Content uploaded by Debra Lew
Author content
All content in this area was uploaded by Debra Lew on Sep 07, 2016
Content may be subject to copyright.
International Comparison of
Wind and Solar Curtailment Ratio
Yo h Ya su da
Kansai University, Japan
yasuda@mem.iee.or.jp
Lori Bird
NREL, USA
Enrico Maria Carlini
Ter na, Ita ly
Ana Estanqueiro
LENG, Portugal
Damian Flynn
University College Dublin, Ireland
Alain Forcione
IREQ, Canada
Emilio Gómez Lázaro
Uni. de Castilla-La Mancha, Spain
Paraic Higgins
Queen’s University Belfast, UK
Hannele Holttinen
VTT, Finland
Debra Lew
GE Energy Management, USA
Sergio Martin-Martinez
Uni. de Castilla-La Mancha, Spain
John McCam
SEAI, Ireland
Nickie Menemenlis
IREQ, Canada
J. Charles Smith
UVIG, USA
Abstract—As the penetration of VRE (variable renewable
energy), mainly wind and photovoltaic energy, has developed
widely and rapidly, curtailment of VRE has taken on
increased interest. This paper introduces a new evaluation tool,
named the “C-P map”, that shows the correlation between
VRE curtailment ratios and energy penetration ratios of VRE
in selected countries/areas. The C-P map can illustrate
historical trends for VRE curtailment ratios in a given
country/area at a glance, and help any comparison between
historical curves. Using the C-P map, this paper classifies the
selected countries/areas into several categories, depending
upon the level and trends of VRE curtailment. The
classification helps to understand how curtailment occurred in
the past and how it may change in the future in the selected
grids.
Keywords- wind power; photovoltaic; VRE (Variable
Renewable Energy); curtailment ratio; penetration ratio
I. INTRODUCTION
The incorporation of increasing amounts of renewable
energy in electricity systems primarily from variable
renewable sources, such as wind energy and solar
photovoltaic generation, has become an energy policy
priority in a majority of countries. Ever since the possibility
of large penetrations of variable renewable generation was
first proposed, the ultimate limits on VRE (variable
renewable energy) penetration have been the subject of
research. One early wind energy study suggested that, if all
of the variable renewable output must be utilized, only
modest levels of penetration might be possible. Other
research showed that, if relatively small amounts of the
variable renewable output were foregone or “curtailed”,
significantly higher penetrations would be feasible. One
study that explored the nature of curtailment proposed that it
would increase exponentially with increasing wind energy
penetration [1]-[3].
Curtailment of VRE is now becoming an increasingly
important issue with the increasing penetrations of VRE
worldwide. Previous investigations [4]-[5] have explored
some trends in the curtailment of VRE: Italy, China and
Tex as in t he U.S . have reduced their curtailment ratios from
values in excess of 10%, experienced in the early stages of
development, while Spain and Ireland show slight increases
with increasing VRE penetration ratios, despite efforts to
keep the curtailment ratio less than 1%. After the previous
investigations, it is becoming clear that there are some
countries that have common features in curtailment trends.
It is important to monitor curtailment, as declining
curtailment trends may give evidence for the success of
measures to efficiently integrate VRE. Direct comparison of
curtailment levels in different systems is not necessarily
appropriate, as there are a range of system specific factors
that contribute to curtailment. Prior work has provided some
high level evaluation tools that may be applied to assess the
severity of the challenge for incorporating VRE within
particular power systems and allow objective comparison of
curtailment levels [6]. Söder et al. [7] proposed a “maximal
share of wind power” criterion
Share of wind power =
Max. wind power [MW]
Min. consumption [MW] + possible export [MW]
and applied this to compare wind power penetrations in
Gotland, West Denmark, Schleswig Holstein, Ireland and
New Mexico. Ackermann [8] applied this criterion to
compare wind energy penetration in Denmark, Spain,
Ireland and Texas in 2013 and 2020. Yasuda et al. [9]
proposed the flexibility radar plot as a qualitative indicator
of the flexible system resources that might allow curtailment
to be minimised in systems with high VRE penetrations.
To in vest iga te suc h curtailment trends more
quantitatively, this paper introduces a new evaluation tool,
named the “C-P map”, that shows the correlation between
VRE curtailment ratios (curtailed energy per estimated VRE
energy generated) and energy penetration ratios of VRE
(VRE energy per annual gross generation) in the selected
countries/areas.
The paper also proposes new metrics, namely the “C-P
ratio” and “C-P gradient”. The former is defined as the
quotient of the given curtailment ratio by the given
penetration ratio for the selected grid in the selected year,
whereas the latter is the gradient of the C-P curve at the
given point on the C-P map. Using the C-P map and metrics,
this paper classifies the selected countries/areas into several
categories depending upon the level and trends of VRE
curtailment. The classification helps to understand how
curtailment has occurred in the past and how it may change
in the future for the selected grids.
II. STATI ST IC S F OR VRE CURTAILMENT
In general, there are no mandated rules regarding the
publishing of unified statistical data on VRE curtailment
from TSOs (transmission system operators). The previous
papers investigated various documents created by TSOs,
regulators and national/international organisations to gather
the statistical data on curtailment of VRE in several
countries/areas.
Tab les I - IV show available statistical data for VRE
curtailment in available European countries, several RTOs
(regional transmission operators) in North America, China
and Japan, respectively. It is noted that most cases of
curtailment, excluding Japan, were associated with wind
energy. So far, there is limited information available on
curtailed PV energy across the world. In contrast, several
TABLE I. STATIS TIC AL DATA F OR VRE CURTAILMENT IN EUROPEAN COUNTRIES
Country
Year
Total Generation
(GWh)
Wind
(GWh)
PV
(GWh)
Curtailed Energy
(GWh)
Penetration
Ratio
Curtailment
Ratio
A
B
C
D
E = (B+C)/A
F = D/(B+C)
Denmark
2014
31,905
13,079
596
almost zero
42.9%
0.00%
Germany
2011
613,068
48,883
19,599
0.421
11.2%
0.61%
2012
629,812
50,670
26,380
0.385
12.2%
0.50%
2013
635,267
53,400
30,000
0.127
13.1%
0.15%
Ireland
2011
27,472
4,380
–––
0.106
15.9%
2.42%
2012
27,592
4,010
–––
0.103
14.5%
2.57%
2013
26,041
4,541
–––
0.171
17.4%
3.77%
Italy
2009
292,641
6,543
676
0.700
2.5%
9.70%
2010
302,064
9,126
1,874
0.527
3.6%
4.79%
2011
302,584
9,856
10,688
0.264
6.8%
1.29%
2012
299,277
13,407
18,637
0.166
10.7%
0.52%
2013
289,807
14,811
21,228
0.152
12.4%
0.42%
2014
277,897
14,966
23,299
0.121
13.8%
0.32%
Portugal
2014
52,886
12,103
630
0
24.1%
0.00%
Spain
2008
313,758
31,777
2,562
0.108
10.9%
0.31%
2009
294,620
36,991
5,961
0.070
14.6%
0.16%
2010
301,527
43,692
6,425
0.315
16.6%
0.63%
2011
293,848
42,160
14,882
0.073
19.4%
0.13%
2012
297,559
48,126
16,386
0.121
21.7%
0.19%
2013
285,260
54,338
17,950
1.166
25.3%
1.61%
(note) Only wind, no PV, is curtailed in European countries.
(data source) Tota l g en era ti on, wi nd an d P V e ne rgy : R ef. [10], curtailed Energy: Ref.[5].
TABLE II. STATI STI CA L DATA F OR VRE CURTAILMENT IN NORTH AMERICA
RTO
Year
Total Generation
(GWh)
Wind
(GWh)
PV
(GWh)
Curtailed
Energy (GWh)
Penetration
Ratio
Curtailment
Ratio
A
B
C
D
E = (B+C)/A
F = D/(B+C)
ERCOT
2009
N/A
N/A
N/A
N/A
6.2%
17.1%
2010
N/A
N/A
N/A
N/A
7.8%
7.7%
2011
N/A
N/A
N/A
N/A
8.5%
8.5%
2012
N/A
N/A
N/A
N/A
9.2%
3.7%
2013
N/A
N/A
N/A
N/A
9.9%
1.2%
2014
N/A
N/A
N/A
N/A
10.6%
0.5%
MISO
2010
N/A
N/A
N/A
N/A
3.6%
4.2%
2011
N/A
N/A
N/A
N/A
5.0%
3.4%
2012
N/A
N/A
N/A
N/A
7.1%
2.5%
2013
N/A
N/A
N/A
N/A
9.0%
4.6%
2014
N/A
N/A
N/A
N/A
6.1%
5.5%
Hydro-Québec
2014
216,703
6,670
–––
0
3.1%
0%
(note) Only wind, no PV, is curtailed in the U.S. and Canada.
(data source) Penetration ratio in ERCOT: Ref.[11], Penetration ratio in MISO: Ref.[12],
Curtailment ratio in ERCOT and MISO: Ref.[13]. All data in Hydro-Québec: Ref.[14].
Figure 1. C-P m ap of selected count ries (Group I)
Japanese utilities have estimated possible curtailment for
PV, due to its anticipated rapid growth in the near future.!
III. ANALYSIS USING C-P MAPS
To analy se the statistical VRE curtailment data from
various countries/areas quantitatively, the paper proposes a
new evaluation tool, named the “C-P map”, as a correlation
map between VRE curtailment ratios and VRE penetration
ratios. Figures 1, 2 and 3 show C-P maps for the selected
countries/areas grouped by their historical tendencies.
A. C-P Map of Group I: Well-operated
Figure 1 shows a group of (European) countries, i.e.
Denmark, Portugal, Spain, Ireland and Germany, where
VRE curtailment has been minimised in a “well-operated”
system management paradigm. For every country in this
group, VRE curtailment (only wind, in fact) has been
restrained to less than 4%, despite high VRE penetration
ratios. The fact that all of these countries are ranked in the
top-five countries with the highest VRE penetration ratios in
the world is not a coincidence. It can be understood that the
TSOs in these countries have suitably planned and operated
their grids to accept large volumes of VRE, with available
legislative and regulatory schemes in the last decade.
According to Spain's TSO, Red Eléctrica de España
(REE), deviations from imposed (wind) setpoints are
occurring, during curtailment periods and the recovery of
production once the limitations are released. REE has
proposed a new mechanism to minimise undesired
production losses, by performing dynamic curtailment
management, allowing control centres to modify the
setpoints according to prevailing weather conditions and
maintain production as constant as possible. In this way it
can help maintain power system operations, maximising the
integration of wind energy. REE expects to curtail 1.6 TWh
(2.2% of variable RES) in 2016, and for 2020, it is estimated
that 3.6% of wind and solar generation may be curtailed
[17].!
B. C-P Map of Group II: To b e im pr ov ed and imp ro vi ng
In contrast, Fig.2 shows a second grouping where VRE
curtailment ratios are significantly higher than those in
group I. Although the selected countries/areas, i.e. China,
Italy and two RTOs in the U.S. have relatively low
penetration ratios, the curtailment ratios are plotted in a wide
range, but less than 20%. It is interesting to note, in general,
the countries/areas show decreasing trends in their historical
C-P map curve.
These trends indicate that efforts to improve the
undesirable (curtailment) situation by the given TSOs/RTOs
has gradually borne fruit in the form of negative gradients
for their historical curve. While the fact that China and
TABLE III. STATI ST ICA L DAT A FOR VRE CURTAILMENT IN CHINA
Country
Year
Total Generation
(GWh)
Wind
(GWh)
PV
(GWh)
Curtailed
Energy (GWh)
Penetration
Ratio
Curtailment
Ratio
A
B
C
D
E = (B+C)/A
F = D/(B+C)
China
2012
4,994,038
95,978
6,366
20,820
2.0%
17.12%
2013
5,432,843
134,900
8,700
16,230
2.6%
10.74%
(note) Only wind, no PV, is curtailed in China.
(data source) Tota l g en era ti on, wi nd an d P V e ne rgy in 2 012 : R ef .[1 5], others: Ref.[5].
TABLE IV. STATI STI CA L DAT A F OR VRE CURTAILMENT IN JAPAN (FUTURE ESTIMATES)
Utility
Case
Total Generation
(GWh)
Wind
(GWh)
PV
(GWh)
Curtailed
Energy (GWh)
Penetration
Ratio
Curtailment
Ratio
A
B
C
D
E = (B+C)/A
F = D/(B+C)
Hokkaido
Current situation
(2014)
33,134 663 217 0 2.7% 0.00%
Case I: acceptable
VRE capacity
32,723 1,704 49 5.2% 2.90%
Case II: higher
penetration
32,723 3,001 394 9.2% 13.14%
Tohoku
Current (2014)
83,829
1,631
312
0
2.3%
0.00%
Case I
80,366
11,535
692
14.4%
6.00%
Case II
80,366
15,205
1,728
18.9%
11.37%
Kyushu
Current (2014)
87,783
607
992
0
1.8%
0.00%
Case I
88,736
11,773
494
13.3%
4.20%
Case II
88,736
17,473
1,919
19.7%
10.99%
(note) Only PV, no wind, will be curtailed in estimates from Japanese utilities.
(data source) Ref.[16].
Figure 4. Conceptual illustration of the C-P ratio in a C -P map
Figure 2. C-P m ap of selected countries/area s (Group II)
Figure 3. C-P m ap of selected are as (Group III)
Tex as (ER CO T) display similar trends may be coincidental,
a common cause can be surely considered. China has a
strong national movement to install wind turbines, despite a
delay in the construction of long-distance transmission lines
between the inner continental area and the high population
coastal areas. Also, Texas established a CREZ (Competitive
Renewable Energy Zone) scheme, which promotes
investment for wind power plants in inner desert areas
before the completion of transmission lines to the coastal
metropolitan area. In recent times, wind curtailment and
associated negative electricity prices in Texas have been
dramatically improved, following the gradual completion of
transmission lines in the CREZ scheme [18]. Italy also
overcame an uncomfortable situation with a high
curtailment ratio of nearly 10% in former years, to finally
achieve a low ratio less than 1%.!
C. C-P Map of Group III: Possiblly deteriorating
The final group for the C-P map relates to estimated PV
curtailment in the near future, completed by several utilities
in Japan. The country is now facing strong PV growth after
enforcement of a FIT (Feed-in Tariff) Law in 2012, while
other renewables, including wind, biomass, micro hydro and
geothermal, have not yet developed in a similar fashion, due
to a mismatch of renewable policies, including strict
environmental assessment schemes.
Several Japanese utilities announced their “acceptable
VRE capacity”, i.e. the maximum installed capacity
whereby they can curtail PV and wind energy without
compensation according to the FIT Law and the relevant
Ministerial Order. At a working group of the METI
(Ministory of Economy, Trade and Industory, Japan), held in
December 2014, the utilities published their estimated
results for the case of high PV penetration in the near future.
Figure 3 shows selected results, summarised in Ref. [8],
from the published data, and indicate the possibility of
significantly high curtailment ratios associated with future
PV development.
The results may well be overestimates, and remain a
matter of debate, given that the selected data in Table IV and
Fig. 3 represents the worst case for each utility. However,
the anouncement by the utilities and the METI resulted in
considerable negative impact on the Japanese renewables
market.
IV. C-P RAT I O A ND C-P GRADIENT
The previous section showed C-P maps for selected
countries/areas, divided into several groups according to
historical trends. Here, more quantitative analysis is
performed using new evaluation indicators, named the “C-P
ratio” and the “C-P gradient”.
A. Definition of C-P Ratio
The C-P ratio is a simple metric to show the VRE
curtailment level in a given country/area. A C-P ratio, R, is
simple defined as the quotient of the curtailment ratio C and
the energy penetration ratio P, as following:
. (1)
The C-P ratio, therefore, indicates the location of the
individual plot for a given year of the selected
countries/areas in the C-P map.
Figure 4 illustrates a conceptual map where the C-P
plane was divided into three zones, “green”, “yellow” and
“red” according to the critical C-P ratio, R. In this paper,
each zone is experimentally defined as the area where R is
less than 0.1, greater than 0.1 and less than 0.5, and greater
than 0.5, respectively. The green zone reflects the “well-
operated” countries, such as Denmark, Germany, Ireland,
Portugal, and Spain with high penetration ratios but low
curtailment ratios. The red zone indicates the “to be
improved” situation, with a high curtailment ratio despite a
R≡C
P
Figure 5. Conceptual illustration of the C-P grad ient in a C-P map
(a) closing-up correlation map (b) correlation map in wider area
Figure 6. Correlation maps between the C-P r atios and C-P gra dien ts of sele cted coun tries /areas
low penetration ratio, which includes part of the historical
plots of Group II.
B. Definition of C-P gradient
In contrast, the C-P gradient shows the historical
curtailment trends. The definition of the C-P gradient, G, is
given by the following equation:
G≡ΔC
ΔP
, (2)
where
Δ
C and
Δ
P are the backward difference of C and P,
respectively.!
The C-P gradient G indicates the trend of the historical
curve for the selected countries/areas: a negative G implies
an improving effort to reduce the curtailment ratio, a
positive G greater than 0.5 expresses a warning of a possible
deterioration in the curtailment ratio for the future. Figure 5
illustrates a sample concept with various cases of the C- P
gradient.
C. Quantitative analysis using C-P ratio and C-P gradient
Given that the groupings employed in the previous
section were qualitatively and tentatively proposed, a
quantitative analysis is now performed using the above-
proposed indicators, the C-P ratio and the C-P gradient.
Figure 6 presents the resulting analysis, as the
correlation map between the C-P ratios and C-P gradients
for the selected countries/areas. Figure 7 also illustrates an
evaluation of the correlation map linking the two indicators.
The previous sections introduce three status levels, with
“green”, “yellow” and “red” identified by the C-P ratios, and
three trends including “deteriorating”, “stable”, and
“improving”, according to the C-P gradients. Therefore,
classifications with nine categories can be defined in the
correlation map of the two indicators as follows:
(1a) red but improving,
(1b) red and stable,
(1c) red and even deteriorating,
(2a) yellow and improving,
(2b) yellow and stable,
(2c) yellow and deteriorating,
(3a) green and still improving,
(3b) green and stable,
(3c) green but deteriorating.
By matching the plots and history of the corresponding
map in Fig. 6 to the conceptual map in Fig. 7, trends for
selected countries become clear; e.g. Italy and ERCOT have
both significantly improved and finally achieved a “good”
condition, while China requires more effort for the future.
Ireland is gradually heading towards the yellow zone, as
system stability limits become increasingly active, while
Spain, similar to several utilities in Japan, will reach the red
zone in the near future without reform to their grid planning
and operations. Until now, Hydro-Québec has not conducted
any curtailment, but this could become a real possibility in
the future.
The reasons for curtailment vary among countries/areas
and have changed over time. In Ireland, curtailment is for
reliability/stability reasons. ERCOT and China were in the
process of building transmission that was not yet ready
when the wind plants were installed. MISO has network
congestion in windy regions.
Summarising the results from Fig. 6, a classification into
nine categories is given in Table V, which provides an
overview on worldwide curtailment trends by objective and
quantitative indicators, with information expressed visually
and qualitatively by the C-P maps of Figs. 1, 2 and 3.
For example, Italy and ERCOT have moved from
position (1a) to (3a) in the last five years, which indicates
that an undesirable situation has been successfully
Figure 7. Classification of Curtailment Trend
by the C-P ratio and the C-P gradien t
improved, while China remains in position (1a), and
requires further improvement. Spain has gradually fallen
into (3b) from (3a), and may reach (3c) or (2c) in the near
future. Also, Ireland is at risk of sliding into (2c) without
appropriate countermeasures (which are indeed in hand),
which follows from its synchronously isolated grid. The
latest trend for MISO was slightly surprising and needs
further investigation of what happened and how to resolve
the situation. The Japanese utilities also require effort to
reduce curtailment towards (1b) or (2b) in the near future.
V. CONCLUSIONS
This paper proposed a novel tool, C-P map, and new
indicators, C-P ratio and C-P gradient, for a quantitative
and objective evaluation of curtailment trends from VRE
(variable renewable energy; i.e. wind and solar) in an
international comparison. Using these objective and
quantitative indicators, the historical trends for VRE
curtailment were classified and compared across several
countries/areas.
Also, it would be useful to move the discussion from
curtailment being a “bad” thing, to how advantage can be
taken of curtailment to obtain more services from curtailed
VRE. For example, ERCOT achieves upward primary
frequency reserves from curtailed wind. These ancillary
services are valuable and show that curtailment does not
need to be seen as a waste of energy. Further discussions are
needed to create appropriate market structures with the
optimal VRE curtailment and to provide ancillary services
from VRE plants.
As grid conditions, energy mix and management of
power systems vary, these comparisons should take into
account the operating practices and available flexibility of
each region. Still, experience and knowledge in one
country/area can be informative to others. An international
comparison with objective measures can help gather best
practices. It is hoped that the newly proposed concepts in
this paper, C-P map, C-P ratio and C-P gradient, will enable
improved evaluation of VRE curtailment, and provide help
in planning and operating power systems in order to further
optimise curtailment of wind and solar generation in the
future.
ACKNOWLEDEMENT
This paper is part of R&D collaboration in the
International Energy Agency Wind Implementation (IEA
Wind) Task25 “Operation and Planning of Power Systems
with large amount of wind energy”.
REFERENCES
[1] E. O’Dwyer et al. “The Case for Wind Energy”, CIGRE CE/SC37
(1990)
[2] P. Gardner et al.: “The Impacts of Increased Levels of Wind
Penetration on the Electricity Systems of the Republic of Ireland and
Northern Ireland”, Commission for Energy Regulation (2003)
[3] J. Holt et al.: “Assessment of the impact of wind energy on the
CEGB system”, CEC Brussels (1990).
[4] D. Lew et al.: "Wind and Solar Curtailment", 12th Wind Int egra tion
Workshop, WIW13-1146 (2013)
[5] L. Bird et al.: “Wind and Solar Energy Curtailment: A Review of
International Experience”, Renewable and Sustainable Energy
Review [in review]
[6] H. Holttinen et al.: “Design and operation of power systems with
large amounts of wind power”, Final report of IEA Wind Task 25
Phase one 2006-08 (2009)
[7] L. Soder et al. “Experience from wind integration in some high
penetration areas”, IEEE Trans. on Energy Conversion, Vol.22,
No.1, pp.4-12 (2007)
[8] T. A ck erm an n ed.: “WIND POWER IN POWER SYSTEMS, 2nd edition”,
Wiley (2013)
[9] Y. Ya s u d a et al.: “Flexibility Chart – Evaluation on Diversity of
Flexibility in Various Areas”, 12th Win d In tegr atio n Work shop ,
WIW13-1029 (2013)
[10] International Energy Agency (IEA): “Electricity Information” (2015
web version)
[11] U.S. Energy Information Agency (EIA): “Wind generates more than
10% of Texas electricity in 2014”, today in energy (Feb. 19, 2015)
[12] Mid-Wes t In de pe nde nt System Operator (MISO): “2014 State of the
Market Report for the MISO Electricity Markets” (2015)
[13] Energy Efficiency & Renewable Energy (EERE), U.S. Department
of Energy (DOE): “2014 Wind Technologi es Market Rep ort” (2015)
[14] Hydro-Québec: unpublished data
[15] EIA: International Energy Statistics: China (web version)
[16] Y. Ya s u d a : “ International Comparison on Curtailment Ratio of Wind
and Solar Energy”, Joint Meeting of High Voltage and Environ-
mental & New Energy, IEEJ, HV-15-064 (2015) [in Japanese]
[17] S. Martín-Martínez et al.: “Wind power curtailment analysis under
generation flexibility requirements: the Spanish case study”, IEEE
PES General Meeting, 15PESGM1031 (2015)
[18] EIA: “Fewer wind curtailments and negative power prices seen in
Tex as a ft er m ajo r gr id e xp ans ion ”, today in energy (June 24, 2014)
TABLE V. CLASSIFICATION OF CURLTAILMENT TRENDS IN SELECTED COUNTRIES/AREAS
class
trend
(a)
(b)
(c)
C-P Gradie nt
G < 0
0 ≤ G < 0.5
0.5 ≤ G
status
C-P Ratio
improving
stable
deteriorating
(1)
0.5 ≤ R red
China, Italy(2010),
ERCOT(2010)
MISO Japan (in future)
(2)
0.1 ≤ R < 0.5
yellow –––– Ireland ––––
(3)
R < 0.1 green
Germany, Spain(2009),
Italy(2014), ERCOT(2004)
Hydro-Québec, Denmark,
Portugal, Spain(2013)
––––
!