Conference PaperPDF Available

How much energy storage is needed to incorporate very large intermittent renewables?

Authors:

Abstract

In this paper, we present issues of electricity storage requirements based on comparative studies of various results. Studies using the datasets of Israel and California show that the storage requirement was defined by the seasonal and diurnal patterns of the local demand, and the corresponding variable renewable energy (VRE) resources profile. It was found that when we increase energy supply from VRE, the use of storage and its capacity increases until we reach some threshold. After that threshold, the storage use starts to decline even if we increase the size. An optimally utilized storage of about daily average demand would be sufficient to reach grid penetration of about 90% of the total demands from VRE at 20% total energy loss. Optimizing with other RE resources will be necessary to reach a net zero energy system instead of pushing for penetration of 100% VRE, which will require larger storage size at reduced storage usability. A loose approximation shows that the largest storage requirement for such a VRE was of the order of 6 times average daily demand with a modest increase in energy loss. A diverse Finnish 100% RE system (with 70% from VRE) was reported with energy storage size of about 8.6 times average daily demand and 6% total loss. At similar loss, the same penetration was achieved by a storage size of 0.5 times daily average demand in California, suggesting further optimization in the Finish system could result in further reduction in storage with some increase in curtailment, but might lead to higher total system cost. It was also noted that the mismatch between the VRE and load profile leads to least efficient use of resources if 100% VRE grid was aspired. However, optimal designing for VRE penetration up to 90% complemented with other renewable resources could provide an efficient energy system relying on lower storage size and balancing. We conclude that understanding of the physics and economics of the future energy system is mandatory to build and operate it optimally.
ScienceDirect
Available online at www.sciencedirect.com
Energy Procedia 135 (2017) 283–293
1876-6102 © 2017 The Authors. Published by Elsevier Ltd.
Peer-review under the responsibility of EUROSOLAR - The European Association for Renewable Energy.
10.1016/j.egypro.2017.09.520
Available online at www.sciencedirect.com
ScienceDirect
Energy Procedia 00 (2017) 000–000
www.elsevier.com/locate/procedia
1876-6102 © 2017The Authors. Published by Elsevier Ltd.
Peer-review under responsibility of the Scientific Committee of The 15th International Symposium on District Heating and Cooling.
The 15th International Symposium on District Heating and Cooling
Assessing the feasibility of using the heat demand-outdoor
temperature function for a long-term district heat demand forecast
I. Andrića,b,c*, A. Pinaa, P. Ferrãoa, J. Fournierb., B. Lacarrièrec, O. Le Correc
aIN+ Center for Innovation, Technology and Policy Research -Instituto Superior Técnico,Av. Rovisco Pais 1, 1049-001 Lisbon, Portugal
bVeolia Recherche & Innovation,291 Avenue Dreyfous Daniel, 78520 Limay, France
cDépartement Systèmes Énergétiques et Environnement -IMT Atlantique, 4 rue Alfred Kastler, 44300 Nantes, France
Abstract
District heating networks are commonly addressed in the literature as one of the most effective solutions for decreasing the
greenhouse gas emissions from the building sector. These systems require high investments which are returned through the heat
sales. Due to the changed climate conditions and building renovation policies, heat demand in the future could decrease,
prolonging the investment return period.
The main scope of this paper is to assess the feasibility of using the heat demand outdoor temperature function for heat demand
forecast. The district of Alvalade, located in Lisbon (Portugal), was used as a case study. The district is consisted of 665
buildings that vary in both construction period and typology. Three weather scenarios (low, medium, high) and three district
renovation scenarios were developed (shallow, intermediate, deep). To estimate the error, obtained heat demand values were
compared with results from a dynamic heat demand model, previously developed and validated by the authors.
The results showed that when only weather change is considered, the margin of error could be acceptable for some applications
(the error in annual demand was lower than 20% for all weather scenarios considered). However, after introducing renovation
scenarios, the error value increased up to 59.5% (depending on the weather and renovation scenarios combination considered).
The value of slope coefficient increased on average within the range of 3.8% up to 8% per decade, that corresponds to the
decrease in the number of heating hours of 22-139h during the heating season (depending on the combination of weather and
renovation scenarios considered). On the other hand, function intercept increased for 7.8-12.7% per decade (depending on the
coupled scenarios). The values suggested could be used to modify the function parameters for the scenarios considered, and
improve the accuracy of heat demand estimations.
© 2017 The Authors. Published by Elsevier Ltd.
Peer-review under responsibility of the Scientific Committee of The 15th International Symposium on District Heating and
Cooling.
Keywords: Heat demand; Forecast; Climate change
10.1016/j.egypro.2017.09.520
© 2017 The Authors. Published by Elsevier Ltd.
Peer-review under the responsibility of EUROSOLAR - The European Association for Renewable Energy.
1876-6102
Available online at
www.s cienc edirect .com
ScienceDirect
Energy Procedia 00 (2017) 000–000
www.elsevier.com/locate/procedia
1876-6102 © 2017 The Authors. Published by Elsevier Ltd.
Peer-review under the responsibility of EUROSOLAR - The European Association for Renewable Energy.
*Corresponding author E-mail: solomon.asfaw@lut.fi
11th International Renewable Energy Storage Conference, IRES 2017, 14-16 March 2017,
Düsseldorf, Germany
How much energy storage is needed to incorporate very large
intermittent renewables?
A.A. Solomon
*
, Michel Child, Upeksha Caldera, Christian Breyer
Lappeenranta University of Technology, Skinnarilankatu 34, 53850 Lappeenranta, Finland
Abstract
In this paper, we present issues of electricity storage requirements based on comparative studies of various results. It
was found that when we increase energy from VRE, the use of storage and its capacity increases until we reach some
threshold. After that threshold, the storage use starts to decline even if we increase the size. An optimally utilized
storage of about daily average demand would be sufficient to reach grid penetration of about 90% of the total demands
from VRE. The understanding of th e physics and economics of the future energy system is man datory to build and
operate it optimally.
© 2017 The Authors. Published by Elsevier Ltd.
Peer-review under the responsibility of EUROSOLAR - The European Association for Renewable Energy.
Keywords: Energy storage; Wind Energy; S olar energy; St orage Capacity;
1. Introduction
Driven by an increased interest in transitioning to low-carbon energy systems, energy storage technology has
garnered significant attention. For many experts, energy storage technology is considered one of the disruptive
technologies that could change the way we generate and consume energy. In the past decades [1-18], several research
activities dealing with some scenarios of low-carbon energy future have somehow examined the role of energy storage
technology in the corresponding systems. Many researchers [1-10] estimated storage capacity requirements for
renewable energy based grids that dominantly depends on wind and solar. These studies uses diverging methodologies
Available online at
www.s cienc edirect .com
ScienceDirect
Energy Procedia 00 (2017) 000–000
www.elsevier.com/locate/procedia
1876-6102 © 2017 The Authors. Published by Elsevier Ltd.
Peer-review under the responsibility of EUROSOLAR - The European Association for Renewable Energy.
*Corresponding author E-mail: solomon.asfaw@lut.fi
11th International Renewable Energy Storage Conference, IRES 2017, 14-16 March 2017,
Düsseldorf, Germany
How much energy storage is needed to incorporate very large
intermittent renewables?
A.A. Solomon
*
, Michel Child, Upeksha Caldera, Christian Breyer
Lappeenranta University of Technology, Skinnarilankatu 34, 53850 Lappeenranta, Finland
Abstract
In this paper, we present issues of electricity storage requirements based on comparative studies of various results. It
was found that when we increase energy from VRE, the use of storage and its capacity increases until we reach some
threshold. After that threshold, the storage use starts to decline even if we increase the size. An optimally utilized
storage of about daily average demand would be sufficient to reach grid penetration of about 90% of the total demands
from VRE. The understanding of th e physics and economics of the future energy system is man datory to build and
operate it optimally.
© 2017 The Authors. Published by Elsevier Ltd.
Peer-review under the responsibility of EUROSOLAR - The European Association for Renewable Energy.
Keywords: Energy storage; Wind Energy; S olar energy; St orage Capacity;
1. Introduction
Driven by an increased interest in transitioning to low-carbon energy systems, energy storage technology has
garnered significant attention. For many experts, energy storage technology is considered one of the disruptive
technologies that could change the way we generate and consume energy. In the past decades [1-18], several research
activities dealing with some scenarios of low-carbon energy future have somehow examined the role of energy storage
technology in the corresponding systems. Many researchers [1-10] estimated storage capacity requirements for
renewable energy based grids that dominantly depends on wind and solar. These studies uses diverging methodologies
Available online at
www.s cienc edirect .com
ScienceDirect
Energy Procedia 00 (2017) 000–000
www.elsevier.com/locate/procedia
1876-6102 © 2017 The Authors. Published by Elsevier Ltd.
Peer-review under the responsibility of EUROSOLAR - The European Association for Renewable Energy.
*Corresponding author E-mail: solomon.asfaw@lut.fi
11th International Renewable Energy Storage Conference, IRES 2017, 14-16 March 2017,
Düsseldorf, Germany
How much energy storage is needed to incorporate very large
intermittent renewables?
A.A. Solomon
*
, Michel Child, Upeksha Caldera, Christian Breyer
Lappeenranta University of Technology, Skinnarilankatu 34, 53850 Lappeenranta, Finland
Abstract
In this paper, we present issues of electricity storage requirements based on comparative studies of various results. It
was found that when we increase energy from VRE, the use of storage and its capacity increases until we reach some
threshold. After that threshold, the storage use starts to decline even if we increase the size. An optimally utilized
storage of about daily average demand would be sufficient to reach grid penetration of about 90% of the total demands
from VRE. The understanding of th e physics and economics of the future energy system is man datory to build and
operate it optimally.
© 2017 The Authors. Published by Elsevier Ltd.
Peer-review under the responsibility of EUROSOLAR - The European Association for Renewable Energy.
Keywords: Energy storage; Wind Energy; S olar energy; St orage Capacity;
1. Introduction
Driven by an increased interest in transitioning to low-carbon energy systems, energy storage technology has
garnered significant attention. For many experts, energy storage technology is considered one of the disruptive
technologies that could change the way we generate and consume energy. In the past decades [1-18], several research
activities dealing with some scenarios of low-carbon energy future have somehow examined the role of energy storage
technology in the corresponding systems. Many researchers [1-10] estimated storage capacity requirements for
renewable energy based grids that dominantly depends on wind and solar. These studies uses diverging methodologies
284 A.A. Solomon et al. / Energy Procedia 135 (2017) 283–293
2Solomon A. A.,et al./ Energy Procedia 00 (2017) 000–000
for modeling while also studying the cases at different geographic regions. As regards to modelling techniques, we
could have three major categories. Namely, (i) those estimating required energy capacity for very high shares of
renewable energy with no or little attention to the power capacity of storage [1-3]; (ii) economic models assessing
storage as a key technology in a low-carbon en ergy futur e [4-11]; and (iii) those studying factors affecting storage
design and the corresponding capacity requirements [12-15]. As regards to the diversities in geographic location, it is
possible to find studies covering several parts of the world such as entire regions (or a part) of Europe [1,2,5,9], Japan
[3], Kingdom of Saudi Arabia (KSA) [6], Asia [7], Israel [12,13], USA [4,10,14-16]. Yet, most have a different level
of emphasis on the spatial and temporal resolution as well as the share of wind and solar in the studied scenarios. It is
well known that wind and solar profiles as well as load profiles demonstrate certain common typical characteristics
globally as much as they have certain location specific characteristics. As a result, one may expect convergence
towards a certain global picture about storage requirement from these studies. Unfor tunately, this is not occurring due
to several reasons. In this paper, we will make a closer examination of various results in order to extract insights
regarding storage requirements and their design as well as options to high Variable Renewable Energy (VRE) systems.
This study analyzes results of various hourly models using physical parameters related to storage design criteria.
From several studies, we included studies dealing with the case of the Kingdom of Saudi Arabia (KSA) [10], and
Finland [9] as well as studies based on the Israeli [19] and Californian [5] grids because of the ready availability of
the result data. We have also included 2 other studies [8,4,6] (namely a study covering the eastern United states PJM
interconnection [8] and Europe [4,6]) that are based on an hourly modeling strategy and presented useful typical data
for the intended analysis. We believe th at achieving 100% VRE is not a necessity to reach to 100% RE (net zero
emission energy system). However, in order to clarify the complexity of storage design to the possible breadth and
address the case of some resource constrained regions, we will extend our discussion to 100% VRE grid or energy
systems.
2. Bases for Comparison
Energy storage modeling is currently one of the most unsettled areas due to complexities arising during modeling
of these technologies. The major challenge relates to the requirement of several abstraction parameters to correctly
characterize the technology as well as issues related to uncertainties around the operational policies and pricing of the
future grid. On top of this, large scale modeling results in simplifications that are intended to overcome computational
resource limitations. Due to the above challenges, there are several modeling approaches, which are often difficult to
compare [1-16]. The first of the above three modeling categories deals with estimating the energy capacity of the
storage by assuming a limitless capacity without any/little attention to power capacity of the storage [1-3]. Such studies
are often based on aggregated wind and solar profiles over a large geographic swat such as Europe [1,2]. Such
approaches overrun the possibility of an optimized storage with a disaggregated several node optimization model. The
second group, which is basically an economic optimization, assesses economic performance of storage in the future
energy system [4-11]. There are several versions of these categories but for our interest on large scale renewables, we
will focus on two of the sub-categories, i.e., those relying on sampled time [9,10] and those which model storage
based on the hourly time dynamics [4-8]. Due to the implementation of exogenous limits on the energy-to-power
capacity ratio of the storage and other simplifications, the optimal design of storage is still in question in most of these
models. However, those optimizing on an hourly time scale approach the problem in far-better-way than the other. In
a third group, where storage design requirements were studied (regardless of economics) by increasing renewable
energy from low pen etration to ver y high penetration, several lessons regarding factors affecting energy storage design
for very high penetration were identified. These models were crafted to study both the power and energy capacity
requirements for various levels of penetration and several high penetration enabling strategies such as energy
curtailment [12-15]. These studies identify the important constraints to be considered for a proper design of large scale
energy storage to enable high use of VREs. These constraints are the building blocks of any model that aims to better
design energy systems based on high shares of renewable energy and thus should compromise the bases for our
A.A. Solomon et al. / Energy Procedia 135 (2017) 283–293 285
Solomon A. A., et al./ Energy Procedia 00 (2017) 000–000 3
criterion for comparison. Consequently, we will first present how these constraints affect storage requirements and
briefly state the manner of our comparison in the next section.
3. Factors affecting energy storage capacity
To do a proper compar ison, it is important to know the bases for proper storage design criteria and factors affecting
it. The present consensus seems th at if larger and larger storage is available then higher and higher grid penetration of
renewables would be possible. Under such proposition, the only challenges to achieving high grid penetration with
large storage are economics. But this is half the truth because effective storage size depends on the energy and power
capacity of the storage [12-15], the nature of the local renewable energy resources [1, 2, 14, 15], the level of grid
penetration of renewables [12-15] and several other factors. Understan ding the link between these factors could help
in creating a ground for justifying our economic models performance as their result also depends on these physical
constraints. At the same time, the magnitude of certainty in the physics of the future energy system is far better than
the economic forecast. In the following, a detailed summary of (non-economic) factors affecting storage requirements
will be presented. These are:
3.1. Level of grid penetration:-
Energy storage has little role in increasing bulk penetration of renewables to the energy system at low penetration
of the variable renewable resour ces [14, 19-21]. Note that the term penetration represents the percentage of the total
demand supplied by direct VRE energy plus the stored one. But it emerges as a key player as the VRE system size
increases to a level where energy curtailment becomes prevalent [12-16]. At that time, storage plays the role of
transferring the excess energy to a later time when the generation from VRE becomes lower than the demand. Fig. 1
presents storage size requirements versus VRE penetration to the respective grids produced using a data set from
California (2011) [14, 15] and Israel (2006) [12,13]. California’s grid (which is more than 6 times larger in energy
consumption and 20 times larger in land coverage than Israel) was represented as a 12 load area system as compared
to the single load area system for Israel. The figure shows that the required storage size approximately linearly
increases with penetration as storage emerges as a key player in increasing grid penetration. However, th e increase in
grid penetration starts to level off as we increase storage capacity to accommodate more surplus VRE generation.
Once the storage reaches that turning point, further increase in capacity results in lesser and lesser grid penetration.
This could be seen from the corresponding usefuln ess index (an index that shows the ratio of the energy delivered by
storage to storage energy capacity) presented in the same figure, which shows an initially increasing use of storage
with an increase in capacity that started decreasing after reaching some peak value. In the present study, both cases
represent the condition in which little or no energy curtailment was allowed. Consequently, the above change in
storage use is attributed to the corresponding change in dispatch strategy. When relatively smaller storage capacities
are built, the storage could be dispatched more actively based on the short term generation and demand profiles
(diurnal and weekly conditions), but larger storage capacities store massive energy which could be transferred from
one season to the other often leading to large unused stored energy at the end of the year and causing their least
utilization. The above result points to the importance of aiming at finding properly sized storage while increasing its
use rather than going for a very large storage as discussed in [13]. Note that storage use starts to decrease for storage
capacity much lower than daily average demand of the local grid, specifically above a capacity of about 70% and 22%
of the daily demands of Israel and California, respectively. This was termed as peak Energy Storage Capacity (peak
EC) due to the observed peak usefulness index (UI) value corresponding to these storage capacities. The lower peak
286 A.A. Solomon et al. / Energy Procedia 135 (2017) 283–293
4Solomon A. A.,et al./ Energy Procedia 00 (2017) 000–000
EC for California was due to an effect of wind-solar complementarities and the impact of transmission lines between
the 12 load areas as compared to the solar based and single load area systems of the Israeli grid [14].
Fig. 1. Grid penetration (left axis) and UI (right) versus (Network) Energy storage Capacity. Network energy capacity represents total storage
capacit y over the entir e network.
3.2. Energy curtailment
As a result of the foregoing discussions, we can understand that applying the strict no curtailment strategy to design
a storage system with 100% VRE penetration will lead to over building of both VRE and storage systems, and leave
massive unused stored energy at the end of the year. In a model that allows energy curtailment, it was found that
models naturally starts to curtail energy in order to reduce the storage capacity requirement once it reached sufficiently
large storage (creating another backward inflection at the end of the level off region in Fig. 1). This is first reported
by [15]. In that paper, it was shown that the trend described above was observed for wind energy only scenarios after
the storage capacity reached approximately 6.6 times the daily average demand. It was noted that if further VRE size
increase was enforced, similar conditions could occur for any other wind-solar combinations. In [12-15], it was shown
that energy curtailment increases the use of energy storage to increase VRE grid penetration. For the Israeli and
Californian grids, allowing 20% total VRE energy loss (including the loss due to energy storage efficiency), lead to
grid penetration of 85% of the annual demand or better by storage size termed as “peak EC” in the above [13-15].
Under similar conditions, depending on grid size and diversity of VRE resources, a storage system with a size of daily
average demand can achieve grid penetration as high as 93% of the annual demand as in the case of California. If we
push for 100% grid penetration of VRE, the required storage size will also increase by some daily average demand.
Energy curtailment has also shown additional benefit of reducing the balancing requirement that comes from
conventional power plants, often termed as backup [2,14,15]. Note that the term “balancing” and “backup” capacity
were used by different groups to refer to the same thing; i.e. any generation resources available to fill in the shortfalls
of the VRE generators and storage. Due to some negative connotation that term “backup” is interpreted as fossil fuel
generators. We will use the term “balancing” instead. In 100% RE energy systems balancing could come from a mix
A.A. Solomon et al. / Energy Procedia 135 (2017) 283–293 287
Solomon A. A., et al./ Energy Procedia 00 (2017) 000–000 5
of generators, such as hydropower, biomass, conventional generators running on synthetic natural gas, and marine
resources, if they are co-optimized.
Fig. 2 shows how renewable energy curtailment increase leads to a reduced balancing capacity need while
increasing grid penetration. The figure shows that using the 186 GWh/22 GW storage, which at 20% total energy loss
achieves grid penetration of approximately 85% of the annual demand, the corresponding conventional balancing
capacity was reduced to 59% of the peak demand. The total energy loss stands for the loss due to curtailment plus loss
due to storage efficien cy. The loss due to storage efficiency is about 8% and 3% of the total renewable generation for
the Israeli and California grid, respectively. A decrease in curtailment could lead to an increase in the energy storage
size and vice versa, but the best option is to find the optimal condition by considering various constraints.
Fig. 2. Grid penetration (left axis) and UI (right) versus (Network) Energy storage Capacity. Network energy capacity represents total storage
capacit y over the entir e network.
3.3. Storage design and dispatch
An important question of the future renewable energy based grid is designing a proper storage system (one of the
flexibility options along with demand response, supply side management, sector coupling, transmission
interconnection, etcetera), which consists of identifying a proper storage power and energy capacity to be placed at a
given location in a power grid, and to enable an efficient use of local renewable energy resources. Proper sizing of
energy storage power and energy capacity requires the ability to capture the storage time dynamics because of their
role in matching a time varying renewable power output to a time varying load [14,15]. The hourly study of the Israeli
288 A.A. Solomon et al. / Energy Procedia 135 (2017) 283–293
6Solomon A. A.,et al./ Energy Procedia 00 (2017) 000–000
and California grid shows that the inter-link between the power and energy capacity of the storage depends on the
seasonal and diurnal interaction of VRE resource output and the local electricity demand [12-15]. Economics could
somehow move it to one or the other side based on the resource complementarity, wind/solar and storage cost, energy
curtailment, etcetera subject to these physical constraints. Storage dispatch also has its own impact on storage design
and balancing capacity requirement. As discussed above, the peak ECs are storage systems that are mostly undergoing
charge and discharge cycles in short periods of time resulting in high use. But much larger storage may transfer energy
from one season to the oth er without undergoing significant disch arging, when seen collectivel y, resulting in a limited
role in increasing grid penetration. This will be much worse if the power and energy capacity is not well matched to
the requir ed systems. This indicates that going for very large storage does not always mean large benefits both
economically and technically. As shown above, for a total energy loss at 20% of the VRE generation, a storage system
size of about daily average demand suffices to reach grid penetration of about 90% of the annual demand.
Consequently, to reach a certain penetration level it is important to push for a dispatch approach that will provide an
efficient use of all resources.
3.4. Resources complementarity
Complementarities between wind and solar were shown to give a multi-dimensional advantage to the future grid
as compared to wind/solar technologies as a stand-alone [15]. As regards storage design [15], it was shown that wind-
solar complementarities could lead to higher VRE penetration with smaller storage energy and power capacity as
compared to wind and solar as a stand-alone. For example, to arrive at 52% VRE penetration without any energy
curtailment, the required storage energy capacity for wind and solar as a stand-alone was 30 and 2 folds larger than
the corresponding storage for a 50-50 wind–solar mix, respectively [15]. Overall, it was shown that the both storage
EC and PC are smaller for the wind-solar mixture than any of the technologies as a stand-alone. Similar observation
was reported by [2] regarding the storage Energy Capacity requirement for European grid.
3.5. Relevant reliability and reserve criteria
The present supply reliability conditions, which ar e based on peak demand, may not be applicable to th e future
grid. This is due to the observation that the significant balancing needs were required outside the traditional peak load
time due to the flexible dispatching possibility of the energy storage and the matching of the renewable resources to
the local peak load times [13,15]. In addition, several issues related to the dispatch of such systems are not yet known.
Economic models that involve several energy storage technologies do not always have any clear dispatch merit order
and sometimes even a clear limit on impacts of cycling on the lifetime of the battery is missing. Lack of clarity on this
and other matters, such as market structure, may reduce the accuracy of the storage design. As a side note, we want to
remind the reader that creating a constraint that will enforce the above criteria in an economic model is not easy.
Future work is necessary in order to further understand and synchronize this into economic models.
Returning to the manner of comparison, it is important to focus on studies that are based on hourly models. From
those, we included studies dealing with the case of the Kingdom of Saudi Arabia [6, 23], and Finland [5] as well as
studies based on the Israeli [12,13] and California [14,15] grid because of the ready availability of the result data. We
have also included two other studies [2,4] (namely a study covering the eastern United states PJM interconnection [4]
and Europe [2]) that are based on an hour ly modelin g strategy and those presented useful, typical data for the intended
comparison. Yet, comparison is not easy because the studies dealt with scenarios of their own interest. For example,
the study dealing with Finland’s energy system examines the entire energy system (power, heating/cooling and
transport), while the KSA study has both a power only and integrated energy system scenario (power and desalination).
This is important as integrated energy systems covering all energy sectors will be the future reality that could provide
major flexibility, and that power only scenarios neglect. All other studies considered the case of power systems alone.
In their studies, the economic models consider various storage technologies [4-7] while others use generic storage
with assumed energy efficiencies [1,2,12-15]. As regards to the modeling technique, [4] used a model called Regional
Renewable Electricity Economic Optimization Model (RREEOM); [5] was based on EnergyPLAN; [6] was based on
large high spatial and temporal resolution model detailed in [7]; while studies covering the remaining regions [1,2,12-
A.A. Solomon et al. / Energy Procedia 135 (2017) 283–293 289
Solomon A. A., et al./ Energy Procedia 00 (2017) 000–000 7
15] rely on independent non-economic models. As a result, we have to look at typical parameters (sometimes with
simplified adjustment) that may help us to make informed comparisons to arrive at meaningful conclusions.
4. How large can storage be?
Due to complexities discussed above, there is no simple way to estimate the largest storage capacity one would
need. In this analysis, we have examined the lowest storage size based on technical studies investigating various
conditions to find alternatives while also comparing it to the highest cases approximated using economically optimized
models with projected economic parameters. In the following, we will provide a summary of important data and make
our own estimates based on principles discussed above.
Table 1. Typical parameters and corresponding values
Parameters of interest Region/country of the study
Finland
integrated
[5]
KSA
integrated
[6]
KSA
[6]
Israel
[13]
California
[14]
Europe
[2]
PJM+
[4,22]
VRE Penetration [% of annual
demand]
70 99 98 90 85 100 100
Energy storage capacity [GW h] 3990 37473 38381 113 186 16,000 891
Energy storage capacity [ daily
averag e demand ]
8.6 18.7 21.6 0.83 0.22 1.8 1.2
Total Energy loss [% of total VRE
generati on]
6 (2.5%
storage
loss)
16 (10%
storage loss)
17 (11%
storage
loss)
20 (12%
storage
loss)
20 (3%
storage
loss)
50*>50*,+*
Annual d emand du ring t he studi ed
year [TWh]
105 729 650 50.2 302 3240 276
Usefulness index [a.u.] 6.3X11.6X10X186 220 NA** 24
Storage Efficiency mix2mix2mix275% 75% 100% 81%
Share of other RE resources [% of
annual demand]
30 (hydr o,
biomass)
1
(geothermal
and bioma ss)
2
(geother
mal and
biomass)
1031530 0
Energy sectors investigated power, heat
sector
power,
desalination
power power power power power
+ this is based on their GIV storage scenario, which was selected for data completeness.
* Estimate may not include loss due to storage efficiency
+* estimate based on the given data [4], in [22] it was shown that at 95% penetration, the loss was 51%
X UI could be lower for low efficiency storage because of the resulting larger energy storage need, UI for KSA was readjusted for gas storage
(the impr ovement in UI was to ma ximu m 5 point s). For Finland, direct use of synt hetic ga s makes it less importa nt (but even i f necessary, th e
change will be a maximu m 5 point)
** At 50% curtailment, for a diverse resource such as theirs, a penetration of approximately 80% is possible, one could then see the UI will at
maximum be 40.5.
1 resource type not specified, but considered dispatchable renewable technologies (at the same time, data also show 98% VRE penetration at 25%
total energy loss as well as storage capacity of about 1.3 and 6 times daily average demand, respectively.
2 The dominant capacities and their efficiencies are power-to-gas (electrolyzer, 61% [5], 77% [6], CCGT 58% and OCGT 43%), battery (>90%),
thermal storage = 90%
Table 1 shows typical parameters collected or calculated based on the corresponding studies. The data for [5, 6, 13,
14] was obtained from the result database acquired from the authors, while the estimates corresponding to [2,4] were
290 A.A. Solomon et al. / Energy Procedia 135 (2017) 283–293
8Solomon A. A.,et al./ Energy Procedia 00 (2017) 000–000
made based on information given in the papers. Due to significant differen ce between resource qualities from place
to place, it is not easy to create an index related to VRE generation capacity. But indicators related to their matching
capability to the local demand for energy provide important insights because of similarities of the matching
characteristics [12-15,19-21]. As a result, we collect/create indicators such as VRE penetration, total energy loss,
energy storage as a fraction of daily average demand, usefulness index and storage efficiency. As seen in previous
sub-sections, these parameters are related to factors affecting storage design and use.
Table 1 shows significantly varying parameters by geographic regions covered and the different studies. If we
simply compare the locations targeting 100% or closer from VRE, we see that they correspond to significantly
divergent storage sizes. Note that even if the corresponding storage energy capacities were given by an absolute unit
of GWh, we rely on the capacity given in daily average demand for the purpose of the comparison. Considering the
power only studies, it can be seen that the energy storage capacity per grid varies from 1.2 to 22 times daily average
demand. Note that the KSA integrated scenario relied on comparatively lower storage size, due to a reduction that
occurred as a result of the added flexibility coming from sector coupling. In Budischak et al. [4], it was concluded that
economics prefers huge curtailment over increasing storage sizing and claimed that with certain policies future cost
of VRE electricity could be comparable to the present cost of electricity. However, the cost of shadow thermal backup
capacity in their study was not well accounted for due to the estimation of cost of its electricity at the present price
rate regardless of negligible dispatch time. On the contrary, the data for the KSA study shows total electricity cost
even lower to the present market though with massive storage and approximately 6% curtailment [6]. This is mainly
a consequence of n ot ignorin g energy subsidies for the current energy system and strict cost optimization of the future
100% RE system leading to a balance of RE generation, curtailment, flexibility due to sector coupling and a mix of
short- and long-term storage options. In [2], the storage requirement was assessed by aggregating the wind and solar
resources over entire Europe into one per resource type. They reported smaller storage at the cost of massive energy
curtailment and optimal wind-solar complementarity. When it comes to relatively lower penetration, the Finish study
reports VRE penetration of about 70% of the annual demand with storage energy capacity of 8.6 times daily average
demand and at 6% total energy loss while complemented by other RE generation to reach 100% RE in the energy
system. Note that the study of the Israeli and Californian grids showed a penetration of up to 90% of the annual
demand, with energy capacity less than or approximately equal to daily average demand. Caution should be made not
to make an absolute comparison between each results. For example Finland, which is in a temperate zone, has solar
resources for two-third of a year as compared to Israel and California. The solar resource in this area exists mainly
from Mar ch to October (during spring to autumn ) but the remaining half-year was well complemented by good wind
resources. Compared to California, where both wind and solar show better diurnal and seasonal complementarity, it
could be expected that by comparison the Finish system may require more storage capacity. This may be further
clarified if we compare the Finnish scenario with the achieved 73% VRE penetration for California at approximatel y
6% total loss and storage capacity of about 0.5 times daily average demand. But the possibility for improvement will
be discussed later.
Which one is correct, promoting the very high curtailment or large en ergy storage? The answer is neither. The
significant dissimilarity between results may be the evidence that shows that present economic optimizations are not
good to find a global optimal solution. The reason may be related to poor understanding of the future grid and difficulty
to accurately model some of the parameters. A good example of misunderstanding may be evident from the tendency
to design a grid for 100% VRE penetration and the corresponding result summarized in Table 1, which may be clear
from results presented in [13,14]. At such a perfect 100% VRE penetration, the system should overcome the
mismatching challenge either by building significantly large storage (very significant as compared to the 1 times daily
average capacity and 20% total loss estimated for about 90% penetration) or depend on excessive energy loss of up to
50% of the generation for an exchange in storage capacity reduction. Due to the aforementioned challenges,
researchers employ various tools to arrive at their result. For example, in a study by [4], the requirement that VRE and
storage should supply the load 99.9% of the time may have lead to an over constraining that resulted in such an excess
generation. On the contrary, the model used in [6] does not have such a constraint. Instead, it applies a priori limits
on energy to power ratio of the storage, which could limit the flexibility in arriving at the optimal storage. Though,
that approach is mandator y to correctly account for the cost of storage, at least for curr ent battery tech nologies. No
eviden ce was given on h ow the related non-linearity for such storage was overcame in [4]. Moreover, it is worth n oting
that the KSA study resulted in a system that dominantly depended on solar as compared to the dependence on wind
A.A. Solomon et al. / Energy Procedia 135 (2017) 283–293 291
Solomon A. A., et al./ Energy Procedia 00 (2017) 000–000 9
observed for the PJM area. In short, it is possible to say that a grid optimized for high shares of renewables has yet to
emerge even if the possibility of an economical renewable future is projected in those studies. This may show that
optimizing the grid may guarantee a competitive renewable future. If not, one should expect a grid which will be
extremely material intensive and may face the same sustainability question as the present grid, or alternatively face
challenges of being constructed due to a material production limit in some sector.
The ways to improve this challenge is to pursue a grid optimized for variable renewable energy systems. Yes,
curtailment is mandatory to increase grid penetration and reduce storage needs, but the level of curtailment must have
a limit. It is no surprise that it required very small storage at a curtailment value of about 50% of the total generation
as more than 80% grid penetration was reported [14] for California with out storage. It was also shown that th e physical
benefits of energy curtailment in increasing grid penetration by increasing storage use peaks at approximately 15%
total energy loss [12-15]. The reduction in balancing capacity need levels off when total energy loss approaches 20%
of the total VRE generation [15]. This puts some upper limit on its technical benefit where it starts to significantly
diminish, and the upper limit may be dependent on level of penetration. Note that poor storage efficiency may make
a little more curtailment more economical depending on VRE penetration. However, this should be seen against a
report by Caldera et al. [23], where a 100% VRE system with less curtailment and large power-to-gas storage cost
10% less than the otherwise battery dominated system. In the present studies, the designed large storage are the least
utilized (as can be seen from the UI values). However, one can conclude that well utilized storage size of about daily
average demand could have led to a grid penetration higher than 90% of the annual demand at 20% energy loss. This
shows that the solution to the above problem may require a different approach. First, there is a possibility for an
alternative mix between the two extremes even if we still push reaching closer to 100%. Second, rather than targeting
100% from VRE alone it would be better to optimize the grid to achieve approximately 90% penetration, while other
renewable resources are co-optimized to reach 100% RE (net zero emissions). Thirdly, it is important to work to arrive
at an optimal storage design and dispatch to reduce the significant curtailment or significantly larger storage and
balancing need.
Returning to the question of storage requirement, one can see that a storage approximately equal to the daily
average demand (at 20% total energy loss) is enough to arrive at grid penetration of 90% of the annual demand.
Depending on resources, a further grid penetration target may require increased storage capacity or massive energy
curtailment or a combination of both. The large storage capacity in Finland at (70% VRE penetration) may be due to
wind dr iven storage capacity during the winter season (as wind appears to prefer a higher energy to power ratio than
solar PV). But we have seen that an increase in energy curtailment could reduce the energy storage capacity need
significantly below what is currently reported but might increase the cost for the energy system. We have indicated
above that for wind alone systems some technical model started preferring curtailment over an increase in storage at
energy capacity approximately 6.6 times the daily average demand [15] (note that the corresponding penetration and
total energy loss was 57% and 0%, respectively). For the solar only Israeli grid scenario, at 25% total energy loss,
storage capacity of about 6 daily average deman d was sufficient to reach to grid penetration of about 98% of the
annual demand. Not e that the UI for such storage will be on e of th e lowest. Such storage gives a significant opportun it y
to reach 100% by using smaller other types of resources if an optimal dispatch is implemented. Thus, one may take
this storage capacity (6 times daily average demand) as a loose approximation of the largest storage required for
resource constrained regions with moderate curtailment. Even though, due to its hydro and other renewable resour ces,
California does need to push for that high VRE share as it could be used as a showcase for other resource rich areas.
The data found in the result database shows that when keeping above 25% total energy loss, a penetration higher than
98% was achieved by a storage capacity of about 1.3 times daily average demand.
In summary, the mismatch between the VRE and load profile leads to least efficient use of resources, if a 100%
renewable grid was aspired from wind and solar systems alone. However, optimal designing for VRE penetration up
to 90% complemented with other renewable resources could provide an efficient energy system relying on lower
energy storage and modest curtailment. Thus, designing a future power grid should involve broader principles as
compared to the existing market economic and “reliability of supply” lead design strategy. At the core of this principle
should be the critical material limitations of the future energy system. To properly deal with the challenge, detailed
understanding of the physics of the future grid will be mandatory to design and operate such an efficient system.
Finally, it is important to note that in the context of this paper, discussing large storage capacity as seasonal storage
292 A.A. Solomon et al. / Energy Procedia 135 (2017) 283–293
10 Solomon A. A.,et al./ Energy Procedia 00 (2017) 000–000
(since they are limited to a few days even in a loose approximation), is wron g. However, depending on the level of
penetration and technology type, seasonal dispatching may be necessary for achieving a 100% RE system which
supports the net zero emission target. It is also important to remind that the present economic model, such as the one
reported for KSA [6, 23], shows that if the forecasted economic and technological advancement goes as expected, we
may build storage as large as 22 times daily average demand.
5. Conclusions
We have made a thorough comparison between several studies evaluating storage requirements based on hourl y
models. A number of interesting lessons surfaced at the end of our analysis. First, economic models promoted high
penetration of VRE based on two varying techniques. The first one relied on large storage capacity at low energy
curtailment, while others used significantly smaller storage at the expense of massive VRE curtailment. However,
both groups reported cost of electricity comparable to the present cost. Closer study of the data and a comparison with
other results show that an energy system optimized for VRE has yet to emerge. Th e data set for the Israeli and
Californian grid studies show that a storage energy capacity of about 1 times daily average demand could suffice to
arrive at VRE penetration of approximately 90% of th e annual demand at total energy loss of about 20% of th e total
VRE generation. Depending on resources, a further VRE grid penetration target may require increased storage
capacity, massive energy curtailment or a combinations of both. For example, for the solar only Israeli grid scenario,
at 25% total energy loss, storage capacity of about 6 times daily average demand was required to reach to grid
penetration of about 98% of the annual demand while a capacity of about 1.3 times daily average demand to reach the
same penetration under the same loss requirement for the case of California. The mismatch between the VRE and load
profile lead to the least efficient use of resources if a 100% renewable grid was aspired from wind and solar systems
alone. However, optimal designing for VRE penetration up to 90% complemented with other renewable resources, in
order to reach to 100% RE (net zero emission energy system), could provide an efficient energy system relying on
lower energy storage and modest curtailment. Thus, designing a future energy systems should involve broader
principles as compared to the existing market economic and “reliability of supply” lead design strategy. At the core
of this principle of reaching to a net zero emission energy system should be a commitment to limit material intensity
of the future energy system. To properly deal with the challenge, detailed understanding of the physics of the future
energy system will be mandatory to design and operate such an efficient system.
Acknowledgements
The authors gratefully acknowledge the public financing of Tekes, the Finnish Funding Agency for Innovation, for
the ‘Neo-Carbon Energy’ project under the number 40101/14 and support from LUT internally research platform
REFLEX. The third author would like to th ank Reiner Lemoine-Foun dation for the valuable scholarship.
References
1. Heide, D., M. Greiner, L. von Bremen, C. Hoffmann. Reduced storage and balancing needs in a fully renewable European power system
with excess wind and solar power generation, Renewable Energy 36 (2011) 2515-2523
2. Heide, D., L. Bremen, M. Gr einer, C. H offmann, M. Speckmann, S. Bofinger. Seasonal optimal mix of wind and solar po wer in a future,
highly renewable Europe, Renewable Energy, 35 (2010) 2483-2489
3. Esteban,M., Q. Zhang, A. Utama, Estimation of the energy storage requirement of a future 100% renewable energy system in Japan, Energy
Policy 47 (2012) 22–31
4. Budischak C, D. Sewell, H. Thomson, L. Mach, D. E. Veron,W. Ke mpton. Cost-minimized combinati ons of wind power, solar power and
electrochemical storage, powering the grid up to 99.9% of the time. J Power Sources, 225 (2013) 60–74.
5. Child M. and Breyer C., The role of energy storage solutions in a 100% renewable Finnish energy system, Energy Procedia, 99 (2016) 25-
34
6. Caldera, U., D . Bogdanov, S. Afanasyeva, C. Breyer. Integration of reverse osmosis seawater d esalinati on in the power sector, based on PV
and wind energy, for the Kingdom of Saudi Arabia. Proceedings of 32nd European Photovoltaic Solar Energy Conference, June 20 – 24,
2016, Munich, Germany
A.A. Solomon et al. / Energy Procedia 135 (2017) 283–293 293
10 Solomon A. A.,et al./ Energy Procedia 00 (2017) 000–000
(since they are limited to a few days even in a loose approximation), is wron g. However, depending on the level of
penetration and technology type, seasonal dispatching may be necessary for achieving a 100% RE system which
supports the net zero emission target. It is also important to remind that the present economic model, such as the one
reported for KSA [6, 23], shows that if the forecasted economic and technological advancement goes as expected, we
may build storage as large as 22 times daily average demand.
5. Conclusions
We have made a thorough comparison between several studies evaluating storage requirements based on hourl y
models. A number of interesting lessons surfaced at the end of our analysis. First, economic models promoted high
penetration of VRE based on two varying techniques. The first one relied on large storage capacity at low energy
curtailment, while others used significantly smaller storage at the expense of massive VRE curtailment. However,
both groups reported cost of electricity comparable to the present cost. Closer study of the data and a comparison with
other results show that an energy system optimized for VRE has yet to emerge. Th e data set for the Israeli and
Californian grid studies show that a storage energy capacity of about 1 times daily average demand could suffice to
arrive at VRE penetration of approximately 90% of th e annual demand at total energy loss of about 20% of th e total
VRE generation. Depending on resources, a further VRE grid penetration target may require increased storage
capacity, massive energy curtailment or a combinations of both. For example, for the solar only Israeli grid scenario,
at 25% total energy loss, storage capacity of about 6 times daily average demand was required to reach to grid
penetration of about 98% of the annual demand while a capacity of about 1.3 times daily average demand to reach the
same penetration under the same loss requirement for the case of California. The mismatch between the VRE and load
profile lead to the least efficient use of resources if a 100% renewable grid was aspired from wind and solar systems
alone. However, optimal designing for VRE penetration up to 90% complemented with other renewable resources, in
order to reach to 100% RE (net zero emission energy system), could provide an efficient energy system relying on
lower energy storage and modest curtailment. Thus, designing a future energy systems should involve broader
principles as compared to the existing market economic and “reliability of supply” lead design strategy. At the core
of this principle of reaching to a net zero emission energy system should be a commitment to limit material intensity
of the future energy system. To properly deal with the challenge, detailed understanding of the physics of the future
energy system will be mandatory to design and operate such an efficient system.
Acknowledgements
The authors gratefully acknowledge the public financing of Tekes, the Finnish Funding Agency for Innovation, for
the ‘Neo-Carbon Energy’ project under the number 40101/14 and support from LUT internally research platform
REFLEX. The third author would like to th ank Reiner Lemoine-Foun dation for the valuable scholarship.
References
1. Heide, D., M. Greiner, L. von Bremen, C. Hoffmann. Reduced storage and balancing needs in a fully renewable European power system
with excess wind and solar power generation, Renewable Energy 36 (2011) 2515-2523
2. Heide, D., L. Bremen, M. Gr einer, C. H offmann, M. Speckmann, S. Bofinger. Seasonal optimal mix of wind and solar po wer in a future,
highly renewable Europe, Renewable Energy, 35 (2010) 2483-2489
3. Esteban,M., Q. Zhang, A. Utama, Estimation of the energy storage requirement of a future 100% renewable energy system in Japan, Energy
Policy 47 (2012) 22–31
4. Budischak C, D. Sewell, H. Thomson, L. Mach, D. E. Veron,W. Ke mpton. Cost-minimized combinati ons of wind power, solar power and
electrochemical storage, powering the grid up to 99.9% of the time. J Power Sources, 225 (2013) 60–74.
5. Child M. and Breyer C., The role of energy storage solutions in a 100% renewable Finnish energy system, Energy Procedia, 99 (2016) 25-
34
6. Caldera, U., D . Bogdanov, S. Afanasyeva, C. Breyer. Integration of reverse osmosis seawater d esalinati on in the power sector, based on PV
and wind energy, for the Kingdom of Saudi Arabia. Proceedings of 32nd European Photovoltaic Solar Energy Conference, June 20 – 24,
2016, Munich, Germany
Solomon A. A., et al./ Energy Procedia 00 (2017) 000–000 11
7. Bogdanov, D., and Breyer Ch. , Nort h-East Asian Super Grid for 100% renewable en ergy supply: O ptimal mix of energ y technologi es for
electricity, gas and heat supply options., Energy Conversion and Management 112 (2016) 176–190
8. Child M. and Breyer Ch., Vision and Initial Feasibility Analysis of a Recarbonised Finnish Energy System, Renewable and Sustainable
Energy Review, 66 (2016) 517-536
9. Babrowski, S, P. Jochem, W. Fichtner, Electricity storage systems in the future German energy sector,
An optimization of the German electricity generation system until 2040 considering grid restrictions. Computers & Operations Research,
66(2016) 228–240
10. Mileva A, Nelson J. H, J. Johnston, D. Kammen, Sunshot solar power reduces costs and uncertainty in future low-carbon electricity
systems. Environ Sci Technol, 47 (2013):9053–60.
11. Brouwer, A., M. van den Broek, W. Za ppa, W. C. Turkenbu rg, A. Faaij, Least-cost options for integrating intermittent r enewables in low-
carbon power systems, Applied Energy 161 (2016) 48–74
12. Solomon, A.A. D. Faiman, G. Meron. Properties and uses of storage for enhancing the grid penetration of very large-scale photovoltaic
systems. Energy Policy, 38 (2010), 5208–5222
13. Solomon, A.A., D. Faiman, G. Meron. Appropriate storage for high-penetration grid-connected photovoltaic plants, Energy Policy, 40
(2011) 335–344
14. Solomon, A.A., D.M. Kammen, D. Callaway. The role of large-scale energy storage design and dispatch in the power grid: a study of very
high grid penetration of variable renewable resources. Appl. Energy, 134 (2014), 75–89
15. Solomon, A.A., D.M. Kammen, D. Callawa y. Inv estigating the impact of wi nd–solar complementarit ies on energy stora ge requirement and
the corresponding supply reliability criteria. Appl. Energy, 168 (2016), 130–145
16. Denholm, P., a nd Margolis, R.M., E valuating the limits of solar photovoltaics (PV) in electric power systems utilizing energy stora ge and
other enabling technologies. Energy Policy 35 (2007) 4424–4433.
17. Lund, P, J. Lindgren, J. Mikkola, J. Salpakari . Review of energy syste m flexibility measures to enabl e high levels of varia ble renewa ble
electricity, Renewable and Sustainable Energy Reviews 45(2015) 785–807
18. Konziella, H and T. Bruckner, Flexibility requirements of renewable energy based electricity systems – a review of research results and
methodologies, Renewable and Sustainable Energy Reviews 53(2016) 10–22
19. Solomon, A.A., D. Faiman and G. Meron, Grid matching of large-scale wind energy conversion systems, alone and in tandem with large-
scale ph otovoltai c syst ems: An Israeli case study – Energy Policy, 38(2010) 7070-7081
20. Solomon, A.A., D. Faiman and G. Mer on. The effects on grid matching and ramping requ irements, of single and distri buted P V systems
employing various fixed and sun-tracking technologies. Energy Policy , 38(2010) 5469-5481
21. Solomon, A.A., D. Faiman and G. Meron. An energy-based evaluation of the matching possibilities of very large photovoltaic plants to
the electricity gr id: Israel as a case study. Energy Policy, 38(2010) 5457-5468
22. Pensini, A. C. N. Rasmussen, W. Kempton, Economic analysis of using excess renewable electricity to displace heating fuels. Applied
Energy 131 (2014) 530–543
23. Caldera U., Bogdanov D., Breyer C., Local cost of seawater RO desalination based on solar PV and wind energy - A global estimate,
Desalination, 385 (2016) 207-216
... Some try to find benefits of spatial distribution using one resource or more, others examine complementarity between various resources regardless of geographic distribution [20,[22][23][24][25][26][27][28][29][30]. For instance, while wind/solar complementarity studies are common, studies looking into complementarities of these resource to other renewables such as hydropower [29,31] and marine energy [28] also exists. This paper also evaluates how these benefits are manifested in various studies. ...
... The major reason for that difference is the consideration of existing non-VRE resources in the mix. As compared to a study by [5,[16][17][18]31], which shows that storage with a capacity less than daily average demand suffices to reach up to 90% with moderate dumping, this observed storage capacity is still very large. This may be because of large wind energy stored during winter (in Finland) and the low solar availability during this period. ...
Conference Paper
Resource complementarities carry significant benefit to the power grid due to their smoothing effect on the variable renewable resources output. In this paper, we show that complementarity significantly reduces energy storage requirement by using simulation results generated for Israel, Saudi Arabia, California and Finland. In a complementarity study performed using Israeli and Californian data sets (focusing on the electricity sector alone), the wind-solar complementarities were shown to significantly increase grid penetration as compared to stand-alone wind/solar systems even without the need of energy storage. At even higher grid penetration their complementarity carries significant multidimensional benefits to the power grid. The most important observation was the achievement of very high grid penetration at reduced energy storage and balancing requirements as compared to stand-alone systems. Using specific energy storage capacity (186 GWh/22 GW) and setting the solar share to 0%, 50% and 100% of the total VRE capacity, the 50-50 wind-solar capacity mix has led to significantly higher penetration as compared to the stand-alone systems. For instance, by allowing 15% energy curtailment, it was shown that grid penetration of 63%, 80% and 55% of the annual demand, respectively, can be achieved. This was because of storage being able to follow a flexible dispatch strategy, which makes it applicable for various services depending on the season of the year and the available resources. A study on a 100% renewable energy system of Finland shows that one of the best scenarios was related to a 43%-57% wind-solar capacity mix for a 70% VRE penetration by 2050. A similar study on Saudi Arabia shows that broader resource complementarity and higher level of flexibility obtained through sector coupling has reduced the required storage very significantly. The results indicate that the multiple benefits obtained from resource complementarity should be emphasized during the transition to systems of high renewable energy shares.
Technical Report
Full-text available
Technical Report "Global Energy System based on 100% Renewable Energy – Power Sector", published at the Global Renewable Energy Solutions Showcase event (GRESS), a side event of the COP23, Bonn, November 8, 2017 A global transition to 100% renewable electricity is feasible at every hour throughout the year and more cost effective than the existing system, which is largely based on fossil fuels and nuclear energy. Energy transition is no longer a question of technical feasibility or economic viability, but of political will. Existing renewable energy potential and technologies, including storage can generate sufficient and secure power to cover the entire global electricity demand by 2050 . The world population is expected to grow from 7.3 to 9.7 billion. The global electricity demand for the power sector is set to increase from 24,310 TWh in 2015 to around 48,800 TWh by 2050. Total levelised cost of electricity (LCOE) on a global average for 100% renewable electricity in 2050 is 52 €/MWh (including curtailment, storage and some grid costs), compared to 70 €/MWh in 2015. Solar PV and battery storage drive most of the 100% renewable electricity supply due to a significant decline in costs during the transition. Due to rapidly falling costs, solar PV and battery storage increasingly drive most of the electricity system, with solar PV reaching some 69%, wind energy 18%, hydropower 8% and bioenergy 2% of the total electricity mix in 2050 globally. Wind energy increases to 32% by 2030. Beyond 2030 solar PV becomes more competitive. Solar PV supply share increases from 37% in 2030 to about 69% in 2050. Batteries are the key supporting technology for solar PV. Storage output covers 31% of the total demand in 2050, 95% of which is covered by batteries alone. Battery storage provides mainly short-term (diurnal) storage, and renewable energy based gas provides seasonal storage. 100% renewables bring GHG emissions in the electricity sector down to zero, drastically reduce total losses in power generation and create 36 million jobs by 2050. Global greenhouse gas emissions significantly reduce from about 11 GtCO2eq in 2015 to zero emissions by 2050 or earlier, as the total LCOE of the power system declines. The global energy transition to a 100% renewable electricity system creates 36 million jobs by 2050 in comparison to 19 million jobs in the 2015 electricity system. Operation and maintenance jobs increase from 20% of the total direct energy jobs in 2015 to 48% of the total jobs in 2050 that implies more stable employment chances and economic growth globally. The total losses in a 100% renewable electricity system are around 26% of the total electricity demand, compared to the current system in which about 58% of the primary energy input is lost.
Article
Full-text available
A 100% renewable energy scenario was developed for Finland in 2050 using the EnergyPLAN modelling tool to find a suitable, least-cost configuration. Hourly data analysis determined the roles of various energy storage solutions. Electricity and heat from storage represented 15% of end-user demand. Thermal storage discharge was 4% of end-user heat demand. In the power sector, 21% of demand was satisfied by electricity storage discharge, with the majority (87%) coming from vehicle-to-grid (V2G) connections. Grid gas storage discharge represented 26% of gas demand. This suggests that storage solutions will be an important part of a 100% renewable Finnish energy system.
Conference Paper
Saudi Arabia is in the midst of redefining the vision for the country's future and creating an economy that is not dependent on fossil fuels. This work presents a pathway for Saudi Arabia to transition from the 2015 power structure to a 100% renewable energy based system by 2050 and analyse the benefits of integrating the power sector with the growing desalination sector. It is found that Saudi Arabia can transition to a 100% renewable energy power system by 2040 whilst meeting the growing water demand through seawater reverse osmosis (SWRO) desalination plants. The dominating renewable energy sources are PV single-axis tracking and wind power plants with 210 GW and 133 GW, respectively. The levelised cost of electricity (LCOE) of the 2040 system is 48 €/MWh. By 2050, PV single-axis tracking dominates the power sector due to the further reduction in the capital costs alongside cost reductions in supporting battery technology. This results in 80% share of solar PV in the total electricity generation. Battery storage is required to meet the total electricity demand and by 2050, accounts for 48% of the total electricity demand. The LCOE is estimated at 38 €/MWh, required capacity of PV single-axis tracking is 369 GW and wind power plants 75 GW. In the integrated scenario, due to flexibility provided by the SWRO plants, there is a reduced demand for battery storage and power-togas (PtG) plants. In addition, the ratio of the energy curtailed to the total energy generated is lower in all time periods from 2020 to 2050, in the integrated scenario. As a result, the annual levelised costs of the integrated scenario is found to be 2%-4% less than the non-integrated scenario.
Conference Paper
Integrating high shares of renewable energy (RE) sources in future energy systems requires a variety of storage solutions and flexibility measures. In this work, a 100% RE scenario was developed for Finland in 2050 for all energy sectors using the EnergyPLAN modelling tool to find a least-cost system configuration that suited the national context. Hourly data was analysed to determine the roles of various energy storage solutions, including stationary batteries, vehicle-to-grid (V2G) connections, thermal energy storage and grid gas storage for Power-toGas (PtG) technologies. V2G storage and stationary batteries facilitated use of high shares of variable RE on a daily and weekly basis. Thermal energy storage and synthetic grid gas storage aided in resolving seasonality issues related to variable RE generation plus facilitated efficient use of other forms of RE, such as biomass, and Combined Heat and Power to maintain the reliability and independence of the energy system throughout the year. In this scenario, 30 GWp of installed solar PV, 35 GWe of onshore wind power and 5 GWe of offshore wind power are supported by 20 GWh of stationary Lithium-ion batteries, 150 GWh of V2G storage (Li-ion), 20 GWhth of thermal energy storage, and 3800 GWhth of grid gas storage. Discharge of electricity and heat from storage represented 15% of end-user demand. Thermal storage discharge was 4% of end-user heat demand. In the power sector, 21% of end-user demand was satisfied by electricity storage discharge, the majority of this (87%) coming from V2G connections. Grid gas storage discharge represented 26% of gas demand. These observations suggest that storage solutions will be an important part of a 100% renewable Finnish energy system.
Article
This study demonstrates how seawater reverse osmosis (SWRO) plants, necessary to meet increasing future global water demand, can be powered solely through renewable energy. Hybrid PV–wind–battery and power-to-gas (PtG) power plants allow for optimal utilisation of the installed desalination capacity, resulting in water production costs competitive with that of existing fossil fuel powered SWRO plants. In this paper, we provide a global estimate of the water production cost for the 2030 desalination demand with renewable electricity generation costs for 2030 for an optimised local system configuration based on an hourly temporal and 0.45° × 0.45° spatial resolution. The SWRO desalination capacity required to meet the 2030 global water demand is estimated to about 2374 million m3/day. The levelised cost of water (LCOW), which includes water production, electricity, water transportation and water storage costs, for regions of desalination demand in 2030, is found to lie between 0.59 €/m3–2.81 €/m3, depending on renewable resource availability and cost of water transport to demand sites. The global system required to meet the 2030 global water demand is estimated to cost 9790 billion € of initial investments. It is possible to overcome the water supply limitations in a sustainable and financially competitive way.
Article
In order to define a cost optimal 100% renewable energy system, an hourly resolved model has been created based on linear optimization of energy system parameters under given constrains. The model is comprised of five scenarios for 100% renewable energy power systems in North-East Asia with different high voltage direct current transmission grid development levels, including industrial gas demand and additional energy security. Renewables can supply enough energy to cover the estimated electricity and gas demands of the area in the year 2030 and deliver more than 2000 TW hth of heat on a cost competitive level of 84 €/MW hel for electricity. Further, this can be accomplished for a synthetic natural gas price at the 2013 Japanese liquefied natural gas import price level and at no additional generation costs for the available heat. The total area system cost could reach 69.4 €/MW hel, if only the electricity sector is taken into account. In this system about 20% of the energy is exchanged between the 13 regions, reflecting a rather decentralized character which is supplied 27% by stored energy. The major storage technologies are batteries for daily storage and power-to-gas for seasonal storage. Prosumers are likely to play a significant role due to favourable economics. A highly resilient energy system with very high energy security standards would increase the electricity cost by 23% to 85.6 €/MW hel. The results clearly show that a 100% renewable energy based system is feasible and lower in cost than nuclear energy and fossil carbon capture and storage alternatives.
Article
Large power sector CO2 emission reductions are needed to meet long-term climate change targets. Intermittent renewable energy sources (intermittent-RES) such as wind and solar PV can be a key component of the resulting low-carbon power systems. Their intermittency will require more flexibility from the rest of the power system to maintain system stability. In this study, the efficacy of five complementary options to integrate intermittent-RES at the lowest cost is evaluated with the PLEXOS hourly power system simulation tool for Western Europe in the year 2050. Three scenarios to reduce CO2 emissions by 96% and maintain system reliability are investigated: 40%, 60% and 80% of annual power generation by RES. This corresponds to 22%, 41% and 59% of annual power generation by intermittent-RES. This study shows that higher penetration of RES will increase the total system costs: they increase by 12% between the 40% and 80% RES scenarios. Key drivers are the relatively high investment costs and integration costs of intermittent-RES. It is found that total system costs can be reduced by: (1) Demand response (DR) (2-3% reduction compared to no DR deployment); (2) natural gas-fired power plants with and without Carbon Capture and Storage (CCS) (12% reduction from mainly replacing RES power generation between the 80% and 40% RES scenarios); (3) increased interconnection capacity (0-1% reduction compared to the current capacity); (4) curtailment (2% reduction in 80% RES scenario compared to no curtailment); (5) electricity storage increases total system costs in all scenarios (0.1-3% increase compared to only current storage capacity). The charging costs and investment costs make storage relatively expensive, even projecting cost reductions of 40% for Compressed Air Energy Storage (CAES) and 70% for batteries compared to 2012. All scenarios are simulated as energy only markets, and experience a "revenue gap" for both complementary options and other power generators: only curtailment and DR are profitable due to their low cost. The revenue gap becomes progressively more pronounced in the 60% and 80% RES scenarios, as the low marginal costs of RES reduce electricity prices.
Article
It is expected that an energy system faces increasing flexibility requirements in order to cope with increasing contributions from variable renewable energy sources (VRE). In general, the instant balance of temporal and spatial inequalities of the electricity system can be achieved by many compensating measures. However, a thorough and precise quantification of the flexibility demand of a VRE based energy system turns out to be a complex task. So far, literature on energy economics and engineering has provided analyses concerning various aspects of the system requirements for flexibility. Accordingly, this review paper primarily aims to categorize the scientific approaches that have been used in "flexibility demand" studies. In this context, we classify exemplary study results from the German and European energy systems into technical, economic, and market potential categories to enhance their comparability. Moreover, we conduct a methodological evaluation of the literature findings to determine further research requirements. Against this background we also discuss a conceptual framework to quantify the market potential of flexible technologies.