Flexibility in thermal

power plants
With a focus on existing coal-fired power plants

STUDY
Flexibility in thermal
power plants

IMPRINT

STUDY STUDY BY:

Project management:
Flexibility in thermal power plants – Prognos AG
With a focus on existing coal-fired power plants Europäisches Zentrum für Wirtschaftsforschung
und Strategieberatung
Goethestraße 85
COMISSIONED BY: 10623 Berlin

Agora Energiewende Contact:

Anna-Louisa-Karsch-Straße 2 F. Ess
10178 Berlin | Germany Telephone: +41 (0)61 32 73-401
T +49 (0)30 700 14 35-000 Email: florian.ess@prognos.com
F +49 (0)30 700 14 35-129
www.agora-energiewende.de F. Peter
info@agora-energiewende.de
Comparison of thermal power plant technology
Project lead: Dimitri Pescia (chapter 3) and Retrofit options to increase
dimitri.pescia@agora-energiewende.de flexibility of coal-fired power plants (chapter 4):
Proofreading: WordSolid, Berlin Fichtner GmbH & Co. KG
Layout: oekom verlag, Munich Sarweystrasse 3
Cover image: istock.com/photosoup 70191 Stuttgart

Contact:
Dr. F. Klummp
Telephone: +49 (0)711 89 95-401
Email: florian.klumpp@fichtner.de

Please cite as:

Agora Energiewende (2017):
115/04-S-2017/EN Flexibility in thermal power plants –
Publication: June 2017 With a focus on existing coal-fired power plants
Preface

Dear readers, feasible. Since flexibility (rather than baseload gen-

eration) is the paradigm that shapes modern power
Due to immense cost reductions over the last decades, systems, increasing the flexibility of existing coal and
wind and solar power are contributing more and more gas power plants should be partially understood as
to the decarbonisation of power systems around the necessary for bringing them up-to-date.
globe. However, given their specific characteristics,
these technologies fundamentally change electricity Flexibility does not make coal clean, but making
systems and markets. More variable power produc- existing coal-fired plants more flexible enables the
tion increases the flexibility requirements placed integration of more wind and solar power in the
on the overall power system, both on the supply and system. In the mid- to long-term, coal power plants
demand sides. will be replaced by clean solutions like storage
technologies that provide power when the wind and
Often it is claimed that existing conventional power sun are scarce.
plants, especially coal power plants, cannot cope with
the weather-dependent generation of wind and solar I hope you find this report inspiring!
power. As a result, there is a rising level of renewable Yours sincerely,
energy curtailment in some power systems. How-
ever, this report shows that making existing power Patrick Graichen
plants more flexible is technically and economically Executive Director of Agora Energiewende

Key Findings at a Glance

Existing thermal power plants can provide much more flexibility than often assumed, as experi-
ence in Germany and Denmark shows. Coal-fired power plants are in most cases less flexible
compared to gas-fired generation units. But as Germany and Denmark demonstrate, aging hard
1 coal fired power plants (and even some lignite-fired power plants) are already today providing large
operational flexibility. They are adjusting their output on a 15-minute basis (intraday market) and
even on a 5-minute basis (balancing market) to variation in renewable generation and demand.

Numerous technical possibilities exist to increase the flexibility of existing coal power plants.
Improving the technical flexibility usually does not impair the efficiency of a plant, but it puts
more strain on components, reducing their lifetime. Targeted retrofit measures have been
2 implemented in practice on existing power plants, leading to higher ramp rates, lower minimum
loads and shorter start-up times. Operating a plant flexibly increases operation and maintenance
costs — however, these increases are small compared to the fuel savings associated with higher
shares of renewable generation in the system.

Flexible coal is not clean, but making existing coal plants more flexible enables the integration
of more wind and solar power in the system. However, when gas is competing with coal, carbon
pricing remains necessary to achieve a net reduction in CO2. In some power systems, especially
3 when gas is competing against coal, the flexible operation of coal power plants can lead to
increased CO2 emissions. In those systems, an effective climate policy (e.g. carbon pricing) remains
a key precondition for achieving a net reduction in CO2 emissions.

In order to fully tap the flexibility potential of coal and gas power plants, it is crucial to adapt power
markets. Proper price signals give incentives for the flexible operation of thermal power plants.
4 Thus, the introduction of short-term electricity markets and the adjustment of balancing power
arrangements are important measures for remunerating flexibility.

3
Agora Energiewende | Flexibility in thermal power plants

4
Content
Policy Summary 11

WORK PACKAGE 1: 
by Prognos AG

1. Introduction and background 21

2. The Effects of Expanded Renewables on Conventional Generation 23

2.1 Increasing requirements for flexible operation 23
2.2 The impact of renewables on the cost and utilisation of existing
thermal power plants 26
2.3 The Merit-Order Effect 27
2.4 Balancing power 28
2.5 Congestion management and renewables curtailment 29

WORK PACKAGE 2: 
by Fichtner GmbH & Co. KG

3 Thermal power plant technology – Flexibility and comparison

of generation technologies 33
3.1 Fundamentals of thermal power plant design and operation 34
3.2 Operational flexibility 42
3.3 Comparison of flexibility parameters in different generation technologies 46

4. Retrofits to increase flexibility of coal-fired power plants –

Options, potential and limitations 57
4.1 Key components for flexibility retrofits 58
4.2 Trade-offs between flexibility and efficiency 74
4.3 Potential and limitations of flexibility retrofits 76

5
Agora Energiewende | Flexibility in thermal power plants

WORK PACKAGE 3:
by Prognos AG

5. The Impact of Flexibility on Power Plant Profitability and CO2 Emissions 83

5.1 Flexibility impacts on power plant operations 83
5.2 Effects on CO2 emissions 88
5.3 Market design requirements to enhance the flexible operation of
thermal power plants 94

6. Profiles for Selected Countries: South Africa & Poland 99

6.1 South Africa 99
6.2 Poland 104

7. Conclusion 109

8. Bibliography 111

6
Index of figures
Figure S1: Power generation from nuclear, hard coal and lignite power plants and demand
in Germany, 23 to 30 March 2016 12

Figure S2: Qualitative representation of key flexibility parameters of a power plant 12

Figure S3: Minimum load and ramp rates of different hard coal power plants 13

Figure S4: Ramp rates and start-up times of different power plant technologies 13

Figure S5: Illustrative profil margin of a coal power plant in a short-term market with high shares
of renewables under different flexibility and must-run conditions 16

Figure 1: Flexibility requirements with high share of renewables. Example load curves for
two weeks during the winter in Germany. 24

Figure 2: Sorted hourly load change with and without the impact of renewable energy in Germany.
Example load curves for two weeks during winter in Germany. 24

Figure 3: Conventional power generation in Germany during ten days in November 2016 25

Figure 4: Decreasing power prices on the wholesale market due to increasing shares of renewable
energy production (Merit-Order Effect) 27

Figure 5: Relationship between monthly wind generation and monthly redispatch volumes
in Germany (Jan. 2013 – Oct. 2016) 29

Figure 6: Structure of Chapter 3 33

Figure 7: Energy conversion process for fuel-fired thermal power plants 34

Figure 8: Overview of thermal generation technologies by fuel type 36

Figure 9: Schematic view of a water-steam power process 37

Figure 10: Simplified representation of the state changes of water in a water-steam circuit 37

Figure 11: Schematic of a gas turbine in open cycle configurations 39

Figure 12: Schematic view of a combined cycle gas turbine configuration 40

Figure 13: Schematic view of steam extraction from a steam turbine providing heat to a district
heating system 41

Figure 14: Overview of flexibility characteristics 42

Figure 15: Qualitative representation of a power plant load curve with key power variables 43

Figure 16: Net power for a simplified start-up from standstill until nominal operation 44

Figure 17: Qualitative depiction of a power plant load curve highlighting the ramp rate 45

Figure 18: Ramp rates and start-up times of thermal power plants in comparison 47

Figure 19: Comparison of power plants with most commonly used technologies and power plants
with state-of-the-art technologies for each generation type with regard to minimum load
(values based on Table 1) 50

7
Agora Energiewende | Flexibility in thermal power plants

Figure 20: Comparison of power plants with most commonly used technologies and power plants
with state-of-the-art technologies for each generation type with regard to the average
ramp rate (values based on Table 1) 50

Figure 21: Comparison of power plants with most commonly used technologies and power plants
with state-of-the-art technologies for each generation type with regard to start-up time
(hot <8 h) (values based on Table 1) 51

Figure 22: Comparison of power plants with most commonly used technologies and power plants
with state-of-the-art technologies for each generation type with regard
to start-up time (cold >48 h) (values based on Table 1) 52

Figure 23: Structure of Chapter 4 57

Figure 24: 20 subsystems in a coal-fired power plant 58

Figure 25: Schematic illustration of a boiler in a tower construction 59

Figure 26: Tangential firing burner configuration with 4 burners on one stage 60

Figure 27: Schematic illustration of coal supply to burner 62

Figure 28: Net grid feed-in for indirect (IF) and direct firing (DF) configurations 62

Figure 29: Coal mill and burner arrangement of a boiler in tangential firing configuration
with four burner stages (single-mill operation) 63

Figure 30: Operation of four burner stages (left) in comparison to a single burner stage
in single-mill mode (right) 64

Figure 31: Load curves for pre-retrofit and post-retrofit of Unit G and H at Weisweiler 65

Figure 32: Charging a thermal energy storage system and its influence on net power fed into the grid 67

Figure 33: Simplified illustration of a coal-fired power plant with a gas turbine employed for feed water
pre-heating 68

Figure 34: Basic principle of BoilerMax application 69

Figure 35: Comparison of two start-ups at Zolling, one with BoilerMax and one without BoilerMax 70

Figure 36: Influence of repowering on net power and ramp rate 72

Figure 37: Influence of relative pressure on allowable temperature change rate in K/min
for a high pressure header designed with two different materials 73

Figure 38: Relationship between operating point and plant efficiency 74

Figure 39: Major coal-fired plant subsystems where retrofits were performed to improve flexibility 79

Figure 40: Hard coal power plant operation before and after retrofitting with lower minimum load,
increased ramp rates and reduced start-up time in a 48-hour example period 85

Figure 41: Hard coal power plant in must-run operation in a 48 hour example period 85

Figure 42: Hard coal power plant with temporary shut-down in a 48 hour example period 86

Figure 43: Hard coal power plant with lowered minimum load and increased ramp rates
in must-run operation in a 48 hour example period 86

8
 STUDY | Flexibility in thermal power plants

Figure 44: Hard coal power plant with lowered minimum load and increased ramp rates and
shorter start-up time in a 48 hour example period 88

Figure 45: Relationship between plant output and efficiency of hard coal and CCGT gas power plants
(600 MW nominal power) at different operating points (illustrative) 90

Figure 46: CCGT operation versus hard-coal plant operation in a 48 hour example period 91

Figure 47: CO2 emissions of CCGT and hard coal power plants under similar dispatch conditions
but with different flexibility features during 2 example days 92

Figure 48: Main challenges for the remuneration of flexibility 95

Figure 49: Balancing demand due to schedule leaps (hourly and quarter hourly) 96

Figure 50: Average balancing demand in Germany for each 15 minutes interval of the day
in 2012 to 2015 97

Figure 51: Power generation in South Africa by source, 1990–2014 100

Figure 52: Power supply in South Africa by source, 2014 101

Figure 53: Age distribution of coal power plants in South Africa 102

Figure 54: Technical performance of coal power plants in South Africa 102

Figure 55: Ramp rates in comparison 103

Figure 56: Power generation in Poland by source, 1990 to 2014 105

Figure 57: Power generation in Poland by source, 2014, in TWh 105

Figure 58: Age distribution of coal and lignite fired power stations in Poland 107

Figure 59: Unit size distribution for Polish coal and lignite fired power stations 107

9
Index of tables
Table S1: Summary of some retrofit options, their effect on flexibility parameters and
their limitations 14

Table 1: Comparison of most commonly used and state-of-the-art power plants for each generation
technology with regard to flexibility 48

Table 2: Comparison of three state-of-the-art coal-fired power plants in Poland and Germany 53

Table 3: Net efficiency and specific CO2 emissions for the most commonly used generation
technologies at nominal operation 54

Table 4: Net efficiency and specific CO2 emissions for state-of-the-art generation technologies
at nominal operation 54

Table 5: Summary of analysed retrofit options, their effect on flexibility parameters

and their limitations 76

Table 6: Potential and limitations of retrofit options for reducing minimum load 77

Table 7: Potential and limitations of retrofit options for reducing start-up time 78

Table 8: Potential and limitations of retrofit options for increasing ramp rate 78

Table 9: Plant parameters and market environment for the following illustrative examples 84

Table 10: Specific CO2 emissions for a range of fuels 89

Table 11: CO2 emissions for CCGT and hard coal fired power stations in different operational modes
(data from Figure 47) 92

10
 STUDY | Flexibility in thermal power plants

Policy Summary

The goal of limiting global warming to well below to effectively integrating large shares of renewables.
2°C can only be achieved if energy systems are almost This is especially true in systems characterised by
completely decarbonised over the long run. Renew- few other flexibility options and/or very high shares
able energies, especially wind and solar PV, are play- of existing inflexible power plants, for instance in
ing a fundamental role to reach this goal. They have Poland and South Africa. In those countries, existing
witnessed rapid expansion in power systems world- conventional power plants will continue to play a role
wide thanks to the immense cost reductions of the during the transition to a deeply decarbonised power
last decade. Because of their variable output and zero system. However, the generation output of these
marginal generation costs, these technologies alter power plants will need to adjust to the generation of
the characteristics of electricity systems and markets. variable renewables.
Steeper and more variable residual loads increase the
flexibility requirements placed on the overall power In the long run, however, fossil-fuel power plants,
system, both on the supply and demand sides. especially coal-fired plants, will need to be replaced
altogether with less CO2 intensive technologies if
In several countries the development of renewable international emission-reduction targets are to be
energy is hampered after reaching a certain pene- met.
tration level, because of the belief that the existing
power system cannot cope with the weather- 1. Existing coal power plants can
dependent generation of wind and solar power. technically provide much more
As a result, renewable energy curtailment has been on flexibility than many think,
the rise in various power systems, with priority given
as shown by experiences in countries
to baseload operation of conventional generation
like Germany and Denmark.
technologies. While it is true that conventional power
systems were not built to adjust to quickly changing In countries like Germany, hard coal-fired power
patterns on the supply side, system operators around plants, and to some extent lignite-fired power plants,
the world have learned to apply different flexible are already providing significant operational flexibil-
resources that complement growing shares of ity, adjusting their output to variation in renewable
variable renewable energy. There are many potential energy feed-in and demand (see figure 1).
sources of flexibility, including cross-border energy
trading, demand side management, storage tech- At the power plant level, operational flexibility is
nologies, flexible biomass/biogas, and the flexible characterised by three main features: the overall
operation of conventional generation technologies, bandwidth of operation (ranging between minimum
like gas and coal. and maximum load), the speed at which net power
feed-in can be adjusted (ramp rate), and the time
Regarding coal-fired power plants, it is widely required to attain stable operation when starting up
assumed that they cannot be operated to flexibly from standstill (start-up time) (see figure 2).
adapt to varying system loads without costly rede-
sign measures or losses in efficiency. However, State-of-the-art power plants have significantly
the contrary is the case, as we show in this report. improved flexibility characteristics. As illustrated
In actual fact, augmenting the flexibility of conven- in Figure S3 (left), state-of-the-art hard coal
tional power plants represents a major strategy for power plants can operate at minimum load levels

11
Agora Energiewende | Flexibility in thermal power plants

of 25–40 percent of nominal load. State-of-the-art mainly for baseload operation, especially in countries
lignite power plants can achieve minimum loads of like China or India, can have much higher minimum
35–50 percent of nominal load. By contrast, power load levels, significantly limiting the bandwidth of
plants built ten to twenty years ago in industrial- their operation. The ramp rate of state-of-the-art
ised countries had minimum load levels of 40 per- coal power plants (hard coal and lignite) can reach
cent (hard coal) to 60 percent (lignite). Retrofitting 6 percent of nominal load per minute, equalling or
can reduce minimum loads even further; in Germany, exceeding the ramp rate of the most-common CCGTs.
for example, minimum load levels of 12 percent have The ramp rate of the most-common hard coal power
been achieved. Older coal power plants designed plants in industrial countries is significantly lower,

Power generation from nuclear, hard coal and lignite power plants and demand in Germany,
23 to 30 March 2016 Figure S1

100
90
80
70
60
[GW]

50
40
30
20
10
0
26. Mar

28. Mar

30. Mar
29. Mar
24. Mar
23. Mar

25. Mar

27. Mar

demand Hard Coal Lignite Nuclear

Agora Energiewende (2017)

Qualitative representation of key flexibility parameters of a power plant Figure S2

100 PNom
Ramp rate
PNet [% PNom ]

PNet

30 PMin

Minimum load
0
Start-up time t [h]

Fichtner (2017)

12
 STUDY | Flexibility in thermal power plants

as can be seen in figure 3 (right). The same is true of Even though the flexibility features of state-of-the
old coal power plants in countries like South Africa, art coal power plants are significantly better than
where the ramp rates per minute usually do not those of older power plants, it must be pointed out
exceed 1 percent per minute. Start-up times, that coal-fired power plants are in general less flex-
both hot and cold, are also significantly reduced in ible than gas-fired generation units, especially in
state-of-the-art designs. regard to start-up times and ramp rates.

Minimum load and ramp rates of different hard coal power plants Figure S3

Minimum load of different hard coal power plants Ramp rates of hard coal power plants in South Africa
(as a percentage of nominal capacity) compared to most-commonly used and state-of-the-art designs
70 7

60 6
[% of Nominal Capacity

50 5
per minute]

40 4
[%]

30 3

20 2

10 1

0 0
old hard coal-fired most-commonly example of hard coal most-commonly state-of-the-art
power plants used and state- Retrofit power plants used hard coal hard coal
in China of-the-art hard in Germany in South Africa power plants power plants
and India coal power plants (Bexbach Unit)

DEA, NREL, Fichtner (left), Prognos, Fichtner (right)

Ramp rates and start-up times of different power plant technologies Figure S4

OCGT CCGT CCGT hard coal-fired

(state-of- (state-of- (commonly power plant
the-art) the-art) used) (state-of-the-art)
1
PNet [% PNom ]

start
0
0 1 2
t [h]

Fichtner (2017) based on (VDE, 2012)

13
Agora Energiewende | Flexibility in thermal power plants

2. Numerous technical possibilities helps to incorporate higher shares of renewables

exist to increase the flexibility into the power system. Retrofitting measures have
of existing coal power plants. been successfully implemented even in older power
plants, significantly enhancing their technical
Improving the technical flexibility
flexibility. For example, engineering and control
usually does not impair the efficiency
system upgrades at the Weisweiler hard coal power
of a plant, but it puts more strain plant in Germany allowed the minimum load levels of
on components, reducing their the two 600 MW generation units to be reduced by
lifetime. 170 MW (Unit G) and 110 MW (Unit H). This retrofit
also had a positive effect on the ramp rate, which
Targeted retrofit measures have been implemented was increased by 10 MW/min. The retrofit cost about
in practice on existing power plants, leading to 60 million euros per generation unit. At the Bexbach
higher ramp rates, lower minimum loads, and shorter hard coal power plant (721 MW), the minimum load
start-up times (see table S1). Important enabling fac- was reduced from 170 MW (22 % of PNom) to 90 MW
tors for success include also the adoption of alternate (11 % of PNom) by switching from two mill to single
operation practices as well as rigorous inspection mill operation. Boiler fire stability and the allowable
and training programs. Reducing minimum load thermal stress on components are the two main lim-
levels has proven to bring the most benefits, as it itations to improved flexibility. Nevertheless, as the

Summary of analysed retrofit options, their effect on flexibility parameters and their limitations Table S1

Option Minimum Start-up Ramp Limitations

load time rate

Indirect Firing   Fire stability

Switching from two-mill to single-mill operation  Water-steam circuit

Control system and plant engineering upgrade   Fire stability/

thermal stress

Auxiliary firing with dried lignite ignition burner   Fire stability and
boiler design

Thermal energy storage for feed water pre-heating  N/A

Repowering   N/A

Optimized control system  Thermal stress

Thin-walled components/special turbine design  Mechanical and

thermal stresses

“New” turbine start  Turbine design

Reducing wall thickness of key components  Mechanical and

thermal stresses

Fichtner (2017)

14
 STUDY | Flexibility in thermal power plants

above examples show, meaningful improvements can of their power plants, taking into account reduced
be achieved. plant lifetimes. Flexibility can also increase operation
and maintenance costs. From a system perspective,
The advanced age and limited operational flexibility however, these increased costs are relatively small
of existing coal power plants are a key driver of compared to the fuel savings associated with higher
modernisation measures. The net benefit of flexibility shares of renewable generation in the system.
retrofitting depends on factors specific to the power
plant and power system. Countries with large and 3. Flexible coal is not clean, but
aging coal-power fleets that were designed for making existing coal plants more
baseload operation have a large upside potential for flexible enables the integration
retrofit measures to increase efficiency and flexi-
of more wind and solar power in
bility. Improving the technical flexibility of a power
the system. However, when gas is
plant usually does not come at the expense of lower
efficiency or higher CO2 emissions. In many cases —  competing with coal, carbon pricing
for example, when pre-cast gas turbines are used —  remains necessary to achieve a net
flexibility measures can even improve the efficiency reduction in CO2.
of a coal-fired power station.
Power system effects are complex and the flexible
The investment costs required for flexibility operation of coal power plants without carbon con-
retrofitting must be considered specifically on a straints can, in some particular scenarios, increase
case-by-case. They can be roughly estimated at 100 CO2 emissions. In principle, the flexible operation of
to 500 €/kW (as the examples in chapter 4 show). coal power plants can have two conflicting effects on
Retrofitting usually increases the technical lifetime of CO2 emissions. On the one hand, the flexible opera-
a power plant by about 10–15 years.1 In comparison, tion of a coal-fired power plant can reduce its overall
overnight construction costs for new coal fired power CO2 emissions, since the plant generally produces less
stations with lifetimes of more than 40 years range electricity over the year. On the other hand, lower-
between 1,200 €/kW to more than 3,000 €/kW if CCS ing the minimum load through retrofit measures can
technology is implemented.2 reduce the efficiency of a power plant at low load lev-
els, increasing the specific CO2 emissions. (This effect
Flexible operation reduces the lifetime of a power is mitigated, however, by avoidance of expensive and
plant. Thick-walled components are especially CO2-intensive shutdown and start-up).
affected by thermal stress, which is exacerbated by
higher ramp rates and multiple start-ups. Model A comprehensive assessment of a power plant’s
calculations indicate that the lifetime of an old coal CO2 emissions must take into account characteristic
power plant is substantially decreased when sub- market and dispatch conditions as well as complete
jected to flexible operation. In Germany, some power operation cycles, without focusing only on the low-
plant operators deliberately push the flexibility limits est operating points. A comprehensive perspective
reveals that in many systems the benefits of greater
1 See NREL 2012: Cost-Benefit Analysis of Flexibility
flexibility outweighs the CO2 emission drawbacks of
Retrofits for Coal and Gas-Fueled Power Plants, low load operation, especially when one considers the
http://www.nrel.gov/docs/fy14osti/60862.pdf
expanded deployment of renewables in the system.
2 See Fraunhofer ISI et al: Estimating energy system costs of sectoral
RES and EE targets in the context of energy and climate targets
for 2030, http://www.isi.fraunhofer.de/isi-wAssets/docs/x/en/
However, in markets with a mixed portfolio of coal
projects/REScost2030-Background-Report-10-2014_clean.pdf power plants and other lower emission technologies

15
Agora Energiewende | Flexibility in thermal power plants

such as natural gas, coal retrofits improve the com- icantly positive effect on the overall emissions of the
petitive position of coal plants compared to other power plant fleet.
technologies. In such systems, increasing the flexi-
bility of coal-fired power plants can have a negative It is also important to state that, under similar dis-
impact on CO2 emissions at the plant level. Therefore, patch conditions, flexible coal power plants emit more
the goal of limiting CO2 emissions in the power sec- CO2 per MWh of electricity compared to gas power
tor must be addressed with effective CO2 abatement plant generation, even when taking into account the
policy. lifecycle emissions of the fuels.

Increased plant-level CO2 emissions after retrofit- 4. In order to fully tap the flexibility
ting can occur, for example, if partial load operation potential of thermal power plants,
prevents the coal-fired power plant from shutting it is crucial to adapt power markets.
down during periods of non-profitable operation
(however, this drawback is mitigated by avoidance of The economics of retrofitting existing coal power
CO2-intensive start-up). In such a case, the coal-fired plants are significantly influenced by the availabil-
power plant stays in the market due to its improved ity of remuneration options for flexibility. In other
competitive position compared to less CO2-inten- words, a market design that hampers investment in
sive gas plants. This has a negative impact on overall flexibility constrains the appropriate retrofitting of
CO2 emissions – unless the plant is a must-run plant coal power plants (not to mention the investment in
that would have stayed operational anyway in order alternative flexibility options). Proper price signals
to provide system services. In this latter case, which should remunerate the flexible operation of thermal
is likely in a system with very high share of coal, power plants. In short-term markets with a high
more flexible operation will generally have a signif- share of renewables, the profit margins earned by

Illustrative profil margin of a coal power plant in a short-term market with high shares
of renewables under different flexibility and must-run conditions Figure S5

140

120

100

80
[k€]

60

40

20

0
must-run and no-must-run but must-run but no-must-run and
inflexible limited flexibility increased flexiblity increased flexiblity

Prognos, Agora Energiewende (2017)

16
 STUDY | Flexibility in thermal power plants

flexible coal-fired power plants can be significantly (e.g. intraday trading) as well as the adjustment of
improved. To some extent, this can offset losses suf- balancing power arrangements. With these changes,
fered because of reduced utilisation (as a consequence integrating renewables into the power system
of the expansion of renewables). Indeed, reduced becomes easier and more economically efficient, and
minimum load is in many cases key for shoring up wasteful renewable energy curtailment is avoided.
profitability.
In this way, improving the operational flexibility of
Whether and to what extent flexibility retrofitting coal power plants can, together with other flexibil-
measures are profitable varies on a case-by-case ity measures, support the expansion of renewables
basis in relation to plant characteristics and the mar- during the transition toward a decarbonised power
ket environment (e.g. age of the plant, market share system. A crucial determinant of the need to retrofit
of renewables, general market design, remuneration coal power plants is the availability of alternative
options for flexibility). However, experience in flexibility options, including other flexible conven-
Germany shows that when the market is properly tional generation (gas, flexible hydro), demand-side
designed to remunerate flexibility, flexibility retro- flexibility and cross-border energy trading. The
fitting is likely to be profitable. quality and availability of these options varies con-
siderable between countries due to structural, eco-
With high shares of renewable power generation, nomic, and geographic factors. However, in countries
electricity markets should be designed to support with power sectors dominated by coal, improving the
market actors that provide valuable flexibility operational flexibility of coal power is an important
options. Necessary measures include the introduc- and highly viable option for bolstering the adoption of
tion of shorter-term electricity markets and products renewables.

17
Agora Energiewende | Flexibility in thermal power plants

18
Effects of Expanded
Renewables
on Conventional
Generation
WORK PACKAGE 1

WRITTEN BY
Prognos AG
Europäisches Zentrum für Wirtschaftsforschung
und Strategieberatung
Goethestraße 25
10623 Berlin
Telephone: +49 (0)30 52 00 59-200
Fax: +49 (0)30 52 00 59-201
www.prognos.com

Contributing authors:
F. Ess
Telephone: +41 (0)61 32 73-401
Email: florian.ess@prognos.com

F. Peter
Agora Energiewende | Flexibility in thermal power plants

20
 STUDY | Flexibility in thermal power plants

1. Introduction and background

The basis for international climate policy changed existing inflexible power plants, especially coal-fired
significantly with the adoption of the Paris Agree- plants.
ment in December 2015. The goal of limiting global
warming to well below to 2°C can only be achieved Historically, conventional generation capacities were
if energy systems are almost completely decarbon- built to follow rather predictable electricity demand
ised over the long term. The decarbonisation of the patterns. This paradigm favoured the construction
power system is essential in this regard, as fossil fuels of a mix of generation resources dominated by largely
remain the dominant source of power generation inflexible power plants, operating as baseload power
worldwide, and are responsible for a large share of (more than 80 percent of the year) and fired by lig-
global greenhouse gas emissions. Renewable energy nite, hard coal or nuclear energy. Today, the priority
such as wind power and solar photovoltaic are play- given to these inflexible power plants has become a
ing a fundamental role in the transformation of the major force curbing the development of renewables,
power system. These technologies have experienced especially in countries that rely on large share of coal
tremendous cost reductions in recent years and are power production (such as South Africa and Poland).
becoming cost-competitive with conventional tech- In these countries, existing conventional power plants
nologies for new investment. However, renewables will continue to play a role during the transition
are characterized by variable and uncertain out- toward a fully decarbonised power system. However,
put, increasing the need for flexibility in the power the generation output of these power plants will need
system. Indeed, enhancing supply and demand-side to adjust to the generation of variable renewables.
flexibility will be crucial for integrating higher shares
of renewables in a cost-efficient and reliable way. Existing coal power plants can technically provide
much more flexibility than many think, as this report
This study addresses an important concern that is will show. In countries like Germany and Denmark,
typically raised when discussing power systems targeted retrofit-measures have been implemented
with a high share of renewables. Once the develop- on existing power plants, significantly enhanc-
ment of renewables reaches a certain level, concerns ing their technical flexibility. Furthermore, effective
grow that existing conventional power plants cannot market incentives — including intraday electricity
be operated with sufficient flexibility. As a result, markets —  have been introduced in order to remu-
there are calls to limit the addition of new fluctuating nerate the provisioning of flexibility. Such measures
renewable capacity to the system. One clear problem have enabled renewable generation to be integrated
that is connected to this issue is the high level of more easily and in an economically efficient way,
renewable energy curtailment that occurs in certain thus limiting wasteful curtailment.
power systems  —  for example, in some provinces
in China, where priority is given to conventional Together with other flexibility measures, improv-
baseload generation. ing the flexibility of thermal power plants can enable
higher shares of renewable production during the
Making existing conventional power plants more transition to a decarbonised power system. In the
flexible is therefore a key prerequisite for integrating long run, however, fossil-fuel power plants, especially
large shares of renewables more effectively. This coal power plants, will need to be replaced altogether
is especially true in systems characterized by few with less CO2 intensive technologies if international
other flexibility options and/or very high shares of climate targets are to be met.

21
Agora Energiewende | Flexibility in thermal power plants

The main aim of this study is to provide a broad technical characteristics related to the flexibility of
analysis on possible flexibility measures for thermal thermal power plants. The third part analyses some
power generation while focusing on coal power retrofit measures to increase the flexibility of coal
plants. In doing so, we consider technical and eco- power plants, including their technical and economic
nomic factors related to increasing the flexibility of parameters. Fourth, our findings with regard to chal-
conventional power plants.3 The study is divided in lenges and opportunities are discussed and put into
four parts: The first part analyses major challenges perspective by spotlighting the situation in South
related to the integration of large shares of renew- Africa and Poland, two countries with large coal
ables. The second part describes in detail current power generation shares.

3 Note that the flexibility challenge of combined heat and power

plants (CHP) is not addressed in detail in this report.
Brief information on this topic can be found in section 3.1.5.

22
 STUDY | Flexibility in thermal power plants

2. The Effects of Expanded Renewables

on Conventional Generation

An increasing share of variable renewable energy countries. A third option is to store electricity using
such as wind and PV has a direct impact on the oper- conventional storage technologies (e.g. hydro storage)
ation of conventional power plants. Conventional or new technologies (e.g. batteries). Last but not least,
power plants need to operate more flexibly, meaning increasing power plant flexibility make a key contri-
they have to ramp up and down more frequently and bution to greater system flexibility, and, by extension,
more quickly, operate often at partial loads and have promote the integration of renewables.
to be turned on and off with greater regularity. More-
over, a rising share of renewables also decreases the Historically, conventional power plants have been
market profitability of conventional generation due designed to serve electricity demand pattern that
to the so-called Merit-Order Effect. In addition, it has is characterized by relatively low variability as well
indirect impacts on conventional power plants, as it as prototypical daily, weekly and seasonal profiles.
increases the demand for balancing and congestion In the absence of variable renewables, this leads to an
management in the power system. optimal generation mix with a high share of base-
load power plants (i.e. running more than 80 percent
2.1 Increasing requirements for flexible of the year). However, renewable generation is highly
operation variable, and to some extent less predictable. With a
high share of variable renewables, a large proportion
In a power system characterised by increasing of conventional generation can no longer operate as
shares of renewable power generation, the flexibil- baseload capacity and must be run with greater flex-
ity requirements placed on existing conventional ibility.
capacities rise significantly. The main cause of an
increased need for flexibility is the variable nature The need for flexibility and the challenges faced by
of power generation from wind power and photo- conventional power plants are illustrated in Figure 1.
voltaics (PV). Both technologies depend on weather The left side of the figure (a) shows the hourly struc-
conditions, daily and seasonal changes, and therefore ture of electricity demand (load) over two weeks.
cannot generate “on demand” like conventional power On the right side (b) the same two weeks are plotted  —  
plants. Furthermore, renewables have almost no mar- but with an annual share of 40 percent renewables in
ginal costs. This means that they produce “for free” the system. This “residual load” profile is derived by
whenever the primary resource (i.e. wind or sun) is subtracting hourly renewable generation from hourly
available. These factors entail a fundamental trans- electricity demand.
formation of power systems, because of the need to
respond flexibly to variation in renewables feed-in. In a system with no variable renewables, conven-
tional power plants serve demand based on the load
Several options currently exist to provide more curve (see figure 1a). In systems with high shares of
system flexibility for the integration of renewables. wind and PV, conventional plants must serve the load
Encouraging demand-side flexibility (e.g. more flex- not covered by variable renewables, i.e. the resid-
ible manufacturing processes) is one option. Another ual load curve. Therefore, their operation has to be
is to promote grid development, so that power can be significantly more flexible (figure 1b). Whereas the
transported with greater ease between regions and load only ranges between 47 and 84 GW in a sys-

23
Agora Energiewende | Flexibility in thermal power plants

Flexibility requirements with high share of renewables. Example load curves for two weeks
during the winter in Germany. Figure 1

a) Load b) Residual load 40 % RES

100,000 100,000

80,000 80,000
[MWh/h]

[MWh/h]
60,000 60,000

40,000 40,000

20,000 20,000

0 0

−20,000 −20,000

300
320
20
40
60
80
100
120
140
160
180
200
220
240
260
280
0
20
40
60
80
100
120
140
160
180
200
220
240
260
280
300
320

[hour] [hour]

Prognos

Sorted hourly load change with and without the impact of renewable energy in Germany.
Example load curves for two weeks during winter in Germany. Figure 2

9,000

6,000

3,000

0
[MW/h]

−3,000

−6,000

−9,000

−12,000
0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320
[hour]

Load Residual load 40% RES

Min: −9,502 MW/h Min: −11,308 MW/h
Max: 4,964 MW/h Max: 6,280 MW/h
Average: 2,219 MW/h Average: 2,595 MW/h

Source: Authors’ figure.

Prognos

24
 STUDY | Flexibility in thermal power plants

Conventional power generation in Germany during ten days in November 2016 Figure 3

100
Power generation and consumption [GW]

80

60

40

20

0
14.11. 15.11. 16.11. 17.11. 18.11. 19.11. 20.11. 21.11. 22.11. 23.11. 24.11.
Conventional Solar Wind Wind Hydro Biomass Electricity
power plants Onshore Offshore Consumption
100
Conventional powergeneration [GW]

80

60

40

20

0
14.11. 15.11. 16.11. 17.11. 18.11. 19.11. 20.11. 21.11. 22.11. 23.11. 24.11.
Hard coal Lignite Nuclear Pumped hydro Natural gas Other

Agora Energiewende (Agorameter)

tem without renewables during these two example mum and maximum hourly load changes (Figure 2). In
weeks, in a system with 40 percent variable renew- our example, the overall average hourly load change
ables, the residual load can fall to minus 12 GW (due increases from 2,219 MW per hour to 2,595 MW per
to temporary surpluses from renewable generation) hour. This represents an increase of about 17 percent.
and rise to 70 GW within few days. Residual-load If intermittent renewable shares reach even greater
ramp rates (i.e. load changes in one or more consec- levels, the observed load changes will also increase
utive hours) are also considerably higher than the accordingly.
variations in electricity demand. Figure 2 shows in
greater detail the hourly load changes during the two The German power system provides a good exam-
example weeks. The addition of intermittent renew- ple on how conventional power generation can
ables leads to a significant change in both the mini- adjust output in a power system characterised by

25
Agora Energiewende | Flexibility in thermal power plants

considerable amounts of renewable generation.4 2.2 The impact of renewables on

Figure 3 illustrates power generation in Germany the cost and utilisation of existing
during 10 days in November 2016, which includes thermal power plants
high feed-in from wind power during the weekend
of November 20–21. In addition to imposing the need for greater flexibil-
ity, the increasing integration of variable renewables
As can be seen in figure 3, conventional power gener- has economic impacts both on the utilisation of
ation from gas power plants as well as hard-coal conventional power plants and their profitability.
and lignite plants declines significantly when there The flexible operation of coal power plants is techni-
is high renewable feed-in (and low power demand). cally possible (as will be shown in the following sec-
In this example, the flexible operation of conventional tions), yet reducing the utilisation of capital-inten-
power plants enables the integration of large amounts sive technology has negative impacts on profitability.
of renewables, especially on November 20, when However, these impacts can be mitigated with a
both wind energy and solar PV contribute to up to properly functioning market (see sections 5.1 and 5.3).
60 percent of power demand during a few hours.
In general, when additional capacity is introduced
This study examines the following flexibility factors to a power system  —  whether wind, solar, or any other
for conventional power plants in a system with a sig- type of power plant  —  the output and revenues of
nificant share of renewables: other power plants tend to be reduced. In contrast to
thermal power plants, however, wind and solar power
→→ Ramp rates: Because load variations in the residual plants produce electricity only if the wind blows or
load are larger than variations in electricity the sun shines. This means that their output is varia-
demand (load), conventional power plants have ble and does not react to changes in demand for elec-
to be faster in adjusting their generation over the tricity. This has two important implications for the
course of one or more consecutive hours. utilisation of existing conventional power plants:

→→ Minimum load: Renewable generation during one →→ First, the structure of residual demand (defined as
hour can amount to nearly 100 percent of demand, demand minus renewables feed-in) is changed,
even if shares of renewable generation are much leading to a change in the use of existing power
lower over the whole year. Therefore, conventional plants. In the long run, this also produces a change
generation must adjust to lower operating thresh- in the cost-optimal mix of residual power plants.
olds than are adequate in a system without a sig- This is often described as a shift from “base load” to
nificant share of renewables. “mid-merit and peak load”.

→→ Start-up times: At certain times it is necessary →→ Second, conventional thermal power plants may
(and economically beneficial) for conventional still be needed in the system, in order to provide
power plants to shut down temporarily. Start-up capacity during times of high demand, particularly
times after such a shut-down are another crucial when the wind is not blowing and sun is not shin-
factor that determine the flexibility of conventional ing. This is often described as a need for “backup
power generation. capacity”, or alternatively as the need for thermal
capacity with reduced average utilisation.

4 In 2016, variable renewables (wind energy and solar PV)

These two factors impact the fixed costs and variable
accounted for 18 % of power consumption in Germany. costs of the thermal power plants in the system.

26
 STUDY | Flexibility in thermal power plants

Decreasing power prices on the wholesale market due to increasing shares of renewable
energy production (Merit-Order Effect) Figure 4

140
Demand
120
Supply curve
low RED feed-in
100
[EUR/MWh]

80

60
Merit order effect
40 Supply curve
high RED feed-in

20

0
0 20,000 40,000 60,000 80,000 100,000 120,000
[MW]

Prognos

The reduced average utilisation of the thermal power In liberalised markets, the power prices on the whole-
plants leads to higher specific generation costs sale market are determined by supply and demand.
(EUR/MWh). This effect is particularly important for Typically, wholesale markets are organised using
generation technologies that are capital intensive, auctions as pay-as-clear markets.5 To calculate the
like coal-fuelled power plants. Coal and lignite are a market clearing price, the supply curve (merit order)
highly available and low cost energy source in many is first sorted in an ascending order by means of the
countries in the world. In those countries, low fuel variable costs of the power supply units (see figure 4).
costs in combination with rather inflexible power The variable costs are determined by different factors,
plants, designed for baseload operation, increase the such as fuel prices, CO2 costs and opportunity costs.
benefits of inflexible operation. When intermittent Nuclear and lignite power plants typically have low
renewables are incorporated into such a system, variable costs, while hard coal and new CCGT have
the likely response is to simply curtail renewable medium variable costs. OCGT and oil power plants
generation when feed-in is very high. have the highest variable costs. In a second step, the
market clearing price is determined by the intersec-
2.3 The Merit-Order Effect tion of the supply and demand curves.

Beyond impacting the utilisation of thermal power In contrast to thermal power stations, wind and PV
plants (as discussed above), renewables also impact have no variable costs. Therefore, renewable energies
power plant earning in the wholesale market due to
the so-called Merit-Order Effect. 5 In contrast to pay-as-bid markets, each successful bidder
gets/pays the same price.

27
Agora Energiewende | Flexibility in thermal power plants

integrate at the beginning of the merit order, pushing systems, power plant dispatch takes place well ahead
conventional technologies further out on the merit of real-time, increasing weather forecasting errors).
order. This has two effects: on the one hand, the utili-
sation rate of power plants tends to decrease  —  espe- The magnitude of forecasting error depends on the
cially during times of high renewable energy produc- quality of the forecasting methods and the time hori-
tion and low demand (as explained in 2.2). On the other zon for which the forecast is made. While forecasting
hand, the average market clearing price decreases as errors are likely to be significant when made over a
more expensive technologies are less frequently em- period of several hours or days, they are likely to be
ployed. This crowding out effect is termed the Mer- close to zero if made for a period less than an hour.
it-Order Effect. Both effects decrease the profitability Furthermore, the relative size of the deviation is also
of thermal power plants on the wholesale market. likely to decline with a greater geographical distribu-
tion of renewable power plants.
In Germany and other markets, the increasing
penetration rate of renewable energy in combination Other factors also influence balancing demand, such
with low fuel and emission costs and surplus of pro- as schedule leaps (see subsection 5.3) and the size of
duction capacities have placed significant pressure the balancing area.
on conventional generation assets during the past
years. Several power stations have been forced to The impact of these factors can be observed in the
shut down. German balancing system. Yet despite increasing
energy generation from wind and PV, balancing
2.4 Balancing power demand has not increased. This is primarily attribut-
able to the efficiency savings that have been achieved
Renewable generation, being weather-dependent, with the introduction of the International Grid
is subject to forecasting errors. Forecasting errors Control Cooperation, which increased the balancing
increase the need for maintaining and activating bal- area. In addition, the impact of scheduling leaps was
ancing reserves, and can therefore increase balancing reduced by strengthening the trading of quarter-hour
costs. Other factors, however, can decrease balancing power contracts.
costs, partially offsetting the cost impact of increased
renewables (for example, more competitive balancing To evaluate the cost of integrating renewable energy,
markets, better forecasts, liquid intraday markets, both the demand and supply side have to be taken
better cooperation between TSOs, etc.). Balancing into consideration. The market entry of new partici-
power is necessary to guarantee the frequency pants and technologies (thanks to eased prequalifica-
stability of electrical grids by balancing in real-time tion requirements as well as financial pressures from
power generation and consumption. If the power decreased wholesale market revenues) have recently
system is undersupplied, positive control power has reduced balancing costs in the German market.
to be added, whereas negative control power has to
be activated if the system is oversupplied. In power systems with mostly thermal plants,
balancing costs are estimated at between 0 and
The causes of balancing power demand are vari- 6 EUR/MWh, even at wind penetration rates of up
ous. In systems without renewable generation, the to 40 percent. In power systems with significant
primary causes are unplanned power plant outages, shares of flexible hydro generation, such as the
load forecasting errors and load noise. In systems Nordic region, balancing costs are even lower.6
with variable renewables, errors in forecasting wind
and PV production must be added to the list. (In most 6 See Agora Energiewende (2015).

28
 STUDY | Flexibility in thermal power plants

2.5 Congestion management and gestion, network operators employ various measures.
renewables curtailment These include network switching, countertrading,
redispatch7 of conventional power plants and the
The German experience shows that the expansion curtailment of renewable energy production.
of variable renewables changes power flows in the
grid, which can impact the operation of conventional In countries with priority feed-in for renewable
power plants. energy, curtailment of renewables generation is by
law the last option to be chosen. In these countries,
The production of wind and PV power is location- redispatch regimes, in which network operators
specific. Typically, wind turbines and PV panels request power plants to adjust their production, are
are installed in regions with high wind speeds and usually the favoured solution. This requires power
solar radiation. Often these renewable generation plants to be flexible enough to come back to their
centres are geographically distant from where power schedule after the redispatch, in order to avoid creat-
is actually consumed. As wind and solar radiation ing new imbalances to the system..
cannot be stored and transported directly like coal or
natural gas, the renewable power has to be transmit-
ted. However, the expansion of the transmission and 7 In the event of a redispatch request to conventional power plants,
the asset before the network congestion has to shorten its power
distribution grid has lagged behind the expansion of generation, whereas a different power plant after the bottleneck
renewable capacities. To avoid short-term grid con- balances the shortage by increasing power generation.

Relationship between monthly wind generation and monthly redispatch volumes in Germany
(Jan. 2013 – Oct. 2016) Figure 5

1,400

1,200
Monthly redispatch volume [GWh]

1,000

800

600

400

200

0
0 2,000 4,000 6,000 8,000 10,000 12,000 14,000
Monthly wind production [GWh]

Prognos based on TSOs data (netztransparenz.de)

29
Agora Energiewende | Flexibility in thermal power plants

In other countries, the curtailment of renewables All of the described effects associated with a large
may be chosen as a first option, especially if the share of renewables lead to significant changes in
market design favours baseload operation of thermal the operation of conventional power plants. Fleets of
generation (for example, through long-term contracts power plants that are dominated by rather inflexible
and priority access). This is, for example, the case assets, i.e. that have been mainly designed for base-
in China, where about 15 percent of the total wind load operation, would prefer to see the curtailment
production was curtailed in 2015, with the level of of renewable energy as the key option for assuring
curtailment even reaching 30–40 percent in some system stability. However, this would substantially
provinces, according to the Danish Energy Agency. lower the CO2 savings of increased RES shares, lead to
higher system costs and limit the level of RES that can
Figure 5 illustrates the relationship between wind be incorporated into the system.
energy production and redispatch volumes in the
German markets. The figure clearly shows the impact The following sections describe in detail the techni-
of increasing wind energy production on redispatch cal potential for increasing the flexibility of exist-
volumes. Regional grid constraints are the main ing power plants. These options pave the way for the
reason for increased redispatch volumes. This graph integration of larger shares of renewable energy, even
shows that without redispatch, monthly renewable when the conventional power plant fleet is dominated
curtailment could reach significantly high levels by coal and lignite stations that were installed mainly
(up to 1 TWh in the most windy month). for baseload operation.

30
Comparison of thermal
power plant technology
and
Retrofit options
to increase flexibility of
coal-fired power plants
WORK PACKAGE 2

WRITTEN BY
Fichtner GmbH & Co. KG
Sarweystrasse 3
70191 Stuttgart
Telephone: +49 (0)711 89 95-693
Fax: +49 (0)711 89 95-459
www.fichtner.de/

Contributing author:
Dr. Florian Klummp
Telephone: +49 (0)711 89 95-401
Email: florian.ess@prognos.com
Agora Energiewende | Flexibility in thermal power plants

32
 STUDY | Flexibility in thermal power plants

3. Thermal power plant technology –

Flexibility and comparison
of generation technologies

Structure of Chapter 3 Figure 6

Section 1:
Fundamentals Section 3:
Section 2:
of thermal Comparison of
Operational
power plants generation
flexibility
design and technologies
operation

Fichtner (2017)

The following chapter is organized into three sections

(Figure 6).

Section 1 explains the key terminology and under­ Section 3 compares four relevant thermal generation
lying working principles of thermal power plants. technologies based on their flexibility parameters and
It also provides an overview of pertinent generation CO2 emissions: OCGT and CCGT gas-power plants,
technologies. lignite-fired and hard coal-fired power plants.
It also presents the characteristics of specific coal
Section 2 introduces the concept of operational flex- power plants.
ibility. The scope of the study encompasses three key
parameters that characterize flexibility: minimum
load, start-up time and ramp rate.8

8 Flexible operation can also be characterized by technical parameters

such as fuel flexibility or black start capacity. (Fuel flexibility is the
ability to burn a wide range of fuels with different properties. Black
start capacity describes the ability of restarting a power plant without
requiring the external grid). However, the relevance of such parame-
ters are of secondary importance for assessing the overall flexibility
of the power plants and were therefore not considered in this study.

33
Agora Energiewende | Flexibility in thermal power plants

3.1 Fundamentals of thermal power boiler, the turbine and the generator. Energy losses
plant design and operation occur during each conversion step. Below is a brief
description of the main steps, and where they occur
The following fundamentals are necessary for under- within the power plant.
standing the basic operation of thermal power plants.
Burner/Boiler
3.1.1 Definition of key terminology Chemical energy stored in the fuel is converted into
To have a sound discussion about a topic as complex thermal energy via combustion.
as power plant technology, it is important to provide
a precise definition of the terminology. Turbine
Thermal energy (gas or steam at high temperature and
pressure) is converted into mechanical energy (torque
Thermal power plant on a shaft) through the expansion of the working
A thermal power plant is characterized by an fluid.
energy conversion process in which thermal energy
(e.g. released during fuel combustion) is converted Generator
into electrical energy. Mechanical energy is converted into electricity
through electromagnetic induction.

Figure 7 illustrates the energy conversion process for Cooling tower/Exhaust

fuel-fired thermal power plants. The second law of thermodynamics says that thermal
energy cannot be fully converted into mechanical
This figure shows the conversion of fuel in thermal energy. The non-convertible part (anergy) has to
power plants. Each type of energy conversion takes be released into the environment through a cooling
place in a main power plant component  —  the burner/ tower or through exhaust.

Energy conversion process for fuel-fired thermal power plants  Figure 7

Useful
Heat*
Burner/Boiler Turbine Generator

Chemical Thermal Mechanical Electric

Energy Energy Energy Energy

Cooling Tower/
Exhaust/etc.

Non-Usebale
* Only in combined
Energy heat and power (CHP)
(Losses) configuration

Fichtner (2017)

34
 STUDY | Flexibility in thermal power plants

Cogeneration module (if in use)

Different power plant technologies offer different Load
ways to cogenerate electricity and heat. For instance, In the context of power plant operation, the net
heat can be used directly for some industrial pro- power, PNet, is usually referred to as the load of
cesses or it can be fed into a district heating system. a power plant.
Section 3.1.5 explains this process in more detail.
Nominal load, also referred to as nameplate or
nominal capacity, describes the highest consistent
Efficiency net power output of a power plant operating
In the context of thermal power plants, efficiency, under design conditions. It is denoted by PNom
usually denoted by the Greek symbol η, represents for the remainder of this report.
the share of the fuel’s energy that is converted into
electricity. Part load describes the operation of a power plant
with a net power output that is lower than its
nominal value.
Alternatively, efficiency is formulated as a ratio
between a system’s beneficial output (e.g. net power Minimum load describes the lowest net power
PNet) and its input (e.g. heat flow released through the output a power plant can deliver while maintaining
combustion of fuel Q̇In): stable operation. It is denoted by PMin for the
remainder of this report.
P
η = ———Net
———
Q̇In
Typically, power plants are optimized to have their
The net power is the power that is fed into the grid. highest efficiency at or close to their nominal load.
It is defined as the generator power output, When a power plant has to reduce its electricity gen-
PGen (sometimes called gross output), minus the power eration, it is forced to operate under part load condi-
required to drive auxiliary systems, PAux, such as tions, at a lower efficiency. This in turn leads to higher
pumps, fans and coal mills. CO2 emissions per MWh as described in Section 3.3.2.

Efficiency is closely related to the CO2 emissions 3.1.2 Overview of thermal generation
of a thermal power plant. For a specific amount of technologies
generated electricity, usually denoted in MWh or This report considers four main thermal generation
GWh, less fuel is required when the power plant is technologies:
operated at a higher efficiency, which also translates
into lower specific CO2 emissions. Typical efficiency →→ lignite-fired power plants;
values at nominal loads in thermal power plants vary →→ hard coal-fired power plants;
between 39–60 percent, depending on type and age →→ open cycle gas turbine (OCGT) power plants; and
of the power plant.9 (This is described in detail in →→ combined cycle gas turbine (CCGT) power plants.
Section 3.3.2.)
As shown in Figure 8, fossil-fuelled power plants are
separated by fuel types.

9 The parameters of most commonly used generation technologies and

Coal is the leading fuel used in steam power plants.
state-of-the-art generation technologies are defined in Section 3.3. Lignite and hard coal need to be distinguished

35
Agora Energiewende | Flexibility in thermal power plants

Overview of thermal generation technologies by fuel type Figure 8

Thermal
Generation
Technologies
Coal-fired Gas-fired
Power Plants Power Plants

Hard coal-fired OCGT

Lignite-fired CCGT

Fichtner (2017)

because they greatly influence the characteristics Figure 9 shows a schematic view of a general water-
of a power plant’s operation. steam circuit. Its main components are the pump (1),
the boiler (2), the turbine coupled with the genera-
Natural gas, for the remainder of this report simply tor (3) and the condenser (4).
referred to as gas, is a fuel used in gas-fired power
plants. Gas-fired power plants are characterized The process can be broken down into four steps:
by their operation design, which can either be open
cycle or combined cycle. Step 1: Pressure increase
A pump increases liquid water pressure. Since water
The next three sections analyse the working principles is nearly incompressible, its density undergoes virtu-
of coal-fired power plants (Section 3.1.3), gas-fired ally no change during this step.
power plants (Section 3.1.4) and combined heat and
power (CHP) plants (Section 3.1.5). Other generation Step 2: Heat addition through coal combustion
technologies, such as internal combustion engines or The boiler burns a mixture of air and fossil fuel, such
nuclear power plants, are not analysed in this study. as coal. The thermal energy released through this
process is then transferred to the water, causing the
3.1.3 Basic working principle of coal-fired water to evaporate and turning it into steam. After
power plants all the water has been evaporated, the steam contin-
The underlying working principle of steam tur- ues to be heated in a process known as superheating.
bine-driven power plants, such as coal-fired, nuclear This increases the temperature and specific volume
or concentrated solar power plants, is the water- of the steam.
steam circuit. In thermodynamics, this is referred to
as the Rankine cycle. It is a self-contained working Step 3: Expansion in the turbine
cycle, which means that the working fluid (water) After the heat is added, steam expands in the
experiences different changes in its state but never turbine. The reactive forces of the expanding fluid
leaves the cycle. Steam turbines generate mechani- are used to drive the turbine. This process is driven
cal torque through the expansion of high temperature by a significant pressure difference between the
and high pressure steam. turbine inlet and outlet. At the turbine outlet, both

36
 STUDY | Flexibility in thermal power plants

Schematic view of a water-steam power process Figure 9

Live steam

Air Boiler Generator

Fuel (Step 2)
G
Pump
(Step 1) Turbine
(Step 3)
Water
Condenser
(Step 4) Exhaust steam

Fichtner (2017)

pressure and temperature of the steam decrease working fluid returns to its liquid state by releasing
significantly. heat at a low temperature to a cooling medium, such
as water from a nearby river.
Step 4: Condensation
Since a steam turbine process is a closed cycle, After returning to its liquid state, the water continues
a fourth step is necessary to bring the working fluid the cycle and undergoes the above state changes on a
back to its original liquid state. The non-converti- continuous basis (1–4).
ble part of the thermal energy (anergy), contained in
the steam after expansion, has to be released through The figure 10 depicts the qualitative state changes of
condensation. During the condensation process, the water in a water-steam cycle. The y-axis represents

Simplified representation of the state changes of water in a water-steam circuit Figure 10

2. Heat addition in the boiler

230
Pressure [bar]

1. Pressure increase
3. Expansion in turbine

0.1

4. Condensation
40 46 620
Temperature [°C]

The small rectangles represent the water in a liquid state; the large rectangles, the water in steam state; the blue represents a state at lower
temperature and the red a state at high temperature.
Fichtner (2017)

37
Agora Energiewende | Flexibility in thermal power plants

the working fluid pressure in bar, whereas the x-axis In comparison with hard coal, lignite’s low energy
shows the temperature in degrees Celsius. The tem- density (about 8 MJ/kg) requires a larger boiler and
perature after condensation mainly depends on the flue gas cleaning equipment to reach a specific
cooling medium employed. power output, leading to relatively long and cost-
intensive start-up.
Differentiation of subcritical, supercritical and Due to the relatively low energy density of lignite,
ultra-supercritical water-steam circuits it is not economically feasible to transport it over
Water-steam circuits can be operated below or above long distances. Hence, lignite-fired power plants are
the critical point of water specified by its critical usually constructed close to mining areas
pressure and temperature (pc = 221,2 bar; Tc = 374,15 °C).
Three types of water-steam circuits exist and are dif- Hard coal-fired
ferentiated based on their live steam parameters: Hard coal-fired power plants show a greater flexibil-
subcritical: 160 bar / 535 °C ity than lignite-fired power plants. Their component
supercritical: 240 bar / 540 °C dimensions are smaller, mainly due to larger energy
ultra-supercritical: 285 bar / 600 °C density (about 25–32 MJ/kg) and lower water con-
tent (about 2–7 percent) relative to lignite.
Higher temperature and pressure during operation Before the hard coal is blown into the boiler of
require advanced materials but also yield higher effi- the power plant it is finely grained in the bowl mills
ciencies. and dried with a hot air stream to reduce its water
content.
Lignite- and hard coal-fired power plants
Both power plant types use a steam turbine cycle. The
main difference is the coal type, which has significant 3.1.4 Basic working principle of gas-fired
implications on plant operation. State-of-the-art hard power plants
coal-fired units provide up to 900 MW, whereas state- The Joule cycle is the underlying working principle
of-the-art lignite-fired units reach up to 1,050 MW. of gas-fired power plants. One distinguishes between
open cycle and combined cycle configurations. Com-
bined cycle gas turbines employ the Joule as well as
Lignite-fired the Rankine cycles (described in the previous Sec-
Lignite-fired power plants are typically designed to tion 3.1.3).
operate at nominal load for most hours of the year
(i.e. baseload operation) and should only perform 1. Open cycle gas turbine (OCGT)
a few start-ups annually. Gas turbines create mechanical torque by expand-
The high water content of lignite (45–60 percent), ing a mixture of compressed air and flue gas at high
requires a pre-combustion drying procedure in the pressure and temperature. In the open cycle configu-
mills (beater-wheel mills). For this process, hot flue ration, the exhaust stream is released to the environ-
gas (up to 1,000 °C) is fed in.10 ment.

The open cycle gas turbine process is illustrated in

Figure 11. The basic components are the compres-
10 Flue gas describes the gas stream exhausted to the environment through
a flue gas stack or chimney after a combustion process. For fossil-fired sor (1), combustion chamber (2) and the turbine cou-
thermal power plants, the composition of the flue gas depends on pled with the generator (3). Compressor, gas turbine
the type and characteristics of fuel that is combusted and the com-
bustion characteristics. The main constituents of flue gas are nitro-
and generator are mounted on a common shaft.
gen (N2), oxygen (O2), water vapor (H2O) and carbon dioxide (CO2).

38
 STUDY | Flexibility in thermal power plants

Schematic of a gas turbine in open cycle configurations Figure 11

Fuel
Combustion
Air Chamber
Inlet (Step 2)

Generator

Compressor Turbine
(Step 1) (Step 3) Exhaust

Fichtner (2017)

The process can be broken down into three steps: The components of a CCGT are similar to gas and
steam turbine power plants. A heat recovery steam
Step 1: Compression generator (HRSG) is used instead of an externally
During operation, ambient air is sucked into the fired boiler. It transfers thermal energy from the
machine by the compressor and brought to a higher exhaust gas of the gas turbine to the water of the
pressure level. steam turbine cycle. Figure 12 shows a schematic
view of a CCGT.
Step 2: Heat addition through gas combustion
The compressed air enters the combustion chamber The process can be broken down into three steps:
and is mixed with the fuel (i.e. natural gas).
The thermal energy released during combustion Step 1: OCGT process
causes an increase in gas temperature and volume. For typical CCGT configurations, heat input only
takes place during the Joule cycle through fuel com-
Step 3: Expansion in the turbine bustion. The generated electricity in the gas turbine
The hot gas mixture expands in the turbine, which in typically accounts for roughly two thirds of the total
turn exerts torque on the shaft. power generation of the CCGT.

Again, compressor, turbine and generator sit on a Step 2: Heat transfer

common shaft. In this way, the energy transmitted to In an OCGT process, the exhaust gas is released
the shaft by the turbine is used to turn both the gen- directly to the ambient air. In a CCGT, the thermal
erator and the compressor. energy contained in the gas turbine exhaust is
transferred to a water-steam cycle in a heat recovery
2. Combined cycle gas turbine (CCGT) steam generator (HRSG).
A combined cycle gas turbine (CCGT) uses the waste
heat of the gas turbine exhaust to drive a water- Step 3: Steam turbine process
steam circuit. Hence, a CCGT is a combination of The thermal energy from the exhaust gas is used to
a gas turbine and a steam turbine. generate steam and operate the water-steam circuit.
The turbines of the CCGT can have individual gener-

39
Agora Energiewende | Flexibility in thermal power plants

Schematic view of a combined cycle gas turbine configuration Figure 12

Fuel Step 1
Air Joule cycle
Combustion
Inlet Chamber

Generator

Compressor Turbine

Exhaust Heat Recovery Steam

Generator (HRSG)
Gas (Step 2)

Water
Live steam
Pump Generator

G
Condenser
Turbine

Step 3
Exhaust steam Rankine cycle

Fichtner (2017)

ators, as depicted in Figure 12, or drive a common one, • OCGT are typically operated in pure peak load
referred to as a single shaft configuration. The steam operation. Their efficiency reaches up to 40 per-
turbine typically provides about a third of the total cent, they display high fuel cost and they require
power generation of a CCGT power plant. very low CAPEX.
• CCGTS are typically operated at a medium load.
Their efficiency reaches up to 60 percent,
Gas-fired power plants they have medium fuel costs and they require
Gas-fired power plants are usually designed to pro- low CAPEX.
vide medium to peak load to the grid, due to their
relatively high level of flexibility and to their cost Both technologies (OCGT and CCGT) can also be oper-
structure (low capital expenditure (CAPEX), high fuel ated in CHP mode.
cost). However, the operation may change in the
future depending on fuel and CO2 emission prices.

40
 STUDY | Flexibility in thermal power plants

3.1.5 Brief description of Combined Heat turbine or use a so-called back-pressure steam
and Power (CHP) turbine.
Combined heat and power, also referred to as cogen-
eration, describes the simultaneous generation of In practice it is very common to operate CCGTs, hard
electricity and useful heat. It significantly improves coal-fired plants and lignite-fired plants in CHP
the overall utilization of fuel by substantially reduc- mode in Germany for economic and environmental
ing the amount of waste heat. reasons. CHP operation depends on the existence of
heat demand by, say, district heating or process heat.
In CHP plants, partially expanded steam at medium
temperature is extracted from the steam turbine. According to the (AG Energiebilanzen, 2016), 17 per-
The thermal energy in the steam is then transferred cent of net electricity generation in Germany in 2015
to another medium in a separate network, which was provided by cogeneration plants.11 At today’s
supplies customers with heat either through district industrial power plants (serving on-site consumption
heating or for heat-intensive industrial processes of electricity and heat), almost 75 percent of electric-
(process heat). ity is generated through gas-fired units. A OCGT in
combination with a HRSG is commonly used when
Figure 13 shows a simplified schematic view of a dis- high temperature process heat is required.
trict heating system supplied with heat from a water-
steam circuit. Flexibility of cogeneration power plants
Typically, cogeneration plants are partly operated in a
Theoretically, all thermal power plants can be heat controlled mode. To ensure a constant supply of
operated in cogeneration mode. With OCGTs, thermal energy to their customers, they are required
a HRSG can be used to generate process heat using to run at a certain load (“must-run capacity”), mak-
hot flue gases (up to 550 °C). CCGTs, hard coal- ing them rather inflexible. This means that they are
fired power plants and lignite-fired power plants
have two options: extract steam from the steam 11 This value includes so-called mini-cogeneration facilities.

Schematic view of steam extraction from a steam turbine providing heat

to a district heating system Figure 13

Hot & high pressure

steam Steam turbine

G
Low temperature and
low pressure steam Buildings

Low temperature feed District heating grid

water

Fichtner (2017)

41
Agora Energiewende | Flexibility in thermal power plants

limited in responding to changing electrical power 3.2 Operational flexibility

demands.
The section discusses the concept of operational
Large thermal energy storages can be used to reduce flexibility. For ease of reading, it is simply referred to
the inflexibility of CHP power plants. Heat produc- as “flexibility” for the remainder of the report.
tion and consumption can be partly decoupled  —  in
times of high renewable production, say. This allows
the cogeneration plant to react flexibly to changes in Flexibility
power demand. The flexibility of a power plant can be described as
its ability to adjust the net power fed into the grid,
Typical capacities for thermal heat storages range its overall bandwidth of operation and the time
from 20 MWh to 1,500 MWh and have storage required to attain stable operation when starting up
volumes of 500 to 45,000 m3. The discharge duration from a standstill.
of the different thermal energy storages vary by
size and discharge capacity. For example, a large
atmospheric thermal energy storage with a discharge The key parameters characterizing the flexibility of
capacity of 1,500 MWh and a water volume of a thermal power plant are illustrated in Figure 14:
30,000 m3 has a discharge duration of about
6 hours (Kraft, 2015). This means that the power Sections 3.2.1 to 3.2.3 below elaborate on each flexi-
plant can in principle stop generation for up to bility parameter. Section 3.2.4 describes the influence
6 hours while providing a constant heat of 250 MW of flexible operation on the lifetime costs and on the
to its consumers through the discharge of its operation and maintenance (O&M) costs of a thermal
thermal storage. power plant.

Overview of flexibility characteristics Figure 14

Minimum load
(Section 3.2.1)

Operational Start-up time

Flexibility (Section 3.2.2)

Ramp rate
(Section 3.2.3)

Fichtner (2017)

42
 STUDY | Flexibility in thermal power plants

3.2.1 Minimum load Limitations

The minimum load, PMin,describes the lowest possible The lower the load, the more difficult it is to ensure a
net power a power plant can deliver under stable stable combustion without supplemental firing.13
operating conditions. It is measured in percentage of
nominal load, % PNom. Figure 15 shows a qualitative
load curve for a power plant with key power varia- 3.2.2 Start-up time
bles. The start-up time is defined as the period from start-
ing plant operation until reaching minimum load.
In this figure, the minimum load is assumed to be The start-up time of different generation technol-
30 % of nominal power PNom. The net power PNet fed ogies vary greatly. Other factors influencing the
into the grid can range from minimum load to nom- start-up time are down time (period when the power
inal load. The range between minimum and nominal plant is out of operation) and cooling rate. Figure 16
load is called part load operation. illustrates the time for a simplified start-up.

After start-up initiation (t0), no power is fed into the

Impact on flexibility grid until t1. After t1, the net power gradually starts
The lower the minimum load, the larger the range of to increase. As mentioned above, the start-up time
generation capacity. A low minimum load can avoid is defined as the period from the start of plant oper-
expensive start-ups and shutdowns. ation (t0) until minimum load is reached (t2 ). Gener-
ally, steeper load curve slopes translate into shorter
Disadvantages start-up time. The following types of start-ups are
At minimum load, the power plant operates at lower defined according by (Gostling, 2002) for power plants:
efficiency12.

13 Supplemental firing describes the process of combusting expensive

12 A typical issue with low load operation is also its impacts on the auxiliary fuels, such as heavy oil or gas, in addition to pulverized
SOx and dust emissions. This dimension is not studied in details in coal. This stabilizes the flame in the boiler. Such fuels are usually
this report. See for example NREL (2014) for more information. required during the start-up procedure of coal-fired power plants.

Qualitative representation of a power plant load curve with key power variables Figure 15

100 PNom
PNet [% PNom ]

PNet

30 PMin

0
t [h]

Fichtner (2017)

43
Agora Energiewende | Flexibility in thermal power plants

Net power for a simplified start-up from standstill until nominal operation Figure 16

100 PNom
PNet [% PNom ]

PMin

0
t0 t1 t2
t [h]
Start-up time

Fichtner (2017)

Hot start-up: Limitations

The power plant has been out of operation for less The allowable thermal gradient in Kelvin per
than 8 hours. minute, K/min, for thick-walled components limits
the start-up time speed. The state of develop-
Warm start-up: ment of the automation can also be a limiting
The power plant has been out of operation for factor.14
between 8 and 48 hours.

Cold start-up: 3.2.3 Ramp rate

The power plant has been out of operation for more The ramp rate describes how fast a power plant can
than 48 hours. change its net power during operation. Mathemat-
ically, it can be described as a change in net power,
Generally, a cold start puts a larger strain on plant ΔPNet, per change in time, Δt.
components than a hot start due to the greater
temperature differences that occur during the ΔPNet
Ramp rate = ——————
start-up. Δt

Normally the ramp rate is specified in MW per min-

Impact on flexibility ute, MW/min, or in percentage of nominal load per
The shorter the start-up time, the quicker a power minute, % PNom/min. In general, ramp rates heavily
plant can reach minimum load. depend on generation technology, as will be discussed

Disadvantages 14 An increase of temperature causes thermal expansion in metals.

Faster start-up times put greater thermal stress During a cold start-up, the temperature changes with time, from initially
ambient temperature until reaching nominal operating temperature.
on plant components, thereby reducing their Temperature changes spatially as the wall thickness of the components
lifetime. vary. The different states of thermal expansion result in thermal
stress. Normally, an allowable thermal gradient with regard to time
in Kelvin per minute, K/min, is provided to keep thermal stress below
a damaging threshold.

44
 STUDY | Flexibility in thermal power plants

Qualitative depiction of a power plant load curve highlighting the ramp rate Figure 17

100
Slope = Ramp rate
PNet [% PNom ]

30

0
t [h]

Fichtner (2017)

in Section 3.3. Figure 17 presents a qualitative load 3.2.4 Implications of flexible operation on

curve over time, with the ramp rate visually inter- lifetime and operation and maintenance
preted as the slope. (O&M) cost of thermal power plants
The following section describes the impact that a
more flexible operation has on the lifetime of a ther-
Impact on flexibility mal power plant and its associated costs.
A higher ramp rate allows a power plant operator
to adjust net power more rapidly to meet changes 1. Impact of flexible operation on lifetime
in power demand. Flexible operation (high ramp rates and multiple
starts) has significant influence on the lifetime of a
Disadvantages power plant (Ziems, et al., 2012). Thick-walled com-
A rapid change in firing temperature results in ponents are especially affected by thermal stress,
thermal stress for plant components. which can be derived from ramp rates and start-ups.
Load changes of over 50 % of PNom (from 40 % to 100 %
Limitations of PNom ) and cold starts put the highest strain on
The allowable thermal stress for thick-walled these components.
components and the allowable unsymmetrical
deformations limit the ramp rate. For coal-fired However, the specific lifetime consumption depends
power plants, the storage behaviour of the steam on many parameters (change in temperature, pres-
generator, the quality of fuel used for combustion sure, etc.) and is different for each component.15 The
(which has a direct effect on temperature variation) specific influence of flexible operation and associated
and the time lag between coal milling and turbine
response can act as limiting factors.
15 The lifetime consumption is used to capture the effect of power
plant operation on the life of components. Critical processes such
as starts or load changes of over 50 % PNom are usually assigned
a specific lifetime consumption value as a percentage of the compo-
nent’s life. For example, if a start-up causes a lifetime consumption
of 0.005 % for a given component, 20,000 starts could be performed
before replacement was needed.

45
Agora Energiewende | Flexibility in thermal power plants

lifetime consumption can be calculated using detailed 0.5–1.3 $/MWh in a system without renewables to
modeling. 1.0–3.0 $/MWh in a system with 33 percent variable
renewables. To put this into perspective this amounts
Such a modeling was performed for baseline mode to an increase of approx. 2–5 percent of total variable
and a dynamic operation mode (50 more starts per operation and maintenance cost (27–28 $/MWh).
year and a ramp rate twice as high as the baseline
operation mode) for a hard coal-fired power plant in From a system perspective, these increased costs are
Rostock. The dynamic operation mode increases the relatively small compared to the fuel savings associ-
accumulated annual lifetime consumption from 0.4 % ated with wind and solar generation.
to 3.24 % (an increase by a factor of 8). To put this in
real terms, the unit would have a theoretical life- The lifetime of a plant greatly depends on external
time of 250 years in the baseline scenario and only factors (electricity price, CO2, fuel, etc.). If a compo-
31 years in the dynamic operation scenario (Ziems, nent needs to be replaced, however, significant costs
et al., 2012). (>1 million euros) arise.

In practice, frequent physical component checks 3.3 Comparison of flexibility parameters

(e.g., X-ray examination, crack testing and micro- in different generation technologies
structure examination) are necessary to verify
component health, as modeling outcomes are “only” This section compares the four thermal generation
theoretical. technologies discussed above with regard to flexibil-
ity and CO2 emissions. Once again, these technologies
In Germany, some power plant operators deliber- are:
ately push flexibility even though it reduces plant life.
In part, this has to do with the shift in energy policy →→ lignite-fired power plants;
away from coal for the next decades. This explains →→ hard coal-fired power plants;
the higher flexibility of German power plants relative →→ open cycle gas turbine (OCGT) power plants; and
to other countries. →→ combined cycle gas turbine (CCGT) power plants.

Generally, it is not possible to put lifetime consump- To ensure proper comparison, only larger generation
tion in monetary terms. The reason is that lifetime units (300 MW and more) are considered. The state of
consumption and the associated loss in revenues development also plays a critical role in the compar-
largely depend on future earnings, future plant oper- ison:
ation, future maintenance, repair strategies, and the
like. 1. Most commonly used technologies
“Most commonly used technologies” refer to typical,
2. Impact of flexible operation on O&M costs existing plant designs. Generally speaking, today’s
Lifetime consumption of thick-walled components commonly used technologies are power plants built
is not directly linked to O&M costs. The affected 10–20 years ago with a state-of-the-art design at the
components (headers, etc.) in the HP (high pressure) time.
line are typically designed to be used over the entire
lifetime of the plant (typically 40 years). Accord- 2. State-of-the-art technologies
ing to NREL (2014), more cycling of fossil-fueled “State-of-the-art technologies” describe the best
power plants in systems with high shares of varia- technology commercially available when investing
ble renewables can increase the cycling costs from in a new power plant project today.

46
 STUDY | Flexibility in thermal power plants

It should be noted that the average values for each →→ Part 2: Detailed comparison between state-of-
generation technology can vary from region the-art and most commonly used technologies
to region. A “most commonly used” design in The second part provides a more in-depth compar-
a developed industrial country such as Germany built ison of most commonly used and state-of-the-art
10–20 years ago might be more advanced generation technologies with regard to flexibility.
than a comparable power plant in a less developed
country. →→ Part 3: Comparison of three specific coal-fired
power plants
Section 3.3.1 summarizes the flexibility parameters Part three focuses on specific coal-fired power
for each generation technology. plants in Germany and Poland and compares their
flexibility parameters.
Section 3.3.2 discusses the net efficiencies and spe-
cific CO2 emissions for each technology. Part 1: General comparison of the four
technologies
3.3.1 Flexibility parameters Table 1 provides a summary of the flexibility param-
This section presents and compares the flexibility eters (minimum load, ramp-rate and start-up time)
parameters of the four generation technologies. It has of most commonly used and state-of-the-art power
three parts: plants for each generation technology (OCGT, CCGT,
hard coal- and lignite-fired power plants). The main
→→ Part 1: General comparison of the finding is that gas-fired power plants (OCGT and
four technologie CCGT) have a higher operational flexibility relative to
The first part provides a general comparison of coal-fired units. As Figure 18 shows, start-up time is
flexibility parameters. significantly shorter and ramp rates are higher than
for hard coal- and lignite-fired power plants.

Ramp rates and start-up times of thermal power plants in comparison Figure 18

OCGT CCGT CCGT hard coal-fired

(state-of- (state-of- (commonly power plant
the-art) the-art) used) (state-of-the-art)
1
PNet [% PNom ]

start
0
0 1 2
t [h]

Fichtner (2017) based on (VDE, 2012)

47
Agora Energiewende | Flexibility in thermal power plants

Comparison of most commonly used and state-of-the-art power plants for each generation
technology with regard to flexibility Table 1

Property OCGT CCGT Hard coal-fired Lignite-fired

power plant power plant

Most commonly used power plants

Minimum load [% PNom] 40–50 % 40–50 % 25–40 %a 50–60 %

Average ramp rate [% PNom per min] 8–12 % 2–4 % 1.5–4 % 1–2 %

Hot start-up time [min] or [h] 5–11 minb 60–90 min 2.5–3 h 4–6 h

Cold start-up time [min] or [h] 5–11 minc 3–4 h 5–10 h 8–10 h

State-of-the-art power plants

Minimum load [% PNom] 20–50 % 30–40 % 25e–40 %f 35g–50 %

(20 % with SCd)

Average ramp rate [% PNom per min] 10–15 % 4–8 % 3–6 % 2–6h %

Hot start-up time [min] or [h] 5–10 min i 30–40 min 80 min–2.5 h 1.25j–4 h

Cold start-up time [min] or [h] 5–10 min i 2–3 h 3–6 h 5 k–8 h

a Source: (Heinzel, Meiser, Stamatelopoulos, & Buck, 2012)

b Large heavy-duty gas turbines such as the Siemens SGT5-4000F typically have longer start-up
times. A fast start takes about 11 minutes and a normal start about 30 minutes.
c The amount of fuel that can be burned at the maximum continuous rating of the appliance multiplied by the net calo-
rific value of the fuel and expressed as megawatts thermal. The thermal input is specified by the manufacturer of a plant.
d SC (sequential combustion): Some state-of-the-art CCGT power plants are equipped with sequential com-
bustion, which enables a very low load operation without exceeding emission limits.
e See (Then, 2016)
f Minimum load: 25–30 % in “recirculation mode” and 35–40 % in “once-through mode.”
g See Boxberg “unit R”, with a minimum load of 35 %.
h See the “Belchatów II Unit 1” power plant in Poland or the Boxberg power plant in Germany, both with a ramp rate of up to 6 % Pnom.
i Large heavy-duty gas turbines such as the Siemens gas turbine SGT5-8000H typically have longer
start-up times. A fast start takes about 11 minutes and a normal start about 30 minutes.
j See the Boxberg power plant “unit R” with a start-up time (hot) of 75–85 minutes.
k See the Boxberg power plant “unit R” with a start-up time (cold) of 290–330 minutes.

Fichtner (2017) ; Original sources: (VDE, 2012), (Steck & Mauch, 2008) and (Balling, 2010). The technical data is from OEMs.

Table 1 highlights the following aspects for the This is mainly due to combustion stability issues,
most-commonly used power plants: which are more pronounced in the larger boiler
designs present in lignite-fired power plants.
Minimum load
Hard coal-fired power plants can reach the lowest Average ramp rate
minimum load with 25 percent of nominal load. Lig- In terms of average ramp rates, the OCGT configu-
nite-fired power plants, however, provide the least ration provides the greatest flexibility with 8–12 %
flexibility, with 50–60 percent of the nominal load. of nominal power per minute. The OCGT configura-

48
 STUDY | Flexibility in thermal power plants

tion can respond significantly faster than the CCGT technologies. The ramp rate of state-of-the-art coal
configuration due to the thermal inertia of the steam power plants (hard coal as well as lignite) can meet or
generator and the steam turbine (Cziesla, et al., 2013). exceed the ramp rate of most-commonly-used CCGT
gas-fired plants.
Coal-fired power plants have relatively low ramp
rates due to large component dimensions and time Start-up time
lag between an increase in fuel input and turbine The start-up time of state-of-the-art technology can
response (Cziesla, et al., 2013). be much lower than those of most-commonly used
technology, with the exception of OCGTs. In particu-
Start-up time lar, the reduction of start-up time can be as much as
Like the average ramp rate, hot start-up times vary several hours for lignite-fired coal power plants. The
greatly between technologies. Both gas turbine con- hot start-up time of new hard-coal power plants are
figurations can start significantly faster than coal- approaching those of most-commonly-used CCGTs.
fired plants.
However it must be pointed out that even for state-
For a gas turbine, the start-up time consists of the of-the-art power plants, coal-fired power plants
time required to bring the turbine into a rotary (hard coal as well as lignite) are still less flexible
movement, the time to start the ignition, the time relative to gas-fired generation units, especially with
to achieve nominal rotational speed and the time to regard to start-up time and ramp rate.
synchronize the generator.
Part 2: Detailed comparison of state-of-the-art
For coal-fired power plants, however, the start-up technologies with most commonly used
process is far more complex. It requires the operation technologies
of auxiliary systems, such as cooling pumps, fans and
burners. Additionally, it takes more time for larger 1) Minimum Load
components to reach the required temperature levels With most commonly used technologies, hard coal-
to begin operation. fired power plants can reach the lowest minimum
load with 25–40 % of PNom, as shown in Figure 19.
As for state-of-the-art power plants, Table 1 shows OCGT and CCGT both have a slightly higher minimum
that a significant improvement of flexibility can be load, ranging between 40–50 % of PNom. The most
achieved when compared to most-commonly used commonly used lignite-fired power plants have the
technologies : highest minimum load with 50–60 % of PNom.

Minimum load As Figure 19 shows, most state-of-the art technol-

The minimum load of state-of-the art power plants ogies can achieve significant improvements relative
can be reduced to 20 percent of nominal load for to most commonly used power plants. Technolog-
OCGT and down to 35 percent of nominal load for ical advancement significantly reduced the mini-
lignite. This represents a significant improvement mum load in state-of-the-art OCGT and CCGT power
relative to most commonly used technologies. plants. They reach the lowest minimum load with
20–50 % and 20–40 % (with sequential combustion)
Ramp rate of PNom respectively.
The ramp rate of most flexible state-of-the-art
power plants can be up to 2–3 times higher than Lignite-fired power plants with state-of-the-art
the ramp rate of less flexible most commonly used designs have significantly reduced minimum loads,

49
Agora Energiewende | Flexibility in thermal power plants

Comparison of power plants with most commonly used technologies and power plants
with state-of-the-art technologies for each generation type with regard to minimum load
(values based on Table 1) Figure 19

65
55
Minimum load
[% PNom]

45
35
25
15
commonly used

commonly used

commonly used

commonly used
state-of-the-art

state-of-the-art

state-of-the-art

state-of-the-art
OCGT CCGT Hard Coal Lignite
with sequential combustion

Fichtner (2017)

Comparison of power plants with most commonly used technologies and power plants with
state-of-the-art technologies for each generation type with regard to the average ramp rate
(values based on Table 1) Figure 20

16
14
Average ramp rate
[% PNom per min]

12
10
8
6
4
2
0
commonly used

commonly used

commonly used

commonly used
state-of-the-art

state-of-the-art

state-of-the-art

state-of-the-art

OCGT CCGT Hard Coal Lignite

Fichtner (2017)

50
 STUDY | Flexibility in thermal power plants

from 50–60 percent to 35–50 percent. But they still ramp rates to CCGT power plants, reaching 1.5–4 % of
provide the least flexibility with regard to minimum PNom per minute for hard coal-fired power plants
load. with most commonly used technologies, whereas
state-of-the-art power plants improved to 3–6 %
2) Ramp rate of PNom. Of all generation technologies, lignite-fired
Figure 20 compares the average ramp rate of power power plants with most commonly used technolo-
plants with most commonly used technologies and gies have the lowest average ramp rates, 1–2 % of PNom
power plants with state-of-the-art technologies. per minute. But state-of-the-art lignite-fired power
As can be seen in this figure, OCGT power plants pro- plants can ramp up significantly faster, with an
vide the highest ramp rate, reaching 8–12 % of PNom average ramp rate reaching 2–6 % PNom per minute
per minute for most commonly used power plants (versus 1–2 % for most commonly used technologies).
and 10–15 % of PNom per minute for state-of-the-art
power plants. 3a) Start-up time (hot)
Figure 21 illustrates the difference between power
The ramp rate of CCGT power plants is about two to plants with most commonly used and state-of-
four times slower than in OCGT power plants. the-art technologies with regard to hot start-up
However, the ramp rate of state-of-the-art CCGT time. In both categories, OCGT has by far the short-
(4–8 % of PNom per minute) shows significant est hot start-up time among the different genera-
improvement relative to the most commonly used tion technologies (5–11 minutes)  —  followed by CCGT,
CCGT technology (ramp rate of 2–4 % of PNom per hard coal-fired power plants and lignite-fired power
minute). Hard coal-fired power plants have similar plants.

Comparison of power plants with most commonly used technologies and power plants with
state-of-the-art technologies for each generation type with regard to start-up time (hot <8 h)
(values based on Table 1) Figure 21

400
Start-up time (hot <8 h)

350
300
250
[min]

200
150
100
50
0
commonly used

commonly used

commonly used

commonly used
state-of-the-art

state-of-the-art

state-of-the-art

state-of-the-art

OCGT CCGT Hard Coal Lignite

Fichtner (2017)

51
Agora Energiewende | Flexibility in thermal power plants

The range of hot start-up time for OCGT decreases for most commonly used technologies (5–11 minutes)
only slightly, from 5–11 minutes (most commonly and for state-of-the-art technologies (5–10 minutes),
used) to 5–10 minutes (state-of-the-art). The hot followed by CCGT, hard coal-fired power plants and
start-up time of CCGT is nearly halved between most lignite-fired power plants. The cold start-up time of
commonly used power plants (hot start-up time of CCGT improved significantly between most com-
60–90 minutes) and state-of-the-art power plants monly used power plants (180–240 minutes) and
(hot start-up time of 30–40 minutes). state-of-the-art power plants (120–180 minutes).

The hot start-up time for hard coal-fired power plants Most commonly used hard coal-fired and lignite-
improved from 150–180 minutes to 80–150 minutes fired power plants have the longest cold start-up
in the state-of-the-art design category. Lignite- time and therefore the lowest flexibility of all the
fired power plants decreased their hot start-up time generation technologies under comparison. The
considerably, from 240–360 minutes (commonly cold start-up time of hard coal-fired power plants
used) to 75–240 minutes (state-of-the-art). range between 300–600 minutes. Lignite-fired
power plants lie between 480–600 minutes and
3b) Start-up time (cold) thus tend to start slower than hard coal-fired power
Figure 22 compares power plants with most com- plants. The cold start-up time of hard coal-fired
monly used technologies and with state-of-the-art power plants with state-of-the-art design takes
technologies with regard to cold start-up time. 180–360 minutes less. State-of-the-art lignite-fired
OCGT provides the shortest cold start-up time, both power plants have a range of 300-480 minutes.

Comparison of power plants with most commonly used technologies and power plants with
state-of-the-art technologies for each generation type with regard to start-up time (cold >48 h)
(values based on Table 1) Figure 22

700
Start-up time (cold >48 h)

600
500
400
[min]

300
200
100
0
commonly used

commonly used

commonly used

commonly used
state-of-the-art

state-of-the-art

state-of-the-art

state-of-the-art

OCGT CCGT Hard Coal Lignite

Fichtner (2017)

52
 STUDY | Flexibility in thermal power plants

Part 3: Comparison of three specific coal-fired Boxberg Power Plant, Germany

power plants Boxberg is a lignite-fired power plant in the eastern
This section compares three state-of-the-art coal- part of Germany with a total installed electric capac-
fired power plants in terms of flexibility. ity of 2,575 MW. The latest unit “unit R”, completed in
2012, has an electric capacity of 675 MW. It uses the
Belchatów Power Plant (new unit), Poland latest advances in material research and in boiler and
Belchatów power station is Europe’s largest power turbine technologies (LEAG, n.d.). The minimum load
station and is listed as one of the world’s largest fos- of this new unit is 35 percent of the nominal load and
sil power stations. With a total installed capacity of has a ramp rate between 4.6–6 percent. The start-up
4,400 MW it generates almost 20 percent of the total time under hot and cold conditions range between
power output in Poland (SGS Industrial Services, 75–85 and 290–333 minutes.
2011). The new lignite-fired power unit “Belchatów
II Unit 1” with 858 MW was completed in 2011. With Table 2 provides an overview of the coal-fired power
a minimum load of 45 percent and a ramp rate of plants’ flexibility parameters.
2–6 percent it can be classified as a state-of-the-
art lignite-fired power plant. The start-up-times are Table 2 indicates that these power plants lie in the
140 minutes (hot) and 360 minutes (cold). range of state-of-the-art power plant designs.

Walsum Power Plant, Germany 3.3.2 CO2 emissions

“Unit 10” of Walsum Power Plant was completed in This section compares net efficiency and specific
2013. This new hard coal-fired unit has an installed CO2 emissions of thermal generation technologies.
capacity of 725 MW (STEAG GmbH, n.d.). The mini- Net efficiency indicates power plant operation
mum load is 35 percent of nominal load and the ramp at nominal load. Average annual net efficiency is
rate ranges between 3.5–6 percent. The start-up time lower than net efficiency, since power plants are
(hot) is 66 minutes and thus slightly shorter than the sometimes operated at part load (when net efficiency
average typical time range given in Table 1 (80 min- decreases).
utes). The start-up time (cold) is about 290 minutes.

Comparison of three state-of-the-art coal-fired power plants in Poland and Germany Table 2

Name Belchatów Walsum Boxberg

(Poland) (Germany) (Germany)

Fuel type Lignite Hard coal Lignite

Minimum load [% PNom] 45 % (45–50%)* 35 % (25–40 %) 35 % (45–50 %)

Average ramp rate [% PNom per min] 2–6 % (2–6 %) 3.5–6 % (3–6 %) 4.6–6 % (2–6 %)

Hot start-up time [min] or [h] 140 min (1.25–4 h) 66 min (80 min–2.5 h) 75–85 min (1.25–4 h)

Cold start-up time [min] or [h] 360 min (5–8 h) 290 min (3–6 h) 290–330 min (5–8 h)

* The values in italics represent the average values for state-of-the-art power plants and are based on Table 1
Fichtner (2017)

53
Agora Energiewende | Flexibility in thermal power plants

Net efficiency and specific CO2 emissions for the most commonly used generation technologies
at nominal operation  Table 3

Property OCGT CCGT Hard coal-fired Lignite-fired

power plant power plant

Net efficiency [%] 39.5 % up to 59 % 43 % 42.5 %

Fuel specific CO2 emissions 202–300 202–300 325–350 340–410

[g CO2/kWhth]

CO2 emissions of electricity 511–759 342–508 756–814 800–965

generation [g CO2/kWhel]

Fichtner (2017), Prognos (2016), INAS (2014)

CO2 emissions for each technology are determined by hard coal-fired power plants are very similar with
the specific net efficiency and specific fuel emis- regard to net efficiency but show considerable
sions. The overall life cycle CO2 emissions for each difference in specific CO2 emissions. This mostly has
fuel depend on the carbon intensity of each energy to do with the high specific CO2 emissions of lignite.
source and on the technologies used for exploration
and transportation. Table 4 summarizes the values for state-of-the-art
generation technologies.
Table 3 summarizes the values for the most commonly
used generation technologies. State-of-the-art CCGT configurations have the high-
est efficiency of all the generation technologies under
Table 3 shows that CCGT have higher net efficiency, consideration. Hard coal-fired power plants achieved
with values of up to 59 percent. Both lignite- and the greatest improvement between the two develop-

Net efficiency and specific CO2 emissions for state-of-the-art generation technologies
at nominal operation  Table 4

Property OCGT CCGT Hard coal-fired Lignite-fired

power plant power plant

Net efficiency [%] 39.7 % 60 % 46 % 43 %

Fuel specific CO2 emissions 202–300 202–300 325–350 340–410

[g CO2/kWhth]

CO2 emissions of electricity 509–756 337–500 707–761 791–953

generation [g CO2/kWhel]

Fichtner (2017), Prognos (2016), INAS (2014)

54
 STUDY | Flexibility in thermal power plants

ment stages, increasing net efficiency by 3 percent.

The net efficiency of OCGT increased only marginally,
by 0.3 percent, from the most commonly used tech-
nologies to state-of-the-art designs.

The low specific CO2 emissions from OCGT are on

account of its high efficiency and the fuel charac-
teristics of natural gas. Once again, lignite causes the
highest specific CO2 emissions.

The impact of flexible operation on efficiency and

CO2 emissions of power plants is discussed in more
details in Sections 4.2 and section 5.2.

55
Agora Energiewende | Flexibility in thermal power plants

56
 STUDY | Flexibility in thermal power plants

4. Retrofits to increase flexibility

of coal-fired power plants –
Options, potential and limitations

Structure of Chapter 4 Figure 23

Section 1: Section 2: Section 3:

Key Trade-off Potential and
components between limitations
for flexibility flexibility and of flexibility
retrofits efficiency retrofits

Fichtner (2017)

This chapter explores retrofits on key power plant But first a general definition:
components to improve flexibility. (Whenever pos-
sible, available options are presented and supported
by quantitative data.) It also discusses the trade-offs Retrofit
between flexibility and efficiency and elaborates on In the field of power plant technology, a retrofit is
the potential and limitations of flexibility retrofits. defined as a modernization or upgrade of power
plant components or subsystems. In general,
The structure of this chapter is presented a retrofit is performed as part of a major overhaul
in Figure 23. and usually requires a power plant standstill lasting
multiple weeks.

Retrofits are performed for various reasons, such

as improving plant efficiency, increasing flexibility
or extending the lifetime of components. This chapter
focuses solely on retrofits aiming to increase opera-
tional flexibility.

57
Agora Energiewende | Flexibility in thermal power plants

4.1 Key components for flexibility 3 – Control and communication system
retrofits This subsystem is the “operating system” of the power
plant and comprises all components for control and
To gain a better understanding of coal-fired power communication between subsystems. Among other
plant operation, it is helpful to look at its subsystems. things, it enables the control of the temperature and
Figure 24 shows a schematic view of a coal-fired pressure inside the boiler.
power plant divided into 20 subsystems. Each subsys-
tem fulfils a crucial role in the power plant. 5 – Oil and fuel supply for ignition
To initiate coal combustion, the air volume in the
Research has shown that retrofits on the following interior of the boiler needs to be brought to a certain
subsystems are the most effective means for increas- temperature and pressure. This is typically done
ing plant flexibility: by burning auxiliary fuels, such as oil or gas. This

20 subsystems in a coal-fired power plant Figure 24

16) 2) Energy derivation and

3) Control and 19) Auxiliary 8) Boiler Steam auxiliary power supply
communication system and 20) Ancilliary turbine
system heavy machinery systems system

G
~
5) Oil and ignition fuel supply

17)
Generator 1) Grid &
system distriburion
system

9) Coal mills, 18) Cooling water

coal bunker system
4) Conventional and allocation
fuel supply system

14) Chemical flue 10) Combustion

gas treatment air system
system
12) Electrical 13) Denox
precipitator 15) Steam,
water and 7) Water supply and
gas cycle disposal

11) Flue gas system 6) Ash handling plant

and slag-removing
device

Legend of Material Flows: Circulation Water Solid fuels

Steam Other substances (ash)
Untreated Water Non-flammable gases
Air Oil

Klumpp (2009)

58
 STUDY | Flexibility in thermal power plants

subsystem plays a crucial role during the start-up of There are two main methods to remove the ash
coal-fired power plants. produced by coal combustion: slag tap and dry ash
removal.
8 – Boiler
The main task of the boiler is to turn feed water into With dry ash removal, combustion takes place in
steam. Therefore, it also referred to as the steam a furnace with small dimensions and little cool-
generator. Today, steam is typically generated in ing. In slag-tap furnaces, the temperature is higher
a single-pass, once-through boiler often in tower than the melting temperature of the ash. This pro-
construction design (see Figure 25). The radiative duces molten ash, which is diverted and then released
heating surface (the inner boiler surface, shown red as fusion granulate.
in Figure 25) have pipes mounted inside, where the
water evaporates. The convection tube banks, where In the case of dry ash removal, ash is discharged via
the steam is overheated, are mounted vertically above the bottom hopper and by means of an electrostatic
the burner-stages. Steam temperatures are limited precipitator. The ash is swept out with the flue gas,
to 560/600 °C, allowing conventional ferritic tube where it remains in a dry, solid state.
materials to be used.
Due to their high combustion temperatures, slag-
tap furnaces produce high emissions of thermal NOx,
Schematic illustration of a boiler which despite combustion modification measures
in a tower construction Figure 25 can barely be kept below the emission limits defined
by federal environmental regulations16. The more
ambitious the limits, the tighter the constraints on
115 m firing temperature.

The advantage of the slag-tap firing system is that

the ash can be recovered completely as marketable
slag, a common industrial building material.

The burners are operated using pulverized coal from

the coal mills (subsystem 9 in Figure 24). The advan-
Convection tubes tage of pulverized coal is that it burns similarly to gas
(Strauss, 2016). This technology can be used for most
types of coal. The pulverized coal is transported via
an air stream from the coal mills to the burner. In the
Burner stage
burner, the coal is combusted together with air from
the coal mill (primary air) and additional air for com-
Hot flue gases bustion (secondary air).

There are two types of burner constructions:

jet burners and vortex burners. Jet burners are
most commonly used in a tangential firing config-

Fichtner (2017) based on Strauss (2016) & Scheffknecht (2005)

16 see (NREL 2014) for more information on the
impact of cycling on NOx emissions.

59
Agora Energiewende | Flexibility in thermal power plants

Tangential firing burner configuration with 4 burners on one stage  Figure 26

Top view Simulation

Individual
burner

Strauss (2016) & Heinzel, et al. (2012)

uration (Figure 26). The primary and secondary air For hard coal-fired power plants, the vertical roller
stream mix due to the velocity difference of the two or bowl mill is used to produce pulverized coal. Since
jets. the water content of hard coal (2–7 %) is significantly
lower than lignite (45–60 %), the drying process is
The colors in the simulation in the figure above rep- much less energy intensive. A hot air stream is suffi-
resent temperature. The highest temperatures (red cient enough to drive out remaining water. After the
and orange) are achieved in the air stream where the milling process, coal dust is blown into the boiler.
pulverized coal combusts.
In general, tube mills are more flexible than beater-
Vortex burners feed in the air concentrically. The wheel mills. Tube mills use a rotating cylinder to
mixture of both air streams is influenced by their pulverize the coal. Bowl mills are considered the most
velocity difference. Unlike jet burners, a vortex burner inert of the three types (Scheffknecht, 2005).
can be installed as a single burner in the boiler, which
permits a more unconstrained design (Strauss, 2016). 15 - Steam, water and gas cycle
This subsystem is closely linked with the boiler and
9 – Coal mills, coal bunker and allocation system the steam turbine. Its functions include the pre-heat-
In this subsystem, the raw coal is milled into pulver- ing of the feed water.
ized coal (PC).
Before the feed water enters the boiler it is pre-
For lignite-fired power plants, the coal is milled via heated by different heat exchangers. Usually, this is
beater-wheel mills and dried with hot flue gas (up to done by extracting hot steam from the steam turbine
1,000 °C). and cooling it in the heat exchangers. The tempera-

60
 STUDY | Flexibility in thermal power plants

ture of the feed water increases as it flows through Fire instability can occur for different reasons, such
the exchangers. Pre-heating the feed water is an as sudden changes in firing rate or fuel quality,
important process in optimizing power plant effi- improper fuel-air ratios or uneven flows of pulver-
ciency. ized coal (Sarkar, 2015). In low load operations, fire
can become instable when the hot flue gases do not
16 - Steam turbine completely ignite the inflowing pulverized coal.
The steam turbine converts pressure and thermal
energy into mechanical — i.e. rotational — energy and Under those constraints, the minimum load of hard-
is situated in the machinery hall. Unlike gas turbines, coal power plants with dry ash removal is typically
which rotate in a hot flue gas flow, steam turbines 25–40 % of PNom. For slag-tap firing systems, the
rotate in vaporized water. minimum load is around 40 % because the temper-
ature required to maintain the flow of liquid ash is
In large power plants, steam turbine systems contain higher. For lignite-fired power plants, it is between
high-pressure, intermediate-pressure and low-pres- 40–50 % because lignite must be dried during milling.
sure sections. The steam turbine is mounted on
a common shaft connected to the generator (subsys- Several retrofit options exist for overcoming many of
tem 17 in Figure 24), which transforms mechanical these technical limitations:
energy into electrical energy.
Option 1: Indirect firing
Options for improving operational flexibility are pre- Indirect firing (IF) involves the use of a pulverized
sented below. coal (PC) storage facility, a so-called dust bunker, sit-
uated between coal mills and burners. This decouples
4.1.1 Options for decreasing minimum load the direct supply chain between mills and burners
Before proceeding, it is useful to recall why a de- (Figure 27).
creased minimum load benefits power plant opera-
tion. Decoupling has the following effects:

a. stable fire at low load because of faster response to

Reasons for decreasing minimum load fire instabilities;
Decreasing minimum load is beneficial because b. reduced net power feed-in because coal mill opera-
it provides a larger range of generation capacity. tion is held at nominal levels during low loads; and
This helps plant operators maintain operation when c. higher ramp rate during operation thanks to
power demand is low and avoid expensive start-up reduced time lag between mills and burners.
and shutdown procedures. From a system stand-
point, reducing the minimum load of conventional Effects a and b help decrease the minimum load that
power plants allows a greater share of renewables is fed into the grid. Effect c will be discussed in Sec-
by avoiding potential curtailment. tion 4.1.3.

With direct firing (DF), mills must reduce their load

Reducing the minimum load in hard coal-fired power during low load power plant operation (at night,
plants is subject to certain technical limitations. say). With indirect firing, mills can run at nominal
According to (Heinzel, et al., 2012) these limitations load even if the pulverized coal is not immediately
are fire stability (see explanation below), flame con- required because it can be stored in the dust bunker.
trol, ignition, unburned coal and CO emissions. This allows the auxiliary power needed for milling to

61
Agora Energiewende | Flexibility in thermal power plants

Schematic illustration of coal supply to burner Figure 27

After retrofit: Indirect Firing

Raw Coal Dust

Burner
coal mill bunker

Before retrofit: Direct Firing

Fichtner (2017)

ramp up when the load is low (at night, say). By main- Direct firing requires coal mills to operate under
taining nominal mill operation when load is low, this part load during periods of low power plant load.
reduces the net power fed into the grid, as illustrated The resulting drop in efficiency leads to an increase
in Figure 28. in specific CO2 emissions. In indirect firing, coal mills
maintain nominal load and can run at optimal effi-
The figure shows the qualitative reduction in mini- ciency. This translates into a reduction of specific CO2
mum load fed into the grid PMin for indirect and direct emissions.
firing configurations. The difference between PMin,IF
and PMin,DF results from the difference in milling According to (Jeschke, et al., 2012), implementing
power ΔPMills. indirect firing in combination with a staged vortex
burner retrofit can decrease the minimum stable

Net grid feed-in for indirect (IF) and direct firing (DF) configurations Figure 28

100
PNet [% PNom ]

30 PMin,DF
∆PMills
PMin,IF
0
t [h]

Fichtner (2017)

62
 STUDY | Flexibility in thermal power plants

firing rate from 25–30 % to 10 %. Indirect firing is inside the boiler (Figure 29). In the direct firing con-
also applicable to other burners, such as jet burners. figuration, reducing the net power of a power plant
In general, firing rate and net power are proportional. requires the burners and the coal mills to both run
A reduction of the firing rate therefore leads to a sim- at part load. At a certain firing rate, the fire becomes
ilar reduction of minimum load. Another advantage of instable, requiring the power plant controller to limit
reaching a low stable fire is that the need for ignition the low load operation in order to avoid damaging
fuels, such as oil or gas, can be reduced by 95 %. pressure pulses that can occur inside the boiler. The
fire stability typically represents the lowest threshold
Option 2: Switching from two-mill to single-mill for low load operation.
operation
Coal mills grind lignite or hard coal to pulverized At a certain net power output, it is feasible to shut
coal (PC). The PC is transported via air stream (pri- down some of the mills (typically 4 to 6 in number)
mary air) to the burners, where it is then combusted and have the remaining mills operate closer to their

Coal mill and burner arrangement of a boiler in tangential firing configuration

with four burner stages (single-mill operation) Figure 29

Burner stage 4

Burner stage 3

Burner stage 2

Burner stage 1

Coal mill

Fichtner (2017) based on Heinzel, et al. (2012)

63
Agora Energiewende | Flexibility in thermal power plants

design point. Since coal mills typically supply a single Relative to two-mill operation, single-mill opera-
burner stage with PC, turning off a mill leads to tion can significantly reduce the minimum load while
a boiler operation with a reduced number of burning increasing operational stability. The limitations for
stages. minimum load operation are shifted from the boiler
side (mainly flame stability) to other sections of the
Figure 29 shows a technical drawing of a mill/burner power plant, such as the water-steam circuit.18
arrangement in a boiler of a hard coal-fired power
plant (Heinzel, et al., 2012). The purple crosses mark Experiments at Heilbronn Unit 7 and Bexbach, both
mills that are turned off. The pink arrows illustrate hard coal-fired power plants in Germany (start of
the flow of air conveying the pulverized coal from operation in 1985 and 1983, respectively) (Heinzel,
mill 4 to the burner stage 4, where it is blown into the et al., 2012) have shown that a reduction of minimum
interior of the boiler (combustion chamber). load to 12.5 % PNom was possible by switching from
a two- to a single-mill operation. In fact, it was found
In single-mill operation, only the highest burner stage that single-mill operation achieved greater fire sta-
is operated for the benefits of releasing heat “higher” bility than two-mill operation since both the burner
in the boiler (Figure 30).17 stage and the mill can operate closer to their design
point. Since the end of 2011, single-mill operation is
being used commercially in both power plants.
17 According to (Heinzel, et al., 2012), operating the highest burner stage
in combination with a large air excess compensates for lower steam and
flue gas temperatures by creating a colder flame and more flue gas. At Bexbach (721 MW PNom) the minimum load was
reduced from 170 MW (two-mill operation) to 90 MW
in single-mill operation (12.5 % PNom). It was found
Operation of four burner stages (left) that the process variables were more stable in sin-
in comparison to a single burner stage gle-mill than in two-mill operation. For proper
in single-mill mode (right)  Figure 30 monitoring of burner stage 4 in single-mill opera-
tion, additional flame controllers had to be installed.
No auxiliary firing is required for stable operation
Flue gas

DeNOx

at 90 MW net power. However, to increase the load

LuVo

from 90 MW (ramp up), auxiliary firing with oil is

necessary (Heinzel, et al., 2012).
Overheater

At Heilbronn Unit 7 (800 MW PNom) single-mill

operation achieved a reduction of the minimum
load from 200 MW (two-mill operation) to 100 MW
(12.5 % PNom). The fire was found to be more sta-
heating surfaces:
Evaporator

ble than in two-mill operation. Two additional flame

Radiative

controllers were installed on each burner stage to

achieve improved flame monitoring. A substantial
task for implementing single-mill operation was the

18 On the boiler side, the lower load requires switching from varia-
ble pressure to minimum pressure operation. To maintain appro-
priate pressure levels in the water-steam circuit, steam flow at
Fichtner (2017) based on Heinzel, et al. (2012) the mid-pressure turbine inlet can be held back. DeNOX (flue
gas denitrification) operation remained unproblematic.

64
 STUDY | Flexibility in thermal power plants

adjustment of control technology and boiler safety Example 1:

(Heinzel, et al., 2012). Weisweiler lignite-fired power plant, Germany
Unit G and H at Weisweiler, each with 600 MW, PNom,
Option 3: Upgrade of control system in received a digital control system and other plant
combination with plant engineering upgrades engineering retrofits.
Control technology plays a crucial part in power plant
operation. It allows navigation between different According to (Frohne, 2012) & (RWE Power AG,
loads and ensures stable operation by adjusting all 2012), the retrofit at Unit G decreased minimum
relevant process variables. In the context of coal- load by 170 MW and resulted in an increase of ramp
fired power plants, the control system monitors and rate. (For more, see Section 4.1.3.) The total cost of
controls the temperature and pressure inside the the retrofit was 60 million euros. The retrofit at
boiler, the feed-water mass flow in the water-steam Unit H reduced the minimum load from 400 MW to
circuit, the load point of the coal mills and the turbine 290 MW. The total cost amounted to 65 million euros
valve positions. (RWE Power AG, 2011).

An upgrade of the control system improves precision, Figure 31 shows the difference from before and after
reliability and speed of control. For instance, the retrofit in terms of nominal power, minimum
it allows operation closer to the material limitations power and ramp rate of Units G and H at Weisweiler.
of important components, such as the boiler. This can
mean operation at very high temperatures without The minimum power is significantly lower pre-retro-
significantly reducing material lifespan. An upgrade fit, while the ramp rate (slope of the curve) increases.
of the control system is usually combined with plant
engineering upgrades, such as retrofits of the boiler or Example 2:
the turbine or other components. Lignite-fired power plant Neurath, Germany
According to (Schulze & Hoffmann, 2013), an upgrade
to the control system and plant engineering compo-
nents including the boiler, condenser and the cool-

Load curves for pre-retrofit and post-retrofit of Unit G and H at Weisweiler  Figure 31

100
PNet [% PNom ]

PMin,Pre Retro pre-retrofit

post-retrofit
PMin,Post Retro

0
t [h]

Fichtner (2017) based on Frohne (2012)

65
Agora Energiewende | Flexibility in thermal power plants

ing tower at Unit E (600 MW PNom) of the Neurath According to (Michels, 2016), operating the dried lig-
lignite-fired power plant decreased the minimum nite ignition burner for auxiliary firing reduced the
load from 440 to 290 MW. Additionally, efficiency minimum load from 36 % to 26 % PNom.
improved by 0.6 % and the ramp rate increased. (See
Section 4.1.3.) The total cost of this retrofit amounted Another advantage of operating the burner with dried
to 70 million euros (RWE Power AG, 2011). lignite is that it reduces the need for high quality and
expensive fuels, such as heavy oil or gas. Accord-
Based on (Schulze & Hoffmann, 2013), a retrofit of ing to (FDBR, 2012) auxiliary firing can additionally
the control system and plant engineering at Neurath improve the overall efficiency of the power plant.
Unit D (600 MW PNom) decreased minimum power
from 440 MW to 260 MW and increased the ramp Option 5: Thermal energy storage for feed
rate. (See Section 4.1.3.) In addition, the retrofit water pre-heating
allowed positive and negative control power to be Thermal energy storage can be used to store heat and
delivered to the market. Previously, only negative release it at a later point in time. It presents an inter-
primary control power could be achieved (by throt- esting concept for influencing net power without
tling the turbine inlet valve). Now, condensate stop changing the firing rate in the boiler (subsystem 15 in
operation enables positive primary control power as Figure 24).
well. Unit D also gained prequalification for 75 MW of
secondary control power. In a typical configuration, the feed water is pre-
heated in a heat exchanger with steam extracted
Option 4: Auxiliary firing with dried lignite from the steam turbine. This increases the overall
ignition burner efficiency of the power plant and offsets the loss of
Auxiliary firing describes the process of stabilizing turbine power caused by the steam extraction.
the fire in the boiler by combusting auxiliary fuels,
such as heavy oil or gas, in addition to the PC-fired Releasing or absorbing heat to or from the feed
main burners. This allows for an overall lowering of water has, therefore, a direct influence on net power
the stable firing rate in the boiler. Auxiliary firing because it influences the amount of steam extracted
can also be used for rapid increases to the firing rate, from the turbine.
which have a positive influence on the ramp rate.
(See Section 4.1.3.) The operation of a storage system consists of
charging and discharging cycles.
Since fire stability in the boiler usually limits the
minimum load, auxiliary firing can support the mini- Charging is done by transferring heat from the feed
mum load reduction. water to the storage system. To maintain a con-
stant feed water temperature, more steam must be
As part of a research project at the Jänschwalde lig- extracted from the steam turbine, leading to a reduc-
nite-fired power plant, the ignition burners (combus- tion in net power. Crucial for reducing the minimum
ting heavy oil and gas) were replaced with a type that load is that charging take place during periods when
runs on dried lignite. The finely milled dried lignite is loads are low (at night, say).
carried through the burner by an air stream. Plasma
(induced by microwaves) ignites the lignite at the Figure 32 shows how charging a thermal energy stor-
lance near the burner exit. The goal of the project was age (TES) system can reduce minimum power.
to use the ignition burner also for auxiliary firing.

66
 STUDY | Flexibility in thermal power plants

Charging a thermal energy storage system and its influence on net power fed into the grid Figure 32

100
PNet [% PNom ]

30 PMin
PMin,TES
0
t [h]
charging

Fichtner (2017)

The minimum load achieved during the charging pro-

cess is lower than in the normal configuration. It is Reasons to decrease start-up time
important to note that the reduction of net power has Power plant operators want to decrease start-up
no influence on the firing rate in the boiler. time because it enables a more rapid response
to power demand. Start-up procedures are complex
According to (Schmidt & Schuele, 2013), the use and expensive since they usually require auxiliary
of a hot water storage system that can operate for fuel, such as oil or gas, during the ignition period.
2–8 hours can reduce the minimum power fed
into the grid by 5–10 percent (Schmidt & Schuele,
2013). Discharging the stored thermal energy can There are various technical factors that limit the
temporarily increase net power by 5 percent without reduction of start-up time. Thick-walled components
increasing the firing rate. allow higher operating parameters (steam tempera-
ture and pressure, say), which increase efficiency. But
Smaller hot water tanks (operation for less than quick temperature changes in thick-walled compo-
30 minutes) can be used to improve the ramp rate nents induce thermal stress, which acts as a limiting
(Schmidt & Schuele, 2013). Section 4.1.3 will factor for the start-up time. With “thinner” compo-
discuss options for improving the ramp rate in nent designs, flexibility can be higher but efficiency
more detail. is usually lower.

4.1.2 Options for decreasing start-up time Several options exist for shortening start-up time in
Before presenting retrofit options, let’s first recall power plants that have not been built with flexibility
why a decreased start-up time benefits power plant in mind. Four of these retrofit options are described
operators. in the following section: repowering, predictive boiler
operation, advanced turbine design and enhanced
turbine start-up.

67
Agora Energiewende | Flexibility in thermal power plants

Simplified illustration of a coal-fired power plant with a gas turbine employed for feed water
pre-heating Figure 33

Existing power plant

Repowering

Gas turbine(s)
HP recovery preheater LP recovery preheater

Repowering cycle

The black lines connecting the grey components respresent the water-steam circuit.
Fichtner (2017) based on Jeschke, et al. (2012).

Option 1: Repowering An increase in gas turbine power output directly

Repowering involves placing a gas turbine upstream increases the heat transfer to the feed water of the
of the water-steam circuit in coal-fired power plants. water-steam circuit. This reduces the steam extrac-
The thermal energy in the exhaust stream of the gas tion needed from the steam turbine, which translates
turbine is then transferred to the feed water via heat into higher steam turbine output. (See Section 4.1.1,
exchangers (see Figure 33). Option 5.)

Gas turbines can ramp up significantly faster than In terms of start-up performance, repowering is
coal-fired power plants. For hot starts, state-of- especially helpful because the gas turbine can pro-
the-art OCGT designs require about 5–10 minutes, vide power while the water-steam circuit is still
whereas hard coal-fired power plants take from heating up. In 2006 and 2007, two gas turbines with
80 minutes to 2.5 hours. According to (Jeschke, et al., 190 MW of net power each were installed in Units G
2012), repowering increases the gross output of the and H at Weisweiler. Pre-heating the feed water with
power plant, improves total efficiency and start-up gas turbine exhaust increased the net power (of the
performance and increases ramp rate. (For more, see coal-fired unit) by 80 MW (+ 6.6 % PNom), because less
Section 4.1.3.) steam had to be extracted from the steam turbine.

68
 STUDY | Flexibility in thermal power plants

The total investment amounted to 150 million euros The parameters include among others fuel costs and
(RWE Power AG, n.d.). thermal stress on thick-walled components (Franke &
Weidmann, 2008) .
In sum, repowering
BoilerMax has been installed in several E.ON power
• increases the net power of the coal-fired power plants in Germany, including the 450 MW coal-fired
plant; unit Zolling 5. The start-up time is shortened by
• improves flexibility; and 33 percent, as can be seen in Figure 35.
• increases efficiency, which leads to lower
specific CO2 emissions. Once installed in the control system, BoilerMax
allows plant operators to shorten plant start-up time.
Option 2: Optimized control systems A shorter start-up time normally implies higher
Predictive controller solutions such as ABB’s thermal stress for the materials. The tool also provides
BoilerMax are used for the online optimization of plant operators with the opportunity to choose
start-ups. Such control systems use dynamic optimi- between different start-up options, allowing them
zation, which beat the performance of conventional to adjust the specific start-up to the current market
control systems. BoilerMax optimises several param- situation.
eters to shorten boiler start-up time (Figure 34).

Basic principle of BoilerMax application Figure 34

Optimization Goals

Real boiler

Optimizer

Inputs Iteration
Boiler optimized
model setpoint
and inputs
Optimized
Process variables
inputs and
variables Cost function
= Minimum!

Nonlinear Model Predictive Control with

cyclic tracking to the real process

ABB AG Power Systems Division (2013)

69
Agora Energiewende | Flexibility in thermal power plants

Comparison of two start-ups at Zolling, one with BoilerMax and one without BoilerMax Figure 35

300
Generator capacity [MW]

250

200

150

100

50

0
0 10 20 30 40 50 60
Time [h]
15 minutes faster start-up

Generator capacity with “Boiler Max”

Generator capacity without “Boiler Max”

Franke & Weidmann (2008)

Option 3: Thin-walled components/special be rather thick- or thin-walled. With thick-walled

turbine design components, steam temperature and pressure can be
The quicker a start-up, the faster the temperature of higher than power plants that have thin-walled com-
thick-walled components rises. Thermal stress on ponents. This increases efficiency but decreases flex-
thick-walled components of the boiler system, such ibility. With a “thin-walled” component design, power
as headers, limits temperature fluctuations.19 plant flexibility is higher but efficiency decreases
because steam temperature and pressure are lower.
For quicker start-ups, the wall thickness of thick-
walled components needs to be reduced (Alstom, Siemens new steam turbine, the SST5-6000,
2013). This can be achieved by using high-grade is designed for supercritical steam power plants
materials such as ferritic martensitic steel P92, which with a power range between 600 and 1,200 MW per
can better cope with thermal stress, or by using spe- unit. A typical set consists of a four-casing arrange-
cial designs. ment with separate high pressure, intermediate
pressure and two low pressure turbines. Smaller
When designing a power plant, future operators units (<500 MW), like the SST-5000, have also been
need to evaluate if they want the power plant to be designed with higher operational flexibility.
more flexible or more efficient. Plant operators need
to decide if they want power plant components to High parameter values (temperature and pressure)
for steam require a specially designed turbine like
19 A header is a component in which steam is collected after
the SST5-6000. The high pressure cylinder of the
having passed through the overheating phase in the boiler. SST5-6000 is achieved using a bypass cooling

70
 STUDY | Flexibility in thermal power plants

system. The turbine has a barrel-type construction 4.1.3 Options for increasing ramp rate
with an inner casing. A small amount of cooling steam Recall why an increased ramp rate benefits power
passes through radial bores into a small annulus plant operation.
between the inner and outer casings. The cooling
steam is led through the inner casing, reducing the
surface temperature. Lower surface temperatures Reasons to increase the ramp rate
reduce creep stress and protect the inner surface Power plant operators are interested in increasing
of the outer casing. In this way, the wall-thickness ramp rates because it allows dynamic adjustments
of the outer casing can be reduced for faster heat-ups to net power. This is especially important in power
and better start-up performance. systems with rising shares of renewables.

Such a turbine was installed in the Lünen power sta-

tion, a hard coal-fired power plant in Germany with The previous two sections presented several retrofit
750 MW of installed capacity. The total costs of this options for reducing minimum load (Section 4.1.1) and
new power plant were 1.4 billion euros. Lünen started start-up time (Section 4.1.2). As this sections shows,
commercial operation at the beginning of 2014 several of those retrofit options also have a positive
(Trianel, n.d.). impact on power plant ramp rate.

Option 4: “New” turbine start Option 1: Repowering

In most cases, steam turbine start-ups require The repowering option, described in Section 4.1.2.,
a steam temperature that is higher than the metal has important implications for the ramp rate. Once
temperature. Due to its mass, the steam turbine cools again, repowering involves installing a gas turbine in
down fairly slowly. If the power plant has been out of a coal-fired power plant upstream of the water-steam
operation for only a couple of hours, the restart must circuit. Heat exchangers transfer the thermal energy
be delayed until the steam temperature reaches the in the exhaust stream from the gas turbine to the feed
turbine temperature. water.

In the past, steam turbine start-ups followed the Usually, the ramp rate is limited by the allowable
static performance curves of the boiler and did not thermal stress for thick-walled components. Addi-
take ramp rates into account. As a result, the “hot” tional limitations are caused by the fuel quality
turbine hindered overall hot start performance. and the time lag between coal milling and turbine
response present in the direct firing configuration.
To solve this problem, a new dynamic approach was
introduced: allow “cold” steam to enter the steam In a normal coal-fired power plant, burning coal
turbine as quickly as possible after shutdown. This provides the only heat source for the water-steam
enables the turbine to start with the boiler while it’s circuit. With the repowering option, a second heat
still ramping up. This approach can reduce the hot source can be used to pre-heat the feed water. This
start-up time by 15 minutes (Quinkertz, et al., 2008). makes it possible to achieve a greater change in heat
input per time, which translates into a faster ramp
rate.

Figure 36 depicts the influence of the gas turbine on

net power output. It shows the difference between

71
Agora Energiewende | Flexibility in thermal power plants

Influence of repowering on net power and ramp rate  Figure 36

Pr = Pc + PGT + ∆RR × ∆t
Pr (t) Legend
PNet [% PNom ]

PGT: Net power of gas turbine

Pr: Net power with repowering
Pc (t) Pc: Net power without repowering
(conventional)
∆RR: Difference in ramp rate
PGT between “conventional” and
“with repowering” configuration
in MW/min
∆t: Difference in time (t1 –t0)
t0 t [h] t1
With repowering Conventional

Fichtner (2017)

gas turbine repowering and a conventional configu- (Schulze & Hoffmann, 2013) report that the ramp rate
ration. increased by 6 MW/min to 12 MW/min (2 % PNom) as
part of the Unit D retrofit at Neurath.
With repowering, the ramp rate is greater (hence the
steeper slope) because an additional heat source is Option 3: Reducing the wall thickness
available to pre-heat the feed water. This means that of key components
after an equivalent period of ramping up a larger net As discussed earlier, the wall thickness of compo-
power can be reached with the turbine than with the nents is an important parameter because it influences
traditional configuration. The difference in net power the allowable temperature change rate. The tempera-
between the two configurations is given by the net ture change rate describes the change in temperature
power of the gas turbine, PGT, and the difference in per change in time at a specific location in the wall in
ramp rate, ΔRR. Kelvin per minute, K/min. Since temperature changes
induce thermal stress, each material is assigned a
Option 2: Upgrading control systems and plant maximum allowable value. Exceeding this value
engineering reduces the material’s lifespan.
This option has already been described in
Section 4.1.1. Here, the benefits of the retrofit on In general, reducing wall thickness increases the
ramp rate are presented. allowable temperature change rate. This translates
into a faster start-up by boosting the ramp rate. Wall
The retrofits at Weisweiler’s Unit G — a new digital thickness can be reduced by using superior materials
control and communication system and upgrades to or by increasing the number of specific components,
its plant engineering — not only reduced the minimum such as switching from a 2-line to a 4-line design
power; they also had a positive effect on the ramp (Jeschke, et al., 2012).
rate. According to (Frohne, 2012), the ramp rate
increased by 10 MW/min. The total retrofit at Unit G Research conducted by (Jeschke, et al., 2012) has
cost of 60 million euros. shown that using a superior material such as Alloy 617
instead of P92 allows high pressure headers with

72
 STUDY | Flexibility in thermal power plants

Influence of relative pressure on allowable temperature change rate in K/min for a high pressure
header designed with two different materials  Figure 37

Header di=180mm
10
Allowable temperature change rate

8
10 % Design
Superior material
Wall thickness 40mm, Alloy 617
6
[K/min]

7 % Design
4
Normal material
Wall thickness 52mm, P92
2

0
0 20 40 60 80 100 125
Pressure/Pnom [%]

Fichtner (2016) based on Jeschke, et al. (2012)

23 %-thinner walls. This increases the allowable tem- Option 4: Auxiliary firing with dried lignite
perature change rate by 60 percent in the load regime ignition burner in booster operation
of 50–100 percent. The option of auxiliary firing with a dried lignite
ignition burner was presented in Section 4.1.1 as
Figure 37 shows the influence of relative pressure on means for decreasing the minimum load.
the allowable temperature change rate in K/min for a
high pressure header using two different construc- The ignition burner can also be used during opera-
tion materials. tion to increase firing power and increase net power
and ramp rate. This type of operation is referred to
At 100 percent relative pressure — that is to say, as booster operation. It requires a dust bunker to be
at nominal operation — an allowable temperature independent of the inertia of the milling process (see
change rate of about 8 K/min can be achieved when Option 1 in Section 4.1.1)
using Alloy 617 at a thickness of 40 mm. The differ-
ence from using P92 at a thickness of 52 mm results Booster operation helps reduce time lag (partially
in an allowable temperature change rate of only caused by the milling process) between the rise in the
about 5 K/min. firing rate and turbine response. Normally, the lag is
around 20–60 s for hard coal-fired and 30–60 s for
According to (Jeschke, et al., 2012), the use of the lignite-fired power plants (Scheffknecht, 2005).
superior material would increase the plant’s ramp
rate by 3 percent.

73
Agora Energiewende | Flexibility in thermal power plants

4.2 Trade-offs between flexibility and range of possible net power outputs. It can also
efficiency avoid expensive and CO2-intensive shutdowns and
start-ups.
This section discusses the relationship between flex-
ibility and efficiency of coal-fired power plants. In In general, operating a thermal power plant in part
doing so, it answers a key question: load leads to lower efficiency relative to the nominal
load. A decrease of efficiency translates into an
increase of specific CO2 emissions (g CO2/kWh).
Key question: Figure 38 illustrates this effect. Three operating
“Do retrofits that aim to improve flexibility have points (OP) are depicted: the nominal OP, the mini-
a negative impact on power plant efficiency and, mum OP pre-retrofit and the OP post-retrofit.
by extension, on specific CO2 emissions?”
The efficiency continuously drops the more operation
is shifted from nominal conditions to part load.
The section tracks the flexibility parameters de- The effect of minimum load reduction is illustrated
scribed in this report: minimum load, start-up time by the shift of the minimum operating point (from the
and ramp rate. lila to the pink dot). Reducing the net power output
by about 20 percentage points (ΔPretrofit) decreases
1. Reducing minimum load efficiency by about 2–5 percentage points (η Pretrofit).
The minimum load is considered to be the most cru-
cial flexibility parameter. Reducing the minimum This effect translates into higher specific CO2 emis-
load provides the power plant operator with a wider sions at very low load. However, when operat-

Relationship between operating point and plant efficiency  Figure 38

OP minimum OP minimum OP nominal

post-retrofit pre-retrofit
ηnom

ηpre retrofit
∆ηretrofit ≈ 2–5% points
Efficiency [%]

ηpost retrofit

∆Pretrofit ≈ 20 % points

Ppost retrofit Ppre retrofit Pnom

Power [MW]

Fichtner (2017)

74
 STUDY | Flexibility in thermal power plants

ing at a very low load, expensive and CO2 intensive Summary

shutdowns and start-ups can be avoided. In addi- Retrofit measures do not have a negative effect on
tion, start-ups put strain on the components and efficiency.21 In many cases, retrofits to increase flexi-
reduce their life span. For instance, a hot start-up bility improved plant efficiency. (See Options 1, 3, 4 of
at a 750 MW hard coal-fired power plant requires Section 4.1.1 and Option 1 of Section 4.1.2.)
approximately 1,820 MWh of thermal energy. This
is about the same quantity required to operate the However, lowering the minimum load can reduce
power plant at nominal load for approximately an the efficiency of the power plant at very low load,
entire hour. The fuel needed for the start-up trans- increasing specific CO2 emissions at low load oper-
lates into roughly 620 tons of CO2 emissions. The ating points. To measure this effect fully, CO2 emis-
associated fuel cost amounts to around 15,000 euros sions must be assessed over the power plant’s entire
(at a coal price of about 30 euros per ton and exclud- operation instead of focusing on the lowest operat-
ing the cost for CO2 certificates). It should be noted ing points. All in all, the flexibility gained by thermal
that a retrofit that decreases the minimum load has power plants outweighs in most cases the drawbacks
no effect on higher operating loads. of CO2 emissions at low operating points, and this
advantage will only grow as the share of renewables
Given that the penetration of renewables such as increases. These effects will be discussed in more
wind and PV will continue to rise, fossil-fired power details in section 5.2.
plants will be needed to respond quickly to changing
power demand. From this perspective, it can be better
to maintain low load operation than to shut down,
since even a hot start for state-of-the-art hard coal-
fired power plants takes between 80 minutes and
2.5 hours and leads to significant CO2 emissions.20

2. Reducing the start-up time

For each of the options reviewed, start-up time
reduction measures were found to have no effect on
efficiency.

3. Increasing ramp rate

For the options reviewed, increasing ramp rate had
no negative effect on efficiency. In fact, repowering
and other measures actually improved overall plant
efficiency.

21 The designers of new power plants face, however, a conflict

between flexibility and efficiency. Achieving high efficiency at
20 In a case-by-case evaluation, the effect of decreasing mini- nominal load means generating high-temperature and high-pres-
mum load on a single plant can increase absolute CO2 emissions sure steam. Components such as headers must have a certain
due to increased usage and improved market competitiveness thickness to handle these conditions. This reduces the allowable
post-retrofit. This is discussed in more details in section 5.2.3. temperature change rate, reducing power plant flexibility.

75
Agora Energiewende | Flexibility in thermal power plants

4.3 Potential and limitations of flexibility 1. Reducing minimum load

retrofits The increasing volatility of feed-in from renewable
energy sources leads to more frequent start-ups
Flexibility retrofits are an important way of modify- and shutdowns of coal-fired plants and other con-
ing coal-fired power plants for increasingly volatile ventional power stations (Balling, 2010). Tradition-
power demand. ally, coal-fired power plants, especially lignite-fired
ones, have been designed for base load operation.
This section assesses the potential and limitations A more flexible operation schedule puts more strain
of flexibility retrofits for coal-fired power plants. on components and necessitates more start-ups,
Table 5 summarizes the retrofit options discussed in which are energy- and CO2-intensive (Section 4.2).
Section 4.1 and shows the key flexibility parameters
that improved as a result of their implementation. Minimum load reduction retrofits have clear poten-
tial. They can reduce the number of start-ups and
Minimum load reduction, decreasing start-up time shutdowns by allowing the power plant to stay
and increasing ramp rate are discussed separately in online at very low loads. Even though efficiency in
this section. The material is then summarized at the part load, especially when loads are very low, is lower
end. relative to operation at nominal load, CO2 emissions
can be avoided because of the reduced number of

Summary of analysed retrofit options, their effect on flexibility parameters and their limitations Table 5

Option Minimum Start-up Ramp Limitations

load time rate

Indirect Firing   Fire stability

Switching from two-mill to single-mill operation  Water-steam circuit

Control system and plant engineering upgrade   Fire stability/

thermal stress

Auxiliary firing with dried lignite ignition burner   Fire stability and
boiler design

Thermal energy storage for feed water pre-heating  N/A

Repowering   N/A

Optimized control system  Thermal stress

Thin-walled components/special turbine design  Mechanical and

thermal stresses

“New” turbine start  Turbine design

Reducing wall thickness of key components  Mechanical and

thermal stresses

Fichtner (2017)

76
 STUDY | Flexibility in thermal power plants

start-ups.22 Furthermore, flexible operation yields caused by the water-steam circuit described in Sec-
a higher penetration rate of renewables without tion 4.1.1.
compromising grid stability. This, in turn, reduces
the CO2 emissions of the power system generally Table 6 provides a summary of the potential and limi-
(see section 5.2). tations of each retrofit option. For a detailed descrip-
tion of all options for minimum load reduction, please
The limitations of minimum load reduction are refer to Section 4.1.1.
usually posed by the fire stability in the boiler, as
described in Section 4.1.1. Currently, the minimum 2. Reducing start-up time
load of state-of-the-art hard coal- and lignite-fired Due to the increased share of fluctuating power
power plants lies between 25–40 % and 35–50 %. feed-in from renewables, the number of start-ups
In case of extremely low load operation (such as the and shutdowns in coal-fired plants and other con-
12 % of PNom achieved with single-mill operation in ventional power stations is expected to rise.
Bexbach and Heilbronn Unit 7), the limitations are
Start-ups and shutdowns are energy intensive,
require expensive ignition fuels (such as heavy oil
22 In a case-by-case evaluation, the effect of a decreasing minimum
load for a single plant can increase absolute CO2-emissions because of
and gas) and put a high level of strain on components.
increased usage and improved market competitiveness after retrofit. Decreasing the start-up time reduces the need for

Potential and limitations of retrofit options for reducing minimum load Table 6

Option Potential Limitations

Indirect firing A reduction of minimum stable firing rate from 25–30 % Fire stability
to 10 % (with burner retrofit) was achieved (Jeschke, et al.,
2012). This leads to a corresponding reduction in
minimum load.

Switching from two-mill On average, these retrofits reduced minimum load from 23 % Water-steam circuit
to single-mill operation to 12 % of PNom (Heinzel, et al., 2012).

Control system and plant On average, these retrofits reduced minimum load from Fire stability
engineering upgrades 71 % to 47 % of PNom . The total cost of the retrofits at units G
and H at Weisweiler amounted to 60 and 65 million eu-
ros, respectively. At Neurath the total cost of the retrofit at
unit E amounted to 70 million euros (RWE Power AG, 2012),
(Frohne, 2012), (Schulze & Hoffmann, 2013).

Auxiliary firing with This option reduced the minimum load from 36 % to 26 % Fire stability
dried lignite ignition of PNom (Michels, 2016).
burner

Thermal energy storage A reduction of minimum load by 5–10 % employing a hot N/A
for feed water water storage system that can operate for 2–8 hours is
pre-heating deemed realistic (Schmidt & Schuele, 2013).

Fichtner (2017)

77
Agora Energiewende | Flexibility in thermal power plants

Potential and limitations of retrofit options for reducing start-up time Table 7

Option Potential Limitations

Repowering In general, repowering has a positive influence on start-up N/A

behaviour, as the gas turbine can ramp significantly faster
(Jeschke, et al., 2012).
The implementation of two gas turbines at Units G and H at
Weisweiler with 190 MW each (31 % of PNom of the coal unit)
increased net power by 80 MW (+6.6 PNom) per unit. The total
investment amounted to 150 million euros.

Optimized control system This retrofit reduced start-up time by 33 % (15 minutes) Thermal stress
(Franke & Weidmann, 2008).

Thin-walled components/ Utilizing superior materials allows for thinner walls Mechanical and
special turbine design in components such as headers. Thinner walls allow faster thermal stresses
start-ups.

“New” turbine start This retrofit reduced the hot start-up by 15 minutes Turbine design
(Quinkertz, et al., 2008).

Fichtner (2017)

Potential and limitations of retrofit options for increasing ramp rate Table 8

Option Potential Limitations

Repowering Repowering has been shown to increase ramp rates. Modern N/A
power plants achieve ramp rates of up to 6 % PNom/min.

Control system and plant These retrofit options increased ramp rates by +6 MW/min Thermal stress
engineering upgrade (600 MW PNom) and +10 MW/min (600 MW PNom) at Neurath
and Weisweiler (Frohne, 2012), (Schulze & Hoffmann, 2013).
The total cost of the retrofits are given in Table 6.

Reducing the wall This retrofit increased the ramp rate by 3 % (Jeschke, et al., Mechanical and
thickness of key 2012). thermal stresses
components

Auxiliary firing with dried Increasing the firing rate at constant boiler load with booster Boiler design,
lignite ignition burner in operation has potential for rapidly increasing net power booster operation
booster operation (Michels, 2016).

Fichtner (2017)

78
 STUDY | Flexibility in thermal power plants

those fuels because a stable fire with pulverized coal The major limitations for increased ramp rates are
can be achieved faster. In addition, plant operators caused by thermal and mechanical stress during
can reduce their response time to power demand in ramping. This stress reduces component life and must
case of plant standstill. be accounted for during component design. Gener-
ally, there is a trade-off between thick-walled design
The limitations are mainly caused by the allowable for high efficiency and thin-walled design that per-
thermal and mechanical stress for thick-walled com- mits a higher temperature change rate and therefore
ponents such as headers. Table 7 provides a summary higher ramp rates.
of the potential and limitations of each retrofit option.
For a detailed description of all options for start-up Table 8 provides a summary of the potential and
time reduction, see Section 4.1.2. limitations of each retrofit option. For a detailed
description of all options for increasing ramp rate,
3. Increasing ramp rate see Section 4.1.3.
Increasing ramp rate is particularly important for
grid stability given increasing shares of fluctuating Conclusion
renewable feed-in. The faster generating units can Retrofits for increasing flexibility were performed
adjust their net power, the easier it becomes for (grid) at numerous coal-fired power plants in recent years.
operators to balance supply and demand. These retrofits significantly improved the flexibility

Major coal-fired plant subsystems where retrofits were performed to improve flexibility  Figure 39

8) Boiler 16)
Steam
3) Control and turbine
communication system
system
5) Oil and ignition fuel supply

9) Coal mills,
coal bunker
and allocation
system
15) Steam,
water and
gas cycle

Fichtner (2017) based on Klumpp (2009)

79
Agora Energiewende | Flexibility in thermal power plants

of coal-fired power plants with regard to minimum Few retrofit options portrayed in Section 4.1 provide
load, start-up time and ramp rate. Besides improving information about financial expenditures. In terms
flexibility, the retrofits mostly had a positive influ- of economic viability, each retrofit has to be analysed
ence on plant efficiency, which lowered specific CO2 on a per plant basis. Generally, it is not possible to say
emissions. whether a retrofit will be economically viable without
knowing the role of the power plant within the elec-
Figure 39 summarizes the major subsystems where tricity mix, within the electricity market and within
retrofits were performed to improve flexibility. the country-specific energy road map. This dimen-
sion will be further assessed in the next section.
Most retrofits can be implemented independently of
coal type or ash removal system. The main limitations
to flexibility improvements are caused by boiler
fire stability and by the allowable thermal stress
on components. But meaningful improvements can
nevertheless be attained within the boundaries of
these limitations.

80
Impact of flexibility on
Power Plant Profitability
and CO2 Emissions
and
Country Profiles
South Africa and Poland
WORK PACKAGE 3

WRITTEN BY
Prognos AG
Europäisches Zentrum für Wirtschaftsforschung
und Strategieberatung
Goethestraße 25
10623 Berlin
Telephone: +49 (0)30 52 00 59-200
Fax: +49 (0)30 52 00 59-201
www.prognos.com

Contributing authors:
F. Ess
Telephone: +41 (0)61 32 73-401
Email: florian.ess@prognos.com

F. Peter
Agora Energiewende | Flexibility in thermal power plants

82
 STUDY | Flexibility in thermal power plants

5. The Impact of Flexibility on Power Plant

Profitability and CO2 Emissions

5.1 Flexibility impacts on power plant If plants have to stay in the market (e.g. to provide
operations system services), more flexibility has direct economic
value for the operator. Moreover, switching off a
As discussed in section 2, power systems with sig- power plant entails start-up costs. Therefore, a trade-
nificant shares of renewable generation require off exists between avoiding losses from negative
more flexibility to cope with fluctuating generation. prices and the costs associated with start-up.
If markets are adequately designed, flexibility needs
are reflected in electricity prices at the wholesale The following example illustrates the revenue effects
level. of increased coal power plant flexibility. It assumes
the plant is selling electricity in a marginal-cost-
The structure and functioning of electricity markets based day-ahead market. We additionally assume a
varies from country to country. Electricity markets power system with a significant share of renewables
generally comprise long-term (derivative) markets, but with a considerable volume of conventional
day-ahead markets and intraday markets. These generation from thermal power plants.23
markets segments are complemented by markets
and arrangements for ancillary services (i.e. in order Figure 40 illustrates two coal power plants with dif-
to maintain system stability in real-time). Flexible ferent flexibility characteristics but the same effi-
generation capacities are able to earn revenues, ciency standards. The solid line represents a coal
depending on their specific characteristics, in day- power plant without retrofitting and limited flexibil-
ahead and intraday markets as well as in markets ity. In comparison, the dashed line represents a coal
for ancillary services. However, day-ahead markets power plant with retrofitting and improved flexibility
currently account for the majority of the volume of characteristics, namely higher ramp-rates and lower
all market segments and have the greatest impact on minimum load. Because of high shares of renewable
power plant operations and revenues. generation, the power plants face periods of low and
even negative electricity prices.
Most day-ahead markets are currently based on a
marginal-cost approach. Since renewables have low Table 9 shows the characteristics of a typical coal
or almost zero marginal costs, electricity prices tend fired power station, constructed during the 1970s in
to be significantly lower when renewable generation Europe, with and without increased flexibility fol-
is high (due to the so-called Merit-Order Effect dis- lowing retrofitting. The assumptions for the illustra-
cussed in section 2). tive CCGT plant, constructed in the 1990s, is required
for the later analysis of CO2 emissions (see sec-
Taking this into consideration, an increasing share tion 5.2).
of renewables and low residual load will lead to more
times with low or even negative electricity prices at
the wholesale level. Conventional power plants are
thus encouraged to avoid operation during times with 23 Additional revenues for power plants from increased flexibil-
ity can also be derived from intraday markets and balancing power
negative prices or when prices fall below the plant’s markets. However, day ahead markets are usually responsible for
marginal operating costs in order to limit losses. more than 80 % of the revenues of a coal fired power station.

83
Agora Energiewende | Flexibility in thermal power plants

Plant parameters and market environment for the following illustrative examples Table 9

Plant specification Hard coal Hard coal CCGT

Limited Increased
flexibility flexibility

Unit nominal capacity 600 MW 600 MW 600 MW

Minimum load in % of nominal capacity 40 % 25 % 40 %

Minimum load in MW 240 MW 150 MW 240 MW

Net efficiency at nominal load 40 % 40 % 52 %

Net efficiency at minimum load 34.5 % 31 % 40 %

Start-up-costs in euro/MW 80 80 40

Specific CO2 emissions of fuel in g/kWhth 330 330 202

Variable operation costs in euro/MWhel 2.0 2.0 1.0

Start-up-time hot start in h 2 2 1

Market environment

Fuel price in euro/MWhth 10 10 15.8

CO2 price in euro/tonne 10 10 10

Marginal generation costs in euro/MWhel 35.3 35.3 35.3

Assumptions and calculations from Prognos

Figures 41 to 44 illustrate the economic effects of a)  Inflexible generation

more flexible operation. To assess power plant eco- Figure 41 illustrates the operation of an inflexible coal
nomics, we consider profit margins, total generation power plant. Due to its limited flexibility in “must
costs and specific generation costs during an illustra- run” operation, the plant has to stay in the market and
tive time span of 48 hours with a typical hourly price experiences losses during times with low or negative
formation for markets with large shares of renewable prices. The following example was calculated for
energy.24 a power plant based on the parameters and market
environment summarised in Table 9.

Because the minimum load of the plant is limited

to 40 percent, it only realises a profit margin of
46,800 euros, and suffers losses during times of low
24 Profit margin equals total generation costs minus total earnings
from the electricity sales. The total generation costs include or negative prices. Specific generation costs reach
the marginal costs of operation and the costs for starting the power 36.70 euros/MWh.
plant, fixed costs are not considered. The specific generation
costs are derived from the ratio of the total generation costs
and the electricity produced in that observed 48 hours.

84
 STUDY | Flexibility in thermal power plants

Hard coal power plant operation before and after retrofitting with lower minimum load,
increased ramp rates and reduced start-up time in a 48-hour example period  Figure 40

80 700
Improved
ramp rate
Hourly electricity price [EUR/MWh]

60 600

Plant operation [MW]

40 Reduced 500
start-up-
time
20 400

0 300

−20 200

−40 Reduced 100

minimum load
−60 0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47
[Hours of observed time period]

Day ahead spot market price, left axis Flexible plant operation after retrofit
Inflexible plant operation, right axis

Prognos (2017)

Hard coal power plant in must-run operation in a 48 hour example period Figure 41

80 700
Hourly electricity price [EUR/MWh]

60 600
Plant operation [MW]

40 500

20 400

0 300

−20 200

−40 100

−60 0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47
[Hours of observed time period]
Profit margin of operation: 46,800 Euro
Day ahead spot market price, left axis Total generation costs: 740,400 Euro
Inflexible plant operation, right axis Spec. generation costs: 36.7 Euro/MWh

Prognos (2017)

85
Agora Energiewende | Flexibility in thermal power plants

Hard coal power plant with temporary shut-down in a 48 hour example period  Figure 42

80 700
Hourly electricity price [EUR/MWh]

60 600

Plant operation [MW]

40 500

20 400

0 300

−20 200

−40 100

−60 0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47
[Hours of observed time period]
Profit margin of operation: 84,900 Euro
Day ahead spot market price, left axis Total generation costs: 621,900 Euro
Inflexible plant operation, right axis Spec. generation costs: 42.5 Euro/MWh

Prognos (2017)

Hard coal power plant with lowered minimum load and increased ramp rates
in must-run operation in a 48 hour example period Figure 43

80 700
Hourly electricity price [EUR/MWh]

60 600
Plant operation [MW]

40 500

20 400

0 300

−20 200

−40 100

−60 0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47
[Hours of observed time period]
Profit margin of operation: 116,100 Euro
Day ahead spot market price, left axis Total generation costs: 726,900 Euro
Inflexible plant operation, right axis Spec. generation costs: 36.7 Euro/MWh

Prognos (2017)

86
 STUDY | Flexibility in thermal power plants

b)  No must-run but limited flexibility the two starts can be avoided. The overall genera-
If the plant is able to shift to a more flexible mode tion costs are lower compared to the must-run case
of operation, the first possible approach would be to because overall less fuel is used during times of mini-
avoid negative prices and shut down temporarily mum load operation even when considering the lower
during times with negative prices. However, the efficiency in low load operation. Such an operation
plant loses part of its earnings due to shut-down and pattern could also be the result of measures to opti-
start-up times. Figure 42 illustrates the same power mise the market. This would be the case, for exam-
plant in a more flexible mode of operation with tem- ple, if the losses incurred from negative prices do not
porary shut-down during times of negative prices. exceed the costs of an additional start.

Because losses during times of negative prices can be d)  Flexible operation without must-run
avoided, the profit margin increases to 84,900 euros, Figure 44 shows the optimal dispatch of a retrofitted
while specific generation costs also increase to power station when no must-run scheme is enforced.
42.50 euros/MWh due to additional start-up costs. The reduced minimum load mitigates losses during
The trade-off between avoiding losses from negative times of negative prices. The increased ramp rate
prices and reduced revenues during times of start-up and the reduced start-up time leads to more flexible
and shut-down highlights the benefits of operation operation compared to a plant with weaker flexibility
at lower minimum load levels and of improved ramp characteristics. The profit margin (122,160 euros)
rates. is the highest of the analysed cases, but the gap
gradually decreases, and is rather small compared
c) Higher operational flexibility with must-run to flexible operation under must-run conditions
condition (116,100 euros).
Furthermore, some conventional plants have to stay
in operation because of their relevance for sys- As can be seen from this example, the decision to run
tem services or heat supply (“must-run” conditions). a plant using a flexible mode of operation depends
In this situation, reducing the minimum load is a on the earnings associated with more flexible opera-
key solution for optimising power plant earnings tion. Therefore, to allow power plant operators to fully
while limiting losses. Reducing minimum load can be harness the benefits of flexibility, market conditions
achieved with a range of retrofit measures, which are have to be designed adequately (see subsection 5.3).
described in section 4. Figure 43 illustrates the case
of a coal power plant that is able to reduce its mini- From this analysis, some preliminary conclusions
mum load to 25 percent of its nominal capacity while can be drawn: When implemented in a market
also increasing its ramp rate. environment with high shares of renewables and
wholesale markets based on marginal costs, increas-
As Figure 43 shows, in must-run operation the total ing the flexibility of a thermal power plant improves
profit margin is 116,100 euros, a figure that is consid- the economic situation of the plant, as compared to
erably higher than profits before retrofitting because inflexible operation.
the plant is able to generate additional earnings dur-
ing some hours after the price drop. In comparison →→ Reducing minimum load is the measure with the
to Figure 41 (with higher must-run operation), the most positive profitability impact for a thermal
plant is also able to limit its losses in times of negative power plant in most cases.
prices because of its ability to operate with a reduced
minimum load. The specific generation costs are →→ The question whether a specific flexibility invest-
lower compared to the case with two starts because ment is profitable or not cannot be answered in

87
Agora Energiewende | Flexibility in thermal power plants

Hard coal power plant with lowered minimum load and increased ramp rates and
shorter start-up time in a 48 hour example period Figure 44

80 700
Hourly electricity price [EUR/MWh]

60 600

Plant operation [MW]

40 500

20 400

0 300

−20 200

−40 100

−60 0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47
[Hours of observed time period]
Profit margin of operation: 122,200 Euro
Day ahead spot market price, left axis Total generation costs: 672,800 Euro
Inflexible plant operation, right axis Spec. generation costs: 41.5 Euro/MWh

Prognos (2017)

general. Specific plant parameters and market The CO2 emissions of a power plant are crucially
environments (e.g. age of the plant, renewable determined by the type of fuel used. A proper
shares, general market design, remuneration approach for measuring emissions is to assess
options for flexibility) require a case-by-case the overall life-cycle greenhouse gas emissions of
determination. the fuel in question. Those emissions depends on
the type of fuel, extraction techniques, and sup-
5.2 Effects on CO2 emissions ply routes (see 5.2.1). The emissions released specif-
ically by the power plant depends on its efficiency
The flexible operation of coal power plants due to (the higher the efficiency, the lower the emissions).
an increased share of renewables also influences Furthermore, this efficiency varies when the power
plant-specific CO2 emissions (since power plants face plant is operated at partial loads. This aspect is dis-
lower full-load hours and are more often operated cussed in section 5.2.2. Finally, in order to compare
at partial loads). In general, coal power plants produce the emissions of different technologies (e.g. flexible
more CO2 emissions per unit of output compared coal versus CCGTs gas power plants), the technologies
to other forms of conventional power generation must be compared under similar dispatch conditions.
(e.g. natural gas power plants). However, the key An illustrative example is given in section 5.2.3.
question is whether the flexible operation of coal
power plants contributes to an overall reduction in
CO2 emissions in the economic and political environ-
ment of a specific country.

88
 STUDY | Flexibility in thermal power plants

Specific CO2 emissions for a range of fuels Table 10

Fuel Natural gas Hard coal Lignite

Range of specific 202–300 325–350 340–410

emissions [gCO2/kWth]

Lower limit specification Pipeline gas Bituminous coal Pulverised lignite

Upper limit specification Shale gas Anthracite Raw lignite

Prognos (2017)

5.2.1 Life-cycle emissions of different fuels 5.2.2 Effect of partial loads on CO2 emissions
The greenhouse gas emissions of power plants are In section 4.2, we discussed the relationship
not only a consequence of burning fuel (whether coal, between partial load operation and the efficiency
natural gas or oil), but depends also on the overall life of a power plant. The efficiency of a power plant
cycle emissions of each specific fuel. Overall life cycle (as a percentage) indicates how much electric
emissions depend on the following aspects: energy (kWh_electric) is produced from the total
• exploration and extraction technology, energy content of the fuel (kWh_thermal). The rate of
• fuel processing and transport, efficiency varies depending on the operational mode
• use of the fuel (e.g. power generation) and of the power plant. It is highest at the plant’s nominal
post-production processes. load and decreases when the plant operates at partial
loads. This leads to an increase in the specific CO2
Depending on these parameters, the CO2_eq content of emissions (gCO2/kWh) of the power plant at low load
the fuel can vary significantly, as shown in table 10. levels, as illustrated in the following figure. It must be
As can be seen, lignite and hard coal have in general noted, however, that this efficiency drop only occurs
higher life-time greenhouse gas (GHG) emissions during partial load operation. It does not represent
than natural gas. However, natural gas has a broader the average efficiency of the plant over the full year
range of associated GHG emissions content, varying (which is likely to be much closer to the efficiency
from 200 to 300 g CO2_eq/kWhth, depending on the at nominal load).
types of gas and extraction techniques (shale gas,
LNG, pipeline gas, etc.). The CO2_eq content of shale gas As can be seen in Figure 45, the net efficiency of
is about 50 percent higher than that of pipeline gas, a typical older coal power plant (40 percent) at nomi-
positioning shale gas close, but still below, bitumi- nal load is considerably lower than the net efficiency
nous hard coal (325 g CO2_eq/kWh_thermal). The CO2 of a CCGT (52 percent). This implies that the specific
content of pipeline gas is, however, far below that CO2 emissions for the coal plant are considerably
of coal (both hard coal and lignite). higher at nominal loads. However, the efficiency of
a CCGT falls much more significantly than the effi-
ciency of a coal power plant when it operates at very
low load levels (in this example, minus 12 percentage
points for the CCGT versus 5.5 percentage points for
the coal power plant).

89
Agora Energiewende | Flexibility in thermal power plants

Relationship between plant output and efficiency of hard coal and CCGT gas power plants
(600 MW nominal power) at different operating points (illustrative) Figure 45

60

efficiency at different operating points [%]

52 %
50

40 % 40%
40
35%
31% 30

20

10
minimum minimum
nominal load before load after
load retrofit retrofit
0
600 550 500 450 400 350 300 250 200 150 100
operating point of a 600 MW power plant [MW]

gas power plant hard coal power plant

Agora Energiewende based on assumptions from Table 9

5.2.3 Comparing CO2 emissions of different than CCGT gas power plants.26 In order to meaning-
technologies under similar dispatch fully compare the CO2 emissions of different power
plants, we need to assess their operation under simi-
In systems with an increasing share of renewables, lar dispatch conditions, but with different flexibility
the yearly utilisation hours of coal power plants is parameters, while also taking into account variation
reduced, moving from pure baseload operation (above in efficiency as a function of the load at any given
7,000 hours) to more mid-merit operation (between time.
4,000 and 7,000 hours).25 This can reduces the overall
emissions of the power plant (since it produces less Considering the above, we conducted a compari-
power). This development makes coal power plants son of conventional power plants using different
competitive with CCGT gas power plants. There- fuels. In the following example, the CO2 emissions
fore, a key question is whether coal power plants from a coal power plant (using hard coal) are com-
under flexible operation emit more or less emissions pared with the CO2 emissions of a CCGT power plant.

26 As shown in table 10, the specific emissions of OCGT power plants are in

the same range as those of coal power plants. A more detailed compar-
ison of the overall emissions released by these two technologies is not
25 In reserve operation schemes (e.g. strategic reserves) the annual utili- particularly relevant in the present context, however, as OCGTs have
zation of coal fired power stations can drop even further. However, the a rather different function in the power system. As peak power plants,
need for flexible plant characteristics still exists with such schemes. their utilization rates are limited to several hundred hours a year.

90
 STUDY | Flexibility in thermal power plants

We considered different operational modes, both with →→ (b) Without must-run (two stops), but with limited
or without must-run levels. For the coal power plants flexible operation, the CO2 emissions of the coal
we also considered two operational modes, without power plant are reduced to 15.4 kt. Each start-up
retrofitting (limited flexibility) and after retrofitting procedure is emissions-intensive, significantly
(increased flexibility). Table 9 shows the technical increasing the CO2 emissions per kWh of electric-
parameters and other general assumptions for this ity produced. However, since the plant is offline for
specific example. Because the marginal costs of both several hours, cumulative emissions are lower.
plants are equal within the chosen framework, CCGT
plant operation mirrors coal plant dispatch, but has →→ (c) In must-run but flexible operation (the must-
a shorter start-up time and faster ramp rate (see fig- run level is reduced accordingly to 150 MW), the
ure 46). coal power plant emits 17.0 kt, compared to 7 kt
(pipeline gas) or 10.4 kt (shale gas) for the CCGT
Given the market conditions shown in Figure 46, unit.
we obtain the following results over a 48 hour period:
→→ (d) After retrofitting (which enables increased
→→ (a) In must-run operation (without retrofitting), the ramp rates, lower minimum loads and reduced
cumulative CO2 emissions of the coal power plant start-up times), the coal power station generates
are 17.4 kt, whereas the emissions of the CCGT more electricity in the example 48 hour period
plant are 9.0 kt (pipeline gas) or 13.3 kt (shale gas). (16,200 MWh), almost equalling the output of the

CCGT operation versus hard-coal plant operation in a 48 hour example period  Figure 46

80 700
Hourly electricity price [EUR/MWh]

60 600
Plant operation [MW]
40 500

20 400

0 300

−20 a) 200
c)
−40 100
d) b)
−60 0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47
[Hours of observed time period]
Coal plant operation:
a) Before retrofit must run
b) Before retrofit 2 stops Day ahead spot market price, left axis
c) After retrofit must run Coal power plant output, right axis
d) After retrofit 2 stops CCGT power plant output, right axis

Prognos (2017)

91
Agora Energiewende | Flexibility in thermal power plants

CO2 emissions of CCGT and hard coal power plants under similar dispatch conditions
but with different flexibility features during 2 example days  Figure 47

20

18

16

14

12
[kt CO2]

10

2 hard coal gas

0
inflexible coal inflexible coal flexible coal flexible coal CCGT without CCGT with
power plant but without with lower without must-run must-run
with high must-run must-run must-run
must-run

Agora Energiewende based on Prognos

CO2 emissions for CCGT and hard coal fired power stations in different operational modes
(data from Figure 47) Table 11

Plant type and operation mode Electricity CO2 emissions in Specific CO2
production tonnes emissions
in MWh in g/kWhel

Hard coal no retrofit must-run 20,160 17,369 862

Hard coal retrofit must-run 19,800 17,054 861

CCGT must-run using shale gas 21,600 13,336 617

CCGT must-run using pipeline gas 21,600 8,980 416

Hard coal no retrofit 2 stops 14,640 15,432 1,054

Hard coal retrofit 2 stops 16,200 16,280 1,004

CCGT 2 stops using shale gas 16,560 10,405 628

CCGT 2 stops using pipeline gas 16,560 7,007 423

Prognos (2017)

92
 STUDY | Flexibility in thermal power plants

CCGT (16,560 MWh). Overall emissions compared →→ In specific cases, increasing the flexibility of a coal
to the non-retrofitted plant are higher (15.4 versus power plant may lead to higher overall emissions.
16.2 kt), but emissions per kWh are lower due This can happen if part load operation avoids to
to reduced fuel use in the start-up procedure. stop a plant during periods of non-profitable oper-
However, overall emissions are lower than that ation, without being compensated by avoiding the
of an inflexible coal power plant with a must-run CO2-intensive start-up processes. This under-
operational mode. scores the need for effective CO2 abatement policy
that encourages plant operators to consider emis-
Although the flexible operation of the coal power sions when making operational decisions.
plant reduces its overall CO2 emissions, the emissions
produced by a CCGT plant operating under similar →→ In power systems dominated by coal generation, a
circumstances are always clearly lower. However, significant share of coal power plants are needed
when the CCGT is fuelled with natural gas that has to deliver system services and therefore operate
high lifecycle CO2 emissions, the difference in over- under must-run conditions. In such a system, the
all emissions between the CCGT and hard coal power flexible operation of coal power plants will have a
plants becomes smaller. significantly positive effect on the overall emis-
sions of the power plant fleet.
Considering the foregoing, some initial conclusions
can be drawn: While natural gas power plants generally cause lower
CO2 emissions than coal power plants, shifting from
→→ Power generation technologies have to be consid- coal to natural gas in certain countries may not be
ered under similar dispatch conditions in order to a viable option, particularly if the country is highly
compare cumulative CO2 emissions. dependent on coal. Indeed, when coal power domi-
nates the market, established economic and political
→→ Lifecycle emissions depend on type of fuel and interests may prevent such a transition. Yet techni-
associated exploration and transportation technol- cal path dependencies are also an important hurdle,
ogies. Therefore, at the plant level, specific case- as tremendous investments in natural gas infrastruc-
by-case evaluations have to be carried out. ture may be needed to use natural gas as a bridge
technology on the road to a fully decarbonised power
→→ In general, coal fired power generation always system. By the same token, building new gas-based
leads to more CO2 emissions compared to the use infrastructure as an interim solution could lead to
of natural gas, even when the use of shale gas is new path dependencies, thus undermining the tran-
considered. sition to a fully decarbonized system in the long-run.
In such countries, increasing the share of renewables
→→ Under must-run conditions, decreased minimum while simultaneously encouraging the flexible oper-
load levels lead to significantly lower CO2 emissions ation of existing coal plants is likely to be the most
for all types of fuels. viable political and economic strategy.

→→ Without must-run, the overall level of emissions

can drop significantly, as the power plant stops
generating during several hours. However, the
specific emissions (g CO2/kWh) of the power plant
increase significantly, as start-up processes are
CO2-intensive.

93
Agora Energiewende | Flexibility in thermal power plants

5.3 Market design requirements →→ Wholesale market: increasing RES shares trans-
to enhance the flexible operation form the residual load curve, thus placing increased
of thermal power plants flexibility requirements on conventional power
plants. Moreover, RES can decrease profitability
The development of renewables has become one due to the Merit-Order Effect.
of the key driver for decarbonising energy systems.
Enhancing the flexibility of power systems is there- →→ Balancing market: RES can increase balancing
fore crucial for integrating higher shares of variable demand.
renewable energy in a cost efficient and reliable way.
Against this backdrop, the power market needs to →→ Congestion management: RES can increase redis-
incentivise rather than hampers flexibility. Specifi- patch measures.
cally, the power market must be designed to encour-
age the full exploitation of technical potentials for These market segments are strongly interdependent.
increasing flexibility. Accordingly, inefficiency in one market segment can
undermine efficiency in other segments, hampering
Regulatory and market arrangements that provide overall flexibility, as the following example makes
clear price signals for the further development of clear: A coal power plant in Germany with a net
renewables are increasingly important in countries capacity of 500 MW, a minimal load of 40 percent
seeking to incorporate larger RES shares. Exten- (200 MW), generation costs of 15 EUR/MWh and hot
sive attention has been devoted to the interrelation- start time of 150 minutes plans to provide 50 MW in
ships between market design and flexibility. The IEA the market for negative secondary balancing power.
has identified three market-design challenges for Negative balancing power is activated if real-time
the remuneration of flexibility (cf. Figure 48). These generation exceeds demand. Generation units typi-
challenges relate to (a) the capital intensive nature of cally provide negative balancing power by reducing
renewables, (b) the limited predictability and varia- their generation output. In the German balancing
bility of renewable output and (c) the fact that gener- market, the regulations for secondary balancing
ation is decentralised. power require balancing power to be fully activated
within five minutes. Furthermore, the market design
The aim of this subsection is not to explore this dis- stipulates:
cussion in detail, but rather to increase awareness for
this topic by giving examples in which market design →→ the contracted capacity must be available for a
can incentivise flexibility.27 period of seven days;

In section 5.2, we discussed how renewables impact →→ seven days should typically pass between the end
different aspects of the power system, placing new of an auction round – so-called “gate closure time”
requirements on the operation of existing thermal – and real time; and
power plants. The market segments impacted by RES
are: →→ two products (with a 12h duration) can be chosen:
peak and off-peak.

27 For further reading and a more detailed discussion, we recom-

mend several studies recently published by Agora Energiewende: Thus, if a power plant wants to provide negative
Power Market Operations and System Reliability (2014); secondary balancing power, it has to provide the
The Power Market Pentagon (2016); Refining Short-Term
Electricity Markets to Enhance Flexibility (2016); and
capacity for seven days. Moreover, since the start
The Integration Costs of Wind and Solar Power (2015). time (150 minutes) exceeds the required activation

94
 STUDY | Flexibility in thermal power plants

Main challenges for the remuneration of flexibility  Figure 48

Limited predictability Decentralized and

Capital intensity
and variability scattered generation

→ Cost recovery: investment → Price volatility → Coordination between

incentives • More volatile prices generation and grids
• Adequate investment signals • Product definition • Increased investment
• Implications for the design of (e.g., peak/off-peak) looses demand requires new
energy markets, capacity relevance approach to TSO and DSO
markets, support schemes regulation
→ Spot market design
• Locational price signals for
→ Cost of capital: optimal risk • Reduced gate closure centralized & decentralized
allocation • Higher frequency generators needed
• Exposure to risk, including • Both day-ahead and
policy risk, is a fundamental → Prosumers
intra-day
factor determining total • Retail prices becomes
system costs if the system is → Assurance of system stability investment signal
capital-intensive • Need for new ancillary • Base for taxes and grid fee
• Trade-off between policy services products, e.g. erodes
flexibility and regulatory risk providing system inertia • Many small producers need
• Redesign ancillary services access to wholesale markets
to allow renewable participa-
tion

IEA (2016)

period for balancing capacity (5 minutes), the power However, this market design becomes inefficient
plant must be active in the wholesale market (day- in a system with a high share of renewables, as it
ahead market) in order to provide balancing power. discourages flexibility. Over a longer period with
low power demand and very high shares of RES,
The market regulations mentioned above were wholesale prices on the day-ahead market can easily
established in a market environment with almost fall below the actual generation costs of coal power
no renewable energy production. Within these (15 EUR/MWh). During such times, keeping a coal
boundaries, the system is reasonable and efficient: power plant running is not efficient. However, due
On the one hand, early gate closure (7 days before to its balancing obligations, the power plant has to
real time) and long contracting periods (7 days) offer contribute 250 MW to the market (200 MW minimal
higher planning security for grid operators. On the load plus 50 MW negative balancing regulation). This
other hand, this regulatory arrangement incentivises can result in the curtailment of renewable generation.
baseload capacity to run 24/7. In our example, the Moreover, this must-run capacity increases the
coal power plant would be encouraged to run baseload flexibility demands placed on the remaining power
in the day ahead market and reduce its capacity if system assets.
negative balancing power is requested, due to its low
marginal generation costs. In this way, it can be more efficient to provide nega-
tive balancing power with other assets such as wind
power. In order to do so, however, the market design

95
Agora Energiewende | Flexibility in thermal power plants

must be refined with shorter contracting periods, enables buying and selling power up to 45 minutes
shorter product durations (e.g. 4 hours instead of before delivery. In contrast to the day-ahead market,
12 hours) and later gate closure (i.e. closer to real- power also can be traded in schedules of 15/30 min-
time). Furthermore, freeing the coal power plant from utes instead of hourly schedules. Here, again, liquid
its must-run balancing obligations would allow it to intraday markets have effects on balancing markets:
act more flexibly on the day-ahead market. On the one hand, late gate closure reduces fore-
casting errors for renewable energy and therefore
As described above, the design of the balancing decreases balancing requirements. On the other hand,
market can have substantial impacts on day-ahead 15-minute products reduce the balancing demand
market dispatch and thus on the flexible operation of by diminishing the so-called schedule leaps. This
power plants. In addition to the day-ahead market, interrelationship will be explained in the following
most countries have additionally introduced a second sub-section.
short-term wholesale market with later gate closure
and shorter products. This so-called intraday market

Balancing demand due to schedule leaps (hourly and quarter hourly)  Figure 49

(a) (b)

Saw-tooth
pattern
[MW]

Time Time
07:00

08:00

09:00

17:00

18:00

19:00

(c)
Actual load curve
Load schedule 1h
Load schedule 1/4h
Balancing demand 1h
[MW]

Balancing demand 1/4h

Time
07:00

08:00

09:00

Prognos (2017)

96
 STUDY | Flexibility in thermal power plants

Schedule leaps In the first 30 minutes of the morning hours (fig-

Power consumption and generation from wind and ure 49a), the scheduled load exceeds the actual load.
PV changes continuously, whereas trading (scheduled Negative balancing power is required. In the second
production) is done in discrete steps, e.g. 60-min- half of the hour, the situation reverses, and positive
ute intervals. Balancing demand due to deviations balancing power is required. In the evening hours
between scheduling and actual loads or production (figure 49b) when the load gradient is negative, we
are called schedule leaps or schedule jumps. see a mirror image of the same trend. This results in
the typical saw-tooth pattern for balancing power
The figure demonstrates the phenomenon of sched- demand – with right-tilted spikes in the morning
ule leaps by examining load procurement for a single hours and left-tilted spikes in the evening hours.
power purchaser. To cover the load in its balancing
group, the power purchaser procures the required Figure 49c shows that the load procurement interval
power in hourly intervals on the wholesale mar- has a significant impact on the magnitude of bal-
ket. On an hourly average, purchased energy equals ancing power demand. If the market design offers
demand. However, within each hour, the actual load 15-minute products, electricity can be purchased and
deviates from the purchased (scheduled) load, thus scheduled on a quarter-hour basis — which signifi-
causing balancing demand. cantly lowers the demand for balancing power.

Average balancing demand in Germany for each 15 minutes interval of the day in 2012 to 2015  Figure 50

1000
Less distinctive
spikes
(2014/2015)
500
Average of balancing power [MW]

−500

−1000
Right-tilted spikes Left-tilted spikes
in the morning in the evening
−1500 (2012/2013)

−2000
00:00 – 00:15
00:30 – 00:45
01:00 – 01:15
01:30 – 01:45
02:00 – 02:15
02:30 – 02:45
03:00 – 03:15
03:30 – 03:45
04:00 – 04:15
04:30 – 04:45
05:00 – 05:15
05:30 – 05:45
06:00 – 06:15
06:30 – 06:45
07:00 – 07:15
07:30 – 07:45
08:00 – 08:15
08:30 – 08:45
09:00 – 09:15
09:30 – 09:45
10:00 – 10:15
10:30 – 10:45
11:00 – 11:15
11:30 – 11:45
12:00 – 12:15
12:30 – 12:45
13:00 – 13:15
13:30 – 13:45
14:00 – 14:15
14:30 – 14:45
15:00 – 15:15
15:30 – 15:45
16:00 – 16:15
16:30 – 16:45
17:00 – 17:15
17:30 – 17:45
18:00 – 18:15
18:30 – 18:45
19:00 – 19:15
19:30 – 19:45
20:00 – 20:15
20:30 – 20:45
21:00 – 21:15
21:30 – 21:45
22:00 – 22:15
22:30 – 22:45
23:00 – 23:15
23:30 – 23:45

Time [quarter hours of the day]

2012 2013 2014 2015

50 Hertz

97
Agora Energiewende | Flexibility in thermal power plants

Figure 50 shows the average balancing demand for

each quarter hour of the day in 2012 to 2015. The
typical saw-tooth demand pattern caused by load
schedule jumps can be clearly observed in empirical
data on German balancing demand in 2012 and 2013:
Balancing demand is characterised by right tilted
spikes in the morning and left tilted spikes in the
evening hours.

In 2014 and 2015 the pattern is less pronounced. This

is attributable to the rise of the intraday market in
German power trading, which significantly reduced
the structural demand for balancing power associated
with schedule leaps.

98
 STUDY | Flexibility in thermal power plants

6. Profiles for Selected Countries:

South Africa & Poland

In this section, questions concerning the flexibility of tions are not foreseen before 2022. Additional contri-
conventional power plants are discussed while spot- butions are foreseen from natural gas CCGT (2.4 GW),
lighting the market environment in South Africa and natural-gas OCGT (3.8 GW), cogeneration and
Poland, two countries with large coal power shares. imports (mainly from hydro power plants in Mozam-
bique and potentially also from Zambia, Zimbabwe
6.1 South Africa and Zaire). By contrast, the draft version of the new
IRP recommends the addition of 18 GW of PV, 37 GW
Energy and climate policy of wind, 20 GW of nuclear, 34 GW of natural gas
The primary vehicle for electricity policy in South power plants, 2.5 GW of imported hydro and 15 GW of
Africa is the Integrated Resources Plan (IRP), which is coal by 2050.
part of the overall Integrated Energy Plan (IEP).
In the field of climate policy as a whole, some con-
The main objective of the IRP is to provide sustaina- ditional commitments exist. South Africa has com-
ble long-term electricity planning while considering mitted itself to achieving emissions reductions of
technical, economic and social constraints and exter- 34 percent from business as usual by 2020 and
nalities (DoE South Africa, 2016). The IRP is designed reductions of 42 percent by 2025. Specific climate
as a “living plan” that can be adapted to changing policies include a carbon tax (implementation is
market conditions when necessary. The first IRP planned in 2017) and carbon budgets at the com-
was designed for the period from 2010 to 2030 and pany level (planned for the period from 2016 to 2020).
remains the official government plan for new genera- As part of the Paris Agreement, South Africa has
tion capacity. In November 2016, an update of published Intended Nationally Determined Contribu-
the IRP 2010 was published as a draft for public con- tions (INDC) and desired emission reductions. In this
sultation, which will take place in 2017. This update connection, it has communicated a peak, plateau and
takes into consideration new economic and tech- decline trajectory for its greenhouse gas emissions,
nical developments and enlarges the timeframe to with emissions slated to range between 398 and
2050. The IRP is also considered to be the regulatory 614 Mt CO2_eq in 2025–2030 and decline in the long
framework with the largest impact on South African term to 212 to 428 Mt CO2_eq by 2050.
climate policy.
Power generation
The 2010 IRP sets forth a fixed target for new renew- South Africa has a long tradition of power gener-
able capacity: namely, 17.8 GW by 2030, including ation from coal power plants, which cover about
1 GW of solar CSP, 8.4 GW of solar PV and 8.4 GW of 90 percent of power needs. As the country has large
wind energy (DoE South Africa, 2013). The IRP also hard coal resources, all coal power plants are fuelled
foresees new coal and nuclear power capacities. Spe- with domestic hard coal. Major expansion of the coal
cifically, 10 GW of new coal power plants should be fleet occurred in the 1960s and 70s due to economic
built by 2020 (Eskom, the main power producer in growth and the substitution of oil with electric-
South Africa, committed to constructing these plants ity after the oil crisis in the 1970s. This large-scale
before the IRP process). Some 9.6 GW of new nuclear capacity expansion subsequently resulted in overca-
power are planned, although nuclear capacity addi- pacities in the late 1980s because electricity demand

99
Agora Energiewende | Flexibility in thermal power plants

Power generation in South Africa by source, 1990–2014  Figure 51

300
260 253
245
250
193
200 187
167
[TWh ]

150

229 242 232

100 193
173
156
50

0
1990 1995 2000 2005 2010 2014

Coal Oil Hydro Nuclear Renewables Other

IEA (2016)

growth failed to meet forecasts. As a result, some Renewable capacities (including flexible hydro from
overcapacities were temporarily shut down. How- pumped storage) have been introduced over the last
ever, in the late 1990s and early 2000s, a range of 5–10 years, though their shares still remain quite
power plants were reactivated following forecasts of low. Figure 52 shows statistics on power supply for
higher future demand. 2014. More than 88 % (232 TWh) of South African
power was generated by coal power plants, which
South Africa’s coal power plants are generally located were mainly operated in baseload mode. Nuclear
near coal mines and remote from large cities. This is power, which is the second largest individual source,
proving to be a liability due to the country’s aging grid accounted for just 5 % (15 TWh) of power generation.
infrastructure, and security of supply is now a major Meanwhile, renewables (including hydro) represented
concern. 2.4 % (over 6 TWh) of electricity generation in 2014.

Figure 51 shows the development of power genera- Because of the dominance of coal power production,
tion in South Africa from 1990 to 2014. South African specific CO2 emissions from power generation in
power generation was and remains dominated by South Africa are as high as 900 g CO2/kWh. By con-
coal power. Over the last 20 years, increasing demand trast, specific CO2 emissions in Germany amount
for electricity has been mainly covered by new or to 500 g CO2/kWh. CCS is often seen as an option
recommissioned coal power plants. for decarbonising electricity generation, but major
challenges exist due to costs, uncertain geological
conditions and the large distances between power

100
 STUDY | Flexibility in thermal power plants

Power supply in South Africa Coal production

by source, 2014  Figure 52 Coal production in South Africa is mainly based on
hard coal and amounts annually to around 300 Mt.
Bituminous coal accounts for 98.6 % of coal produc-
1,1 1,1 tion. South Africa does not produce lignite.
4,1
11,7
13,8 Proven coal reserves in South Africa are estimated
at around 35 million tonnes, which comprise 3 % of
global reserves and 95 % of African reserves. Sixty per
cent of coal production is used for power generation,
followed by synthetic fuels (20 %) in industrial use.
Besides domestic use, more than 20 % of coal produc-
232,0
tion is exported — mainly to the Pacific and Atlantic
steam coal market.

Hard coal Nuclear energy The IRP forecasts increasing coal production (mainly
Hydro power Wind for electricity generation), which raises a number
of challenges. Some restrictions exist due to infra-
Solar PV Imports and other
structure problems, including in particular a lack
of rail capacity. Furthermore, new coal mines will
IEA (2016)
require extensive exploration and feasibility studies,
because high-grade coal from the Central Basin will
be depleted by 2040. Against the backdrop of South
Africa’s reliance on coal power and rising electricity
plants and possible storage facilities (which often demand, coal shortfalls are an increasing risk for
exceed 600 km). energy security. The first coal supply shortages are
expected to occur after 2018 if major investments are
Beyond climate concerns, South Africa’s growing not realised (IEA CIAB, 2016).
power demand and ageing power plant fleet pose
significant challenges, particularly with a view to Main characteristics of coal fired generation
security of supply. This is reflected by the narrowing South Africa’s coal fired power stations are located
margin between peak load and available capacity. in several multi-block sites, and are mainly found
in one province, Mpumalanga. This province is
The country’s coal power plants are old, poorly main- also the epicentre of South African coal production.
tained and often pushed to their maximum capacity. As most of the country’s coal power plants are located
The controlled load shedding that was implemented a considerable distance from demand centres in
after the collapse of a coal silo at the Majuba Power the south-west and south-east, robust grid infra-
Station in 2014 testifies to the poor state of South structure is required to assure security of electricity
Africa’s energy infrastructure. The early retirement supply in all regions.
of coal power stations is therefore constrained by
security of supply problems. Moreover, grid infra- Most of South Africa’s power plants were constructed
structure is weak and outdated. between the late 1970s and the early 1990s. With
an average age of about 35 years, coal power plants
in South Africa are relatively old compared to other

101
Agora Energiewende | Flexibility in thermal power plants

Age distribution of coal power plants in South Africa  Figure 53

4,500

4,000

3,500

3,000
[MW per class]

2,500

2,000

1,500

1,000

500

0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59
age of power plants [years]

Eskom (2016)

Technical performance of coal power plants in South Africa Figure 54

Efficiency of thermal power plants Ramp-rate coal power plants

40 0,7
38 % 0,67
0,6
35 35 % 35 %
0,5
0,48
0,44
[% P/Min]

0,4
[%]

30 30 %
0,3

0,2
25
0,1 0,10

20 0

Eskom (2016)

102
 STUDY | Flexibility in thermal power plants

countries (e.g. 20–25 years for coal power plants in ity requirements in South Africa are likely to increase
Germany). Figure 53 shows the age structure of the in the future. As discussed above, more intermittent
South African power plant fleet. renewable generation significantly increases the
flexibility requirements placed on power systems.
Coal power plants in South Africa show an average Accordingly, the flexibility of conventional power
plant efficiency of about 35 percent, which is well plants (and electricity demand) is sure to become
below the 40 percent average in most industrialised more important in South Africa in the coming years.
countries. Furthermore, the coal fleet is dominated
by slag tap firing boilers, which generally reduces the South Africa’s coal power plants currently lag far
flexibility of the existing coal fleet because of higher behind the flexibility standards that are common
minimum load requirements. for most commonly used hard coal plants elsewhere.
As a result, there is a large potential for retrofit-
South Africa’s coal power plants thus display below- ting measures to increase efficiency and flexibil-
average minimum load levels as well as slow start-up ity, which would reduce coal consumption and CO2
times and ramp rates. The ramp rates of the country’s emissions. A range of options for increasing flexibil-
coal power plants range between 0.1 % and 0.7 % of ity was described in section 4. Flexibility retrofitting
nominal capacity per minute. This is considerably in South Africa would require investment costs below
lower than the standard for hard coal power plants 500 €/kW, as current examples in that section show.
(e.g. 1.5 to 4 % per min as seen in chapter 3). Figure 53
compares ramp rate data. Specific figures on mini- Because South Africa’s coal power plants mainly
mum loads are not publicly available. operate as baseload plants, flexibility retrofitting
could also help to lower CO2 emissions. Moreover,
With more than 50 GW of intermittent renewable such retrofitting would help to reduce coal consump-
generation planned by 2050 in the new IRP, flexibil- tion, easing coal supply concerns.

Ramp rates in comparison Figure 55

Ramp rates of hard coal power plants in South Africa compared

to most-commonly used and state-of-the-art designs
7
6
[% of Nominal Capacity

5
per minute]

4
3
2
1
0
hard coal most-commonly state-of-the-art
power plants used hard coal coal power plants
in South Africa power plants

Eskom (2016)

103
Agora Energiewende | Flexibility in thermal power plants

Alternative flexibility options neighbouring countries. Large additional investment

Beyond increasing the flexibility of coal power plants, would be needed to use power imports as a flexibility
a range of other flexibility options exist. The avail- tool.
ability and viability of different flexibility options
depend on the underlying conditions in each country 6.2 Poland
and must be determined on a case-by-case basis.
Energy and climate policy
Pumped storage and hydro storage power plants rep- Climate policy in Poland is principally determined
resent one flexibility option in South Africa. South by the climate policy of the EU. Poland has com-
Africa currently has total capacity of 3.5 GW in mitted itself to limiting non-ETS GHG emissions to
these technologies. Additional capacities in pumped 14 percent (over 2005) and to increasing the share of
storage are planned in the years up to 2025 (around renewables in gross final consumption by 15 percent
3 GW). Currently, planned pumped storage plants are by 2020.
expected to cost between 500 and 1,500 €/kW.
In the field of energy policy, the National Renewable
Another option is “demand side management” (DSM), Energy Action Plan in Poland seeks to achieve
which aims to increase the flexibility of electricity renewable shares of 19 % in generation, 17 % in the
consumers. The South African utility Eskom is heating/cooling sector and 10 % in the transportation
currently providing incentives for demand side sector by 2020. Reductions in CO2 emissions in the
management through its EEDSM (Energy Efficiency ETS sector are to be realised primarily through the
Demand Side Management) incentive program. construction of efficient coal-fired power plants and
An extremely wide range of DSM flexibility options by building new multi-fuel CHP plants.
are available. Furthermore, the costs associated with
different options diverge considerably. Decentralised The strategy paper “Energy Policy of Poland until
storage — for example, using PV systems in combi- 2030” (EPP) defines the current framework for Polish
nation with batteries — is an additional solution that energy policy after 2020. Published in 2009, the EPP
must be mentioned. Such technology could be an seeks to achieve the following:
extremely viable option in the future, particularly if • Improved energy efficiency
the costs of decentralised storage decrease further. • Enhanced security of fuel and energy supplies
• Diversification of power generation with
Gas power plants are another option for making nuclear energy
conventional generation more flexible. The IRP is • Increased use of renewable energy,
considering the addition of OCGT and CCGT plants in including biofuels
the future. However, the country’s lack of gas infra- • Establishment of competitive fuel and
structure limits the expansion of flexible gas power energy markets
capacity. Furthermore, new gas power plants could • Reduction of environmental impact of
lead to a potential lock-in situation and thus impede the energy sector
the transition to a fully decarbonised power system
in the long-run. The current government is expected to publish an
updated version of the EPP in 2017. This revision
In principal, the importation of electricity is an should reflect the EU’s 2030 energy policy targets
option for addressing regional imbalances in power while forecasting developments in the Polish energy
supply and demand. However, South Africa’s grid sector up to 2050 (IEA 2016).
infrastructure is weak, with limited connections to

104
 STUDY | Flexibility in thermal power plants

Power generation Power generation in Poland by source,

Figure 56 shows the development of power generation 2014, in TWh Figure 57
in Poland from 1990 to 2010. As can be seen from the
data, power generation in Poland is dominated by coal 0,1
power, which accounted for 85 % of power generation
in 2014. Over the course of the last 20 years, growth
1,6
in coal power generation has remained relatively flat. 20,4

Additional power demand has been mainly covered 5,3

by new natural-gas power plants and the expansion

of renewables. In 2016, gross electricity produc- 77,4
tion amounted to 166.6 TWh, including 22.8 TWh
(13.7 percent) from renewables. 54,2

Figure 57 shows power generation by source in 2014.

As can be seen from the data, coal generation is
subdivided into hard coal (77.4 TWh) and lignite Hard coal Lignite
(54.2 TWh). Renewable generation mainly consists of Gas Oil
wind power (8 TWh in 2014) and biomass (10 TWh in RES Others
2014). Because of Poland’s dependency on coal gen-
eration and widespread use of lignite, specific CO2 IEA (2016), IEA CIAB (2016), authors’ calculations
emissions amount to around 1000 g CO2/kWh.

Power generation in Poland by source, 1990 to 2014 Figure 56

180
157 159 159
160 145
136 139
140
120
100
[TWh]

80
138 143 138
131 133 132
60
40
20
0
1990 1995 2000 2005 2010 2014

Coal Oil Gas Hydro Nuclear Renewables Other

IEA (2016)

105
Agora Energiewende | Flexibility in thermal power plants

Some studies have forecasted that Poland will experi- closing down of unprofitable hard coal mines seems
ence shortages at peak-load times in the near future. inevitable, and lignite mines are bound to be depleted
However, Poland’s expansion planning up to 2020 before 2030.28 Lignite production is expected to drop
should place security of supply levels well above that to roughly 10 Mt by 2033 if no new pits are opened.
of other European countries. On the downside, this
expansion will depend heavily on coal power, which If coal demand remains at current levels, Poland’s
could create technology lock-in problems, especially domestic mining will have to be drastically restruc-
in the context of rising CO2-prices lowering the prof- tured by 2030. Otherwise, Poland is likely to become
itability of coal power plants. a significant coal importer.

Poland is also planning to meet additional capacity Main characteristics of coal fired generation
requirements with two new nuclear power plants More than 80 % of Poland’s coal power plants were
with a total capacity of 6 GW. In the polish debate, constructed between the late 1960s and 1990 (see
nuclear power is considered a good option for avoid Figure 58). Half of Poland’s power plant fleet is
import dependency, because domestic coal produc- more than 30 years old, and needs to be replaced or
tion is restricted. However, the construction sites are upgraded soon. Compared to other large coal fleets in
not yet set and commercial operation is not expected countries like Germany, the fleet is 10 years older on
before 2029. average.

Coal production Given the average technical life time of coal and lig-
Coal production in Poland comprises around 140 Mt nite stations is somewhere between 50 and 60 years,
per year and is subdivided into lignite (64 Mt) and Poland will face a major challenge in modernising
hard coal (73 Mt). Poland is the second largest pro- its power plant fleet within the next two decades.
ducer of lignite in Europe after Germany but by far The strengthened EU air pollution standards for
the largest producer of hard coal in Europe. power plants that will be enforced by 2021 increase
the pressure on the Polish power sector to take near-
Hard coal production in Poland is currently down term action.
from considerably higher levels in the 20th century,
when import quotas restricted coal imports. Poland While few data on the technical aspects of the Polish
has 60 billion tonnes of proven hard coal reserves. fleet are publicly available, figures on Polish coal
However, the country’s industrial reserves are much consumption indicate that the average efficiency
lower, amounting to about 4 billion tonnes. Despite of whole power plant fleet is well below 40 %. Most
these considerable reserves, hard coal production in boilers in operation were built by the Polish company
Poland is characterised by poor efficiency and com- Rafako and use the pulverised coal firing technology.
petitive disadvantages to imports from Russia, Czech
Republic and Ukraine. The Polish energy sector also faces similar challenges
in the area of district heating due to the high share of
Around 55 % of primary energy consumption is based CHP plants that are operated using hard coal and (and
on coal and most of the coal production of Poland is to a lesser extent) lignite (see Figure 59). Cogeneration
used for domestic consumption.

Poland’s coal and lignite industries will face major 28 Deloitte 2016, POLISH POWER SECTOR RIDING ON THE WAVE
OF MEGATRENDS, https://www2.deloitte.com/content/dam/
challenges in the coming decade. Indeed, Poland could Deloitte/pl/Documents/Reports/pl_FAE_POLISH_POWER_
face a coal and lignite production gap by 2030. The SECTOR_RIDING_ON_THE_WAVE_OF_MEGATRENDS.pdf

106
 STUDY | Flexibility in thermal power plants

Age distribution of coal and lignite fired power stations in Poland  Figure 58

2,500

2,000
[MW per class]

1,500

1,000

500

0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75
age of power plants [years]

Authors’ figure based on data from Platt

Unit size distribution for Polish coal and lignite fired power stations  Figure 59

5,000
4,500
4,000
3,500
3,000
[MW per class]

2,500
2,000
1,500
1,000
500
0
0 50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850
power plant class [MW]

Power generation Cogeneration

Prognos based on data from Platt

107
Agora Energiewende | Flexibility in thermal power plants

units usually have smaller average sizes than plants Alternative flexibility options
devoted solely to power generation. Aside from measures to improve the flexible oper-
ation of coal power, Polish flexibility options are
The high share of CHP plants in Poland poses prob- limited. As of today, gas and pumped storage hydro
lems for the flexibility needs of the power system. plants provide only a small potential, because their
During the heating period the need for heat supply installed capacity is below 1 GW.
puts CHP plants in must-run operation if heating
needs cannot be covered by back-up boilers or differ- Grid connections to neighbouring electricity mar-
ent sources like industrial waste heat. kets are also limited. During peak times, Poland only
has about 2 GW of exchange capacity to its connected
With an increasing share of renewables, new heat neighbours, including Germany, Sweden, the Czech
storage solutions can help to increase the flexibility Republic and Lithuania. The European Ten Year Net-
of CHP plants. In Denmark and Germany numerous work Development Plan (TYNPD) details projects for
heat storage systems have been integrated into exist- enhancing interconnector capacities to Germany,
ing district heating systems to improve the operation Sweden and Lithuania.
of CHP plants. Energy production in the condensing
mode provides another flexibility reserve in the CHP Another option for increasing flexibility is “demand
sector. This option is especially valuable during peak- side management” (DSM), which aims to increase
load periods in the summer, and offers a flexibility the flexibility of electricity consumers. The utility
potential of at least 1–2 GW. company PGE is offering business clients services to
manage their DSM potential. Various pilot projects
also aim to implement DSM solutions in the house-
hold and commercial sectors. Despite the absence of
good data on the country’s total DSM potential, one
estimate places it at 1.2 GW (Forum Energii).

108
 STUDY | Flexibility in thermal power plants

7. Conclusions

This paper provided a broad analysis on possible Market design and remuneration mechanisms
flexibility measures for thermal power generation, for flexibility
focusing on coal power plants. In doing so, we dis- The economics of retrofitting existing coal power
cussed technical and economic factors related to plants are significantly influenced by the availability
increasing the flexibility of those power plants,while of remuneration options for flexibility. In the absence
also considering specific conditions in two countries of such options, the market design will hamper
(South Africa and Poland). Based on this discussion, investment in coal power flexibility and alternative
some preliminary conclusions can be drawn: flexibility tools. With rising renewable shares, mar-
kets should be tailored to promote the integration of
Energy and climate policy actors that provide valuable flexibility options.
With the ratification of the Paris agreement, decar-
bonisation of the power sector has become a top pri- Alternative flexibility options
ority for a range of countries. However, enhancing The specific benefits of coal power retrofitting are
the flexibility of power systems is crucial if renewa- influenced by the availability of alternative flexibility
ble generation is to be considerably expanded. A pri- options, including flexible generation from conven-
mary option in this regard is to operate power plants tional power plants (e.g. gas, flexible hydro), demand-
more flexibly. Existing coal power plants can contrib- side flexibility and cross-border energy trading.
ute to this flexibility need through targeted retrofit The availability of these options varies considerable
measures. In addition to enabling higher renewable between countries due to structural, economic, and
shares in the power system, coal power plant retro- geographic factors.
fitting can help to reduce CO2 emission, in power sys-
tems characterized by very high shares of baseload or Coal production
must-run coal power plants. The threat of shortfalls in domestic coal production is
constraining the development of coal power plants in
The structure of the existing power plant fleet a number of countries. However, concerns regarding
The advanced age and limited flexibility of existing the long term profitability of the coal industry and a
coal power plants are the two main drivers of mod- lack of good sites have led to decreasing investment
ernisation measures. Countries with old power plants in the development of new coal mines. Coal plants in
that are designed for baseload operation can profit baseload operation consume tremendous amounts of
significantly from retrofitting measures to improve highly specific types of coal and make tight coal sup-
the efficiency and flexibility of their coal plants. ply situations foreseeable in the future if consump-
While the costs of flexibility retrofitting have to be tion remains high. If coal power plants can increase
considered on a case-by-case basis, they can be the flexibility of their operation while increasingly
roughly estimated at 100 to 500 €/kW (see section 4). acting as a back-up for renewable generation, coal
Overnight construction costs for new coal fired consumption can be reduced. This would extend
power stations range from 1,200 €/kW to more than the longevity of existing coal mines while reducing
3,000 €/kW if CCS technology is installed. the need for new exploration. As decarbonisation
progresses over the long run, coal power plants could
be gradually phased-out or maintained as a strategic
reserve, thus reducing coal consumption and emis-
sions even further.

109
Agora Energiewende | Flexibility in thermal power plants

110
 STUDY | Flexibility in thermal power plants

8. Bibliography

ABB AG Power Systems Divison, 2013. Boiler Max – Franke, R. & Weidmann, B., 2008. Unter Dampf –
Optimize power plant start-up. Mannheim: s.n. Anfahroptimierung von Dampfkesseln in
E.ON-Kraftwerken. s.l.:ABB Technik.
AG Energiebilanzen, 2016. Auswertungstabellen.
[Online] Frohne, A., 2012. Kraftwerke im Umbruch –
Available at: www.ag-energiebilanzen.de/10-0- Antworten auf die Anforderungen der Energiewende.
Auswertungstabellen.html Berlin: Dialogforum “Retrofit und Flexibilisierung
[Accessed 12 December 2016]. konventioneller Kraftwerke”.

Agora Energiewende (2015). The Integration Cost of GE, n.d. Heat Recovery Steam Generators. [Online]
Wind and Solar Power. An Overview of the Debate Available at: powergen.gepower.com/content/dam/
on the Effects of Adding Wind and Solar Photovoltaic gepower-pgdp/global/en_US/images/product/
into Power Systems. heat%20recovery%20steam%20generators/hrsg-
approach.jpg
Alstom, 2013. Ausbalanciertes Portfolio, Kiel: 13. [Accessed 18. November 2016].
Technisches Presse Colloquiums.
Gostling, J., 2002. Two Shifiting of Power Plant:
Alstom, n.d. s.l.:s.n. Damage to Power Plant Due to Cycling – A brief
overview, s.l.: s.n.
Balling, L., 2010. Flexible future for combined cycle.
Modern Power Systems, December. Heinzel, T., Meiser, A., Stamatelopoulos, G.-N. &
Buck, P., 2012. Einführung Eimühlenbetrieb in
Clerens, P. et al., 2015. Thermal power in 2030 – den Kraftwerken Bexbach und Heilbronn Block 7.
Added value for EU energy policy, Brussels: European VGB PowerTech, November, pp. 79–84.
Power Plant Suppliers Association.
INAS, 2014. s.l.:s.n.
Cziesla, F. et al., 2013. Fossil befeuerte
Grosskraftwerke in Deutschland – Stand, Tendenzen, Jeschke, R., Henning, B. & Schreier, W., 2012.
Schlussfolgerungen, s.l.: VDI. Flexibility through highly-efficient technology.
VGB PowerTech, May, pp. 64–68.
Cziesla, F. et al., 2009. Advanced 800+ MW Steam
Power Plants and Future CCS Options, Katowice: s.n. Klumpp, F., 2009. Energiewirtschaftliche Bewertung
anfahrbedingt verursachter Instandhaltungskosten
FDBR, 2012. Anpassung thermischer Kraftwerke von Steinkohlekraftwerken. Stuttgart: s.n.
an künftige Herausforderungen im Strommarkt –
Massnahmenkatalog, Düsseldorf: Fachverband Kraft, A., 2015. Virtuelle Kraftwerke und
Anlagenbau Energie. Umwelt. Prozessindustrie. Wärmespeicher. s.l.:s.n.

Fichtner, 2017. Stuttgart: s.n. LEAG, n.d. Geschäftsfeld Kraftwerke. [Online]

Available at: kraftwerke.vattenfall.de/boxberg
[Accessed 20. October 2016].

111
Agora Energiewende | Flexibility in thermal power plants

Michels, A., 2016. Trockenbraunkohle erhöht die Sarkar, D., 2015. Thermal Power Plant – Design and
Flexibilität, Bonn: BINE Informationsdienst – Operation. s.l.:Elsevier.
FIZ Karlsruhe.
Scheffknecht, G., 2005. Vorlesung Dampferzeugung.
Prognos, 2017. Berlin: s.n. Stuttgart: s.n.

Quinkertz, R., Ulma, A., Gobrecht, E. & Wechsung, M., Schmidt, G. & Schuele, V., 2013. Anpassungen
2008. USC Steam turbine technology for maximum thermischer Kraftwerke an künftige
efficiency and operational flexibility, Kuala Lumpur: Herausforderungen im Strommarkt. Berlin: Alstom.
Siemens.
Schulze, U. & Hoffmann, H., 2013. Vom Grundlast­
RWE Power AG, 2011. Revision von Block E im betrieb zur flexiblen Leistungsbereitstellung.
Kraftwerk Neurath: Investitionen von 70 Millionen s.l.:10. Jahrestagung Kraftwerke – Rückgrat der
Euro steigern Effizienz und Flexibilität. [Online] Energieversorgung.
Available at: www.rwe.com/web/cms/de/2320/rwe-
power-ag/presse-downloads/pressemitteilungen/ SGS Industrial Services, 2011. Case study. s.l.:SGS SA.
pressemitteilungen/?pmid=4006805
[Accessed 13. October 2016]. Siemens, 2016. Siemens Gas Turbines. [Online]
Available at: www.energy.siemens.com/co/pool/hq/
RWE Power AG, 2011. Revision von Block H im power-generation/gas-turbines/downloads/gas-
Kraftwerk Weisweiler: Investitionen von 65 Millionen turbines-siemens.pdf
Euro steigern Effizienz und Flexibilität. [Online] [Accessed 18. November 2016].
Available at: www.rwe.com/web/cms/de/2320/rwe-
power-ag/presse-downloads/pressemitteilungen/ STEAG GmbH, n.d. Kraftwerk Duisburg-Walsum.
pressemitteilungen/?pmid=4006953 [Online]
[Accessed 12. October 2016]. Available at: www.steag.com/s-kraftwerke-walsum.
html
RWE Power AG, 2012. Kraftwerk Weisweiler: [Accessed 20. October 2016].
Mit 60 Millionen Euro wird „Gustav“ fit gemacht für
das Zusammenspiel mit Wind und Sonne. [Online] Steck, M. & Mauch, W., 2008. Technische
Available at: www.rwe.com/web/cms/de/2320/rwe- Anforderungen an neue Kraftwerke im Umfeld
power-ag/presse-downloads/pressemitteilungen/ dezentraler Stromerzeugung. 10. Symposium
pressemitteilungen/?pmid=4008214 Energieinnovation, 13.–15. February.
[Accessed 12. October 2016].
Strauss, K., 2016. Kraftwerkstechnik zur Nutzung
RWE Power AG, n.d. Kraftwerk Weisweiler. [Online] fossiler, nuklearer und regenerativer Energiequellen.
Available at: www.rwe.com/web/cms/de/60142/rwe- 7. Auflage ed. Dortmund: Springer Berlin Heidelberg .
power-ag/energietraeger/braunkohle/standorte/
kw-weisweiler/ Then, O., 2016. What drives conventional power plants
[Accessed 13. October 2016]. in Germany and Europe?. Delhi: VGB PowerTech.

RWE Power AG, n.d. Kraftwerk Weisweiler. [Online]

Available at: Kraftwerk Weisweiler
[Accessed 19. October 2016].

112
 STUDY | Flexibility in thermal power plants

Trianel, n.d. Daten und Fakten zum Kraftwerk. [Online] Wietschel, M. et al., 2015. Energietechnologien der
Available at: www.trianel-luenen.de/kraftwerk/ Zukunft. Wiesbaden: Springer Vieweg.
daten-und-fakten/
[Accessed 19 October 2016]. Ziems, C. et al., 2012. Kraftwerksbetrieb bei
Einspeisung von Windparks und Photovoltaikanlagen,
VDE, 2012. Erneuerbare Energie braucht flexible Rostock: VGB Powertech.
Kraftwerke – Szenarien bis 2020, Frankfurt am Main:
s.n.

113
Publikationen von Agora Energiewende

IN ENGLISH

The Cost of Renewable Energy

A critical assessment of the Impact Assessments underlying the Clean Energy for All Europeans-Package

Future Cost of Onshore Wind

Recent auction results, long-term outlook and implications for upcoming German auctions

The Energiewende in a nutshell

10 Q & A on the German energy transition

Energy Transition in the Power Sector in Europe: State of Affairs in 2016

Review on the Developments in 2016 and Outlook on 2017

Assessing the Winter Package in Light of the Energy Union Objectives

FAQ EEG – Energiewende: What do the new laws mean?

Ten questions and answers about EEG 2017, the Electricity Market Act, and the Digitisation Act

Reducing the cost of financing renewables in Europe

A proposal for an EU Renewable Energy Cost Reduction Facility (“RES-CRF”)

Refining Short-Term Electricity Markets to Enhance Flexibility

Stocktaking as well as Options for Reform in the Pentalateral Energy Forum Region

Energy Transition in the Power Sector in Europe: State of Affairs in 2016

Review on the Developments in 2016 and Outlook on 2017

The Power Market Pentagon

A Pragmatic Power Market Design for Europe’s Energy Transition

Energy Transition in the Power Sector in Europe: State of Affairs in 2015

Review of the Developments and Outlook for 2016

Projected EEG Costs up to 2035

Impacts of Expanding Renewable Energy According to the Long-term Targets of the Energiewende

Eleven Principles for a Consensus on Coal

Concept for a stepwise decarbonisation of the German power sector (Short Version)

The Integration Costs of Wind and Solar Power

An Overview of the Debate of the Effects of Adding Wind and Solar Photovoltaics into Power Systems

12 Insights on Germany’s Energiewende

A Discussion Paper Exploring Key Challenges for the Power Sector

114
Publikationen von Agora Energiewende

IN GERMANY

Die deutsche Braunkohlenwirtschaft

Historische Entwicklungen, Ressourcen, Technik, wirtschaftliche Strukturen und Umweltauswirkungen

Charta für eine Energiewende-Industriepolitik

Ein Diskussionsvorschlag von Agora Energiewende und Roland Berger

Neue Preismodelle für Energie

Grundlagen einer Reform der Entgelte, Steuern, Abgaben und Umlagen auf Strom und fossile Energieträger

Smart-Market-Design in deutschen Verteilnetzen

Entwicklung und Bewertung von Smart Markets und Ableitung einer Regulatory Roadmap

Energiewende und Dezentralität

Zu den Grundlagen einer politisierten Debatte

Wärmewende 2030
Schlüsseltechnologien zur Erreichung der mittel und langfristigen Klimaschutzziele im Gebäudesektor

All publications are available on our website: www.agora-energiewende.de

115
115/04-S-2017/EN

Agora Energiewende develops

evidence-based and politically viable
strategies for ensuring the success of
the clean energy transition in
Germany, Europe and the rest of the
world. As a think tank and policy
laboratory we aim to share knowledge
with stakeholders in the worlds
of politics, business and academia
while enabling a productive exchange
of ideas. Our scientifically rigorous
research highlights practical policy
solutions while eschewing an
ideological agenda. As a non-profit
foundation primarily financed through
philanthropic donations, we are not
beholden to narrow corporate or
political interests, but rather to our
commitment to confronting climate
change.

Agora Energiewende
Anna-Louisa-Karsch-Straße 2 | 10178 Berlin
P +49 (0)30 700 14 35-000
F +49 (0)30 700 14 35-129
www.agora-energiewende.de
info@agora-energiewende.de

Agora Energiewende is a joint initiative of the Mercator Foundation and the European Climate Foundation.