cases, shown in Figure 4.12, follow the trends for the electrical efficiencies. The Trent cycles had the highest electrical

35.80 36.1 Cyclone Cyclone, PEvGT 45 % 36.0 35.70 35.9 35.60 35.8 35.7 35.50 FEvGT 35.6 STIG 35.5 35.40 FSEvGT 35.4 PEvGT Cost of electricity [mills/kWh] electricity of Cost Cost of electricity [mills/kWh] electricity of Cost 35.30 35.3 33.10 33.3 GTX100 GTX100, PEvGT 50 % 33.2 33.00 33.1 32.90 33.0 32.80 32.9 32.8 32.70 32.7 32.60 32.6

Cost of electricity [mills/kWh] electricity of Cost 32.50 [mills/kWh] electricity of Cost 32.5 32.00 32.3 Trent Trent, PEvGT 15 % 31.80 32.1 31.9 31.60 31.7 31.40 31.5 31.3 31.20 31.1

31.00 Cost of electricity Cost of electricity 30.9 30.80 30.7 0 1020304050 low med high β [%] Temperature differences Figure 4.12. Costs of electricity for all cases and costs of electricity when the minimum temperature differences in the heat exchangers are varied.

67 Advanced Power Cycles with Mixtures as the Working Fluid

STIG 16.6 46.2 0.218 76.7 16.3 4.7 1.6 - 0.6 - 10,814 652 35.77 FEvGT 15.7 49.5 0.192 72.1 14.2 7.4 1.6 4.0 0.7 - 11,279 719 35.35 Cyclone or inlet mass flow rate of rate of mass flow or inlet FSEvGT 15.7 49.6 0.195 70.2 14.0 11.0 1.5 2.7 0.6 - 11,588 737 35.63

=45 %) β ( 16.7 48.5 0.226 71.2 15.6 7.1 1.5 2.1 1.1 1.3 11,665 699 35.44 PEvGT STIG 54.4 47.4 0.217 73.7 19.2 5.6 1.2 - 0.3 - 28,828 530 32.98 FEvGT 52.4 49.8 0.211 69.7 17.9 7.9 1.2 2.9 0.4 - 30,057 574 32.62 GTX100 FSEvGT 52.5 49.9 0.215 67.1 17.5 12.1 1.1 1.8 0.3 - 31,226 594 32.93

=50 %) β PEvGT ( 55.1 49.5 0.235 68.6 18.8 8.5 1.1 1.6 0.7 0.7 31,032 564 32.55 STIG 65.1 50.0 0.124 80.6 13.7 3.8 1.7 - 0.3 - 34,121 524 31.65 FEvGT 72.2 50.8 0.232 70.9 19.0 6.4 1.2 2.0 0.6 - 38,810 538 31.54 Trent

2 FSEvGT 72.5 50.9 0.237 68.3 18.6 10.2 1.1 1.2 0.6 - 40,267 555 31.79

=15 %)

β ( PEvGT 70.2 52.1 0.180 73.5 16.6 6.7 1.3 0.7 0.6 0.5 37,425 533 30.90 Table 4.9. Detailed results for the Trent, GTX100 and Cyclone cases. ] e

1 [kg/kg] air

] /m % [

water el Degree of humidification: mass flow rate of water evaporated in the humidification tower and/or boilers divided by the compress divided tower and/or boilers the humidification in water evaporated rate of mass flow humidification: Degree of costs are included. The installation m Cost of electricity [mills/kWh] Gas turbine [%] Water recovery [%] Heat exchangers [%] Chimney [%] Humidification tower [%] Pumps [%] Ejector [%] Investment cost [kUSD] Spec. invest. cost [USD/kW η P [MW] Equipment cost/Total equipment cost [%] air. 1 2

68 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

efficiencies and the lowest costs of electricity. The largest difference in cost of electricity between different cycle configurations was found for the Trent, where the difference between the highest and the lowest cost of electricity was 3 % (i.e., 0.9 mills/kWh). For the GTX100 and Cyclone cases, this difference was about 1 % (i.e., 0.4 mills/kWh). For both the Trent and GTX100 cases, a part-flow cycle had the lowest cost of electricity, while for the Cyclone cases, a full-flow case had the lowest cost of electricity, although the difference to the part-flow case with the lowest cost of electricity was small. Detailed results for the part-flow cases with the lowest cost of electricity, the full-flow cases and the steam-injected cycles are presented in Table 4.9. For all cycles, the turbine accounted for the major part of the equipment cost, about 70 % in most cases. The water recovery system was expensive, since a system with indirect cooling of the exhaust gas with water that is indirectly cooled by ambient air was used in the analysis. To investigate the influence of the heat exchanger characteristics on the thermodynamic and economic results of the cycles, the heat exchanger temperature differences were varied in a sensitivity analysis. The investigated temperature differences were 20 °C, 30 °C and 40 °C for gas-gas heat exchange, 10 °C, 15 °C and 20 °C for gas-liquid heat exchange and 5 °C, 10 °C and 15 °C for the boiler pinch point. For all cases, the variations of the power outputs and electrical efficiencies were between 0 % and 2 % when the heat exchanger temperature differences were changed and the variations of the total and specific investment costs were between 1 % and 8 %. The variations of the cost of electricity were between 0 % and 1.5 %, as shown in Figure 4.12. Variations of the assumptions used for calculating the cost of electricity influenced the cost of electricity more than the heat exchanger temperature differences. For example, increasing the fuel price from 10 USD/MWh fuel to 15 USD/MWh fuel raised the cost of electricity for all cases by 30 %. When the number of annual full-load hours was increased from 6,000 hours/year to 8,000 hours/year, the cost of electricity was reduced by 10 %. An increase of the interest rate from 6 % to 8 % raised the cost of electricity with 5 %. The results for the Trent and GTX100 cycles were compared with the performance of the competing gas turbine technology for mid-size base-load power generation, that is, the combined cycle. In Table 4.10, data for the evaporative cycles with the lowest costs of electricity are compared with literature data for combined cycles with dual-pressure steam bottoming cycles (GTW, 2001). The combined cycle cost data were updated with inflation indexes and the costs of electricity were calculated with the assumptions in Table 4.8. The data in the table cannot be compared directly, since they are based on different assumptions; however, a qualitative assessment is possible. The costs of electricity for the evaporative and combined cycles were in the same range, while the investment cost was 16-25 % lower and the specific investment cost was 12-15 % lower for the evaporative cycles.

69 Advanced Power Cycles with Mixtures as the Working Fluid

Table 4.10. Comparison between PEvGT cycles and combined cycles based on the Trent and GTX100. Trent GTX100 Combined PEvGT Combined PEvGT cycle (β=15 %) cycle (β=50 %) P [MW] 74.21 70.2 62.01 55.1 1 1 ηel [%] 52.7 52.1 54.0 49.5 Investment cost [kUSD] 44,7771 37,425 41,2551 31,032 Specific investment cost 604 533 665 564 [USD/kWe] Cost of electricity [mills/kWh] 31.9 30.9 32.5 32.6 1(GTW, 2001)

4.4 Discussion

In the second part of this thesis, evaporative gas turbines with part- and full-flow humidification for base-load power generation have been investigated thermodynamically and economically. The power outputs of the evaporative cycles studied varied between 16 MWe and 108 MWe and the electrical efficiencies varied between 47 % and 53 %.

4.4.1 Method and Assumptions

4.4.1.1 Thermodynamic Analysis The assumptions made for the simulations influence the results of the analysis. For example, in Papers V and VI, the compressor intake air mass flow rate in the evaporative cycles was the same as in the simple cycle gas turbines. Hence, the expander size in the evaporative cycles was greatly increased, since the working fluid was humidified. Thus, the gas turbine models used in the simulations were not exactly comparable to the gas turbines that are available from the manufacturers. However, the gas turbine models had the same characteristics and level of technology as the manufactured gas turbines. In Paper VII, the expander size was kept constant in all cycles by a constant expander outlet volumetric flow rate. This method resulted in reduced compressor sizes for the evaporative cycles compared with the simple cycle gas turbines. It is problematic to define a “constant expander size”, since different turbine parameters could be considered for this condition. The method used in this study ensured that the expander size was not excessively increased, so that the thermodynamic and economic parameters assumed for the calculations were still valid. Different thermodynamic property models were used in the cycle simulations in GateCycle, where an ideal mixture model was utilized, and in the humidification tower model, where a real gas model was utilized. The water concentration and

70 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

temperature of the saturated tower outlet air stream calculated in the tower model corresponded to a state below the dew point in the GateCycle model. If the values from the tower model were used directly in GateCycle, water evaporation in the humid air superheater would result in an unrealistic heat load. Therefore, the GateCycle dew point was calculated for the water concentration from the tower model, and this value was used as the tower outlet temperature. The GateCycle dew point was at the most 3.5 °C higher than the tower model dew point. The sensible energy required to heat the humid air stream 3.5 °C is small compared with the energy required to heat the humid air stream from below the dew point to the same final temperature. The solution to the problem of dissimilar thermodynamic property models in different cycle components is to calculate the evaporative cycle in a simulation program where a real gas model for the thermodynamic properties and the tower model have been implemented. In Paper VII, a simplified model was utilized for calculation of the ejector. To accurately design the ejector, more advanced methods are required; however, these methods were deemed to complicated for a parametric study of many cycle cases. However, a low value for the ejector efficiency was adopted to achieve a conservative design. The steam-injected full-flow evaporative configuration was included as a comparison for the part-flow cases. For base-load power generation, the full-flow evaporative cycle with only a humidification tower is more appropriate.

4.4.1.2 Economic Analysis In the economic study in Paper VII, the equipment cost data were retrieved from different sources, which introduced uncertainties. The cost to design and develop an evaporative gas turbine was not included, in order to compare the simulated cycles with the conventional technology (i.e., the combined cycle) on a mature-cost basis. A water recovery system where the exhaust gas is indirectly cooled in a heat exchanger and the cooling water is cooled indirectly by ambient air was chosen, since this system has no water consumption and can be used if cooling water from the environment is unavailable. The costs for the ejector, pumps and stack (Ulrich, 1984) are valid for the year 1982 and were updated to the year 2002 with an inflation index; therefore, these costs are relatively uncertain. However, these components accounted for only 1-4 % of the total installed equipment cost. The original cost for the water recovery system was in Swedish currency and had to be converted to US currency. This poses a problem, since the exchange rate varies from year to year and does not give the correct cost for manufacturing a piece of equipment in another country. Hence, an exchange rate based on the gross domestic product per inhabitant was used, as explained in Paper VII. The exchange rate calculated with this method was 6.9 SEK/USD, compared with the exchange rate of August 2002, which was 9.7 SEK/USD. Since the water recovery system was the second most expensive cycle component, the exchange rate had a significant impact on the economic results. With an exchange rate based on the gross domestic product, the water recovery system accounted for 14-19 % of the

71 Advanced Power Cycles with Mixtures as the Working Fluid

installed equipment cost, while with the current exchange rate, 10-14 % of the equipment cost was accounted for by the water recovery system. By using the current exchange rate for the water recovery system, the cost of electricity was reduced by approximately 0.5 mills/kWh for the cycles in Table 4.9. To account for the installation cost for a piece of equipment, the method developed by Guthrie (1969) was used for the humidification tower shell, heat exchangers, ejector and pumps. The factors used for the estimation of the installed cost are relatively old and may not be accurate today. Moreover, the component that influences the investment cost the most is the gas turbine, and for this component an installation cost factor supplied by ALSTOM was used. The assumptions made for the calculation of the cost of electricity influence the results; however, the same assumptions were used for all cycles, including the combined cycles for which cost data from literature were used, to facilitate the comparison between the different cycles. The GateCycle models of the Trent and GTX100 are based on data that are rather old compared with the combined cycle literature data. In the comparison between the evaporative and the combined cycles, this is to the disadvantage of the evaporative cycles, since the performance of the gas turbines has been improved since the gas turbine models were created.

4.4.2 Economical and Technical Aspects The high specific power outputs of steam-injected and evaporative gas turbines decrease the specific investment costs compared with combined cycles, which require a costly steam bottoming cycle. The evaporative cycle investment cost can be further reduced by part-flow humidification, since this decreases the heat exchanger areas and tower volumes compared with full-flow cycles, yet retains or increases the electrical efficiency. Part-flow humidification can be combined with steam generation to efficiently recover the exhaust gas energy. When steam is available, steam cooling of the gas turbine can further increase the electrical efficiency. A natural gas-fired evaporative cycle has a higher electrical efficiency than water-injected and steam-injected gas turbines and an efficiency on the same level or higher than a combined cycle. To reach the highest possible efficiency, the evaporative cycle should be intercooled and recuperated; however, intercoolers and recuperators are not standard technology for gas turbines for power generation. Furthermore, an evaporative cycle will experience a flow mismatch between the compressor and expander. A new gas turbine, for example with a compressor from a smaller gas turbine and an expander from a larger gas turbine, could be designed for accommodating the increased flow; however, the cost would probably be substantial, as indicated by the studies on the FT 4000 HAT gas turbine (EPRI, 1993). Nilsson et al. (2001) suggested using a free power turbine and/or multiple spools, although this is most suitable for medium-size and large-size gas turbines, while for single-shaft gas turbines, a compromise for the compressor and turbine design can be made or a gear used between them. In steam-injected gas turbines, the flow mismatch has been solved by using only partial steam addition or by using

72 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

gas turbines with large surge margins. For the evaporative cycle, the CHAT concept has been suggested to enable using standard gas turbine components in a humidified cycle. The CASH and CASHING cycles solve the flow mismatch problem as well. Apart from the gas turbine and the intercooler and recuperator, if included, the rest of the cycle consists of standard components, although the humidification tower has not been used previously in power plants. The humid air reduces the formation of nitrogen oxides in the combustion process without the use of an expensive dry low-NOx combustor. When only low levels of NOx are generated, the need for expensive post-combustion removal of NOx, for example by selective catalytic or non-catalytic reduction, is reduced. In addition, gas turbines with high concentrations of water in the exhaust gas have a potential for less expensive carbon dioxide separation than dry gas turbines. Humidified gas turbines consume water, which must have a sufficiently high quality to not endanger the cycle components. Experiences from steam-injected gas turbines and from the Swedish evaporative gas turbine pilot plant show that gas turbines can be operated with humid air working fluid without corrosion or combustion problems. The water consumption of humidified cycles increases the operational cost. However, it has been shown in several installations that the water can be recovered from the exhaust gas and treated to achieve a quality sufficient for re-use in the cycle. The cost of the makeup water, of a suitable quality, determines if the water should be recovered or not. The power output and efficiency penalties caused by part-load and high ambient temperatures are smaller for humidified than for dry gas turbines, since the humidification rate can be changed to compensate for variations in load and ambient conditions. Moreover, for cogeneration of power and steam or district heating, the steam-injected and evaporative cycles are more flexible than combined cycles. These wet cycles are interesting for distributed power generation on deregulated markets, where small, flexible power plants with high efficiencies and low electricity generation costs are required. Although the use of natural gas will probably increase in the future, solid fuels are less expensive and more abundant, and evaporative cycles could use these fuels efficiently. For reduced natural gas consumption, evaporative cycles can integrate external heat sources (e.g., refuse) effectively. In addition, evaporative cycles fueled with gasified coal or biomass have a potential for higher efficiency and lower investment costs and costs of electricity than combined cycles. Externally-fired evaporative cycles show good performance as well.

4.4.3 Results In the first study on evaporative gas turbines, steam-injected part-flow evaporative cycles were compared with gas turbines with ammonia-water bottoming processes. The study showed that the ammonia-water cycle could use energy at the relatively low temperature level of the exhaust gas from the high pressure ratio Trent more efficiently than the evaporative cycle. For the Trent gas turbine, all ammonia-water combined cycles had higher electrical efficiencies than the evaporative cycle, while

73 Advanced Power Cycles with Mixtures as the Working Fluid

for the GTX100 with a medium pressure ratio, combined cycles with bottoming configurations I and II had lower efficiencies than the evaporative cycle and configuration IV had a higher efficiency. The combined cycle with configuration IV had 1.3-1.4 percentage points higher efficiency than the evaporative cycle for both gas turbines, which shows that this ammonia-water cycle configuration is efficient for both aeroderivative and industrial gas turbines due to the flexibility introduced by the ammonia-water mixture working fluid. The exclusion of the bottoming cycle and the higher specific power output should result in a lower investment cost for the evaporative cycle compared with an ammonia-water combined cycle. On the other hand, the humidification of the working fluid requires gas turbine design changes, which are unnecessary for a combined cycle. In the second study, an exergy analysis was performed on the evaporative cycles from the first study to evaluate the performance of part-flow humidification systems. The highest exergetic efficiency was found for the GTX100 evaporative cycle with full-flow humidification, although the exergetic efficiency was relatively constant from full-flow humidification down to 60 % part-flow. The highest exergetic efficiency for the Trent cycle was found to be with a 20 % part-flow. The largest exergy destruction occurred in the combustor, while the exergy destruction of the heat recovery system was low. The maximum exergetic efficiency for the heat recovery system occurred for the part-flow cases that had the highest cycle exergetic efficiencies, which shows the importance of the heat recovery system. One feature of part-flow evaporative cycles that could impair the potential for heat recovery is that a reduced percentage of part-flow increases the humidification tower water outlet temperature. Nevertheless, the exergetic efficiency of the heat recovery system was high at low part-flows for the Trent cases and almost constant down to 60 % part-flow for the GTX100 cases. Hence, the study showed that part-flow evaporative cycles can recover energy equally well or better than full-flow cycles. In the third study, part-flow humidification was further studied. Steam-injected part-flow evaporative cycles were compared with full-flow and steam-injected cycles for the non-intercooled Trent, GTX100 and Cyclone gas turbines. The results for the Trent and GTX100 cycles with the lowest costs of electricity, which were part-flow cycles, were compared with literature data for dual-pressure combined cycles. The calculated cost of electricity for the evaporative cycles was lower or almost equal to the combined cycle cost of electricity, while the investment cost was 16-25 % lower and the specific investment cost was 12-15 % lower for the evaporative cycles. For the high pressure ratio Trent, the part-flow cycles had lower power outputs and higher efficiencies than the full-flow cycles, while the situation was the opposite for the medium pressure ratio GTX100 and Cyclone. The specific investment costs were lower for the part-flow cases than for the full-flow cases for all gas turbines, since the heat exchanger areas and tower volumes were lower for the part-flow cases. Furthermore, the part-flow cycles based on the GTX100 and Cyclone gas turbines had higher power outputs than the full-flow cycles. Consequently, the part-flow cycles had lower costs of

74 4 Part- and Full-Flow Evaporative Gas Turbines for Power Generation

electricity compared with the full-flow cycles, except for the Cyclone for which the full-flow cycle with only a humidification tower had a slightly lower cost of electricity. This shows that part-flow evaporative cycles are viable, especially for aeroderivative gas turbines with high pressure ratios. However, the difference in cost of electricity between different cycles was only 1-3 % (0.4-0.9 mills/kWh), with the largest differences between Trent-based cycles. The steam-injected cycles had the lowest total and specific investment costs since the cycle configuration was simple, while the costs of electricity were high due to the low electrical efficiencies. A sensitivity analysis showed that variation of the heat exchanger temperature differences did not influence the cost of electricity much, while the assumptions for the calculation of the cost of electricity had a larger influence. When the fuel price is increased, a cycle with a high electrical efficiency experiences a lower rise in the cost of electricity, which is to the advantage of the evaporative cycle.

4.4.4 Suggestions for Future Work The evaporative cycle requires the development of a gas turbine that can accommodate the flow variation between the compressor and expander, with intercooling and recuperation for applications where high electrical efficiencies are necessary. Accurate thermophysical properties for air-water mixtures are required for reliable performance calculations, component design and cost estimation for the evaporative cycle. Ideal models for the air-water mixture are not accurate for the whole pressure and temperature range of an evaporative cycle and real models are required, for which more experimental data on air-water mixtures are needed for model validation. The evaporative cycle is still to be demonstrated in a larger scale to evaluate the full potential of the technology. The possible applications for the evaporative cycle are many, for example base-load power generation and cogeneration of power and steam or hot water. However, Bartlett (2002) identified district heating as a feasible niche market for demonstration of the evaporative cycle.

75

5 Conclusions

In the first part of the thesis, ammonia-water cycles as bottoming processes for natural gas-fired gas and gas-diesel engines were investigated. It was shown that the ammonia-water cycle has a better thermodynamic performance than the steam Rankine cycle for this application. All simulated ammonia-water cycle configurations generated more power than the steam cycles, except for one simple ammonia-water cycle configuration compared with a dual-pressure steam cycle. The best ammonia-water bottoming cycle could generate 40-50 % more power than a single-pressure steam cycle and 20-24 % more power than a dual-pressure steam cycle. An ammonia-water bottoming cycle could add 6-8 percentage points in efficiency to the gas engines, while a single-pressure steam bottoming cycle could add about 5 percentage points. For the gas-diesel engines, the efficiency augmentation was 4-7 percentage points for the ammonia-water bottoming cycles, 4-5 percentage points for a single-pressure steam cycle and 4-6 percentage points for a dual-pressure steam cycle. At this time, small-scale power generation from geothermal heat sources and industrial waste heat appears to be the most likely applications for the ammonia-water cycle, since some plants for these applications have already been constructed. If the costs of fuel and electricity increase, recovering different waste heat sources for power generation will become more interesting. The waste heat from reciprocating engine power plants could be one such source and the ammonia-water cycle is an advantageous alternative compared with the steam cycle for this application. In the second part of this thesis, evaporative gas turbines with part- and full- flow humidification were investigated. It was shown that a combined cycle with an ammonia-water bottoming process could use exhaust gas energy at the temperature level of an aeroderivative high pressure ratio gas turbine more efficiently than a steam-injected part-flow evaporative gas turbine. For an industrial medium pressure ratio gas turbine, an evaporative cycle had a higher electrical efficiency than the combined ammonia-water cycles, except for an ammonia-water cycle configuration with reheat. Compared with the ammonia-water combined cycle, the specific investment cost should be lower for the evaporative cycle since it excludes the turbine of the bottoming cycle; however, the evaporative cycle requires design modifications for the gas turbine that are unnecessary for a combined cycle. An exergy analysis of the steam-injected part-flow evaporative cycles showed that the full-flow case for the industrial gas turbine had the highest exergetic efficiency, although the efficiency was almost constant down to a part- flow of 60 %, and for the aeroderivative gas turbine, the 20 % part-flow case had the highest exergetic efficiency. The exergy destruction in the heat recovery system was low and a heat recovery system exergetic efficiency showed that part-flow systems could recover energy as efficiently as, or more efficiently than, full-flow

77 Advanced Power Cycles with Mixtures as the Working Fluid

systems. An economic analysis showed that the specific investment costs were lower for part-flow than for full-flow evaporative cycles for one aeroderivative gas turbine and two industrial gas turbines. The lower specific investment costs are due to reduced heat exchanger areas and tower volumes for the part-flow cases compared with the full-flow cases, while the power outputs of part- and full-flow cases were similar. In addition, the part-flow cycles had lower costs of electricity compared with the full-flow cycles, except for the small industrial gas turbine for which a full-flow cycle had a slightly lower cost of electricity. Compared with combined cycle literature data, the costs of electricity for the part-flow evaporative cycles were lower or equal, while the total and specific investment costs were significantly lower for the evaporative cycles. This shows that part-flow evaporative cycles can compete with the conventional mid-size power generation systems used today.

78

6 References

Aagren, N. D., Cavani, A. and Westermark, M. O. (1997a). New Humidifier Concept in Evaporative Gas Turbine Cycles. In: Proceedings of TAIES’97, Thermodynamic Analysis and Improvement of Energy Systems, Beijing, China, Jun. 10-13, 1997. 270-276. ISBN 7-5062-3264-Z/TK. Aagren, N. D., Cavani, A. and Westermark, M. O. (1997b). New Humidification Concept for Evaporative Gas Turbine Cycles Applied to a Modern Aeroderivative Gas Turbine. In: Proceedings of the ASME Advanced Energy Systems Division. AES-Vol. 37, 223-230. ISBN 0-7918-1845-4. Agarwal, A. K. and Das, L. M. (2000). Biodiesel Development and Characterization for Use as a Fuel in Compression Ignition Engine. In: Proceedings of the 2000 Spring Technical Conference of the ASME Internal Combustion Engine Division, San Antonio, Texas, USA, Apr. 9-12, 2000. ICE-Vol. 34-3, 17-25. ASME Paper No. 2000-ICE-286. ISBN 0-7918-1664-8. Ahnger, A. (1996). Advanced Diesel Combined Cycle (DCC) Plant. In: Proceedings of Power-Gen International ‘96, Orlando, Florida, USA, Dec. 4-6, 1996. Amano, Y., Hashizume, T., Usui, A., Takeshita, K., Akiba, M. and Tanzawa, Y. (2001). Experimental Results of an Ammonia-Water Mixture Turbine System. In: Proceedings of the 2001 International Joint Power Generation Conference, New Orleans, Louisiana, USA, Jun. 4-7, 2001. ASME Paper No. JPGC2001/PWR-19004. Annerwall, K. and Svedberg, G. (1991). A Study on Modified Gas Turbine Systems with Steam Injection or Evaporative Regeneration. In: Proceedings of 1991 ASME Cogen-Turbo, Budapest, Hungary, Sep. 3-5, 1991. IGTI-Vol. 6, 1-7. ISBN 0-7918-0650-2. Anonymous (1989). Kraftvärme – En teknisk och ekonomisk utredning om optimal el- och värmeproduktion i liten och medelstor skala. Stockholm, Sweden: Vattenfall. FUD- rapport. U (S) 1989/33. ISSN 1100-5130. In Swedish. Anonymous (1992). Diesel Combined Cycle Plant Burns Compressed Natural Gas. Power Engineering, 96(6), 56. ISSN 0032-5961. Anonymous (1998). Kohinoor – Combined Cycle Power Master. Energy News, No. 2, 10-12. Wärtsilä NSD Corporation. ISSN 0782-9779. Anonymous (2001). The TopHat Turbine Cycle. Modern Power Systems, 21(4), 35-37. ISSN 0260-7840. Anxionnaz, R. (1952). Improvements in or Relating to Gas Turbine Plants. British Patent No. 676,008. Aronis, N. and Leithner, R. (2002). Combined Cycle with Low Quality Heat Integration and Water Injection into the Compressed Air. In: Proceedings of ECOS 2002, Berlin, Germany, Jul. 3-5, 2002. Vol. II, 1135-1142. ISBN 3-00-009533-0.

79 Advanced Power Cycles with Mixtures as the Working Fluid

Ashley, S. (1998). Fuel-Saving Warship Drives. Mechanical Engineering, 120(8), 65-67. ISSN 0025-6501. Badeer, G. H. (2000). GE Aeroderivative Gas Turbines – Design and Operating Features. http://www.gepower.com/publications/en_us/pdf/GER3695e.pdf. GE Power Systems. GER-3695E (10/00). Access date Nov. 25, 2002. Bartlett, M. (2002). Developing Humidified Gas Turbine Cycles. Stockholm, Sweden: Dept. of Chemical Engineering and Technology, Div. of Energy Processes, Royal Institute of Technology. Doctoral Thesis. ISRN KTH/KET/R--165--SE. Bartlett, M. A. and Westermark, M. O. (2001a). Modelling of Alkali Contaminant Flows in Evaporative Gas Turbines. In: Proceedings of ASME Turbo Expo 2001, New Orleans, Louisiana, USA, Jun. 4-7, 2001. ASME Paper No. 2001-GT-0096. Bartlett, M. A. and Westermark, M. O. (2001b). Experimental Evaluations of Air Filters and Metal Ion Migration in Evaporative Gas Turbine Cycles. In: Proceedings of the 2001 International Joint Power Generation Conference, New Orleans, Louisiana, USA, Jun. 4-7, 2001. ASME Paper No. JPGC2001/PWR-19119. Bartlett, M. A., Wikman, K., Holmgren, K. and Westermark, M. O. (2002). Effective Waste Incineration in Hybrid Cycles for CHP Applications – A Cycle and System Study. In: Proceedings of ECOS 2002, Berlin, Germany, Jul. 3-5, 2002. Vol. II, 804-813. ISBN 3-00-009533-0. Bartlett, M. A. and Westermark, M. O. (2003a). A Study of Humidified Gas Turbines for Short-Term Realisation in Mid-Sized Power Generation - Part A: Non-Intercooled Cycle Analysis. Accepted for publication in: Proceedings of ASME Turbo Expo 2003, Atlanta, Georgia, USA, Jun. 16-19, 2003. ASME Paper No. GT2003-38402. Bartlett, M. A. and Westermark, M. O. (2003b). A Study of Humidified Gas Turbines for Short-Term Realisation in Mid-Sized Power Generation - Part B: Intercooled Cycle Analysis and Final Economic Evaluation. Accepted for publication in: Proceedings of ASME Turbo Expo 2003, Atlanta, Georgia, USA, Jun. 16-19, 2003. ASME Paper No. GT2003-38403. Bathie, W. W. (1996). Fundamentals of Gas Turbines. Second edition. John Wiley and Sons. ISBN 0-471-31122-7. de Biasi, V. (1999). 12-MW Demo Plant Proposed to Prove Out CHAT Technology. Gas Turbine World, 29(3), 22-25. ISSN 0747-7988. de Biasi, V. (2000a). LM6000 Sprint Design Enhanced to Increase Power and Efficiency. Gas Turbine World, 30(4), 16-19. ISSN 0747-7988. de Biasi, V. (2000b). Aquarius Design Recovers Water from Exhaust for Steam Injection. Gas Turbine World, 30(2), 14-17. ISSN 0747-7988. de Biasi, V. (2001a). 11-MW UGT 10000 Gas Turbine in Service on Ukrainian Pipeline. Gas Turbine World, 31(6), 14-18. ISSN 0747-7988. de Biasi, V. (2001b). DOE Evaluating CHAT for Next Generation Gas Turbine Program. Gas Turbine World, 31(3), 12-14, 16-17. ISSN 0747-7988.

80 6 References

Bisio, G. (1992). Industrial Waste Heat Recovery: The Ideal . In: Proceedings of the Florence World Energy Research Symposium, Florence, Italy, Jun. 7-12, 1992. 503-519. ISBN 1-56072-082-4. Bjorge, R. W., Corman, J. C. and Smith, R. W. (1996). Kalina Cycle Application to Gas Turbine Combined Cycles. In: Proceedings of Power-Gen Europe ‘96, Budapest, Hungary, 1996. 108-123. Bjorge, R. W., Boericke, R., O’Connor, M. F. and Smith, R. W. (1997). Kalina Combined Cycle Power Plant Design and Performance Characteristics. In: Proceedings of Power-Gen Europe ‘97, Madrid, Spain, 1997. 217-235. Blanco, G. and Ambs, L. L. (2002). Water Recovery Systems for Steam-Injected Gas Turbines: Size Optimization and Life Cycle Savings. In: Proceedings of the 2002 International Joint Power Generation Conference, Phoenix, Arizona, USA, Jun. 24-26, 2002. ASME Paper No. IJPGC2002-26049. Bolland, O. and Stadaas, J. F. (1993). Comparative Evaluation of Combined Cycles and Gas Turbine Systems with Water Injection, Steam Injection, and Recuperation. In: Proceedings of the International Gas Turbine and Aeroengine Congress and Exposition, Cincinnati, Ohio, USA, May 24-27, 1993. ASME Paper No. 93-GT-57. Bram, S. and de Ruyck, J. (1997). Exergy Analysis Tools for ASPEN Applied to Evaporative Cycle Design. Energy Conversion and Management, 38(15-17), 1613-1624. ISSN 0196-8904. Bärring, M., Gustafsson, J.-O., Nilsson, P.-A., Ohlsson, H. and Olsson, F. (2000). El från nya anläggningar. Stockholm, Sweden: Elforsk. Report No. 00:01. In Swedish. Capstone (2003). http://www.microturbine.com/. Access date Feb. 8, 2003. Carcasci, C., Facchini, B. and Harvey, S. (1998). Modular Approach to Analysis of Chemically Recuperated Gas Turbine Cycles. Energy Conversion and Management, 39(16-18), 1693-1703. ISSN 0196-8904. Cataldi, G. (2001). Dry Air-Cooling for Water Recovery in Humidified Gas Turbine Cycles. Stockholm, Sweden: Dept. of Chemical Engineering and Technology, Div. of Energy Processes, Royal Institute of Technology. Master Thesis. Chase, M. W., Jr. (1998). NIST-JANAF Thermochemical Tables. Fourth edition. Journal of Physical and Chemical Reference Data, Monograph No. 9. ISBN 1-56396-831-2. Cheng, D. Y. and Nelson, A. L. C. (2002). The Chronological Development of the Cheng Cycle Steam Injected Gas Turbine During the Past 25 Years. In: Proceedings of ASME Turbo Expo 2002, Amsterdam, The Netherlands, Jun. 3-6, 2002. ASME Paper No. GT-2002-30119. Chiesa, P., Lozza, G., Macchi, E. and Consonni, S. (1995). An Assessment of the Thermodynamic Performance of Mixed Gas-Steam Cycles: Part B - Water-Injected and HAT Cycles. Journal of Engineering for Gas Turbines and Power, 117(3), 499-508. ISSN 0742-4795.

81 Advanced Power Cycles with Mixtures as the Working Fluid

Cohn, A. (1993). Power Plant Cycles Featuring Air Humidification. EPRI Journal, 18(Oct.-Nov.), 43-47. ISSN 0362-3416. Cohn, A., Ehrlich, S., Darguzas, J. and Giermac, E. (1999). Compressed Air Energy Storage with Humidification: An Economic Analysis. In: Proceedings of the International Gas Turbine and Aeroengine Congress and Exhibition, Indianapolis, Indiana, USA, Jun. 7-10, 1999. ASME Paper No. 99-GT-328. Cordeiro, V. and Jensen, J. B. (1996). Two-Stroke Diesels Meet Macau Electric Power Needs. Diesel and Gas Turbine Worldwide, 28(6), 22-26. ISSN 0278-5994. Dalili, F. (2003). Humidification in Evaporative Power Cycles. Stockholm, Sweden: Dept. of Chemical Engineering and Technology, Div. of Energy Processes, Royal Institute of Technology. Doctoral Thesis. ISRN KTH/KET/R--171--SE. Dalili, F., Andrén, M., Yan, J. and Westermark, M. (2001). The Impact of Thermodynamic Properties of Air-Water Vapor Mixtures on Design of Evaporative Gas Turbine Cycles. In: Proceedings of ASME Turbo Expo 2001, New Orleans, Louisiana, USA, Jun. 4-7, 2001. ASME Paper No. 2001-GT-0098. Dalili, F. and Westermark, M. (2001). First Experimental Results on Humidification of Pressurized Air in Evaporative Power Cycles. In: Proceedings of the 36th Intersociety Energy Conversion Engineering Conference, Savannah, Georgia, USA, Jul. 29 - Aug. 2, 2001. Paper No. 2000-CT-05. ISSN 0146-955X. Dalili, F. and Westermark, M. (2002). Experimental Study on a Packed Bed Humidifier in an Evaporative Gas Turbine. In: Proceedings of the 2002 International Joint Power Generation Conference, Phoenix, Arizona, USA, Jun. 24-26, 2002. ASME Paper No. IJPGC2002-26106. Dalili, F. and Westermark, M. (2003). Experimental Results on Humidification of Compressed Air in a Tubular Humidifier for Evaporative Cycles. Accepted for publication in: Proceedings of ASME Turbo Expo 2003, Atlanta, Georgia, USA, Jun. 16-19, 2003. ASME Paper No. GT2003-38034. Day, W. H. and Rao, A. D. (1993). Redefined Natural Gas HAT Cycle Produced Higher Output. Modern Power Systems, 13(6), 26-28. ISSN 0260-7840. Day, W. H., Kendrick, D., Knight, B., Bhargava, A., Sowa, W., Colket, M., Casleton, K. and Maloney, D. (1999). HAT Cycle Technology Development Program. In: Advanced Turbine Systems Annual Program Review Meeting, Nov. 8-10, 1999. http://www.netl.doe.gov/publications/proceedings/99/99ats/2-8.pdf. Access date Dec. 4, 2002. Dejfors, C., Thorin, E. and Svedberg, G. (1998). Ammonia-Water Power Cycles for Direct-Fired Cogeneration Applications. Energy Conversion and Management, 39(16-18), 1675-1681. ISSN 0196-8904. Desideri, U. and Bidini, G. (1996). Study of Possible Optimisation Criteria for Geothermal Power Plants. In: Proceedings of ECOS’96, Stockholm, Sweden, Jun. 25-27, 1996. 313-319. ISBN 91-7170-664-X.

82 6 References

Eastop, T. D. and McConkey, A. (1993). Applied Thermodynamics for Engineering Technologists. Fifth edition. Harlow, UK: Longman Scientific and Technical, Longman Group Limited. 282. ISBN 0-582-09193-4. El-Masri, M. A. (1988a). A Modified, High-Efficiency Recuperated Gas Turbine Cycle. Journal of Engineering for Gas Turbines and Power, 110(2), 233-242. ISSN 0742-4795. El-Masri, M. A. (1988b). A Flexible, Efficient Gas-Turbine Cogeneration Cycle with a Novel Dual-Mode Heat Recovery System. In: Proceedings of 1988 ASME Cogen-Turbo, Montreux, Switzerland, Aug. 30 – Sep. 1, 1988. IGTI-Vol. 3, 229-237. El-Sayed, Y. M. and Tribus, M. (1985a). A Theoretical Comparison of the Rankine and Kalina Cycles. ASME Advanced Energy Systems Division, AES-Vol. 1, 97-102. New York, USA: ASME. El-Sayed, Y. M. and Tribus, M. (1985b). Thermodynamic Properties of Water-Ammonia Mixtures - Theoretical Implementation for Use in Power Cycle Analysis. ASME Advanced Energy Systems Division, AES-Vol. 1, 89-95. New York, USA: ASME. Enick, R. M., Klara, S. M. and Marano, J. J. (1995). A Robust Algorithm for High- Pressure Gas Humidification. Computers and Chemical Engineering, 19(10), 1051-1061. ISSN 0098-1354. Enick, R. M., McIlvried, H. G., Gale, T. K. and Klara, J. M. (1997). The Modeling of LEBS-Kalina Power Cycles. In: Proceedings of the 1997 International Joint Power Generation Conference, Denver, Colorado, USA, Nov. 2-5, 1997. PWR-Vol. 32, 55-67. ISBN 0-7918-1578-1. EPRI (1986a). Technology Assessments of Advanced Power Generation Systems II - Kalina Bottoming Cycle. Palo Alto, USA: Electric Power Research Institute. EPRI AP-4681, Project 2528-4, Final Report. EPRI (1986b). State-of-the-Art Survey of Diesel Bottoming Cycles. Palo Alto, USA: Electric Power Research Institute. EPRI AP-4420, Project 2528-2, Final Report. EPRI (1991). A Comparison of Humid Air Turbine (HAT) Cycle and Combined-Cycle Power Plants. Palo Alto, USA: Electric Power Research Institute. EPRI IE-7300, Project 2999-7, Final Report. EPRI (1993). A Feasibility and Assessment Study for FT 4000 Humid Air Turbine (HAT). Palo Alto, USA: Electric Power Research Institute. EPRI TR-102156, Project 3251-05, Final Report. Exergy (2001). The Kalina Cycle Bottoming Cycle for Diesel Combined Cycle. http://www.exrg.com/papers/ExergyDieselCC.pdf. Access date Dec. 19, 2002. Exergy (2003). http://www.exrg.com/. Access date Feb. 20, 2003. Fischer, A. C., Frutschi, H. U. and Haselbacher, H. (2001). Augmentation of Gas Turbine Power Output by Steam Injection. In: Proceedings of ASME Turbo Expo 2001, New Orleans, Louisiana, USA, Jun. 4-7, 2001. ASME Paper No. 2001-GT-0107.

83 Advanced Power Cycles with Mixtures as the Working Fluid

Foster-Pegg, R. W. (1989). Turbo-S.T.I.G. - The Turbocharged Steam Injected Gas Turbine Cycle. In: Proceedings of the International Gas Turbine and Aeroengine Congress and Exposition, Toronto, Ontario, Canada, Jun. 4-8, 1989. ASME Paper No. 89-GT-100. Frutschi, H. U. and Plancherel, A. (1988). Comparison of Combined Cycles with Steam Injection and Evaporisation Cycles. In: Proceedings of 1988 ASME Cogen- Turbo, Montreux, Switzerland, Aug. 30 – Sep. 1, 1988. IGTI-Vol. 3, 137-145. Frutschi, H. U. and Wettstein, H. E. (1991). A Thermodynamic Comparison of Steam Injection Gas Turbines and Combined-Cycle Plants for Cogeneration Applications. In: Proceedings of 1991 ASME Cogen-Turbo, Budapest, Hungary, Sep. 3-5, 1991. IGTI-Vol. 6, 25-30. ISBN 0-7918-0650-2. Furuya, T. and Mori, Y. (2001a). An Application of Kalina Combined Cycle to Refuse Incineration Power Generation (1st Report). Memoirs of the Faculty of Engineering, Kyushu University, 61(2), 41-59. ISSN 1345-868X. Furuya, T. and Mori, Y. (2001b). An Application of Kalina Combined Cycle to Refuse Incineration Power Generation (2nd Report). Memoirs of the Faculty of Engineering, Kyushu University, 61(3), 85-100. ISSN 1345-868X. Gajewski, W., Lezuo, A., Nürnberg, R., Rukes, B. and Vesper, H. (1989). Der Kalina–Prozess – Thermodynamische Grundlagen und Realiserbarkeit. VGB Kraftwerkstechnik, 69(5), 477-483. ISSN 0372-5715. In German. Gallo, W. L. R., Bidini, G., Bettagli, N. and Facchini, B. (1995). The Evaporator Process Simulation and the HAT Cycle (Humid Air Turbine) Performance. In: Proceedings of 1995 ASME Cogen-Turbo Power, Vienna, Austria, Aug. 23-25, 1995. ASME Paper No. 95-CTP-59. Gasparovic, N. and Stapersma, D. (1973). Gas Turbines with Heat Exchanger and Water Injection in the Compressed Air. Combustion, 45(6), 6-16. ISSN 0010-2172. GateCycle, versions 5.3 and 5.41.0.r, GE Enter Software, 1490 Drew Avenue, Suite 180, Davis, CA 95616, USA. Gröndalen, O. (1987). Organiska Rankine Cykler (ORC) – ett sätt att generera elenergi. In: Värme- och kraftsymposium, ÅF+IVA 3 dagars energivecka, Nov. 20, 1987. In Swedish. GTW (1999). Gas Turbine World 1999-2000 Handbook, 20. ISSN 0747-7988. GTW (2001). Gas Turbine World 2001-2002 Handbook, 22. ISSN 0747-7988. Guthrie, K. M. (1969). Data and Techniques for Preliminary Capital Cost Estimating. Chemical Engineering, 76(6), 114-142. ISSN 0009-2460. Hewitt, G. F., Shires, G. L. and Bott, T. R. (1994). Process Heat Transfer. CRC Press. 220, 229, 381. ISBN 0-8493-9918-1. Hofstädter, A., Frutschi, H. U. and Haselbacher, H. (1998). Effects of Steam Reheat in Advanced Steam Injected Gas Turbine Cycles. In: Proceedings of the International Gas Turbine and Aeroengine Congress and Exhibition, Stockholm, Sweden, Jun. 2-5, 1998. ASME Paper No. 98-GT-584.

84 6 References

Horlock, J. H. (1998). The Evaporative Gas Turbine [EGT] Cycle. Journal of Engineering for Gas Turbines and Power, 120(2), 336-343. ISSN 0742-4795. Húsavík Energy (2003). http://www.oh.is/reports.asp?tID=7. Access date Feb. 19, 2003. Hyland, R. W. and Wexler, A. (1983). Formulations for the Thermodynamic Properties of Dry Air from 173.15 K to 473.15 K, and of Saturated Moist Air from 173.15 K to 372.15 K, at Pressures to 5 MPa. ASHRAE Transactions, 89(2A), 520-535. ISSN 0001-2505. IAPWS (1996). Release on the IAPWS Formulation 1995 for the Thermodynamic Properties of Ordinary Water Substance for General and Scientific Use. The International Association for the Properties of Water and Steam. http://www.iapws.org/. Access date Feb. 8, 2003. Ibrahim, O. M. and Klein, S. A. (1993). Thermodynamic Properties of Ammonia- Water Mixtures. ASHRAE Transactions, 99(1), 1495-1502. ISSN 0001-2505. IEA (2002). World Energy Outlook 2002. Paris, France: International Energy Agency. ISBN 92-64-19835-0. IPSEpro, version 3.0, SimTech Simulation Technology, Riesstrasse 120, A-8010 Graz, Austria. Isaksson, A., Maunsbach, K., Svedberg, G. and Lior, N. (1999). Integration of Power Generating Processes Using Waste Heat in a Pulp Mill. In: Proceedings of the Second Biennial Johan Gullichsen Colloquium, Helsinki, Finland, Sep. 9-10, 1999. Ji, X. and Yan, J. (2002a). Humidity, Enthalpy, and Entropy for the Humid Air from 298 to 573 K and up to 200 bar. In: Proceedings of the 16th European Conference on Thermophysical Properties, London, UK, Sep. 1-4, 2002. Ji, X. and Yan, J. (2002b). Saturated Thermodynamic Properties for the Air-Water System at Elevated Temperature and Pressure. Manuscript submitted for publication. Johnke, T. and Mast, M. (2002). Power Boosters - Technologies to Enhance Gas Turbine Power Output on Demand. Siemens Power Journal Online, May. http://www.siemenswestinghouse.com/download/pool/mast_engl_3.pdf. Access date Nov. 12, 2002. Kalina, A. I. (1983). Combined Cycle and Waste Heat Recovery Power Systems Based on a Novel Thermodynamic Energy Cycle Utilizing Low-Temperature Heat for Power Generation. In: Proceedings of the 1983 Joint Power Generation Conference, Indianapolis, Indiana, USA, 1983. ASME Paper No. 83-JPGC-GT-3. Kalina, A. I. (1984). Combined-Cycle System with Novel Bottoming Cycle. Journal of Engineering for Gas Turbines and Power, 106(4), 737-742. ISSN 0742-4795. Kalina, A. I. (1989). The Kalina Cycle Technology Applied to Direct-Fired Power Plants. In: Proceedings of the Joint ASME/IEEE Power Generation Conference, Dallas, Texas, USA, Oct. 22-26, 1989. ASME Paper No. 89-JPGC/Pwr-24.

85 Advanced Power Cycles with Mixtures as the Working Fluid

Kalina, A. I. (1991). Kalina Cycles and System for Direct-Fired Power Plants. ASME Advanced Energy Systems Division, AES-Vol. 25/HTD-Vol. 191, 41-47. New York, USA: ASME. ISBN 0-7918-0868-8. Kalina, A. I. (1993). Recent Improvements in Kalina Cycles: Rationale and Methodology. In: Proceedings of the American Power Conference, Chicago, Illinois, USA, Apr. 13-15, 1993. Vol. 55-I, 191-195. ISSN 0097-2126. Kalina, A. I. and Leibowitz, H. M. (1987). Applying Kalina Technology to a Bottoming Cycle for Utility Combined Cycles. In: Proceedings of the Gas Turbine Conference and Exhibition, Anaheim, California, USA, May 31 – Jun. 4, 1987. ASME Paper No. 87-GT-35. Kalina, A. I. and Leibowitz, H. M. (1988). The Design of a 3 MW Kalina Cycle Experimental Plant. In: Proceedings of the Gas Turbine and Aeroengine Congress, Amsterdam, The Netherlands, Jun. 6-9, 1988. ASME Paper No. 88-GT-140. Kalina, A. I. and Leibowitz, H. M. (1989). Application of the Kalina Cycle Technology to Geothermal Power Generation. In: Geothermal Resources Council Transactions, 1989. Vol. 13. 605-611. ISSN 0193-5933. Kalina, A. I., Leibowitz, H. M., Markus, D. W. and Pelletier, R. I. (1991). Further Technical Aspects and Economics of a Utility-Size Kalina Bottoming Cycle. In: Proceedings of the International Gas Turbine and Aeroengine Congress and Exposition, Orlando, Florida, USA, Jun. 3-6, 1991. ASME Paper No. 91-GT-365. Kalina, A. I. and Pelletier, R. I. (1996). Basic Principles and Structure of the Kalina Cycle Systems for Power Generation. In: Proceedings of the 1996 International Joint Power Generation Conference, Houston, Texas, USA, Oct. 13-17, 1996. Vol. 2. Kouremenos, D. A. and Rogdakis, E. D. (1990). The Temperature – Entropy (or Enthalpy) and the Enthalpy – Entropy (Mollier) Diagram of the Kalina Cycle. ASME Advanced Energy Systems Division, AES-Vol. 19, 13-19. New York, USA: ASME. ISBN 0-7918-0580-8. Krause, A., Tsatsaronis, G. and Sauthoff, M. (1998). On the Cost Optimization of a District Heating Facility Using a Steam-Injected Gas Turbine Cycle. In: Proceedings of ECOS'98, Nancy, France, Jul. 8-10, 1998. Vol. I, 283-290. ISBN 2-905-267-29-1. Lagerström, G. and Xie, M. (2002). High Performance and Cost Effective Recuperator for Micro-Gas Turbines. In: Proceedings of ASME Turbo Expo 2002, Amsterdam, The Netherlands, Jun. 3-6, 2002. ASME Paper No. GT-2002-30402. Larson, E. D. and Williams, R. H. (1987). Steam-Injected Gas Turbines. Journal of Engineering for Gas Turbines and Power, 109(1), 55-63. ISSN 0742-4795. Layne, A., Williams, M., Samuelsen, S. and Hoffman, P. (2000). Hybrid Heat Engines: The Power Generation Systems of the Future. In: Proceedings of ASME Turbo Expo 2000, Munich, Germany, May 8-11, 2000. ASME Paper No. 2000-GT-0549.

86 6 References

Lazzeri, L. (1997). Application of Kalina Cycle as Bottoming Cycle for Existing Geothermal Plants. In: Proceedings of the Florence World Energy Research Symposium, Florence, Italy, Jul. 30 - Aug. 1, 1997. 389-396. ISBN 88-86281-21-8. Lazzeri, L. and Bruzzone, M. (1995). Geothermal Plant Efficiency Enhancement by Means of the Use of Kalina Cycle. In: Proceedings of the 30th Intersociety Energy Conversion Engineering Conference, Orlando, Florida, USA, Jul. 30 – Aug. 4, 1995. Vol. 2, 453-457. ISSN 0146-955X. Lazzeri, L. and Diotti, F. (1998). Experiences in Geothermal Production via New Products Development and Effective Production Delocalization. In: Proceedings of Power-Gen Europe ’98, Milan, Italy, Jun. 9-11, 1998. Vol. IV, 123-132. Leibowitz, H. M. (1993). Operating Experiences on the 3 MW Kalina Cycle Demonstration Plant. In: Proceedings of the American Power Conference, Chicago, Illinois, USA, Apr. 13-15, 1993. Vol. 55-I, 173-178. ISSN 0097-2126. Leibowitz, H. M. and Markus, D. W. (1990). Economic Performance of Geothermal Power Plants Using the Kalina Cycle Technology. In: Geothermal Resources Council Transactions, Proceedings of the 1990 International Symposium on Geothermal Energy, Kailua-Kona, Hawaii, USA, Aug. 20-24, 1990. Vol. 14, Part II, 1037-1042. ISSN 0193-5933. Leibowitz, H. and Mirolli, M. (1997). First Kalina Combined-Cycle Plant Tested Successfully. Power Engineering, 101(5). ISSN 0032-5961. Leibowitz, H. M. and Mlcak, H. A. (1999). Design of a 2 MW Kalina Cycle Binary Module for Installation in Husavik, Iceland. In: Geothermal Resources Council Transactions, Oct. 17-20, 1999. Vol. 23, 75-80. ISBN 0-934412-83-9. Lindquist, T. (2002). Evaluation, Experience and Potential of Gas Turbine Based Cycles with Humidification. Lund, Sweden: Dept. of Heat and Power Engineering, Div. of Thermal Power Engineering, Lund Institute of Technology, Lund University. Doctoral Thesis. ISRN LUTMDN/TMHP--02/1005--SE. Lindquist, T. O., Rosén, P. M. and Torisson, T. (2000a). Theoretical and Experimental Evaluation of the EvGT-Process. In: Proceedings of ASME Advanced Energy Systems Division – 2000. AES-Vol. 40, 457-469. ISBN 0-7918-1907-8. Lindquist, T. O., Rosén, P. M. and Torisson, T. (2000b). Evaporative Gas Turbine Cycle – A Description of a Pilot Plant and Operating Experience. In: Proceedings of the ASME Advanced Energy Systems Division – 2000. AES-Vol. 40, 511-520. ISBN 0-7918-1907-8. Lindquist, T., Thern, M. and Torisson, T. (2002). Experimental and Theoretical Results of a Humidification Tower in an Evaporative Gas Turbine Cycle Pilot Plant. In: Proceedings of ASME Turbo Expo 2002, Amsterdam, The Netherlands, Jun. 3-6, 2002. ASME Paper No. GT-2002-30127. Luckey, J. (2000). Booming Market. Energy Markets. 5(10), 66, 68, 70, 72, 74.

87 Advanced Power Cycles with Mixtures as the Working Fluid

Lupandin, V. V., Romanov, V. I., Krivutsa, V. A. and Lupandin, V. V. (2001). Design, Development and Testing of a Gas Turbine Steam Injection and Water Recovery System. In: Proceedings of ASME Turbo Expo 2001, New Orleans, Louisiana, USA, Jun. 4-7, 2001. ASME Paper No. 2001-GT-0111. Macchi, E. and Poggio, A. (1994). A Cogeneration Plant Based on a Steam Injection Gas Turbine with Recovery of the Water Injected: Design Criteria and Initial Operating Experience. In: Proceedings of the International Gas Turbine and Aeroengine Congress and Exposition, The Hague, The Netherlands, Jun. 13-16, 1994. ASME Paper No. 94-GT-17. Macchi, E., Consonni, S., Lozza, G. and Chiesa, P. (1995). An Assessment of the Thermodynamic Performance of Mixed Gas-Steam Cycles: Part A - Intercooled and Steam-Injected Cycles. Journal of Engineering for Gas Turbines and Power, 117(3), 489-498. ISSN 0742-4795. MAN B&W (2003). http://www.manbw.com/web/viewers/news/SubjectViewer. aspx?id=83. Access date Feb. 20, 2003. Marston, C. H. and Hyre, M. (1995). Gas Turbine Bottoming Cycles: Triple- Pressure Steam versus Kalina. Journal of Engineering for Gas Turbines and Power, 117(1), 10-15. ISSN 0742-4795. Martinka, M. (1940). Combustion Engine and a Method for the Operation thereof. US Patent No. 2,186,706. Maunu, J. (1999). Wärtsilä Finland Oy, P.O. Box 252, FIN-65101 Vaasa, Finland. Personal communication. McCarthy, S. J. and Scott, I. (2002). The WR-21 Intercooled Recuperated Gas Turbine Engine – Operation and Integration into the Royal Navy Type 45 Destroyer System. In: Proceedings of ASME Turbo Expo 2002, Amsterdam, The Netherlands, Jun. 3-6, 2002. ASME Paper No. GT-2002-30266. McDonald, C. F. and Wilson, D. G. (1996). The Utilization of Recuperated and Regenerated Engine Cycles for High-Efficiency Gas Turbines in the 21st Century. Applied Thermal Engineering, 16(8-9), 635-653. ISSN 1359-4311. McIlvried, H. G., Ramezan, M., Enick, R. M. and Venkatasubramanian, S. (1998). Exergy and Pinch Analysis of an Advanced Ammonia-Water Coal-Fired Power Cycle. In: Proceedings of the ASME Advanced Energy Systems Division. AES-Vol. 38, 197-203. ISSN 1071-6947. McNeely, M. (2001). Engine Order Survey - Engine Orders Go Through the Roof. Diesel and Gas Turbine Worldwide, October. http://www.dieselpub.com/ww/ww_ power2001.htm. Access date Feb. 2, 2002. Melin, P. and Simonsson, N. (2001). Evaporative Gas Turbine Application Study. Stockholm, Sweden: Vattenfall Utveckling AB. Master Thesis.

88 6 References

Mirolli, M., Hjartarson, H., Mlcak, H. A. and Ralph, M. (2002). Testing and Operating Experience of the 2 MW Kalina Cycle Geothermal Power Plant in Husavik, Iceland. In: Proceedings of the World Renewable Energy Congress VII, Cologne, Germany, Jun. 29 - Jul. 5, 2002. Moran, M. J. (1989). Availability Analysis: A Guide to Efficient Energy Use. Corrected edition. New York, USA: ASME Press. ISBN 0-7918-0009-1. Mori, Y., Nakamura, H., Takahashi, T. and Yamamoto, K. (1983). A Highly Efficient Regenerative Gas Turbine System by New Method of Heat Recovery with Water Injection. In: Proceedings of the 1983 Tokyo International Gas Turbine Congress, Tokyo, Japan, 1983. 297-303. Paper No. 83-Tokyo-IGTC-38. Nakamura, H., Takahashi, T., Yamamoto, K. and Sayama, N. (1985). Regenerative Gas Turbine Cycle. US Patent No. 4,537,023. Nakamura, H., Takahashi, T., Narazaki, N., Yamamoto, K. and Sayama, N. (1987). Regenerative Gas Turbine Cycle. US Patent No. 4,653,268. Nakhamkin, M., Swensen, E., Wilson, L. M., Gaul, G. and Polsky, M. (1996). The Cascaded Humidified Advanced Turbine (CHAT). Journal of Engineering for Gas Turbines and Power, 118(3), 565-571. ISSN 0742-4795. Nakhamkin, M., Swensen, E. C., Scheibel, J. R. and Cohn, A. (1998). CHAT Technology: An Alternative Approach to Achieve Advanced Turbine Systems Efficiencies with Present Combustion Turbine Technology. In: Proceedings of the International Gas Turbine and Aeroengine Congress and Exhibition, Stockholm, Sweden, Jun. 2-5, 1998. ASME Paper No. 98-GT-143. Nguyen, H. B. and den Otter, A. (1994). Development of Gas Turbine Steam Injection Water Recovery (SIWR) System. Journal of Engineering for Gas Turbines and Power, 116(1), 68-74. ISSN 0742-4795. Niemi, S. (1997). Survey of Modern Power Plants Driven by Diesel and Gas Engines. Espoo, Finland: Technical Research Centre of Finland. VTT Research Notes 1860. ISSN 1235-0605. ISBN 951-38-5155-9. Nilsson, K.-J. (2002). ALSTOM Power Sweden, SE-612 82, Finspong, Sweden. Tel. +46 [0] 122 810 00. Personal communication. Nilsson, P.-A. (editor) (1996). EvGT – Evaporativ gasturbin, Processanpassning 70-80 MWe anläggning. Lund, Sweden: Dept. of Heat and Power Engineering, Div. of Thermal Power Engineering, Lund Institute of Technology, Lund University. Technical report. Slutrapport, Block 3, Etapp 1. In Swedish. Nilsson, U., Eidensten, L. and Tollin, J. (2001). Applications for Evaporative Gas Turbines. In: Proceedings of Power-Gen Europe 2001, Brussels, Belgium, May 29-31, 2001. NIST (2001). http://webbook.nist.gov/chemictry/form-ser.html. Access date Sep. 4, 2001. Oelkrug, T. (1999). Design of Packed Towers for Evaporative Gas Turbine Air Humidifica- tion. Stockholm, Sweden: Dept. of Chemical Engineering and Technology, Div. of Energy Processes, Royal Institute of Technology. Student project.

89 Advanced Power Cycles with Mixtures as the Working Fluid

O’Keefe, W. (1995). Engine/Generators Reconfigured to Compete in the Next Century. Power, 139(10), 52, 54-56, 58, 61-62. ISSN 0032-5929. Olsson, E., Åbom, J. and Andersson, K.-G. (1998). Development of a Wood Powder Burning Diesel Engine. In: Proceedings of the 22nd CIMAC International Congress on Combustion Engines, Copenhagen, Denmark, May 18-21, 1998. Vol. 4, 1085-1091. Olsson, E., Desideri, U., Stecco, S. S. and Svedberg, G. (1991). An Integrated Gas Turbine-Kalina Cycle for Cogeneration. In: Proceedings of the International Gas Turbine and Aeroengine Congress and Exposition, Orlando, Florida, USA, Jun. 3-6, 1991. ASME Paper No. 91-GT-202. Olsson, E. K., Thorin, E. B., Dejfors, C. A. S. and Svedberg, G. (1994). Kalina Cycles for Power Generation from Industrial Waste Heat. In: Proceedings of the Florence World Energy Research Symposium, Florence, Italy, Jul. 6-8, 1994. 39-49. ISBN 88-86281-04-8. Olsson, F. (1999). Cogeneration Based on Gasified Biomass – A Comparison of Concepts. Lund, Sweden: Dept. of Heat and Power Engineering, Div. of Thermal Power Engineering, Lund Institute of Technology, Lund University. Thesis for Degree of Licentiate of Engineering. ISRN LUTMDN/TMVK--7033--SE. Osenga, M. (1984). Slow-Speed Diesel Installed for U.S. Power Generation. Diesel and Gas Turbine Worldwide, 16(2), 40-41. ISSN 0278-5994. De Paepe, M. and Dick, E. (1999). Water Recovery in Steam Injected Gas Turbines: A Technological and Economical Analysis. European Journal of Mechanical and Environmental Engineering, 44(4), 195-204. ISSN 0035-3612. De Paepe, M. and Dick, E. (2000). Cycle Improvements to Steam Injected Gas Turbines. International Journal of Energy Research, 24(12), 1081-1107. ISSN 0363-907X. Park, Y. M. (1988). A Generalized Equation of State Approach to the Thermodynamic Properties of Ammonia-Water Mixtures with Applications. Ann Arbor, USA: The University of Michigan. Doctoral Thesis. Park, Y. M. and Sonntag, R. E. (1990). A Preliminary Study of the Kalina Power Cycle in Connection with a Combined Cycle System. International Journal of Energy Research, 14(2), 153-162. ISSN 0363-907X. Rao, A. D. (1989). Process for Producing Power. US Patent No. 4,829,763. Rao, A. D. and Day, W. H. (1996). Mitigation of Greenhouse Gases from Gas Turbine Power Plants. Energy Conversion and Management, 37(6-8), 909-914. ISSN 0196-8904. Regan, J. W., von Hein, R. J., Peletz, L. J., Jr., Wesnor, J. D. and Bender, D. J. (1996). ABB’s LEBS Activities – A Status Report. Windsor, USA: ABB Combustion Engineering. Technical Report.

90 6 References

Regan, J. W., von Hein, R. J. and Wesnor, J. D. (1997). ABB’s LEBS Technologies: Practical Solutions for Controlling Air Emissions and Increasing Efficiency. In: Proceedings of the International Technical Conference on Coal Utilization and Fuel Systems, 1997. Vol. 22, 173-181. Reid, R. C., Prausnitz, J. M. and Poling, B. E. (1987). The Properties of Gases and . Fourth edition. New York, USA: McGraw-Hill. 667. ISBN 0-07-051799-1. Rensfelt, E. (1987). Förgasning och större dieselmotorer. VVS och energi, No. 6, 26-28, 30, 32. ISSN 0280-9524. In Swedish. Rice, I. G. (1993a). Steam-Injected Gas Turbine Analysis: Part I - Steam Rates. In: Proceedings of the International Gas Turbine and Aeroengine Congress and Exposition, Cincinnati, Ohio, USA, May 24-27, 1993. ASME Paper No. 93-GT-132. Rice, I. G. (1993b). Steam-Injected Gas Turbine Analysis: Part II - Steam-Cycle Efficiency. In: Proceedings of the International Gas Turbine and Aeroengine Congress and Exposition, Cincinnati, Ohio, USA, May 24-27, 1993. ASME Paper No. 93-GT-420. Rice, I. G. (1993c). Steam-Injected Gas Turbine Analysis: Part III - Steam- Regenerated Heat. In: Proceedings of the International Gas Turbine and Aeroengine Congress and Exposition, Cincinnati, Ohio, USA, May 24-27, 1993. ASME Paper No. 93-GT-421. Richter, H. J. and Lobachyov, K. V. (1996). Addition of Highly Efficient Bottoming Cycles for the Santa Clara Molten Carbonate Fuel Cell Power Plant. In: Proceedings of ECOS’96, Sweden, Jun. 25-27, 1996. 599-606. ISBN 91-7170-664-X. Robson, F. L. and Seery, D. J. (1998). Advanced Aeroderivative Gas Turbines in Coal-Based High Performance Power Systems (HIPPS). In: Proceedings of the International Gas Turbine and Aeroengine Congress and Exhibition, Stockholm, Sweden, Jun. 2-5, 1998. ASME Paper No. 98-GT-131. Rosén, P. M. (2000). Evaporative Cycles – In Theory and in Practise. Lund, Sweden: Dept. of Heat and Power Engineering, Div. of Thermal Power Engineering, Lund Institute of Technology, Lund University. Doctoral Thesis. ISRN LUTMDN/TMVK--1020--SE. Rukes, B. and Taud, R. (2002). Status and Perspectives of Fossil Power Generation. In: Proceedings of ECOS 2002, Berlin, Germany, Jul. 3-5, 2002. Vol. I, 23-43. ISBN 3-00-009533-0. de Ruyck, J., Allard, G. and Maniatis, K. (1996). An Externally Fired Evaporative Gas Turbine Cycle for Small Scale Biomass Gasification. In: Proceedings of the 9th European Bioenergy Conference, Copenhagen, Denmark, Jun. 24-27, 1996. Vol. 1, 260-265. ISBN 0-08-042849-5. Rydstrand, M. C., Westermark, M. O. and Bartlett, M. A. (2002). An Analysis of the Efficiency and Economy of Humidified Gas Turbines in District Heating Applications. In: Proceedings of ECOS 2002, Berlin, Germany, Jul. 3-5, 2002. Vol. II, 695-703. ISBN 3-00-009533-0.

91 Advanced Power Cycles with Mixtures as the Working Fluid

Saidi, A., Sundén, B. and Eriksson, D. (2000). Intercoolers in Gas Turbine Systems and Combi-Processes for Production of Electricity. In: Proceedings of ASME Turbo Expo 2000, Munich, Germany, May 8-11, 2000. ASME Paper No. 2000-GT-0234.

Schorr, M. M. (1992). NOx Emission Control for Gas Turbines: A 1992 Update on Regulations and Technology. In: Proceedings of 1992 ASME Cogen-Turbo, Houston, Texas, USA, Sep. 1-3, 1992. IGTI-Vol. 7, 1-12. ISBN 0-7918-0670-7. SEA (2002). Energiläget 2002. Eskilstuna, Sweden: Swedish Energy Agency (Statens energimyndighet). In Swedish. Simonsson, N., Anheden, M., Eidensten, L. and Tollin, J. (2002). Evaporative Gas Turbines - Humidified Air Gives Flexible Power for the Future. In: Proceedings of Power-Gen Europe 2002, Milan, Italy, Jun. 11-13, 2002. Smith, R. W., Ranasinghe, J., Stats, D. and Dykas, S. (1996). Kalina Combined Cycle Performance and Operability. In: Proceedings of the 1996 International Joint Power Generation Conference, Houston, Texas, USA, Oct. 13-17, 1996. PWR-Vol. 30, 701-728. Sorenson, S. C. (2000). Dimethyl Ether in Diesel Engines – Progress and Perspectives. In: Proceedings of the 2000 Spring Technical Conference of the ASME Internal Combustion Engine Division, San Antonio, Texas, USA, Apr. 9-12, 2000. ICE-Vol. 34-3, 65-74. ASME Paper No. 2000-ICE-292. ISBN 0-7918-1664-8. Stambler, I. (2002). DOE Shifting from Design Studies to Technology Oriented Programs. Gas Turbine World, 32(2), 17-18, 20-21, 23. ISSN 0747-7988. SteamTab, ChemicaLogic Corporation, 8 Cedar Street, Suite 56, Woburn, Massachusetts 01801-6362, USA. Stecco, S. S. and Desideri, U. (1989). A Thermodynamic Analysis of the Kalina Cycles: Comparisons, Problems and Perspectives. In: Proceedings of the International Gas Turbine and Aeroengine Congress and Exposition, Toronto, Ontario, Canada, Jun. 4-8, 1989. ASME Paper No. 89-GT-149. Stenhede, T. (1998a). Lean Burn Gas Engines in Cogeneration Applications. Wärtsilä NSD Corporation. Brochure. Stenhede, T. (1998b). Wärtsilä Sweden AB, P.O. Box 920, SE-461 29 Trollhättan, Sweden. Personal communication. Sumimoto Metals (2003). http://www.sumitomometals.co.jp/e/business/ strategies-for-preserving-the-environment.html. Access date Feb. 20, 2003. Szargut, J. (2002). Cogeneration of Network Heat in the Set of a Humid Air Turbine. Energy, 27(1), 1-15. ISSN 0360-5442. Takeshita, K., Amano, Y., Hashizume, T., Usui, A. and Tanzawa, Y. (2002). Experimental Results of an Ammonia-Water Mixture Turbine System (Part 2: Effect of the Ammonia Mass Fraction). In: Proceedings of the 2002 International Joint Power Generation Conference, Phoenix, Arizona, USA, Jun. 24-26, 2002. ASME Paper No. IJPGC2002-26064.

92 6 References

Thorin, E. (2000). Comparison of Correlations for Predicting Thermodynamic Properties of Ammonia-Water Mixtures. International Journal of Thermophysics, 21(4), 853-870. ISSN 0195-928X. Thorin, E. (2001). Thermophysical Properties of Ammonia-Water Mixtures for Prediction of Heat Transfer Areas in Power Cycles. International Journal of Thermophysics, 22(1), 201-214. ISSN 0195-928X. Thorin, E., Dejfors, C. and Svedberg, G. (1998). Thermodynamic Properties of Ammonia-Water Mixtures for Power Cycles. International Journal of Thermophysics, 19(2), 501-510. ISSN 0195-928X. Tillner-Roth, R. and Friend, D. (1998). A Formulation of the Thermodynamic Properties of the Mixture {Water + Ammonia}. Journal of Physical and Chemical Reference Data, 27(1), 63-96. ISSN 0047-2689. Tuzson, J. (1992). Status of Steam-Injected Gas Turbines. Journal of Engineering for Gas Turbines and Power, 114(4), 682-686. ISSN 0742-4795. Uji, S. (1999). IM400 IHI-FLECS Flexible Electric Cogeneration System. In: Proceedings of the 1999 International Joint Power Generation Conference, Burlingame, California, USA, Jul. 25-28, 1999. PWR-Vol. 34, 337-341. ISBN 0-7918-1615-X. Ulrich, G. D. (1984). A Guide to Chemical Engineering Process Design and Economics. John Wiley and Sons. 288, 301, 310. ISBN 0-471-08276-7. Utriainen, E. (2001). Investigation of Some Heat Transfer Surfaces for Gas Turbine Recuperators. Lund, Sweden: Dept. of Heat and Power Engineering, Div. of Heat Transfer, Lund Institute of Technology, Lund University. Doctoral Thesis. ISRN LUTMDN/TMHT--1005--SE. Valdimarsson, P. (2002). Cogeneration of District Heating Water and Electricity from a Low Temperature Geothermal Source. In: Proceedings of the 8th International Symposium on District Heating and Cooling, Trondheim, Norway, Aug. 14-16, 2002. Valenti, M. (2002). Reaching for 60 Percent. Mechanical Engineering, 124(4), 35-39. ISSN 0025-6501. Verneau, A. (1984). Recovery from Exhaust Gas on a Diesel Engine. In: Proceedings of the International VDI-Seminar, ORC-HP-Technology: Working Fluid Problems, Zurich, Switzerland, Sep. 10-12, 1984. VDI Berichte. 501-513. ISSN 0083-5560. ISBN 3-18-090539-5. Wagner, W. and Kruse, A. (1998). Properties of Water and Steam: The Industrial Standard IAPWS-IF97 for the Thermodynamic Properties and Supplementary Equations for Other Properties. Springer-Verlag. ISBN 3-540-64339-7. Wahlberg, P.-E. (2001). Design and Comparison between a Finned Tube Humidifier and a Packed Bed Humidifier Concerning Performance and Costs. Stockholm, Sweden: Dept. of Chemical Engineering and Technology, Div. of Energy Processes, Royal Institute of Technology. Master Thesis.

93 Advanced Power Cycles with Mixtures as the Working Fluid

Wei, A. (2002). Technologies for Next Generation Turbine Systems. In: Proceedings of Turbine Power Systems Conference and Condition Monitoring Workshop, Galveston, Texas, USA, Feb. 25-27, 2002. http://www.netl.doe.gov/publications/proceedings/02/ turbines/wei.pdf. Access date Nov. 7, 2002. Westermark, M. (1998). Method and Device for Generation of Mechanical Work and, if Desired, Heat in an Evaporative Gas Turbine Process. International Patent Application, Publication No. WO9801658. Westermark, M. (2001). Process Water Supply for Wet Gas Turbine Cycles - Condensate Recovery versus Deionisation of Surface Water. Stockholm, Sweden: Dept. of Chemical Engineering and Technology, Div. of Energy Processes, Royal Institute of Technology. Technical Report. Wilson, D. G. (1984). The Design of High-Efficiency Turbomachinery and Gas Turbines. The MIT Press. ISBN 0-262-23114-X. Winell, F. and Svedberg, G. (1997). Biobaserade bränslen för förbränningsmotorer - en jämförande litteraturstudie angående teknik, ekonomi, miljö och säkerhet. Stockholm, Sweden: Dept. of Chemical Engineering and Technology, Div. of Energy Processes, Royal Institute of Technology. Technical Report. TRITA-KET R71. In Swedish. Wolk, R. H. and Cohn, A. (1993). New Power Plant Cycles Utilizing Air Humidification and Turbomachinery. In: Proceedings of the American Power Conference, Chicago, Illinois, USA, Apr. 13-15, 1993. Vol. 55-I, 568-572. ISSN 0097-2126. Wärtsilä (2003). http://www.wartsila.com/english/index.jsp?cid=pp_ref_WPPP&. Access date Feb. 20, 2003. Yan, J., Eidensten, L. and Svedberg, G. (1995). An Investigation of the Heat Recovery System in Externally Fired Evaporative Gas Turbines. In: Proceedings of the International Gas Turbine and Aeroengine Congress and Exposition, Houston, Texas, USA, Jun. 5-8, 1995. ASME Paper No. 95-GT-72. Yan, J., Eidensten, L. and Svedberg, G. (1996). Externally Fired Evaporative Gas Turbine with a Condensing Heat Exchanger. In: Proceedings of the International Gas Turbine and Aeroengine Congress and Exhibition, Birmingham, UK, Jun. 10-13, 1996. ASME Paper No. 96-GT-77. Yan, J., Ji, X. and Jonsson, M. (2003). Thermodynamic Property Models for the Simulation of Advanced Wet Cycles. Accepted for publication in: Proceedings of ASME Turbo Expo 2003, Atlanta, Georgia, USA, Jun. 16-19, 2003. ASME Paper No. GT2003-38298. Zervos, N. G. (1992). Preliminary Design and Economics of a 100 MWe Coal- Fired Kalina Cycle Power Plant. In: Proceedings of the 1992 International Joint Power Generation Conference, Atlanta, Georgia, USA, Oct. 18-22, 1992. ASME Paper No. 92-JPGC-FACT-20. Zervos, N. G. (1993). Updated Design and Economics of the Kalina Cycle for Solid Fuel Applications. In: Proceedings of the American Power Conference, Chicago, Illinois, USA, Apr. 13-15, 1993. Vol. 55-I, 179-184. ISSN 0097-2126.

94 6 References

Ziegler, B. and Trepp, C. (1984). Equation of State for Ammonia-Water Mixtures. International Journal of Refrigeration, 7(2), 101-106. ISSN 0140-7007. Ågren, N. (2000). Advanced Gas Turbine Cycles with Water-Air Mixtures as Working Fluid. Stockholm, Sweden: Dept. of Chemical Engineering and Technology, Div. of Energy Processes, Royal Institute of Technology. Doctoral Thesis. ISRN KTH/KET/R--120--SE. Ågren, N. D. and Westermark, M. O. (2001a). Design Study of Part Flow Evaporative Gas Turbine Cycle: Performance and Equipment Sizing, Part 1: Aeroderivative Core. In: Proceedings of ASME Turbo Expo 2001, New Orleans, Louisiana, USA, Jun. 4-7, 2001. ASME Paper No. 2001-GT-0112. Ågren, N. D. and Westermark, M. O. (2001b). Design Study of Part Flow Evaporative Gas Turbine Cycle: Performance and Equipment Sizing, Part 2: Industrial Core. In: Proceedings of ASME Turbo Expo 2001, New Orleans, Louisiana, USA, Jun. 4-7, 2001. ASME Paper No. 2001-GT-0113. Ågren, N., Westermark, M., Bartlett, M. and Lindquist, T. (2002). First Experiments on an Evaporative Gas Turbine Pilot Power Plant: Water Circuit Chemistry and Humidification Evaluation. Journal of Engineering for Gas Turbines and Power, 124(1), 96-102. ISSN 0742-4795.

95

7 Nomenclature

Abbreviations a bottoming cycle configuration using only the exhaust gas from the reciprocating engine as its heat source AC aftercooler b bottoming cycle configuration using, if possible, the exhaust gas, charge air, jacket water and lubricating oil streams from the reciprocating engine as its heat sources cI ammonia-water cycle configuration with one pressure level in the HRVG, described in Section 3.1.2 and Papers I-IV (Kalina, 1983) cII ammonia-water cycle configuration with one pressure level in the HRVG and more internal heat recovery than configuration I, described in Section 3.3.1.2 and Papers I-IV (El-Sayed and Tribus, 1985a) cIII ammonia-water cycle configuration with one pressure level in the HRVG and two separators, described in Section 3.3.1.2 and Papers I- IV (Kalina, 1983) cIV ammonia-water cycle configuration with reheat, described in Section 4.3.2.1 and Paper V (where it is called configuration III) (Kalina and Leibowitz, 1987) CAES compressed air energy storage CASH compressed air storage with humidification CASHING compressed air storage with humidification integrated with natural gas CHAT cascaded humidified advanced turbine cnd hp high-pressure condenser cnd lp low-pressure condenser cs carbon steel DCSS distillation-condensation subsystem, including separator(s), reheater(s) and condensers DEA deaerator eco, ECO economizer section of a HRVG or HRSG EPRI Electric Power Research Institute, a US research consortium eva evaporator section of a HRVG or HRSG EvGT evaporative gas turbine FEvGT evaporative gas turbine with full-flow humidification, where all the compressed air from the compressor (after removal of the expander cooling air) is passed through the humidification system FGC flue gas condenser

97 Advanced Power Cycles with Mixtures as the Working Fluid

FSEvGT evaporative gas turbine with full-flow humidification and boilers for steam injection FWH feedwater heater gen generator HAT humid air turbine, the evaporative gas turbine cycle patented by Rao (1989) HRSG heat recovery steam generator (economizer, evaporator and superheater sections) HRVG heat recovery vapor generator (economizer, evaporator and superheater sections) H-SH humid air superheater HTU height of a transfer unit [m] (used for calculation of the humidification tower height) IGCASH integrated gasification compressed air storage with humidification IGCC integrated gasification combined cycle IGHAT integrated gasification humid air turbine mills thousandths of a USD NH3 ammonia NOx nitrogen oxides (i.e., nitrogen monoxide, NO, and nitrogen dioxide, NO2) PEvGT part-flow evaporative gas turbine, where only a fraction of the compressed air from the compressor is passed through the humidification system, with boilers for steam injection R1 steam Rankine cycle with one pressure level in the HRSG R2 steam Rankine cycle with two pressure levels in the HRSG rht reheater SEK currency of Sweden sep separator sh superheater section of a HRVG or HRSG ss stainless steel S-SH steam superheater STIG steam-injected gas turbine tbn turbine USD US dollars, currency of the USA 1-p HRVG or HRSG with one evaporative pressure level 2-p HRVG or HRSG with two evaporative pressure levels 16V25SG gas engine model with 16 cylinders and a bore of 25 cm 18V34SG gas engine model with 18 cylinders and a bore of 34 cm 18V32GD gas-diesel engine model with 18 cylinders and a bore of 32 cm 18V46GD gas-diesel engine model with 18 cylinders and a bore of 46 cm

98 7 Nomenclature

Parameters a [kg vapor/kg tot] cp specific [kJ/kg, K] E exergy flow rate [MW] FM module factor (used in Guthrie’s method to calculate the installed cost) m mass flow rate [kg/s] P net electric power output (i.e., the net power consumptions of pumps and fans and mechanical and generator efficiencies of vapor or steam turbines and gas turbines have been taken into account) [MW] p pressure [bar] Q heat rate of fuel or heat source [MW] T temperature [K] t temperature [°C] U overall heat transfer coefficient [W/m2, °C] V volumetric flow rate [m3/s] Wloss percentage of the expander power output that is lost in transferring the energy from the blades to the shaft, for example due to shaft/bearing and heat losses [%] x ammonia mass fraction [kg ammonia/kg mixture] β fraction of the compressed air mass flow rate sent to the humidi- fication system after the removal of the expander cooling air [%] ∆ difference η efficiency

Subscripts bas basic mixture, ammonia concentration in the low-pressure condenser in the ammonia-water cycle ca charge air e electricity eg exhaust gas el electrical h heat hi high evaporative pressure level in a steam Rankine cycle in in to the bottoming cycle is isentropic jw jacket water lo lubricating oil lw low evaporative pressure level in a steam Rankine cycle max maximum mech mechanical min minimum out out from the bottoming cycle

99 Advanced Power Cycles with Mixtures as the Working Fluid

tot total power output or efficiency of engines or gas turbine and bottoming cycle work working mixture, ammonia concentration in the HRVG and turbine in the ammonia-water cycle 0 reference state I first law, energy analysis II second law, exergy analysis

Expressions gross power output power output with the power requirements by pumps and fans taken into account, without consideration of mechanical or generator efficiencies for any components [MW] specific fuel consumption fuel consumption divided by the power generation [kg fuel/kWh] specific investment cost investment cost divided by the power output [USD/kWe] specific power output power output divided by the compressor intake mass flow rate of air for a gas turbine cycle or divided by the engine intake mass flow rate of air for a reciprocating engine with a bottoming cycle [kJ/kg]

100

8 Acknowledgments

First of all, I would like to express my gratitude to my supervisor Professor Jinyue Yan for his encouraging and stimulating guidance during the work with this thesis. I am also grateful to Professor Mats Westermark for useful ideas and discussions on evaporative gas turbine cycles and Professor Gunnar Svedberg for giving me the opportunity to begin doctoral studies. The discussions with and help from Dr. Eva Thorin during the work with the ammonia-water cycles were very useful. I would also like to thank my past and present colleagues at the Division of Energy Processes for providing a pleasant and creative working environment. The financial support from the Swedish Energy Agency (Statens energimyn- dighet), within the national research program Thermal Processes for Power Generation, and from the Royal Institute of Technology, through a doctoral studentship (“excellenstjänst”), is gratefully acknowledged. Last, but not least, I wish to thank my family for their support. I am especially indebted to Janne - I am grateful for all your support and encouragement.

101