Energy 116 (2016) 265e280

Contents lists available at ScienceDirect

Energy
journal homepage: www.elsevier.com/locate/energy

Modelling concentrated solar power (CSP) in the Brazilian energy

system: A soft-linked model coupling approach
Rafael Soria a, d, *, Andre

F.P. Lucena a, Jan Tomaschek b, Tobias Fichter c, Thomas Haasz b,
Alexandre Szklo a, Roberto Schaeffer a, Pedro Rochedo a, Ulrich Fahl b, Jürgen Kern c
a
Energy Planning Program, Graduate School of Engineering, Universidade Federal do Rio de Janeiro, Centro de Tecnologia, Bloco C, Sala 211, Cidade
Universita
ria, Ilha do Funda
~o, 21941-972 Rio de Janeiro, RJ, Brazil
b
Institute of Energy Economics and the Rational Use of Energy, University of Stuttgart, Hessbruehlstr. 49a, D-70565 Stuttgart, Germany
c
German Aerospace Center, Institute of Engineering Thermodynamics, Department of Systems Analysis and Technology Assessment, Pfaffenwaldring 38-40,
70569 Stuttgart, Germany
d
Department of Mechanical Engineering, Escuela Polit

n de Guevara E11-253, 17-01-2759 Quito, Ecuador
ecnica Nacional, Ladro

a r t i c l e i n f o a b s t r a c t

Article history: Brazil is looking for innovative alternatives to supply the country's increasing electricity demand and
Received 18 April 2016 provide flexibility to cope with higher shares of variable renewable energy (VRE). Concentrated solar
Received in revised form power (CSP) can help solving this double challenge. Three energy planning tools, namely MESSAGE-
16 August 2016
Brazil, TIMES-TiPs-B and REMIX-CEM-B, have been combined to analyze the opportunities that CSP
Accepted 11 September 2016
Available online 30 September 2016
plants offer to the power system and to the wider energy system of the country. This work shows that
CSP can be a cost-effective option under stringent mitigation scenarios. CSP can provide firm energy and
dispatchable capacity in the Northeast region of Brazil, optimally complementing wind and PV gener-
Keywords:
CSP
ations. Moreover, CSP can offer additional flexibility to the Northeast power system of the country,
Model coupling especially during winter, when the hydrological period is dryer. Results show synergies between CSP and
Brazil other power supply technologies with small cost differences between the baseline and CSP-forced
Energy system scenarios.
Power system © 2016 Elsevier Ltd. All rights reserved.
Energy planning tool

1. Introduction renewable energy (VRE).

Although renewable energy sources e hydropower in particular
e contribute to the major share of Brazil's electricity supply, the
remaining hydropower potential in the country is limited due to
environmental and social constraints [99]. Hence, fossil fuels,
specially coal and natural gas, are likely to play a larger role in the
Brazilian electricity system [63,77]. Building wind and photovoltaic
(PV) power plants is an alternative for Brazil to supply its growing
energy demand and reduce its greenhouse gas (GHG) emissions. A
challenge for Brazil, as well as for other countries that pursue high
penetrations of wind and PV in their power systems, is how to
supply electricity in a secure, reliable and cost effective way with
minimum GHG emissions, while providing the system with the
required flexibility that allows integrating high shares of variable

cnica Nacional, Facultad de Ingeniería
* Corresponding author. Escuela Polite

nica, Ladro
Meca
n de Guevara E11-253, 17-01-2759 Quito, Ecuador.
E-mail address: rafael.soria.energia@gmail.com (R. Soria).
Concentrated solar power (CSP) plants with thermal energy
storage (TES) and/or back-up systems can provide firm and flexible
generation capacity based on renewable energy resources to bal-
ance out unreliable and fluctuating power generation from VRE.
Within this framework, the deployment of CSP plants in Brazil is an
attractive option to achieve GHG mitigation goals and to provide
some of the additional flexibility that the Brazilian power and en-
ergy system will certainly need.
Although the Brazilian Government is interested in developing
CSP plants under the current auctions scheme [39], the official
power system expansion plan does not consider CSP plants as a
cost-effective option. One of the reasons for that is the high capital
cost of specific CSP components [67,68]. The second reason is more
related to methodological limitations during the power expansion
planning process. The latter view is one of the contributions of this
work, which is valid not only for Brazil but also for various other
countries with high CSP potential that may face problems to value
the real CSP contribution to their power systems. Currently CSP is
not adequately modeled within the energy planning tools that

http://dx.doi.org/10.1016/j.energy.2016.09.080
0360-5442/© 2016 Elsevier Ltd. All rights reserved.
266 R. Soria et al. / Energy 116 (2016) 265e280
Brazilian energy planning institutions use, especially regarding the
use of TES and back-up systems. The paradigm shift from conven-
tional hydro-thermal-generation-based power system to VRE and
CSP based systems brings various technical challenges that the
planning and operation stages in Brazil must overcome.
Adequate energy planning tools must be applied to assess the
potential role of CSP in the power system and in the wider energy
system. Connolly et al. [24], Deane et al. [27] and Hidalgo et al. [46]
assessed the ability of major models to analyze the integration of
renewable energy in power systems, finding that none of them
were able to properly address all issues involved. An analysis car-
ried out by Despre
s et al. [30] showed that, up until now, energy
modelling tools and power system tools have been designed to
meet separate objectives and have not combined their advantages.
For example, using outputs from power sector models might help
to calibrate long-term energy modelling tools. On the other hand,
long-term energy models provide economic assumptions to power
sector models. Frequently a good approach is combining, in a soft-
linked manner, tools to optimize the expansion of the power sys-
tem with tools to optimize the dispatch of a set of power plants.
Some works have been done in that sense. For example [27],
combined the TIMES energy model with the PLEXOS tool for power
systems to improve and develop understanding of energy systems
with high penetration of wind power. Sullivan et al. [113] devel-
oped capacity reserve and flexibility requirements equations in the
MESSAGE energy model using information generated in a NREL
unit commitment model.
Part of the motivation of this work is derived from a view that
greater insights and progress can be gained by profiting from the
strengths of multiple modelling tools rather than trying to incor-
porate them all into one single comprehensive model. A literature
review did not show a soft-link between energy models and a
power system tool able to model CSP plants in a detailed way, to
understand the real flexibility that this technology may offer to
power systems. Additionally, few scientific studies have analyzed
the role of CSP in the Brazilian energy and power systems,
combining long-term expansion and short-term operation points of
view. For example [67,68], modeled the long-term expansion of the
Brazilian power and energy system until 2050, forcing CSP capac-
ities in the power system according to solar incentives scenarios.
Castro [19] optimized the dispatch of the Brazilian power system
projected to 2022 according to the official energy expansion plan
[38], forcing CSP capacities in the power system. Nevertheless,
there is not a comprehensive study using a combination of
expansion and operation energy planning tools to verify the role of
CSP for the Brazilian energy sector. The work presented in the paper
at hand helps to fill this gap.
Therefore, the objective of this study is to apply a combination of
long-term expansion and short-term operation energy planning
tools to assess the role of worldwide mature and commercial CSP
technologies as part of the future Brazilian energy system and to
understand CSP capabilities to provide an extra source of flexibility
to the Brazilian power system that allows integrating large shares of
VRE. Thus, this work assesses with an appropriate combination of
tools the synergies between CSP mature technologies and others
energy supply options, especially for electricity supply.
Three different tools are combined to provide further insight
about the possible synergies between CSP and other power gen-
erators in the Brazilian energy system: MESSAGE, TIMES and
REMIX-CEM. MESSAGE-Brazil has been applied to evaluate CSP
options in the expansion of the energy system as a whole. TIMES-
TiPS-B was developed to evaluate the role of CSP in the expan-
sion of the power sector. REMix-CEM is an appropriate tool to
model CSP technology. Instead of using REMix-CEM in its original
application as a detailed power sector specific capacity expansion
model [41,42], the developed REMix-CEM-B was applied as a unit
commitment optimization tool, to analyze the operation of the
proposed generator fleet (provided by the expansion tools),
focusing on the role of the CSP plants to the power dispatch.
Within the worldwide mature and commercial CSP technolo-
gies, parabolic trough collector (PTC) plants are the most developed
and most mature of all CSP commercially operating plants
[7,10,58,61,124]. The solar power tower (SPT) is the second most
deployed technology and is of increasing importance as a result of
its advantages [9,18,124]. The main SPT advantage is that its DNI
concentrating ratio achieves very high working temperatures,
thereby increasing the thermodynamic efficiency at which heat is
converted into electricity, and reducing the cost of TES. The use of
molten salt, typically a mixture of sodium and potassium nitrates
(KNO3 and NaNO3), as heat storage media in PTC and SPT systems is
worldwide common [58,81,124]. In fact, molten salt can be used as a
heat carrier and storage medium, especially in SPT plants due to its
high process working temperature and the proximity of molten salt
tanks to the central receiver, decreasing the TES cost [18,90,91,124].
Several studies present novel CSP configurations to increase ther-
mal power cycle efficiencies and reduce the levelised cost of elec-
tricity (LCOE). To improve power block efficiency it is necessary to
increase the process temperature. One option is the utilization of
water/steam as heat transfer medium (HTF), the so-called direct
steam generation (DSG) [40,47,124]. DSG can be used in PTC, SPT
and Fresnel. There are also advancements in novel treatments for
TES [61,95,121]. Heat can be stored in either sensible, latent or
thermochemical storage. Latent heat storage, using phase change
materials (PCMs), and thermo-chemical heat storage (TCS) are still
at an early stage of laboratory and pilot research despite their
attractive application for long term energy storage [121]. The use of
powders as HTF for heat capture, conveying and storage in CSP-SPT
plants is another important research-topic [121e123]. Preliminary
assessments from tests at pilot scale confirm the advantages of the
upflow bubbling fluidized bed and demonstrate that particulate
suspensions enable major savings in investment and operating
costs (increased power cycle efficiency, an increased storage den-
sity, reduced thermal power requirements, reduced heliostat field
size, reduced parasitic power consumption and increased plant
capacity factor) [121e123]. Additionally, different renewable fuels
and configurations to hybridize CSP plants are being studied, for
example solid wood ([26,109]; Soria et al., 2015), municipal and
industrial residues [92e94], sugar cane bagasse [15,16] and bio-
methane or syngas (from biomass gasification) to fuel Brayton
turbines in novel SPT plants [3,25,96].
In this work, we consider as the reference CSP technology the
worldwide most mature and commercial one, which is based on
PTC, wet cooling and use of synthetic oil as HTF in the primary
circuit. The reason for this choice, identified as one of the chal-
lenges for the Brazilian power sector, is presented in Section 2.
Thus, the novelty of this work lies on the soft-link between
REMIX-CEM-B, which is a power system modelling tool appropriate
for modelling CSP plants with TES and back-up systems using high
temporal and spatial resolution, with an integrated energy plan-
ning tool (MESSAGE-Brazil) and a power sector expansion tool
(TIMES-TiPS-B). Providing accurate and recent data for modelling
the Brazilian power sector is also a contribution of this work.
Although the case study focuses on Brazil, the methodological
approach proposed here can be applied to other realities with
similar challenges. As Brazil is one of the main protagonists in the
global climate policy discussions [102], what happens in the Bra-
zilian energy sector is relevant and worth studying.
The next section presents the challenges for the Brazilian power
system related to the increase of VRE and the contribution of CSP
technology. Then, Section 3 thereafter presents a brief theory
 R. Soria et al. / Energy 116 (2016) 265e280
framework about the impacts caused by high VRE penetration to
the grid and the contribution of CSP to minimize those impacts.
Section 4 presents the basic data and methods of the study, de-
tailing which were the participating models and how the model
coupling was performed. Main data and assumptions to model the
Brazilian energy and power system are also described in Section 4.
Main results and discussion are presented in Section 5. Finally,
Section 6 assesses the possible role of CSP in Brazil.
2. A challenge for the Brazilian power sector: high shares of
VRE and CSP
The Northeast is the Brazilian region where VRE has grown
faster due to its large potential and where lies most of the
country's CSP potential. For these reasons, this subsection will
provide more details about the Northeast power system in Brazil,
illustrating the problems that the country faces. As to today, this
region's wind power installed capacity totals 5.4 GW and the
prospects, considering all wind farms contracted by the end of
August 2015 by regular auctions, is to install at least some
13.3 GW more by 2019 [2,23,35]. Although the PV installed ca-
pacity in the Northeast region by December 2014 was negligible
(4.6 MWp), this technology shows promising expansion pros-
pects in the country in the long term. Recent auctions held in
Brazil (August 2015) contracted a total capacity of 1.3 GWp of
centralized PV in the Northeast subsystem to begin operation
between 2016 and 2017 [2,23,35]. A significant deployment of
distributed generation (DG) based on PV systems is also expected
in the medium to long terms. On the other hand, the current
electric power system of Brazil's Northeast region is character-
ized by low flexibility due to high penetration of inflexible power
plants, such as coal-fired power plants and non-flexible natural
gas combined cycles (CCGT), including cogeneration of heat and
power (CHP) (See Table A1). Additionally, hydroelectric power
plants in the Northeast have low reservoir capacity, which also
creates difficulties for the integration of large quantities of VRE
generation. In the long term, it is expected that the Northeast
subsystem will increasingly rely on inflexible power plants, due
to the planned construction of run-of-the river hydroelectric
power plants and possibilities of building coal-fired power plants
[37,38]. Moreover, until very recently there had been some dis-
cussions about the possibility of constructing nuclear power
plants in the Northeast, lowering the flexibility of the system.
However, this possibility was not considered in the Brazilian
official expansion plan until 2022 [38]. Thus, Brazil is rapidly
increasing the participation of VRE in its power generation mix,
which can boost problems1 in the power system with frequency
response and operational reserve services.
On the other hand, Brazil has a large area (97,700 km2) with
good technical potential for CSP deployment. Burgi [14] analyzed
the Brazilian techno-economic CSP potential, in terms of installable
capacity, and concluded that it varies between 85 GW and 346 GW,
according to the CSP configuration. Approximately 70% of this po-
tential is located in the Northeast region (See Fig. 1), which justifies
its selection in our study, where also most of the country's wind
projects are currently installed and are expected to be installed in
the short to medium term. Although the techno-economic CSP
potential in Brazil is huge, so far there is no CSP project operating in
the country. Brazilian universities and research centers are just
starting to investigate and develop small CSP pilot projects. Gov-
ernment has no official plan to develop CSP at a commercial scale.
1
Section 3 describes the impacts to power systems associated to high penetra-
tion of VRE.
267
Thus, only mature and worldwide commercial CSP technology,
whose financial risk is low, could play a role in the short and mid-
terms in Brazil. Although the reference CSP technology considered
in this study (PTC, wet cooled, using synthetic oil as HTF in the
primary circuit) is not the most advanced one, it has proven to be
bankable worldwide. This mature technology would play an
important role during the CSP introductory phase in Brazil, devel-
oping a solid CSP market and creating the conditions that can
ensure a major role for novel and more efficient CSP technologies in
the long-term.
3. A challenge for CSP technology: providing additional
flexibility to power systems with high penetration of VRE
Significant increases in centralized VRE sources, like central-
ized wind and PV, may threaten the adequacy (in generation and
flexibility) and reliability of power systems [31,32,66,104]. Some
of the local impacts (very much related to distribution systems)
occur on voltage control, fault current, harmonic distortion and
flicker [13,82,97,107,120]. System-wide impacts on power systems
include the imbalance between load and generation, reactive
power generation, and reduced frequency control
[49e51,107,108]. To cope with these impacts, power systems will
need to have sufficient flexibility to maintain the demand-
generation balance at all time (frequency response services),
and operational reserve services at a reasonable cost over
different time scales [66,103].
Traditionally, flexibility to meet peak load in short time scales
is provided by partially loaded synchronised plants and fast start-
up plants, such as open-cycle gas turbines (OCGT). Besides flex-
ible generation, a modern point of view considers alternative
flexibility sources such as flexible demand, flexible energy stor-
age, strengthening of cross border interconnections, grid expan-
sion, flexible district heating and flexible heat pumps
[12,69e71,112,126,127]. Several types of energy storage technol-
ogies are being largely researched, for example power-to-gas
(P2G), pumped hydro storage (PSH), compressed air energy
storage (CAES), flywheel energy storage system (FESS), hydrogen-
based energy storage system (HESS), TES in CSP plants and bat-
teries in plug-in hybrid electric vehicles (PHEV) and pure electric
vehicles (EVs) [12,107,116,126,127].
Actually, TES in CSP plants have a high market potential for
increasing flexibility in power systems [5,62,116]. It facilitates en-
ergy dispatch management during the whole day (depending on
the TES size), and increases the capacity factor of the system and its
ability to meet peak loads, as well as to operate in the baseload with
firm energy from solely solar origin [4,7,105,116]. In addition, with
the same power block, CSP plants can operate as hybrid systems
that employ different configurations [55,74,125]. Hybridization
enables the solar plant to operate partially using a back-up fuel [79],
that can be fossil (typically natural gas) or non-fossil (biomass or
biogas) [5,92,109,116]. In sum, CSP plants with TES and/or back-up
systems can provide the firm and flexible generation capacity based
on renewable energy resources to balance out unreliable and
fluctuating power generation from VRE.
4. Basic data and methods
4.1. Participating models and model coupling
Combining tools to optimize the expansion of the power and
energy system with tools to optimize the dispatch of power plants is a
good approach to improve the modelling of VRE and the integration
of flexibility sources to power systems [46]. This is the main reason
why three different energy planning tools (MESSAGE-Brazil, TIMES-
268 R. Soria et al. / Energy 116 (2016) 265e280
Fig. 1. High CSP potential regions in Northeast Brazil.
TiPS-B and REMix-CEM-B) have been combined in a soft linked2
multi-model network to analyze the penetration of CSP in the Bra-
zilian energy system (See Fig. 2). The Energy Planning Program (PPE/
COPPE), from the Universidade Federal do Rio de Janeiro (UFRJ), used
the Integrated Energy Planning Model MESSAGE-Brazil. The Institute
of Energy Economics and the Rational Use of Energy (IER), from the
University of Stuttgart, used the TIMES-TiPS-B (The Integrated
MARKAL EFOM System - Power System Model for Brazil) planning
tool. The Department of Systems Analysis and Technology Assess-
ment of the Institute of Engineering Thermodynamics, from the
German Aerospace Center (DLR Stuttgart), used REMix-CEM-B
(Renewable Energy Mix - Capacity Expansion Model-Brazil).
These models differ from each other in terms of their modelling
approach, sectoral scope, geographical coverage, time resolution,
availability of technological options, dispatch details, etc. In general
terms: MESSAGE eBrazil optimizes the expansion of the whole
energy sector, TIMES-TiPS optimizes the expansion of the power
sector and REMIX-CEM-B optimizes the dispatch of a set of power
plants previously calculated by the expansion tools. Table A2 pre-
sents a cross section comparison about the general configuration
and assumptions of each tool. A complete comparison of model
features and a detailed SWOT analysis (strengths, weaknesses,
opportunities and threats) can be found in Ref. [110].
REMix-CEM has the capability to optimize, besides the invest-
ment and dispatch decision, the configuration of candidate CSP
units (size of solar field, TES, back-up boiler system) within the
capacity expansion optimization process. In this work, REMix-CEM
was used in a first step to determine possible CSP plant configu-
rations for solar only operation at sites in the Northeast region of
2
More details about model coupling approach are described by Hidalgo et al.
[46].
Brazil, where lies most of the CSP potential. Three different CSP
configurations with TES and fossil back-up system were calculated
using REMix_CEM. Besides, a hybridized CSP-biomass plant was
also proposed based on the findings of [109]; totalling at the end
four types of CSP plants to be inserted as inputs (investment op-
tions) into the expansion models (see section 4.2). Hence, infor-
mation about total annual and seasonal power generation as well as
hourly dispatch3 for the proposed CSP configurations were for-
warded to the expansion optimization models MESSAGE-Brazil and
TIMES-TiPs-B.
MESSAGE-Brazil is an integrated energy planning tool, that
models not only the power sector but also all the relationships
between the energy conversion chains within the energy sector. A
detailed description of MESSAGE-Brazil is presented by Lucena
et al. [65], Nogueira et al. [77] and Soares et al. [107]. Additionally,
TIMES-TiPS-B was developed to model the long-term expansion of
the Brazilian power sector. Therefore, the capacity expansion was
calculated by MESSAGE-Brazil and TIMES-TIPS-B for ambitious GHG
mitigation scenarios. The least-cost expansion results from the
most diverse and ambitious scenario for CSP (2DS_GEN) were then
fed into REMIX-CEM-B, which now was applied to optimize the
power system operation dispatch for the proposed generator fleets,
following the sequence presented in Fig. 2.
At this stage, an analysis was performed to see whether the
operational requirements could be achieved with the investment
portfolio provided to REMIX-CEM-B. Due to the fact that the
Northeast region of Brazil is the electrical subsystem that has the
largest penetration of VRE generation and where most of the CSP
techno-economic potential is located [14] the optimization of
dispatch was made only for this region. It is important to highlight
3
Dispatch strategy: maximum electricity generation.
 R. Soria et al. / Energy 116 (2016) 265e280
Thermal production
Solar tools Hourly CSP
Hourly DNI
(INSEL, SAM) performance
REMIX-CEM
Hourly ambient
temperature
Fig. 2. Interactions between MESSAGE-BRAZIL, TIMES-TiPS-B and REMIX-CEM-B in the Brazilian application.
that the full cycle shown in Fig. 2 was followed only once for
illustrative purposes. However, the results present important in-
sights about the role of CSP in Brazil and the advantages of having
applied the modelling approach proposed by Fig. 2.
The model network has been used to produce six scenarios for
the Brazilian energy mix up to 2050 under different CO2 price and
technology availability.
4.2. Main data and assumptions
MESSAGE-Brazil is an integrated energy model, which considers
the exogenous useful energy demand vector as input (electricity for
captive uses, heat, driving force, passenger/km and t/km), accord-
ing to [45]. The useful energy demand vector is in line with pop-
ulation growth estimated by the Brazilian Institute of Geography
and Statistics [52] and GDP data from the World Energy Outlook
[54]. For a complete description of the methodology to estimate the
useful energy demand vector used in MESSAGE-Brazil see Nogueira
et al. [77] and Soares et al. [107]. Based on the optimization of the
energy transformation through the whole energy chain, MESSAGE-
Brazil calculates the total electricity consumption. On the other
hand, TIMES-TiPS-B used an exogenous useful energy demand
vector calculated by applying national annual growth rates of
electricity demand, derived from MESSAGE-Brazil (See Table 1),
over the verified 2010 total electricity consumption of 465 TWh
[36], which included industrial self-electricity production. Both
expansion models used a real annual discount rate of 10% p.a.
The daily load profiles, by region, were harmonized across models
using hourly values from 2013 provided by the National Operator of
the Power System [86]. Fig. A1 presents the hourly load by region in
days of maximum national peak load in summer and winter.
The power generation chain within the expansion models are
based on the structure of MESSAGE-Brazil. For further details see
Nogueira et al. [77] and Soares et al. [107]. Within the power sector,
all models included hydropower plants, nuclear plants, wind farms,
solar facilities (PV and CSP) and thermal power plants fueled with
sugarcane bagasse, fuel oil, diesel, coal or natural gas. Carbon capture
(CCS) facilities are only included in MESSAGE-Brazil as add-on op-
tions into existing coal fired thermal power plants (retrofit option)
and as a capture ready option to be installed in new coal fired power
plants (greenfield option). The power plant database was updated,
validated and harmonized between the models (See Table A3).
Table A3 presents the evolution of capital costs of power generation
technologies, including for the CSP reference units, until 2050.
269
Inputs for TIMES and MESSAGE
Technical data Electricity demand Scenarios
Dispatch for a typical day
TIMES-TiPS-B Total cost
O&M cost CO2 emissions
Capacity expansion
CSP configuration
MESSAGE_BR
Results for other energy
Capacity factor sectors
Definition of a
Installed
portfolio of Analysis of results:
capacity by
projects (MW), by capacity (MW), etc.
technology
technology.
Capacity factors of hydro power plants by region are considered
within MESSAGE-Brazil and TIMES-TiPS-B (See Table A4). The
Table A5 shows the scenario of transmission capacity between re-
gions considered in this study. This key information is the driver to
model seasonal fluctuations of imports/exports between regions.
Data for wind speed and ambient temperature for each region
were considered according to the model's needs using data from
Meteonorm [72] and SWERA/NREL [114]. Hourly time-series of DNI
and global horizontal irradiation (GHI), for high solar (PV and CSP)
potential regions (hot-spots), in the Northeastern Brazil, were
generated by the German Aerospace Center (DLR) using the Solar
Energy Mining (SOLEMI) service. SOLEMI is a high-quality irradi-
ance database, mainly based on METEOSAT satellite-data, with a
nominal resolution of 2.5 km in the visible channel and 5 km in the
infrared channel and half-hourly temporal resolution [28,73]. For
derivation of the direct normal irradiance SOLEMI applies the
method of Schillings et al.[100]; which was validated by Schillings
et al. [101] using ground measured data. For the work at hand,
SOLEMI generated DNI and GHI hourly time series based on half-
hourly images from Meteosat-7 satellite, using data from
20024[44]. This deterministic approach was used as a first
approximation that must be improved in further studies.
MESSAGE-Brazil used this information to input an availability
resource profile in the required time resolution for each region,
while TIMES-TiPS and REMIX considered hourly series. REMIX-CEM
used these hourly data in a more detailed temporal and spatial
resolution to model the performance of VRE power plants installed
around “hot spots” in the Northeast region and their interaction
with conventional thermal, hydro and CSP generators [42,44].
Four types of parabolic trough CSP power plants were proposed
(See Table 2). Using INSEL5 and SAM6 tools data for the proposed
4
The 2002 annual DNI value was the closest value to the average annual DNI in
the period 2001e2003. During the elaboration of this work the most updated DNI
data provided by DLR/SOLEMI for Northeast Brazil corresponded to the period
2001e2003.
5
The INSEL tool [59] was used to calculate normalized hourly generation profiles
for the thermal power generation of SM1 (solar multiple equal to one) solar field of
1 MW CSP plant, following the methodology of [75]. These electricity generation
time-series served as input for the capacity expansion optimization model REMix-
CEM-B. These data were used to model CSP-4hTES, CSP-8hTES and CSP-12hTES
units. INSEL was also used to calculate normalized hourly electricity generation
time-series for fixed-mounted utility-scale PV plants at the identified PV hot-spots.
6
The System Advisor Model (SAM), developed by NREL [80] was used to model
CSP-BIO, following the proposal of Soria et al. (2015).
270 R. Soria et al. / Energy 116 (2016) 265e280

Table 1
Annual growth rate of electricity demand.

2010e2015 2015e2020 2020e2025 2025e2030 2030e2035 2035e2040 2040e2045 2045e2050

Annual growth rate of electricity demand (%) 1.6 2.8 2.6 2.5 2.4 2.4 2.2 2.0

Note: Growth rates based on integrated energy modelling until 2050, without considering industrial self-production.

Table 2
Technical characteristics of the modeled CSP power plants.

Technical specification CSP-4hTES CSP-8hxTES CSP-12hTES CSP-BIOb

Gross capacity of turbine (MW) 150 150 150 33

Solar multiple 1.6 1.9 2.2 1.2
Thermal energy storage (full load hours) 4 8 12 0
Back-up system capacity (% of gross cap.) 100 100 100 30c
Back-up fuel Fossil Fossil Fossil Biomass
Max. share annual generation by back-up system (%) 0 0 0 49
Total land area (km2) 4.1 4.9 5.5 0.84a
Turbine gross efficiency @ full load (%) 39.3 39.3 39.3 37.0
Annual electricity production (GWh) 413 491 554 139
Capacity factor (%) 31.5 37.4 42.2 51.4

Notes:
a
Total project area, including the crop, equals 51 km2[109]. TES: Thermal energy storage. CSP-BIO: CSP hybridized with biomass fueled boiler.
b
Defined within SAM model by Soria et al. [109] instead of REMix-CEM-B.
c
Operating in “supplemental operation mode” (firm capacity).

CSP power plants were previously calculated to set a standard restrictions to minimum and maximum annual imports and exports
performance within expansion tools (See Table 2 and Fig. 3). Back- between Northeast and other regions were set based on historical
up fossil boilers were considered for the first three conventional values [88]. Typical capacity credits were considered by TIMES-TiPS-
CSP plants to guarantee the possibility of having a more flexible B (see Table A7) within capacity balance equation.
operation. Although under real situations those power plants can
operate using a fossil back-up system, a restriction of no electricity
4.3. Scenario description
production from fossil back-up was set in the expansion planning
tools to maximize the renewable electricity production. Subse-
A baseline scenario (six-degree scenario - 6DS) plus five capacity
quently, to take advantage of main capabilities of REMIX-CEM for
expansion trajectories were calculated by MESSAGE-Brazil and
modelling CSP, the tool endogenously calculated, for each CSP plant,
TIMES-TiPS-B which include a combination of mitigation scenarios
optimal specific values for the solar multiple, the TES size and the
and CSP constraints. The baseline scenario considers business-as-
use of the back-up system as described in the previous section.
usual assumptions at national level and includes climate and en-
As MESSAGE-Brazil model was assembled in an integrated form
ergy policies enacted prior to 2010. Mitigation scenarios are pro-
to assure inter-sectorial consistency, the opportunity cost for each
gressively stringent. Mitigation policy in climate change scenarios
energy source was calculated across the whole energy conversion
were implemented in the models through CO2 price applied to all
chain. On the other hand, for TIMES-TiPS-B and REMIX-CEM-B the
GHGs, with values related to a 6
C, 4
C and 2
C global warming
fuel costs were important input data (harmonized with fuel cost
scenarios of [55] (namely, 6DS, 4DS and 2DS scenarios e Table 3).
calculated by MESSAGE-Brazil whenever possible). Table A6 pre-
The scenarios are further divided into two different sets: sce-
sents harmonized fuel cost.
narios with CO2 price and scenarios with CO2 price plus a constraint
Minimum installed capacity restrictions were implemented
of minimum CSP installed capacity until 2050 (Table 4) (scenario
based on the existing and contracted power plants until December
for CSP capacity contracted by auctions). Minimum CSP installed
2014. Existing capacities were implemented based on ANEEL [1],
capacity scenarios are feasible to happen in the long term due to the
CCEE [21] and ONS [87]. Contracted capacities were considered ac-
large CSP potential and the national interest to develop this tech-
cording to CCEE [22]; EPE [128] and SRHE [111]. Finally, following the
nology. In fact, CSP has been considered one of the critical tech-
guidelines of the 10-year Expansion Plan (PDE 2022) new power
nologies for sustainable energy development in Brazil [78]. In all
plant and power line projects (See Table A5) were also considered in
scenarios and sensitivities, the optimization tools allocated the CSP
the long term [38]. Specific costs and parameters for operational
constraint's maximum value into the four considered CSP tech-
restrictions by technology (ramp-up, ramp-down, minimum up-
nologies according to a total minimum cost criteria.
time, minimum down-time, minimum and maximum load rate,
A sensitivity analysis was made for the scenarios listed in
etc.) required for the unit commitment optimization model were
Table 5. A constraint for minimum electricity production from CSP
provided by DLR following the standard REMIX-CEM data base
was tested to compare the results of this constraint with those
values. Specific operational restrictions for some power plants in the
achieved by the minimum CSP installed capacity constraint. Finally,
Northeast region were implemented in REMIX-CEM-B when
some sensitivity tests on the technology portfolio were analyzed for
possible, following the Newave7 v.17 data base [83e85]. Operational
each scenario: a) when CCS is not available; b) when low-cost
restrictions to existing and new transmission capacities between
biomass is an option to hybridize CSP plants (CSP-BIO); and 3)
regions were implemented in REMIX-CEM-B according to ONS mid-
when solar water heating (SWH) is considered. A total of seven
term planning [88,89]. Additionally, within REMIX-CEM-B,
sensitivity tests were analyzed (Table 5). The availability of low-cost
local biomass, such as Mimosa tenuiflora or commonly called
jurema-preta, to hybridize CSP plants was presented by Soria et al.
7
Medium term expansion planning model for the Brazilian power system. [109]. In fact, the Northeast semiarid region has large amounts of
 R. Soria et al. / Energy 116 (2016) 265e280 271

Fig. 3. Left: Monthly seasonal CSP generation profile (Total annual electricity production ¼ 100%) Right: Average hourly CSP generation profile (Total daily electricity
production ¼ 100%). (Operating solely on solar power, including TES, except for CSP-BIO).

Table 3 Table 4
CO2 price paths by scenario (US$2013/t CO2). Brief description of the scenarios used in this study.

CO2 price paths Year Scenarios CO2 price path CSP constraint Computed by

2020 2030 2040 2050 Baseline 6DS

4DS 4DS No constraint
6DS 0 0 0 0
2DS 2DS MESSAGE-Brazil
4DS 30 40 50 60
6DS_CAP 6DS Minimum CSP installed and
2DS 50 100 140 170
4DS_CAP 4DS capacity: TIMES-TiPS-B
Source: Based on IEA [55]. 2DS_CAP 2DS 1.5 GW in 2020, 10 GW in 2030,
18 GW in 2040, 30 GW in 2050

jurema-preta due to its rapid natural regeneration and adaptation

within the caatinga ecosystem [8,20]. Large degraded areas in the Table 5
caatinga can be recovered with this native species [6,43,115]. Brief description of the sensitivity analysis.
Total primary energy is calculated only by MESSAGE-Brazil. The Sensitivity tests CO2 price Description of the Computed by
model comparison for the different models/scenarios results is path sensitivity constraints
made only for the power sector. 6DS_GEN 6DS Min. electricity
4DS_GEN 4DS generation from CSP:
2DS_GEN 2DS 15 TWh in 2030, MESSAGE-Brazil
5. Results and discussion 45 TWh in 2040 and
110 TWh in 2050
4DS_w/o CCS_CSP BIO 4DS CCS is not an
It is important to highlight the relativity of the results presented
2DS_w/o CCS_CSP BIO 2DS investment option.
below, which may change due to variations in costs (fuel and in- Hybrid CSP-BIO power MESSAGE-Brazil
vestments) and climate (e.g. climate change scenarios). The most plant added as a
important finding in this study is not the absolute value of the re- technological option
sults, but to understand what can be the role of CSP in Brazil. 2DS_SWH 2DS SWH is a technological
option.
As for primary energy, according to the results of MESSAGE- 2DS_CAP_SWH 2DS Combination of TIMES-TiPS-B
Brazil (Fig. 4), in the baseline scenario 6DS, primary energy con- 2DS_SWH and
sumption increases more than two folds from 2010 to 2050 (2.2 2DS_CAP scenarios
times increase). According to the results of the primary energy
consumption until 2050 obtained by MESSAGE-Brazil, the 6DS expansion option in stringent mitigation scenarios. MESSAGE-Brazil
scenario would be increasingly based on fossil fuels. Although the calculated that this happens by 2040 in the “2DS_w/o CCS_CSP BIO”
share of coal in the primary energy mix would remain worthwhile with high CO2 price (140 US$/t CO2 in 2040), when CCS is not available
in stringent mitigation scenarios (4DS and 2DS), it would lose its as an option and when CSP plants use low-cost biomass for its hy-
importance, being replaced by large amounts of biomass, some bridization. In a similar fashion, TIMES-TiPS-B showed with the
natural gas and marginal contributions of nuclear, hydro and VRE. sensitivity analysis results that CSP is a cost-effective option by 2050
Fig. 5 presents results of the expansion of power generation with medium CO2 price (above 60 US$/t CO2).
until 2050 calculated by both expansion tools. Both expansion Based on the results calculated by MESSAGE-Brazil, the left
models agree that under the applied cost assumptions and ac- panel of Fig. 7 presents the electricity cost and emission factor in
cording to an exclusively least cost expansion criteria, the expan- power sector for different scenarios. The average electricity cost of
sion of power generation in 6DS would be increasingly based on 2DS scenarios in 2050, would be 3% higher than in 2010. The right
coal, especially after 2030 when any additional large hydro power panel of Fig. 7 shows the CSP expansion, by type of technology, until
plant would be installed. In stringent mitigation scenarios (4DS and 2050. In a “2DS_w/o CCS_CSP BIO” scenario the CSP technology
2DS) coal fired plants equipped with CCS would still play an becomes a cost-effective option for Brazil, due to the use of low-
important role in the power sector, but additional hydro and VRE cost biomass for CSP hybridization (CSP-BIO), expanding 17 GW
power plants (especially wind power) would gain force. For the from 2040 until 2050. In 2DS_GEN scenario, following the mini-
applied costs assumptions, TIMES-TiPS-B showed that CSP can mum total cost criteria for the wide energy system, the CSP
become a cost-effective option by 2050 with medium CO2 price expansion is based on two traditional technologies (4 h and 8 h of
(above 60 US$/t CO2). TES). Results show that 5 GW of CSP_4hTES are installed until 2030,
Fig. 6 presents the results for power generation in the sensitivity and after 2030 only CSP_8hTES plants are installed, summing
tests. For the applied costs assumptions, CSP is a cost-effective
272 R. Soria et al. / Energy 116 (2016) 265e280

Fig. 4. Primary energy consumption in analyzed scenarios.

34 GW in 2050. This result is related to two effects: learning rates provide cheaper electricity than CSP_4hTES plants. In the 2DS_CAP
that reduce investment costs in the long term, and greater elec- scenario, which set a minimum of 30 GW of CSP until 2050, a total
tricity production for CSP_8hTES. After 2035, CSP_8hTES plants of 31.5 GW is installed. This result shows that in 2DS_CAP scenario

Fig. 5. Power generation calculated for scenarios 6DS, 4DS and 2DS (a) and for scenarios 6DS_CAP, 4DS_CAP and 2DS_CAP (b).
 R. Soria et al. / Energy 116 (2016) 265e280 273

Fig. 6. Power generation for the sensitivity tests.

Fig. 7. Left: Electricity cost and grid emission factors in 2DS variations. Right: CSP expansion in 2DS variations.

CSP_4hTES was the cost-effective option in 2050. chosen for the optimization of the dispatch of power plants
Based on results using MESSAGE-Brazil for each CO2 price sce- calculated by REMIX-CEM-B. All CSP installed capacity was located
nario, the additional cost in the energy system8 was calculated by in the Northeast region. The main purpose of this stage was to
comparing each scenario's cost to the respective unconstrained
scenarios. For instance, the 6DS_CAP scenario was 0.31% more
expensive than the 6DS scenario (See Fig. 8). In all cases, for similar
total CSP installed capacities (31.5 GW in 2DS_CAP and 34 GW in
2DS_GEN), setting CSP electricity targets was cheaper than setting
minimum CSP installed capacity goals. One reason for that is the
optimized mix of CSP technologies in 2DS_GEN scenario.
Interestingly, findings also show that the cost differentials be-
tween scenarios are not large. Actually, optimization tools always
look for corner solutions, and, hence, undermine the expansion of CSP
plants in Brazil, even though this expansion may occur at low addi-
tional costs while providing benefits to the electric power system.
Another purpose of this exercise was to verify whether elec-
tricity demand in the Northeast could be supplied by the expansion
plan calculated by the expansion planning tools for this region. Due
to the variety of CSP technologies in the results of 2DS_GEN sce-
nario (calculated only by MESSAGE-Brazil), that scenario was

8
The cost of the wide energy system includes not only power sector cost, but also
production of primary energy, refinery, transport sector, etc. For this reason it is not Fig. 8. Energy system additional cost by CO2 price scenario (in comparison with the
worth providing absolute numbers. Instead, relative numbers were calculated. respective unconstrained scenario).
274 R. Soria et al. / Energy 116 (2016) 265e280

Fig. 9. Monthly shares of power generation in NE region- 2DS_GEN scenario. Left: 2030. Right: 2050.

Fig. 10. Typical dispatch in 2030 and 2050, NE region- 2DS_GEN scenario. Left: summer week; Right: winter week.

understand the role of CSP in the hourly dispatch in the Northeast
region, where most of VRE generation would be integrated.
REMIX-CEM-B calculated the monthly shares of power generation
in the Northeast region in 2030 and 2050 in the 2DS_GEN scenario
(Fig. 9), based on an hourly unit commitment optimization.
Fig. 10 exemplifies the optimum dispatch using the expansion
plan calculated by MESSAGE-Brazil for the region for a summer and
winter week in 2030 and 2050, respectively. REMIX-CEM-B
restricted the use of biomass back-up to 20% of annual power
generation, for each hour to take advantage of all CSP strengths: TES
and back-up system together providing firm and flexible generation
capacity. Interestingly, the complementarity between wind and CSP
solar dispatch can be noticed, particularly in a typical winter day,
while in the summer hydro operates mostly as a base load plant
being complemented by coal and natural gas plants, besides power
imports. In 2050, CSP plants almost eliminate the need for imports
and the dispatch of fossil fuel plants.
Scenarios are largely dependent on the assumptions for the
future evolution of main drivers, such as energy demand and fuel
price. Although premises such as technology costs, performance
and availability were harmonized across models, differences in
model scope, time and spatial resolution are the main reasons for
divergences in results from both expansion tools. As the MESSAGE-
Brazil model is an integrated planning tool and TIMES-TiPS-B fo-
cuses only on power sector, some divergences become evident in
their results. Nevertheless, both expansion tools showed similar
trends through 2050. Both models showed that CSP is highly
complementary with wind and PV generation in CAP and GEN
scenarios. This was verified on an hourly resolution in 2030 and
2050 using REMIX-CEM-B.
REMIX-CEM-B showed (Fig. 10) some hours with deficit when
optimizing the dispatch of expansion plans calculated by both
expansion tools. In this sense, further study is necessary to repeat
the loop showed in Fig. 2 several times until converge in a solution
without deficit risk. In this case, the new expansion planning cycle
should take into consideration the outcomes from REMIX-CEM-B.
6. Conclusions
MESSAGE, TIMES and REMIX-CEM are potential tools to improve
the modelling of large penetration of VRE in the Brazilian power
system. These energy planning tools provide outputs to understand
the opportunities that CSP offers to the power system and to the wider
energy system. This novel modelling approach (soft-link between
energy planning tools) not only allowed improving VRE representa-
tion, but it proved to be a good approach to understand the role that
 R. Soria et al. / Energy 116 (2016) 265e280 275

CSP technology can play in the Brazilian energy and power sectors. example, in the Southeast and Mid-West regions of the country,
The methodological approach allowed understanding the where a good CSP potential is present.
different opportunities for CSP in Brazil in the short to medium
term, consolidating this technology as a feasible alternative for
Acknowledgments
GHG mitigation in the country. This work verified that under the
applied cost assumptions CSP is a cost-effective expansion option in
We thank the Brazilian Federal Agency of Coordination for the
stringent mitigation scenarios (4DS and 2DS). MESSAGE-Brazil
Improvement of Higher Education Personnel (CAPES) (grant
showed that if there is a carbon mitigation commitment, CSP is a
number 3995/14-3), the Deutsche Gesellschaft für Internationale
cost effective option by 2040 in the “2DS_w/o CCS_CSP BIO” sce-
Zusammenarbeit (GIZ) and the German Academic Exchange Service
nario, characterized by high CO2 price, when CCS is not available as
(DAAD), for their specific support in the CSP area, financing the
an option and when CSP plants use low-cost biomass for its hy-
project “CSP Tools” under the New Integrated Partnerships pro-
bridization. The share of CSP electricity in the national power ma-
gram (i-NoPa). We also thank to the National Council for Scientific
trix in 2050 in this scenario can reach 6%. Using low-cost biomass
and Technological Development (CNPq), and the Studies and
hybridization is a technical and economic feasible option for the
Projects Financing Agency (FINEP) for the support of our research
country that would decrease the cost of the electricity generated by
activities. Finally, we thank to Anna-Lena Fuchsand Massimo
CSP plants. Marginal CSP_4hTES capacities were also part of the
Moser from DLR-Stuttgart by the support in early stages of this
cost-effective options in 2DS_CAP scenario in 2050. Additionally,
work.
TIMES-TiPS-B showed that if there is a carbon mitigation commit-
ment (medium and high CO2 price), traditional CSP power plants
with TES and fossil back-up can be a cost effective option by 2050 to Appendices
provide dispatchable electricity. In the longer term, TES is a feasible
option, especially when carbon tax policies are implemented. Re-
sults from REMIX-CEM-B showed that CSP can provide firm energy
and dispatchable capacity in the Northeast region of Brazil, opti- Table A1
mally complementing wind and PV generation. In the same way, it Existing capacity of thermal power plants in Northeast in December 2014.
was verified that CSP offered additional flexibility to the Northeast
Energy source Technology MW Share (%)
power system, especially during winter and after 2030. Results
showed several positive synergies between CSP and other tech- Natural gas CCGT 880 31
CCGT-CHP 647 23
nologies in the power system. Finally, in all 2DS scenarios the
GT 124 4
emission factor of the power system remained below 100 gCO2/ Coal ST 720 26
kWh throughout the whole period of analysis. Oil IC 429 15
Further studies using a methodology to calculate hourly DNI in Note: GT: gas turbine; ST: steam turbine; IC: internal combustion engine.
specific localities, like the proposed by Zhang et al. [124]; are Source: ANEEL [1] and ONS [85].
required to evaluate the CSP potential in specific places across the
country for which there is no available9 hourly DNI series, for

Table A2
Cross section comparison between applications of the models to Brazil.

TIMES-TiPS-B MESSAGE-Brazil Remix-Brazil NE

Methodology Linear Programing Linear Programing Mixed Integer Linear Programming

Sectoral scope Focus on power sector with simple Integrated model that represents the entire Focus on power sector
residential sector representation energy system
CSP technology CSP modeled with TES types and dispatch CSP-PT, 12hs HTS; CSP hybridization with CSP module allowing configuration optimization
representation options biomass. Regional specific capacity factors (solar field, storage, back-up boiler)
Technology changes Exogenous learning curves Exogenous learning curves Exogenous learning curves
Storages Intra-day storage, endogenous storage Single option of TES for CSP Multiple options for TES; hydro reservoir;
optimization batteries
Load management and Limited (energy service demand for Limited (energy service demand) Not available (electricity demand)
energy efficiency residential hot water endogenous)
Time resolution 432 time slices; 6 seasons; 3 typical days; 4 seasons; 5 intra-day periods Hourly resolution
24 h
Time horizon 2010e2050; 5 year steps 2010e2050; 5 year steps One year, pre-defined
User-constraint/policy Available, depends on scenario definition Available, depends on scenario definition Not available but implementable
options
Non-energy CO2 only CO2; sugar production None
commodities
Geographical coverage 5 regions (South-Central; Northwest, 4 Regions (South-Southeast-Centre; Northeast; 2 Regions (Northeast; External)
North; Southeast; external) North; External)
Computational Short running time Short running time Long running time
efficiency
Power system Energy and capacity balance, including Energy balance only, based on intra-day energy Electricity and capacity balance, including
operation dispatch on typical day basis production profiles dispatch optimization
Unit commitment No associated parameters available No associated parameters available Available

9
SOLEMI/DLR cannot provide hourly DNI data for this region because it is out of
Meteosat satellite's coverture.
276 R. Soria et al. / Energy 116 (2016) 265e280

Southeast South
60,000 20,000
15,000
40,000

MW
MW
10,000
20,000
5,000
0 0
0 3 6 9 12 15 18 21 24 0 3 6 9 12 15 18 21 24
h h

Northeast North
15,000 8,000

6,000
10,000
MW

MW
4,000
5,000
2,000

0 0
0 3 6 9 12 15 18 21 24 0 3 6 9 12 15 18 21 24
h h

Fig. A1. Hourly load profile by region in summer and winter (days of maximum national peak load) Source: Own elaboration, based on [86].

Table A3
Cost and performance of the electricity generation technologies.

Power plant options Investment cost Variable O&M Fixed O&M cost Conversion Availability
(US$/kW) cost (US$/MWh) (US$/kW/year) efficiency

2010 2050 2010 2050 2010 2050 2010 2050

Coala
Domestic coal-FBC 3690 2500 4.6 3.1 39 26 0.38 0.42 0.85
FBC with CCS 4190 3000 8.1 6.2 91 78 0.22 0.34 0.85
Pulverized imported coal -PC 2000 2000 5.6 5.6 38 38 0.40 0.45 0.85
PC with CCS 2500 2500 9.1 9.1 90 90 0.23 0.36 0.85
IGCC (imported coal) 2400 2400 3.5 3.5 28 28 0.40 0.48 0.85
IGCC with CCS 2600 2600 7.1 7.1 54 54 0.35 0.42 0.85
Co-firing of domestic coal and biomass 3690 2500 4.6 3.1 39 26 0.35 0.40 0.85
Natural gas (NG)b
OCGT 800 600 3.5 3.5 20 15 0.35 0.38 0.90
CCGT 1190 1000 3.5 3.5 13 11 0.50 0.55 0.85
CCGT with CCS 3090 3090 3.5 3.5 23 23 0.43 0.43 0.85
Flexible CCGT 1300 1300 3.5 3.5 13 13 0.55 0.58 0.85
Hydroelectricc
Small hydroelectric (<30 MW) 2936 2936 e e 65 65 e e *
Medium hydroelectric (>30 MW; <300 MW) 2513 2513 e e 58 58 e e *
Large hydroelectric (>300 MW) 2091 2091 e e 52 52 e e *
Nucleard 4000 4000 0.8 0.8 136 136
Biomasse
Bagasse with backpressure turbines (22 bar) 800 800 5.6 5.6 e e 0.25 0.25 0.90
Bagasse with CEST - existing 959 959 4.8 4.8 e e 0.25 0.25 0.90
Bagasse with CEST - new 2712 2392 4.6 4.6 e e 0.30 0.30 0.90
Bagasse with BIG/GT 1009 1009 4.8 4.8 e e 0.40 0.40 0.80
Biomass -steam turbine 3600 2500 6.3 6.3 50 50 0.28 0.28 0.60
Municipal solid waste 7050 6210 e e 211 186 0.28 0.28 0.74
Oilf
Diesel 1000 1000 14.3 14.3 e e 0.35 0.35 0.35
Fuel oil 1070 1070 14.3 14.3 e e 0.30 0.33 0.55
Non-conventional RE
Solar PV-USg 4300 1300 e e 51 15 e e 0.17
Solar PV-DGg 5300 2000 e e 22 8 e e 0.17
Wind onshoref 1810 1547 e e 42 36 e e 0.35
Wind offshoref 5000 3000 e e 60 36 e e 0.40
Wavef 6000 4500 e e 20 20 e e 0.15
Solar CSP-4hTESh 5208 3315 e e 85 54 e e 0.32
Solar CSP-8hTESh 6312 3912 e e 103 64 e e 0.37
 R. Soria et al. / Energy 116 (2016) 265e280 277

Table A3 (continued )

Power plant options Investment cost Variable O&M Fixed O&M cost Conversion Availability
(US$/kW) cost (US$/MWh) (US$/kW/year) efficiency

2010 2050 2010 2050 2010 2050 2010 2050

h
Solar CSP-12hTES 7254 4422 e e 118 72 e e 0.42
Solar CSP-BIOi 5856 3641 5.0 5.0 65 65 0.57 0.57 0.51

Notes: * Availability of hydro power plants will be presented in Table A4. Variable O&M cost does not include fuel cost. PC: pulverized coal. FBC: fluidized-bed coal. IGCC:
integrated gasification combined cycle. CCS: carbon capture and storage. OCGT: open cycle gas turbines. CCGT: combined cycle gas turbines. BIG/GT: biomass integrated
gasification/gas turbines. CEST: condensing-extraction steam turbine. n.a.: not applicable. RE: renewable energy. US: utility scale. DG: distributed generation. CSP: concen-
trated solar power. TES: thermal energy storage. CSP-BIO: CSP power plant hybridized with biomass, without TES.
Sources:
a
Hoffmann et al. [48]; Rochedo and Szklo [98] and Soares et al. [107].
b
Black & Veatch [11]; Deutch et al. [29]; EIA [33,34]; IEA [56]; UNFCCC.Secretariat [118].
c
Lucena [64] and Lucena et al. [65].
d
Cabrera-Palmer and Rothwell [17]; Deutch et al. [29] and NEA/IEA [76].
e
EIA [34]; IRENA [60] e Soares [106].
f
Soares et al. [107].
g
IEA [57].
h
Fichter et al. [42]; IEA [58] Trieb et al. [116]. Data provided by DLR for this study. A back-up fossil boiler, that guarantees flexibility, is considered within the investment
costs. Nevertheless, for this study a restriction of no electricity production from fossil back-up was set.
i
Soria et al. [109].

Table A4
Table A7
Availability factor of hydroelectric plants in Brazil.
Capacity credit for renewable energy used by TIMES-TiPS-B.
Regions
Power plant Capacity credit
Hydroelectric plant South, Southeast, Midwest North, Northeast
Wind 0.18
Small/Medium 0.57 0.60
PV 0.10
Large 0.52 0.54
CSP 0.90
Source: Lucena [64] and Lucena et al. [65]. Hydroelectric 0.60
Biomass 0.90

Source: Soares [106] and own calculations.

Table A5
Scenario of transmission capacity between regions e MW.
References

Year NE - IMP IMP - NE SE/CO - NE NE - SE/CO [1] ANEEL. Banco de Informaço ~ es da Geraç~ ao Ele
trica -BIG. Rio de Janeiro:
Age ^ncia Nacional de Energia Ele
trica; 2014. http://www2.aneel.gov.br/
2010 3341 3894 1000 600
aplicacoes/capacidadebrasil/capacidadebrasil.cfm (accessed 10.09.14).
2015 3341 3894 1000 600 [2] ANEEL. Banco de Informaço ~ es da Geraç~ ao Ele
trica -BIG. Rio de Janeiro:
2020 4026 8310 918 2208 Age ^ncia Nacional de Energia Ele
trica; 2015. Available in, http://www2.aneel.
2025 4026 8310 918 2208 gov.br/aplicacoes/capacidadebrasil/capacidadebrasil.cfm (accessed 10.08.15).
2030 5526 9810 2418 3708 [3] Aora. Aora - green energy - local power. Israel: Aora-solar. 2015. Available in,
2035 5526 9810 2418 3708 http://aora-solar.com/product/ (accessed 10.08.16).
2040 5526 9810 2418 3708 [4] Arce P, Medrano M, Gil A, Oro
E, Cabeza LF. Overview of thermal energy
2045 7526 11,810 4418 5708 storage (TES) potential energy savings and climate change mitigation in
2050 7526 11,810 4418 5708 Spain and Europe. Appl Energy 2011;88(8):2764e74.
[5] Arvizu D, Balaya P, Cabeza L, Hollands T, J€ ager-Waldau A, Kondo M, et al.
Note: Transmission capacities until 2022 are based on the official expansion plan Special report on renewable energy sources and climate change mitigation
[38]. For the long term, based on historical transmission capacity expansions, a -SRREN. Cambridge and New York: IPCC; 2011. Special Report of IPCC.
conservative scenario is adopted in this work. IMP: Imperatriz Electrical Substation. [6] Azeve ^do M, Bakke I, Bakke O, Oliveira A. Crescimento de pla ^ntulas de jurema
preta (Mimosa tenuiflora (Wild) Poiret) em solos de a
reas degradadas da
caatinga. Eng Ambient 2012;9(3):150e60.
[7] Baharoon DA, Rahman HA, Omar WZW, Fadhl SO. Historical development of
concentrating solar power technologies to generate clean electricity effi-
Table A6
ciently e a review. Renew Sustain Energy Rev 2015;41:996e1027.
Fuel price scenario (without tax).
[8] Bakke I, Bakke O, Pereira A, Salcedo I. Regeneraç~ ao natural da Jurema preta
Fuel 2010 (US$/GJ) 2050 (US$/GJ) em
area sob pastejo de bovinos. Rev Caatinga 2006;19(3):228e35.
[9] Behar O, Khellaf A, Mohammedi K. A review of studies on central receiver
a
Natural gas 7.61 7.61 solar thermal power plants. Renew Sustain Energy Rev 2013;23:12e39.
National coala 1.77 1.84 [10] Bijarniya JP, Sudhakar K, Baredar P. Concentrated solar power technology in
Imported coala 3.30 3.39 India: a review. Renew Sustain Energy Rev 2016;63:593e603.
Nuclearb 2.20 2.20 [11] Black, Veatch. Cost and performance data for power generation technologies.
Sugar cane bagassea 1.39 1.39 Prepared for NREL. Black & Veatch; 2012. Available in, http://bv.com/docs/
Biomass (Eucalyptus)c 1.86 1.86 reports-studies/nrel-cost-report.pdf (accessed 10.08.15).
Biomass (Mimosa tenuiflora)d 0.48 0.48 [12] De Boer HS, Grond L, Moll H, Benders R. The application of power-to-gas,
pumped hydro storage and compressed air energy storage in an electricity
Imported liquefied natural gasa 13.31 13.31
system at different wind power penetration levels. Energy 2014;72:360e70.
Fuel oil (international price)e 15.17 28.70 € ttger D, Go€ tz M, Theofilidi M, Bruckner T. Control power provision with
[13] Bo
Diesel (international price)e 20.85 33.73
power-to-heat plants in systems with high shares of renewable energy
Sources. sources e an illustrative analysis for Germany based on the use of electric
a
Soares et al. [107]. boilers in district heating grids. Energy 2015;82:157e67.
b
Cabrera-Palmer and Rothwell [17]; Deutch et al. [29] and NEA/IEA [76]. [14] Burgi A. Avaliaç~ ao do potential te
cnico de geraç~
trica termossolar no
ao ele
c Brasil a partir de modelagem em SIG e simulaç~ ao de plantas virtuais [Master
Hoffmann et al. [48].
d thesis]. Rio de Janeiro: Universidade Federal do Rio de Janeiro; 2013.
Soria et al. [109].
e Available in:, http://www.ppe.ufrj.br/ppe/production/tesis/salvi.pdf
Viebahn et al. [119] and -high fossil price scenario.
278 R. Soria et al. / Energy 116 (2016) 265e280

(accessed 20.08.15). geraça ~o - Empreendimentos heliote
rmicos. 2014.

[15] Burin E, Buranello L, Giudice P, Vogel T, Go €rner K, Bazzo E. Boosting power [40] Feldhoff J, Schmitz K, Eck M, Schnatbaum-Laumann L, Laing D, Ortiz-Vives F,
output of a sugarcane bagasse cogeneration plant using parabolic trough et al. Comparative system analysis of direct steam generation and synthetic
collectors in a feedwater heating scheme. Appl Energy 2015;154:232e41. oil parabolic trough power plants with integrated thermal storage. Sol En-
[16] Burin E, Vogel T, Multhaupt S, Thelen A, Oeljeklaus G, Go € rner K, et al. ergy 2012;86(1):520e30.
Thermodynamic and economic evaluation of a solar aided sugarcane bagasse [41] Fichter T. REMix-capacity expansion model: supporting capacity expansion
cogeneration power plant. Energy 2016. In press. Available in: http://www. and grid integration of renewable energy technologies in the Middle East
sciencedirect.com/science/article/pii/S0360544216308453 (accessed and North Africa - combining capacity expansion and system operation
26.09.16) (in press). optimization. unpublished. Ph.D. thesis. University of Stuttgart; 2015.
[17] Cabrera-Palmer B, Rothwell G. Why is Brazil enriching uranium? Energy [42] Fichter T, Trieb F, Moser M. Optimized Integration of Renewable Energy
Policy 2008;36(7):2570e7. Technologies Into Jordan's Power Plant Portfolio. Heat Transf Eng
[18] C
aceres G, Anrique N, Girard A, Degre ve J, Baeyens J, Zhang HL. Performance 2014;35(3):281e301.
of molten salt solar power towers in Chile. J Renew Sustain Energy [43] Figueiredo J, Araújo J, Pereira O, Bakke I, Bakke O. Revegetation of degraded
2013;5(5):53142. Caatinga sites. J Trop For Sci 2012;24(3):332e43.
[19] Castro G. Avaliaça ~o do valor da energia proveniente de usinas heliote
rmicas [44] Fuchs AL. Integration of CSP and utility-scale PV in Northeast Brazil -
com armazenamento no a ^mbito do sistema interligado nacional [Master resource assessment and electricity system modeling, 2014 [Master thesis].
thesis]. Rio de Janeiro: Universidade Federal do Rio de Janeiro; 2015. Stuttgart: University of Stuttgart; 2014.
Available in:, http://www.ppe.ufrj.br/ppe/production/tesis/castro.pdf [45] Herreras Martínez S, Koberle A, Rochedo P, Schaeffer R, Lucena A, Szklo A,
(accessed 20.08.15). et al. Possible energy futures for Brazil and Latin America in conservative and
[20] Cavalcante R, Oliveira A, Santos A, Oliveira R, Vena ^ncio I, Trugilho P, et al. stringent mitigation pathways up to 2050. Technol Forecast Soc Change
Potential Energe
tico da Madeira de Espe
cies Oriundas de Plano de Manejo 2015;98:186e210.
Florestal no Estado do Rio Grande do Norte. Cie ^ncia Florest 2013;23(2): [46] Hidalgo Gonza
lez I, Ruiz Castello P, Sgobbi A, Nijs W, Quoilin S, Zucker A,
491e502. et al. Addressing flexibility in energy system models. Luxembourg: European
[21] CCEE. Relato
rio Geral da Geraç~ ao e Garantia Física de usinas em Operaç~ ao Commission - Joint Research Centre (JRC); 2015.
Comercial - Agosto 2014. Arquivo “Relato
rio das Usinas em Operaç~ ao Com- [47] Hirsch T, Khenissi A. Systematic Comparison on Power Block Efficiencies for
ercial” de 30/09/2014. CCEE; 2014a. Available in:, http://www.ccee.org.br/ CSP Plants with Direct Steam Generation. Energy Procedia 2014;49:
portal/faces/acesso_rapido_footer/biblioteca_virtual?_adf.ctrl- 1165e76.
state¼19tdwfgyrn_4&contentId¼CCEE_178396&assunto¼Garantia%20F% [48] Hoffmann BS, Szklo A, Schaeffer R. An evaluation of the techno-economic
C3%ADsica&pag¼1&_afrLoop¼2050462667165806. potential of co-firing coal with woody biomass in thermal power plants in
[22] CCEE. Resultado consolidado dos leilo ~es - julho 2014. Relato
rio mensal. Rio the south of Brazil. Biomass Bioenergy 2012;45:295e302.
de Janeiro: Ca ^mara de Comercializaça ~o de Energia Ele
trica; 2014b. [49] Holttinen H. Estimating the impacts of wind power on power system-
[23] CCEE. Resultados consolidados dos leilo ~es. CCEE; 2015. Available in: http:// sdsummary of IEA Wind collaboration. Environ Res Lett 2008;3(2):025001.
www.ccee.org.br/portal/faces/oquefazemos_menu_lateral/leiloes?_adf.ctrl- 
[50] Holttinen H, Orths A, Hulle F, Kiviluoma J, Lange B, OMalley M, et al. IEA
state¼nlzh7g16d_4&_afrLoop¼196478740272230#%40%3F_afrLoop% Wind RP 16. Expert group report on recommended practices - 16. Wind
3D196478740272230%26_adf.ctrl-state%3D125tzsdlvf_4 (accessed integration studies - 1.Edition 2013. Expert group report. Finland: Technical
02.09.15). Research Centre of Finland -VTT; 2013a.
[24] Connolly D, Lund H, Mathiesen BV, Leahy M. A review of computer tools for [51] Holttinen H, Robitaille A, Tande J, Orths A, Estanqueiro A, Pineda I, et al.
analysing the integration of renewable energy into various energy systems. Summary of experiences and studies for Wind Integration e IEA Wind Task
Appl Energy 2010;87(4):1059e82. 25. In: Proceedings of WIW2013, IEA wind task 25; 2013. p. 10.
[25] Corona B, San Miguel G. Environmental analysis of a Concentrated Solar [52] IBGE. Projeça ~o da Populaça ~o do Brasil por sexo e idade: 2000-2060. Rio de
Power (CSP) plant hybridised with different fossil and renewable fuels. Fuel Janeiro: Instituto Brasileiro de Geografia e Estatística; 2010. Available in:
2015;145(0):63e9. http://www.ibge.gov.br/home/estatistica/populacao/projecao_da_
[26] Cot A, Ametller A, Vall-Llovera J, Aguilo
J, Arque
J. Termosolar Borges: a populacao/2013/default.shtm (accessed 10.08.14).
thermosolar hybrid plant with biomass. In: Proceedings Venice 2010. Venice, [54] IEA. World energy Outlook 2013. Paris: International Energy Agency; 2013.
Italy: Environmental Sanitary Engineering Centre - CISA; 2010. Available in: [55] IEA. Energy technology perspectives 2014- harnessing Electricity's potential.
http://tutzingwaste.org/pub/Solar/SolarLiterature/Cot_Thermosolar_Borge_ Paris: IEA; 2014a.
A_thermosolar_hybrid_plant_with_biomass_2010.pdf (accessed 27.01.15). [56] IEA. World energy investment Outlook. Paris: IEA; 2014b.
[27] Deane JP, Chiodi A, Gargiulo M, O
Gallacho
ir BP. Soft-linking of a power [57] IEA. Technology roadmap - solar photovoltaic energy. Paris: IEA; 2014c.
systems model to an energy systems model. In: Energy, 8th World energy [58] IEA. Technology roadmap - solar thermal electricity. Paris: IEA; 2014d.
system conference, wesc 2010, vol. 42, no. 1; 2012. p. 303e12. [59] INSEL. INSEL - integrated simulation environment language. Stuttgart:
[28] DLR. SOLEMI - solar energy mining. Stuttgart: DLR-Institute of Technical INSEL; 2014. Available in: http://www.insel.eu/index.php?id¼138&L¼1
Thermodynamics Section Systems Analysis and Technology Assessment; (accessed 10.08.16).
2010. Available in: http://www.dlr.de/tt/Portaldata/41/Resources/ [60] IRENA. Renewable energy technologies: cost analysis series. Power sector -
dokumente/institut/system/projects/SOLEMI_Handout_2010_09.pdf issue 1/5-Biomass, vol. 1. Bonn: IRENA; 2012.
(accessed 10.08.16). [61] Liu M, Steven-Tay N, Bell S, Belusko M, Jacob R, Will G, et al. Review on
[29] Deutch J, Forsberg C, Kadak A, Kazimi M, Moniz E, Parsons J, et al. Updated of concentrating solar power plants and new developments in high tempera-
the MIT 2003-The future of nuclear power. An interdisciplinary MIT study. ture thermal energy storage technologies. Renew Sustain Energy Rev
Massachusetts: Massachusetts Institute of Technology Energy Initiative; 2016;53:1411e32.
2009. Available in: http://web.mit.edu/nuclearpower/pdf/nuclearpower- [62] Lovegrove K, Wyder J, Agrawal A, Boruah D, McDonald J, Urkalan K.
update2009.pdf (accessed 10.07.14). Concentrating solar power in India. Report commissioned by the Australian
[30] Despre
s J, Hadjsaid N, Criqui P, Noirot I. Modelling the impacts of variable Government and prepared by IT Power. Australia: Department of Climate
renewable sources on the power sector: reconsidering the typology of en- and Energy Efficiency; 2011. Available in:, http://s3.amazonaws.com/
ergy modelling tools. Energy 2015;80:486e95. zanran_storage/www.itpowergroup.com/ContentPages/2520785388.pdf
[31] Drouineau M, Assoumou E, Mazauric V, Maïzi N. Increasing shares of inter- (accessed 10.08.16).
mittent sources in Reunion Island: Impacts on the future reliability of power [63] Lucena A, Clarke L, Schaeffer R, Szklo A, Rochedo P, Nogueira L, et al. Climate
supply. Renew Sustain Energy Rev 2015;46:120e8. policy scenarios in Brazil: A multi-model comparison for energy. Energy
[32] Drouineau M, Maïzi N, Mazauric V. Impacts of intermittent sources on the Econ 2015;56:564e74.
quality of power supply: The key role of reliability indicators. Appl Energy [64] Lucena A. Proposta Metodolo
gica para Avaliaça ~o da Vulnerabilidade a s
2014;116:333e43. Mudanças Clim
aticas Globais no Setor Hidroele
trico [Ph.D. thesis]. Rio de
[33] EIA. Assumptions to the annual energy Outlook 2013 -electricity market Janeiro: Universidade Federal do Rio de Janeiro COPPE/UFRJ; 2010. Available
module Annex -. Washington, D.C.: U.S. Energy Information Administration; in:, http://www.ppe.ufrj.br/ppe/production/tesis/andre_frossard.pdf
2013. (accessed 28.09.15).
[34] EIA. Electricity market module - Annex of the electricity market module of [65] Lucena A, Schaeffer R, Szklo A. Least-cost adaptation options for global
the national energy modeling system - assumptions to AEO2014. Washing- climate change impacts on the Brazilian electric power system. Glob Environ
ton, D.C.: U.S. Energy Information Administration; 2014. Change 2010;20(2):342e50.
[35] EPE. Resultados do 20
Leila ~o de energia nova A-5-Resumo Vendedor. EPE; [66] Ma J, Silva V, Belhomme R, Kirschen DS, Ochoa LF. Evaluating and Planning
2014a. Available in: http://www.epe.gov.br/leiloes/Documents/Leil%C3%B5es Flexibility in Sustainable Power Systems. IEEE Trans Sustain Energy
%202014/Resumo_Vendedor_20len_a5.pdf (accessed 27.01.15). 2013;4(1):200e9.
[36] EPE. Balanço energe
tico nacional 2014-ano base 2013. Rio de Janeiro: EPE; [67] Malagueta D, Szklo A, Borba B, Soria R, Arag~ ao R, Schaeffer R, et al. Assessing
2014b. incentive policies for integrating centralized solar power generation in the
[37] EPE-MME. Plano Nacional de Energia 2030. Rio de Janeiro: MME; 2007. Brazilian electric power system. Energy Policy 2013;59:198e212.
[38] EPE-MME. Plano Decenal de Expans~ ao de Energia 2022. Rio de Janeiro: EPE; [68] Malagueta D, Szklo A, Soria R, Dutra R, Schaeffer R, Borba B. Potential and
2013. impacts of Concentrated Solar Power (CSP) integration in the Brazilian
[39] EPE-MME. Instruço ~es para solicitaça
~o de cadastramento e habilitaça ~o te

cnica electric power system. Renew Energy 2014;68(August 2014):223e35.
com vistas a  participaça ~o nos leilo~es de energia ele
trica - Expans~ ao da [69] Martinsson F, Riekkola A, Lindblom J, Wrake M. Modelling the Swedish
 R. Soria et al. / Energy 116 (2016) 265e280
residential and service sectors in TIMES: a feasibility study. Stockholm: IVL
Swedish Environmental Research Institute Ltd; 2014.
[70] Mata E,
Kalagasidis AS, Johnsson F. A modelling strategy for energy, carbon,
and cost assessments of building stocks. Energy Build 2013;56:100e8.
[71] Mckenna R, Merkel E, Fehrenbach D, Mehne S, Fichtner W. Energy efficiency
in the German residential sector: A bottom-up building-stock-model-based
analysis in the context of energy-political targets. Build Environ 2013;62:
77e88.
[72] Meteonorm. Meteonorm - weather stations world-wide database. Meteot-
est; 2014. Available in: http://www.meteonorm.com/ (accessed 28.09.16).
[73] Meyer R, Hoyer C, Schillings C, Trieb F, Diedrich E, Schroedter M. SOLEMI: a
new satellite-based service for high-resolution and precision solar radiation
data for Europe, Africa and Asia. Stuttgart: DLR; 2003.
[74] Montes MJ, Rovira A, Mun ~ oz M, Martínez-Val JM. Performance analysis of an
Integrated Solar Combined Cycle using Direct Steam Generation in parabolic
trough collectors. Appl Energy 2011;88(9):3228e38.
[75] Moser M. Combined electricity and water production based on solar energy
[Ph.D. thesis]. Stuttgart: Universita €t Stuttgart; 2015. Available in:, http://elib.
uni-stuttgart.de/bitstream/11682/2382/1/Moser_M._2014_Diss_Combined_
Electricity_And_Water_Production_Based_On_Solar_Energy.pdf (accessed
27.06.16).
[76] NEA/IEA. Projected Costs of Generating Electricity. 2010. Paris.
[77] Nogueira L, Lucena AFP, Rathmann R, Rochedo P, Szklo A, Schaeffer R. Will
thermal power plants with CCS play a role in Brazil's future electric power
generation. Int J Greenh Gas Control 2014;24:115e23.
[78] Nogueira L, Rochedo P, Portugal-Pereira J, Hoffmann B, Arag~ ao R, Milani R,
et al. Critical Technologies for Sustainable Energy Development in Brazil:
Technological Foresight Based on Scenario Modelling. J Clean Prod 2016;130:
12e24.
[79] NREL. Potential for Renewable Energy in the San Diego Region. San Diego,
California: National Renewable Energy Laboratory NREL; 2005.
[80] NREL. Solar Advisor Model Reference Manual for CSP Trough Systems. Col-
orado: National Renewable Energy Laboratory (NREL); 2009. Available in:
https://sam.nrel.gov/ (accessed 27.06.16).
[81] NREL. Concentrating Solar Power Projects. NREL; 2016. Available in: http://
www.nrel.gov/csp/solarpaces/ (accessed 27.06.16).
[82] Olsson M, Sjo €stedt G. Systems Approaches and Their Application: Examples
from Sweden. Springer Science & Business Media; 2004. p. 340. ISBN:
1402023693.
[83] ONS. Plano da Operaça ~o Energe
tica 2013/2017 PEN 2013-Revisa ~o 1-Suma
rio
Executivo. Rio de Janeiro: ONS; 2013a.
[84] ONS. Plano da Operaça ~o Energe
tica 2013/2017 PEN 2013 -Volume I - Con-
~ es de atendimento. Rio de Janeiro: ONS; 2013b.
diço
[85] ONS. Plano da Operaça ~o Energe
tica 2013/2017 PEN 2013 -Volume II -
Relato
rio Complementar. Rio de Janeiro: ONS; 2013c.
[86] ONS. Curva de carga hor
aria do sistema ele
trico Brasileiro, por subregia~o,
correspondente a s usinas simuladas e despachadas individualmente pela
ONS no ano 2013. NOS; 2014a.
[87] ONS. Reporte da modalidade de operaça ~o de usinas - Excel file“Modalidade
para o Site - Usinas Individualizadas - 01/10/2014”. ONS; 2014b.
[88] ONS. Interca ^mbio entre regio ~ es - Dados disponíveis on-line no site oficial da
ONS. ONS; 2014c.
[89] ONS. Histo
rico de vazo~ es - Se
rie de vazo ~es naturais me
dias di
arias 1931-
2012. ONS; 2014d.
[90] Parrado C, Girard A, Simon F, Fuentealba E. 2050 LCOE (Levelized Cost of
Energy) projection for a hybrid PV (photovoltaic)-CSP (concentrated solar
power) plant in the Atacama Desert, Chile. Energy 2016a;94:422e30.
[91] Parrado C, Marzo A, Fuentealba E, Ferna
ndez A. 2050 LCOE improvement
using new molten salts for thermal energy storage in CSP plants. Renew
Sustain Energy Rev 2016b;57:505e14.
[92] Peterseim J, Hellwig U, Tadros A, White S. Hybridisation optimization of
concentrating solar thermal and biomass power generation facilities. Sol
Energy 2014;99:203e14.
[93] Peterseim J, White S, Hellwig U, Tadros A. Concentrating Solar Power/Energy
from waste hybrid plants - creating synergies. In: Proceedings SolarPaces,
SolarPaces; 2012. Available in: http://cms.solarpaces2012.org/proceedings/
paper/9045ba83984 (accessed 29.01.15).
[94] Peterseim J, White S, Tadros A, Hellwig U. Concentrated solar power hybrid
plants, which technologies are best suited for hybridisation? Renew Energy
2013;57:520e32.
[95] Prieto C, Cooper P, Ferna
ndez A, Cabeza L. Review of technology: Thermo-
chemical energy storage for concentrated solar power plants. Renew Sustain
Energy Rev 2016;60:909e29.
[96] PSA. Plataforma Solar de Almería - Annual Report 2015. Almería - Espanha:
PSA - CIEMAT; 2015. Available in: http://www.psa.es/es/techrep/2015/
ANNUAL_REPORT_2015.pdf (accessed 10.08.16).
[97] Rather ZH, Liu C, Chen Z, Bak CL, Thogersen P. Dynamic security assessment
of Danish power system based on decision trees: Today and tomorrow.
PowerTech 2013. Grenoble: Anais IEEE; 2013.
[98] Rochedo PRR, Szklo A. Economic analysis under uncertainty of coal fired
capture-ready power plants. Int J Greenh Gas Control 2013;12:44e55.
[99] Schaeffer R, Szklo A, Lucena AFP, Soria R, Chavez M. The Vulnerable Amazon:
The Impact of Climate Change on the Untapped Potential of Hydropower
Systems. IEEE Power Energy Mag 2013;11(3):22e31.
[100] Schillings C, Mannstein H, Meyer R. Operational method for deriving high
279
resolution direct normal irradiance from satellite data. Sol Energy
2004a;76(4):475e84.
[101] Schillings C, Meyer R, Mannstein H. Validation of a method for deriving high
resolution direct normal irradiance from satellite data and application for
the Arabian Peninsula. Sol Energy 2004b;76(4):485e97.
[102] Schmitz H. Who Drives Climate-relevant Policies in the Rising Powers?.
England: IDS; 2016. Available in: http://www.ids.ac.uk/publication/who-
drives-climate-relevant-policies-in-the-rising-powers (accessed 22.06.16).
[103] Silva V. Value of flexibility in systems with large wind penetration [Ph.D.
thesis]. Imperial College London; 2010. Available in:, https://tel.archives-
ouvertes.fr/tel-00724358/document (accessed 22.06.15).
[104] Silva V, Prime G, Hinchliffe T, Lafond D, Rehulka F, Lopez M. Challenges of the
representation of near term electricity system flexibility in energy system
models. SETIS - European Comission; 2014. Available in: https://setis.ec.
europa.eu/system/files/Abstract%20-%2004%20Silva%20%28EDF%29.pdf
(accessed 13.06.15).
[105] Skumanich A. CSP: Developments in heat transfer and storage materials.
Renew Energy Focus 2010;11(5):40e3.
[106] Soares B. Modelagem integrada da introduça ~o de veículos leves conecta
veis a

rede ele
trica no sistema energe
tico Brasileiro [Ph.D. thesis]. Rio de Janeiro:
Universidade Federal do Rio de Janeiro COPPE/UFRJ; 2012. Available in:,
http://www.ppe.ufrj.br/ppe/production/tesis/bruno_borba.pdf (accessed
09.06.15).
[107] Soares B, Szklo A, Schaeffer R. Plug-in hybrid electric vehicles as a way to
maximize the integration of variable renewable energy in power systems:
The case of wind generation in northeastern Brazil. Energy 2012;37(1):
469e81.
[108] Soder L, Hofmann L, Orths A, Holttinen H, Wan YH, Tuohy A. Experience from
wind integration in some high penetration areas. IEEE Trans Energy Convers
2007;22(1):4e12.
[109] Soria R, Portugal-Pereira J, Szklo A, Milani R, Schaeffer R. Hybrid concen-
trated solar power (CSP)ebiomass plants in a semiarid region: A strategy for
CSP deployment in Brazil. Energy Policy 2015a;86:57e72.
[110] Soria R, Szklo A, Schaeffer R, Lucena A, Tomaschek J, Fichter T, et al. CSP
Tools: Comparison and application of planning tools for grid integration of
CSP in Brazil. In: The NoPa case. GIZ; 2015b. ISBN: 978-85-88468-19-1. Pag.
68-94. Brasília: Available in: https://www.daad.de/medien/the_nopa_
case_-_reflections_and_achievements.pdf.
[111] SRHE. Notícia do site do Governo do Estado de Pernambuco: Pernambuco
promove primeiro leil~ ao de energia solar do País. In: Secretaria de Recursos
Hídricos e Energe
ticos do Governo do Estado de Pernambuco; 2013. Avail-
able in: http://www.srhe.pe.gov.br/index.php?option¼com_
content&view¼article&id¼519:pernambuco-promove-primeiro-leilao-de-
energia-solar-do-pais&catid¼40:newsflash&Itemid¼72 (accessed 27.07.14).
[112] Stanojevic V, Silva V, Pudjianto D, Strbac G, Lang P, Macleman D. Application
of storage and Demand Side Management to optimise existing network ca-
pacity. In: 20th international conference and exhibition on electricity dis-
tribution - Part 1, 2009. CIRED 2009; 2009. p. 1e4.
[113] Sullivan P, Krey V, Riahi K. Impacts of considering electric sector variability
and reliability in the MESSAGE model. Energy Strategy Rev Future Energy
Syst Mark Integr. Wind Power 2013;1(3):157e63.
[114] SWERA/NREL. Solar and Wind Energy Resource Assessment - SWERA. United
States: NREL; 2007. Available in: https://maps.nrel.gov/swera/ (accessed
27.07.14).
[115] Tavares S. Recuperaç~ ao de a
reas degradadas da Caatinga com as espe
cies
nativas Jurema preta (Mimosa tenuiflora) com e sem acúleos e favela (Cni-
doscolus quercifolius) com e sem espinhos [Master thesis]. Patos: Uni-
versidade Federal de Campina Grande; 2012. Available in:, http://www.
bibliotecaflorestal.ufv.br/handle/123456789/6639?show¼full (accessed
27.07.14).
[116] Trieb F, Fichter T, Moser M. Concentrating solar power in a sustainable future
electricity mix. Sustain Sci 2014;9(1):47e60.
[118] UNFCCC.SECRETARIAT. Updated compilation of information on mitigation
benefits of actions, initiatives and options to enhance mitigation ambition.
Geneva: United Nations Office at Geneva; 2014. Technical paper. Addendum.
[119] Viebahn P, Lechon Y, Trieb F. The potential role of concentrated solar power
(CSP) in Africa and EuropeeA dynamic assessment of technology develop-
ment, cost development and life cycle inventories until 2050. Energy Policy
2011;39(8):4420e30.
[120] Welsch M, Deane P, Howells M, Gallacho
Rogan F, Bazilian M, et al.

ir BO,
Incorporating flexibility requirements into long-term energy system models
e A case study on high levels of renewable electricity penetration in Ireland.
Appl Energy 2014;135:600e15.
[121] Zhang H, Baeyens J, C
aceres G, Degre ve J, Lv Y. Thermal energy storage:
Recent developments and practical aspects. Prog Energy Combust Sci
2016a;53:1e40.
[122] Zhang H, Benoit H, Gauthier D, Degre ve J, Baeyens J, Lo
pez I, et al. Particle
circulation loops in solar energy capture and storage: Gasesolid flow and
heat transfer considerations. Appl Energy 2016b;161:206e24.
[123] Zhang H, Degre ve J, Baeyens J, Wu S. Powder attrition in gas fluidized beds.
Powder Technol 2016c;287:1e11.
[124] Zhang H, Baeyens J, Degre ve J, Cace
res G. Concentrated solar power plants:
Review and design methodology. Renew Sustain Energy Rev 2013;22:
466e81.
[125] Zhang Y, Smith SJ, Kyle GP, Stackhouse Jr PW. Modeling the potential for
280 R. Soria et al. / Energy 116 (2016) 265e280

thermal concentrating solar power technologies. Energy Policy 2010;38(12):
7884e97.
[126] Zhao H, Wu Q, Hu S, Xu H, Rasmussen CN. Review of energy storage system
for wind power integration support. Applied Energy 2015a;137:545e53.
[127] Zhao P, Wang J, Dai Y. Capacity allocation of a hybrid energy storage system
for power system peak shaving at high wind power penetration level. Renew
Energy 2015b;75:541e9.
[128] EPE. 20
Leil~
ao de energia nova A-5 - Resumo Vendedor - Resultados 2014 e
Resultados. 2014. Available in: http://www.epe.gov.br/leiloes/Documents/
Leil%C3%B5es%202014/Resumo_Vendedor_20len_a5.pdf (accessed 26.09.16).