Overview

Wind integration: experience,

issues, and challenges
Hannele Holttinen∗

The challenge of wind integration is to make best use of the variable and uncertain
power source while maintaining the continuous balance between consumption
and generation and high level of reliability in the power system. There is already
experience of operating power systems with large amounts of wind power and
integration studies give estimates on wind power impacts. Power systems are
equipped to handle variability and uncertainty that comes from the electricity
consumption, the load. Short-term wind forecasting is required to manage large
amounts of wind power. The main impacts of wind integration are investments
in grid infrastructure and efficiency losses in power plants when following the
increased variations and uncertainty in the power system. Wind power will lower
emissions while replacing energy produced by fossil fuels and can also replace
some power plant capacity. However, wind’s lower capacity value compared to
conventional power plants is one integration impact of wind power, meaning
higher total installed capacity in power systems with high wind penetration.
Managing options for wind integration impacts includes proper wind power plant
grid-connection rules, increasing transmission capacity and increasing flexibility
that is available from generation plants and demand side. Further development of
models and tools is required to study how entire power systems can be operated
during the hours and days of very high penetration levels covering 60–80% of
load. C 2012 John Wiley & Sons, Ltd.

How to cite this article:

WIREs Energy Environ 2012, 1: 243–255 doi: 10.1002/wene.18

INTRODUCTION installed wind power capacity as a percentage of peak

load (capacity). In this paper, the energy penetration
W ind power is a strongly growing renewable
electricity technology that has high technical
potential worldwide and high targets of deployment
level is used.
There is already experience in operating power
in many countries. Integrating variable and uncertain systems with 10–20% penetration levels and a lot
wind power production to the electricity system is one of studies on the impacts of wind power to power
limiting factor in using the high technical potential. systems. This paper is about integrating wind power
The concerns that wind power brings for system op- in larger power systems and is mainly based on three
eration are how to maintain power system reliability summary publications on the issue.1–3 For smaller au-
and the balance between load and generation. tonomous systems where wind provides a large part
Integration challenge depends on the penetra- of the energy for wind–diesel systems, or in small is-
tion level of wind power in the power systems. The land systems see, e.g., Refs 4 and 5.
penetration of wind power can be expressed by vari-
ous measures. Usually, either energy or capacity met-
rics are used: yearly wind power production as a per- THE CHALLENGE OF WIND POWER
centage of yearly electricity consumption (energy) and
TO POWER SYSTEMS
The goal of power system operations is to maintain
∗
Correspondence to: [email protected] the continuous balance between load and generation
VTT, Technical Research Centre of Finland, Espoo, Finland with high reliability. Possible impacts are presented in
DOI: 10.1002/wene.18 Figure 1 in two dimensions—impacts in different area

Volume 1, November/December 2012 

c 2012 John Wiley & Sons, Ltd. 243
Overview wires.wiley.com/wene

have more challenges than larger, well-interconnected

power systems that can use all flexibility options in
power plants and have a strong grid to start with.

Existing Experience on Wind Integration

There is already valuable experience of operating
power systems with high wind penetrations.3,8–11
West Denmark and North Germany have more than
30% of yearly load covered by wind generation. They
are smaller parts of larger systems and operate at
times with wind power covering more than the load
F I G U R E 1 | Impacts of wind power on power systems, displayed (more than 100% wind penetration). Spain, Portu-
by time and spatial scales relevant for the studies. Primary reserve is gal, and South Australia have about 15% of yearly
here denoted for reserves activated in seconds (system-wide frequency load covered by wind generation. They are regions
activated reserve; regulation; automatically activated reserve of the that are not very well-connected parts of larger sys-
balancing zones). Secondary reserve is here denoted for reserves tems and have coped with periods of 54–71% wind
activated in 10–15 min (minute reserve, load following reserve, penetration levels. Ireland is a small power system
manually activated). with about 10% penetration level on a yearly basis
and currently coping at times with 50% penetration
size (locally and systemwise) and impacts in different levels.
timescales. Impacts can been seen during operation (in Coping with these levels of penetration is made
seconds to days) and need to be taken into account possible by online real-time wind-generation data to-
when planning the adequacy of future generation ca- gether with constantly updated forecasts of expected
pacity and transmission grid years ahead. production in system operators’ control rooms.1 In
The integration cost is the additional cost of cases where wind power plants are from small units
the design and operation of the nonwind part of the in the distribution system, decentralized control cen-
power system when wind power is added to the gen- tres enable collecting online data and possibility to
eration mix. Often integration costs are associated control wind power plants.11,12
only with wind power; however, actually other gen-
eration can also cause integration costs as increased
transmission and increase in contingency reserve to
Integration Studies: Issues and Setup
The specific issues investigated in the wind integration
accommodate a large unit, for example. Integration
studies vary and the methods applied have evolved
cost can sometimes also be defined as difference in
over time, with studies building upon the experience
cost compared to some alternative power production
gained in previous studies. Best practices are emerging
that could be added to a system. Cost allocation is
and models are being improved.1,13–15 The studies
another point—often system costs are allocated to all
cover different penetrations and systems and show a
users in tariffs. Part of the integration costs may be
wide range of results. The main issues studied are as
covered by the wind power producers as connection
follows:
costs and network and imbalance tariffs.
The impacts of wind power can be different for
• Impacts on balancing in different timescales:
different power systems because power system char-
any increase needed in short-term reserves or
acteristics are not the same. Adding large amounts of
ramping requirements, scheduling, and effi-
wind power plants will make the power systems more
ciency of conventional power plants;
complex and requires new tools and operational prac-
tices. Variability and uncertainty in wind generation • Impacts on grid reinforcement needs and grid
is managed mainly by using flexible resources in the stability; and
power systems.6,7 Flexibility refers to the ability to • Impacts on generation adequacy—how to
change power output level or consumption according meet the peak loads.
to the system needs when keeping the continuous bal-
ance of consumption and generation. In general, sys- Relevant issues to be taken into account when
tems where the balance needs to be kept in smaller ar- assessing the impacts of wind power on the power
eas with power plants operating with fixed schedules system are as follows14 :

244 
c 2012 John Wiley & Sons, Ltd. Volume 1, November/December 2012
 WIREs Energy and Environment Wind integration

BOX 1: WIND TURBINE CAPABILITIES ARE to areas that are part of larger areas, it is chal-
IMPROVING lenging to simulate possibilities that transmission ca-
pacity to neighboring areas can have in a way that
Wind turbines can provide support the grid. Allowing higher would not under- or overestimate the system flex-
instantaneous penetration levels means that wind power ibility in managing variability.19 Larger area stud-
plant capabilities in providing (part) of the ancillary services ies capture the impacts of wind power taking into
need to be improved. account the possibilities for cross-border trade and
One reliability concern has been that large amounts of wind balancing.20–23
power can trip off the grid because of a short disturbance of Most studies so far have concentrated on the
the grid (voltage drop).24 This problem has been addressed technical costs of integrating wind into the power sys-
by new grid connection rules (grid codes) requiring fault- tem The benefit when adding wind power to power
ride-through (FRT) capability from wind turbines.11,12,25–27 systems is reducing the total operating costs and emis-
The grid codes often require wind turbines to provide re- sions as wind replaces fossil fuels. To set the costs
active power and in some regions also to take part in volt- in scope, integration costs of wind power should be
age and frequency control.8,11,12 It should be noted, how- compared to something, such as the production costs
ever, that when providing frequency (active power) control or market value of wind power, or integration cost
(Figure 2), this will be at the expense of lost production of other generation forms.1 A fair comparison should
from the turbines. Even if wind power would probably be keep system reliability in the same level and some-
the last to provide such services, it is important to have that how take into account any changes in costs and emis-
option in the systems during high penetration events. sion savings. Cost–benefit analysis has been used in
Ireland.18

• What is the main setup for the assessment

or simulation: is wind power replacing other WIND IMPACTS ON BALANCING AND
production or capacity? Is the power sys- RESERVES
tem operation or remaining generation mix
Power systems have controllable generation to handle
optimized when wind power generation is
significant variability and uncertainty in loads over
added?
timescales from seconds to days. The generation is
• What is the wind input used: How well does scheduled first based on forecasted demand, then re-
the wind data represent the future distribu- dispatched closer to delivery with updated informa-
tion of wind power plants in the power sys- tion on forecasts, and, finally, systems carry operat-
tem area? How is wind power simulated, ing reserves that will respond either automatically or
what timescale effects on variability and pre- manually on 5–15 min notice to keep the balance of
dictability have been taken into account? It demand and generation at all times. System operators
is important to have the wind data syn- follow the change in total demand, not the variation
chronous with load data to capture their from a single generator or customer load.
correlation (see also Box 1). The variations of wind power do not correlate
• How is the uncertainty in the wind genera- with how the load varies from one time step to an-
tion forecasts handled with respect to the load other. Adding this new component of variability to a
forecast uncertainty? Are they combined us- power system will not result in just adding up the total
ing proper statistical methods? Are updated and extreme variability of both, because the extreme
forecasts for wind power closer to delivery variations are not likely to coincide. Power systems
hour taken into account? carry reserves to handle rare events occurring, for ex-
• What is the level of detail in the simulation— ample, once a year. Carrying reserve for even more
time resolution, assumptions on pricing (mar- improbable events such as the extreme of wind and
ket/technical cost)? What has been taken into load and a large outage occurring all at the same in-
account when modeling thermal and hydro stant would mean extra costs for reserve that would
units and transmission possibilities? not be used in practice. ‘Backup’ generating plants
dedicated to wind plants, or to any other generation
In smaller systems, the studies can and need plant or load, are not required, and would be quite
to be more detailed as frequency control is more costly use of power-generation resources.2,3 Example
challenging.16–18 When integration studies are made of the variability of load and the net load, the load

Volume 1, November/December 2012 

c 2012 John Wiley & Sons, Ltd. 245
Overview wires.wiley.com/wene

F I G U R E 2 | Outline of possible active power control functions from wind power plants. The plots show the possible power and the actual
achieved power when different control functions are activated (Reproduced by permission of Jesper Runge Kristoffersen.)

areas decreases the number of hours of zero output.

One wind power plant can have zero output for more
than 1000 h in a year, whereas the output of aggre-
gated wind power in a very large area is always greater
than zero. Aggregation also means that the full total
installed capacity will not be reached at any instant,
as the wind will not blow hard enough over large
areas of several hundred kilometers simultaneously.
Also, as the technical availability of single turbines
is of the order of 95% of time, there will always be
some turbines standing still when a fleet of hundreds
or thousands of turbines is considered.
The variability also decreases as the timescale
decreases. The second and minute variability of large-
scale wind power is generally small. Over several
hours, however, there can be great variability even
for distributed wind power: low wind and high wind
days can still clearly be seen (Figure 4).
Storm events can result in extreme variation
from wind power: when wind speeds are high enough
to require wind turbines to shut down from full
power, to protect the wind turbine. These events are
quite rare: usually one or two times in 1–3 years de-
pending on location. Large storm fronts take 4–6 h
to pass over several hundred kilometers so, aggrega-
F I G U R E 3 | Wind power will add to the variability that power tion of wind capacity turns the sudden interruption of
systems experience. One week of hourly data from West Denmark power into a multihour downward ramp.2 Short-term
(January 10–16, 2005) that has 24% wind penetration level on yearly forecasts of wind power are critical in managing these
basis, showing the variability of load and wind (upper graph) and
situations. Large wind-power plants can be required
resulting net load: Net load = load–wind power (lower graph). Source
to operate at partial loads during storm events to pre-
of data: http://www.energinet.dk.
vent large ramps. The impact can also be reduced by
changing the controls of wind turbines: preventing all
minus the wind power produced, is shown from real turbines from shutting down during the same minute,
data from West Denmark that has 24% wind pene- and by reducing the output more slowly as winds in-
tration level in a yearly basis (Figure 3). crease over cutout wind speeds.
Wind energy forecasting can be used to predict
wind energy variability in advance through a variety
The Variability and Uncertainty of of methods based on numerical weather-prediction
Wind Power models and statistical approaches. Wind forecast-
The relative variability of wind will decrease with the ing has been developed since the 1990s and is still
aggregation of more wind power plants (Figure 4). developing.28 The overall shape of wind generation
Aggregating wind generation over larger geographic can be predicted most of the time, but significant

246 
c 2012 John Wiley & Sons, Ltd. Volume 1, November/December 2012
 WIREs Energy and Environment Wind integration

F I G U R E 4 | Variability of wind power will smooth out with aggregation of wind power plants. Real data from Germany where the data are
from the same time period and are normalized to the mean output of each group of wind turbines. (Reproduced by permission of ISET.)

F I G U R E 5 | Predictability of wind power is better for shorter time F I G U R E 6 | Decrease of forecast error of prediction for aggregated
horizons and for larger areas/several sites. Example of average wind power production due to spatial smoothing effects. Error
absolute prediction error for a single wind-power plant and four reduction = ratio between root-mean-square error (rmse) of regional
distribute wind power plants when forecasting horizon is from one to prediction and rmse of single site, based on results of measured power
36 h ahead. generation of 40 wind farms in Germany. (Reproduced by permission
of Energy & Meteo Systems).

errors can occur in both the level and in the timing

of wind generation. Wind forecast accuracy improves capacity but drop to 6–8% for day-ahead wind fore-
for shorter time horizons (Figure 5). There is a strong casts for a single control area and to 5–7% for day-
aggregation benefit for wind forecasting, as shown in ahead wind forecasts for all of Germany [measured
Figure 6, aggregation over a 750 km region reduces as a root-mean-square error (RMSE)]. Combining dif-
forecasting error by about 50%. Typical wind fore- ferent wind forecasting models into an ensemble wind
cast errors for representative wind power forecasts for forecast also can improve wind forecasting accuracy
a single wind project are 10–15% of installed wind by up to 20%.1,28,29

Volume 1, November/December 2012 

c 2012 John Wiley & Sons, Ltd. 247
Overview wires.wiley.com/wene

F I G U R E 7 | Results for the increase in reserve requirement due to wind power, presented as percent of installed wind capacity, for different
wind penetration levels.

Both accuracy and uncertainty of short-term variations seen by the system. The range of results
forecasts, are important information for system oper- for reserve requirement increase due to wind power
ators, when allocating the reserves needed to manage is wide (Figure 7): increased reserve capacity should
the real-time operation. be 1–15% of installed wind power capacity at 10%
penetration and 4–18% of installed wind power ca-
pacity at 20% penetration.1 Timescales used in the
Frequency Control and Reserves estimation explain much of the differences in results
The reserve requirement addresses the more short- as can be seen from German results from 2010 that
term flexibility for power plants that can follow un- calculate the reserve requirement based on either day-
predicted net load variations. Wind power will also ahead, 4 h ahead or hour-ahead uncertainties.32 Ger-
increase the need for flexibility for power plants that man Dena estimates only show the average day-ahead
can follow the scheduled net load. Operational prac- uncertainty (for up and down reserves separately).26
tices, such as markets scheduling on hourly level or In Minnesota33 and California,34 day-ahead uncer-
scheduling at 5–15 min, will also have an impact tainty has been included in the estimate. UK study35
on how much reserves are needed during the op- combines the 4 h ahead variability of wind to load
erating hour. Experience shows that when reaching uncertainty–using wind forecasting would give lower
penetration levels of 5–10%, an increase in the use results.1 For others, the effect of variations during the
of short-term reserves is observed, especially for re- operating hour is considered,36 with Ireland37 and
serves activated on 10–15 min timescale. So far, no Sweden38 including the 4 h ahead uncertainty sepa-
new reserve capacity has been built specifically for rately. All studies show increasing trend of reserve
wind power.8,30 In Portugal and Spain, new pumped requirements as wind penetration increases. Even if
hydro is planned to be built to increase the flexibil- the aggregation benefits of wind power can decrease
ity of the power system, and this is mainly driven by the reserve requirements, in the studies it is often as-
wind power to enable more than 15% penetration sumed that the smoothing effect reaches its maximum
levels.11 In the highest wind penetration countries— already at lower penetration levels, and adding wind
Denmark, Spain, and Portugal—no significant fre- in the same area will no longer decrease the variability
quency impacts have been observed that are the result nor forecast error levels.
of wind-power variation.31 With high wind penetration, it will be beneficial
In wind integration studies, the increase in to allocate reserve requirements dynamically. If allo-
short-term reserve requirement is mostly estimated cation is estimated once a day for the next day instead
by comparing the reliability of the system before and of using same reserve requirement for all days, the low
after the addition of wind. A basic approach is to com- wind days will induce less requirements for the system
bine the variability or forecast errors of wind power and thus the reserve allocation can be increased only
with that of load and in some cases also power plant for days when wind variability and uncertainty are at
outages, and to investigate the increase in the largest highest.15,30

248 
c 2012 John Wiley & Sons, Ltd. Volume 1, November/December 2012
 WIREs Energy and Environment Wind integration

Increasing reserve requirement is usually calcu- Increase in balancing costs at wind penetrations
lated for the extreme cases. Even if there is increase in of up to 20% amounted to roughly 1–4 €/MWh of
reserve requirements, this does not necessarily mean wind power produced.1,33,35–37,45,47–51 If intercon-
new investments for reserve capacity, rather genera- nection capacity is allowed to be used also for balanc-
tors that were formerly used to provide energy could ing purposes, the balancing costs are lower compared
now be used to provide reserves.1 to the case where balancing is made only in the area—
increasing area size will have aggregation benefits of
wind power and also add more balancing power. The
Electricity Markets two points for Greennet Germany and Denmark45 at
There is good experience from Denmark, Spain, Ire-
the same wind penetration level reflect that balancing
land, and New Zealand with balancing wind power
costs increase when neighboring countries get more
variations through forecasting and liquid day-ahead
wind52 (Figure 8).
and balancing markets.39 For West Denmark, the bal-
Not all case studies presented results quantified
ancing cost from the Nordic day-ahead market has
as MW of increase in reserve requirements or mone-
been 1.4–2.6 €/MWh for a 24% wind penetration of
tary values for increase in balancing costs. The Irish
gross demand.1
All Island Grid Study shows that the net benefit for the
There is already some experience on how wind
power systems, going from 2 to 6 GW wind, will be
power impacts the day-ahead electricity market prices
€13/MWh, as the operational costs of the electricity
during hours with a lot of wind, the market prices
system fall compared to the base case.18
are lowered as wind energy displaces power sources
with higher marginal costs.40–42 At high wind penetra-
tions, wind power will increase the volatility in mar-
ket prices as wind energy will not always be available CURTAILMENTS OF WIND POWER
to displace higher marginal cost generators.43 In the GENERATION
long run, however, the average effect of wind energy
Challenging situations seen in system operation so
on wholesale electricity prices is not as clear because
far are from high wind-power generation during low-
the relationships between investment costs, operation
load situations, when wind power penetration lev-
and maintenance costs, and wholesale price signals
els exceed 50%.11 Wind is usually last to be cur-
will begin to influence decisions about the expansion
tailed. However, when all other units are already
of transmission interconnections, conventional gener-
at minimum (and some shut down) system opera-
ator retirement, and the type of new generation that
tors sometimes need to curtail wind power to control
is built.10,40
frequency.8 Denmark has solved part of the curtail-
ments by increasing flexible operation of combined
heat and power plants and by lowering the minimum
BALANCING COSTS OF WIND POWER generation levels used in the thermal plants.1
Balancing costs reflect increased use of reserves and In Ireland, some curtailments have been due to
less efficient scheduling of conventional power plants. concerns of low inertia53 and consequently suscep-
Impact on efficiency of conventional power plants tibility to instability in the system because of high
needs simulations of power system scheduling and instantaneous wind penetration and low system load.
dispatch and they are mostly based on comparing Currently, the issue of low inertia is unique to small
costs of system operation without wind and adding systems such as Ireland and Crete in Greece.3 In Ire-
different amounts of wind. The studies show a signif- land, they are working on solutions to go up to 75%
icant reduction of operational costs (fuel usage and penetration levels.17
costs) due to wind power even at higher penetration
levels so the integration effort will not offset the emis-
sion savings of wind power.44 To capture the integra- Does Wind Integration Need Storage?
tion cost means capturing the difference of full credit Storage is nearly always beneficial to the grid, but this
for operating cost reduction compared with cost for benefit must be weighed against its cost. All variation
system operation with efficiency penalties due to in- and uncertainty in power systems is being handled
creased variability and uncertainty. One way of cap- in power system level. This is because of lower costs
turing cost of variability is by comparing simulations when variability is aggregated before being balanced.
with flat wind energy to varying wind energy.33,45 Storage is most economic when operated to maximize
However, the two simulated cases can also result in the economic benefit for the entire system. Additional
other cost differences than just the variability cost.46 wind generation could increase the value of energy

Volume 1, November/December 2012 

c 2012 John Wiley & Sons, Ltd. 249
Overview wires.wiley.com/wene

F I G U R E 8 | Results from estimates for the increase in balancing and operating costs due to wind power for different wind penetration levels.
The currency conversion used here is 1 € = 0.7 £ and 1 € = 1.3 US$. For UK, according to a 2007 study, the average cost is presented here, the
range in the last point for 20% penetration level is from 2.6 to 4.7 €/MWh.35

storage in the grid as a whole, but storage would Wind power is normally not the only driving
continue to provide its services to the grid.2 force for grid investments but it is a major fac-
There are many studies that specifically looked tor (Ireland,59 Germany,60 Europe,61 and the United
at the cost effectiveness of electricity storage to as- States,62 ).24 The cost of grid reinforcements due
sist in integrating wind.54–58 These studies found that to wind power is very dependent on where the
for wind penetration levels up to 30% the cost effec- wind power plants are located relative to load and
tiveness of building new electricity storage is still low grid infrastructure. Portugal reported million 145 €
for other options than hydropower with large reser- (70 €/kW) increase in grid infrastructure investments
voirs and pumped hydro. With higher wind penetra- in the period 2004–2009 for increasing wind pene-
tion levels the extra flexibility that energy storage can tration from 3% (1400 MW) to 13% wind energy
provide will be beneficial for the power system op- penetration (3500 MW).24 In several studies grid re-
eration, provided it is economically competitive with inforcement costs roughly vary from 0 €/kW to 270
other forms of flexibility, e.g., from thermal power €/kW reflecting different systems, countries, grid in-
plants and demand side. frastructure, and calculation methodologies.1
A challenge for transmission planning is to re-
solve the scheduling conflict where wind plants can
WIND IMPACTS ON TRANSMISSION
be permitted and constructed in 2–3 years and it
GRID may take 5–10 years to plan, permit, and construct
Grid planning for future wind energy targets brings a transmission line. Some transitional solutions can
needs to reinforce the grid as well as building new allow wind power plants to connect to the existing
lines, both inside the country/region as well as inter- grid even if there will be times when the grid is not
connection to neighboring countries/regions. Trans- strong enough to transmit all generation produced.3
mission is important both for enabling transfer of Dynamic line ratings, taking into account the cooling
the generated electricity to loads but is also required effect of the wind together with ambient temperature
to gain aggregation benefits in variability and uncer- in determining the transmission constraints, can in-
tainty of wind power. To assess the impacts of wind crease transmission capacity and delay the need for
power to the transmission grid involves steady-state network expansion. By curtailing the generation in
load flow and transient stability simulations of the critical situations, grid equipment such as overhead
network for specific snapshots situations. Network lines or transformers can be protected from over-
contingency situations are studied to meet the criteria loads. As wind power will produce most of the time
of power system operation and safety established by at part load, the critical situations often result in only
the system operator. small production losses and in these cases it can be

250 
c 2012 John Wiley & Sons, Ltd. Volume 1, November/December 2012
 WIREs Energy and Environment Wind integration

F I G U R E 9 | Capacity credit of wind power, results from eight studies.1,26,33,66,67 The Ireland estimates were made for two power system
configurations, with 5 GW and 6.5 GW peak load.

cost effective to curtail and thus lower the connection some capacity to be relied upon during peak loads.
costs.63 However, these transitional solutions are in- Ensuring power adequacy is typically done using reli-
sufficient for large amounts of wind power and can ability analysis, which is based on loss of load proba-
result in high curtailments, such as in Texas with 17% bility (LOLP) or loss of load expectation (LOLE). The
of all potential wind energy generation curtailed in use of these approaches allows the system planner to
2009.64 determine the power adequacy level and also the con-
tribution that each generating plant makes toward
Wind Power Impacts on Stability resource adequacy, called capacity credit or capacity
There will be technical and operational implications value.2,65 The availability of high-quality chronolog-
for the power system at times of high shares of ical synchronized data that captures the correlation
wind power. The power system should sustain dis- with load data is of paramount importance and the
turbances, such as the loss of largest power plant or robustness of the calculations is highly dependent on
line so that the frequency and voltage remain stable. the volume of this data.
Wind power is asynchronous generation that does not So far, wind power has been built as additional
have the same inherent, physical support to the power generation to power systems, and thus no problems
system inertia as synchronous machines. In the small with having adequate power capacity to cover peak
island system of Ireland, the issues to enable more loads have been reported.8
than 30% wind penetration level has been studied.17 The results from studies estimating the capac-
The issues that could be mitigated were power bal- ity value for wind power range from 5% to 40%
ancing with instantaneous reserves, voltage stability, of wind-rated capacity (Figure 9). The wide range of
transient, and small-signal stability. The fundamen- capacity credit assigned to wind reflects the differ-
tal issues regarding frequency stability when losing ences in the timing of wind energy delivery (when the
a large unit or power line as well as large amounts wind blows) relative to system peak loads.66 Aggre-
of wind power tripping during network faults need gating larger areas benefits the capacity credit of wind
further analyses. power.1,67
In some reports the term ‘capacity cost’ is used.
The meaning of this is the cost for the difference
between lower capacity credit for wind power and
CAPACITY VALUE OF WIND POWER higher capacity credit for a conventional power plant.
Power system planning includes determining genera- It is not straightforward to calculate how a ‘re-
tion capacity needs for the future. Wind power is often duced value’ transforms to a cost for wind power.
considered as an energy source, but it can also provide It is important to use the lowest investment cost

Volume 1, November/December 2012 

c 2012 John Wiley & Sons, Ltd. 251
Overview wires.wiley.com/wene

generating capacity as back-up not to overestimate offset the emission savings of wind power. The capac-
this cost.68 ity value of wind power is less than for conventional
power plants and will reduce at higher penetration
levels.
Reaching high wind penetration levels chal-
CONCLUSION AND OUTLOOK
lenges generation owners and transmission operators
The natural variability of wind power makes it dif- to better utilize technology and existing assets to pro-
ferent from other generating technologies. This raises vide flexibility. More flexibility from the nonwind
questions about how wind power can be successfully generation fleet would include reduced minimum gen-
integrated into the grid. Power systems are equipped eration levels, greater ramp rates, quicker start times,
to handle variability and uncertainty that comes from and designs that allow frequent cycling without in-
the load. There is already a lot of experience of op- creasing material fatigue or reducing component life-
erating power systems with large amounts of wind times. This will require incentives or requirements and
power and integration studies have also offered valu- markets and tariffs also need to be designed to reward
able insights into how high wind penetrations can be increased flexibility. Integrating wind power genera-
successfully achieved. Model and tool development is tion into power systems can be aided by enlarging
still required to study how entire power systems can balancing areas and moving to subhourly scheduling,
be operated with occasionally high penetration levels which enable grid operators to access a deeper stack
at 60–80% of load. of generating resources and to take advantage of the
Wind integration means investments in grid in- smoothing of wind power output due to geographic
frastructure and increased use of balancing power diversity. The on-going developments on smart grids,
that can lead to efficiency losses in power plants. demand response, and plug-in hybrids as well as con-
There will be significant reduction of operational tinued improvements in new conventional-generation
costs (fuel usage and costs) due to wind power even at technologies will give new opportunities for wind in-
higher penetration levels so integration effort will not tegration.

ACKNOWLEDGMENTS
This work is part of International Energy Agency Implementing Agreement for Wind Energy
(IEAWIND) research collaboration Task 25 Design and Operation of Power Systems with Large
Amounts of Wind Power http://www.ieawind.org/AnnexXXV.html. All participating members
are greatly acknowledged for providing information on wind integration studies and experience.
Juha Kiviluoma is acknowledged for his comments on the text.

REFERENCES
1. Holttinen H, Meibom P, Orths A, van Hulle F, Power M, et al. Integration of RE into electric power
Lange B, O’Malley M, Pierik J, Ummels B, Tande JO, systems. In: Edenhofer O, Pichs-Madruga R, Sokona
Estanqueiro A, et al. Design and operation of power Y, Seyboth K, Matschoss P, Kadner S, Zwickel T,
systems with large amounts of wind power. Final re- Eickemeier P, Hansen G, Schlömer S, et al., eds. Spe-
port, IEA WIND Task 25. VTT Tiedotteita—Research cial Report on Renewable Energy Sources and Climate
Notes 2493. Phase one 2006–2008. Espoo, VTT; Change Mitigation. Cambridge/New York, NY: Cam-
2009. 200 p. + app. 29 p. Available at: http://www. bridge University Press; 2011, p. 1075.
vtt.fi/inf/pdf/tiedotteet/2009/T2493.pdf. (Accessed 4. Lundsager P, Baring-Gould I. Isolated systems with
May 14, 2012). wind power. In: Ackermann T, ed. Wind Power in
2. Milligan M, Porter K, DeMeo E, Denholm P, Holt- Power Systems. Chichester, England: John Wiley &
tinen H, Kirby B, Miller N, Mills A, O’Malley Sons; 2005, Ch. 14.
M, Schuerger M, et al. Wind power myths de- 5. Caralis G, Zervos A. Analysis of the combined use
bunked. IEEE Power Energy Mag 2009, 7:89–99. of wind and pumped storage systems in autonomous
DOI: 10.1109/MPE.2009.934268. Greek islands. Renew Power Gener 2007, 1:49–60.
3. O’Malley M, Flynn D, Holttinen H, Outhred H, van 6. International Energy Agency. Empowering variable
Hulle F, Bazilian M, Denny E, Infield D, Keane A, renewables. options for flexible electricity systems;

252 
c 2012 John Wiley & Sons, Ltd. Volume 1, November/December 2012
 WIREs Energy and Environment Wind integration

2008. Available at: http://www.iea.org/g8/2008/ 18. All Island Grid Study. Workstream 4: Analysis of Im-
Empowering_Variable_Renewables.pdf. (Accessed pacts and Benefits. Dublin, Ireland: Department of
May 14, 2012). Communications, Energy and Natural Resources and
7. International Energy Agency. Harnessing Variable Re- UK Department of Enterprise, Trade and Investment;
newables: a Guide to the Balancing Challenge. Paris, 2008, 82.
France: OECD/IEA; 2011. Available at: www.iea.org. 19. Holttinen H. Overview of integration studies—
(Accessed May 14, 2012). methodologies and results. In: Ackermann T,
8. Söder L, Hofmann L, Orths A, Holttinen H, Wan ed. Wind Power in Power Systems. 2nd Ed.
Y.-H, Tuohy A. Experience from wind integration in John Wiley & Sons; 2011 (In press). Available
some high penetration areas. IEEE Trans Energy Con- at: http://www.windpowerinpowersystems.info. (Ac-
vers 2007, 22:4–12. cessed May 14, 2012).
9. EirGrid, Annual renewable report: powering a sus- 20. EWIS European Wind Integration Study. Towards
tainable future. Dublin, Ireland: Eirgrid; 2010, 32. a successful integration of wind power in European
Available at: http://www.eirgrid.com/media/Annual electricity grids. Final Report TREN/07/FP6EN/
%20Renewable%20Report%202010.pdf. (Accessed S07.70123/038509, 2010, 182. Available at: http://
May 14, 2012). www.wind-integration.eu/downloads/library/EWIS_
10. Wiser R, O’Malley M, Infield D, Smith JC, Holttinen Final_Report.pdf. (Accessed May 14, 2012).
H. Near-term grid integration issues. In: IPCC Spe- 21. van Hulle F, Tande JO, Uhlen K, Warland L,
cial Report on Renewables (SRREN). Cambridge/New Korpås M, Meibom P, Sörensen P, Morthorst PE,
York, NY: Cambridge University Press. 2011, Cutululis N, Giebel G, et al. Tradewind Integrat-
Ch. 7.5. ing Wind, Developing Europe’s Power Market for
11. Estanqueiro A, Mateus CB, Pestana R. Operational the Large Scale Integration of Wind. Brussels, Bel-
experience of extreme wind penetrations. In: Pro- gium: Tradewind; 2009. Available at: http://www
ceedings of the 9th International Workshop on .trade-wind.eu/fileadmin/documents/publications/Final
Large-Scale Integration of Wind Power into Power _Report.pdf. (Accessed May 14, 2012).
Systems as well as on Transmission Networks 22. EWITS (Eastern Wind Integration and Transmis-
for Offshore Wind Power Plants. Québec City, sion Study). Prepared for the National Renewable
Canada: Energynautics; 2010, 34–39. Available at: Energy Laboratory by EnerNex, 2011, 242. Avail-
www.windintegrationworkshop.org. (Accessed May able at: http://www.nrel.gov/docs/fy11osti/47078.pdf.
14, 2012). (Accessed May 14, 2012).
12. Morales A, Robe X, Sala M, Prats P, Aguerri C, Tor- 23. Western Wind and Solar Integration Study. Prepared
res E, Advanced grid requirements for the integration for NREL by GE Energy. Golden, CO: National Re-
of wind farms into the Spanish transmission system, newable Energy Laboratory; 2010, 536 . Available
Renew Power Gener 2008, 2:47–59. at: http://www.nrel.gov/wind/systemsintegration/wwsis
13. Smith JC, Milligan M, DeMeo E, Parsons B. Utility .html. (Accessed May 14, 2012).
wind integration and operating impact state of the art. 24. Smith JC, Osborn D, Zavadil R, Lasher W, Gómez-
IEEE Trans Power Syst 2007, 22:900–908. Lázaro E, La Mancha C, Trötscher T, Tande
14. Söder L, Holttinen H. On methodology for modelling JO, Korpås M, Van Hulle F, et al. Transmission
power system impact on power systems. Int J Global planning for wind energy: status and prospects.
Energy Issues 2008, 29:181–198. In: Proceedings of the 9th International Workshop
15. Milligan M, Ela E, Lew D, Corbus D, Wan Y. Ad- on Large-Scale Integration of Wind Power into
vancing wind integration study methodologies: impli- Power Systems as well as on Transmission Net-
cations of higher levels of wind. In: Presented at Wind- works for Offshore Wind Power Plants. Québec City,
Power. Dallas, TX; 2010. Available as NREL/CP- Canada: Energynautics; 2010, 371–382. Available at:
550-48944, 46 p. Available at: http://www.nrel.gov/ www.windintegrationworkshop.org. (Accessed May
docs/fy10osti/48944.pdf. (Accessed May 14, 2012). 14, 2012).
16. Corbus D, Schuerger M, Roose L, Strickler J, 25. Tsili M, Papathanassiou S. A review of grid code
Surles T, Manz D, Burlingame D, Woodford D. technical requirements for wind farms. Renew Power
Oahu wind integration and transmission study: sum- Gener 2009, 3:308–332.
mary report, 2010. NREL/TP-5500-48632. Available 26. Dena. Planning of the grid integration of wind
at: http://www.nrel.gov/docs/fy11osti/48632.pdf. (Ac- energy in Germany onshore and offshore up to
cessed May 14, 2012). the year 2020 (Dena Grid study), 2005 Deutsche
17. EirGrid. All island TSO facilitation of renewables Energie-Agentur Dena, March 2005. English sum-
studies, prepared for Eirgrid by Ecofys. Dublin, mary and full German version. Available at: http://
Ireland: Eirgrid; 2010, 77 pp. Available at: http://www. www.dena.de/fileadmin/user_upload/Publikationen/
eirgrid.com/renewables/facilitationofrenewables/. (Ac- Energiedienstleistungen/Dokumente/dena-grid_study
cessed May 14, 2012). _summary.pdf. (Accessed May 14, 2012).

Volume 1, November/December 2012 

c 2012 John Wiley & Sons, Ltd. 253
Overview wires.wiley.com/wene

27. FERC Federal Energy Regulatory Commission, US. 36. Holttinen H. The impact of large scale wind power pro-
Order 661-A. Available at: http://www.ferc.gov/ duction on the Nordic electricity system. PhD disserta-
industries/electric/indus-act/gi/wind.asp. (Accessed De- tion. VTT Publications 554. Espoo, Finland; VTT Pro-
cember 12, 2005). cesses, 2004, 82 p. + app. 111 p. Available at: http://
28. Giebel G, Brownsword R, Kariniotakis G, Dernhard www.vtt.fi/inf/pdf/publications/2004/P554.pdf. (Ac-
M, Draxl C. The state of the art in short-term cessed May 14, 2012).
prediction of wind power: a literature overview. 37. ILEX, UCD, QUD, UMIST. Operating reserve require-
Deliverable D-1.2 of Anemos.plus, 2011. Avail- ments as wind power penetration increases in the Irish
able at: http://www.prediktor.dk/publ/GGiebelEtAl- electricity system. Dublin, Ireland: Sustainable Energy
StateOfTheArtInShortTermPrediction_ANEMOSplus_ Ireland; 2004.
2011.pdf. (Accessed May 14, 2012). 38. Axelsson U, Murray R, Neimane V. 4000 MW wind
29. Lange M, Focken U, Meyer R, Denhardt M, Ernst B, power in Sweden—impact on regulation and reserve
Berster F. Optimal combination of different numerical requirements. Elforsk Report 05:19, Stockholm; 2005.
weather models for improved wind power predictions. Available at: http://www.elforsk.se. (Accessed May 14,
In: Sixth International Workshop on Large-Scale Inte- 2012).
gration of Wind Power and Transmission Networks 39. Ackermann T, Ancell G, Diness Borup L, Børre Eriksen
for Offshore Wind Farms. Delft, The Netherlands; P, Ernst B, Groome F, Lange M, Möhrlen C, Orths AG,
2006. O’Sullivan J, et al. Where the wind blows. IEEE Power
30. Gil A, de la Torre M, Dominguez T, Rivas R. Influence Energy Mag 2009, 72009:65–75.
of wind energy forecast in deterministic and proba- 40. Sensfuß F, Ragwitz M, Genoese M. The merit-order
bilistic sizing of reserves. In: Proceedings of the 9th effect: a detailed analysis of the price effect of renew-
International Workshop on Large-Scale Integration of able electricity generation on spot market prices in Ger-
Wind Power into Power Systems as well as on Trans- many. Energy Policy 36:3076–3084.
mission Networks for Offshore Wind Power Plants.
41. Munksgaard J, Morthorst PE. Wind power in the Dan-
Québec City, Canada: Energynautics; 2010, 83–
ish liberalized power market—policy measures, price
88. Available at: www.windintegrationworkshop.org.
impact and investor incentives. Energy Policy 2008,
(Accessed May 14, 2012).
36:3940–3947.
31. Eto, Undrill JH, Mackin J, Daschmans P, Williams
42. Clifford E, Clancy M. 2011. Impact of wind gen-
R, Haney B, Hunt B, Ellis R, Illian J, Martinez HC,
eration on wholesale electricity costs in 2011 SEI
et al. Use of frequency response metrics to assess the
and Eirgrid, 2011. Available at http://www.seai.ie/
planning and operating requirements for reliable
Publications/Statistics_Publications/Energy_Modelling
integration of variable renewable generation.
_Group/Impact_of_Wind_Generation_on_Wholesale_
Lawrence Berkeley National Laboratory, LBNL-
Elec_Costs/Impact_of_Wind_Generation_on_Wholesa
4142E, Berkeley, CA; 2010.
le_Electricity_Costs_in_2011.pdf.
32. Dobschinski J, De Pascalis E, Wessel A, von Bremen L,
43. Jónsson T, Pinson P, Madsen H. On the market impact
Lange B, Rohrig K, Saint-Drenan Y-M. The potential
of wind energy forecasts. Energy Econ 2010, 32:313–
of advanced shortest-term forecasts and dynamic pre-
320.
diction intervals for reducing the wind power induced
reserve requirements. In: European Wind Energy Con- 44. UKERC, 2006. The costs and impacts of intermittency:
ference EWEC’2010. Warsaw, Poland; 2010. an assessment of the evidence on the costs and impacts
of intermittent generation on the British electricity net-
33. EnerNex/WindLogics 2006. Minnesota Wind Integra-
work. Imperial College, London: Energy Research Cen-
tion Study Final Report. Vol I. Prepared for Minnesota
tre; 2006. ISBN 1 90314 404 3.
Public Utilities Commission, Nov. 2006. Available
at: http://www.puc.state.mn.us/portal/groups/public/ 45. Meibom P, Weber C, Barth R, Brand H. Operational
documents/pdf_files/000664.pdf. (Accessed May 14, costs induced by fluctuating wind power production in
2012). Germany and Scandinavia. Renew Energy Gener 2009.
3:75–83.
34. Porter K, Intermittency Analysis Project Team. Inter-
mittency Analysis Project: Summary of Final Results. 46. Milligan M, Kirby B. Calculating wind integra-
California Energy Commission, PIER Research Devel- tion costs: separating wind energy value from inte-
opment & Demonstration Program. 2007. CEC-500- gration cost impacts. NREL Technical report TP-
2007-081. Available at: http://www.uwig.org/CEC 550-46275; 2009. Available at: http://www.nrel.gov/
-500-2007-081.pdf. docs/fy09osti/46275.pdf. (Accessed May 14, 2012).
35. Strbac G, Shakoor A, Black M, Pudjianto D, Bopp T. 47. Ilex Energy, Strbac G. Quantifying the system costs of
Impact of wind generation on the operation and de- additional renewables in 2020. DTI 2002.
velopment of the UK electricity systems. Electr Power 48. Zavadil R. Wind integration study for public ser-
Syst Res 77:1143–1238. vice company of Colorado. 2006. Available at:

254 
c 2012 John Wiley & Sons, Ltd. Volume 1, November/December 2012
 WIREs Energy and Environment Wind integration

http://www.uwig.org/CRPWindIntegrationStudy.pdf. http://www.eirgrid.com/media/Grid%2025.pdf. (Ac-

(Accessed May 14, 2012). cessed May 14, 2012).
49. EnerNex/WindLogics. Xcel North study (Minnesota 60. Dena Grid Study II. Integration of Renewable
Department of Commerce), 2004. Available at: Energy Sources into the German Power Supply
http://www.uwig.org/XcelMNDOCStudyReport.pdf. System Until 2020. Deutsche Energie-Agentur Dena,
(Accessed May 14, 2012). 2010. Available at: http://www.dena.de/fileadmin/user
50. Shiu H, Milligan M, Kirby B. Jackson K. 2006. Cali- _upload/Publikationen/Sonstiges/Dokumente/Summary
fornia Renewables Portfolio Standard Renewable Gen- _dena_Grid_Study_II.pdf. (Accessed May 14, 2012).
eration Integration Cost Analysis. California Energy 61. ENTSO-E: Ten-Year Network Development Plan
Commission, PIER Public Interest Energy Research 2010–2020. European network of transmission
Programme. system operators for electricity, 2010. Avail-
51. PacifiCorp. Integrated Resource Planning, 2005. Avail- able at: https://www.entsoe.eu/index.php?id=232. (Ac-
able at: http://www.pacificorp.com/es/irp.html. (Ac- cessed May 14, 2012).
cessed May 14, 2012). 62. American Transmission Company. 10 Year As-
52. Holttinen H, Meibom P, Orths A, Lange M, O’Malley sessment, 2008. Available at: www.atc10yearplan.
M, Tande JO, Estanqueiro A, Gomez E, Söder L, com. (Accessed May 14, 2012).
Strbac G, et al. Impacts of large amounts of wind power 63. Jupe SCE, Michiorri PC, Taylor A. Coordinated
on design and operation of power systems, results of output control of multiple distributed genera-
IEA collaboration. Wind Energy 2011, 14:179–192. tion schemes. Renew Power Gener 2010, 4:283–
DOI:10.1002/we.410. 297.
53. Dudurych IM. Statistical analysis of frequency re- 64. Wiser R, Bolinger M, Barbose G, Darghouth N, Hoen
sponse of island power system under increasing wind B, Mills A, Seel J, Porter K, Buckley M, Fink S. 2011
penetration. In: Power and Energy Society General Wind technologies market report. USDoE, July 2011.
Meeting. Minneapolis, MN: IEEE; 2010, 1–29. Available at: http://eetd.lbl.gov/ea/ems/reports/lbnl-
54. Ummels BC, Pelgrum E, Kling WL. Integration of large- 4820e.pdf. (Accessed May 14, 2012).
scale wind power and use of energy storage in the 65. Keane A, Milligan M, Dent CJ, Hasche B, D’Annunzio
netherlands’ electricity supply. Renew Power Gener C, Dragoon K, Holttinen H, Samaan N, Söder L,
2008, 2:34–46. DOI:10.1049/iet-rpg:20070056 O’Malley M. Capacity value of wind power. IEEE
55. Denholm P, Ela E, Kirby B, Milligan M. The role of Trans Power Syst 2011, 26:564–572.
energy storage with renewable electricity generation. 66. GE Energy. The effects of integrating wind power on
NREL Report No. TP-6A2-47187, 2010, 61 pp. Avail- transmission system planning, reliability, and opera-
able at http://www.nrel.gov/docs/fy10osti/47187.pdf. tions. Report on Phase 2, Prepared for The New York
56. Denholm P, Hand M. Grid flexibility and storage re- State Energy Research and Development Authority,
quired to achieve very high penetration of variable re- City, State; 2005. Available at: http://www.nyiso.com/
newable electricity. Energy Policy. 2011; 39. public/services/planning/special_studies.jsp. (Accessed
57. Nyamdash B, Denny E, O’Malley M. The viability of May 14, 2012).
balancing wind generation with large-scale energy stor- 67. Tande JO, Korpås M. Impact of large scale wind power
age. Energy Policy 2010, 38:7200–7208. on system adequacy in a regional hydro-based power
58. Black M, Strbac G. Value of bulk energy storage for system with weak interconnections. In: Proceedings of
managing wind power fluctuations. IEEE Trans En- Nordic Wind Power Conference NWPC, Espoo, Fin-
ergy Convers 2007, 22:197–205. land; 2006, 22–23.
59. Grid 25. A strategy for the development of Ireland’s 68. Amelin M. Comparison of capacity credit calcula-
electricity grid for a sustainable and competitive fu- tion methods for conventional power plants and wind
ture. Eirgrid, Dublin, Ireland; 2008, 47 p. Available at: power. IEEE Trans Power Syst 2006, 24:685–691.

Volume 1, November/December 2012 

c 2012 John Wiley & Sons, Ltd. 255