AckermannTh. (ed) Wind Power in Power Systems
Подождите немного. Документ загружается.
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 18_CHA17 .3D – 381 – [365–382/18]
17.12.2004 10:39PM
national Symposium on Distributed Generation: Power System and Market Aspects, Stockholm, 2–4
October 2002.
[10] IRENE (2001) ‘Integration of Renewable Energy in the Electrical Network in 2010 – IRENE’, EC funded
project, ALTENER XVII/4.1030/Z/99-115.
[11] Landberg, L. (2000) ‘Prediktor – an On-line Prediction System’, presented at Wind Power for the 21st
Century, Special Topic Conference and Exhibition, Convention Centre, Kassel, Germany, 25–27
September 2000.
[12] Landberg, L., Joensen, A., Giebel, G., Madsen, H., Nielsen, T. S. (2000a) ‘Zephyr: The Short-term
Prediction Models’, presented at Wind Power for the 21st Century; Special Topic Conference and
Exhibition, Convention Centre, Kassel, Germany, 25–27 September 2000.
[13] Landberg, L., Joensen, A., Giebel, G., Madsen, H., Nielsen, T. S. (2000b) ‘Short Term Prediction
Towards the 21st Century’, in Wind Forecasting Techniques: 33 Meeting of Experts, Technical Report
from the International Energy Agency, R&D Wind, Ed. S.-E. Thor, FFA, Sweden, pp. 77–85.
[14] Nielsen, T. S., Madsen, H., Tofting, T. (2000) ‘WPPT: A Tool for On-line Wind Power Prediction’, in
Wind Forecasting Techniques: 33 Meeting of Experts, Technical Report from the International Energy
Agency, R&D Wind, Ed. S.-E. Thor, FFA, Sweden, pp. 93–116.
[15] Sa
´
nchez, I. et al. (2002) ‘SIPREO
´
LICO – A Wind Power Prediction System Based on Flexible Combina-
tion of Dynamic Models. Application to the Spanish Power System’, in First Joint Action Symposium on
Wind Forecasting Techniques, International Energy Agency, R&D Wind, Ed. S.-E. Thor, FFA, Sweden.
[16] Tande, J., Landberg, L. (1993) ‘A 10 Sec. Forecast of Wind Turbine Output with Neural Networks’,
presented at European Wind Energy Conference 1993.
Useful websites
Anemos Project, http://anemos.cma.fr.
Danish Meteorological Institute, http://www.dmi.dk; see also http://www.dmi.dk/eng/ftu/index.html.
Deutscher Wetterdienst (German Weather Service), http://www.dwd.de.
HIRLAM, http://www.knmi.nl/hirlam; see also Danish Meteorological Institute.
HONEYMOON, http://www.honeymoon-windpower.net.
Institut fu
¨
r Solare Energieversorgungstechnik, http://www.iset.uni-kassel.de.
Prediktor, http://www.prediktor.dk; see also Risø National Laboratory.
Previento, http://www.previento.de.
Risø National Laboratory, http://www.risoe.dk.
Zephyr: see Risø National Laboratory.
Wind Power in Power Systems 381
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 18_CHA17 .3D – 382 – [365–382/18]
17.12.2004 10:39PM
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 19_CHA18 .3D – 383 – [383–410/28]
17.12.2004 10:41PM
18
Economic Aspects of Wind
Power in Power Systems
Thomas Ackermann and Poul Erik Morthorst
18.1 Introduction
Wind power can reach a significant penetration level in power systems, as the
examples in this part of the book have shown. To define the optimum penetration
level of wind power is hardly a technical problem; it is, rather, an economic issue.
Considering that the power system in most countries was designed around conven-
tional generation sources such as coal, gas, nuclear or hydro generation (see
Chapter 3), the integration of wind power may require a redesign of the power
system and/or a change in power system operation. Both will affect the costs of the
overall power supply.
Many of the conflicts between network owners or network operators and wind
project developers worldwide are therefore often only indirectly related to technical
issues being concerned rather with the question of what costs wind power causes and
how those costs should be split between wind farm operator, network owner or network
operator and other network customers.
In this chapter we discuss some of the relevant aspects – namely, network connection,
network upgrade and system operati on co sts. The chapter includes examples from
different countries of how costs for network connection and network upgrading can
be divided. We also use empirical data to illustrate the effect of wind power on power
prices and balancing costs.
Wind Power in Power Systems Edited by T. Ackermann
Ó 2005 John Wiley & Sons, Ltd ISBN: 0-470-85508-8 (HB)
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 19_CHA18 .3D – 384 – [383–410/28]
17.12.2004 10:41PM
18.2 Costs for Network Connection and Network Upgrading
Different countries use different approaches regarding the determination of connection
costs for new wind turbines and wind farms. In some countries, the regulator sets the
connection charges , independent of the actual connection costs. In most re-regulated
market environments the connection costs have to represent the actual costs for the
connection. However, the approaches differ when determining what should be included
in the specific network connection costs. Three different approaches can be distin-
guished: deep, shallowish and shallow connection charges. Table 18.1 shows the net-
work connection charging principals for renewable energy projects in the EU. The
differences between the three approaches will be explained next.
18.2.1 Shallow connection charges
The shallow co nnection cost approach includes only the direct connection costs, that is,
the costs for new servi ce lines to an existing network point, and partially also the costs
for the transformer that is needed to raise the voltage from the wind farm to the voltage
in the distribution or transmission network. The details of the shallow connection
approach, however, can vary significantly between countries, as can be seen from the
following examples.
Example: Denmark
In the Danish approach, public authorities assign an area to a wind farm. The wind farm
owner pays all network costs within this area (e.g. for the interconnection of the wind
turbines). On the border of this area, the distribution network company defines a
connection point (CP; see Figure 18.1). All costs from this connection point, including
the costs for the connection point (transformer station), are paid by the network
company, which can ‘socialise’ all these costs.
According to Danish regulations, network integration costs can be socialised (i.e. the
costs are equally divided between all network customers). For wind power the network
integration costs that can be socialised include those from the CP up to the 20 kV
network. Today, new rules even allow the socialisation of network integration costs
associated with wind power for a network voltage level of up to 100 kV.
In Denmark, the interconnection of a wind farm consists therefore typically of a
newly built connection, usually an underground cable, between the CP a nd a strong
Table 18.1 Network connection charging principals in the EU
Approach Country
Deep Greece, Ireland, Luxembourg, Netherlands, Spain, UK
Shallowish UK
a
Shallow France, Belgium, Denmark, Germany, Netherlands, Italy, Portugal, Sweden
a
Proposed by the regulator Ofgem and expected to start 1 April 2005 (Ofgem, 2003).
384 Economic Aspects
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 19_CHA18 .3D – 385 – [383–410/28]
17.12.2004 10:41PM
point within the distribution network, most likely an existing substation. This has the
advantage that it solves a number of technical issues:
.
As no other customers are connected to the new line, larger voltage variations can be
allowed; flicker as well as harmonics are a minor concern.
.
The comparatively long underground cable will result in a decrease of voltage vari-
ations at the strong connection point in the distribution network.
.
In cases where local generation exceeds local demand, the protection system needs only
to be redesigned between the CP in the distribution network and the transmission system.
.
Power flow management will be required only if local generation exceeds the rating of
the equipment between the CP in the distribution network and the transmission system.
Table 18.2 provides an overview of the annual network integration costs for combined
heat and power (CHP) projects and wind power in the Western Danish system. The
costs include the costs for building a network connection between the CP and the
existing network as well as possible network upgrades within the Danish power system.
Table 18.2 illustrates the following:
.
Network Integration Costs per megawatt can vary significantly. For example, for
wind power they can vary between DKr1.1 million per megawatt and DKr0.18
million per megawatt (E15 000–E24 000 per megawatt). This shows that network
integration costs are very case-specific.
.
Also, the average costs for CHP – DKr0.14 million per megawatt (E18 860 per
megawatt) – are almost 50 % lower than the average costs for wind power –
DKr0.26 million per megawatt (E35 000 per megawatt). The reason is that CHP is
usually built close to populated areas and hence are closer to strong CPs than are wind
farms, which are usually built at a larger distance from any population.
Example: Germany
In Germany, the distribution network company will define the capacity that can be
connected at point A, B or at the substation (see Figure 18.2). For the calculation of the
available capacity, the distribution network company takes into account voltage limits,
increases in short-cir cuit levels and power flow issues, for instance.
Distribution
network
Substation
20/10 kV
A
B
Wind farm
CP
Transmission
network
Figure 18.1 Danish connection approach. Note:CP¼ connection point
Wind Power in Power Systems 385
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 19_CHA18 .3D – 386 – [383–410/28]
17.12.2004 10:41PM
The wind farm owner has to pay for the costs for the connection between the wind
farm and the network CP, including the costs for the CP (i.e. the substation). Hence a
German wind project will incur higher costs for being connected to the network than a
similar Danish project. The German project will have to include these costs when
calculating the overall financial feasibility of the project.
The distribution network company will mainly pay for the redesign of the protection
system and the voltage control system. Only if a larger number of wind farms are
connected within the same area may the distribution network company incu r costs for
network upgrades (e.g. between the distribution system and the transmission system). In
Germany, all these costs can in principal be socialised (i.e. all wind power related
Table 18.2 Annual integration cost for combined heat and power (CHP) and wind power in the
Western part of Denmark (exchange rate: 1 E ¼7:45 DKr)
Year CHP Wind power
MW
installed
total cost
per year
(DKr millions)
cost
per MW
(DKr millions)
MW
installed
total cost
per year
(DKr millions)
cost
per MW
(DKr millions)
1993 30 26 0.86 0 0 —
1994 180 58 0.32 15 2 0.13
1995 230 25 0.10 10 11 1.1
1996 250 19 0.07 70 60 0.8
1997 230 20 0.08 125 65 0.52
1998 120 3 0.02 280 52 0.18
1999 80 7 0.08 210 68 0.32
2000 40 11 0.27 210 131 0.62
2001 30 2 0.06 420 51 0.12
2002 20 19 0.95 220 76 0.34
Total 1300 190 0.14 1950 516 0.26
— Not applicable
Source: Eltra, http://www.eltra.dk.
Distribution
network
Substation
20/10 kV
A
B
Wind farm
CP
Transmission
network
Figure 18.2 German network connection approach. Note:CP¼ connection point
386 Economic Aspects
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 19_CHA18 .3D – 387 – [383–410/28]
17.12.2004 10:41PM
network upgrade costs are shared by all network customers). As wind power in
Germany is developed mainly along the coastl ine see Plate 1, in Chapter 2, wind power
related network upgrade costs are higher in coastal areas than in other areas. Hence the
German approach assures that the network related wind power costs are divided equally
between all network users in Germany.
Example: Sweden
In Sweden, distribution network companies (DNCs) are required to have a license from
the regulator in order to be able to operate a network within a certain geographical area.
Only one license for a geographical area is available, hence if a renewable distributed
generation (DG) project (e.g. example a wind farm) is under development the distribu-
tion company that has the license for this geographical area must build a network
connection to this new power plant. The distribution network company can charge
the new power plant with the costs for the construction of this new line (connection
costs). In addition, the wind farm has to pay for using this distribution network (use-of-
system charges). However, if the wind farm increases capacity-related costs for the DNC
(e.g. capacity costs for the use of the transmission system) the DNC is not allowed to
charge the wind farm for this. It is not uncommon for wind farms to increa se capacity-
related costs, because wind power is often developed in areas with low population
density (i.e. in areas with good wind speeds). Hence, the locally generated power cannot
be used entirely in the distribution network. It is therefore exported to the transmission
network, which may cause additional costs. The disadvantage of the Swedish approach
is that those additional costs must be shared by all distribution network customers of
this particular distribution network company and are not covered by the wind farm
operator or all Swedish network customers. Hence, DNCs have comparatively low
incentives to connect larger wind farms.
18.2.2 Deep connection charges
Connection costs include the connection assets (transformer) as well as all or parts of the
costs for necessary network reinforcements (i.e. network reinforcements at the transmis-
sion and distribution level).
The ap proach may cause the following problems:
.
It is not always possible to determine exactly which costs are related to the connection
of an additional generation project (i.e. a wind power project) or which costs might be
caused by load increase or by ageing equipment.
.
The network is usually upgraded in steps. The voltage level is, for example, increased
from 20 kV to 60 kV. Hence a wind farm project might have to pay for a network
upgrade that exceeds the actual requirements. Hence, any generation project that is
connected later and benefits from this network reinforcement will not be charged for
it; hence it uses the network for ‘free’ (the free-rider prob lem). In some cases, different
wind power projects have voluntarily agreed to share the costs of a network upgrade,
even though the actual rules do not require this .
Wind Power in Power Systems 387
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 19_CHA18 .3D – 388 – [383–410/28]
17.12.2004 10:41PM
In practice, deep connection charges caused significant conflicts between network
owners or network operators, on the one hand, and wi nd farm developers, on the other
hand, about the appropriate costs for network reinforcements.
18.2.3 Shallowish connection charges
Shallowish connection charges are a combination of deep charges and shallow charges,
as the connection charges include a contribution to reinforcement costs based upon the
proportion of increased capacity required by the connectee.
Example: UK
The English and Welsh regulator Ofgem concluded, after long investigations, that deep
connection charges for wind power projects or other DG projects cannot be considered
fair as new larger electricity customers that require a network upgrade when being
connected to the distribution network usually pay only a shallow connection charge.
(Ofgem, 2003). Hence demand and generation is treated differently, which can be
considered unfair. In addition, it can also be argued that distributed and centralised
generation are generally treated differently, because usually centralised generation pays
only shallow connection fees (i.e. a large fossil fuel power station will not pay for
network upgrade costs).
As a result of the findings, Ofgem has decided to introduce the following changes in
England and Wales from 1 April 2005:
.
Wind power or other DG requir ing connection to a distribution netwo rk will be
treated in a similar way to demand (i.e. it will no longer face deep connection charges).
.
Connection charges will include a contribution to reinforcement costs based upon the
proportion of increased capacity required by the connectee. This policy will apply to
the connection of both demand and generation.
18.2.4 Discussion of technical network limits
Network upgrades are required because certain limits (e.g. capacity or stability limits) of
the existing network are reached. It is often assumed that network limitations are of a
purely technical nature. However, technical limitations are often closely related to
economic aspects. For instance, distribution network operators usually have to take
into account the loadability of the installed equ ipment when defining the possible wind
power capacity that can be connected. Overloaded components can lead to voltage
stability problems or to an interruption of the supply caused by the tripping of the
overloaded equipment. Voltage stability problems are hardly acceptable; however,
overloading of equipment may not automatically result in equipment failure. Some
network components may tolerate short-term overloading, but the lifetime of the
equipment will most likely be reduced when short-term overloading is acceptable.
The impact that an overloading of components has on the remaining equipment
lifetime is influenced by numerous factors and is statistical in character. The exact effect
388 Economic Aspects
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 19_CHA18 .3D – 389 – [383–410/28]
17.12.2004 10:41PM
of overloading on the lifetime of equipment is difficult to determine. That is the reason
why most network companies do not allow any overloading during normal operation.
Economically, a short overloading of some equipment may be economically efficient for
the distribution company as it may be able to generate an additional income (i.e. via
network tariffs) without causing any upgrade costs.
In relation to wind power , this discussion can be very important. As wind power
generation as well as local demand varies, certain extreme events that define the
technical limits of a network occur comparatively seldom. Typical extreme events in
Europe are nights with very low local loads and extreme high wind speeds, for instance.
Such events usually occur for only a few hours per year; hence the most economic
solution to avoid network upgrades would either be
.
to accept short-term overloading of network equipment or
.
to reduce wind power production during such extreme events.
In the first case, the network operator would accept the economic risk associated with
overloading equipment over a short period, because the operator would receive a higher
income via network tariffs for the remaining time of the year.
(1)
In the secon d case, the
wind power producer would spill some wind generation; however, in retur n the network
operator would accept the connectio n of more wind power capacity to the network than
it would in a situation without spilling in extreme events.
(2)
In general, it seems difficult to include the above aspects into the definition of
network interconnection charges. On the one hand, shallow connection charges provide
little incentives for wind farm ope rators to consider voluntary spilling but may provide
incentives for network companies to accept short-term overloading. On the other hand,
shallowish charges may provide incentives for wind farm operators to consider volun-
tary spilling if this results in avoiding contributions to reinforcement costs.
18.2.5 Summary of network interconnection and upgrade costs
The a bove exampl es reflect very different general principals towards wind power and the
related interconnecti on and upgrade costs. The German and Danish approach aims at
simplicity and allows the socialisation of all wind power related upgrade costs. The
principal idea behind the socialisation of wind power related upgrade costs is that wind
power is a measure to create a sustainable energy sup ply for the benefit of all market
participants. Hence, all market participants pay a share of the relevant costs (i.e. the
network upgrade costs). The development in Germany and Denmark also shows that
(1)
A few network operators in Europe therefore accept short-term overloading of network equipment caused
by wind power. The equipment in question (i.e. transformers and overhead lines) is usually located in open
terrain (i.e. the equipment is exposed to wind). Overloading is accepted because overloading is assumed to
occur only during situations with high wind speeds (i.e. the overloaded equipment will be additionally cooled
by the wind).
(2)
Such cases are found in Europe. However, wind farm operators may encounter problems in financing such
wind farms, because investors may not accept the uncertainty regarding the likelihood of spilling and hence the
financial impact of such an agreement.
Wind Power in Power Systems 389
//INTEGRAS/KCG/P AGIN ATION/ WILEY /WPS /FINALS_1 4-12- 04/0470855088_ 19_CHA18 .3D – 390 – [383–410/28]
17.12.2004 10:41PM
dividing the relevant costs between all market participant s reduces interconnection
barriers, which emerge when only some market participants have to cover the additional
costs (as is currently the case in Sweden). The fast development of wind power in
Germany and Denmark over the past decade is partly related to the fact that shallow
interconnection charges have reduced interconnection barriers significantly.
The disadvantage of shallow connection charges is that there are no incentives for
developing wind power projects in areas where network upgrade costs are low. Hence,
with a shallow connection approach, network upgrade costs are not part of the overall
economic decision of where to develop wind power projects, and that may lead to an
overall inefficient economic solution.
(3)
The shallowish approach may help to find a
more overall efficient economic solution. This will, however, depend very much on how
the contribution to the reinforcement costs is actually calculated. Deep connection
charges, however, are hardly justified because neither conventional power plants nor
customers pay deep connection charges.
The final decision regarding which approach to implement depends on the overall
policy aim (e.g. promoting the fast development of an emerging sustainable energy
resource or providing incentives for the development of a sustainable energy resource
under consideration of network upgrade costs). If the second approach is applied, all
network customers should be treated in a similar way (i.e. conventional generation
sources as well as customers should pay a contribution to the network reinforcement
costs they cause).
18.3 System Operation Costs in a Deregulated Market
The intermittency of wind power affects system operation. It causes costs that are
associated with balancing wind power in order to provide a continuous balance of
demand and supply in a power system. The balancing task can be divided into primary
and secondary control. Primary control refers to the short-term, minute-by-minute
control of the power system frequency. Secondary control assumes the capacity tasks
of the primary control 10 to 30 minutes later and thereby frees up capacity that is used
for primary control. In the following, the relevant economic issues related to wind
power and primary and secondary control are discussed (for an illustration, see also
Figure 8.3, page 148).
Other system operation costs related to quality of supply other than frequency (e.g.
voltage control, system restoration or restart costs following a blackout) are not
discussed because it is usually very difficult to determine clearly what causes these costs.
If a wind farm directly requires additional voltage control equipment, this could be
considered network upgrade costs. The actual operating costs of the voltage control
equipment, however, are difficul t to determine and to allocate to a wind farm, for
instance.
(3)
It can, of course, be argued that a sustainable energy supply system takes into account external costs. A
sustainable energy supply system therefore also aims at an overall efficient economic solution. However,
shallow connection charges still provide no incentives for developing wind power projects in areas where
network upgrade costs are low.
390 Economic Aspects