GIZ BR ESOF Product 6

ENERGY SYSTEMS
OF THE FUTURE:
Integrating variable renewable energy
sources in Brazil's energy matrix
PRODUCT 6:
SUMMARY REPORT
Energy systems of the future: Final Report
Integrating variable renewable energy sources in Brazil's energy matrix
Processing No.: 15.2126.9-001.00
Energy Systems of the Future: Integrating Variable Renewable Energy Sources

in Brazil's Energy Matrix
Document Title: Product 6 – Final Report

Título do Documento: Produto 6 – Relatório Final
Prepared by: Lahmeyer International GmbH

Elaborado por: Tractebel Engineering S.A.
PSR Soluções e Consultoria em Energia Ltda
Authors: Leonardo Rese, Rebecca Sinder (Tractebel)

Autores:
Date: November 2019

Data: Novembro 2019
Coordination: Florian Geyer (GIZ)

Coordenação: Juarez Castrillon Lopes (EPE), in memoriam
Karim Karoui (Tractebel)
Leonardo Rese (Tractebel)
Livio Teixeira Filho (MME)
Marcelo Prais (ONS)
Rafael Kelman (PSR)
Renata Carvalho (EPE)
Roberto Castro (GIZ)
This study was carried out within the scope of the German Cooperation for Sustainable
Development, through the Deutsche Gesellschaft für Internationale Zusammenarbeit
(GIZ), within the Energy Systems of the Future Program. On the Brazilian side, the
Program has as its political coordinating partner the Ministry of Mines and Energy
(MME), also counting on the participation of other relevant institutions in the national
electricity sector, such as the Empresa de Pesquisa Energetica (EPE) and the Operador
Nacional do Sistema Eletrico (ONS), technical implementing partners of this study.
Este estudo foi elaborado no âmbito da Cooperação Alemã para o Desenvolvimento

Sustentável, por intermédio da Deutsche Gesellschaft für Internationale
Zusammenarbeit – GIZ, dentro do Programa Sistemas de Energia do Futuro. Pelo lado
brasileiro, o Programa tem como parceiro coordenador político o Ministério de Minas e
Energia (MME), contando também com a participação de outras relevantes instituições
do setor elétrico nacional, destacando aqui a Empresa de Pesquisa Energética (EPE)
e o Operador do Sistema Elétrico Nacional (ONS), parceiros executores técnicos deste
estudo.
Processing No.: 15.2126.9-001.00
Legal Information
1. All indications, data and results of this study were compiled and carefully reviewed
by the author(s). However, errors regarding the content cannot be avoided.
Consequently, neither GIZ or the author(s) can be held responsible for any direct or
indirect claim, loss or damage resulting from the use or reliance placed on the
information contained in this study, or directly or indirectly resulting from errors,
inaccuracies or omissions of information in this study.
2. Duplication or reproduction of all or parts of the study (including the transfer of data
to media storage systems) and distribution for non-commercial purposes is
permitted, provided GIZ is cited as the source of the information. For other
commercial uses, including duplication, reproduction or distribution of all or parts of
this study, written consent from GIZ is required.
Informações Legais
1. Todas as indicações, dados e resultados deste estudo foram compilados e

cuidadosamente revisados pelo(s) autor(es). No entanto, erros com relação ao
conteúdo não podem ser evitados. Consequentemente, nem a GIZ ou o(s) autor(es)
podem ser responsabilizados por qualquer reivindicação, perda ou prejuízo direto
ou indireto resultante do uso ou confiança depositada sobre as informações
contidas neste estudo, ou direta ou indiretamente resultante dos erros, imprecisões
ou omissões de informações neste estudo.
2. A duplicação ou reprodução de todo ou partes do estudo (incluindo a transferência
de dados para sistemas de armazenamento de mídia) e distribuição para fins não
comerciais é permitida, desde que a GIZ seja citada como fonte da informação.
Para outros usos comerciais, incluindo duplicação, reprodução ou distribuição de
todo ou partes deste estudo, é necessário o consentimento escrito da GIZ.
Processing No.: 15.2126.9-001.00
List of Participants / Lista de Participantes
Steering Committee / Comitê Gestor

Florian Geyer (GIZ)
Juarez Castrillon Lopes (EPE)
Karim Karoui (Tractebel)
Leonardo Rese (Tractebel)
Marcelo Prais (ONS)
Rafael Kelman (PSR)
Renata Carvalho (EPE)
Roberto Castro (GIZ)
Technical Committee / Comitê Técnico – EPE

Bruno Cesar Mota Macada; Bruno Scarpa Alves da Silveira; Caio Monteiro
Leocadio; Carolina Moreira Borges; Cristiano Saboia Ruschel; Daniel José
Tavares de Souza; Fabiano Schmidt; Fabio de Almeida Rocha; Flavio Alberto
Figueredo Rosa; Gabriel Konzen; Glaysson de Mello Muller ; Gustavo
Brandão Haydt de Souza; Gustavo Pires da Ponte; Gustavo Valeriano Neves
Luizon; Igor Chaves; Jean Carlo Morassi; João Henriques Magalhães
Almeida; Jorge Trinkenreich; Jose Filho da Costa Castro; José Marcos
Bressane; Leandro Moda; Leandro Pereira de Andrade; Lucas Simões de
Oliveira; Luiz Felipe Froede Lorentz; Marcelo Willian Henriques Szrajbman;
Marcos Vinícius G. da S. Farinha; Maria Cecilia Pereira de Araújo; Maxwell
Cury Junior; Paulo Fernando de Matos Araujo; Pedro Americo Moretz-Sohn
David; Renata de Azevedo Moreira da Silva; Renato Haddad Simões
Machado; Rodrigo Ribeiro Ferreira; Rodrigo Rodrigues Cabral; Roney
Nakano Vitorino ; Samir de Oliveira Ferreira; Sergio Felipe Falcão Lima;
Simone Quaresma Brandão; Thais Pacheco Teixeira; Thiago de Faria Rocha
Dourado Martins; Tiago Campos Rizzotto; Tiago Veiga Madureira.
Technical Committee / Comitê Técnico – ONS

André Snaider; Angela Barbosa Greenhalgh; Elder Sales de Santanna; Karine
Rejane de Oliveira Franca Louzada; Lillian Monteath; Paulo Eduardo Martins
Quintão; Roseane de Souza Nunes; Tatiane Moraes Pestana Cortes; Vitor
Silva Duarte.
Technical Committee / Comitê Técnico – Consultant / Consultor

Achim Schreider; Alexis Bonneschky; Atom Mirakyan; Enrique Salazar; Felix
Knicker; Francois Botreau; Jörg Großmann; Julia Hoepp; Julio Sanchez; Kai-
Uwe Horn; Matthias Drosch; Oliver Heil; Ralf Bucher; Stefan Drenkard
(Lahmeyer International). Christian Merckx; François Promel; Françoise
Dassy; Guillaume Brunieau; Karim Karoui; Leonardo Rese; Loïc Maudoux;
Lucas Manso da Silva; Pieter Tielens; Rebecca Sinder; Rodolfo Bialecki; Stijn
Cole (Tractebel). Alessandro Soares; João Marcos; Juliana Pontes; Julio
Alberto; Lucas Okamura; Maria Luján Latorre; Mario Veiga Pereira; Martha
Rosa; Maynara Aredes; Rafael Kelman; Silvio Binato; Tainá Martins (PSR).
Processing No.: 15.2126.9-001.00
Presentation
The study “Integrating Variable Renewable Energy Sources in Brazil's Energy Matrix”
was conceived within the scope of German Cooperation for Sustainable Development,
through Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ), within the
“Energy Systems of the Future” Program. On the Brazilian side, the Program has as its
political coordinating partner the Ministry of Mines and Energy (MME), also counting on
the participation of other relevant institutions in the national electricity sector, with an
emphasis to the Empresa de Pesquisa Energetica (EPE) and the Operador Nacional
do Sistema Eletrico (ONS), technical implementing partners of this study.
The study was structured into five main products: Technical Regulation Studies (Grid
Codes); Energy Studies; Power System Studies; Methodology Studies and Technology
Studies, which, all together, have as an end result an analysis of the impacts of the
integration of large amounts of variable renewable energy sources in the National
Interconnected System (SIN). The study takes into account the analysis of energy and
power aspects and considers technological and cost trends, as well as a methodological
proposal and analytical tools for studies of this nature.
In order to carry out the study, an international bidding process was carried out, in which
the Consortium formed by the companies Lahmeyer International, Tractebel and PSR
was awarded to carry out the work.
One aspect that should be highlighted was the active participation of EPE and ONS
experts in the project execution, who, together with the contracted consultant, made
their knowledge available in the preparation of the products, as well as in the
participation of the various training sessions that were carried out during the work.
Objective
The study aims to analyse the impacts of the integration of large amounts of variable
renewable energy sources in the National Interconnected System and has the following
main objectives: i) to review the planning practices for the integration of renewable
energy sources in Brazil; ii) identify any gaps in current planning practices in Brazil with
respect to international practices; iii) to propose improvements in terms of
methodologies and analytical tools for the planning of the Brazilian electrical system;
and iv) carry out a case study using the methodologies and analytical tools proposed in
the study. Additionally, technical training sections carried out by the Consultant to
experts of EPE and ONS teams constituted an important part of this project.
Enjoy the reading.

Processing No.: 15.2126.9-001.00
Apresentação
O estudo sobre a Integração de Fontes Renováveis Variáveis na Matriz Elétrica

Brasileira foi concebido no âmbito da Cooperação Alemã para o Desenvolvimento
Sustentável, por intermédio da Deutsche Gesellschaft für Internationale
Zusammenarbeit – GIZ, dentro do Programa Sistemas de Energia do Futuro. Pelo lado
brasileiro, o Programa tem como parceiro coordenador político o Ministério de Minas e
Energia (MME), contando também com a participação de outras relevantes instituições
do setor elétrico nacional, destacando aqui a Empresa de Pesquisa Energética (EPE)
e o Operador do Sistema Elétrico Nacional (ONS), parceiros executores técnicos deste
estudo.
O estudo foi estruturado em cinco produtos principais: Estudos Regulatórios (Códigos
de Rede); Estudos Energéticos; Estudos Elétricos; Estudos Metodológicos e Estudos
Tecnológicos, que, integrados entre si, tem como resultado final uma análise dos
impactos da integração de grandes quantidades de fontes renováveis de energia no
Sistema Interligado Nacional (SIN), principalmente solar e eólica, avaliando aspectos
energéticos e elétricos e considerando tendências tecnológicas e de custos, bem como
uma proposta metodológica e de ferramentas analíticas para estudos desta natureza.
Para execução do estudo, foi realizada uma licitação internacional onde o Consórcio
formado pelas empresas Lahmeyer International, Tractebel e PSR sagrou-se vencedor
do certame para realização do trabalho.
Um aspecto que deve ser ressaltado foi a participação ativa dos colaboradores da EPE
e ONS, que, juntos com a consultoria contratada, disponibilizaram seu conhecimento
na elaboração dos produtos, bem como na participação das diversas capacitações que
foram realizadas durante o trabalho.
Objetivo
O estudo visa analisar os impactos da integração de grandes quantidades de fontes
renováveis de energia no Sistema Interligado Nacional e possui os seguintes objetivos
principais: i) revisar as práticas de planejamento para a integração de fontes renováveis
de energia no Brasil; ii) identificar eventuais lacunas das práticas de planejamento
atuais no Brasil com respeito às práticas internacionais; iii) propor melhorias em termos
de metodologias e ferramentas analíticas para o planejamento do sistema elétrico
brasileiro; e iv) realizar um estudo de caso aplicando as metodologias e ferramentas
analíticas propostas no estudo. Adicionalmente, ações de capacitação técnica das
equipes da EPE e ONS constituem parte importante deste projeto.
Boa leitura.
Processing No.: 15.2126.9-001.00
Acknowledgements
The Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ), thanks the

Empresa de Pesquisa Energética (EPE and the Operador Nacional do Sistema Elétrico
(ONS), the focal points of the respective institutions, Juarez Castrillon Lopes, Renata
Carvalho and Marcelo Prais, and all members of the working groups that during the
study made their time and knowledge available in order to achieve results of technical
excellence. Last but not least, we would like to thank the contracted Consultant
Consortium, Lahmeyer International, Tractebel and PSR, for the excellent work carried
out and for the way in which the negotiations have always been conducted. Special
thanks are due to the consultants Rafael Kelman (PSR) and Leonardo Rese (Tractebel),
for their availability, attention and technical quality provided to the work.
Many thanks to all.
Special thanks to the friend and colleague Juarez Castrillon Lopes
At the conclusion of this study, the project team could not fail to pay
a special tribute to our friend and colleague Juarez Castrillon Lopes,
who toasted us with his presence, joy, friendship and knowledge
during the execution of the study, but which unfortunately left us on
15 April 2020.
From the first conversations in the conception of this study, Juarez
has always shown himself as an idealizer and encourager of work.
The first ideas for this study were written by him in conversations
over coffee, on napkins, which were later refined until the elaboration
of the terms of reference detailing the content of the study. From the
beginning defending a greater interaction between system operation and planning, he
motivated the joint participation of EPE and ONS in all discussions about the study.
During its execution, Juarez always kept the work in the direction that the final objective
was reached, participating in all study working groups, always actively, enriching the
discussions with his high technical knowledge and creating a fraternal work environment
within the team. His always active participation was one of the factors that led to the
end of delivering a study of technical excellence. Juarez will definitely be missed by his
friends and colleagues who had the opportunity to share the life and work with him.
This acknowledgment does not intend to reflect everything that this electrical engineer,
promoter of wind energy and “Botafoguense”, contributed to the electricity sector in
these almost 45 years of dedication, but rather, to leave a simple and sincere tribute to
those who had the opportunity to enjoy his presence in the execution of this study.
Thank you very much from the entire team.

Processing No.: 15.2126.9-001.00
Agradecimentos
A Deutsche Gesellschaft für Internationale Zusammenarbeit – GIZ, agradece aos

titulares da Empresa de Pesquisa Energética e do Operador do Sistema Elétrico
Nacional, aos pontos focais das respectivas instituições, Juarez Castrillon Lopes,
Renata Carvalho e Marcelo Prais e a todos os membros dos grupos de trabalho que
durante a realização do estudo disponibilizaram seu tempo e conhecimento no sentido
de alcançarmos um estudo de excelência técnica. Por fim, não menos importante,
agradecemos ao Consórcio Consultor contratado, Lahmeyer International, Tractebel e
PSR, pelo excelente trabalho executado e pela forma em que as tratativas sempre
foram conduzidas. Agradecimento especial registramos aos consultores Rafael Kelman
(PSR) e Leonardo Rese (Tractebel), pela disponibilidade, atenção dispensada e
qualidade técnica fornecida ao trabalho.
Um muito obrigado a todos.
Agradecimento especial ao amigo e colega Juarez Castrillon Lopes
Na conclusão deste estudo, a equipe não poderia deixar de registrar

uma homenagem especial ao nosso amigo e colega Juarez
Castrillon Lopes, que nos brindou com sua presença, alegria,
amizade e conhecimento durante a execução do estudo, mas que
infelizmente nos deixou em 15 de abril de 2020.
Desde as primeiras conversas na concepção deste estudo, até
mesmo nas informais, Juarez sempre se mostrou como um
idealizador e incentivador do trabalho. As primeiras ideias foram por
ele escritas em conversas durante cafés, em guardanapos, que
depois foram sendo refinadas até a elaboração de um termo de referência detalhando
o conteúdo do estudo. Defendeu, desde do começo, uma maior interação entre
operação e planejamento, motivando a participação conjunta da EPE e do ONS em
todas as discussões.
Já durante sua execução, Juarez sempre manteve a condução dos trabalhos no
sentido que o objetivo final fosse alcançado, participando de todos os grupos de
trabalho do estudo sempre de forma ativa, enriquecendo as discussões com seu alto
conhecimento técnico e criando um ambiente fraterno de trabalho dentro da equipe.
Sua participação sempre ativa foi um dos fatores que propiciaram ao final entregar um
estudo de excelência técnica. Juarez sem dúvida vai deixar muitas saudades entre
seus amigos e colegas que tiveram a oportunidade de conviver e trabalhar com ele.
Este agradecimento não pretende refletir tudo que este engenheiro eletricista,
incentivador da energia eólica e Botafoguense contribuiu para o setor elétrico nestes
quase 45 anos de dedicação, mas sim, deixar uma simples e sincera homenagem
daqueles que tiveram a oportunidade de desfrutar de sua presença na execução deste
estudo.
Muito obrigado de toda a equipe.

Energy systems of the future:
Integrating variable renewable energy
sources in Brazil's energy matrix
Product 6:
Final Report
Final Report
Gesellschaft für
Internationale Zusammenarbeit
Brazil
RESTRICTED
November 2019
20-26-00173
Final Report
Transaction 81212141
number:
Project Number 15.2126.9-001.00
Client Deutsche Gesellschaft für Internationale Zusammenarbeit
Country Brazil
Project title Energy systems of the future:

Services Consultancy Services
Consultant: Lahmeyer International GmbH

Friedberger Straße 173
61118 Bad Vilbel, Germany
Tractebel Engineering
Avenue Simon Bolivar 34-36,
1000 Brussels, Belgium
PSR Soluções e Consultoria em Energia Ltda

Praia de Botafogo 228 / 1701-A Botafogo
22250-145 - Rio de Janeiro, Brasil
Date November 2019
Revision Date Status Author Verified Approved

R. Kelman (PSR)
01 27. Nov 2019 Final L. Rese (TE) L. Rese (TE) L. Rese (TE)
R. Sinder
November 2019 i
20-26-00173
© Lahmeyer International GmbH, 2020
The information contained in this document is proprietary, protected and solely for the use of the Client
identified on the cover sheet for the purpose for which it has been prepared. Lahmeyer International GmbH
and its consortium partners Tractebel Engineering S.A. and PSR Soluções e Consultoria em Energia Ltda.
undertake no duty, nor accept any responsibility, to any third party who may wish to rely upon this document.
Save to the extent agreed otherwise with the Client all rights are reserved and no section or element of this
document may be removed from this document, reproduced, electronically stored or transmitted in any form
without written permission of Lahmeyer International GmbH and Tractebel Engineering S.A. and PSR
Soluções e Consultoria em Energia Ltda.
November 2019 ii
20-26-00173
Processing No.: 15.2126.9-001.00
TABLE OF CONTENTS
1 INTRODUCTION 11
2 SUMMARY REPORT 12
2.1 Brazilian power system and future prospect 12
2.2 Technical regulations for VRE integration in Brazil 13
2.3 VRE technology landscape 14
2.4 Energy planning studies 19
2.5 Power System Studies 25
2.6 Methodology Studies 36
3 MAIN REPORT 49
3.1 Introduction 49
3.2 Product 1: Technical Regulation Studies 51
3.3 Product 2: Energy Studies 64
3.4 Product 3: Power System Studies 85
3.5 Product 4: Methodology Studies 116
3.6 Product 5: Technology Studies 129
3.7 Capacity Building 147
4 REFERENCES 153
November 2019 i
20-26-00173
Processing No.: 15.2126.9-001.00
LIST OF FIGURES
November 2019 ii
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 2-1: Overview of the energy study methodology 20
Figure 2-2: The evolution of percentage of total installed capacity 24
Figure 2-3: Total hourly production for a week in August 24
Figure 2-4: Coupling of energy and power system simulation models 26
Figure 2-5: Methodology for power system studies (overview) 43
Figure 3-1 Product structure and flow 50
Figure 3-2: General composition of a grid code (adapted from [4]) 51
Figure 3-3: Aspects to be considered when drafting a grid connection code (adapted from [4])
52
Figure 3-4: Typical grid code development process [4] 53
Figure 3-5: Recommended model management process 61
Figure 3-6: Overview of Product 2 methodology 64
Figure 3-7: Regional grid modelling 66
Figure 3-8: Diagram of the process flow of computational model 68
Figure 3-9: Capacity additions per source from 2026 to end of horizon 68
Figure 3-10: Capacity additions per source and region (GW) 69
Figure 3-11: The evolution of the Installed capacity [GW] 69
Figure 3-12: Comparative % of total installed capacity 70
Figure 3-13: The evolution of % of total installed capacity 70
Figure 3-14: International comparison of unit emission per energy consumption 71
Figure 3-15: Main energy transfers 71
Figure 3-16: Sensitivity 1 - results for different projected VRE cost decreases 72
Figure 3-17: Sensitivity 4 - VRE scenarios uncorrelated with inflows 73
Figure 3-18: VRE uncorrelated with inflows and illimited wind developments in the South 74
Figure 3-19: Sensitivity 6: improved forecasting of the VRE production 74
Figure 3-20: Overview of the cut years for the planning exercise 74
Figure 3-21: Timing of investment decisions (GW) 75
November 2019 iii

20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-22: Transmission network study procedure 76
Figure 3-23: SDDP run with full transmission and no flow monitoring identifies reinforcement
candidates 77
Figure 3-24: Solution method for OptNet 77
Figure 3-25: Transmission reinforcements 79
Figure 3-26: Total production for a week in August 80
Figure 3-27: Number of OCGT start-ups per month for different hydrology years 80
Figure 3-28: Hourly marginal costs for August 18th for Albras 230kV busbar (North region) 81
Figure 3-29: Hourly marginal costs for August 18th for Milagres 500kV busbar (Northeast
region) 81
Figure 3-30: Hourly Marginal costs for August 18th for Bandeirantes 88kV busbar (Southeast
region) 81
Figure 3-31: North>Northeast energy exchanges for a normal hydrological year (GW) 82
Figure 3-32: Power flow distribution of DC Link for Xingu to Estreito 82
Figure 3-33: Power flow distribution of DC Link for Xingu to Terminal Rio 82
Figure 3-34 Power system studies 85
Figure 3-35: Process for the development of the static model 86
Figure 3-36: ANATEM to Eurostag data file conversion (Step 1) 87
Figure 3-37: Conversion process for the user-defined models (CDU to macroblock) 88
Figure 3-38: Dynamic model validation procedure 88
Figure 3-39: Coupling of energy and power system simulation models (generation data, PDE
2026 database) 90
Figure 3-40: WPP dynamic model – overview 91
Figure 3-41: WPP dynamic model - WTG model overview 92
Figure 3-42: WPP dynamic model - PPC model overview 92
Figure 3-43: SPP dynamic model – overview 93
Figure 3-44: Detailed SPP dynamic model structure [2] 93
Figure 3-45: Selection of relevant operating conditions – overview 94
November 2019 iv
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-46: System inertia assessment - overview of the methodology 97
Figure 3-47: Total synchronous inertia – histograms 98
Figure 3-48: (Potential) Total system inertia –histogram 99
Figure 3-49: Total synchronous inertia per region - histogram (Scenario 4) 100
Figure 3-50: PFC performance evaluation – Scenario 4 (1500 MW contingency) 101
Figure 3-51: Static studies - overview of the methodology (Product 3) 102
Figure 3-52: Static analysis, Step 1 - synthesis of network expansion results 104
Figure 3-54: Dynamic analysis - overview of the methodology 108
Figure 3-55: Methodology for power system studies (overview) 123
Figure 3-56: LCOE for different markets (left) and expected cost decrease (right) [12] 130
Figure 3-57: Size and capacity evolution of wind turbines over time (onshore and offshore) [13]
130
Figure 3-58: Average rotor diameter and rated capacity increase, 2010-2016 [14] [15] [16]
131
Figure 3-59: Global installed PV capacity forecast 132
Figure 3-60: Global weighted average total installed costs of utility PV and potential of cost
reduction [17] 133
Figure 3-61: Evolution of system efficiency [18] 135
Figure 3-62: Simplified Scheme of Demand’s flexible Adjustment to meet supply [19] 138
Figure 3-63: Optimized performance and new opportunities for grid and ancillary services [20]
139
Figure 3-64: Trends affecting supply and needs for operating reserve [21] 140
Figure 3-65: Forecasting value creation chain 141
Figure 3-66: Services provided by Generation, Storage and Demand Response [23] 146
November 2019 v
20-26-00173
Processing No.: 15.2126.9-001.00
LIST OF TABLES
Table 2-1: Wind Power CAPEX, OPEX and potential of cost reduction 14
Table 2-2: LCOE for different markets (left) and expected cost decrease (right) 15
Table 2-3: Selected operating conditions for the power system studies 28
Table 2-4: Synthesis of system inertia evaluation results (synchronous inertia only) 29
Table 2-5: Synthesis of system inertia evaluation results (total potentially available inertia) 29
Table 2-6: Transmission expansion summary - transmission lines (static studies) 30
Table 2-7: Transmission expansion summary - transformers (static studies) 31
Table 2-8: Transmission expansion summary – shunt compensation (static studies) 31
Table 2-9: Energy studies - summary of methodological recommendations 37
Table 2-10: Summary of the methodological recommendations for power system studies 44
Table 2-11: Power system studies - summary of methodological recommendations 45
Table 3-1: Overview of technical requirements for VRE connection according to VRE share
(adapted from [4]) 53
Table 3-2: Main characteristics of the systems for which the grid codes were evaluated 54
Table 3-3: Benchmark of the Brazilian grid code with respect to the international practice
(synthesis of analyses) 62
Table 3-4: Economic Parameters of the Candidate Projects 65
Table 3-5: Sensitivity 1 - main assumptions 71
Table 3-6: Sensitivity 6 – results per region (GW) 75
Table 3-7: Synthesis of transmission expansion results per region 77
Table 3-8: Number of lines built in each region per voltage level 78
Table 3-9: Investments in intra-regional transmission lines per voltage level 78
Table 3-10: Selected operating conditions for the power system studies 94
Table 3-11: Selected extreme operating conditions – generation mix 95
Table 3-12: Synthesis of system inertia evaluation results (synchronous inertia only) 101
November 2019 vi
20-26-00173
Processing No.: 15.2126.9-001.00
Table 3-13: Synthesis of system inertia evaluation results (total potentially available inertia)
101
Table 3-14: Transmission expansion summary - transmission lines (static studies) 105
Table 3-15: Transmission expansion summary - transformers (static studies) 106
Table 3-16: Transmission expansion summary – shunt compensation (static studies) 106
Table 3-17: Samples of DSA simulation results 111
Table 3-18: Frequency stability simulation results (Case 8) 114
Table 3-19: Energy studies - summary of methodological recommendations 118
Table 3-20: Summary of the methodological recommendations for power system studies 124
Table 3-21: Power system studies - summary of methodological recommendations 125
Table 3-22: Wind Power CAPEX, OPEX and potential of cost reduction 129
Table 3-23: Summary of technology trends 131
Table 3-24: PV CAPEX, OPEX and potential of cost reduction 133
Table 3-25: Summary of technology improvements for different PV technologies 135
Table 3-26: CSP CAPEX, OPEX and potential of cost reduction 136
Table 3-27: List of proposed topics for the training sessions 151
November 2019 vii

20-26-00173
Processing No.: 15.2126.9-001.00
List of Acronyms
Acronym Definition
AC Alternating Current
AGC Automatic Generation Control
ANAFAS Programa de Análise de Faltas Simultâneas (CEPEL)
ANAREDE Programa de Análise de Redes (CEPEL)
ANATEM Programa de Análise de Transitórios Eletromecânicos (CEPEL)
AR Acre and Rondônia Subsystem
BESS Battery Electricity Storage Systems
BFCL Bridge Fault Current Limiter
BIPS Brazilian Interconnected Power System
BoS Balance of System
CAES Compressed Air Energy Storage
CAPEX Capital Expenditure
CCGT Combined Cycle Gas Turbine
CdTe Cadmium Telluride
CDU Controlador Definido pelo Usuário
CEPEL Electrical Energy Research Center (“Centro de Pesquisas de Energia Elétrica”)
CIGRE International Council on Large Electric Systems

CMP Communication Management Plan
CR Central Receiver tower
CRS Central Receiving Systems
CSP Concentrating Solar Power
CW Central-West Subsystem
DC Direct Current
DC-LF DC Load Flow
DC-SA DC Security Assessment
DFIG Doubly Fed Induction Generator
DG Distributed Generation
DLR Dynamic Line Rating
DNI Direct Normal Irradiation
DR Demand Response
DS Digital Substation
DSO Distribution System Operator
EPE Empresa de Pesquisa Energética
ESS Energy Storage System
FACTS Flexible AC Transmission System
November 2019 viii

20-26-00173
Processing No.: 15.2126.9-001.00
Acronym Definition
FCLs Fault Current Limiters
FRT Fault Ride Through
GE General Electric Company
GF “Garantia Física”
GHI Global Horizontal Irradiation
GIS Geographical Information System
GIZ Deutsche Gesellschaft für Internationale Zusammenarbeit
HPP Hydro Power Plant
HTF Heat Transfer Fluid
HV High Voltage
HVAC High-Voltage, Alternating Current
HVDC High-Voltage, Direct Current
ICT Information & Communication Technology
IEC International Electrotechnical Commission
IEEE Institute of Electrical and Electronic Engineers
IPSO Integrated Power System Optimizer
IRENA International Renewable Energy Agency
IT Itaipu Subsystem
LCC Line-Commutated Converter
LCOE levelized costs for electricity
LF Linear Fresnel
LVRT Low Voltage Ride Through
MAD Madeira Subsystem
MAPE Mean Absolute Percentage Error
MBFCL Modified Bridge Type Fault Current Limiter
MME Ministry of Mines and Energy
N North Subsystem
NE Northeast Subsystem
NREL National Renewable Energy Laboratory
O&M Operation and Maintenance
OCGT Open Cycle Generation
ONS Operador Nacional do Sistema Elétrico
OPEX Operational Expenditure
OPF Optimal Power Flow
PDE Plano Decenal de Energia
PF Power Factor
PFC Primary Frequency Control
PHS Pumped Hydro Storage
November 2019 ix
20-26-00173
Processing No.: 15.2126.9-001.00
Acronym Definition
PMU Phasor Measurement Units
PoC Point of Connection
PPA Power Purchase Agreement
PSP Pumped Storage Power Plant
PSS Power System Stabilizer
PT Parabolic Trough
PV Photovoltaic
S South Subsystem
SDBR Series Dynamic Braking Resistor
SDDP Stochastic Dynamic Dual Programming
SE Southeast Subsystem
SIN Sistema Interligado Nacional (National Interconnected System)
SPP Solar Power Plants
STATCOM Static Synchronous Compensator
SVC Static VAR Compensator
T&D Transmission and Distribution
TE Tractebel Engineering
TES Thermal Storage System
ToR Terms of Reference
TPP Thermal Power Plant
TSO Transmission System Operator
UK United Kingdom
VRE Variable Renewable Energy
VSC Voltage Source Converter
WAMPAC Wide-Area Measurement, Protection and Control
WAMS Wide Area Monitoring System
WECC Western Electricity Coordinating Council
WGs Working Groups
WPP Wind Power Plants
WS Workshop
November 2019 x
20-26-00173
Processing No.: 15.2126.9-001.00
1 INTRODUCTION
The project “Energy systems of the future: Integrating variable renewable energy sources in Brazil's
energy matrix” aims at studying the impact of the integration renewable energy sources to the Brazilian
interconnected system (SIN) in both expansion and operation planning standpoints.
The general objective of the project, as specified in the ToR, is “to improve the prerequisites for systematic
integration of renewable energy and energy efficiency into the Brazilian Energy System”.
In this assignment, a pilot study on integrating renewable energies into the Brazilian energy system will be
performed. It will cover both operation and expansion planning aspects. More specifically, the objectives of
the project are:
· Perform an assessment of the current practices on RES integration in Brazil;
· Perform an assessment of the international practices on RES integration;
· Carry out a gap analysis between the international and the National practices in RES integration;
· Carry out an expansion planning exercise composed by energetic and power system analyses con-
sidering power system operation aspects;
· Propose upgrades to the current practices in Brazil based on the results of the gap analysis and the
detailed energy and power system studies.
In order to accomplish the aforementioned objectives, the project is organized in eight (8) products, as fol-
lows:
· Product 0: Work Methodology
· Product 1: Technical Regulation Studies
· Product 2: Energy Studies
· Product 3: Power System Studies
· Product 4: Methodology Studies
· Product 5: Technology Studies
· Product 6: Final Report
· Product 7: Workshops
This report comprises the Product 6 of the project, which is a compilation of the main results, conclusions
and recommendations obtained from products 1 to 5. The report is organized as follows:
· Chapter 1: Introduction;
· Chapter 2: Summary Report;
· Chapter 3: Main Report;
· Chapter 4: References.
November 2019 11
20-26-00173
Processing No.: 15.2126.9-001.00
2 SUMMARY REPORT
2.1 Brazilian power system and future prospect

The Brazilian electrical power system has a total installed capacity of about 160,000 MW. The generation
mix is predominantly hydro, complemented by nuclear and conventional thermal power plants, biomass,
wind and solar PV. Hydroelectric power plants (mix of both run-of-river and with reservoir) correspond to
about 65% of the installed capacity. The total energy storage capacity of the existing hydro power plants is
about 290 GW-avg (about 70% of the storage capacity is located in the South-East/Centre-West regions).
The total wind and solar PV installed capacity are of about 12,800 MW and 1,275 MW, respectively.
The National Interconnected System (SIN) is very large and complex, comprising more than 130,000 km of
HV and UHV transmission lines (above 230 kV). HVDC links are also existing, the most important are the
HVDC lines which connected the hydropower plants Itaipu (14,000 MW), Santo Antonio and Jirau (installed
capacity of 3,568 MW and 3,150 MW respectively) in the Madeira river and two additional HVDC links con-
necting the hydropower plant of Belo Monte (11,233 MW). Brazil is asynchronously interconnected to the
power systems of Argentina, Paraguay and Uruguay. Brazil is synchronously interconnected to the power
system of Venezuela, both operating with the nominal frequency of 60 Hz. New interconnections with the
neighbour countries Bolivia, Guyana and Peru, are under study, in the same way as increasing the power
exchanges through the existing interconnections.
Wind and solar PV generation faced a fast development growth in the last years in Brazil. This fast growth
in VRE integration, associated with the increasing constraints for the construction of new hydro power plants
with reservoir and large transmission infrastructure, are raising concerns from the power system expansion
and operation planning points of view. It is consensus among the Brazilian electricity sector experts that the
national power system is going through a transition phase towards a generation matrix less dependent on
hydro generation.
In addition to the aforementioned, Brazil's participation in COP 21 was highlighted internationally for its am-
bitious commitments to the United Nations Framework Convention on Climate Change (UNFCCC), through
the so-called Intended Nationally Determined Contributions (INDCs). In other to achieve the goal of reducing
Greenhouse Gas (GHG) emissions by 43% below 2005 levels in 2030, the Brazilian energy sector must fulfil
the following targets:
· Reach an estimated 45% share of renewable energy in the energy mix by 2030, including:
o Expand the use of renewable sources - besides hydro power - in the total energy matrix,
from 28% to 33% share by 2030;
o Expand the domestic use of non-fossil energy sources, increasing the share of renewable
energy, as well as hydro power, in the supply of electrical energy to at least 23% by 2030,
considering a greater share of wind, biomass and solar sources.
· Reach 10% efficiency gains in the electricity sector by 2030.
· Increase the share of sustainable bioenergy in the Brazilian energy matrix to about 18% by 2030,
expanding consumption of biofuels, increasing the supply of ethanol, also through the increase of
the share of advanced biofuels (second generation), and increasing the share of biodiesel in diesel
fuel mixture.
The facts mentioned above pose several challenges to the planning and operation of the Brazilian intercon-
nected power system in this future configuration. It will require an evolution of the current expansion and
operation planning practices.
November 2019 12
20-26-00173
Processing No.: 15.2126.9-001.00
2.2 Technical regulations for VRE integration in Brazil

The function of a grid connection code applied to VRE is to provide a set of technical requirements for the
connection of wind and solar PV power plants to a given power system. This code is also in charge of
establishing a fair treatment for all VRE power plants to enable the technical requirements for the power
plant to access the grid. This helps to ensure the fair treatment of generator owners and operators in terms
of grid connection, while maintaining system stability and reliability. By providing appropriate technical and
legal rules for VRE generators, VRE connection codes are instrumental in supporting the effectiveness of
energy policies for VRE integration, while maintaining the reliability of the system to a technical-economic
optimum.
The successful integration of VRE sources in power systems depends on the integration and balance of
technology, operation and regulation. Therefore, policy-makers and regulators must support the develop-
ment and implementation of VRE grid connection codes in association with the national renewable energy
goals by:
· Ensuring that grid connection codes include appropriate requirements for VRE;
· Consulting with all relevant stakeholders as grid connection requirements have an impact on most
stakeholders involved in the power system;
· Anticipating technical requirements based on future VRE targets. Grid codes should not only con-
sider the architecture of today’s power system, but already anticipate the future system requirements
in line with national VRE goals.
· Setting a predictable grid code revision process in order to avoid regulatory instability;
· Learning from other countries’ experience and best practices, as well as lessons from frontrunner
countries in integrating a high share of VRE in their power system while considering the peculiarities
of the Brazilian power system and requirements;
· Joining regional initiatives to harmonize requirements and share resources. Brazil large system and
therefore VRE market justifies engaging in international standardization processes by sharing expe-
riences, deploying regional infrastructure for verification and certification processes, and harmoniz-
ing requirements resulting in cost reduction due to market scale for technology suppliers.
The analysis of the current version of the Brazilian grid code for the connection of VRE generators to the
grid shows that it covers all relevant dimensions of VRE integration and is in line with international grid codes
practice. The following tracks for improvement were identified:
· Define process and rules to enable reactive power compensation during hours with 0 MW power
production;
· Define mechanisms to customize fault ride through requirements by regions;
· Define minimum performance requirements for power plant control function in terms of voltage, ac-
tive and reactive power, power factor, etc.;
· Formalize the processes and rules for carrying out commissioning & compliance tests;
· Formalize the process and rules for model development and validation;
· Define equipment and related protection requirements during unbalanced faults;
· Define specific connection and performance requirements for battery electricity storage systems.
November 2019 13
20-26-00173
Processing No.: 15.2126.9-001.00
2.3 VRE technology landscape

2.3.1 Wind power
Brazil is one of the largest wind energy markets in the world and the largest in the Central and South America
region. In the last 7 years, the countrywide installed capacity has been quadrupled and expects a further
increase up to 19 GW up to the end of 2021, which has doubled since the end of 2016. In this time period,
it will be most likely that the turbine manufacturers which are present in Brazil will introduce their new 4 MW
platforms for this market, which will consequently result in a higher rated capacity per turbine compared with
the 2 MW/turbine in 2016.
Another expectation for future years will be the increase of the capacity per wind farm, taking into account
that the benchmark capacity of the power grid utilization fee increased in 2018. Thus, it will be most likely
that the size of contracted wind farms will follow the worldwide trend and will increase to about 100 MW in
the coming years. Assuming that the average capacity of the turbines installed in the future will be in the
range of 3-4 MW, an outage of only 3-4 turbines would already exceed 10 MW of the wind farm capacity
(assuming a 100 MW wind farm). The requirement of sound implementation of condition monitoring and
preventive maintenance will therefore become more important in the near future.
Due to increasing competition and decreasing prices for the investment, a significant evolution of the onshore
wind turbine technology is not expected. Considering the market share and strategy of the main players of
the Brazilian market (GE, Siemens-Gamesa RE), the Doubly Fed Induction Generator (DFIG) should remain
the dominant technology in the next years as this is the most economical solution and can comply with most
of worldwide grid codes. If the grid code requirements tighten in the future, the installation of additional
technical measures within the wind farm, such as battery storage or the STATCOM system, might become
required on the electrical equipment of the plant. These would lead to approximately 2.5-3.5% additional
investment of the wind farm.
Based on the findings of the Brazilian as well of the international wind energy market, the following cost
figures should be retained for the expansion planning in Brazil.
Table 2-1: Wind Power CAPEX, OPEX and potential of cost reduction
Parameter Value
CAPEX (MUSD/MW) 1.35
OPEX (USD/MW/a) 50,000
Potential of CAPEX reduction until 2035 (%) 30
As consequence of the decreasing investment and operation costs and due to increasing efficiency, the
(levelized) costs for electricity generation (LCOE) by wind energy decreased significantly in recent years.
Depending on project size, location, technology and market, the costs range from about 40 USD per MWh
in the US up to 150 USD per MWh in Thailand and Japan. For comparison, the costs of the Brazilian market
are in the range of 45 USD per MWh to 90 USD per MWh as per Bloomberg (see figure below). When
looking at the average LCOEs (right graph), the lowest costs can be expected in Brazil due to the high wind
conditions and the low price for local content.
November 2019 14
20-26-00173
Processing No.: 15.2126.9-001.00
Table 2-2: LCOE for different markets (left) and expected cost decrease (right)1
By end of 2021, about 3 GW of the installed turbine capacity in Brazil will have an operation time of 10 years
or older. By 2026, about 50% of all turbines will be in operation for more than 10 years. Although the design
life time of a wind turbine is typically 20 years, experiences with repowering in European countries have
shown that in some cases wind farms are repowered before the end of the design lifetime, e.g. after 10-14
years due to economic reasons. Thus, a repowering process can be expected for Brazil in about 5-7 years.
In general, a high availability is essential to secure the predicted energy yield. Unfortunately, no statistics
are available for the Brazilian market, but due to the young age of the Brazilian wind turbine fleet it is ex-
pected that the availability is in a similar range as the typical international benchmark. The main reason for
the absence of reliable statistics is that wind farm owners typically are not required to share information on
downtimes2, for the reason of downtime with the utility and regulator.
2.3.2 Solar power

For the review of solar technologies, existing data and projections of the solar industry from different sources,
such as International Renewable Energy Agency (IRENA), US National Renewable Energy Laboratory
(NREL), the Ministry of Mines, and Energy of Brazil, has been analysed in order to define the technologies,
configurations and cost developments that could likely be applied to Brazil within the next years.
In general, electricity generation at utility scale level based on solar technologies is positioned in the global
markets and is currently a technically and economically feasible alternative competing with conventional
power generation.
The information analysed led to the following conclusions for the expansion planning:
· Despite the important advantages of Concentrating Solar Power (CSP) equipped with Thermo-Elec-
trical Energy Storage (TES) in terms of dispatchability the drastic price reduction experienced by Pho-
tovoltaic Power (PV) in recent years alongside with its simplicity have made this technology the obvi-
ous option amongst the solar technologies. Therefore, it is expected that the market in Brazil will still
be implementing PV instead of CSP for several years, and the time when CSP could rival PV in Brazil
is rather uncertain.
· For the case of competitive PV plant auctions at utility scale level, it is expected that current and future
projects will be based on crystalline and thin film (CdTe) technologies as these are the technologies
that can bring lower prices while guaranteeing bankability.
· The use of tracking systems will depend on the local conditions of a specific project. For the case of
Brazil, there is a clear tendency towards the use of single-axis trackers to improve the plant’s capacity
1
Source: IHS 2015
2
As per information from ONS, only in cases more than ten percent of the installed capacity of a wind farm are out of
order, the wind farm owner is obligated to inform the utility.
November 2019 15
20-26-00173
Processing No.: 15.2126.9-001.00
factors. Fixed mounting systems could be competitive in areas with high Global Horizontal Irradiation
(GHI).
· Implementation of batteries for PV for the purpose of generation shifting or to firm up the PV capacity
(for production at peak load) will not be economically viable within the planning period up to 2035.
However, the development of battery prices is to be monitored closely and such systems could be
included in the expansion planning if a substantial price reduction is achieved in the next years.
· The use of bi-facial PV modules is in its first steps. Prices and actual performance are not yet known
by the industry, and there are not enough suppliers to cause sufficient competition at international
competitive auctions. However, this development shall be monitored and included in the expansion
plan if it becomes a market standard.
· CAPEX for PV is expected to keep dropping until 2025 by 20%. It is assumed that the reduction will
likely be maintained linearly until 2035. This assumption needs to be monitored and reviewed when
updating the expansion plan as it is currently difficult to predict the market.
· In terms of performance of PV modules, it is expected that by 2027 mono and multi-crystalline, as
well as CdTe thin film modules will achieve the record efficiencies in the laboratories and that these
efficiencies will become the market standard.
Cost reductions for PV are expected to be substantial in the coming years and will be mainly driven by:
· Economies of scale (larger plants);
· Technology improvements such as efficiency improvements with the corresponding reduction in land,
mounting and maintenance costs;
· Reduction of Balance of System (BoS) costs.
Reduction of BoS costs are to play a major role in the next years according to IRENA. Until 2025 an average
reduction in the total CAPEX of 57% is forecasted with a major reduction expected to come from the BoS
and thus it could be applied for all technologies and tracking systems. However, this reduction was based in
2015 and the market has already experienced a strong price reduction in 2016 and 2017. A conservative
value for reduction until 2025 will thus be around 20%.
Beyond 2025, it is difficult to foresee how costs for utility scale PV will behave. For the purpose of this
document a similar linear reduction should be assumed, resulting in a 40% cost reduction until 2035. In
terms of considering different rates of cost decrease, different scenarios could be explored within the expan-
sion planning with a total cost reduction of e.g. 30% and e.g. 50% until 2035.
In terms of expected technological breakthroughs, main performance improvements will come from improve-
ments on the different generations of PV systems. One important improvement already occurring is the
volume increase of second-generation PV systems, i.e. thin film, which are reaching commercial levels. It
could be expected that third generation PV systems (Concentrated PV, CPV) could also reach that level,
however for the time being they are still at the demonstration level. The one logical assumption is that in the
future those record efficiencies, or premium modules will be achieved by commercial modules in addition to
further increases in efficiencies due to other technology improvements.
Another important technology breakthrough is occurring with the innovation of bifacial PV modules. Basically,
bifacial PV modules can produce electrons from both sides of the module. Other advantages are the ex-
tended durability of the module and the reduction of specific costs of the balance of system as less area is
required to produce the same amount of electricity. Under certain conditions bifacial modules can produce
up to 30% more electricity than normal modules.
November 2019 16
20-26-00173
Processing No.: 15.2126.9-001.00
The availability of a PV plant depends directly on the question - how reliable the plant is and how well it is
maintained. The minimum availability accepted commercially for PV systems is of 99% throughout its plant
lifetime.
2.3.3 Technology resources for mitigating RES variability impacts

Supply-level mitigation:
Large amount of Renewable Energy Sources (RES) generation in the dispatch triggers a requirement for
operating reserve requirements among the conventional thermal and hydro power plants, which can react
to changes in the RES power output.
Brazil has the advantage that it is characterized by a large number of hydro power plants with reservoirs,
which are very flexible in operation. The water storage can act as a “large battery” towards changes in the
output of RES. Since numerous HPPs have short penstocks, they are also able to provide very fast acting
frequency response (primary reserve). For this reason, exploiting the operational flexibility potential of the
already existing HPPs should have first priority for the integration of large-scale RES.
Grid-friendly RES with advanced inverters:
Renewable energies and, in particular, inverter-based generation (solar and wind) are expected to continue
growing at even higher rates. Thus, challenges such as reduced system inertia, unused renewable energy
capacity due to poor forecasting and increased congestion due to bi-directional energy flow and high energy
transfer will become increasingly relevant.
Inverters are being designed with those challenges in mind:
· Advanced Inverters support fast frequency control, by simulating an artificial inertia;
· Advanced communication networks to facilitate the mitigation of the stochastic nature of RES, and
to allow real time data exchange between Transmission System Operators (TSOs) and Distribution
System Operators (DSOs) for a smarter operation of the grid;
· Build on better semiconductor technologies such as Silicon Carbide (SiC) and Gallium Nitride (GaN)
due to the decrease of manufacturing costs.
Due to high competition in the market, especially by the Chinese, prices are expected to continue to drop in
the upcoming years and inverters are expected to offer more and more functionalities to mitigate inverter-
based generation related issues.
VRE production forecasting:
The need for wind and solar power forecasting in power system operations will grow. It will become important
to efficiently utilize the information provided by advanced forecasting models for the following functions:
· Unit commitment: Forecasting is used in the unit commitment process to help avoid costs and inef-
ficiencies due to unnecessary starts and stops of thermal generators. Such models can play a sig-
nificant role in reducing costs while maintaining system security under increased uncertainty and
variability. In addition, it is also important to consider the close interaction between operating reserve
requirements and unit commitment policy.
· Dispatch: Short-term solar and wind power forecasts can support the control of the generation from
solar and wind power systems and enforce curtailment of power generation in situations where this
is needed, either from an economic or reliability perspective.
· Operating reserves: Appropriate forecasting can address the optimal determination of reserve re-
quirements under high penetration of wind and solar power.
November 2019 17
20-26-00173
Processing No.: 15.2126.9-001.00
The forecasting accuracy of wind power depends on many factors like topography, climate, available data,
forecasting methods or horizon. The lowest forecasting error for wind power, as mentioned by Giebel 2017,
for 24h ahead forecast was 10-13% Normalized Root Mean Square Error (NRMSE), which could be used
as an orientation for accuracy improvement. Based on the experience of the Consultant it is expected that
the wind power forecast accuracy will be improved by some 1 to 2% points within the next two decades.
Like in the case of the wind power forecasting, PV power forecasting depends also on many factors like site
specific factor: e.g. climate, methods used or forecasting time horizon. Currently, there is no PV power fore-
casting system in place in Brazil. However, the best forecasting accuracy reached yet (see previsions sec-
tions about best praxis of PV power forecasting) can be an orientation, namely for clear sky days about
3.38%. A rough interval for MAPE value, for orientation can be from 4 to 11%, depending on the day type
‘cloudless sky’ or ‘clouded day’.
Grid-level mitigation
An extension of transmission lines and High Voltage Direct Current (HVDC) linking different areas with dif-
ferent wind and solar power availability and different load centres can increase the utilization of RES energy
and the network stability. These technologies and related functions are already widely used in Brazil and
therefore not further developed in the frame of the present technology landscape.
It is also worthwhile to mention smart grid technologies sometimes already used in Brazil such as:
· The benefits of Dynamic Line Rating (DLR) include but are not limited to improved system reliability
and safety, reduced and or deferred capital expenditure and increased efficiency of generation re-
sources. It can increase the transmission capacity and mitigate the variability.
· Digital substation technology represents a step-change innovation and is close to the market break-
through. It represents a smart grid enabling technology necessary for a power infrastructure based
on renewable energy infeed. Digital substations are critical for the efficient integration or renewable
generation by centralized monitoring of Transmission & Distribution (T&D) networks and automated
grid management.
· Flexible AC Transmission System (FACTS) devices play a key role for integrating RES in a power
system. They can provide fast continuous control of power flow in the transmission system by con-
trolling voltages at critical buses, by changing the impedance of transmission lines, or by controlling
the phase angles between the ends of transmission lines.
· The placement of Fault Current Limiter (FCLs) is important in limiting fault current and augmenting
stability of the power system. Given the maturity of FCLs, it is advisable to consider them for utiliza-
tion in grid regions with large amounts of RES generation.
Energy storage systems (ESS)

Electricity Storage systems (ESS) technologies are expected to play important and different roles towards
the energy system transition in the future. The analysis shows that the cost of Battery ESS technologies will
decrease significantly, while their performance will increase in the coming years. No single storage system
can meet all the criteria to become the ideal energy storage system. So, the relevance of each ESS tech-
nology depends primarily on their expected response time and the targeted type of application:
· For shorter response times (up to 1 s) and discharge duration (minutes to a few hours), for power
quality improvement, for smoothing intermittency of RES or voltage stabilization;
· For medium response time (minute) and discharge duration (hours-day), for energy management,
for time/peak load shifting;
November 2019 18
20-26-00173
Processing No.: 15.2126.9-001.00
· For a response time (minutes) and discharge duration (hours-days), for emergency back-up or sea-
sonal energy storage.
Due to their falling costs, quick charge time, high power density, high energy efficiency, high reliability, high
cycle times, low carbon footprint and low disposal issues, BESS systems appear therefore promising. As
per IRENA 2017 forecasts, the estimate for the energy installation costs are expected to decrease from
between 150 and 1,050 USD/kWh in 2016 to between 75 and 480 USD/kWh by 2030. This would represent
a decline of between 50% and 66% depending on the technology.
In the frame of the present project, the following results have been identified:
· BESS are technologically as well as commercially ready to provide Primary Frequency Response
(PFR).
· BESS are interesting for the provisioning of RES output smoothening (ramp-rate control).
· BESS will not be competitive for a long time for firming up RES power (generation shifting) even
though their costs are expected to decrease.
· However, Brazil’s large amount of HPPs with reservoirs are generally very suitable to firm up RES
power output. Aside some aspects to be analysed such as correlation between wind and PV power
in the different regions and hydrological restrictions in the operation of the reservoirs, PSPs generally
allow for a flexible operation with some four to ten hours of autonomy. Due to their good controlla-
bility, they can operate as peak shaver and provide secondary control power. However, given the
large amounts of HPPs with reservoirs in Brazil which cater for those services, PSPs will not be
needed for such operation.
Demand-side technologies
Demand Response (DR) comprises a broad range of automated and manual initiatives to modify electricity
consumption in reaction to demand-supply imbalances or exceptionally high power prices. Utilities leverage
the power conserved through demand curtailment or displacement to meet a surging electricity demand from
a particular part of the grid.
DR activities can be categorized on basis of their timing and impact on the customer. When properly imple-
mented and integrated in the system, Demand Response (DR) incorporated in a system with a high share
of variable generation sources has the potential to provide services that are equivalent (or even superior) to
services provided by traditional resources.
A Demand Response (DR) program was launched in the North and North-East regions of Brazil in January
2018. According to information from the Brazilian side, until summer 2018, there was only little interest of
potential participants in the DR program. For the large number of potential DR participants, there was no or
little knowledge about what kind of flexibilities their demand may support, how this could be implemented,
or what the financial costs and benefits of this were.
To this end, an awareness raising campaign could be started, addressing the stated concerns. This measure
may help increase the reach of the DR program.
2.4 Energy planning studies

2.4.1 Background
The renewable energy sources variable nature brings along some complicating issues when integrated into
the energy system. In this context, the main objective of the energy planning studies was to present expan-
sion planning methodologies and best practices related to the integration of VRE sources in the Brazilian
November 2019 19
20-26-00173
Processing No.: 15.2126.9-001.00
energy matrix. It also identified improvements on the current energy planning practices in Brazil to prepare
for a high increase in the insertion of variable renewable sources in the coming years.
2.4.2 Methodology
The methodology and the computational tools proposed in the energy study are depicted in Figure 2-1. As
presented, the methodology is divided into the following three main components:
· The modelling of VRE resources;
· The optimal capacity expansion and operation planning;
· The transmission planning.
Figure 2-1: Overview of the energy study methodology

These three main components are described as follows:
VRE Modelling
The first step was to identify wind and solar generation candidate projects. A set of candidate locations, in
geographical coordinates, was extracted from projects that were not selected in the past auctions. The ra-
tionale is that these projects are likely to be competitive in terms of their levelized cost of energy (LCOE)
and connection costs to the grid.
Global databases, such as MERRA-2 data, were used to obtain about 30 years of historic hourly wind speed
and solar radiation on the set of candidate locations. This data was refined and calibrated based on the
actual energy production records (2-3 years) of existing power plants located in the same region as the
candidate projects. The calibrated 30-year records of wind speed and solar radiation were used to estimate
the energy production using the engineering characteristics assumed for new candidate projects.
Stochastic modelling of renewables and inflows
The expected operational cost was calculated over a set of inflow/renewable scenarios. Due to the spatial
correlation of wind and solar production in different regions, as well as the spatial correlation between inflows
and wind/solar in some regions, scenarios for wind, solar (with hourly resolution) and inflow (on a monthly
basis) were jointly produced by PSR’s TSL (“Time Series Lab”), for existing plants and candidate projects.
Co-optimize new electricity and gas investments
November 2019 20
20-26-00173
Processing No.: 15.2126.9-001.00
The power system planning was performed by using PSR’s OptGen model, which determined the optimal
set of generation and transmission reinforcements. The main goal was to minimize the present value of the
sum of investment costs and the expected value of the thermal plant operation costs plus penalties for load
supply shortages.
Probabilistic reserve constraints
Two reserve components were considered in the capacity planning model. The first component is defined
ex-ante as a percentage of the hourly demand to offset forecasting errors and natural fluctuations throughout
the day. The objective is that flexible resources, such as hydro plants, fast response units and batteries will
respond to the short demand variability. The second component is a Dynamic Probabilistic Reserve (DPR),
which is related to the variability of VRE and it is meant to secure the system operation against deviations
between forecasted and verified VRE productions.
Transmission grid modelling in the generation planning
One important feature of renewables is the diversity of seasonal and geographical patterns. This diversity
allows the planner to take advantage of “portfolio effects” to reduce the variability of renewable production
and, thus, reduce the need for fast generation reserve. On the other hand, the incorporation of geographically
diverse production requires investments on transmission capacity. Although the capacity planning method-
ology can, in principle, handle the detailed joint optimization of both generation and transmission systems,
the resulting computation effort can be very high. Therefore, the integrated planning of regional interchange
capacities and generation provides a more adequate trade-off between accuracy and computational effort.
For the portfolio selection purposes, the grid was modelled considering seven regions. The objective was to
model trade-offs related to where the projects should be built.
Optimal stochastic operation
The optimal capacity expansion is a result from the iterative solution of two modules:
· An investment module (OptGen) that produces a trial plan of generation and regional interconnection
reinforcements;
· An operation module (SDDP) which calculates the expected operation cost associated with the trial
plan produced by the investment module.
The SDDP calculates the optimal operation policy of a system composed of storage (usually hydro plants),
renewables, thermal generation and transmission network. Due to uncertainties on future inflows and re-
newable production, the optimal operation must be modelled as a multistage stochastic optimization prob-
lem. The SDDP operation model uses a technique called “stochastic dual dynamic programming” which was
developed by PSR and has been adopted worldwide to solve these type of scheduling problems.
Optimal network planning under uncertainty
The OptGen carries out a joint optimization of investments on new generation and regional interchange
capacity. In this step, NetPlan uses an AC power flow model to determine the least-cost set of transmission
reinforcements in each region.
2.4.3 Case study data assumptions

The study consists of a prospective analysis of the Brazilian power system considering a massive insertion
of VRE in a future configuration. The departing operating point was the system configuration defined by the
Ten-Year Brazilian Expansion Plan 2017-2026 prepared by EPE (PDE 2026). All new generation and trans-
mission capacity planned up to this horizon were assumed to be in full operation.
Market size
November 2019 21
20-26-00173
Processing No.: 15.2126.9-001.00
The end of horizon yearly demand was assumed as twice as the demand observed in 2017 (~600 TWh),
thus 1200 TWh. The modifications were sufficiently large to provide insights for the planning and operation
of the Brazilian power system due to the increased participation of VRE in the energy matrix.
The hourly demand data of year 2015, provided by ONS, was used as a profile for future years. The 2015
hourly profile per region was used to disaggregate monthly load block values into hourly values. This year
is considered as “well behaved”, in contrast with recent years that experienced load shedding that distort
typical data.
Supply options
New candidate projects (to be built after 2026) include coal, open cycle and combined cycle natural gas,
nuclear, biomass cogeneration, wind farms, solar PV power plants and storage devices. Hydropower is ex-
pected to expand as in PDE 2026 but not after that.
The characteristics of projects considered in the optimization include technical and economic information,
such as:
· Investment Cost;
· Fuel variable cost, based on contract (different types are allowed, from fully flexible to baseload);
· Fixed O&M cost;
· Variable O&M cost;
· Heat rate;
· Start-up cost;
· Ramping constraints;
· Lifetime.
Parameters vary per technology. In some cases, such as sugarcane biomass production, production is sea-
sonal (only during harvest period). Similarly, in the case of wind power, increased wind speeds take place
during the dry period.
Candidate wind power projects considered in this study are based on the projects authorized to participate
in the energy auctions after considering the requirements enabled by EPE. A list of 800 candidates was
considered in the simulations, considering the project’s data for the “A-4” auction of 2018, which includes
location, connecting substation and reliable yearly production estimated by the vendors. Technical data,
such as the turbine power curve and the project hub height were based on the forecasted evolution of wind
power technologies. Models of turbines were selected for candidate projects based on wind class according
to IEC61400-1: 2005.
The capacity expansion model OptGen considered the wind power candidate projects of 22 clusters of wind
data that were needed to make a spatial representation more adequate to the energy planning and to con-
sider the need to capture spatial variability at low computational cost.
After the candidate selection, the capacity expansion determined by OptGen was disaggregated within the
region using data from the individual projects located in the region from the complete list of 800 projects for
the both transmission reinforcement and grid operation studies that use a more detailed network.
As in the case of wind power, PV solar projects considered as candidates to the generation expansion model
were also based on information from the energy auctions. In this case, around 400 candidates were consid-
ered from the A-4 auction of 2018. For the 400 registered projects in this auction, information was obtained
as to the substation to which they would be connected and their physical guarantee, related to an annual
production with high reliability.
November 2019 22
20-26-00173
Processing No.: 15.2126.9-001.00
There are several available technologies for the energy storage, with different response times, storage vol-
umes and costs. In this study, the Lithium-Ion and Sodium-Sulphur (NaS) batteries were considered to pro-
vide short-term energy storage and transfer power from low to high demand hours in order to reduce oper-
ational costs. By the end of the horizon the investment cost of Lithium-ion and NaS batteries is expected to
decrease 60% from current costs.
Distributed generation
The rooftop solar photovoltaic was considered as a source of Distributed Generation (DG) in this study.
Biomass cogeneration and small hydropower plant connected to the distribution concession were included
indirectly through the load profile, considered by ONS for the monthly operation.
The total DG capacity forecast value provided by EPE for the end of horizon was 30 GW. Although the DG
capacity was given, the production is variable and modelled by scenarios using the same approach as for
utility scale VRE, using MERRA2/NASA solar hourly radiation time series. The process of producing elec-
tricity scenarios from MERRA2 was then calibrated to match with the reference production factors provided
by EPE for each state.
Problem dimensions
The capacity expansion model OptGen uses seasons, daily load profiles and scenarios of VRE production
and hydrology to decide on optimal expansion considering energy interchanges among regions, reserves
requirements, and integrating investment (binary variables).
Operative decision variables, such as hydropower, thermal power, reservoir storage, water flow through the
turbines, energy exchanges and several others are defined accordingly. The planning problem was modelled
as a mixed integer programming (MIP) optimization with 3.5 million variables (1 million integer) and 3.2
million constraints. Solving a problem of this dimension is computationally challenging and requires an ade-
quate infrastructure to run the model in the Cloud and then post-process several gigabytes of results to
provide insights for power system expansion and operation activities.
2.4.4 Results
Capacity expansion
The wind power is the leading technology in terms of capacity addition, with respect to PDE 2026, exceeding
41 GW, mostly installed the Northeast region (33.3 GW), with smaller amounts in the South region (8 GW).
The solar photovoltaic capacity expansion is around 20 GW, with 12.6 GW in the Southeast, 5.6 GW in the
South and 1.4 GW in other regions. In addition to the centralized solar, the model considers 30 GW of solar
distributed generation (DG) by the end of the study horizon. Therefore, the total solar PV addition (utility
scale + DG) amounts are nearly 50 GW. A total of 2.4 GW of Li-Ion battery capacity (2.4 GWh of energy)
was selected for battery storage due to reserve constraints in the Northeast.
An additional 1.4 GW was installed for open cycle gas turbines (OCGT) due to its higher dispatch flexibility
and capacity to provide reserves. A total of 8 GW of combined cycle gas turbines (CCGT) was added in the
Southeast, with the selection of the cheaper, and less flexible contract throughout. For technical reasons,
such as little dispatch flexibility, strong ramping constraints, and/or economic reasons, some technologies
were not selected, such as coal, biomass and nuclear.
The annuity of the new capacity investments after 2026 amounts to 7.4 billion USD. The yearly power
system operation cost amounts to 5.6 billion USD for the whole system (existing projects in 2026 and pro-
jects included later) and the total cost 13 billion USD.
In terms of capacity, hydropower that currently accounts for 69.8% of total installed capacity decreases its
share to 38.3% by the end of the horizon. Despite this reduction it is key to support the VRE growth, given
the flexibility to accommodate the dispatch. Wind power grows from 6.8% to 22.1% and solar power from
nearly zero to 19% (utility-scale plus DG), as depicted in Figure 2-2.
November 2019 23
20-26-00173
Processing No.: 15.2126.9-001.00
The results show a substantial growth of the VRE participation in the Brazilian power matrix for economic
and technical reasons, with no subsidies. Figure 2-3 illustrates a weekly operation in August, an average
hydrology scenario. As the graph shows, the production in the end of the horizon is nearly 50% by hydro
power, despite the reduction to 39% of share in terms of capacity. Wind, biomass, solar PV and DG contrib-
ute, respectively, to 26%, 3%, 6% and 4% of the generation.
The Brazilian electricity matrix would be 90% renewable and nuclear. The associated CO2 emission related
to fossil fuel production (remaining 10%) with respect to the total consumption is low. In summary, the strong
share of non-fossil generation expected for the end-of-horizon contributes to a very “clean” electricity matrix.
Installed Capacity [%]

Hydro 69.8%
38.3%
22.1%
19.0%
Biomass 8.8%
Wind 6.8% 5.9%
Solar 0.0%
2016 End of
Horizon
Figure 2-2: The evolution of percentage of total installed capacity
Thermal Biomass Wind Solar Hydro Batteries
180
160
140
120
100
GW
80
60
40
20
0
1 25 49 73 97 121 145
Figure 2-3: Total hourly production for a week in August
November 2019 24
20-26-00173
Processing No.: 15.2126.9-001.00
Transmission expansion
The reinforcements of high voltage grid (equal or higher than 230 kV) required to accommodate all the new
capacity added by Optgen included 234 lines/transformers with a total length exceeding 23,000 km and
nearly 10 billion USD of investments.
2.4.5 Conclusions
The main objective of this study was to identify and assess state-of-the-art methodologies and computational
tools that may be useful to EPE and ONS in their studies and decision-making process. The focus was on
Variable Renewable Energy (VRE) generation sources such as wind and photovoltaic solar for the following
reasons:
· The cost of these sources has been falling steeply, becoming economically competitive with more
traditional sources such as hydro and biomass;
· Brazil has a large potential for VRE sources;
· The Brazilian large and diversified hydro storage capacity and robust transmission network allow
the penetration of VREs;
· VREs are mostly non-dispatchable and are located in a wide range of sites, hence the distribution
of power flows in the high-voltage grid may be very different from the current situation. As a conse-
quence, it is necessary to consider jointly generation, transmission and hourly resolution in the plan-
ning and operational studies;
· The substantial insertion of VRE sources requires a detailed probabilistic modelling of generation
reserve requirements;
· Peak supply reliability methodologies and criteria should also be revised to consider the fluctuations
of VRE production.
These aspects were evaluated with a comprehensive generation planning and operation study, where de-
mand at the horizon was assumed to be twice the current value. This distant horizon would allow the impacts
and insights from VRE penetration to be more clearly perceived.
2.5 Power System Studies

This chapter presents the main conclusions from the different tasks covered in Product 3 of this project. The
conclusions are split per topic covered in this product.
2.5.1 Development of simulation models

A key aspect when performing power system planning under high VRE penetration levels is to couple energy
and electrical studies. The underlying reason for this coupling is because VRE penetration requires more
detailed operation of the energetic operation of the system (detailed representation of the technical charac-
teristics of generating units and full representation of the network) and the analysis through power system
studies of a higher quantity of system operating conditions.
In order to effectively perform power system planning activities with high shares of VRE sources, it is required
to seamlessly integrate the energy and electrical study models. A non-exhaustive list of parameters and data
that are needed for both energy and electrical studies is given as follows:
· Transmission network parameters: topology, impedance of branches, rated capacity of equipment,
transfer capacity limits between subsystems, etc.;
· Generating units’ data: , , , , connection node, etc.;
· Load data: connection node, distribution over time (demand profile), distribution over the nodes of
the system, etc.
November 2019 25
20-26-00173
Processing No.: 15.2126.9-001.00
The scheme employed in this project for the development of a the coupled “energy-electrical” model data-
base is depicted in Figure 2-4. It is strongly recommended to put in place a common database and model
management system and process in order to guarantee the quality of the results of the studies, as well as
ensuring the efficiency of the performed planning activities.
Figure 2-4: Coupling of energy and power system simulation models
2.5.2 Dynamic modelling of wind and solar PV power plants

Utility-scale wind and solar PV power plants may consist of tens to hundreds of individual generating units
(wind turbines or solar PV inverter stations). Usually a supervisory control at plant level is applied to regulate
the terminal behaviour (at the point of connection – PoC) of the entire power plant, by dispatching commands
for active and reactive power or voltage control to the individual units. The detailed representation of wind
and solar PV power plants for system-wide studies results is very large models which are computationally
demanding to be simulated.
It is well-known in the power system community that the representation of wind and solar PV power plants
in high level of details for system-level studies does not bring added value to the studies. It is therefore a
common practice around the world to simplify these models in order to make them computationally tractable
while retaining the main dynamic characteristics of the plants.
For WPP modelling, the Consultant implemented the equivalent model for WPPs defined in the IEC 61400-
27-1 (Edition 1.0 2015-02) standard [1]. This model is built based on the following assumptions:
· The representation of the wind power plant is simplified in order to represent the wind turbines and
the internal collector network by an equivalent wind turbine at the PoC of the plant to the grid.
· The chosen wind turbine is the Type 4 one (full converter).
· The model is developed to represent wind power generation in studies of:
o Large-disturbance short-term voltage stability phenomena.
November 2019 26
20-26-00173
Processing No.: 15.2126.9-001.00
o Other dynamic short-term phenomena such as rotor angle stability, frequency stability and
small-disturbance voltage stability.
· The model is applicable for dynamic simulations of power system events such as short-circuits (low
voltage ride through), loss of generation or loads, and system separation of one synchronous area
into more synchronous areas.
· The model is specified for fundamental frequency, positive sequence response.
· The model is not intended for:
o Long-term stability analysis.
o Investigation of sub-synchronous interaction phenomena.
o Investigation of the fluctuations originating from wind speed variability in time and space.
· The model includes wind power plant level controls.
· The model does not include additional equipment such as SVCs and STATCOMs.
For SPP modelling, the Consultant implemented an equivalent PV power plant model based on the works
performed by the “Renewable Energy Modelling” task-force set up by the Western Electricity Coordinating
Council (WECC) in 2012 [2]. This model is built based on the following assumptions:
· The representation of the solar PV power plant is simplified in order to represent the PV inverters
and the internal collector network by an equivalent inverter at the PoC of the plant to the grid.
· The model is developed to represent solar PV power plants in studies of:
o Large-disturbance short-term voltage stability phenomena.
o Other dynamic short-term phenomena such as rotor angle stability, frequency stability and
small-disturbance voltage stability.
· The model is applicable for dynamic simulations of power system events such as short-circuits (low
voltage ride through), loss of generation or loads, and system separation of one synchronous area
into more synchronous areas.
· The model is specified for fundamental frequency, positive sequence response.
· The model is not intended for:
o Long-term stability analysis.
o Investigation of sub-synchronous interaction phenomena.
o Investigation of the fluctuations originating from wind speed variability in time and space.
· The model includes solar PV plant level controls.
· The model does not include additional equipment such as SVCs and STATCOMs.
The Consultant therefore recommends the use of the aggregated wind and solar PV power plant models in
system planning studies. Moreover, it is also recommended that a model development and maintenance
scheme as the one recommended in Product 1 of this project be implemented for the modelling of WPP’s
and SPP’s.
2.5.3 Selection of relevant operating conditions

Performing power system studies in systems with high VRE shares require the analysis of a greater number
of system operating conditions. The underlying reason is the variable nature of VRE sources, resulting in a
November 2019 27
20-26-00173
Processing No.: 15.2126.9-001.00
much wider diversity of system operating conditions than in systems composed only by conventional gener-
ation3. This large amount of operating conditions to be simulated requires:
· A methodology to select the relevant operating conditions for detailed power system analysis and
identify the new weak points of the system;
· High performance tools able to simulate a large number of operating conditions in reasonable
computation times.
In this project, the selection of the relevant operating conditions starts by a detailed analysis of the results of
the simulations of the hourly operation of the system. This analysis aims at understanding the main charac-
teristics of the energetic operation of the system in order to support the selection of the operating conditions
for which detailed power system studies need to be carried out. Once this analysis is concluded, the selection
of the relevant operating conditions is performed using two complementary approaches, as follows:
· Selection of extreme operating conditions;
· Selection of “likely” operating conditions.
Extreme operating conditions are defined as operating points that represent maximum stress for the system
from different points of view.
For the selection of “likely” operating conditions, different configurations of the clustering and the pre-pro-
cessing algorithms have been tested without success for the case of this project. However, the Consultant
already employed similar approach to other systems (e.g. Argentina and Bolivia) with a great level of suc-
cess. For the case of the Brazilian system, more research is needed in order to obtain results with the right
level of confidence for this kind of application. Because of that, the Consultant decided to not consider the
“likely” operating conditions for the power system studies in this project. However, the variety of extreme
operating conditions selected for these studies seem to be sufficient for the objectives of the project.
As a result, operating conditions presented in Table 2-3 are selected for the power system studies4.
Table 2-3: Selected operating conditions for the power system studies
# Extreme Operating Conditions Scenario 4 Scenario 8
1 Annual peak load 20XX-02-12 15:00 20XX-02-12 15:00
2 Annual minimum load 20XX-06-28 07:00 20XX-06-28 07:00
3 Maximum export: NE 20XX-09-13 15:00 20XX-07-08 07:00
4 Maximum import: NE 20XX-03-14 23:00 20XX-04-07 23:00
5 Maximum export: S 20XX-04-11 23:00 20XX-06-26 17:00
6 Maximum import: S 20XX-02-16 09:00 20XX-03-14 11:00
7 Maximum import: SE 20XX-04-20 18:00 20XX-08-18 18:00
7.a At highest instantaneous VRE share 20XX-09-11 18:00 20XX-05-15 12:00
7.b At highest instantaneous VRE share in SE 20XX-04-22 08:00 20XX-05-15 11:00
7.c At lowest instantaneous VRE share 20XX-05-05 20:00 20XX-06-18 19:00
3
In these systems, the main source of variability to the system operation is the load.
4
In Product 3, static and dynamic analyses have been carried out only for Scenario 4.
November 2019 28
20-26-00173
Processing No.: 15.2126.9-001.00
7.d At lowest instantaneous VRE share in SE 20XX-04-20 18:00 20XX-08-18 18:00
8 Highest VRE/Load share 20XX-09-06 10:00 20XX-07-19 12:00
8.a At high load conditions 20XX-09-11 11:00 20XX-08-18 11:00
8.b At low load conditions 20XX-07-19 07:00 20XX-06-21 07:00
2.5.4 Analysis of system inertia

A detailed analysis of the system inertial response considering simplified mathematical models for the sys-
tem has been carried out. Even if the analyses make use of a simplified model for the description of the
physical phenomena behind the system inertial response, the values and insights provided by these anal-
yses are very useful as a first assessment of the impact of VRE integration on the system frequency stability.
Table 2-4 and Table 2-5 present a synthesis of the results presented in this section considering only the
synchronous inertia and the total potentially available inertia, respectively. The ROCOF values and the times
to reach the first UFLS stage were calculated considering a generation imbalance equal to 1500 MW,
which represents the loss of the biggest generating unit of the system.
Table 2-4: Synthesis of system inertia evaluation results (synchronous inertia only)
Synchronous Inertia System ROCOF Time to First UFLS Stage 1
Scenario Mean Min. Max. Mean Min. Max. Mean Min. Max.
[MW∙s] [MW∙s] [MW∙s] [Hz/s] [Hz/s] [Hz/s] [s] [s] [s]
2 407,172 188,465 639,443 0.115 0.070 0.239 13.6 6.3 21.3
4 394,747 157,419 640,790 0.120 0.070 0.286 13.2 5.2 21.4
5 410,171 147,735 644,363 0.115 0.070 0.305 13.7 4.9 21.5
6 425,865 191,636 633,445 0.108 0.071 0.235 14.2 6.4 21.1
7 392,924 166,303 639,088 0.120 0.070 0.271 13.1 5.5 21.3
8 386,954 155,312 625,923 0.122 0.072 0.290 12.9 5.2 20.9
9 393,817 625,923 646,387 0.120 0.070 0.264 13.1 5.7 21.5
2 421,439 201,291 646,893 0.111 0.070 0.224 14.0 6.7 21.6
4 409,905 174,532 652,110 0.115 0.069 0.258 13.7 5.8 21.7
5 423,881 164,941 652,693 0.111 0.069 0.273 14.1 5.5 21.8
6 438,676 206,195 638,801 0.105 0.070 0.218 14.6 6.9 21.3
7 407,365 184,467 647,489 0.115 0.069 0.244 13.6 6.1 21.6
8 402,520 170,066 632,435 0.117 0.071 0.265 13.4 5.7 21.1
November 2019 29
20-26-00173
Processing No.: 15.2126.9-001.00

9 408,087 187,245 650,121 0.116 0.069 0.240 13.6 6.2 21.7
In which regards the performance of the primary frequency control, it can be seen that in the worst case,
the frequency nadir is equal to 59.45 Hz (at 5.4 seconds). Moreover, for the majority of the operating
conditions the frequency nadir lies within the range of 59.6 and 59.9 Hz (99.8% of the hours) while the time
instant of the frequency nadir lies within the range between 3 and 4.5 seconds.
The results obtained in this analysis indicate that, despite the high VRE shares obtained for the studied
scenarios, the system inertia remains at sufficiently high values for the majority of the analysed operat-
ing conditions. This is partly explained by the fact that the planning scenario considered in this study con-
siders a high demand growth and the high VRE capacity expansion is majorly related supplying this addi-
tional demand rather than replacing conventional generation.
This is a point of attention when comparing the future expansion of the Brazilian system with, for example,
the European experience in VRE integration. In that case, the demand growth is moderate and the fast VRE
expansion comes in replacement of conventional generation, which leads to a potentially high degradation
of the system inertia.
2.5.5 Static analysis

In a first step, successive DC load flow and DC static security assessment (N-1) analyses in order to have
been performed in order to:
· Take into account power flow limits in transmission equipment outside the main transmission net-
work ( < 230 );
· Take into account N-1 contingencies.
A second step was performed in order to design the network reactive power compensation strategy by spec-
ifying additional network reinforcements and adjusting the system operation using an optimal power flow
(OPF) tool. As expected, no significant reinforcements in terms of additional transmission lines and/or trans-
formers were required. It can be explained by the fact that the majority of the active power transfer problems
are solved in the previous step. Therefore, the great part of the network expansion and reinforcement options
that are to be decided in this step are shunt reactive power compensation devices to allow proper voltage/re-
active power control in the network.
Table 3-14, Table 3-15 and Table 3-16 summarise the network expansion defined in the static studies (trans-
mission lines, transformers and shunt compensation, respectively). It must be emphasized that these net-
work reinforcements come on top of the expansion plan defined in Product 2 (energy studies).
Table 2-6: Transmission expansion summary - transmission lines (static studies)

New Transmission Lines
Region
Total < 230 kV 230 kV 345 kV 500 kV 525 kV
CW 68 51 9 1 7 0
N 46 38 5 0 3 0
NE 165 75 77 0 13 0
S 158 95 58 0 3 2
SE 363 288 9 40 25 1
November 2019 30
20-26-00173
Processing No.: 15.2126.9-001.00

Region
Total 800 547 158 41 51 3
Table 2-7: Transmission expansion summary - transformers (static studies)

Main Transmission Sub Transmission Three-Winding
Region Total Interface
Network Networks Transformers
CW 77 1 58 6 12
N 71 1 56 7 7
NE 155 16 32 103 4
S 67 2 8 28 29
SE 131 7 86 17 21
Total 501 27 240 161 73
Table 2-8: Transmission expansion summary – shunt compensation (static studies)

Total Capacity [Mvar]
Region
Shunt Capacitors Shunt Reactors
CW 534 -4
N 521 -506
NE 201 -6
S 1817 -625
SE 3136 -536
Total 6209 -1677
A full load flow and static security assessment analysis was performed for the final network configuration
for the selected 14 operating conditions. The results show that the use of the proposed methodology and
OPF formulation lead to adequate voltage profiles across all voltage levels while maximizing the reactive
power margins of the generating units and dynamic var compensators. No violation of voltage magnitude
limits and no network overloads are observed in normal operation (N condition). The average of transmis-
sion losses for the 14 operating conditions is 5.6% of the total generation (mini-mum losses equal to
3.7% for Case 4 and the maximum corresponding to 8.6% for Case 2). Case 1 presents the highest number
of overloaded transformers, with a total of 3 transformers with 124.7% of over-load each. Case 4 presents
two overloaded transformers, the first one with 120.1% of maximum overload caused by 22 different contin-
gencies and the second one with 145.6% of maximum overload caused by 23 different contingencies. Cases
2, 3, 6, 7, 7a, 7b, 8a and 8b do not present over-loaded transformers. All overloaded transformers are con-
centrated in the Northeast region.
A total of about 2470 contingencies are considered for the static security assessment. The possible
contingencies consider the simple loss of a circuit (transmission line or transformer) at the main transmission
network ( ≥ 230 ). Case 8b presents the highest number of overloaded lines, with a total of 11 lines
concentrated in the Northeast region, with overloads ranging from 110.8% to 120.0%. Cases 1, 2, 3, 5, 6, 7,
7a, 7b, 7c, 8 and 8a do not present overloaded lines in the main transmission network. The overloaded lines
are more concentrated in the Northeast region, with the exception of two lines concentrated in the South
region. The problems remaining in the final network configuration are not considered critical (overloads not
too important and limited to a few contingencies). Additional network reinforcements could be proposed to
solve those problems.
November 2019 31
20-26-00173
Processing No.: 15.2126.9-001.00
The short-circuit current analysis focused on the 230 and 500/525 kV networks. As shown in the maxi-
mum short-circuit current plots, the 230 kV busbars present short-circuit currents up to 30 kA. In the North
(N) and Central-West (CW) regions, these currents are more concentrated around 5 kA. On the other hand,
in the Northeast (NE), Southeast (SE) and South (S) regions the short-circuit cur-rent values are more dis-
persed. The 500/525 kV busbars present short-circuit currents up to 50 kA. In the Northeast (NE) region,
these currents are more concentrated around 20 kA. In the South (S) region these currents are more con-
centrated between 20 kA and 30 kA. It is not observed a generalized decrease in the short-circuit power
throughout the network. For some specific cases, the short-circuit currents are above 50 kA, indicating the
need for a revision regarding possible equipment capacity violations. This is explained by the following facts:
i) strong expansion of the transmission network, reducing electrical distances; ii) significant additions of gen-
eration (even if predominantly VRE), increasing the short-circuit current contributions in the network; iii) sys-
tem expansion based on a significant demand growth and not by replacing conventional generation by VRE
under a more or less constant load condition.
A special attention was dedicated to the analysis of the short-circuit currents at the terminals of the
HVDC converter stations. The major risk of insufficient short-circuit power lies in the operation of the in-
verters. In this case, it can be seen that the SCR is sufficiently high for the inverter operation (ensuring the
commutation angle higher than the minimum commutation angle) of all HVDC links (even the bi-directional
ones). Therefore, for the operating conditions analysed in this study, the lack of short-circuit power for the
operation of the HVDC links does not seem to be a major concern. However, dedicated studies to analyse
the operation of these links in such future operating conditions is recommended (e.g. analysis of multi-infeed
problems, etc.).
In which regards the system strength for the operation of VRE power plants, the results show that 16 VRE
power plants (total of 21,180 MW of installed capacity) are exposed to low SCR at their points of connection
(9 WPPs and 7 SPPs). In order to solve the low SCR problems at the PoC of these power plants, synchro-
nous condensers have been allocated in key points of the network aiming at increasing the SCR levels and
allow secure and stable operation of these power plants. A total of 22 synchronous condensers (150 MVA,
-90/+150 Mvar each) have been allocated in points of the network close to the main VRE hubs in the NE
and S regions, mainly.
In which regards the impacts of DPV in sub-transmission and transmission networks, the analyses carried
out by the Consultant show that:
· For the operating conditions where there is no injection from DPV, the average loading of the sub-
transmission circuits remains within the range between 29.3% and 34.2%. For the cases where
there is DPV injection, the average loading of the sub-transmission circuits is between 24.1% and
39.6%.
· In the hours without DPV injection, the number of transformers presenting reverse power flows are
between 7.1% and 7.7% of the total number of interface transformers. However, for the operating
conditions with non-null DPV injection, the number of interface transformers presenting reverse
power flows increase to a range between 7.7% and 11.6%. Therefore, an increase in the reverse
power flows is observed under the presence of DPV.
The Consultant recommends a more detailed analysis in order to reach a stronger conclusion on this subject,
taking into account the different load levels and other generating units (in addition to DPV) connected to the
sub-transmission networks.
2.5.6 Dynamic analysis

The results of small-signal stability analysis show that in the original configuration there were unstable
and/or poorly damped inter-area modes for the operating conditions 1, 2, 3, 8, 8.a and 8.b. A PSS tuning
November 2019 32
20-26-00173
Processing No.: 15.2126.9-001.00
exercise was carried out in order to stabilize the system and improve the performance of oscillations damp-
ing.
The PSS type adopted in this study is the IEEE PSS2B. The units considered for PSS allocation and tuning
are mostly the new combined- and open-cycle power plants. In addition to those units, the future units of
Angra 3 NPP and Tabajara HPP, as well as the existing units of Termonorte 2, have their PSS tuned to
provide positive damping to the inter-area modes (0.5 to 0.65 Hz). Results for the updated system configu-
ration (with new PSS) show the effectiveness of the implemented PSS configuration. Some poorly damped
local modes remain after the PSS tuning exercise. However, the tuning of PSS for solving problems with
local oscillation modes is out of the scope of this study.
Considering the scenarios with very high VRE penetration analysed in this study, it is recommended further
investigation on the possibility of inclusion of supplementary control signals in WPP and SPP in order to
contribute to the damping of inter-area modes. This is valid especially in the cases where there are few
synchronous units in operation in the N/NE systems, cases in which damping the inter-area mode has proven
to be a challenge in this study.
A total of 1484 transient stability simulations have been performed in this study (106 faults for 14 operat-
ing conditions). During the DSA simulations, voltage control issues were encountered in the cases of higher
VRE penetration (2, 3, 8, 8.a and 8.b). In order to solve those issues, additional generators that were not
dispatched in those cases but that are capable of operating as synchronous condensers were put in service
to counteract the problems.
In the final system configuration, the system remains stable for all simulated faults and for all operating
conditions. It must be highlighted the importance of the synchronous condensers in securing the system
operation in cases of high instantaneous VRE penetration. It is of utmost importance that sufficient short-
circuit power remains available in the moments of high VRE generation in order to avoid instability of the
VRE power plant control loops and, as a consequence, undesirable disconnection of these units during
transients. Moreover, the behaviour of the VRE power plants (SPP and WPP) during and just after faults
must be certified during the commissioning phase of those projects. The fault current injection during faults
is essential to contribute to system stability and the respect to the fault ride-through characteristics too.
A total of 98 frequency stability simulations have been performed in this study (7 contingencies for 14
operating conditions). The frequency stability of the system for the different operating conditions was as-
sessed by simulating the loss of the largest generating unit in operation in each subsystem. The simulation
results indicate that even for operating conditions with very high instantaneous VRE penetration the system
remains stable and the system frequency remains always significantly above the first stage of the UFLS
scheme. The minimum frequency nadir from all simulations is equal to 59.72 Hz, which is significantly
above the first threshold of the UFLS scheme (58.5 Hz). It indicates a sufficient frequency stability margin
on the system.
The behaviour of the different generation technologies in the PFC performance is a point of attention and
must be carefully taken into account in the planning studies. The following is to be highlighted from the
simulations:
· Nuclear and CPPP units (ST-based) present a quick response just after the contingency but that
fades out quickly.
· OCGT and CCGT units present a good PFC performance (fast reaction towards frequency devia-
tions). However, special attention should be paid to the fact that these units present a higher contri-
bution to the PFC in the first seconds after the contingency and then after about 10 seconds their
power outputs decrease towards a stable level.
· HPP’s present the slowest PFC control performance amongst the different types of units providing
PFC in the system. Their responses are limited by the constructive characteristics of the hydro power
November 2019 33
20-26-00173
Processing No.: 15.2126.9-001.00
plants and the non-minimum phase behaviour of the units (compromise between performance and
stability of the turbine.
· Biomass and small hydro units provide PFC support. However, in the model only a few of these
units have an associated dynamic model and therefore their contribution to the PFC is not captured
in this study.
· WPP’s, SPP’s and distributed generation are not presented in the results because these units do
not participate in the PFC.
The importance of the interconnections for the performance of the PFC is also highlighted in this study.
It is possible to see that for a contingency in a given subsystem, all tie-lines contribute to bringing additional
power from outside the given subsystem in order to compensate for the generation loss. However, in order
to guarantee that the system benefits from the reserves shared among the different subsystems, adequate
transmission reliability margins must be secured in the operational planning5.
Additional sensitivity analyses of HVDC contingencies have been carried out. It is shown that all pole
contingencies (normative contingencies) result in a stable system operation while respecting the dynamic
performance criteria. It is also shown that in all simulated cases the loss of the bipoles of Belo Monte 1, Belo
Monte 2 and “Milagres-Silvania” result in a stable system operation without the need for SPS activation,
which represents a significant result in terms of system operation security. This is a result of the significant
network reinforcements (including in the interconnections between subsystems) resulting from this study.
2.5.7 General recommendations

This subsection presents the general recommendations derived from the different activities performed in
Product 3 divided in the following categories:
· System modelling and data management;
· Simulation infrastructure;
· Expansion planning;
· System operation;
· Capacity building.
System modelling and data management

· Seamless integration between energy and power system simulation models.
· Implementation of automatic data consistency tests and reporting.
· Implementation of a model management platform with the capabilities of version control, modifica-
tion tracing, validation and approval process, etc.
· Thorough review of the turbine/governor models of the Brazilian power system dynamic model da-
tabase to ensure adequate representation of these equipment for frequency stability and PFC per-
formance studies.
· Make use of aggregated wind and solar PV power plant models in system planning studies. Moreo-
ver, it is also recommended that a model development and maintenance scheme as the one recom-
mend-ed in Product 1 of this project be implemented for the modelling of WPP’s and SPP’s.
5
It is expected that an increase in VRE penetration would lead to the need of increased transmission reliability mar-
gins in order to allow safe operation of the system.
November 2019 34
20-26-00173
Processing No.: 15.2126.9-001.00
Computational infrastructure
· High performance power system simulation tools.
· Cloud-based infrastructure.
· Integrated database management system.
· Results processing tools based on advanced data analytics techniques.
· Use of advanced visualization tools, including intensive use of GIS-based applications.
Expansion planning
· Transmission infrastructure is key for successful VRE integration in the Brazilian system. The trade-
off between VRE resource quality and additional transmission capacity investment should be as-
sessed not only from the point of view of interconnection between subsystems, but also within each
subsystem (e.g. 230 kV and 500 kV reinforcement needs in NE and S regions).
· Impact of distributed generation on system planning
· Integrated transmission, sub-transmission and distribution network planning
· Further develop data analytics approaches to select critical and likely operating conditions for power
system studies.
· Take into account in the planning exercise advanced and/or new transmission equipment to allow
secure system operation as an alternative for massive investments in new transmission lines.
· Improve analysis of system strength in the methodologies for short-circuit current analysis in every
planning exercise.
· Improve frequency stability and net transfer capacity analyses in the standard planning methodolo-
gies.
· Ensure that the maximum loss of infeed due to loss of connection networks of wind and/or solar PV
power plants remains lower or equal to the sizing incident used as a planning and operation crite-
rion6.
· Adoption of state-of-the-art and highly flexible optimal power flow tools in the planning process.
· Implement a “quantification and qualification centred” expansion planning framework in order to
quantify and qualify the system impacts of VRE integration and the related costs.
System operation
· Ensure adequate transmission reliability margins in order to allow effective frequency regulation
support from the conventional generating units (mainly located in the CW and SE regions) to the
areas with higher instantaneous VRE penetration (NE and S).
· Integrate the proposed methodology for system inertia evaluation in the short-term operation plan-
ning in order the system inertia in the generation scheduling phase (including assessment of PFC
performance using a “single-node equivalent” representation).
· Guarantee the availability of the operating reserves (primary and secondary).
· Implement a process of regular certification of generating unit performance for power frequency
control applications (primary and secondary) and its adequacy to ROCOF requirements.
· Online monitoring of system inertia, electromechanical oscillations damping and transmission mar-
gins.
6
Aims at avoiding the connection of large wind and/or solar PV power plants (or groups of them through the same
connection system) through radial connection networks that might results in loss of infeed larger than the normative
contingencies.
November 2019 35
20-26-00173
Processing No.: 15.2126.9-001.00
· Ensure regular evaluation of the adequacy of special protection schemes (SPS), power system sta-
bilizers (PSS) and power oscillation damping (POD) controllers in face of highly variable generation
dispatch profiles.
· Regularly evaluate the adequacy of the UFLS schemes taking into account the rapid deployment of
solar PV distributed generation.
· Regularly review the definition, minimum performance requirements, sizing and allocation of oper-
ating reserves (primary, secondary, tertiary, etc.).
· Keep grid codes up-to-date with respect to the system evolution (see Product 1 for further infor-
mation).
· Put in place effective grid congestion management in order to minimize curtailment of VRE produc-
tion due to network- or stability-related constraints.
Capacity building
· Develop advanced knowledge of the principles of operation and details of controls and capabilities
of inverter-based grid interface technologies.
· Teams in charge of energy studies should receive a training on the basics of power systems oper-
ation, stability and control in order to understand the implications of decisions taken at the energy
studies level on operational performance of the system.
· Teams in charge of power studies should receive a training on the basics of energy studies in order
to understand the implications of decisions taken at the energy studies level on the system opera-
tion, as well as to enable the teams to provide the right feedbacks for the energy studies.
· Intensive training of system operators in a full-dynamic dispatching training simulator (DTS)7 should
be put in place in order to allow the operators to better understand the impacts of the VRE penetra-
tion in the dynamic performance of the system.
2.6 Methodology Studies

Based on the findings of products 1, 2, 3 and 5, the Consultant propose specific recommendations for fine-
tuning the guidelines, methodologies and criteria to be in the planning activities (both expansion and opera-
tion) of Brazil. These recommendations are summarised next.
2.6.1 Energy studies

summarizes the methodological recommendations to be used in energy planning practices considering the
rapid transformation of power systems due to factors such penetration of VRE, distributed energy resources,
and others.
7
Also known as Operators Dynamic Training Simulators (ODTS).
November 2019 36
20-26-00173
Processing No.: 15.2126.9-001.00
Table 2-9: Energy studies - summary of methodological recommendations

# Theme Current Methodology8 Methodology used in Study Possible improvement Analysis/Challenges
Challenges involve modelling, as

(1) Use generation & transmission some reliability modules would
reliability assessment and integra- need to integrate the planning pro-
tion with planning tools. cess and data acquisition (unavail-
ability rates based on historic out-
Two software are used, one for
ages per unit).
generation system expansion
planning and the other for a sto-
A model for system expansion and
chastic operation planning. There (2) Use of model for project finan- The procedure needs to be de-
another for operation with hourly
is an iterative process between the cial analysis to incorporate in- signed from scratch as it is pres-
steps (instead of load blocks), indi-
models until the marginal cost of vestor's risks so that complete ex- ently not made by EPE. Financial
vidualized projects (instead of ag-
expansion is equal to expected pansion plan is compatible with results from auctions can be used
gregate reservoirs per region) and
Models used in value of the marginal cost of oper- developer's decisions. to infer risk profiles.
a probabilistic modelling of hourly
planning study ation (condition for global opti-
1 renewables production. A special-
and general mum). The plan is then evaluated
ized model was used for transmis-
methodology to assure there is sufficient capac-
sion expansion using a linear
ity to meet the peak demand for a (3) robust grid planning model, in- Coordination efforts by EPE’s
power flow model for multiple sce-
reliable supply. This verification cluding N-1 criterion, while consid- energy and electric areas for inte-
narios of production, instead of se-
links capacity-related constraints ering multiple dispatch scenarios. grated planning.
lected conditions, to support a ro-
with energy-related constraints.
bust optimization.
For instance, capacity of hydro
producers is conditioned to the en-
ergy availability.
Modelling issues and database
(4) Reactive power investment op- preparation of resources of reac-
timization. tive power and their investment
unit costs.
8
Methodology used on PDE 2026
November 2019 37
20-26-00173
Processing No.: 15.2126.9-001.00

Monthly profile and contribution in
each load block are estimated out- (1) Enhancement of both models
Representation of VRE needs to
side models. for expansion planning and opera-
Use of location of projects that improve in both models (expan-
tion to include probabilistic VRE
Preparation of Wind power: based on data from participated in electricity auctions, sion planning and system opera-
production. (2) Preparation of sce-
VRE candidate projects enabled to take part in the calibration of existing energy rec- tion). The representation should
narios using either a statistical
projects for ex- electricity auctions. Production in ords with MERRA database for ex- move away from monthly mean
2 method or machine learning. (3)
pansion each load block based on AMA tension of time series, use of production (zero variance) to
Hourly or sub-hourly data acquisi-
planning stud- database (actual production). So- Bayesian network for the genera- hourly or sub-hourly scenarios that
tion from VRE projects would need
ies lar PV power: reference values of tion of multivariate scenarios of can incorporate the modulation of
to be organized for planning pur-
auctions used. Production contri- hourly VRE production each source and corresponding
poses and integrated with en-
bution in each load block based on variability.
hanced planning models.
INPE database.
There are older inventories with
Screening of hydro projects based Hydropower inventory studies con-
assumptions that need revision.
on inventories and feasibility stud- sidering both techno-economic
The study did not focus on new The task is to update these studies
Hydropower ies. Sufficient lead time for project and socioenvironmental issues
hydropower plants in addition to under a new paradigm, with social
3 candidate pro- preparation, permits, and others to supported by specialized computer
those already considered in the and environmental considerations
jects participate in the 10-year horizon. model for the preparation of sensi-
PDE. from early stage planning, follow-
Selection based on viability & soci- ble candidates for expansion stud-
ing a more participatory process
oenvironmental complexity. ies.
for the discussion of alternatives.
As an assumption for planning

Investigation of trade-offs between
study resources of region A were Evaluation and definition of spe-
Multi-area Allows resources of region A to multi-area x single-area supply of
not allowed to supply reserves of cific criteria to measure the robust-
4 spinning re- supply reserves of region B, fol- reserves and best practices with
region B, even though this is part ness of the system and define the
serve lowing ONS practice. respect to system
of normal operation practice car- minimum desirable requirements.
expansion planning studies.
ried out by ONS.
An important improvement given

Exogenous scenario of distributed the continuous growth of DG. An
Use of endogenous approach
Distributed EPE utilizes external scenarios of generation imported to study. Allo- analysis of the requirements
based to coherently integrate DG
5 Energy Re- DG based on specific models cation of DG capacity made in shows a large challenge to make
growth scenarios within the plan-
sources (such as diffusive BASS model) "frontier buses" that connect the this operational (GIZ is currently
ning framework.
HV grid with the distribution utility. supporting a study with this objec-
tive).
Limit the capacity of candidate The challenge is to include trans-
Maximum
projects selected in each region mission bottlenecks and others in
6 VRE per No limitation imposed No limitation imposed
breaking up larger clusters into the selection of the number of re-
region
smaller ones if needed. gions of investment decisions in
November 2019 38
20-26-00173
Processing No.: 15.2126.9-001.00

addition to wind patterns/profiles
used in the study.
EPE makes probabilistic assess-

Firm capacity margin with respect Include a system reliability module Introduce the model as a compo-
Reliability of ments of the supply criteria9 (mar-
to peak annual demand and firm that provides feasibility cuts to the nent of planning model and treat
7 expansion ginal cost of expansion equal to
energy margin with respect to expansion model to meet reliability the data, based on equipment out-
planning average marginal cost of opera-
mean demand criteria (LOLP, EENS). ages, for planning purposes.
tion)
This model is required to separate

the factors of project unavailability
(forced outages and scheduled
maintenance). There are at least
Project Include model to determine repre- two advantages: (1) reliability crite-
Use of unavailability rates combin-
maintenance sentative maintenance schedules ria can be introduced in the plan-
8 ing forced and maintenance out- Same as current practice
of the projects to avoid degrading ning model (as seen); (2) dispatch
modelling ages.
supply reliability. scenarios of improve integration
with transmission planning / elec-
trical studies and electrical studies
because nominal capacity is con-
sidered.
A better representation of trans-

mission bottlenecks is the motiva-
Use of community identification al- tion for this recommendation.
Spatial gorithm to determine which buses The number of regions is decided
Seven systems were used, with
representation: EPE uses 11 systems and of the complete HV grid should be after simulating the operation of
9 some differences with respect to
number of sys- 3 fictitious nodes associated in each region for a policies made with a reduced num-
current practice.
tems given number of regions (input ber of regions with the full network
data). and examining impacts on costs.
Association of buses to regions is
based on nodal values.
9
In December 2019, the National Energy Policy Council (CNPE) approved the use of new supply criteria in Brazil. This improvement was necessary to adapt the power
sector to the new technological reality that has been established and the new market design that is intended for the electricity sector
November 2019 39
20-26-00173
Processing No.: 15.2126.9-001.00

Use of simplified representation
The objective is to improve the
with energy transfers among re-
modelling of constraints on the HV
Regional gions limited by informed amounts Use of linear representation that
grid in the expansion planning pro-
10 exchange (based on reliability criteria used in Same as current practice approximates an aggregate effect
cess, thus reducing the gap be-
limits transmission studies). Sum of en- of Kirchhoff second law.
tween energy planning studies and
ergy flows in different from-to ex-
electric studies.
changes also considered.
It is important to establish criteria

for the number of hours that Run
Spinning All Run-of-river plants can follow Introduce hydrology-related con- of River plants can be sustained at
A fraction of the capacity of each
reserve for run- load, except those with high sea- straints, especially in large pro- a given capacity. This may be a
11 plant is available for the supply of
of-river hydro sonality, which contribute with their jects with a strong seasonal pat- function of water inflows due to
reserve for the units in operation.
plants mean monthly production. tern such as Belo Monte (11GW) seasonal effect. This is true for
both reserve allocation and supply
of peak demand.
Large database required for a bot-

tom up demand modelling. Ad-
Incorporation of Demand Re- vantage of modelling effects that
Demand forecasted on macro-eco- sponse and the ability to model are with specific to some sectors
Demand Profile based on 2015 hourly curve
12 nomic variables. Hourly profile structural changes in the demand (e.g. effect of a change in the
forecast by ONS per region
based on verified real data. profile for the different classes of structure of the residential tariff) or
consumption. cross-sectorial, such as an in-
crease on efficiency standards of
motors.
Suppose demand increases with

temperature (due to air condition-
Temperature
Use of temperature and other cli- ing) and that wind velocity is re-
effect on
matic variables in the preparation duced in hot days. Clearly the
13 demand and - -
of scenarios of demand and VRE combined effect on supply and de-
VRE
production. mand is more challenging if these
production
factors are treated jointly than in-
dependently.
Pumped hydro storage (PHS) is a

Energy stored during surplus Preparation of more candidates for
Two technologies were used in special case because projects
Energy power for use in moments with system expansion considering ca-
14 study (Li-Ion 1h battery and NaS must be located in locations with
storage smaller availability with respect to pacity, stored energy, ramping
6h thermal storage) large water head between upper
load, including losses. constraints and location in the grid.
and lower reservoirs.
November 2019 40
20-26-00173
Processing No.: 15.2126.9-001.00
The Global Wind Atlas, for in-

Mesoscale VRE models can be stance, introduces a high-resolu-
Mesoscale
used to downscale reanalysis tion topographic data, such as
15 wind and - -
data, such as MERRA2 that used hills, ridges and land use in a wind
solar models
in study. flow model. Data availability and
security are a challenge.
Risk premiums may be added to

Project costs are either based on
Project project costs. Ideally an interactive
references per technology for Inclusion of risk premiums in pro-
developer process should be made until the
16 "general projects" or come from Same as current practice ject costs when carrying out a cen-
vs. central expansion plan is also financially
specific studies, as hydropower vi- tralized expansion planning.
planner viable considering the developer’s
ability studies.
risk perspective.
Criteria should be defined and in-

corporated in model (such as the The main challenge is to design
use of small variable production the criteria for VRE curtailment
VRE Curtail- Optional cost of curtailment (not
17 - costs) to determine a priority if cur- quickly before it becomes a critical
ment Criteria used in study)
tailment is necessary, to guaran- and controversial issue with the in-
tee a fair operation for the various crease of VRE in the next years.
producers.
Represent the “unknown un-
A careful design of scenarios un-
knowns”, or “black swans” i.e. se-
der unusual, though possible
vere events that are extremely un-
events, is required. As an exam-
likely from the standpoint of sto-
ple, the Ministry of Energy and
The sector models known chastic models, but that may actu-
Resilience Mines of Chile used simulation
18 unknowns, such as water inflows Same as current practice ally happen. For each extreme
Constraints models to answer what-if ques-
to dams. (but feasible) scenario selected by
tions to evaluate the resilience of
the planner the system must have
the national energy system. The
enough resources to ensure load
motivation was the 2010 earth-
supply for a given period even at
quakes with a magnitude of 8.8.
very high operating cost.
Design a module for the prepara- A user-friendly web-based module
EPE already presents maps with
Presentation of PSR produced maps with the re- tion of customized graphs based for the preparation of user-se-
19 expansion but no interface for ma-
results sults of planned expansion. crude output data produced by lected graphs with the results of
nipulation of output results
MDI and other models. his/her interest.
November 2019 41
20-26-00173
Processing No.: 15.2126.9-001.00
2.6.2 Power system studies

The integration in power systems of variable renewable energy (VRE) sources pose the following specific
challenges in comparison to the conventional synchronous machine-based generation: the intermittency and
variability of the VRE, their non- or poor dispatchability and the power conversion technology that is more
and more inverter-based (i.e. without physical inertia). These three peculiarities have a number of impacts
that are not anymore negligible when the level of penetration becomes significantly high in the entire system
or in a specific region.
Planning a power system (in short-, mid- and long-term horizons) for massive integration of VRE sources
requires in-depth analyses of the power system performance in both steady-state and dynamic conditions
in order to ensure system adequacy, security and quality of supply.
The proposed methodology for power system studies in view of system planning for VRE integration is com-
posed by 4 main blocks, as described in Figure 2-5. Not all elements of this methodology were applied in
the Product 3 of this project. Nevertheless, those steps that were not implemented in the project are still
included as recommended methodological improvements for future studies.
· Analysis of system inertia and primary frequency control (PFC) performance;
· Selection of relevant operating conditions for static and dynamic analyses;
· Static analyses;
· Dynamic analyses.
It must be emphasized that all the analyses within the framework of the power system studies take into
account the entire network and not only the main transmission system. In other words, network loading and
voltage/var control are analysed for both the main network as well as the sub-transmission and distribution
networks that are modelled in the database.
Details of the methodologies for each main block are given in the following subsections.
November 2019 42
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 2-5: Methodology for power system studies (overview)
November 2019 43
20-26-00173
Processing No.: 15.2126.9-001.00
It must be emphasized that all the analyses performed within the framework of the power system studies
comprise the entire network and not only the main transmission system. In other words, it means that net-
work loading and voltage/var control are analysed for both the main network as well as the sub-transmission
and distribution networks that are modelled in the database.
Details of the methodologies for each main block are given in the reports of products 3 and 4.
Table 2-10 presents a summary of the methodological recommendations for power system studies proposed
in this project. It must be emphasized that some of the recommended studies were not performed within the
framework of this project but are recommended to be carried out as an evolution of the planning methodology
employed in this project.
Table 2-10: Summary of the methodological recommendations for power system studies
Level of Importance for
Study Performed in P3?
Future Planning Studies
Selection of relevant operating conditions Yes +++
Analysis of system inertia Yes +++
DC load flow and DC static security assessment Yes ++
Operation optimization (OPF) Yes +++
Static security assessment Yes +++
Short-circuit current analysis Yes ++
System strength analysis No ++
Transient stability – DSA Yes +++
Transient stability – CCT No +
Frequency stability Yes +++
Small-signal stability Yes +
Transfer capacity limits No +++
Table 2-11 summarizes the methodological recommendations to be used in power system studies consid-
ering the rapid transformation of power systems due to factors such penetration of VRE, distributed energy
resources, and others.
November 2019 44
20-26-00173
Processing No.: 15.2126.9-001.00
Table 2-11: Power system studies - summary of methodological recommendations

(1) Integrate the input data for all

static simulations in a single input The only challenge is related with
parameter file (e.g. integrate IT. Ensuring compatibility of old
ANAREDE and ANAFAS data models is key.
files).
The CEPEL power system simula- The power system simulation soft-
(2) Implement automatic data con-
tion software package is currently ware package developed by
sistency and quality checks for Definition of the checks to be em-
employed in power system analy- Tractebel was employed in the
every simulation module. Two lev- ployed in line with the types of
sis. The following computation study. The following computation
els of checks should be imple- models in use in Brazil.
modules are used: modules have been used:
Models used in mented: warnings and errors.
planning study - Load flow and contingency analy- - Load flow and contingency analy-
1 sis; sis;
and general
methodology - Optimal power flow; - Optimal power flow; (3) Fully customizable OPF tool, IT-related challenges, as well as
- Short-circuit current calculation; - Short-circuit current calculation; allowing the user to select con- stronger requirements in terms of
- Electromechanical stability simu- - Extended term electromechanical straints and control variables per technical expertise by the user to
lation; stability simulation; equipment and/or groups of equip- correctly configure the optimization
ment based on pre-defined filters. problems.
- Small-signal stability analysis. - Small-signal stability analysis.
Complex IT challenges for the

(4) Integration of energy and
specification of the database,
power system analysis database,
which must be done in close col-
including model versioning control
laboration with the experts in en-
and model approval process.
ergy and power system analyses.
Revision of the equivalent models

Use of equivalent VRE power of existing VRE power plants to
Thorough review of the existing
plant models developed by the ensure that the power plant con-
models of wind and solar PV
planner based on the individual Use of standard models for wind trollers are modelled as well as
power plants, which might trigger
2 VRE modelling models of the power plants. Not and solar PV power plants based that the model represents correctly
further interactions with the power
every model includes the repre- on [1] and [2], respectively. the behaviour of the power plant at
plant owners in order to validate
sentation of the power plant con- the point of connection. For future
the models.
trollers. power plants, use the kind of mod-
els adopted in this study.
10
November 2019 45
20-26-00173
Processing No.: 15.2126.9-001.00

Perform a thorough review of the
turbine/governors of the existing
Models used for system expansion
units in order to ensure that the
planning are the same as the ones
Models of existing and already models reflect the reality of the
used by the system operator (vali-
planned units are the ones availa- units (e.g. Pmin, Pmax, valve
Dynamic model- dated models as per the grid Thorough review of the existing
ble in the ANATEM database from opening limits, rate limits, etc.).
ling of conven- code). dynamic models, which might trig-
3 PDE 2026. Generating units with For new units, employ standard
tional power For future power plants resulting ger model validation campaigns
missing models have been cor- models or models based on similar
plants from the expansion planning exer- with specific power plant owners.
rected by associating a standard units already existing in the data-
cise, not all new units have an as- model to each of the units. base (ensuring that the associated
sociated dynamic model (e.g. new
controls are properly tuned ac-
open and combined-cycle units).
cording to the parameters of the
units).
Modelled as constant current

Adoption of more detailed dynamic
Modelling of dis- sources for the aggregated unit at
4 Not modelled. models such as the one investi- Still an active topic in R&D.
tributed solar PV the distribution feeder connection
gated in [3].
point.
Extensive use of statistics,

heatmaps, duration curves, histo-
grams, etc.
Use of graph layout algorithms for Explore advanced power system
Requires performant computa-
quick visualization of simulation re- simulation results visualization
Duration curves, histograms, tional resources, as well user skills
Presentation of sults and analysis of network to- tools and follow-up research pro-
5 power flow system diagrams, and for the definition of the right visual-
results pology on the performance of the jects in the domain such the initia-
simulation results figures izations for the different types of
system. tives of NREL in advanced power
information being analysed.
Advanced dynamic simulation systems data visualization.
plots using the advantage of sav-
ing the time series of every state
variable of the dynamic model.
Strong IT challenge to implement
Extensive customization of the Customize the applications for the a Python API for every computa-
API-oriented Not available in current version of simulation tools via the use of the recurrent power system simulation tion module. Strong requirements
6
tools CEPEL software package. Python API’s available with the analyses performed by the planner for the users of the tools in terms
computation engines. in a day-to-day basis. of programming and scripting
skills.
November 2019 46
20-26-00173
Processing No.: 15.2126.9-001.00
Pre-defined operating conditions Employ data analytics-based

Selection of Op- Extreme operating conditions se-
based on: hydrological season methods to identify “likely” operat-
erating Condi- lected from the results of the simu-
7 (wet/dry), import/export level per ing conditions from large and Still an active topic in R&D.
tions for Power lations of the hourly energy opera-
subsystem, average VRE dis- dense datasets of generation dis-
System Studies tion.
patch. patch results.
November 2019 47
20-26-00173
Processing No.: 15.2126.9-001.00
2.6.3 Power system planning database

the system, etc.
The main recommendations for the development and implementation of a power system planning database
are given as follows:
· Common database for energy and power system simulation models;
· User access control to the database (read/write permissions, password control, modification track-
ing, etc.);
· Model validation and approval process;
· Common parameters between electrical and energy simulation models must be linked;
· Automatic check of mismatches between electric and energy simulation models;
· Allow data exchange between different computation modules (input data and simulation results);
· Model version control;
· Cloud-based.
November 2019 48
20-26-00173
Processing No.: 15.2126.9-001.00
3 MAIN REPORT
3.1 Introduction
The project “Energy systems of the future: Integrating variable renewable energy sources in Brazil's
energy matrix” aims at studying the impact of the integration renewable energy sources to the Brazilian
interconnected power system (BIPS) in both expansion and operation planning standpoints.
The general objective of the project, as specified in the ToR, is “to improve the prerequisites for systematic
integration of renewable energy and energy efficiency into the Brazilian Energy System”.
In this assignment, a pilot study on integrating renewable energies into the Brazilian energy system is per-
formed. It covers both operation and expansion planning aspects. More specifically, the objectives of the
project are:
· Perform an assessment of the current practices on VRE integration in Brazil;
· Perform an assessment of the international practices on VRE integration;
· Carry out a gap analysis between the international and the National practices in VRE integration;
· Carry out an expansion planning exercise composed by energetic and power system analyses con-
sidering power system operation aspects;
· Propose upgrades to the current practices in Brazil based on the results of the gap analysis and the
detailed energy and power system studies.
In order to accomplish the aforementioned objectives, the project is organized in eight products, as follows:
· Product 0: Work Methodology
o Outline of the detailed methodology proposed by the Consultant, including a detailed de-
scription of the work plan, as well as the strategy for cooperating and interacting with all
stakeholders.
· Product 1: Technical Regulation Studies
o Brief analysis of Brazilian and international technical regulations, particularly systems with
high VRE shares in the electricity matrix, diagnosing problems and indicate solutions al-
ready employed in Brazil and elsewhere in the world.
· Product 2: Energy Studies
o Detailed assessment of the suitability and/or the need to adapt the Brazilian power system
to cope with the expanding shares of VRE from the energy standpoint: in other words, con-
sidering the criteria, methodologies, models and data currently used for planning the expan-
sion and operations of Brazil’s power generation system.
· Product 3: Power System Studies
o Detailed assessment of the suitability and/or need to adapt the system to keep pace with
the expansion of VRE penetration in the BIPS, from the electrical studies standpoint: in other
words, considering the criteria, methodologies, models and data used to plan the expansion
and operations of Brazil’s transmission system. These studies must be compatible with
those of Product 2, addressing the same VRE penetration scenarios, in addition to intermit-
tence and complementarity among renewable sources of energy connected to the BIPS.
· Product 4: Methodology Studies
o On the basis of the findings of products 2 and 3, provide: recommendations for fine-tuning
the guidelines and criteria used in energy and electrical studies over short-term, mid-term
November 2019 49
20-26-00173
Processing No.: 15.2126.9-001.00
and long-term horizons; recommendations for modifications to the marginal costs calcula-
tion methodology; a database of input data to be considered for energy and electrical studies
in the mid-term and long-term horizons.
· Product 5: Technology Studies
o Studies exploring trends in the technological evolution of the VRE and their impacts on the
energy potential of these sources in Brazil. They must also define (or redefine) the technol-
ogy learning curve for these sources in terms of technologies and costs based on the evo-
lution of this technology at the international level.
· Product 6: Final Report
o Final report synthesizing the findings of the studies conducted for products 1 to 5 in a di-
dactic manner.
· Product 7: Workshops
o Streamlining project management through aligning team accomplishments and expecta-
tions, providing a platform for presenting and discussing study outcomes and findings.
The project structure is presented in Figure 3-1.
Figure 3-1 Product structure and flow
November 2019 50
20-26-00173
Processing No.: 15.2126.9-001.00
3.2 Product 1: Technical Regulation Studies

In this product the Consultant perform a brief analysis of Brazilian and international technical regulations in
order to diagnose problems and indicate solutions already employed in Brazil and elsewhere in the world.
This overview focuses exclusively on grid connection codes and on the provisions relevant to the connection
of wind and solar Photovoltaic (PV) generators of any size, considered as VRE generators. The content of
this subsection is based on [4].
3.2.1 Grid Codes: A General Overview

Grid codes are a legally binding instrument to set the rules for the power system and energy market opera-
tions, ensuring operational stability, security of supply and adequate functioning electricity markets. A grid
code is usually composed by a set of sub-codes, as depicted in Figure 3-2.
Grid Connection Interconnection

Operation Code Planning Code Market Code
Code Code
Synchrhonous
Operational Security Generation Planning
Generator Planning Code Market Rules Code
Code Code
Connection Code
Network Capacity
VRE Generator Operational Planning Network Planning Allocation and
Connections Code
Connection Code and Scheduling Code Code Congestion
Management Code
Load Frequency System

Demand Connection
Operations Code Control and Reserve Interconnection
Code
Code Code
Interchange
HVDC Connection Emergency
Scheduling and
Code Procedure Code
Balancing Codes
Operator Training
Data Exchange Code
Code
Metering Code
Operator Training
Code
Figure 3-2: General composition of a grid code (adapted from [4])
Technical requirements in grid codes are determined by the need to maintain the reliability, security and
quality of the power supply and fulfil the following objectives:
· The electrical power needs of all consumers must be met reliably;
· Voltage and frequency must be maintained within set limits to avoid damaging equipment connected to
the grid;
· The system must be able to recover quickly from system disturbances;
· At all times the system must operate without endangering the public or operating staff.
The sub-code that is the focus of VRE integration analyses is the Grid Connection Code11. The function of
a grid connection code applied to VRE is to provide a set of technical requirements for the connection of
11
In this project, the focus of the grid code analysis is on the Connection Codes and on the provisions related to the
connection of VRE generators (wind and solar PV).
November 2019 51
20-26-00173
Processing No.: 15.2126.9-001.00
wind and solar PV power plants to a given power system. This code is in charge of establishing a fair treat-
ment for all VRE power plants in which regards the technical requirements for the power plant to access the
grid. This helps to ensure the fair treatment of generator owners and operators concerning grid connection
while maintaining system stability and reliability. By providing appropriate technical and legal rules for VRE
generators, VRE connection codes can support the effectiveness of energy policies for VRE integration.
Policy-makers and regulators have to support the development and implementation of VRE grid connection
codes in association with the national renewable energy goals, by:
· Ensuring that grid connection codes include appropriate requirements for VRE;
· Consulting with all relevant stakeholders;
· Setting a predictable and reliable grid code revision process in place;
· Anticipating technical requirements based on future VRE targets;
· Learning from other countries experience;
· Joining regional initiatives to harmonize requirements and share resources.
3.2.2 VRE Connection Code Development and Implementation

When developing a VRE grid code, a code applicable to another area, country or region cannot be trans-
posed word for word since many requirements made for VRE generators depend on specific power system
needs. Figure 3-3 shows the aspects that need to be taken into account when developing a VRE connection
code.
Power System Interconnection

Voltage Levels
Size (Capacity) Level
Distribution and Characteristics of

Flexibility of Load Conventional Energy Policy
and Generation Generating Units
Expansion Market Size for Operational

Planning VRE Practices
Figure 3-3: Aspects to be considered when drafting a grid connection code (adapted from [4])
The process of developing and maintaining a grid code is outlined in Figure 3-4. A grid code must be sensi-
tive to the developing needs of the power system being regularly revised on the basis of feedback and
implementation and experience.
November 2019 52
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-4: Typical grid code development process [4]
3.2.3 Technical Requirements for VRE Integration

Requirements set out in VRE grid codes cover a variety of issues. However, the main driver for new require-
ments for VRE generators is the level instantaneous share of VRE with respect to the load during system
operation (VRE share). Table 3-1 presents the minimum set of technical requirements for VRE connection
according to the level of VRE share in a given system12.
Table 3-1: Overview of technical requirements for VRE connection according to VRE share (adapted from
[4])
VRE PENETRATION
TECHNICAL REQUIREMENTS
LEVEL
• Protection
• Power quality
Always needed
• Power reduction during over-frequency
• Commissioning and compliance tests
• Communication
Low VRE share • Adjustable reactive power
• Constraining active power (active power management)
• LVRT including current contribution

Higher VRE share
• Simulation models
12
It should be noticed that the classification according to VRE share are system-dependent. These values should be
seen from the perspective of the expected impact of VRE in a given system for different penetration levels. All listed
requirements remain necessary when the next VRE share level is reached.
November 2019 53
20-26-00173
Processing No.: 15.2126.9-001.00
VRE PENETRATION
LEVEL
• Voltage control
• Active power gradient limitation
Very high VRE share
• Reduced output duration mode for reserve provision
• Synthetic inertia
• Stand-alone frequency control

• Full integration into general frequency control scheme
Exclusive use of VRE
• Stand-alone voltage control
• Full integration into general voltage control scheme
3.2.4 International Grid Code Benchmarking

This state of the art compares the grid codes of several operators around the world in order to integrate
different points of view, considering Belgium, Canada, Denmark, Germany, Spain, The Philippines and
United Kingdom (UK).
These countries are chosen to cover:
· Different VRE technologies: solar (PV and CSP), wind, etc;
· Different types of power systems around the world: Europe, America, Asia;
· Significant VRE shares in their energy mix;
· Grids presenting similar characteristics than the Brazilian grid.
This review of the grid codes considers only the requirements on VRE power plants. It first focuses on the
classification applied to the different resources, which can be different for each system operator, as summa-
rized in Table 3-2. The main classification factors are usually:
· The primary source of energy for generation: thermal, wind, solar;
· The size of the generation plant: MVA, MW;
· The connection voltage of the power plant;
· The geographical location of the power plant and of its connection: onshore, offshore;
· The type of connection to the grid: AC line, AC cable, AC submarine cable, HVDC, VSC.
Table 3-2: Main characteristics of the systems for which the grid codes were evaluated13
LONG DIS-
EXPECTED EXPECTED EXPECTED
VRE SHARE TANCE
COUNTRY HVDC INTER- SHARE OF SHARE OF
SHARE OF NPP TRANSMIS-
CONNECTIONS WIND SOLAR
SION
United King-
dom
++ + ++ ++
Denmark +++ + +++
13
The legend of this table is the following. A “+” denotes a country that performs above average in the category of the
column, a “++” represent a country that for with the category is significant and a “+++” denotes a country which has to
modify greatly its power system development, planning and operation to allow its secure operation.
November 2019 54
20-26-00173
Processing No.: 15.2126.9-001.00
LONG DIS-
EXPECTED EXPECTED EXPECTED
VRE SHARE TANCE
COUNTRY HVDC INTER- SHARE OF SHARE OF
SHARE OF NPP TRANSMIS-
CONNECTIONS WIND SOLAR
SION
British Co-
lumbia (Can- + +++ + +
ada)
++
Belgium ++ ++ + +
(rooftop PV)
++
Philippines + ++ ++ (rooftop and
large PV)
+++
Germany ++ + + ++ (rooftop and
large PV)
++
Spain +++ + + +++ (utility-scale
CSP)
After this classification is assessed, the requirements for grid access and operation are investigated, for
each category. The issues examined are:
· Capabilities to stay connected: frequency and voltage tolerances, low voltage ride through, reactive
power capabilities;
· Power quality: phase unbalance, harmonics and flicker;
· Controls: active power/frequency regulation, voltage/reactive power regulation, excitation system, PSS
and damping;
· Protections and earthing philosophy;
· Islanding, synchronisation, black start capacities;
· Information exchange.
In the analysed grid codes, the following common features were observed:
· The main part of grid codes is dedicated to Synchronous Generating Units;
· Exceptions, derogations and specific requirements are required for VRE power plants, especially for
units based on power electronic conversion to produce electricity;
· The location-specific requirements are not included in the grid codes but are agreed between network
operator and plant owner in connection contracts;
· The requirements for less common technologies are usually treated in bilateral agreements and based
on the main grid code part dedicated to Synchronous Generation;
· In each of the grid codes, the categories of requirements are the same... Requirements are often clas-
sified in frequency, voltage, power quality, Q-Range, P/f control, Q/V control and information exchange.
Furthermore, the same requirement is often written following the same technical principle.
However, in these grid codes, the main differences are observed in:
· The type of generating units for which a dedicated grid code is available (conventional, wind, solar,
etc.). This mainly depends on the expected energy mix of the country;
· The thresholds/criteria for categorisation of the generating units (size, voltage level, etc.) highly depends
on the structure of the grid and energy mix;
November 2019 55
20-26-00173
Processing No.: 15.2126.9-001.00
· The numerical values used in requirements, such as values of Q-range, Voltage control droops, also
depend on grid structure;
· The approach chosen for modifying the grid code is different from country to country:
· Some chose one grid code for integrating every type of generation;
· Other prefer dedicated grid codes for Wind Farms, PV, etc.;
· Finally, other countries proposed amendments to existing grid codes, without modification to the
existing requirements.
3.2.5 Analysis of the Brazilian Grid Code

This subsection presents the analysis of the requirements for VRE connection as defined in the current
version of the grid code of Brazil (“Procedimentos de Rede”) [5]. The version of the grid code analysed in
this study is Rev. 2016.12, from 12 December 2016.
The Brazilian regulation for electrical power systems and services establishes the national grid code through
a set of documents entitled “Procedimentos de Rede”. The grid code of Brazil is organized in 26 different
modules. In this Product, the emphasis of the analyses is in the Module 3, which establishes the instructions
and processes for the feasibility of access (including the connection and the use) to the transmission facilities
that are part of the main transmission network.
The technical requirements for the connection of VRE generators to the transmission system are defined in
Sub-Module 3.6 of the Brazilian grid code.
3.2.5.1 Critical review of the Brazilian grid code

The review of the Brazilian grid code is based on a careful analysis of the current version of the Brazilian
grid code and the assessment of this document with respect to the international practice to date. It must be
emphasized that the conclusions and recommendations provided in this report take into account the speci-
ficities of the Brazilian power system.
The following topics were identified as points where an enhancement or fine tuning of the existing grid code
is needed:
· Requirements for WPP only in the current version of the grid code:
· Fast reactive current injection;
· Over-frequency control;
· Synthetic inertia.
· Reactive power compensation during hours with 0 MW power production;
· FRT characteristics;
· Performance of power plant control functions (V, Q, P, PF, etc.);
· Commissioning and compliance tests;
· Model development and validation;
· Contribution to protection duties;
· Requirements for energy storage devices (focus on BESS).
3.2.5.1.1 Requirements for WPP only in the current version of the grid code
Analysis:
November 2019 56
20-26-00173
Processing No.: 15.2126.9-001.00
In the current version of the grid code, there are a few technical requirements that are specified only for wind
power plants, as follows:
The first two requirements are in practice already being required from solar PV power plants. This seems to
be a text imprecision in the grid code that could be easily fixed.
In which regards the requirement for the provision of synthetic inertia, it is imposed exclusively to wind power
plants. It is important to highlight that if this requirement is imposed to solar PV power plants as well, it will
require the PV power plants to operate de-rated with respect to the maximum available power or to be
equipped with a local energy storage device.
Recommendations:
The following recommendations are made by the Consultant with respect to the identified issues:
· Sub-module 3.6, chapter 8.2.1, paragraph 9:
“Os aerogeradores As unidades geradoras de centrais geradoras eólicas ou fotovoltaicas com potência
instalada superior a 10 MW deverão dispor de controladores sensíveis às variações de frequência, que
promovam a redução da potência de saída quando em regime de sobrefrequência na faixa de frequên-
cias de 60,2 Hz a 62,5 Hz.
Este controle deverá ser do tipo proporcional com ganho de 3% / 0,1Hz na base da potência disponível
no aerogerador na unidade geradora no momento.”
· Sub-module 3.6, chapter 8.2.1, paragraph 8:
The technical requirements for the provision of synthetic inertia should remain being imposed to wind
power plants only.
· Sub-module 3.6, chapter 8.8:
“Quando de variações transitórias de tensão, além de cumprir os requisitos de manter-se conectadas
pelo período descrito no item 8.7.1 deste submódulo, os aerogeradores as unidades geradoras deve-
rão ser capazes de dar suporte de tensão à rede elétrica através da injeção de corrente reativa adici-
onal, para tensões de sequência positiva inferiores a 85%, e de absorção de corrente reativa adicional
de sequência positiva para tensões acima de 110%, conforme a Figura 6...”
3.2.5.1.2 Reactive power compensation during hours with 0 MW power production
Analysis:
The current version of the grid code requires that the VRE power plants be able to guarantee 0 Mvar injection
at the PoC when not producing active power. However, for certain regions of the system and for given
operating conditions, there could be an interest to use VRE power plants to provide voltage/reactive power
support when not producing.
In today’s situation, the current requirement is enough for the needs of the system. Moreover, as there is no
ancillary services market in place in Brazil, it is not straightforward to give incentives to VRE power plants to
provide voltage/reactive power support in periods of null production.
However, considering the foreseen growth of VRE in the Brazilian energy mix, it might be that in the future
such support to system duties becomes necessary from VRE power plants.
November 2019 57
20-26-00173
Processing No.: 15.2126.9-001.00
Recommendations:
· Maintain the requirement as it is in the current grid code;
· When dealing with higher VRE shares in the future, it is recommended to perform a dedicated analysis
in order to identify if the need for reactive compensation is a structural or a local issue:
· Structural issue: recommend grid reinforcement (e.g. synchronous condensers, SVCs, reactor/ca-
pacitor banks, etc.);
· Specific issue: provide specific requirements for the power plant to be connected at that specific
location (included in the PPA) or promote the development of a competitive ancillary services mar-
ket.
3.2.5.1.3 FRT characteristic
Analysis:
In the current version of the Brazilian grid code, the FRT requirements are defined at the terminals of the
generating unit (e.g. terminals of the wind turbine or the inverters). In addition to that, a single FRT charac-
teristic is defined for the entire system, independent of the location and/or voltage level.
The main reasons for defining a single FRT characteristic for the entire system are:
· Non-discriminatory nature of the grid code (to not favour specific locations of the system);
· The transient behaviour of the voltage in the different subsystems in today’s grid configuration is very
similar, allowing to set a single FRT characteristic for the entire system.
Recommendations:
· Specify the FRT requirements at the VRE power plant PoC;
· Assess the transient behaviour of the system voltages in every medium-term transmission planning
study in order to anticipate future needs of the system. In case the behaviour of the different regions of
the system becomes significantly different, it might be needed to specify a dedicated FRT characteristic
for each region of the grid.
3.2.5.1.4 Performance of power plant control functions (V, Q, P, PF, etc.)
Analysis:
The requirements for performance of VRE power plant control functions must be clearly defined in the grid
code. These requirements have a significant impact to the design of the power plant and to the selection of
the equipment that compose this plant.
The current version of the Brazilian grid code provides clear requirements for the types of control functions
that need to be implemented. However, it does not explicitly define the performance requirements for the
different control functions, such as:
· Maximum setpoint update delay;
· Response time to a new setpoint;
· Setpoint accuracy;
· Accuracy of the control action;
· Open and closed loop response times;
· Maximum setpoint tracking error.
November 2019 58
20-26-00173
Processing No.: 15.2126.9-001.00
As these requirements have an impact on the design and equipment choice (CAPEX and OPEX impacts),
they must be clearly known in the very early stages of the project development in order to not jeopardize the
feasibility of a project.
Recommendations:
· Define a minimum set of performance indicators and the associated numerical values
· Maximum setpoint update delay;
· Response time to a new setpoint;
· Setpoint accuracy;
· Accuracy of the control action;
· Maximum setpoint tracking error;
· Open and closed loop response time.
The numerical values for the requirements depend on the intrinsic behaviour of the system and on the ca-
pabilities of the VRE generation technologies. Therefore, these values should be set based on technical
studies of the dynamic behaviour of the system, as well as taking into account the existing and forecasted
capabilities of the VRE generation technologies.
3.2.5.1.5 Commissioning & compliance tests
Analysis:
The current version of the Brazilian grid code does not provide specifications for commissioning and com-
pliance tests for VRE power plants. However, these requirements are defined in dedicated documents made
available to power plant developers during the execution of the projects.
The following topics are covered by these documents:
· Testing conditions are specified for the following aspects:
· Active power control;
· Frequency control;
· Reactive power control at the PoC;
· Voltage control at the PoC;
· Power factor control at the PoC;
· Transformer energizing;
· Power quality.
· Variables to be measured and/or monitored.
Recommendations:
· Make the requirements for commissioning and compliance tests available to the project developers in
a very early stage of the project or (preferably) publicly available as an annex of the grid code;
· Details of the following aspects, for each test to be performed, should be specified:
· Purpose of the test;
· Pass criteria (in line with the minimum requirements defined in the grid code);
· Instrumentation and data recording;
November 2019 59
20-26-00173
Processing No.: 15.2126.9-001.00
· Initial conditions for the test;

· Step-by-step test procedures;
· Test reporting procedures.
3.2.5.1.6 Model development and validation
Analysis:
The Brazilian grid code specifies the following responsibilities for the power plant owner in which regards
the provision of simulation models of the VRE power plant (Sub-Module 21.4):
· Provide to ONS the data and information necessary to carry out the data and model validation process
for electrical studies;
· Provide the necessary data and participate in model validation activities;
· Update the computational models used for electrical studies according to the adjustments made during
project commissioning;
· Present simulation results using the updated computational models and the records obtained from the
commissioning tests, considering the adjustments made in the field during the commissioning, in order
to attest the performance of the models.
Despite the provision in the grid code, details on the requirements for these models are not defined. The
following relevant topics are not clearly defined in the grid code:
· Model extent: a trade-off between model complexity and model quality has to be achieved
· Application for power plant design purposes: detailed power plant model and simplified grid model;
· Application for system-level analyses: detailed grid model and simplified power plant model.
· Responsibility for model reduction/simplification:
· Should be in the hands of the system operator given that it requires deep understanding of the
modelled process and the application of the models.
· Simulations to be carried out in the model validation process in line with the commissioning and com-
pliance tests:
· Same initial conditions
· Same pass criteria
Recommendations:
· Power plant owners/project developers should be in charge of providing a full model (including internal
MV network and detailed model of the generating units) of the power plant to the system operator;
· The performance of the model should be assessed via simulations aiming at reproducing the commis-
sioning and compliance tests in a simulated environment;
· The model validation should be subject to expert judgement by the system operator;
· The model reduction/simplification should be a responsibility of the system operator.
Figure 3-5 presents a simplified model management process recommended by the Consultant.
November 2019 60
20-26-00173
Processing No.: 15.2126.9-001.00
•Measurements
(commissioning •Development of
PP •Measurements
tests) a reduced-order
•Models PP System
Owner / •Detailed model (equivalent)
(equipment Owner (used for Operator model
OEM level)
compliance
studies)
Figure 3-5: Recommended model management process
3.2.5.1.7 Injection of asymmetrical currents during faults
Analysis:
Asymmetrical faults are the most common types of faults occurring in transmission networks. However, the
majority of the VRE grid codes do not specify technical requirements for the behaviour of the power plants
under asymmetrical faults.
With increasing VRE shares in the system the importance of VRE power plant behaviour during asymmet-
rical faults is increasing. The Brazilian grid code does not specify the technical requirements for negative
sequence current injection for VRE power plants during asymmetrical faults.
Recommendations:
It is recommended that in in every medium-term transmission planning study the behaviour of the system
with respect to asymmetrical faults be analysed in detail. Specific requirements in the grid code should be
defined as soon as it is detected problems in terms of dynamic performance of the system or malfunctioning
of protection systems under unbalanced faults (e.g. the case of the German grid code).
3.2.5.1.8 Requirements for energy storage devices (focus on BESS)
Analysis:
The majority of the grid codes around the world do not currently define Energy Storage or specify technical
requirements for Storage technologies (Pump Storage Plants aside). However, with the expected growth in
the application of BESS in power systems, there are several ongoing discussions around this topic in differ-
ent countries.
Recommendations:
Considering that the technical requirements for energy storage applications have not yet reached sufficient
maturity level, it is recommended to not define such requirements in the Brazilian grid code in the short-term.
However, it is recommended that ONS follows the international discussions in this subject in order to capi-
talize on the international experience when defining the technical requirements for energy storage to the
Brazilian case.
November 2019 61
20-26-00173
Processing No.: 15.2126.9-001.00
3.2.5.1.9 Synthesis of the analysis

The analysis performed by the Consultant shows that the current version of the Brazilian grid code for the
connection of VRE generation covers all key aspects and is in line with international grid codes practice.
Table 3-3 presents a synthesis of the benchmark of the Brazilian grid code with respect to the international
practice.
Table 3-3: Benchmark of the Brazilian grid code with respect to the international practice (synthesis of
analyses)
VRE PENETRATION BRAZILIAN

LEVEL GRID CODE
• Protection J
• Power quality J
Always needed
• Power reduction during over-frequency J
• Commissioning and compliance tests K
• Communication J
Low VRE share • Adjustable reactive power J
• Constraining active power (active power management) J
• LVRT including current contribution J

Higher VRE share
• Simulation models K
• Voltage control J
• Active power gradient limitation K
Very high VRE share
• Reduced output duration mode for reserve provision L
• Synthetic inertia J
• Stand-alone frequency control N/A

Exclusive use of • Full integration into general frequency control scheme N/A
VRE14 • Stand-alone voltage control N/A
• Full integration into general voltage control scheme N/A
3.2.6 Conclusions and recommendations

Product 1 focuses on the role played by grid codes on the integration of VRE power plants to power systems.
Because of this key role on VRE integration, the Brazilian grid code was reviewed in detail in order to assess
its level of maturity having in mind the main objective of this project: analyse the integration of large shares
of VRE to the Brazilian energy mix.
First, this report presents a general overview of grid codes and its role for VRE integration. It shows how
important grid codes are for enabling the integration of large amounts of VRE to a power system without
jeopardizing the quality and security of the service.
Then, an overview of the international practice in grid code development and implementation is presented.
It stresses the importance of the process when defining technical requirements for the connection of VRE
power plants to the grid in order to not create an unstable regulatory framework or to discourage the devel-
opment of new technologies and solutions.
14
Requirements for cases of exclusive use of VRE are not applicable (N/A) to the Brazilian system reality.
November 2019 62
20-26-00173
Processing No.: 15.2126.9-001.00
This report also describes in general words the common technical requirements for VRE integration as per
the international practice. It also presents the general trends for more sophisticated requirements resulting
from higher shares of VRE integration (e.g. synthetic inertia, contribution to unbalanced faults).
The benchmark of the international practice in grid codes for VRE connection to transmission networks is
presented. This analysis was done based on the grid codes from Belgium, Canada (British Columbia), Den-
mark, Germany, Philippines, and UK. In the analysed grid codes, the following common features were ob-
served:
· The main part of grid codes is dedicated to synchronous generating units;
· Exceptions, derogations and specific requirements are required for VRE power plants, especially for
units based on power electronic conversion to produce electricity;
· The location-specific requirements are not included in the grid codes but are agreed between network
operator and plant owner in connection contracts;
· The requirements for less common technologies are usually treated in bilateral agreements and based
on the main grid code part dedicated to synchronous generation;
· In each of the grid codes, the categories of requirements are the same. Requirements are often classi-
fied in frequency, voltage, power quality, Q-Range, P/f control, Q/V control, information ex-change, etc.
Furthermore, the same requirement is often written following the same technical principle.
The last subsection is dedicated to a thorough review of the Brazilian grid code for VRE connection. This
analysis shows that the current version of the Brazilian grid code for the connection of VRE generation
covers all key aspects and is in line with international grid codes practice. A synthesis of the benchmark of
the Brazilian grid code with respect to the international practice is presented. For the majority of the topics,
the grid code is in line with the international practice. For a few of them, there is room for improvement, such
as:
· Requirements for WPP only in the current version of the grid code:
· Reactive power compensation during hours with 0 MW power production;
· FRT characteristics;
· Performance of power plant control functions (V, Q, P, PF, etc.);
· Commissioning & compliance tests;
· Model development and validation;
· Contribution to protection duties;
· Requirements for energy storage devices (focus on BESS).
It must be highlighted that the grid code should be permanently reviewed in order to enable the deployment
of VRE without jeopardizing the quality and security of the system operation. Careful should be taken in the
grid code update process in order to avoid regulatory instabilities.
November 2019 63
20-26-00173
Processing No.: 15.2126.9-001.00
3.3 Product 2: Energy Studies

The objective of Product 2 is to present expansion planning methodologies and best practices related to the
integration of VRE sources in Brazil’s energy matrix. One of the key objectives of this study is to provide
insights to the planning and operation of the system by an exercise of planning for a sufficiently distant future
of the power system in Brazil, with a VRE market share in the power matrix considerably higher than the
present.
Secondary objectives of this study include the following methodological issues:
· Evolution of VRE technologies and how they affect the operation of the power system;
· Modelling of VRE scenarios, including seasonality, short and long-term variability and correlations;
· Modelling of reserve requirements;
· Transmission network reinforcements required to accommodate VRE growth;
· Integration of energy and electrical planning studies;
· Modelling of distributed generation in the low voltage grid.
The methodological steps and corresponding computational tools proposed in this study are depicted in
Figure 3-6.
Figure 3-6: Overview of Product 2 methodology
· Step 1 – Design of VRE candidate projects: identification of candidate projects for wind and solar gen-
eration;
· Step 2 – Stochastic modelling of renewables and inflows: calculation of the expected operational cost
over a set of inflow/renewable scenarios;
· Step 3 – Co-optimize new electricity and gas investments: determination of the set of generation and
transmission reinforcements along the planning period that minimizes the present value of investment
costs plus the expected value of operation costs;
· Step 4 – Optimal stochastic operation: results from the iterative solution of two modules: an investment
module (OptGen) and an operation module (SDDP);
November 2019 64
20-26-00173
Processing No.: 15.2126.9-001.00
· Step 5 – Optimal network planning under uncertainty: NetPlan uses an AC power flow model which
determines the optimal transmission reinforcements in each region.
3.3.1 Market size

The study consists in a prospective analysis of the Brazilian power system considering a massive insertion
of VRE in a future configuration. The departing point is the system configuration defined by the Ten-Year
Brazilian Expansion Plan 2017-2026 prepared by EPE (PDE 2026). All new generation and transmission
capacity planned up to this horizon are assumed to be in full operation. The end of horizon yearly demand
is assumed as twice the demand observed in 2017 (~600 TWh), thus 1200 TWh.
One should note that the study does not aim to carry out an energy planning, for the short, medium or long
term, as this is a mission for EPE. It is necessary, however, to have a significant load increase to be supplied
by an increased participation of VRE to provide insights for the planning and operation of the Brazilian power
system.
3.3.2 Candidate Projects for Capacity Expansion

Candidate projects for capacity expansion include coal, open cycle and combined cycle natural gas, nuclear,
biomass cogeneration, wind farms, solar PV farms and storage devices. Hydropower is assumed to expand
as in PDE 2026 but not after that. Table 3-4 present the technical and economic parameters of the candi-
dates considered in the expansion planning process.
Table 3-4: Economic Parameters of the Candidate Projects

START-UP
FUEL VARI- COST
INVEST- FIXED O&M VARIABLE
ABLE VARIABLE [HOUR OF
MENT COST PRODUC- LIFETIME
COST O&M COST FUEL CON-
COST [USD/KW TION COST [YEARS]
[USD/MMB [USD/MWH] SUMPTION
[USD/KW] PER YEAR] [USD/MWH]
TU] AT RATED
CAPACITY]
Open cycle 582.00 10.45 10.50 4.00 104.00 0.50 25
gas-fired
CCPP - in- 800.00 5.00 10.50 2.75 42.00 1.50 25
flexible
CCPP - 800.00 7.00 10.50 2.75 57.00 1.50 25
semi-flexi-
ble
CCPP - 800.00 9.00 10.50 2.75 71.00 1.50 25
flexible
Coal-fired 2120.00 33.00 15.00 2.00 - 8.00 25
Nuclear 5000.00 10.00 55.00 - - - 40
Sugarcane 1550.00 0.00 26.00 - 0.00 - 25
biomass
Wind 850.00 - 26.00 - - - 25
Utility- 636.00 - 8.00 - - - 25
scale solar
PV
The PDE 2026 decommissioning plan for the existing generation is also assumed in this study. Therefore,
2000 MW of fuel oil and 900 MW of diesel will be decommissioned until 2026, according to the end of their
contracts. This study also considered the decommissioning of additional 762 MW of fuel oil and of 25 MW
of diesel after 2026.
November 2019 65
20-26-00173
Processing No.: 15.2126.9-001.00
3.3.3 Other Relevant Assumptions
3.3.3.1 Regional grid modelling

For the portfolio selection purposes, the grid is modelled considering seven regions 15 as shown in Figure
3-7. Since the objective at this stage is to model the trade-offs related to where the projects should be built,
only the main transmission limits are represented. In this sense, the simulations considered the regional
transfer capacities of the 2026 configuration of PDE with an addition of 9 GW between the Northeast and
the Southeast and a few changes.
The definition of each subsystem is given as follows:
· Acre and Rondônia (AR) and Madeira (MAD);
· Central-West (CW);
· Itaipu (IT);
· North (N);
· Northeast (NE);
· South (S);
· Southeast (SE).
Figure 3-7: Regional grid modelling
3.3.3.2 Distributed Generation

In this study, rooftop solar photovoltaic is explicitly considered as a source of Distributed Generation (DG).
Biomass cogeneration and small hydropower connected to the distribution concession are included indirectly
through the load profile, considered by ONS on monthly operation planning. The total DG capacity forecast
provided by EPE for the end of horizon is 30 GW. As a result, the 30 GW capacity forecast has a corre-
sponding average production of 4.8 GW, which amounts to 3.5% of the total consumption, considering the
capacity factor of 16%, also provided by EPE.
15
PDE uses 14 regions.
November 2019 66
20-26-00173
Processing No.: 15.2126.9-001.00
3.3.3.3 Automatic generation control

In Brazil, the automatic generation control (AGC) considers only hydropower plants from 400 MW. The fol-
lowing assumptions were assumed for the AGC requirement:
· Primary and secondary reserves shall be respected within each control area.
· New AGC participants include all hydro plants with capacity greater than 400 MW (as in the grid code)
and with a water head higher than Estreito hydropower plant (18.94m). Differently from the current grid
code, this study assumed that natural gas fired plants are also included in the AGC.
· The allocated reserve must be less than 40% of the available capacity of each hydro power plant.
· Allocated reserves must meet, in each hour, the sum of the reserve associated with demand variability
(1% primary + 4% secondary) plus the probabilistic dynamic reserve, related to uncontrolled VRE.
3.3.3.4 Minimum water outflow for Sobradinho

The minimum outflow downstream of the Sobradinho reservoir, in the São Francisco river, considered in the
simulations, is 650 m3/s. This outflow is lower than the current constraint defined by ANA/ONS, although it
is higher than the value that was used during several months of a drought in the São Francisco river without
any consequences to other uses of water. However, given the small share of hydropower in the end-of-
horizon, impacts of this alteration in the results are minor.
3.3.3.5 VRE Modelling

There are several aspects regarding the modelling of renewables in expansion and operation planning stud-
ies. As real-time production data is incomplete and there are confidentiality issues, this study uses reanalysis
data to generate scenarios of electricity production at hourly steps. Fortunately, there are monthly production
values for wind power plants that can be used to calibrate the model based on reanalysis data such that the
actual mean production is reproduced by the model, thus avoiding any bias.
Additionally, statistical properties of the hourly time series, such as spatial correlation and correlations in
different time scales, must be reproduced by the model. Seasonal behaviour of wind and solar power must
also be well represented in addition to the daily wind and solar power profile for different times of the year.
Finally, correlations between VRE and hydropower plants can be represented in the synthetic scenarios.
3.3.4 Capacity Expansion
3.3.4.1 Methodology
The process flow of the energy planning study is illustrated in Figure 3-8. The project starts with the prepa-
ration of candidate projects for renewables from resources such as solar radiation, temperature, wind veloc-
ity and transmission network maps. The Time Series Lab is used to produce scenarios of VRE correlated to
water inflows to the hydro plants.
Given these inputs the first model executed is OptGen, which determines the expansion of generation ca-
pacity and the expansion of regional interchange capacities. The next step consists on the simulation of the
operation of the Brazilian interconnected system for this portfolio. The initial objective of this SDDP execution
is verify if the system operation is reliable for all scenarios of hydrological flows, VRE production and sea-
sonal demand. If there are occasional energy deficits or marginal costs that are too high, it is necessary to
return to OptGen to reinforce the generation expansion. However, if the operation reveals a “healthy” oper-
ation then it is possible to move on to the next step which consists in reinforcing the transmission system.
After data preparation the transmission planning study includes two other steps before optimizing transmis-
sion reinforcement requirements. Initially, SDDP is executed considering a full representation of the high
voltage transmission network without monitoring the circuit flows. The second step consists of estimating
November 2019 67
20-26-00173
Processing No.: 15.2126.9-001.00
the investment cost of each candidate circuit or transformer from its technical parameters. After that, OptNet
is executed to find the list of necessary transmission reinforcements at least cost.
The next process consists in running a detailed model of the SIN operation using hourly times steps. SDDP
optimizes the grid-constrained operation of the power system, determines hourly circuit flows, calculates the
corresponding losses, adding half of each circuit loss between the initial and final bus bars and reiterates
until convergence.
Figure 3-8: Diagram of the process flow of computational model
3.3.4.2 Capacity expansion and operation results

The capacity expansion at the end of the horizon is presented in Figure 3-9 for a hierarchical case and a co-
optimization case, with capacity and reserves decided jointly in the same optimization model. As shown, the
leading technology in terms of capacity addition is wind power, which exceeds 41 GW in both runs, mostly
installed in the Northeast region, with smaller amount in the South-region. Solar photovoltaic expansion
amounts to around 20 GW in both cases. In addition to the centralized solar, the model considers 30 GW of
solar distributed generation (DG) by the end of the study horizon. A little over 2 GW of battery capacity was
selected for battery storage due to reserve constraints in the Northeast.
50
45
40
35
30
25
GW
20
15
10
5
0
Wind Solar PV Interconection OCGT CCGT - Model 1 Li-Ion BES
Hierarchical 41.2 20.1 9.0 2.2 8.0 2.3
Co-optimization 41.7 19.6 9.0 1.4 8.0 2.4
Figure 3-9: Capacity additions per source from 2026 to end of horizon
November 2019 68
20-26-00173
Processing No.: 15.2126.9-001.00
The allocation of the new capacity at the end of the horizon, in each Brazilian region, is presented in Figure
3-10 for the co-optimization run of OptGen.
Wind Solar PV OCGT CCGT Li-Ion BES
CW 0.4
S 8.0 1.4
0.4
SE 12.6 8.0
NE 33.3 5.6 2.4
N 1.0
0 5 10 15 20 25 30 35 40 45
Figure 3-10: Capacity additions per source and region (GW)
The evolution of the installed capacity over the time, considering the co-optimization run reference case, for
2016, PDE 2026 and the end of horizon in this study, is shown in
Figure 3-11 and Figure 3-12. In terms of capacity, hydropower currently accounts for 69% of total installed
capacity. It decreases to 56% by 2026 and to 38% by the end of the horizon. On the other hand, wind power
increases from 6.8% to 22.1% and solar power from nearly zero to 19% (centralized, utility scale + Distrib-
uted Generation). These results indicated a substantial growth of the participation of the VRE in the Brazilian
power matrix for economic and technical reasons, with no subsidies. Hydropower reduces market share but
is the key source that supports the VRE growth given the flexibility to accommodate the dispatch.
End of Horizon
Target-Year
PDE 2026
2016
0 50 100 150 200 250 300
Installed Capacity [GW]
Hydro Biomass Wind Solar Nuclear Gas Fuel Oil Coal Solar DG Battery
Figure 3-11: The evolution of the Installed capacity [GW]
November 2019 69
20-26-00173
Processing No.: 15.2126.9-001.00
End of Horizon
Target-Year
PDE 2026
2016
0% 20% 40% 60% 80% 100%
% of Total Installed Capacity
Hydro Biomass Wind Solar Nuclear Gas Fuel Oil Coal Solar DG Battery
Figure 3-12: Comparative % of total installed capacity
Figure 3-13 presents the total capacity evolution percentage from PDE 2016 to the end of horizon in this
study, considering the hydro, wind, biomass, solar and solar DG power plants.
Hydro 56.5%
38.3%
22.1%
Wind 13.5% 9.6%

Biomass 8.6% 9.4%
Solar 4.6% 5.9%
Solar DG 0.0%
PDE End of
2026 Horizon
Figure 3-13: The evolution of % of total installed capacity
Since renewables contribute with 90% of the total energy production, the associated CO2 emission related
to fossil fuel production with respect to the total consumption is very low. In fact, emissions per unit of elec-
tricity consumption, for the co-optimization run of the reference case, is 51 gCO2 per kWh, which is roughly
one tenth of the world average for the electricity sector based on EIA online database [6], as shown in Figure
3-14.
November 2019 70
20-26-00173
Processing No.: 15.2126.9-001.00
India 900
China 800
UK 500
USA 500
Brazil 51
0 200 400 600 800 1000

gCO2 per kWh emission
Figure 3-14: International comparison of unit emission per energy consumption
3.3.4.3 Regional energy transfers

The main energy transfers (average GW) are shown in Figure 3-15. Northeast is the main exporter while
the Southeast region is the main importer.
Figure 3-15: Main energy transfers
3.3.4.4 Sensitivity Cases

The objective of this session is to analyse how the capacity expansion is affected by changes in certain key
factors and assumptions that are intrinsically uncertain, such as the cost of VRE in the future.
3.3.4.4.1 Sensitivity 1 - VRE investment costs

The motivation of this initial sensitivity is to evaluate what happens to the portfolio of projects if the decrease
of VRE investments costs is lower or higher than the reference case assumption. To answer this question
additional OptGen runs were made, as shown in Table 3-5. Figure 3-16 presents the results of this exercise.
Table 3-5: Sensitivity 1 - main assumptions

CASE WIND SOLAR PV
Reference case 30% 40%
Lower VRE cost decrease 20% 30%
Higher VRE cost decrease 40% 50%
November 2019 71
20-26-00173
Processing No.: 15.2126.9-001.00
50
40
30
20
GW
10
0
S1-Low 40.9 16.2 1.0 9.0 2.2
Reference 41.7 19.6 1.4 8.0 2.4
S1-High 42.8 21.8 2.2 7.0 2.2
Figure 3-16: Sensitivity 1 - results for different projected VRE cost decreases
As seen, there is a small capacity change in the of wind power addition with respect to the reference case
for the lower and higher VRE cost decrease, respectively (-0.8 GW, +1.1 GW) and a higher variation for
solar PV (-3.4 GW, +2.2 GW). Associate impacts to reserves have OCGT respond in the same direction as
the VRE (-0.4 GW, +1.2 GW), whereas CCGT responds in the opposite because the total energy consump-
tion is the same (+1.0 GW, -1.0 GW). This is because as VRE cost is reduced, more is selected by OptGen.
More VRE energy implies in less requirements by inflexible CCGT. Additional reserve required by the system
is then economically provided by an increase of open cycle gas turbine (OCGT) in the portfolio.
3.3.4.4.2 Sensitivity 2 - Open cycle generation

OCGT plants were considered with a minimum technical generation equal to 25% of the nominal capacity.
It was requested a sensitivity analysis to evaluate the impact if these plants were considered fully flexible
(no minimum generation if committed). The results of this run did not show differences with respect to the
reference case.
3.3.4.4.3 Sensitivity 3 - Minimum storage for Li-Ion BESS

The minimum storage for Lithium-Ion Batteries was considered as zero in the reference case, which is not
recommendable by many manufacturers considering today’s technologies. In this context, the main objective
of this sensitive case was to evaluate the use of a minimum storage for Lithium-Ion Batteries of 20% of the
maximum capacity. It is worth noting that the unit cost (USD /kW) in the end-of-horizon assumes a strong
reduction with respect to current cost. By reducing the kWh usable amount and maintaining the CAPEX
(USD), in practice this sensitivity basically increases the USD /kWh value. In this case the results were only
marginally impacted: there was a decrease of 0.6 GW of BESS construction with respect to the reference
case (from 2.4 to 1.8 GW) due to lower competitiveness.
3.3.4.5 Sensitivity 4 - VRE uncorrelated with hydrology

3.3.4.5.1 Case A
Evaluate impacts to the generation portfolio selection if scenarios of VRE production are uncorrelated with
inflows to the hydro power plants. For this sensitivity, the monthly Bayesian network of TSL was prepared
from the MERRA2 database only (i.e. inflows were not inputted to the Bayesian network).
The elimination of the monthly correlations between wind farms production and hydropower plants reduces
the portfolio effect between these sources because, in general, correlations between wind and hydro power
are negative. OptGen selected a portfolio with more wind power (+7.3 GW) and less solar PV (-2.7 GW)
because additional energy is necessary in the dry hydrology scenarios. Interestingly biomass and coal
November 2019 72
20-26-00173
Processing No.: 15.2126.9-001.00
power, that had not been selected in the reference case, entered the portfolio with +7.1 GW and 1.5 GW,
respectively. On the other hand, OCGT increased participation (+1.6 GW) due to a higher variability of the
wind production. The amount of CCGT was reduced (-6 GW) while Li-Ion additions were discarded alto-
gether (-2.4 GW).
3.3.4.5.2 Case B
In this case it is examine what happens if biomass and coal projects are no longer considered as candidates
projects in OptGen. The results of this case are similar to the reference case, despite an overall increase in
investments in the portfolio. Figure 3-17: Sensitivity 4 - VRE scenarios uncorrelated Figure 3-17 compares
results for Cases A and Case B with the reference case.
60
50
40
30
20
GW
10
0
CCGT - Li-Ion
Wind Solar PV Biomass Coal OCGT
Model 1 BES
Reference 41.7 19.6 0.0 0.0 1.4 8.0 2.4
Case A 49.0 16.9 7.1 1.5 3.0 2.0 0.0
Case B 47.8 18.3 0.0 0.0 2.0 8.0 0.0
Figure 3-17: Sensitivity 4 - VRE scenarios uncorrelated with inflows
3.3.4.5.3 Case C
During the discussions of the results of this case it was suggested that the maximum constructed wind power
of the South system should be relaxed as there is enough space for new additions. It was also suggested
that the natural gas for thermal power generation should not be considered in the North region and were for
this reason withdrawn. In this case OptGen reacted by selecting a significantly higher amount of wind power
- with the increase concentrated in the South region - while reducing CCGT from 8 GW to zero and solar PV
from 20 to 15 GW (-5 GW). Biomass was now selected with 2 GW. OCGT and Li-Ion suffered slight changes.
Figure 3-24 shows the results for this case.
60
50
40
30
20
GW
10
0
CCGT -
Wind Solar PV Biomass OCGT Li-Ion BES
Model 1
Reference 41.7 19.6 0.0 1.4 8.0 2.4
Case C 62.8 15.2 2.0 1.2 0.0 2.0
November 2019 73
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-18: VRE uncorrelated with inflows and illimited wind developments in the South
This result indicates the need to use better criteria to limit the capacity of candidate projects that can be
selected in each region. A process should ideally combine layers of relevant information, such as the wind
speed, available area, distance to roads, terrain slope, existence of conservation and indigenous areas, and
others, to perform a “screening” with GIS related tool of the potential that could be explored in each cluster.
With such process, the preparation of candidates for the system expansion is improved, thus limiting the
available resources, including those with high capacity factors.
3.3.4.5.4 Sensitivity 5 - VRE forecasting error

In this sensitivity we examine the effect of improving VRE forecasting ability, which would reduce forecasting
error by half with respect to the reference case, with in a Mean Absolute Percentage Error (MAPE) of 10%
in the Northeast region. OptGen in this case chose almost the same portfolio of projects, except for 2.4 GW
of Li-Ion battery, which were reduced to 0.4 GW, as presented in Figure 3-19.
50
40
30
20
GW
10
0
Intercone
ction
Reference 41.7 19.6 9.0 1.4 8.0 2.4
Better forecast 41.5 19.6 9.0 1.4 8.0 0.4
Figure 3-19: Sensitivity 6: improved forecasting of the VRE production
3.3.4.5.5 Sensitivity 6 - Capacity expansion timing

After validating the results of the end of horizon Reference Case, an additional exercise was made to eval-
uate when these investments would be made. Two intermediate points (P1 and P2), were used between
2026 (the last year of the PDE) and the end of horizon, when demand is twice that of 2017. The first (P1)
has 1/3 of the demand growth of the complete period and the second (P2) with 2/3 of the demand growth,
as depicted in Figure 3-20.
2026 (last year of PDE) P1 P2 End of Horizon

Figure 3-20: Overview of the cut years for the planning exercise
The following procedure was used:

· Step 1: Preparation of yearly VRE and storage CAPEX estimates based on assumptions of the refer-
ence case and the expected cost reduction applied for P1 and P2;
· Step 2: Preparation of DG penetration provided by EPE for these two intermediate years.
November 2019 74
20-26-00173
Processing No.: 15.2126.9-001.00
Period PDE 2026 P1 P2 End of Horizon

DG Capacity (GW) 9 11 22 30
· Step 3: Each case is simulated individually considering all the capacity built in the previous cases as
given.
OptGen was executed for these two intermediate years for the reference case. Results of capacity expan-
sions for the intermediate years are shown in Figure 3-25. Table 3-6 shows results per region, in installed
GW.
Figure 3-21: Timing of investment decisions (GW)
Table 3-6: Sensitivity 6 – results per region (GW)

PERIOD PERIOD PERIOD
REGION SOURCE
2026->1/3 1/3-2/3 2/3-END
CCGT 8.0
SE Solar PV 7.2 5.4
Wind 0.4
Wind 14.9 9.2 9.2
NE BESS Li-Ion 2.4
Solar PV 5.6
CW Solar PV 0.4
N Solar PV 1.0
OCGT 1.4
S
Wind 3.3 4.7
As the Table 3-6 shows:

· Wind power is the technology of choice since the beginning of the horizon in the Northeast and South.
It is competitive to utilize the high capacity-factor clusters to expand this power production with the use
of transmission capacity.
· Solar PV is selected in the years in the middle of the horizon in the Southeast and Centre-West. Despite
a slightly lower capacity factor in these regions with respect to the best areas of the Northeast, this
selection has the advantage of not requiring investments in new energy transfer capacities. In the final
November 2019 75
20-26-00173
Processing No.: 15.2126.9-001.00
period, as electricity demand increases, including in the Northeast, it becomes possible once again to
install new solar projects in this region.
· CCGT plants are selected in the end-of-horizon in the Southeast, and Lithium-Ion storage is selected
in the end-of-horizon in the Northeast. By then, reserve requirements are higher, due to an increased
participation of VRE production and a larger consumption of electricity. With the assumption that all
hydropower plants greater than 400 MW and with a minimum water head will join the automatic gener-
ation control, results show that only in the end of the horizon will it be necessary to secure more reserves
though both natural gas plants (OCGT and CCGT) and Li-Ion batteries.
3.3.5 Detailed Transmission Expansion

In the previous section it was seen that more than 41 GW of wind power and 20 GW of utility scale solar PV
are selected by the expansion planning model, as well as new thermal capacity. In this context, the objective
of this chapter is to evaluate the reinforcements required by the national power grid to accommodate all the
new capacity for the reference expansion case. A three-step procedure is adopted, as shown in Figure 3-22.
Figure 3-22: Transmission network study procedure
The admitted transmission investments include the duplication of existing lines, with the same technical
parameters. The set of substations of the PDE 2026 configuration is maintained in the end of horizon.
Initially, an incremental transmission network is created using PSR format for period 2018-2026, supported
by the Anarede16 database for PDE 2017 -2026. The next step allocates to the high voltage grid, plants not
dispatched by ONS. Then, the new capacity determined by OptGen per region is connected to the busbars.
The next step is to run a SDDP with full transmission to remove the gap between the energy study and the
electrical studies and to identify circuits whose flows may approach or even exceed their nominal capacity,
signalling the natural candidates for reinforcements. The number of parallel circuit candidates is proportional
to the ratio between the maximum flow for all simulated conditions and the actual circuit capacity. Figure
3-23 presents the linear power flows for simulated SDDP conditions without monitoring intra-subsystem
circuit flows identifying candidates for expansion.
16
Anarede is used by EPE and ONS for electrical studies.
November 2019 76
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-23: SDDP run with full transmission and no flow monitoring identifies reinforcement candidates
The OptNet execution uses a full representation of the high voltage network (>138 kV) through a linearized
power flow model. The solution method is concisely shown in Figure 3-24.
Phase 1: Feasibility problem

Determines the infeasible scenarios: Load curtailment > 0
Phase 2: Investment problem ( scenarios)

Solves the expansion problem for critical scenarios within scenarios and finds
an investment decision common to all of them, until all scenarios are solved
Phase 3: Solution feasibility (ℵ scenarios)

Solves ℵ OPF problems: minimize load shedding. If there are no infeasibilities the
solution is suitable for all scenarios.
Phase 4: Eliminating redundancies
Each circuit is tentatively removed in descending cost order and the feasibility is
verified for all scenarios
Figure 3-24: Solution method for OptNet
3.3.5.1 Transmission reinforcement results

Table 3-7 shows the number of circuits added in each of the region, the total distance and investment. As
seen, total investments are estimated at almost USD 10 billion.
Table 3-7: Synthesis of transmission expansion results per region
NUMBER OF LINES / LENGTH INVESTMENT

REGION
TRANSFORMERS [KM] [USD BILLION]
S 5/2 112 0.26
SE 26 / 15 2,063 0.64
CW 3/5 242 0.05
NE 128 / 12 14,397 3.14
N 23 / 10 1,748 1.79
November 2019 77
20-26-00173
Processing No.: 15.2126.9-001.00
NUMBER OF LINES / LENGTH INVESTMENT

REGION
TRANSFORMERS [KM] [USD BILLION]
17
HVAC Tie Line 4 1,100 1.39
HVDC Tie Line 1 bipole 1,800 2.18
SIN 234 23,262 9.45
Table 3-8 shows the number of lines built in each region per voltage level.
Table 3-8: Number of lines built in each region per voltage level
REGION 230 KV 345 KV 440 KV 500 KV TOTAL

S 3 2 5
SE 3 16 2 5 26
CW 3 3
NE 110 18 128
N 5 18 23
Total 124 16 2 43 185
Table 3-9 shows the investments in intra-regional transmission lines per voltage level (transformers are not
included). The values are expressed in millions of US dollars.
Table 3-9: Investments in intra-regional transmission lines per voltage level
INVESTMENT (MUSD) 230 KV 345 KV 440 KV 500 KV TOTAL

S 23 235 258
SE 14 374 31 113 532
CW 34 35
NE 1811 1263 3073
N 96 1677 1774
Total 1978 374 31 3288 5671
Figure 3-24 shows in black where the reinforcements were made.
17
The 1100 km of the HVAC represent the sum of the distances of the circuits connecting the NE to SE regions.
November 2019 78
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-25: Transmission reinforcements
3.3.6 Detailed Simulations of the Hourly Operation

The last step of the energy planning process in to run a detailed model of the SIN transmission constrained
network operation using hourly times steps. New generators (selected by OptGen) and new transmission
elements (selected by OptNet) are included in the system configuration of PDE 2026. SDDP model runs is
executed twice.
This section presents a sample of some of hourly operation results of the Brazilian power grid for the end-
of-horizon, including transmission constraints. The study considers the expansion of generation and trans-
mission projects decided in the previous steps of the process.
Figure 3-26 illustrates the total production of the Brazilian power system for a week in august for an average
hydrologic year (1994) for the end of horizon. As seen, thermal generation in this week is based on inflexible
CCGT and biomass generation. Wind power has a daily variability, with the tendency to increase production
in the night-time. Solar PV from utility scale and distributed generation is clearly seen from the picture and
hydropower responds for most of the flexibility, in this case, required to meet demand in real time.
Thermal Biomass Wind Solar Hydro Batteries
180
160
140
120
100
GW
80
60
40
20
0
1 25 49 73 97 121 145
November 2019 79
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-26: Total production for a week in August
3.3.6.1 Start-ups of thermal power plants

Figure 3-27 illustrates the number of start-ups per month of an OCGT plant for three hydrologic years. It is
clear from the picture that the attributed of flexibility that is provided by the open cycle gas turbines is mostly
utilized during March, with just under one start-up per day on average. Thermal power is more stable in June
or July. CCGT basically has a minimum (inflexible) generation equivalent to 70% of the capacity. OCGT,
which has a higher variable production cost, is usually not required to produce electricity in this period, thus
start-ups become less frequent.
Dry Wet Mean

30
Number of srart-ups per month
25
20
15
10
5
0
1 2 3 4 5 6 7 8 9 10 11 12
Figure 3-27: Number of OCGT start-ups per month for different hydrology years
3.3.6.2 Nodal marginal costs

The hourly marginal costs of three substations of Brazil for a day in August is illustrated in Figure 3-28,
Figure 3-29 and Figure 3-30. Each graph has three curves which relate to dry, average and wet years. Notice
how hydrologic conditions are key to determining the level of marginal cost and that hourly variation is usually
not very large, which is typical of systems with a large amount of hydropower. It is also possible to notice
how dramatically different the marginal costs can be for the hour and day, purely dependent on the busbar
location. It can be seen for the dry hydrology that the Northeast region presented a low marginal cost be-
cause wind power in August was enough to compensate lower hydrology, whereas in the North and South-
east hydro-dependent regions, marginal is higher, from USD 80 to USD 100 per MWh versus less than USD
20 per MWh in the Northeast.
November 2019 80
20-26-00173
Processing No.: 15.2126.9-001.00
Average Wet Dry
120
100
80
USD/MWh
60
40
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Figure 3-28: Hourly marginal costs for August 18th for Albras 230kV busbar (North region)
Average Wet Dry
120
100
80
USD/MWh
60
40
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Figure 3-29: Hourly marginal costs for August 18th for Milagres 500kV busbar (Northeast region)
Average Wet Dry
120
100
80
USD/MWh
60
40
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Figure 3-30: Hourly Marginal costs for August 18th for Bandeirantes 88kV busbar (Southeast region)
3.3.6.3 North-Northeast power exchanges

The North to Northeast energy exchange, given by the total flows on circuits connecting these regions, is
presented in Figure 3-31. Notice the reddish from March to May, which show that the flow reaches maximum
capacity in these months of higher water inflows in the region. Notice also that there is an inversion of flows
in the second semester of the year, with the Northeast becoming an exporter to the North during this period,
November 2019 81
20-26-00173
Processing No.: 15.2126.9-001.00
with higher wind speeds. In the heatmap bluish colours denotes negative values, thus indicating that elec-
tricity flows from the Northeast to the North for most of the second semester at the rated capacity of 5 GW.
Figure 3-31: North>Northeast energy exchanges for a normal hydrological year (GW)
The flow duration curve of each pole of the Belo Monte DC Link, is shown in Figure 3-32 and Figure 3-33.
As seen, more than 92% of the time the flow is either zero or in the direction Xingu to Estreito. Moreover,
80% of the time there is flow is either zero or follows the Xingu to Terminal Rio direction.
5
4
3
2
1
0
GW
0%
2%
5%
7%
10%
12%
14%
17%
19%
21%
24%
26%
29%
31%
33%
36%
38%
41%
43%
45%
48%
50%
53%
55%
57%
60%
62%
64%
67%
69%
72%
74%
76%
79%
81%
84%
86%
88%
91%
93%
95%
98%
-1
-2
-3
-4
-5
Figure 3-32: Power flow distribution of DC Link for Xingu to Estreito

5
4
3
2
1
GW
0
0%
2%
5%
7%
10%
12%
14%
17%
19%
21%
24%
26%
29%
31%
33%
36%
38%
41%
43%
45%
48%
50%
53%
55%
57%
60%
62%
64%
67%
69%
72%
74%
76%
79%
81%
84%
86%
88%
91%
93%
95%
98%
-1
-2
-3
-4
-5
Figure 3-33: Power flow distribution of DC Link for Xingu to Terminal Rio
3.3.7 Conclusions
The objective of this report is to identify and assess state-of-the-art methodologies and computational tools
that may be useful to EPE and ONS in their studies and decision-making process. The focus is on Variable
Renewable Energy (VRE) generation sources such as wind and photovoltaic solar.
November 2019 82
20-26-00173
Processing No.: 15.2126.9-001.00
3.3.7.1 Candidate projects

The use of candidate locations from wind and solar projects not selected in auctions followed by the adjust-
ment of the MERRA hourly data based on regionalized power production of existing plants was found to be
effective in the development of a realistic set of VRE candidate projects.
The study showed - through the sensitivity case which relaxed the constraint of maximum wind power pen-
etration per cluster in the South region, which dramatically changed the results - that it is important to estab-
lish criteria to better assess the feasible capacity of projects that can be selected in each region.
As discussed, a process should ideally combine layers of relevant information, such as the wind speed,
available area, distance to roads, terrain slope, existence of conservation or indigenous areas, and others,
to perform a “screening” with GIS related tool of the potential wind or solar capacity that could be explored
in each cluster. With such process, the preparation of candidates for the system expansion is improved as
they will limit the resources, including those with highest capacity factors, to more sensible levels.
3.3.7.2 Renewables and inflows scenarios

The joint modelling of renewable and inflow scenarios has allowed the representation of multiple time scales
(monthly in the case of inflows and hourly for VRE) and of the spatial correlation among renewables in
different regions. This latter aspect is important because of the “portfolio effect”, i.e. the total variability of
several groups VREs may be reduced if their spatial correlation is small. It is interesting to observe that the
portfolio effect for hydro plants was one of the reasons for the historical development of hydro in eleven
major basins across the country. In other words, a well-proven planning scheme can be extended to the
planning of VREs. The VRE scenarios were also fundamental for the definition of probabilistic reserve gen-
eration criteria for VREs used in the expansion planning study described next.
3.3.7.3 Expansion planning study

The first modelling issue in the expansion planning study was the joint representation of generation and
transmission reinforcements. Due to the very large size of Brazil’s HV network, the optimal generation-trans-
mission plan was determined using a seven-region representation. In conceptual terms, this was equivalent
to ignoring the limits of transmission lines inside each region; only the regional tie-lines, plus their reinforce-
ment options, were represented.
The second modelling issue was the representation of distributed generation (DG) at the distribution level.
This was done through the creation of additional nodes for each distribution company. Due to lack of more
detailed information about DG insertion in Brazil, EPE’s forecasts of DG capacity penetration were used as
a fixed input data. In the future, this modelling could be revisited, possibly by an iterative scheme in which
the system expansion is carried out for a given hypothesis of DG penetration; this DG insertion hypothesis
is then revised with basis on the energy cost component for tariffs in each distribution company; and the
system expansion is re-run with the revised DG insertion.
3.3.7.4 Main insights from the expansion planning study

· Massive penetration of renewables: Wind power and solar PV accounted for most of the generation
additions until the end of horizon with 42GW of wind and 50GW of PV, respectively. (As mentioned,
30 GW of DG were fixed.) As the portfolio selection is driven by expansion and operation cost
minimization, which includes reserves requirements that increase due to the resource intermit-
tence, it is expected that VRE will continue increasing its share in Brazil’s electricity matrix in the
years to come.
· Role of hydropower and gas fired plants to manage VRE variability: the exercise showed that Bra-
zil’s hydropower base made the massive VRE insertion more economic because it reduced the
need for gas-fired plants used for generation reserve.
November 2019 83
20-26-00173
Processing No.: 15.2126.9-001.00
· Seasonal complementarity between wind and inflows in the Northeast reduces the need for hydro
storage: the basic role of hydro storage is to transfer water from the wet period to the dry period.
The amount to be transferred depends, among other factors, on the demand to be served in the
dry period. Because wind generation in the Northeast is higher in the dry season, this means that
the net load (i.e. load minus wind generation) decreases as the amount of wind power in the region
increases. Therefore, the need for seasonal water transfer using reservoirs also decreases.
· Seasonal wind inflow synergy with inflows in the Northeast increases the use of Belo Monte’s
transmission system: during the dry season, only a small fraction of Belo Monte’s generation ca-
pacity and transmission system is used. The idle transmission capacity is then used to bring wind
power from the Northeast to the Southeast through Belo Monte DC Link, thus increasing its value
to the system.
· Commercial impact of VRE curtailment: During the hourly simulation of system operation, some
VRE “spillage” was observed during the high-wind season. This “loss” of available generation is
certainly justifiable from the planning model’s objective of maximizing overall economic benefit.
However, this raises some regulatory and commercial issues. For example, the “Garantia Física”
(GF) of a wind plant is based on its expected power generation. If part of this generation is “spilled”
by the National System Operator (ONS), the plant’s GF may be reduced, which has adverse com-
mercial consequences. This regulatory issue should be addressed.
· Given this expansion, by the end of the horizon roughly 90% of the generation will be clean (88%
renewable, including hydropower and 2% nuclear). Hydropower will remain an important source
with 50% of the total production despite the modest expansion considered in the study (based
solely on the PDE 2026).
· Power sector greenhouse gas emissions: the emission intensity of the Brazilian power sector will
be on the lower end in the global scale, with just around 51 grams of CO2eq per kWh. As a refer-
ence, this is about a tenth of the global average.
· Need to increase the range of generation technologies eligible for AGC: because the reserve re-
quirements increase substantially with the VRE penetration, the current set of plants eligible to be
connected to the AGC will be insufficient. This observation motivated a discussion between the
Consultants and ONS regarding extensions to the current criteria. As a result, it was considered in
this study: (i) the inclusion of OCGT and CCGT in the AGC; (ii) a limit to the maximum individual
allocated reserve (per hydro unit) of 40% for its capacity, to avoid significant efficiency loss; and
(iii) the use of batteries and other energy storage devices as candidates. In the reference case 2.4
GW of capacity - just under 2 GWh of energy - of Lithium-Ion were selected.
November 2019 84
20-26-00173
Processing No.: 15.2126.9-001.00
3.4 Product 3: Power System Studies

The power system studies concern the in-depth simulation of the Brazilian power system in normal and
under contingency conditions. These power system studies aim at assessing in detail the suitability and/or
need to adapt the SIN in order to ensure its operability and maintain the targeted level of reliability with
increased VRE integration. The main objectives of Product 3 are the following:
· Identify the new challenges and related mitigation measures that could affect the reliable operation
of the SIN;
· Present proposals for enhancing the performance of the SIN;
· Indicate actions that could lead enhancing the reliability (adequacy and security) of the SIN and/or
indicate VRE integration limits for each sub-system;
· Investigate possibility of deploying demand-side corrective measures and/or structural solutions un-
derpinning SIN security;
· Assess VRE representation in the transmission planning models currently used;
· Evaluate the need to develop new models and/or upgrade the models currently in use for electrical
studies.
In this project, the power system performance analysis comprises static simulations (AC load flow, static
security assessment, short-circuit currents, reactive power compensation analysis and operation optimiza-
tion) and dynamic simulations (transient stability, voltage stability, frequency stability, small-signal stability
and dynamic security assessment), as depicted in Figure 3-34.
The outcomes of this task include the evaluation of network voltage profiles, network overload levels, critical
instability events, possible future requirements for preventive or corrective steps, such as generation re-
dispatch and network topology switching, adequacy of operating reserves, etc.
Figure 3-34 Power system studies
In this study the Consultant made use of the power system simulation tools developed by Tractebel: Smart
Flow (v2.2), Eurostag (v5.2) and IPSO – Integrated Power System Optimizer. Because of the use of these
tools, a model conversion process needed to be put in place in order to convert the models from the format
used in CEPEL’s tools (ANAREDE, ANAFAS, ANATEM) to the formats of the tools developed by Tractebel.
November 2019 85
20-26-00173
Processing No.: 15.2126.9-001.00
3.4.1 Development of the base static model

For the development of the static model in the Smart Flow format, an 7-step process was employed by the
Consultant, as synthesized in Figure 3-35.
•Mapping of all generating units

Step 1
•Identification of generating units able of operating as synchronous condensers

Step 2
•Development of an Excel/VBA script to convert the static model from PWF to PSA
Step 3 format
•Individualization of step-up transformers

Step 4
•Correction of inconsistencies in network parameters

Step 5
•Inclusion of synchronous machine data for short-circuit calculation purposes

Step 6
•Load flow model conversion validation

Step 7
Figure 3-35: Process for the development of the static model
Step 1: Identification and mapping of all generating units composing the base network model. As the load
flow model provided by EPE still does not make use of the concept of individual equipment modelling 18, it
was necessary to identify all generating units associated with the equivalent power plant injections in the
ANAREDE model in order to make it possible the representation by individual equipment in Smart Flow.
Step 2: Identification of generating units that are capable of operating as synchronous condensers.
Step 3: Automating the conversion of the static model in ANAREDE format (.pwf file) to the Smart Flow
format (.psa file), through the development of a three-stage VBA script.
Step 4: In this study it is adopted the representation of the system by individual equipment with the power
plants explicitly modelled by individual generating units. Because of that, the step-up transformers of the
generating units also need to be represented in an individual manner 19
Step 5: Analysis of the parameters of the static model with the support of the data validation capabilities of
Smart Flow and the experience of the Consultant. The following types of inconsistencies were tackled:
· Rated capacity of lines and transformers: MVA limits for normal operation higher than the MVA limits
for short duration emergency conditions;
· Transformer copper losses higher than the leakage impedance;
· Transformers with very low leakage impedance values 20;
18
It must be emphasized that the capability of individual equipment modelling was originally not supported in
ANAREDE. This is a functionality that is now available, but the official models of the Brazilian system are not yet mak-
ing use of that functionality.
19
Please notice that in the ANAREDE model used for PDE 2026 the step-up transformers – and the power plants –
are represented in an aggregated way
20
The leakage impedance value for these transformers was set equal to the value set in the ANAFAS database.
November 2019 86
20-26-00173
Processing No.: 15.2126.9-001.00
· Lines with very small impedance were converted to circuit breakers21;

· Voltage limit groups were reset according to the values defined in Sub-Module 23.3 of the Brazilian
grid code [5].
Step 6: In Smart Flow all static analysis computation modules use the same data file for the representation
of the system. Different from the tools developed by CEPEL and currently in use as official tools in Brazil,
data for both load flow and short-circuit current computation are stored in the same data file in Smart Flow.
It was therefore necessary to include in the PSA file the information related to the short-circuit impedance of
the generating units.
Step 7: Once the static model in ANAREDE format was converted to Smart Flow and the data inconsisten-
cies and model adaptations (e.g. individual representation of equipment) were corrected, a load flow com-
putation was carried out considering the load flow case configuration from PDE 2026. The objective of this
step was to ensure that the converted model matches the results obtained with the original model.
3.4.2 Development of the base dynamic model

The dynamic model conversion process is divided in two steps. Step 1 comprises the conversion of the
dynamic data file (ANATEM to Eurostag) and Step 2 comprises the conversion of the user-defined models
(CDU to macroblock).
In order to define the DTA file for the PDE 2026 dynamic model, an Excel/VBA script was developed by the
Consultant to read the different data files from ANATEM and convert them to the Eurostag format, as syn-
thesized in Figure 3-36.
Figure 3-36: ANATEM to Eurostag data file conversion (Step 1)
Step 2 is the most labour-intensive and complex task in the model preparation phase, consisting in convert-
ing all user-defined models from ANATEM format (CDU files) to Eurostag macroblocks, as synthesized in
Figure 3-37.
21
In Smart Flow it is possible to explicitly model circuit breakers via its capability of detailed substation topology mod-
elling. A topology pre-processing phase is carried out prior to the execution of the computation modules (load flow, N-
1, etc.) in order to eliminate the null impedance branches from the model, improving the numerical stability of the algo-
rithms.
November 2019 87
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-37: Conversion process for the user-defined models (CDU to macroblock)
Considering the complexity involved in the process of converting the dynamic model to the Eurostag format,
it was needed to perform a detailed model validation procedure in order to ensure that the converted model
presents identical behaviour as the original model, as presented in Figure 3-38.
Figure 3-38: Dynamic model validation procedure
November 2019 88
20-26-00173
Processing No.: 15.2126.9-001.00
Before starting a simulation, Eurostag performs a data consistency check on the synchronous machines
parameters and on the topology and parameters of the user-defined models (macroblocks). During this anal-
ysis, the synchronous machines parameters are verified according to the following rules 22 [7], [8]:
" " (3-1)
≥ > ≥ > ≥
" (3-2)
> > " > >
" (3-3)
> > " >
" (3-4)
< " ≤
Throughout the validation procedure, the following five types of inconsistencies are found in the original
dynamic model (PDE 2026 database):
· Inconsistency 1: > ;
· Inconsistency 2: = ;
· Inconsistency 3: = 0;
· Inconsistency 4: < ;
· Inconsistency 5: = .
3.4.3 Coupling of energy and electrical models

The coupling of energy and electrical studies is a key aspect when performing power system planning under
high VRE penetration levels. In order to effectively perform power system planning activities with high shares
of VRE sources, it is required to seamlessly integrate the energy and electrical study models considering
the generating unit, transmission network, system load and distributed generation database.
The coupling of generating unit data is done in two steps:
· Step 1: existing units and expansion foreseen in the PDE 2026 database;
· Step 2: new units resulting from the generation expansion exercised performed in the Product 2 of
this project.
Figure 3-39 presents the process for coupling the energy and power system simulation models for the exist-
ing units and the expansion foreseen in the PDE 2026 database.
The energetic operation simulation model represents the generating units at power plant level (each power
plant represented by an equivalent generator). However, the power system simulations required a finer rep-
resentation of the generation (representation by individual units). Thus, it is necessary to couple the gener-
ation data of energy and power system models to match the output of the energy studies (Product 2) with
the inputs of the power system studies (Product 3). It is also necessary to ensure that the transmission
network model used in the energetic operation simulation database matches with the model used in the
power system simulation database.
In this project, the coupling of the transmission network model between the energy and power system sim-
ulation models is split in two steps, as follows:
· Step 1: network topology as per the PDE 2026 database;
· Step 2: representation of new network elements as a result of the detailed transmission expansion
planning exercise carried out in Product 2.
22
These inequalities reflect the constructive properties of synchronous machines. The validity of these inequalities can
be easily checked by analysing the mathematical expressions for standard parameters of synchronous machines.
November 2019 89
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-39: Coupling of energy and power system simulation models (generation data, PDE 2026 data-
base)
3.4.4 Dynamic Modelling of Wind and Solar PV Power Plants

Utility-scale wind and solar PV power plants may consist of tens to hundreds of individual generating units
(wind turbines or solar PV inverter stations). It is well-known in the power system community that the repre-
sentation of wind and solar PV power plants in high level of details for system-level studies does not bring
added value to the studies. It is therefore a common practice around the world to simplify these models in
order to make them computationally tractable while retaining the main dynamic characteristics of the plants.
The modelling approach employed by the Consultant for the representation of utility-scale wind and solar
PV power plants for power system analyses are based on works developed in the USA and Europe in the
last years.
November 2019 90
20-26-00173
Processing No.: 15.2126.9-001.00
3.4.4.1 Wind power plants

The proposed equivalent model for wind power plants (WPPs) is built based on the IEC 61400-27-1 (Edition
1.0 2015-02) standard [1]. The representation of the wind power plant is simplified in order to represent the
wind turbines and the internal collector network by an equivalent wind turbine at the PoC of the plant to the
grid. The model is specified for fundamental frequency, positive sequence response.
The WPPs model is developed to represent wind power generation in studies of large-disturbance short-
term voltage stability phenomena and other dynamic short-term phenomena such as rotor angle stability,
frequency stability and small-disturbance voltage stability. The model is applicable for dynamic simulations
of power system events such as short-circuits (low voltage ride through), loss of generation or loads, and
system separation of one synchronous area into more synchronous areas. However, it is not intended for
long-term stability analysis, investigation of sub-synchronous interaction phenomena and investigation of
the fluctuations originating from wind speed variability in time and space. The models are not applicable to
studies of extremely weak systems including situations where wind turbines are islanded without other syn-
chronous generation. This model do not cover phenomena such as harmonics, flicker or any other EMC
emissions.
The WPPs model is composed by the equivalent model of the wind turbine generator (full converter) and its
controls (active power, reactive power, voltage, power factor, synthetic inertia, etc.) and by the power plant-
level controllers (active and reactive power control loops).
Figure 3-40 presents a general overview of the wind power plant model developed for this project. The model
is composed by two sub-modules, namely:
· WTG model: represents the equivalent model of the wind turbine generator and its controls;
· PPC model: represents the power plant-level controllers.
Figure 3-40: WPP dynamic model – overview
The general structure of the WTG model is shown in Figure 3-41. This model is composed by the following
elementary parts:
· Electrical equipment: shunt capacitor, transformer and circuit breaker models;
· Grid protection: WPP protection model;
· Generator model: WTG model (e.g. asynchronous generator, DFIG, full-converter, etc.);
· Mechanical model: WT mechanical model;
· Aerodynamic model: WT aerodynamic model;
· WTG controls: WTG control models (active power, reactive power, voltage, power factor, synthetic
inertia, etc.).
November 2019 91
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-41: WPP dynamic model - WTG model overview
The general structure of the power plant controller (PPC) model is shown in Figure 3-42. This model is
composed by the following elementary parts:
· Active power control: power plant-level active power control loops;
· Reactive power control: power plant-level reactive power control loops.
Figure 3-42: WPP dynamic model - PPC model overview
3.4.4.2 Solar PV power plants

The proposed equivalent model for solar power plants (SPPs) is built based on the works performed by the
“Renewable Energy Modelling” task-force set up by the Western Electricity Coordinating Council (WECC) in
2012 [2]. The representation of the solar PV power plant is simplified in order to represent the PV inverters
and the internal collector network by an equivalent inverter at the PoC of the plant to the grid. The model is
specified for fundamental frequency, positive sequence response.
The WECC generic models are designed for transmission planning studies that involve a complex network,
and a large set of generators, loads and other dynamic components. The main objective of this model is to
assess dynamic performance of the system, particularly recovery dynamics following grid-side disturbances
such as transmission-level faults. This approach does not allow a detailed representation of very fast controls
and a response to imbalanced disturbances. The WECC equivalent representation and simplified dynamic
models are not recommended for evaluation of fault ride-through. Considering that terminal voltage can vary
significantly across the plant, a single machine representation has limitations with respect to assessment of
voltage ride through.
The SPPs model is composed by the generator (inverter) and its local active power and reactive power
control and by the PV power plant level active and reactive power control.
Figure 3-44 presents a general overview of the utility-scale solar PV power plant model developed for this
project. The model is composed by the generator, the generator controls and the power plant control.
November 2019 92
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-43: SPP dynamic model – overview
A more detailed view on the model structure is given in Figure 3-44.

· REGC_A module: represent the Generator/Converter (inverter) interface with the grid;
· REEC_B module: represent the Electrical Controls of the inverters;
· REPC_A module: represent the Plant Controller.
Figure 3-44: Detailed SPP dynamic model structure [2]
3.4.5 Selection of Relevant Operating Conditions

In Product 2, simulations of the hourly operation of the system have been carried out for seven hydro/VRE
scenarios. Considering that there are 8,760 operating conditions (one for each hour of the simulated year)
in each of the simulated scenarios, a total of 61,320 operating conditions needed to be analysed. Performing
detailed power system studies for every single operating condition is an infeasible task. It was therefore
necessary to select the most relevant operating conditions from the point of view of system planning under
high VRE shares. Figure 3-45 depicts the approach adopted by the Consultant to select relevant operating
conditions for power system analyses.
In this project, the selection of the relevant operating conditions starts by a detailed analysis of the results of
the simulations of the hourly operation of the system. This analysis aims at understanding the main charac-
teristics of the energetic operation of the system in order to support the selection of the operating conditions
for which detailed power system studies need to be carried out. Once this analysis is concluded, the selection
of the relevant operating conditions is performed using two complementary approaches, as follows:
· Selection of extreme operating conditions;
· Selection of “likely” operating conditions.
November 2019 93
20-26-00173
Processing No.: 15.2126.9-001.00
The selection of extreme operating conditions is based on pre-defined criteria (deterministic approach), aim-
ing at identifying the most critical system operating conditions from different perspectives. The selection of
likely operating conditions aims at identifying the operating conditions that are more likely to happen during
an operating year and is based on statistical data analysis (probabilistic approach).
Figure 3-45: Selection of relevant operating conditions – overview
The analysis of the hourly operation of the system aims at understanding the system operation patterns and
supporting the selection of relevant operating conditions for the power system studies. In order to do so, the
instantaneous VRE penetration level, the tie-line power flows, the load/generation balance of each subsys-
tem and the composition of the generation dispatch per type of source of Product 2 are analysed. Through
this analysis was concluded that scenarios 4 and 8 are the most critical ones from the point of view of
instantaneous VRE penetration.
Once this analysis is concluded, the selection of the relevant operating conditions is performed by the se-
lection of extreme operating conditions based on a deterministic approach, aiming at identifying the most
critical system operating conditions from different perspectives. As a result, 14 operating conditions for each
of the selected hydro/VRE scenarios (4 and 8) operating conditions were selected for detailed power system
studies, as presented in Table 3-10.
Table 3-10: Selected operating conditions for the power system studies
1 Annual peak load 20XX-02-12 15:00 20XX-02-12 15:00
2 Annual minimum load 20XX-06-28 07:00 20XX-06-28 07:00
November 2019 94
20-26-00173
Processing No.: 15.2126.9-001.00
3 Maximum export: NE 20XX-09-13 15:00 20XX-07-08 07:00
4 Maximum import: NE 20XX-03-14 23:00 20XX-04-07 23:00
5 Maximum export: S 20XX-04-11 23:00 20XX-06-26 17:00
6 Maximum import: S 20XX-02-16 09:00 20XX-03-14 11:00
7 Maximum import: SE 20XX-04-20 18:00 20XX-08-18 18:00
7.a At highest instantaneous VRE share 20XX-09-11 18:00 20XX-05-15 12:00
7.b At highest instantaneous VRE share in SE 20XX-04-22 08:00 20XX-05-15 11:00
7.c At lowest instantaneous VRE share 20XX-05-05 20:00 20XX-06-18 19:00
7.d At lowest instantaneous VRE share in SE 20XX-04-20 18:00 20XX-08-18 18:00
8 Highest VRE/Load share 20XX-09-06 10:00 20XX-07-19 12:00
8.a At high load conditions 20XX-09-11 11:00 20XX-08-18 11:00
8.b At low load conditions 20XX-07-19 07:00 20XX-06-21 07:00
Table 3-11 present the generation mix per geographical region of Brazil, as well as the power flows between
the different subsystems for two of the selected extreme operating conditions, the annual peak load and the
highest VRE/Load share for scenario 4 (see Product 3 report for more details).
Table 3-11: Selected extreme operating conditions – generation mix23
# Scenario 4
23
The figures presented in this table were generated using Kaleidoscope [14].
November 2019 95
20-26-00173
Processing No.: 15.2126.9-001.00
# Scenario 4
The selection of “likely” operating conditions aims at identifying representative states that describe a group
of similar operating conditions, therefore reducing the amount of operating points to be analysed in the power
system studies while covering a wider range of possible operating conditions. In order to perform such a
selection, the Consultant employed cluster analysis (clustering) techniques over a large set of operating
points resulting from the simulations of the hourly operation of the system performed in Product 2. However,
the results are far from acceptable for the given application and were therefore discarded.
3.4.6 Analysis of System Inertia

The concept of inertia is one of the well-known and fundamental principles of classical mechanics used to
describe the motion of objects and how they are affected by applied forces. The way generation units provide
inertia mainly depends on their interfacing technology. Due to the increasing penetration of VRE sources,
many of the traditional sources of inertia might be displaced by converter-connected units providing no in-
herent inertial response. Consequently, the total system inertia tends to decrease.
A preliminary assessment of the system inertia was applied to the Brazilian system. Figure 3-46 presents
an overview of the proposed methodology for the assessment of system inertia in the Brazilian system. The
analysis starts with the disaggregation of the power plant dispatches at generating unit level. This was done
for the 8760 hours of each of the 7 scenarios from the hourly simulations of the system operation (Product
2)24. The next step consists in the calculation of the following quantities for each of the 61320 operating
conditions:
· Total system inertia;
· Equivalent inertia per geographical region;
· System ROCOF.
In addition to the calculation of the aforementioned quantities, the Consultant also calculates the perfor-
mance of the primary frequency control (PFC) in order to determine the frequency nadir (and the time instant
it happens) and post-disturbance steady-state frequency value.
The last step consists in analysing the results produced in the previous steps and translate them into prac-
tical recommendations for the system expansion planning and operation.
24
A total of 61,320 operating conditions (dispatches) are evaluated.
November 2019 96
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-46: System inertia assessment - overview of the methodology
It must be emphasized that the results are obtained using a simplified model for the representation of the
phenomena involved in the inertial response and primary frequency control (PFC). Even if the analyses
make use of a simplified model for the description of the physical phenomena behind the system inertial
response, the values and insights provided by these analyses are very useful as a first assessment of the
impact of VRE integration on the system frequency stability.
Figure 3-47 presents the histograms of the synchronous inertia for the seven scenarios analysed in this
project. It can be seen that scenarios 4 and 8 are the ones that present the lowest synchronous inertia in the
majority of the hours of the year.
Figure 3-48 presents the histograms of the (potential) total system inertia (considering that all available syn-
thetic inertia is in operation at the given moment) for the seven scenarios analysed in this project.
Figure 3-49 presents the statistical distribution of the synchronous inertia of each geographical region for
Scenarios 4 and 8, identified as one of the most critical ones in terms of instantaneous VRE shares. The
results show the following:
· The highest synchronous inertia is located in the SE region, no matter the scenario.
· Despite of significant installed capacity of WPPs, the S region presents relatively high level of syn-
chronous inertia when compared to the installed capacity of the region.
· In hours of very high VRE shares, the synchronous inertia in the NE region falls below 20,000 MW∙s.
· The CW region presents the lowest levels of synchronous inertia between all regions. This is ex-
plained by the lower level of installed capacity in the region.
November 2019 97
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-47: Total synchronous inertia – histograms
November 2019 98
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-48: (Potential) Total system inertia –histogram
November 2019 99
20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-49: Total synchronous inertia per region - histogram (Scenario 4)
The analysis of the PFC performance is performed for the Scenario 4 only. Figure 3-50 presents the dynamic
simulation results for a 1500 MW contingency for the 8760 hours of the operating year in Scenario 4 including
the representation of the first threshold of the UFLS scheme and the minimum and maximum ROCOF val-
ues. It can be seen that, despite the considerable number of hours with high instantaneous VRE penetration,
the frequency nadir is significantly higher than the first UFLS threshold in all operating conditions.
November 2019 100

20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-50: PFC performance evaluation – Scenario 4 (1500 MW contingency)
Table 3-12 and Table 3-13 present a synthesis of the results presented in this section considering only the
synchronous inertia and the total potentially available inertia, respectively. The ROCOF values and the times
to reach the first UFLS stage were calculated considering a generation imbalance equal to 1500 MW, which
represents the loss of the biggest generating unit of the system.
Table 3-12: Synthesis of system inertia evaluation results (synchronous inertia only)
2 407,172 188,465 639,443 0.115 0.070 0.239 13.6 6.3 21.3
4 394,747 157,419 640,790 0.120 0.070 0.286 13.2 5.2 21.4
5 410,171 147,735 644,363 0.115 0.070 0.305 13.7 4.9 21.5
6 425,865 191,636 633,445 0.108 0.071 0.235 14.2 6.4 21.1
7 392,924 166,303 639,088 0.120 0.070 0.271 13.1 5.5 21.3
8 386,954 155,312 625,923 0.122 0.072 0.290 12.9 5.2 20.9
9 393,817 625,923 646,387 0.120 0.070 0.264 13.1 5.7 21.5
2 421,439 201,291 646,893 0.111 0.070 0.224 14.0 6.7 21.6
4 409,905 174,532 652,110 0.115 0.069 0.258 13.7 5.8 21.7
5 423,881 164,941 652,693 0.111 0.069 0.273 14.1 5.5 21.8
November 2019 101

20-26-00173
Processing No.: 15.2126.9-001.00

6 438,676 206,195 638,801 0.105 0.070 0.218 14.6 6.9 21.3
7 407,365 184,467 647,489 0.115 0.069 0.244 13.6 6.1 21.6
8 402,520 170,066 632,435 0.117 0.071 0.265 13.4 5.7 21.1
9 408,087 187,245 650,121 0.116 0.069 0.240 13.6 6.2 21.7
In which regards the performance of the primary frequency control, it can be seen that in the worst case, the
frequency nadir is equal to 59.45 Hz (at 5.4 seconds). Moreover, for the majority of the operating conditions
the frequency nadir lies within the range of 59.6 and 59.9 Hz (99.8% of the hours) while the time instant of
the frequency nadir lies within the range between 3 and 4.5 seconds.
The results obtained in this analysis indicate that, despite the high VRE shares obtained for the studied
scenarios, the system inertia remains at sufficiently high values for the majority of the analysed operating
conditions. This is partly explained by the fact that the planning scenario considered in this study considers
a high demand growth and the high VRE capacity expansion is majorly related supplying this additional
demand rather than replacing conventional generation.
This is a point of attention when comparing the future expansion of the Brazilian system with, for example,
the European experience in VRE integration. In that case, the demand growth is moderate and the fast VRE
expansion comes in replacement of conventional generation, which leads to a potentially high degradation
of the system inertia.
Despite providing very useful insights about the impact of VRE integration on the system frequency stability,
it must be highlighted that it is not possible to extract direct conclusions from the values presented in this
section without performing a more detailed assessment of the system PFC performance.
3.4.7 Static analyses

Planning a power system (in short-, mid- and long-term horizons) for massive integration of variable renew-
able energy sources requires in-depth analyses of the power system performance in both steady-state and
transient conditions in order to ensure the secure operation of the system.
The methodology proposed by the Consultant for the static studies is summarised in Figure 3-51.
•Disaggregation of generation dispatch per generating unit

Step 0
•Analysis of partial transmission expansion plan (energy studies)

Step 1
•Operation optimization (OPF)

Step 2 •Static security assessment (contingency analysis)
•Short-circuit current analysis

Step 3
Figure 3-51: Static studies - overview of the methodology (Product 3)
November 2019 102

20-26-00173
Processing No.: 15.2126.9-001.00
Step 0: A methodology is required in order to disaggregate the dispatch from power plant generating unit
level. In this project, the Consultant employed a methodology to disaggregate the dispatch results through
the solution of an optimization problem. This optimization problem is solved for every power plant of the
system that contains more than one generating unit.
Step 1: Results from the capacity and transmission expansion planning produced in the energy studies
(Product 2) framework need to be refined in order to take into account power flow limits in transmission
equipment outside the main transmission network ( < 230 ) and take into account N-1 contingencies.
Step 2: The reactive power compensation design is performed, specifying additional network reinforcements
and adjusting the system operation using an optimal power flow (OPF) tool.
Step 3: Evaluate the maximum short-circuit currents in order to ensure that the transmission equipment can
withstand the required fault current levels. Also evaluate the system strength by means of calculating the
minimum short-circuit currents and short-circuit ratio [9] at key points of the system (terminals of HVDC
converter stations and connection points of VRE power plants).
3.4.7.1 Step 0: disaggregation of generation dispatch per generating unit

The amount of primary reserve considered for this step equal to 1500 MW (size of the largest generating
unit, Angra 3 – 1405 MW, with a margin of about 7%). The loss of the largest generating unit is considered
to be the sizing incident for the frequency stability analysis and therefore the minimum amount of spinning
reserve for primary frequency control should be around 1500 MW. The distribution of the reserve amongst
the different generating units participating to the primary frequency control follows the criteria defined in the
“Sub-Module 23.3”25 of the Brazilian grid code [10].
3.4.7.2 Step 1: analysis of partial transmission expansion plan

This step consists in performing successive DC load flow and DC static security assessment (N-1) analyses
in order to:
· Take into account power flow limits in transmission equipment outside the main transmission net-
work ( < 230 );
· Take into account N-1 contingencies26.
For each analysis, the loading of all transmission equipment (lines and transformers at all voltage levels) is
analysed for normal (N) and under single contingency (N-1) conditions. Transmission network reinforce-
ments are proposed if the criteria violations are detected and the analyses are then repeated for the rein-
forced network. This process is repeated until no violation of the criteria are detected.
Figure 3-52 presents a synthesis of the network reinforcements done in this step to eliminate possible over-
loads in both N and N-1 situations for all 14 operating conditions under analysis. As presented, 12 rounds
of DC load flow and DC static security assessment (N-1) analyses were needed to eliminate the overloaded
branches.
25
Chapter 5: “Diretrizes e critérios para estudos de reserve de potência operative e de controle carga-frequência”.
26
In this project, only contingencies at the main transmission network ( ≥ 230 ) have been evaluated.
November 2019 103

20-26-00173
Processing No.: 15.2126.9-001.00
Iteration 1 (DC-LF)
New LI: 564 New TF: 461
Iteration 2 (DC-LF)
Iteration 3 (DC-LF)
New LI: 1 New TF: 0
Iteration 4 (DC-LF)
New LI: 4 New TF: 0
Iteration 5 (DC-LF)
New LI: 1 New TF: 0
Iteration 6 (DC-SA)
Iteration 7 (DC-LF)
New LI: 0 New TF: 0
Iteration 8 (DC-SA)
New LI: 9 New TF: 0
Iteration 9 (DC-LF)
New LI: 0 New TF: 0
Iteration 10 (DC-SA)
New LI: 4 New TF: 0
Iteration 11 (DC-LF)
New LI: 0 New TF: 0
Iteration 12 (DC-SA)
New LI: 0 New TF: 0
3.4.7.3 Step 2: operation optimization and static security assessment

This step consists in performing the reactive power compensation design, specifying additional network re-
inforcements and adjusting the system operation using an optimal power flow (OPF) tool28. The great part
of the network expansion and reinforcement options that are to be decided in this step are shunt reactive
power compensation devices to allow proper voltage/reactive power control in the network.
For each analysis, the loading of all transmission equipment (lines and transformers at all voltage levels) is
analysed for normal (N) and under single contingency (N-1) conditions. Transmission network reinforce-
ments are proposed in case of criteria violations and the analyses are then repeated for the reinforced net-
work. This process is repeated until no violation of the aforementioned criteria are detected.
Figure 3-53 presents a synthesis of the network reinforcements done in this step to eliminate possible over-
loads or over-/under-voltage problems in both N and N-1 situations for all 14 operating conditions under
analysis. As presented, 12 rounds of optimal power flow and AC static security assessment (N-1) analyses
were needed to adjust the system operation.
27
DC-LF: DC load flow; DC-SA: DC security assessment.
28
The OPF tool used in this project is Tractebel’s “Integrated Power Systems Optimizer (IPSO). Details about this tool
are presented in the Annex Report.
November 2019 104

20-26-00173
Processing No.: 15.2126.9-001.00
Iteration 1 (OPF - Feasibility)

New LI: 27 New TF: 0 New Shunts: 125




Iteration 6 (AC-SA - Feasibility)

Iteration 7 (OPF - Optimality)

Iteration 8 (AC-SA - Optimality)

3.4.7.4 Transmission expansion summary

Table 3-14, Table 3-15 and Table 3-16 summarise the network expansion defined in the static studies (trans-
mission lines, transformers and shunt compensation, respectively). It must be emphasized that these net-
work reinforcements come on top of the expansion plan defined in Product 2 (energy studies).
Table 3-14: Transmission expansion summary - transmission lines (static studies)

Region
CW 68 51 9 1 7 0
N 46 38 5 0 3 0
NE 165 75 77 0 13 0
S 158 95 58 0 3 2
SE 363 288 9 40 25 1
Total 800 547 158 41 51 3
29
AC-SA: AC security assessment.
November 2019 105

20-26-00173
Processing No.: 15.2126.9-001.00
Table 3-15: Transmission expansion summary - transformers (static studies)

Main Transmission Sub Transmission Three-Winding
Region Total Interface
Network Networks Transformers
CW 77 1 58 6 12
N 71 1 56 7 7
NE 155 16 32 103 4
S 67 2 8 28 29
SE 131 7 86 17 21
Total 501 27 240 161 73
Table 3-16: Transmission expansion summary – shunt compensation (static studies)

Total Capacity [Mvar]
Region
Shunt Capacitors Shunt Reactors
CW 534 -4
N 521 -506
NE 201 -6
S 1817 -625
SE 3136 -536
Total 6209 -1677
A full load flow and static security assessment analysis was performed for the final network configuration for
the selected 14 operating conditions. The results show that the use of the proposed methodology and OPF
formulation lead to adequate voltage profiles across all voltage levels while maximizing the reactive power
margins of the generating units and dynamic var compensators. No violation of voltage magnitude limits and
no network overloads are observed in normal operation (N condition). The average of transmission losses
for the 14 operating conditions is 5.6% of the total generation (minimum losses equal to 3.7% for Case 4
and the maximum corresponding to 8.6% for Case 2). Case 1 presents the highest number of overloaded
transformers, with a total of 3 transformers with 124.7% of overload each. Case 4 presents two overloaded
transformers, the first one with 120.1% of maximum overload caused by 22 different contingencies and the
second one with 145.6% of maximum overload caused by 23 different contingencies. Cases 2, 3, 6, 7, 7a,
7b, 8a and 8b do not present over-loaded transformers. All overloaded transformers are concentrated in the
Northeast region.
A total of about 2470 contingencies are considered for the static security assessment. The possible contin-
gencies consider the simple loss of a circuit (transmission line or transformer) at the main transmission
network ( ≥ 230 ). Case 8b presents the highest number of overloaded lines, with a total of 11 lines
concentrated in the Northeast region, with overloads ranging from 110.8% to 120.0%. Cases 1, 2, 3, 5, 6, 7,
7a, 7b, 7c, 8 and 8a do not present overloaded lines in the main transmission network. The overloaded lines
are more concentrated in the Northeast region, with the exception of two lines concentrated in the South
region. The problems remaining in the final network configuration are not considered critical (overloads not
too important and limited to a few contingencies). Additional network reinforcements could be proposed to
solve those problems.
3.4.7.5 Step 3: short-circuit current analysis

The short-circuit current analysis focused on the 230 and 500/525 kV networks. The 230 kV voltage level
present short-circuit currents up to 30 kA. In the North (N) and Central-West (CW) regions, these currents
are more concentrated around 5 kA. On the other hand, in the Northeast (NE), Southeast (SE) and South
November 2019 106

20-26-00173
Processing No.: 15.2126.9-001.00
(S) regions the short-circuit current values are more dispersed. The 500/525 kV busbars present short-circuit
currents up to 50 kA. In the Northeast (NE) region, these currents are more concentrated around 20 kA. In
the South (S) region these currents are more concentrated between 20 kA and 30 kA. On the other hand,
in the North (N), Southeast (SE) and Central-West (CW) regions these currents are more dispersed. For
some specific cases, the short-circuit currents are above 50 kA, indicating the need for a revision regarding
possible equipment capacity violations. This is explained by the following facts: i) strong expansion of the
transmission network, reducing electrical distances; ii) significant additions of generation (even if predomi-
nantly VRE), increasing the short-circuit current contributions in the network; iii) system expansion based on
a significant demand growth and not by replacing conventional generation by VRE under a more or less
constant load condition.
A special attention was dedicated to the analysis of the short-circuit currents at the terminals of the HVDC
converter stations. The major risk of insufficient short-circuit power lies in the operation of the inverters. In
this case, it can be seen that the SCR is sufficiently high for the inverter operation of all HVDC links (even
the bi-directional ones). Therefore, for the operating conditions analysed in this study, the lack of short-circuit
power for the operation of the HVDC links does not seem to be a major concern. However, dedicated studies
to analyse the operation of these links in such future operating conditions is recommended (e.g. analysis of
multi-infeed problems, etc.).
In which regards the system strength for the operation of VRE power plants, the results show that 16 VRE
power plants (total of 21,180 MW of installed capacity) are exposed to low SCR at their points of connection
(9 WPPs and 7 SPPs). In order to solve the low SCR problems at the PoC of these power plants, synchro-
nous condensers have been allocated in key points of the network aiming at increasing the SCR levels and
allow secure and stable operation of these power plants. A total of 22 synchronous condensers (150 MVA,
-90/+150 Mvar each) have been allocated in points of the network close to the main VRE hubs in the NE
and S regions, mainly.
3.4.7.6 Impacts of the integration of distributed solar PV (DPV)

In this study it is considered a total installed capacity of 30,000 MW of distributed solar PV integrated to the
system at the end of the planning horizon. A high-level analysis of the impact of the DPV integration in the
sub-transmission and transmission networks is performed, which consists in:
· Quantifying the impact of DPV on the loading of sub-transmission (69 ≤ ≤ 161 kV) network cir-
cuits.
· Quantifying the impact of DPV in the power flow direction in the interface transformers30.
These analyses are performed by computing the statistics of power flows in sub-transmission network cir-
cuits and interface transformers for all 14 operating conditions analysed in this project. For the operating
conditions where there is no injection from DPV, the average loading of the sub-transmission circuits remains
within the range between 29.3% and 34.2%. For the cases where there is DPV injection, the average loading
of the sub-transmission circuits is between 24.1% and 39.6%.
For the operating conditions with no injection from DPV, the average loading of the sub-transmission circuits
remains within the range between 29.3% and 34.2%. For the other cases, where there is DPV injection, the
average loading of the sub-transmission circuits is between 24.1% and 39.6%.
In the hours without DPV injection, the number of transformers presenting reverse power flows are between
7.1% and 7.7% of the total number of interface transformers. However, for the operating conditions with non-
30
Interface transformers are defined as transformers connecting the main transmission network to the sub-transmis-
sion networks.
November 2019 107

20-26-00173
Processing No.: 15.2126.9-001.00
null DPV injection, the number of interface transformers presenting reverse power flows increase to a range
between 7.7% and 11.6%. Therefore, an increase in the reverse power flows is observed under the presence
of DPV.
The Consultant recommends a more detailed analysis in order to reach a stronger conclusion on this subject,
taking into account the different load levels and other generating units (in addition to DPV) connected to the
sub-transmission networks.
3.4.8 Dynamic analyses

System stability is one of the main challenges for the power system operation in the presence of VRE [11].
The methodology proposed by the Consultant as state-of-the-art for dynamic analysis under power system
planning for VRE integration is summarised in Figure 3-54. The small-signal stability analysis is the first type
of analysis to be carried out in order to assess if the system is stable and the electromechanical oscillations
are properly damped.
Figure 3-54: Dynamic analysis - overview of the methodology
In this project, the following analyses are not carried out 31:
· Critical clearing time (CCT) computation;
· Voltage Stability;
· Net transfer capacity (NTC) computation.
Small-signal stability analysis is the first type of analysis to be carried out in order to assess if the system is
stable and the electromechanical oscillations are properly damped. It is not recommended to perform the
other types of analyses prior to the small-signal stability because if the system is unstable or the oscillations
are poorly damped, it will have an impact on the transient and frequency stability aspects.
31
Therefore, the methodology for these tasks are not presented in this report. However, they are included in Product 4
as a recommendation for future improvements in the expansion planning methodologies.
November 2019 108

20-26-00173
Processing No.: 15.2126.9-001.00
3.4.9 Small-signal stability

The small-signal stability analysis is carried out using the simulation tool Hercules, an Eurostag add-on.
The approach adopted in this study considers the following steps:
· Eigenvalues computation32:
o Unstable electromechanical modes
o Stable electromechanical modes with damping up to 15% and oscillation frequency up to 2
Hz.
· Analysis of the electromechanical modes:
o Detailed analysis of critical modes ( ≤ 5%) with frequency in the range of inter-area modes
(0.1 to 1.0 Hz) by means of computation of participation factors and mode shapes.
o Analysis of the potential inter-area modes with 5% ≤ ≤ 15%.
· Oscillation damping improvement and/or stabilization:
o For the detected inter-area critical modes, a PSS tuning is performed for the units with higher
participation factors.
o Stable but poorly damped local modes are object of PSS design in this study.
The results of small-signal stability analysis show that in the original configuration there were unstable and/or
poorly damped inter-area modes for the operating conditions 1, 2, 3, 8, 8.a and 8.b. A PSS tuning exercise
was carried out in order to stabilize the system and improve the performance of oscillations damping.
The PSS type adopted in this study is the IEEE PSS2B. The units considered for PSS allocation and tuning
are mostly the new combined- and open-cycle power plants. In addition to those units, the future units of
Angra 3 NPP and Tabajara HPP, as well as the existing units of Termonorte 2, have their PSS tuned to
provide positive damping to the inter-area modes (0.5 to 0.65 Hz). Results for the updated system configu-
ration (with new PSS) show the effectiveness of the implemented PSS configuration. Some poorly damped
local modes remain after the PSS tuning exercise. However, the tuning of PSS for solving problems with
local oscillation modes is out of the scope of this study.
Considering the scenarios with very high VRE penetration analysed in this study, it is recommended further
investigation on the possibility of inclusion of supplementary control signals in WPP and SPP in order to
contribute to the damping of inter-area modes. This is valid especially in the cases where there are few
synchronous units in operation in the N/NE systems, cases in which damping the inter-area mode has proven
to be a challenge in this study.
3.4.10 Transient stability

In this study, only the Dynamic Security Assessment (DSA) part of a full transient stability study is performed.
The objective of the DSA is to assess the security of the system from a dynamic point of view. It can be
seen as an evolution of the static security assessment (N-1 criterion).
32
Hercules computes the eigenvalues and eigenvectors using an Arnoldi method. The initial shifts are automatically
defined in order to guarantee that all eigenvalues inside a given area of the complex plane (defined by the user) are
computed.
November 2019 109

20-26-00173
Processing No.: 15.2126.9-001.00
In this project, the focus of the DSA is on the system stability and voltage recovery after incidents occurring
at the interconnection lines (tie-lines). The sizing incident for the DSA are single-phase-to-ground faults 33
cleared in base protection times.
The following acceptance criteria are evaluated for each of the simulations [10]:
· In any operating condition, the system must remain stable for unexpected trips with or without the
application of single-phase short-circuits, without reclosing, even if there is a loss of any of the ele-
ments of the transmission system.
· In the case of contingencies which cause the trip of part or all of the electrical interconnections
between the subsystems, the resulting subsystems must remain stable.
· The system must be dynamically stable in the small power flow variations in the interconnections.
· In addition to being stable, the system must not be subject to the risk of unacceptable overloads on
equipment, to the violation of voltage ranges, nor to undesirable disconnections of network or load
elements34.
· The minimum voltage for post-disturbance situation, in the first oscillation, cannot be less than 60%
of the nominal operating voltage (63% for 500kV) and, in the other oscillations, it must be greater
than 80% of the nominal operating voltage (84% for 500kV);
· The maximum voltage variation allowed between the beginning and the end of the dynamic simula-
tion must be 10% of the nominal operating voltage, that is, ≥( − 0.1 ∙ ).
· The maximum amplitude of peak to peak effective voltage fluctuations must be 2%, in absolute
value, 10 (ten) seconds after the disturbance is eliminated.
Table 3-17 present the simulation results for a fault at the “Bom Jesus da Lapa (BA)” terminal of the “Bom
Jesus da Lapa (BA) – Janauba (MG)” 500 kV line for the highest VRE/Load share operating condition (Case
8). The following variables are plotted in the different figures, split per geographical region:
· Deviation of rotor angular position for each synchronous machine;
· Voltage at HV AC nodes:
o NE: all 500 kV busbars;
o S: all 500 and 525 kV busbars;
o SE: all 500 kV busbars.
A total of 1484 transient stability simulations have been performed in this study (106 faults for 14 operating
conditions). During the DSA simulations, voltage control issues were encountered in the cases of higher
VRE penetration (2, 3, 8, 8.a and 8.b). In order to solve those issues, additional generators that were not
dispatched in those cases but that are capable of operating as synchronous condensers were put in service
to counteract the problems.
In the final system configuration, the system remains stable for all simulated faults and for all operating
conditions. It must be highlighted the importance of the synchronous condensers in securing the system
operation in cases of high instantaneous VRE penetration. It is of utmost importance that sufficient short-
circuit power remains available in the moments of high VRE generation in order to avoid instability of the
VRE power plant control loops and, as a consequence, undesirable disconnection of these units during
33
The single-phase-to-ground short-circuit is used because, among the possible faults, it is the one with the highest
probability of occurrence in the Brazilian system [72].
34
A first analysis of this criterion is done in the static security assessment presented in Chapter Error! Reference
source not found.. The overloading and over-/under-voltage criteria are checked once again in the DSA in order to
confirm those results.
November 2019 110

20-26-00173
Processing No.: 15.2126.9-001.00
transients. Moreover, the behaviour of the VRE power plants (SPP and WPP) during and just after faults
must be certified during the commissioning phase of those projects. The fault current injection during faults
is essential to contribute to system stability and the respect to the fault ride-through characteristics too.
Table 3-17: Samples of DSA simulation results
Case 8
November 2019 111

20-26-00173
Processing No.: 15.2126.9-001.00
Case 8
November 2019 112

20-26-00173
Processing No.: 15.2126.9-001.00
Case 8
3.4.11 Frequency stability

Frequency stability reflects the ability of the system to face unexpected active power unbalance such as the
sudden loss of power infeed or large loads. Frequency stability is usually ensured through the provision of
operating reserves (primary) and under-frequency load-shedding (UFLS) schemes. High probability events
are secured via the operating reserves, while low probability events are secured with the support of UFLS
or special protection schemes (SPS)35.
The sizing incident is usually determined so as to achieve a trade-off between reserve provision and energy
production. A typical sizing incident is the loss of the largest generating unit of the system plus a margin of
10%. Sufficient operating reserves with adequate technical performance must be secured to cover the sizing
incident without resulting in loss of load.
The under-frequency load shedding (UFLS), only activated in case of non-normative incidents, is organized
in various steps in order to minimize the amount of load shed following frequency transients.
In this study, the frequency stability analysis aims at assessing:
· the amount of operating reserves needed to cover the loss of the largest unit of the interconnected
system and the loss of the largest unit in each subsystem;
· the adequacy of the UFLS scheme already in place in the subsystems.
In order to perform the aforementioned assessments, the loss of the largest generating unit in operation in
each subsystem for each selected operating condition is simulated. A total of 98 frequency stability simula-
tions have been performed in this study (7 contingencies for 14 operating conditions). The frequency stability
of the system for the different operating conditions was assessed by simulating the loss of the largest gen-
erating unit in operation in each subsystem. The simulation results indicate that even for operating conditions
with very high instantaneous VRE penetration the system remains stable and the system frequency remains
always significantly above the first stage of the UFLS scheme. The minimum frequency nadir from all simu-
lations is equal to 59.72 Hz, which is significantly above the first threshold of the UFLS scheme (58.5 Hz). It
indicates a sufficient frequency stability margin on the system.
Table 3-18 presents an example of the frequency stability simulation results for case 8 for the VRE NE
contingency. For more detailed results see Product 3 report.
35
Normative incidents are considered to be tackled in preventive mode while non-normative incidents are tackled in
corrective mode.
November 2019 113

20-26-00173
Processing No.: 15.2126.9-001.00
Table 3-18: Frequency stability simulation results (Case 8)

System Frequency Total PFC Response per Plant Type
The behaviour of the different generation technologies in the PFC performance is a point of attention and
must be carefully taken into account in the planning studies. The following is to be highlighted from the
simulations:
· Nuclear and CPPP units (ST-based) present a quick response just after the contingency but that
fades out quickly.
· OCGT and CCGT units present a good PFC performance (fast reaction towards frequency devia-
tions). However, special attention should be paid to the fact that these units present a higher contri-
bution to the PFC in the first seconds after the contingency and then after about 10 seconds their
power outputs decrease towards a stable level.
· HPP’s present the slowest PFC control performance amongst the different types of units providing
PFC in the system. Their responses are limited by the constructive characteristics of the hydro power
plants and the non-minimum phase behaviour of the units (compromise between performance and
stability of the turbine.
· Biomass and small hydro units provide PFC support. However, in the model only a few of these
units have an associated dynamic model and therefore their contribution to the PFC is not captured
in this study.
· WPP’s, SPP’s and distributed generation are not presented in the results because these units do
not participate in the PFC.
The importance of the interconnections for the performance of the PFC is also highlighted in this study. It is
possible to see that for a contingency in a given subsystem, all tie-lines contribute to bringing additional
power from outside the given subsystem in order to compensate for the generation loss. However, in order
to guarantee that the system benefits from the reserves shared among the different subsystems, adequate
transmission reliability margins must be secured in the operational planning.
3.4.12 General Recommendations

System modelling and data management
· Seamless integration between energy and power system simulation models.
· Implementation of automatic data consistency tests and reporting.
November 2019 114

20-26-00173
Processing No.: 15.2126.9-001.00
· Implementation of a model management platform with the capabilities of version control, modifica-
tion tracing, validation and approval process, etc.
· Thorough review of the turbine/governor models of the Brazilian power system dynamic model da-
tabase to ensure adequate representation of these equipment for frequency stability and PFC per-
formance studies.
· Make use of aggregated wind and solar PV power plant models in system planning studies. Moreo-
ver, it is also recommended that a model development and maintenance scheme as the one recom-
mend-ed in Product 1 of this project be implemented for the modelling of WPP’s and SPP’s.
Computational infrastructure
· High performance power system simulation tools.
· Cloud-based infrastructure.
· Integrated database management system.
· Results processing tools based on advanced data analytics techniques.
· Use of advanced visualization tools, including intensive use of GIS-based applications.
Expansion planning
· Transmission infrastructure is key for successful VRE integration in the Brazilian system. The trade-
off between VRE resource quality and additional transmission capacity investment should be as-
sessed not only from the point of view of interconnection between subsystems, but also within each
subsystem (e.g. 230 kV and 500 kV reinforcement needs in NE and S regions).
· Impact of distributed generation on system planning
· Integrated transmission, sub-transmission and distribution network planning
· Further develop data analytics approaches to select critical and likely operating conditions for power
system studies.
· Take into account in the planning exercise advanced and/or new transmission equipment to allow
secure system operation as an alternative for massive investments in new transmission lines.
· Improve analysis of system strength in the methodologies for short-circuit current analysis in every
planning exercise.
· Improve frequency stability and net transfer capacity analyses in the standard planning methodolo-
gies.
· Ensure that the maximum loss of infeed due to loss of connection networks of wind and/or solar PV
power plants remains lower or equal to the sizing incident used as a planning and operation crite-
rion36.
· Adoption of state-of-the-art and highly flexible optimal power flow tools in the planning process.
· Implement a “quantification and qualification centred” expansion planning framework in order to
quantify and qualify the system impacts of VRE integration and the related costs.
36
Aims at avoiding the connection of large wind and/or solar PV power plants (or groups of them through the same
connection system) through radial connection networks that might results in loss of infeed larger than the normative
contingencies.
November 2019 115

20-26-00173
Processing No.: 15.2126.9-001.00
System operation
· Ensure adequate transmission reliability margins in order to allow effective frequency regulation
support from the conventional generating units (mainly located in the CW and SE regions) to the
areas with higher instantaneous VRE penetration (NE and S).
· Integrate the proposed methodology for system inertia evaluation in the short-term operation plan-
ning in order the system inertia in the generation scheduling phase (including assessment of PFC
performance using a “single-node equivalent” representation).
· Guarantee the availability of the operating reserves (primary and secondary).
· Implement a process of regular certification of generating unit performance for power frequency
control applications (primary and secondary) and its adequacy to ROCOF requirements.
· Online monitoring of system inertia, electromechanical oscillations damping and transmission mar-
gins.
· Ensure regular evaluation of the adequacy of special protection schemes (SPS), power system sta-
bilizers (PSS) and power oscillation damping (POD) controllers in face of highly variable generation
dispatch profiles.
· Regularly evaluate the adequacy of the UFLS schemes taking into account the rapid deployment of
solar PV distributed generation.
· Regularly review the definition, minimum performance requirements, sizing and allocation of oper-
ating reserves (primary, secondary, tertiary, etc.).
· Keep grid codes up-to-date with respect to the system evolution (see Product 1 for further infor-
mation).
· Put in place effective grid congestion management in order to minimize curtailment of VRE produc-
tion due to network- or stability-related constraints.
Capacity building
· Develop advanced knowledge of the principles of operation and details of controls and capabilities
of inverter-based grid interface technologies.
· Teams in charge of energy studies should receive a training on the basics of power systems oper-
ation, stability and control in order to understand the implications of decisions taken at the energy
studies level on operational performance of the system.
· Teams in charge of power studies should receive a training on the basics of energy studies in order
to understand the implications of decisions taken at the energy studies level on the system opera-
tion, as well as to enable the teams to provide the right feedbacks for the energy studies.
· Intensive training of system operators in a full-dynamic dispatching training simulator (DTS)37 should
be put in place in order to allow the operators to better understand the impacts of the VRE penetra-
tion in the dynamic performance of the system.
3.5 Product 4: Methodology Studies

Based on the findings of products 1, 2, 3 and 5, the Consultant propose specific recommendations for fine-
tuning the guidelines, methodologies and criteria to be in the planning activities (both expansion and opera-
tion) of Brazil. These recommendations are summarised next.
37
Also known as Operators Dynamic Training Simulators (ODTS).
November 2019 116

20-26-00173
Processing No.: 15.2126.9-001.00
3.5.1 Energy studies

summarizes the methodological recommendations to be used in energy planning practices considering the
rapid transformation of power systems due to factors such penetration of VRE, distributed energy resources,
and others.
November 2019 117

20-26-00173
Processing No.: 15.2126.9-001.00
Table 3-19: Energy studies - summary of methodological recommendations

Challenges involve modelling, as

(1) Use generation & transmission some reliability modules would
reliability assessment and integra- need to integrate the planning pro-
tion with planning tools. cess and data acquisition (unavail-
ability rates based on historic out-
Two software are used, one for
ages per unit).
generation system expansion
planning and the other for a sto-
A model for system expansion and
chastic operation planning. There (2) Use of model for project finan- The procedure needs to be de-
another for operation with hourly
is an iterative process between the cial analysis to incorporate in- signed from scratch as it is pres-
steps (instead of load blocks), indi-
models until the marginal cost of vestor's risks so that complete ex- ently not made by EPE. Financial
vidualized projects (instead of ag-
expansion is equal to expected pansion plan is compatible with results from auctions can be used
gregate reservoirs per region) and
Models used in value of the marginal cost of oper- developer's decisions. to infer risk profiles.
a probabilistic modelling of hourly
planning study ation (condition for global opti-
1 renewables production. A special-
and general mum). The plan is then evaluated
ized model was used for transmis-
methodology to assure there is sufficient capac-
sion expansion using a linear
ity to meet the peak demand for a (3) robust grid planning model, in- Coordination efforts by EPE’s
power flow model for multiple sce-
reliable supply. This verification cluding N-1 criterion, while consid- energy and electric areas for inte-
narios of production, instead of se-
links capacity-related constraints ering multiple dispatch scenarios. grated planning.
lected conditions, to support a ro-
with energy-related constraints.
bust optimization.
For instance, capacity of hydro
producers is conditioned to the en-
ergy availability.
Modelling issues and database
(4) Reactive power investment op- preparation of resources of reac-
timization. tive power and their investment
unit costs.
38
November 2019 118

20-26-00173
Processing No.: 15.2126.9-001.00

Monthly profile and contribution in
each load block are estimated out- (1) Enhancement of both models
Representation of VRE needs to
side models. for expansion planning and opera-
Use of location of projects that improve in both models (expan-
tion to include probabilistic VRE
Preparation of Wind power: based on data from participated in electricity auctions, sion planning and system opera-
production. (2) Preparation of sce-
VRE candidate projects enabled to take part in the calibration of existing energy rec- tion). The representation should
narios using either a statistical
projects for ex- electricity auctions. Production in ords with MERRA database for ex- move away from monthly mean
2 method or machine learning. (3)
pansion each load block based on AMA tension of time series, use of production (zero variance) to
Hourly or sub-hourly data acquisi-
planning stud- database (actual production). So- Bayesian network for the genera- hourly or sub-hourly scenarios that
tion from VRE projects would need
ies lar PV power: reference values of tion of multivariate scenarios of can incorporate the modulation of
to be organized for planning pur-
auctions used. Production contri- hourly VRE production each source and corresponding
poses and integrated with en-
bution in each load block based on variability.
hanced planning models.
INPE database.
There are older inventories with
Screening of hydro projects based Hydropower inventory studies con-
assumptions that need revision.
on inventories and feasibility stud- sidering both techno-economic
The study did not focus on new The task is to update these studies
Hydropower ies. Sufficient lead time for project and socioenvironmental issues
hydropower plants in addition to under a new paradigm, with social
3 candidate pro- preparation, permits, and others to supported by specialized computer
those already considered in the and environmental considerations
jects participate in the 10-year horizon. model for the preparation of sensi-
PDE. from early stage planning, follow-
Selection based on viability & soci- ble candidates for expansion stud-
ing a more participatory process
oenvironmental complexity. ies.
for the discussion of alternatives.
As an assumption for planning

Investigation of trade-offs between
study resources of region A were Evaluation and definition of spe-
Multi-area Allows resources of region A to multi-area x single-area supply of
not allowed to supply reserves of cific criteria to measure the robust-
4 spinning re- supply reserves of region B, fol- reserves and best practices with
region B, even though this is part ness of the system and define the
serve lowing ONS practice. respect to system
of normal operation practice car- minimum desirable requirements.
expansion planning studies.
ried out by ONS.
An important improvement given

Exogenous scenario of distributed the continuous growth of DG. An
Use of endogenous approach
Distributed EPE utilizes external scenarios of generation imported to study. Allo- analysis of the requirements
based to coherently integrate DG
5 Energy Re- DG based on specific models cation of DG capacity made in shows a large challenge to make
growth scenarios within the plan-
sources (such as diffusive BASS model) "frontier buses" that connect the this operational (GIZ is currently
ning framework.
HV grid with the distribution utility. supporting a study with this objec-
tive).
Limit the capacity of candidate The challenge is to include trans-
Maximum
projects selected in each region mission bottlenecks and others in
6 VRE per No limitation imposed No limitation imposed
breaking up larger clusters into the selection of the number of re-
region
smaller ones if needed. gions of investment decisions in
November 2019 119

20-26-00173
Processing No.: 15.2126.9-001.00

addition to wind patterns/profiles
used in the study.
EPE makes probabilistic assess-

Firm capacity margin with respect Include a system reliability module Introduce the model as a compo-
Reliability of ments of the supply criteria39 (mar-
to peak annual demand and firm that provides feasibility cuts to the nent of planning model and treat
7 expansion ginal cost of expansion equal to
energy margin with respect to expansion model to meet reliability the data, based on equipment out-
planning average marginal cost of opera-
mean demand criteria (LOLP, EENS). ages, for planning purposes.
tion)
This model is required to separate

the factors of project unavailability
(forced outages and scheduled
maintenance). There are at least
Project Include model to determine repre- two advantages: (1) reliability crite-
Use of unavailability rates combin-
maintenance sentative maintenance schedules ria can be introduced in the plan-
8 ing forced and maintenance out- Same as current practice
of the projects to avoid degrading ning model (as seen); (2) dispatch
modelling ages.
supply reliability. scenarios of improve integration
with transmission planning / elec-
trical studies and electrical studies
because nominal capacity is con-
sidered.
A better representation of trans-

mission bottlenecks is the motiva-
Use of community identification al- tion for this recommendation.
Spatial gorithm to determine which buses The number of regions is decided
Seven systems were used, with
representation: EPE uses 11 systems and of the complete HV grid should be after simulating the operation of
9 some differences with respect to
number of sys- 3 fictitious nodes associated in each region for a policies made with a reduced num-
current practice.
tems given number of regions (input ber of regions with the full network
data). and examining impacts on costs.
Association of buses to regions is
based on nodal values.
39
In December 2019, the National Energy Policy Council (CNPE) approved the use of new supply criteria in Brazil. This improvement was necessary to adapt the power
sector to the new technological reality that has been established and the new market design that is intended for the electricity sector
November 2019 120

20-26-00173
Processing No.: 15.2126.9-001.00

Use of simplified representation
The objective is to improve the
with energy transfers among re-
modelling of constraints on the HV
Regional gions limited by informed amounts Use of linear representation that
grid in the expansion planning pro-
10 exchange (based on reliability criteria used in Same as current practice approximates an aggregate effect
cess, thus reducing the gap be-
limits transmission studies). Sum of en- of Kirchhoff second law.
tween energy planning studies and
ergy flows in different from-to ex-
electric studies.
changes also considered.
It is important to establish criteria

for the number of hours that Run
Spinning All Run-of-river plants can follow Introduce hydrology-related con- of River plants can be sustained at
A fraction of the capacity of each
reserve for run- load, except those with high sea- straints, especially in large pro- a given capacity. This may be a
11 plant is available for the supply of
of-river hydro sonality, which contribute with their jects with a strong seasonal pat- function of water inflows due to
reserve for the units in operation.
plants mean monthly production. tern such as Belo Monte (11GW) seasonal effect. This is true for
both reserve allocation and supply
of peak demand.
Large database required for a bot-

tom up demand modelling. Ad-
Incorporation of Demand Re- vantage of modelling effects that
Demand forecasted on macro-eco- sponse and the ability to model are with specific to some sectors
Demand Profile based on 2015 hourly curve
12 nomic variables. Hourly profile structural changes in the demand (e.g. effect of a change in the
forecast by ONS per region
based on verified real data. profile for the different classes of structure of the residential tariff) or
consumption. cross-sectorial, such as an in-
crease on efficiency standards of
motors.
Suppose demand increases with

temperature (due to air condition-
Temperature
Use of temperature and other cli- ing) and that wind velocity is re-
effect on
matic variables in the preparation duced in hot days. Clearly the
13 demand and - -
of scenarios of demand and VRE combined effect on supply and de-
VRE
production. mand is more challenging if these
production
factors are treated jointly than in-
dependently.
Pumped hydro storage (PHS) is a

Energy stored during surplus Preparation of more candidates for
Two technologies were used in special case because projects
Energy power for use in moments with system expansion considering ca-
14 study (Li-Ion 1h battery and NaS must be located in locations with
storage smaller availability with respect to pacity, stored energy, ramping
6h thermal storage) large water head between upper
load, including losses. constraints and location in the grid.
and lower reservoirs.
November 2019 121

20-26-00173
Processing No.: 15.2126.9-001.00
The Global Wind Atlas, for in-

Mesoscale VRE models can be stance, introduces a high-resolu-
Mesoscale
used to downscale reanalysis tion topographic data, such as
15 wind and - -
data, such as MERRA2 that used hills, ridges and land use in a wind
solar models
in study. flow model. Data availability and
security are a challenge.
Risk premiums may be added to

Project costs are either based on
Project project costs. Ideally an interactive
references per technology for Inclusion of risk premiums in pro-
developer process should be made until the
16 "general projects" or come from Same as current practice ject costs when carrying out a cen-
vs. central expansion plan is also financially
specific studies, as hydropower vi- tralized expansion planning.
planner viable considering the developer’s
ability studies.
risk perspective.
Criteria should be defined and in-

corporated in model (such as the The main challenge is to design
use of small variable production the criteria for VRE curtailment
VRE Curtail- Optional cost of curtailment (not
17 - costs) to determine a priority if cur- quickly before it becomes a critical
ment Criteria used in study)
tailment is necessary, to guaran- and controversial issue with the in-
tee a fair operation for the various crease of VRE in the next years.
producers.
Represent the “unknown un-
A careful design of scenarios un-
knowns”, or “black swans” i.e. se-
der unusual, though possible
vere events that are extremely un-
events, is required. As an exam-
likely from the standpoint of sto-
ple, the Ministry of Energy and
The sector models known chastic models, but that may actu-
Resilience Mines of Chile used simulation
18 unknowns, such as water inflows Same as current practice ally happen. For each extreme
Constraints models to answer what-if ques-
to dams. (but feasible) scenario selected by
tions to evaluate the resilience of
the planner the system must have
the national energy system. The
enough resources to ensure load
motivation was the 2010 earth-
supply for a given period even at
quakes with a magnitude of 8.8.
very high operating cost.
Design a module for the prepara- A user-friendly web-based module
EPE already presents maps with
Presentation of PSR produced maps with the re- tion of customized graphs based for the preparation of user-se-
19 expansion but no interface for ma-
results sults of planned expansion. crude output data produced by lected graphs with the results of
nipulation of output results
MDI and other models. his/her interest.
November 2019 122

20-26-00173
Processing No.: 15.2126.9-001.00
3.5.2 Power system studies

The integration in power systems of variable renewable energy (VRE) sources pose the following specific
challenges in comparison to the conventional synchronous machine-based generation: the intermittency and
variability of the VRE, their non- or poor dispatchability and the power conversion technology that is more
and more inverter-based (i.e. without physical inertia). These three peculiarities have a number of impacts
that are not anymore negligible when the level of penetration becomes significant in a specific region of the
power system.
Planning a power system (in short-, mid- and long-term horizons) for massive integration of VRE sources
requires in-depth analyses of the power system performance in both steady-state and transient conditions
in order to ensure system adequacy and security.
The proposed methodology for power system studies in view of system planning for VRE integration is com-
posed by 4 main blocks, as described below and further detailed in Figure 3-55. Not all elements of this
methodology were applied in the Product 3 of this project. Nevertheless, those steps that were not imple-
mented40 in the project are still included as recommended methodological improvements for future studies.
· Analysis of system inertia and primary frequency control (PFC) performance;
· Selection of relevant operating conditions for static and dynamic analyses;
· Static analysis;
· Dynamic analysis.
Figure 3-55: Methodology for power system studies (overview)
It must be emphasized that all the analyses performed within the framework of the power system studies
comprise the entire network and not only the main transmission system. In other words, it means that net-
work loading and voltage/var control are analysed for both the main network as well as the sub-transmission
and distribution networks that are modelled in the database.
Details of the methodologies for each main block are given in the reports of products 3 and 4.
40
The steps that were not implemented in this project are clearly identified in the description of the given methodology.
November 2019 123

20-26-00173
Processing No.: 15.2126.9-001.00
Table 3-20 presents a summary of the methodological recommendations for power system studies proposed
in this project. It must be emphasized that some of the recommended studies were not performed within the
framework of this project but are recommended to be carried out as an evolution of the planning methodology
employed in this project.
Table 3-20: Summary of the methodological recommendations for power system studies
Level of Importance for
Study Performed in P3?
Future Planning Studies
Selection of relevant operating conditions Yes +++
Analysis of system inertia Yes +++
DC load flow and DC static security assessment Yes ++
Operation optimization (OPF) Yes +++
Static security assessment Yes +++
Short-circuit current analysis Yes ++
System strength analysis No ++
Transient stability – DSA Yes +++
Transient stability – CCT No +
Frequency stability Yes +++
Small-signal stability Yes +
Transfer capacity limits No +++
Table 2-11 summarizes the methodological recommendations to be used in power system studies consid-
ering the rapid transformation of power systems due to factors such penetration of VRE, distributed energy
resources, and others.
November 2019 124

20-26-00173
Processing No.: 15.2126.9-001.00
Table 3-21: Power system studies - summary of methodological recommendations

(1) Integrate the input data for all

static simulations in a single input The only challenge is related with
parameter file (e.g. integrate IT. Ensuring compatibility of old
ANAREDE and ANAFAS data models is key.
files).
The CEPEL power system simula- The power system simulation soft-
(2) Implement automatic data con-
tion software package is currently ware package developed by
sistency and quality checks for Definition of the checks to be em-
employed in power system analy- Tractebel was employed in the
every simulation module. Two lev- ployed in line with the types of
sis. The following computation study. The following computation
els of checks should be imple- models in use in Brazil.
modules are used: modules have been used:
Models used in mented: warnings and errors.
planning study - Load flow and contingency analy- - Load flow and contingency analy-
1 sis; sis;
and general
methodology - Optimal power flow; - Optimal power flow; (3) Fully customizable OPF tool, IT-related challenges, as well as
- Short-circuit current calculation; - Short-circuit current calculation; allowing the user to select con- stronger requirements in terms of
- Electromechanical stability simu- - Extended term electromechanical straints and control variables per technical expertise by the user to
lation; stability simulation; equipment and/or groups of equip- correctly configure the optimization
ment based on pre-defined filters. problems.
- Small-signal stability analysis. - Small-signal stability analysis.
Complex IT challenges for the

(4) Integration of energy and
specification of the database,
power system analysis database,
which must be done in close col-
including model versioning control
laboration with the experts in en-
and model approval process.
ergy and power system analyses.
Revision of the equivalent models

Use of equivalent VRE power of existing VRE power plants to
Thorough review of the existing
plant models developed by the ensure that the power plant con-
models of wind and solar PV
planner based on the individual Use of standard models for wind trollers are modelled as well as
power plants, which might trigger
2 VRE modelling models of the power plants. Not and solar PV power plants based that the model represents correctly
further interactions with the power
every model includes the repre- on [1] and [2], respectively. the behaviour of the power plant at
plant owners in order to validate
sentation of the power plant con- the point of connection. For future
the models.
trollers. power plants, use the kind of mod-
els adopted in this study.
41
November 2019 125

20-26-00173
Processing No.: 15.2126.9-001.00

Perform a thorough review of the
turbine/governors of the existing
Models used for system expansion
units in order to ensure that the
planning are the same as the ones
Models of existing and already models reflect the reality of the
used by the system operator (vali-
planned units are the ones availa- units (e.g. Pmin, Pmax, valve
Dynamic model- dated models as per the grid Thorough review of the existing
ble in the ANATEM database from opening limits, rate limits, etc.).
ling of conven- code). dynamic models, which might trig-
3 PDE 2026. Generating units with For new units, employ standard
tional power For future power plants resulting ger model validation campaigns
missing models have been cor- models or models based on similar
plants from the expansion planning exer- with specific power plant owners.
rected by associating a standard units already existing in the data-
cise, not all new units have an as- model to each of the units. base (ensuring that the associated
sociated dynamic model (e.g. new
controls are properly tuned ac-
open and combined-cycle units).
cording to the parameters of the
units).
Modelled as constant current

Adoption of more detailed dynamic
Modelling of dis- sources for the aggregated unit at
4 Not modelled. models such as the one investi- Still an active topic in R&D.
tributed solar PV the distribution feeder connection
gated in [3].
point.
Extensive use of statistics,

heatmaps, duration curves, histo-
grams, etc.
Use of graph layout algorithms for Explore advanced power system
Requires performant computa-
quick visualization of simulation re- simulation results visualization
Duration curves, histograms, tional resources, as well user skills
Presentation of sults and analysis of network to- tools and follow-up research pro-
5 power flow system diagrams, and for the definition of the right visual-
results pology on the performance of the jects in the domain such the initia-
simulation results figures izations for the different types of
system. tives of NREL in advanced power
information being analysed.
Advanced dynamic simulation systems data visualization.
plots using the advantage of sav-
ing the time series of every state
variable of the dynamic model.
Strong IT challenge to implement
Extensive customization of the Customize the applications for the a Python API for every computa-
API-oriented Not available in current version of simulation tools via the use of the recurrent power system simulation tion module. Strong requirements
6
tools CEPEL software package. Python API’s available with the analyses performed by the planner for the users of the tools in terms
computation engines. in a day-to-day basis. of programming and scripting
skills.
November 2019 126

20-26-00173
Processing No.: 15.2126.9-001.00
Pre-defined operating conditions Employ data analytics-based

Selection of Op- Extreme operating conditions se-
based on: hydrological season methods to identify “likely” operat-
erating Condi- lected from the results of the simu-
7 (wet/dry), import/export level per ing conditions from large and Still an active topic in R&D.
tions for Power lations of the hourly energy opera-
subsystem, average VRE dis- dense datasets of generation dis-
System Studies tion.
patch. patch results.
November 2019 127

20-26-00173
Processing No.: 15.2126.9-001.00
3.5.3 Power system planning database

the system, etc.
The main recommendations for the development and implementation of a power system planning database
are given as follows:
· Common database for energy and power system simulation models;
· User access control to the database (read/write permissions, password control, modification track-
ing, etc.);
· Model validation and approval process;
· Common parameters between electrical and energy simulation models must be linked;
· Automatic check of mismatches between electric and energy simulation models;
· Allow data exchange between different computation modules (input data and simulation results);
· Model version control;
· Cloud-based.
November 2019 128

20-26-00173
Processing No.: 15.2126.9-001.00
3.6 Product 5: Technology Studies

Product 5 provides technology related information such as performance and costs as input to the subsequent
planning steps (Product 2, Product 3 and Product 4). It outlines the following aspects:
· Technological evolution of VRE (wind power & solar technologies);
· Solar and wind power sources reliability;
· Technology resources for mitigating variability impacts of VRE (supply side, transmission side and
demand side mitigation).
3.6.1 Technological Evolution of Wind Power

Brazil is one of the largest wind energy markets in the world and the largest in the Central and South America
region. In the last 7 years, the country wide installed capacity has been quadrupled and expects a further
increase up to 19 GW up to the end of 2021, which is a doubling since end of 2016. In this time period, it will
be most likely that the turbine manufacturers which are present in Brazil will introduce their new 4 MW
platforms for this market which will consequently result in a higher rated capacity per turbine compared with
the 2 MW/turbine in 2016.
Another expectation for the next years is the increase of the capacity per wind farm, taking into account that
the benchmark capacity of the power grid utilization fee has been increased in 2018. Thus, it will be most
likely, that the size of contracted wind farms will follow the worldwide trend and will increase to about 100
MW in the next years. The requirement of sound implementation of condition monitoring and preventive
maintenance will therefore become more important in the near future.
Considering the market share and strategy of the main players of the Brazilian market (GE, Siemens-
Gamesa RE), the Doubly Fed Induction Generator (DFIG) in combination with a partial converter will be the
dominant technology in the next years as this is the most economical solution which can comply with most
of worldwide grid codes. In case the grid code requirements will revised and tightened in the future, some
wind turbine types might require the installation of additional technical measures within the wind farm such
as battery storage or STATCOM system which can be handled from the electrical point.
Based on the findings of the Brazilian as well of the international wind energy market, the following cost,
presented in Table 3-22, should be retained for the expansion planning.
Table 3-22: Wind Power CAPEX, OPEX and potential of cost reduction
Wind power
CAPEX (MUSD/MW) 1.35
OPEX (USD/MW/a) 50,000
Potential of CAPEX reduction until 2035 (%) 30
The main reason for the decreasing costs for wind energy in recent years is the improvement of the turbine
technology which is mainly driven by competition. More efficient turbines with larger rotor size, higher hub
height and larger rated capacity have been introduced into the market at a faster pace. It is expected that
this trend will continue in the next years.
For instance, as per recent studies from BNEF 2017, a decrease of the CAPEX of more than 10% is expected
due to economies of scale and further technology improvement which will result in a decrease of the levelized
costs for electricity (LCOE), as depicted in Figure 3-56. Thus, new turbine types are focused on minimizing
turbine costs through new design and economies of scale. As a result, the typical lifecycle of a wind turbine
November 2019 129

20-26-00173
Processing No.: 15.2126.9-001.00
platform is decreasing from more than 12 years for legacy turbines to 8-9 years, depending on the manufac-
turers strategy.
Figure 3-56: LCOE for different markets (left) and expected cost decrease (right) [12]
Figure 3-57 demonstrates the size and capacity evolution of wind turbines of the last 25 years. The average
size of all annual installed wind turbines is smaller than the maximum available size and capacity and de-
pends on market and region. From 2010 to 2016, the largest increase was recorded for Ireland with an
increase of the nameplate capacity of 79% and 53% increase of the rotor diameter. Further countries with
the largest increase of the rotor diameter are Canada (47%) and Germany (36%). The largest increase in
nameplate capacity was observed in Ireland, followed by Germany (42%) and Denmark (42%), as indicated
in Figure 3-58.
Figure 3-57: Size and capacity evolution of wind turbines over time (onshore and offshore) [13]
November 2019 130

20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-58: Average rotor diameter and rated capacity increase, 2010-2016 [14] [15] [16]
When looking at the wind turbine generator, the double-feed induction generator (DFIG) held the dominant
market position in the past with approximately 2/3 of all installations worldwide in 2016. Although its share
declined in the last years, the DFIG, in combination with a partial converter, will most likely remain the leading
onshore generator technology for the next years as it presents the lowest costs generator/converter tech-
nology and is able to meet the requirements of most of the world’s grid codes. Thus, many of the new
onshore turbines of the 4 MW platforms are already designed for this technology combination. However,
new offshore wind turbine developments are almost entirely considering permanent magnet generator tech-
nologies in combination with a full converter which will generally decrease the DFIG share of all installed
wind turbines.
In general, the development is mainly driven by the requirements for grid code compliance, which have been
tighten up due to significant amount of wind energy penetration into the public grid. Currently there is no
clear trend towards full converter or DFIG / partial converter since both technologies have been further de-
veloped over the last years with some additional measures, e.g. implementation of STATCOM, batteries
they are able meet the current grid code requirements. The main conclusions of the technical trends are
summarised in Table 3-23.
Table 3-23: Summary of technology trends
Component Expectation
Rotor diameter of 140-150m will become standard and tower heights of up to 200m might
Turbine size feasible in the future. Due to competition pressure, upgrade of some earlier turbines in
and capacity parallel to the development of new products will continue. Turbine development will con-
tinue to be platform based.
DFIG concept will be the dominant technology due to price pressure. Full converter will be
Electrical
mainly used for offshore technology. With the introduction of the >5-6MW platforms, the
components
share of full converters will probably increase.
November 2019 131

20-26-00173
Processing No.: 15.2126.9-001.00
Component Expectation
Blade improvement is in a continuing process and will lead to higher efficiencies. New ma-
Blade im-
terials are introduced and additional improvements such as introduction of vortex generator
provement
or de-icing technology will become standard.
The majority of the turbines of the 4 MW platform will still be equipped with a gearbox. At
Drive train present, only a few manufacturers apply direct drive technology. Nevertheless, some out-
technology looks expect a re-increase of the direct drive concept in the mid-to long-term for turbines
above 6 MW capacity.
The trend to utilize low wind sites will lead to higher towers. Due to transport limitations as
Tower a result of bigger tower diameter, alternative concepts are currently developed such as
hybrid tower solution or large diameter steel towers.
3.6.2 Technological Evolution of Solar Power
3.6.2.1 Photovoltaics
Among solar technologies, photovoltaics (PV) is the most widely spread, cheap, environmentally friendly
and scalable technology available. PV applications range from household systems to utility scale plants in
the order of several hundreds of MW. Although the Brazilian share in the PV installed capacity worldwide is
still marginal, reduced prices of PV and further building of capacity in the country could motivate Brazil to
keep increasing its PV capacity and become a main player.
The PV industry has established itself during the last decade shifting from developed countries to emerging
economies. In 2014, the big PV market players were mostly in Asia Pacific and Europe. In the last years the
US and China have substantially increased their PV installations and it is expected that by 2020 the total
installed capacity of PV reaches beyond 350 GW, as presented in Figure 3-59.
Figure 3-59: Global installed PV capacity forecast
During the last years the PV sector experienced a substantial reduction on the specific investment cost and
the cost of the electricity generated. In the next years, the PV cost reductions are expected to be substantial
and will be mainly driven by:
· Economies of scale (larger plants);
November 2019 132

20-26-00173
Processing No.: 15.2126.9-001.00
· Technology improvements such as efficiency improvements with the corresponding reduction in

land, mounting and maintenance costs;
· Reduction of Balance of System (BoS) costs.
The reduction of BoS costs are to play a major role in the next years according to IRENA, as presented in
Figure 3-60. Until 2025 an average reduction in the total CAPEX of 57% is forecasted with a major reduction
expected to come from the BoS and thus it could be applied for all technologies and tracking systems.
However, this reduction was based on 2015 and the market has already experienced a strong price reduction
in 2016 and 2017. A conservative value for reduction until 2025 thus will be around 20%.
Beyond 2025 is difficult to foresee how costs for utility scale PV will behave. For the purpose of this document
a similar linear reduction should be assumed, resulting in a 40% cost reduction until 2035, as presented in
Table 3-24. For considering different rates of cost decrease different scenarios could be considered within
the expansion planning with a total cost reduction of e.g. 30% and e.g. 50% until 2035. One should note that
this document focusses on the following PV technologies:
· Silicon wafer-based technology: Mono-crystalline and multi-crystalline cells;
· Thin film: Cadmium telluride (CdTe) modules;
· Bi-facial PV: new technology with promising performance improvement at the module and system
level.
Figure 3-60: Global weighted average total installed costs of utility PV and potential of cost reduction [17]
Table 3-24: PV CAPEX, OPEX and potential of cost reduction
Multi-crystalline Mono-crystalline Thin film (CdTe)
CAPEX - fix
0.9 – 1.0 1.1 - 1.2 0.8 - 0.9
MioUSD/MWpeak,DC
CAPEX – single-axis
1.0 - 1.1 1.2 - 1.3 0.9 – 1.0
MioUSD/MWpeak,DC
CAPEX – dual-axis
1.1 - 1.4 1.3 - 1.7 1.0 - 1.2
MioUSD/MWpeak,DC
November 2019 133

20-26-00173
Processing No.: 15.2126.9-001.00
OPEX all tracking systems

10,000 - 18,000
USD/MWDCpeak
Potential of CAPEX reduction

20
until 2025 (%)
Potential of CAPEX reduction

40
until 2035 (%)
The performance and energy yield of PV systems can be affected by the following technology factors:
· Local conditions: in general, geographical and meteorological conditions determine the performance
of PV systems. Amongst irradiation, temperature, latitude, cloud level, air pollution and extreme
weather is the irradiation the most relevant one in order to design or locate a PV system. The re-
maining aspects can be only studied and analysed for specific project at specific locations;
· DC/AC Ratio or Inverter Sizing Factor (ISF): the relation between the DC peak installed capacity in
the PV plant and the maximum rated AC capacity of the inverter(s). The optimum DC/AC ratio should
be part of an optimization process for a particular plant seeking for the lower cost of electricity for
different ratio combinations. Although high DC/AC ratios might lead to higher CAPEX they can also
lead to lower cost of electricity and higher capacity factors depending on the particular location and
design;
· Availability and degradation: The availability of a PV plant depends directly on how reliable the plant
is and how well it is maintained. The minimum availability accepted commercially for PV systems is
of 99% throughout its plant lifetime. Due to the young age of the Brazilian PV plant portfolio, it is
expected that the availability is in a similar range as the typical international benchmark. Generally
speaking, the values given for the minimum availability accepted commercially are universal market
standard values which will apply for projects in Brazil;
· Scalability: The size of PV power plants is theoretically only limited by the land available. Inverter
sizes allow for different levels of modularity. Chiefly enlarging a PV plant will bring benefits in terms
of economies of scale i.e. reduction of CAPEX and OPEX;
· Energy Storage Systems (ESS) - Batteries: ESS for PV plants are still very capital intensive and
little experience has been gained from commercial applications in utility size PV systems. ESS
above 1 MWh are being implemented majorly using Lithium-ion batteries, however it is uncertain
that costs will drop within the next 5 to 10 years to reach economic viability for utility size systems
and therefore their implementation shall be evaluated in a case-by-case basis and mostly where
electrical grid is not available and hybridization with Diesel systems is not an option;
· Grid support: the inverter is the key component in a PV plant to meet the grid requirements. Cur-
rently, most of the inverters have Low Voltage Ride Through (LVRT) capability and flexible active
and reactive power control capabilities. Inverters also enable monitoring, decision making and con-
trol functions providing the grid with extra support. Basic grid support features of inverters are active
power curtailment, reactive power control, power factor control, voltage/frequency ride-through and
ramp-rate controls;
· Performance improvements: main performance improvements will come from improvements on the
different generations of PV systems. One important improvement already occurring is the volume
increase of second-generation PV systems, i.e. thin film, which are reaching commercial levels.
November 2019 134

20-26-00173
Processing No.: 15.2126.9-001.00
For the technologies considered in this document, efficiencies of commercial modules have reached ex-
pected values e.g. around 20% efficiency for mono-crystalline modules or 18% for multi-crystalline modules.
The one logical assumption is that in the future those record efficiencies, or premium modules will be
achieved by commercial modules in addition to further increase in efficiencies due to other technology im-
provements. Some sources consider that by 2027 it is likely to achieve average module efficiencies of about
20% for multi-crystalline modules, 26% for mono-crystalline modules and 19% for CdTe thin film modules.
Improvements on efficiency shall be translated into improvements on system efficiency and reduction of
specific costs of the PV plant due to reduction of the balance of system.Figure 3-61 shows potential increase
in system efficiency until 2020. It is assumed that this increase will follow the behavior of the efficiency
improvement of the modules as mentioned above.
Figure 3-61: Evolution of system efficiency [18]
Table 3-25 summarizes the potential effects on performance of technology improvements for the considered
PV technologies.
Table 3-25: Summary of technology improvements for different PV technologies
Module efficiency im-

11% gain 30% gain 6% gain
provement by 2027
System efficiency For a fixed system area depending on the plant’s capacity and irradiation a typical Perfor-
improvement mance Ratio (PR) can be assumed, and the power gain is proportional to the module effi-
ciency improvement.
Energy produced = PR x Irradiation on PoA42 x Area x Module efficiency
PR for Brazil can be assumed between 75 and 80% being conservative
Bifacial application Up to 30% more energy produced
3.6.2.2 Concentrating Solar Power (CSP)

CSP systems are basically steam turbine based thermal power plants, and therefore such systems share
the design and complexity inherent to thermal power systems. CSP systems obtain their thermal energy
42
Plane of Array: A combination of direct, diffuse and reflected irradiation on the solar panel.
November 2019 135

20-26-00173
Processing No.: 15.2126.9-001.00
from a solar field used to concentrate the energy in a collector element and achieve appropriate steam
temperatures to operate the steam turbine.
CSP systems use combinations of mirrors or lenses to concentrate direct beam solar radiation to produce
forms of useful energy such as heat and electricity and are divided into two main groups namely point-
focusing and line-focusing. CSP systems are not able to use radiation that has been diffused by clouds, dust
or other factors being more efficient in areas with high shares of clear sky. CSP currently are divided into:
· Parabolic Trough (PT);
· Central Receiver tower (CR);
· Linear Fresnel (LF);
· Solar dishes (e.g. Stirling dish).
Although all technologies are to some extent still being commercialized, PT and CR systems are the ones
that have reached a competitive level of commercialization, prices and competitivity.
The main advantage of CSP systems is the implementation of a Thermal Storage System (TES) allowing
the system to be flexible to deliver firm energy for short periods or even operate 24 hours with large TES
designs, thus dispatchability is achieved. The TES system most widely designed and commercialized is the
indirect two tank molten salt system. While PT uses the molten salt of the TES to heat a thermal oil which in
turn creates steam, the CR systems use typically the molten salt directly reaching higher temperatures of
the steam and thus better efficiencies
CSP cost varies considerably according to specific conditions such as solar irradiation potential at the project
site, development cost, technology type and TES capacity, and regulatory framework. Due to the fact that
CSP has not experienced the same level of implementation than other renewable technologies the industry
expects further achievement of the learning curve of CSP and further potential reduction of its main compo-
nents. For the purposes of this document a lower potential of cost reduction than for PV will be assumed for
CSP, as presented in Table 3-26, since PV market will continue to grow at much bigger steps than CSP.
Table 3-26: CSP CAPEX, OPEX and potential of cost reduction
Parabolic trough Central Receiver System

(Molten Salt)
CAPEX MioUSD/MWpeak 4.5 – 6.0 5.0 - 6.5
OPEX all tracking systems USD/MWDC 45. 0 47.0
Potential of CAPEX reduction until 2025 (%) 15.0
Potential of CAPEX reduction until 2035 (%) Linear
3.6.2.3 Conclusions for the expansion planning

This chapter has analyzed existing data and projections of the solar industry from different sources such as
IRENA, NREL and the Ministry of Mines and Energy of Brazil in order to define the technologies, configura-
tions and cost developments which most likely could be applied to Brazil within the next years, i.e. for the
purposes of expansion planning until 2035.
In general, electricity generation at utility scale level based on solar technologies has positioned in the global
markets and is nowadays a technically and economically feasible alternative competing with conventional
power generation.
November 2019 136

20-26-00173
Processing No.: 15.2126.9-001.00
The information analyzed lead to the following conclusions for the expansion planning:
· Despite of the important advantages of Concentrating Solar Power (CSP) equipped with Themo-
Electrical Energy Storage (TES) in terms of dispatchability the drastic price reduction experienced
by Photovoltaic Power (PV) in recent years alongside with its simplicity have made this technology
the obvious option amongst the solar technologies.
· CSP systems are currently developed and implemented in those regions of the world with very high
Direct Normal Irradiation (DNI) levels such as Chile, South Africa, North African countries or in the
Middle East. It is expected that this tendency will go on until electricity networks require dispatchable
sources, however by that time it is possible that battery systems have also reached competitive
prices as part of a PV system. Therefore, it is expected that the market in Brazil will be still imple-
menting PV instead CSP for the years to come and the moment in which CSP could rival PV in
Brazil is rather uncertain.
· In case of implementation of CSP due to either dispatchability requirements or to a political decision
towards gaining technical capability in CSP, it would be advisable to consider Parabolic Trough (PT)
with thermal oil as Heat Transfer Fluid (HTF) and molten salt based Central Receiving Systems
(CRS) both with medium to large two-tanks molten salt systems. These configurations with installed
capacities above 100 MW will be advisable to achieve competitive prices against other conventional
or other renewable sources.
· For the case of competitive PV plant auctions at utility scale level it is expected that currently and
for the years to come projects will be based on crystalline and thin film (CdTe) technologies as these
are the technologies that can bring lower prices while guaranteeing bankability.
· The use of tracking systems will depend on the local conditions of a specific project. For the case of
Brazil there is a clear tendency towards the use of single-axis trackers to improve the plant’s capacity
factors. Fixed mounting systems could be competitive in areas with high Global Horizontal Irradiation
(GHI).
· Implementation of batteries for PV for the purpose of generation shifting or to firm up the PV capacity
(for production at peak load) will not be economically viable within the planning period up to 2035.
However, the development of battery prices is to be monitored closely and such systems could be
included in the expansion planning if a substantial price reduction is achieved in the next years.
· The use of bi-facial PV modules is in its first steps. Prices and actual performance is not yet known
by the industry, furthermore there are not enough suppliers to cause sufficient competition at inter-
national competitive auctions. This development shall be, however, monitored, and included in the
expansion plan if it becomes a market standard.
· CAPEX for PV is expected to keep dropping until 2025 by 20%. This document assumes that the
reduction will be likely maintained linear until 2035. This assumption needs to be monitored and
reviewed when updating the expansion plan as it is difficult to predict the market as of today.
· In terms of performance of PV modules, it is expected that by 2027 mono and multi-crystalline, as
well as CdTe thin film modules will achieve the record efficiencies in the laboratories and these
efficiencies will become the market standard.
3.6.3 Technology resources for mitigating VRE variability impacts

The support by policies pointed on the enhancement of Variable Renewable Energy (VRE) in terms of energy
supply security and sustainability together with the steady technological and cost-reduction development are
the key drivers for the continuously increasing share of VRE in the power sector worldwide.
November 2019 137

20-26-00173
Processing No.: 15.2126.9-001.00
In a global term, there is a high increase of PV and Wind grid-integration in the world with distinct challenges
and conditions for the electrical power network. This change in the power sector requires innovative ap-
proaches to power system planning, system and market operations, as well as in regulation and public policy.
The following challenges and conditions for the electrical power network can be highlighted:
· Power produced by VRE generators is dependent on the weather resource of sun and wind (not
schedulable);
· Utility scale VRE are generally installed in remote geographical locations where their power output
is highest, occurring large energy flows between load centres and points of generation;
· VRE generation are generally installed in the MV and lower voltage areas, as result a bi-directional
power flow is given where a higher transfer capacity is requested by the distributed generation sites;
· Most of the wind turbine generators and all PV generator-units are converter connected to the grid,
so called inverter-based generators. The substitution of conventional generation units trough in-
verter-based generators reduces system inertia which is an indicator for the sensitivity of the fre-
quency to imbalances.
Alternative solutions to these aspects can be market-based, regulatory, technological or procedural in na-
ture. The technological solutions deal with means as grid expansion, storage or automatization, the opera-
tional ones are e.g. load-management or infeed-management.
This section aims to provide an overview about different technologies, that can be utilized for mitigating the
effects of the variability of VRE infeed. These technologies are clustered according to their location in the
power system in supply level mitigation, transmission level mitigation and demand side management.
Figure 3-62 presents a simplified scheme of demand’s flexible adjustment to meet supply. The scheme is
divided by supply level, with mature generation technologies, grid-friendly VRE and VRE production fore-
casting, by transmission level with investment in new transmission capacity, HVDC and smart grid technol-
ogies, by demand response and by energy storage.
Figure 3-62: Simplified Scheme of Demand’s flexible Adjustment to meet supply [19]
3.6.3.1 Supply level mitigation

3.6.3.1.1 Conventional thermal and hydro power plants
Traditionally, thermal and hydro power plants with synchronous generators provide sufficient inertia to the
system, so that the control power capacities are able to maintain the frequency within its defined bandwidth
November 2019 138

20-26-00173
Processing No.: 15.2126.9-001.00
for a stable operation of the grid. Besides inertia, it is thus important that the connected power plants provide
enough flexibility in ramping, which can be used as control power to bring back the frequency to its nominal
value after a fault in the system.
Large amount of variable renewable energy (VRE) generation in the dispatch triggers a requirement for
operating reserve requirements among the conventional thermal and hydro power plants, which are able to
react to changes in the VRE power output. For this, highly flexible thermal and hydro power plants are
needed, which also provide the inertia needed to keep the frequency controllable.
If ramping capabilities (for the provision of frequency response) are too small and in particular start-up times
are too long, then there is the possibility to retrofit existing thermal power plants (TPPs) or also hydro power
plants (HPPs) with battery electricity storage systems (BESS) to increase their flexibility and to reduce the
thermal stress when operated in a flexible way, for which, originally, they have not been designed for, as
exemplified in Figure 3-63. Due to the absence of moving parts, BESS only rely on the reaction time of their
cell chemistry which enables them to react within milliseconds.
The link ed image cannot be displayed. The file may have been mov ed, renamed, or deleted. Verify that the link points to the correct file and location.
Figure 3-63: Optimized performance and new opportunities for grid and ancillary services [20]
Brazil has the advantage that it is characterized by a large number of hydro power plants with reservoir,
which are very flexible in operation. The water storage can act as a “large battery” towards changes in the
output of VRE. Since numerous HPPs have short penstocks, they are also able to provide very fast acting
frequency response (primary reserve). For this reason, exploiting the operational flexibility potential of the
already existing HPPs should have first priority for the integration of large-scale VRE. If their potential, in
particular for the provision of primary reserve is not sufficient, then the amount of primary reserve being able
to be sourced from TPPs should be studied. Steam turbine generators with a boiler (such as coal, oil or even
biomass fired power plants) can provide primary reserve with a limited duration by reducing or increasing
the steam pressure within the boiler for a couple of minutes, after which less fast-acting plants can take over
and release the steam turbines to restore their original steam pressure.
3.6.3.1.2 Grid-friendly VRE

Renewable energies, in particular inverter-based generation (solar and wind), are expected to continue
growing at even higher rates. The replacement of conventional synchronous generation by inverter based
variable renewables introduces a challenge for frequency management in short time scales due to decreas-
ing inertia. New units replacing retiring conventional units have different constraints with respect to the pro-
vision of operating reserve and balancing. Increasing variability and uncertainty drive the need for increased
operating reserve and balancing in the future. Figure 3-64 presents the trends affecting supply and needs for
operating reserve.
November 2019 139

20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-64: Trends affecting supply and needs for operating reserve [21]
Integrating grid friendly VRE into the system requires the implementation and use of advanced power elec-
tronics, or smart inverters, that provide grid services like voltage and frequency regulation, ride-through,
dynamic current injection, and anti-islanding functionality. Advanced inverters can modify their output to
actively support the grid in many innovative ways. Inverters can continue to operate within wider ranges of
frequency and voltage, can vary their power factor, and can enhance the power quality and performance of
the local grid. In terms of physical hardware, advanced inverters and standard inverters do not differentiate
from each other, the main difference is imposed by the software that manages the advanced functionality.
Development and cost trends show that more solutions are becoming economically viable and inverters are
being designed with those challenges in mind:
· Advanced inverters support fast frequency control, thus simulating an artificial inertia. This is still not
as good as the instant change of active power offered only by the traditional synchronous generator;
· Advanced communication networks are being considered for optimal energy planning to help with
the stochastic nature of VRE, and to allow real time data exchange between TSOs and DSOs for a
better operation of the grid;
· With a decrease of manufacturing costs, better semiconductor technologies such as SiC and GaN
are being adopted.
Due to high competition in the market, especially by the Chinese, prices are expected to continue to drop in
the upcoming years and inverters are expected to offer more and more functionalities to mitigate inverter-
based generation related issues.
3.6.3.1.3 VRE production forecasting

As wind and solar power generation will have a larger share in electrical generation in the future, there is a
need to accommodate their variable nature in dispatching activities. An appropriate forecast of wind and
solar energy availability and power potential production can enhance power system operations in many
ways, improving unit commitment and dispatch efficiency, improving reliability issues or optimising market
clearing decisions.
There are multiple methods and sets of methods used in forecasting systems by different private, public or
academic institutions. An overall forecasting generation procedure is presented in Figure 3-65.
November 2019 140

20-26-00173
Processing No.: 15.2126.9-001.00
Figure 3-65: Forecasting value creation chain
It is becoming important to efficiently utilize the information provided by advanced forecasting models.
1. Unit commitment: Forecasting are used in the unit commitment process to help avoid costs and
inefficiencies due to unnecessary starts and stops of thermal generators, forecasting. Forecasted
uncertainty from wind and solar power generation makes it relevant to consider also alternative
stochastic formulations of in the unit commitment problem. Such models can play a significant role
in reducing costs while maintaining system security under increased uncertainty and variability. In
addition, it is also important to consider the close interaction between operating reserve require-
ments and unit commitment policy.
2. Dispatch: Short-term solar and wind power forecasts can support the control of the generation from
solar and wind power systems and enforce curtailment of power generation in situations where this
is needed, either from an economic or reliability perspective. It is used in real-time dispatch and
market-clearing decisions. It is also useful for increasing grid reliability, identifies the risk and poten-
tial for rapid and sustained change in power output within a specific time interval.
3. Operating reserves: Appropriate forecasting can support to address the optimal determination of
reserve requirements under high penetration of wind and solar power. Use the uncertainty infor-
mation from the forecast in the operating reserves requirements can be determined more precisely,
frequently and closer to real-time.
3.6.3.2 Transmission level mitigation

3.6.3.2.1 Investment in new transmission capacity
Renewable power penetration from remote location constrained generators to the grid is facing a set of
inherent, regulatory, economic and technical challenges.
An extension of transmission lines linking different areas with different wind and solar power availability and
different load centres can increase the utilisation of VRE energy and the network stability. The electricity
generation by VRE from different locations will allow to improve the system load factor of generated electric-
ity since the availability of VRE in identified locations differs. On the other side, the electricity demand pat-
terns differ in different load centres as well, so an interconnection of different load centres with different VRE
generation parks will improve the total system performance increasing the utilisation of renewable energies.
3.6.3.2.2 High Voltage Direct Current (HVDC)

High voltage direct current (HVDC), which has traditionally been used to carry electricity in massive quanti-
ties over large distances, is not a new technique, but has regained interest due to the challenges of connect-
ing renewables to the grid and the emergence of the new VSC (voltage source converter) technology.
November 2019 141

20-26-00173
Processing No.: 15.2126.9-001.00
The integration of VRE is creating a demand for grid expansion and grid reinforcement. HVDC is offering a
possibility to answer these demands (for larger transmission distances) with overall lower transmission sys-
tem cost compared to HVAC and lower transmission losses while at the same time removing stability related
limitations to the amount of power or transmission distance.
VSC technology is increasingly becoming economically viable as losses are decreasing and expected to
match the ones of the LCC and voltage/power levels are increasing. This advancement is accelerated by
the demand to interconnect the DC transmission lines and thus moving from individual point-to-point HVDC
systems to multi-terminal systems and in the future to a HVDC meshed grid.
As costs continue to decrease and technologies becoming better, it is evident that HVDC will play an even
bigger role in the reinforcement of the current AC grids to help with stability and congestion issues created
by the increased penetration of inverter based generation.
3.6.3.2.3 Wide Area Monitoring System (WAMS)

Due to general increase of renewable energies share, the transmission and distribution lines have to cope
with constantly decreasing stability. The voltage, frequency and current fluctuations demand for better net-
work controllability and faster reactions to occurring differences. If this is not guaranteed, several outages
and blackout could be the consequence. For overcoming this, Brazil is currently in the process of installing
a wide area monitoring system (WAMS) with in the first phase phasor measurement units (PMUs) in 31
substations and 181 line terminals and two operating centers.
WAMS is a technology developed to minimize the probability of the described forced outages controlling the
complete network to enable intelligent and quick mitigation of discrepancies. Despite the initial investment
costs, WAMS can reduce the overall costs for transmissions operators and utilities due to less maintenance
costs or reduced ancillary services needs and costs. WAMS will enable prediction of outages and prevent
them in advance by appropriate strategies. However, the most important point is that WAMS will not be
sufficient to enable VRE integration alone, there is a need for further smart grid technologies.
3.6.3.2.4 Smart Grid technologies
There are various levels of smart grid technology application, which all have to be considered in context with
the appropriate grid behind. There is not the one solution for a smart grid, it always should fit individually to
existing facilities. Nevertheless, some characteristics of smart grid applications are important:
· Self-healing: it detects, analyzes, responds and restores grid failures rapidly and automatically;
· Stable Network: controls and monitors fluctuations in its grid;
· Provide power quality needed by 21st-century users: it adapts itself to the modern needs of
industry and small costumers;
· Lower capacity needed: through demand side management less capacity is needed;
· Network automation: high grade of network automation and observability;
· Telecommunication systems: reliable telecommunication infrastructure based on several telecom-

munication technologies;
· Handle stochastic demand and respond to smart appliances;
· Provide self-correction, reconfiguration, and restoration;
· Protocols, standard and smart algorithms to improve smart communication and transportation
systems.
November 2019 142

20-26-00173
Processing No.: 15.2126.9-001.00
In general, these characteristics are required due to the challenges caused by the rising share of VRE in the
operation of the grid. The following subsections will give an overview of common technologies which support
the integration of VRE in the future: dynamic line rating, digital substation technology, flexible AC transmis-
sion systems and fault current limiters.
3.6.3.2.4.1 Dynamic line rating (DLR)

The main idea behind dynamic line rating is to increase the efficiency of transmission lines without physical
extension and safely use existing transmission lines’ transmission capacity based on real conditions in which
power lines operate. The parameters which influence the capacity of transmission lines are mostly environ-
mental dependent, like wind speed and wind direction, ambient temperature, solar radiation, line tempera-
ture, line tension, line sag and line current. For using the untapped potential of transmission lines, measure-
ments of respective parameters are carried out. Real-time data are used and dynamic matching to just in
time can be made possible.
The main benefits of the DLR are as follows:
· Increased transmission system efficiency (depending on environmental parameters);
· Decreased or deferred capital costs through optimized utilization of existing assets;
· Decreased system congestion costs;
· Increased situational awareness and operational flexibility of the transmission system;
· Fast option for increasing the transmission capacity of existing lines.
Whereas the main disadvantages of the DRL are as follows:
· Non-negligible costs;
· Can be maintenance intensive;
· Increased Ohmic losses when operated above conventional transmission capacity (close to ther-
mal limit).
The benefits of dynamic line rating (DLR) include but are not limited to improved system reliability and safety,
reduced and or deferred capital expenditure and increased efficiency of generation resources. It can in-
crease the transmission capacity and mitigate the variability. With the purpose of sustainable increase of
transmission capacity, mostly building a new line of exchanging the conductor type will be the more eco-
nomical solution. However, with respect to the controllability of the line, in particular for the integration of
large scale VRE, DLR can be an interesting option. It is advised to follow the DLR project currently in imple-
mentation in Brazil, closely and to evaluate the experiences made. Based on this return of experience it can
be decided whether DLR can bring benefits it implemented on other lines in the Brazilian grid.
3.6.3.2.4.2 Digital substation (DS)
The digital substation does not differ in its main task from the conventional substation. Hence digital substa-
tion includes additional digital solutions to the conventional main task, by adding new “smart” components
and processes, a positive impact to availability, safety and cost-effectiveness is achieved.
Advantages of DS can be characterized by providing a higher operational flexibility and allow operation
closer to over injection to the grid or overvoltage. Furthermore, cost-benefit analyses show clear cost ex-
penses reduction by risk- based or condition-based maintenance and fewer constraints by preventive meas-
urements for outages. The concept enables an improved asset utilization by stress-dependent determination
of remaining life. There is also potential for reduction in CAPEX, but these are highly project specific [22].
The impacts on performance of the general power system especially on transmission lines are as follows:
· Greater flexibility will be enabled through closer to limits with transitional overloading. Enhanced
asset utilization is enabled which safes costs and further grid enlargement which triggers less envi-
ronmental disturbance;
November 2019 143

20-26-00173
Processing No.: 15.2126.9-001.00
· Permanent monitoring provides large data base enabling more efficient maintenance which again
enables cost savings by building the ground for enlargement of the life cycle of asset equipment;
· Greater availability and safety of substations allows the constant use of connected lines and in-
creases the transmission rate of connected VRE;
· DS can significantly reduce operation costs and risk.
The digital substation technology represents a step-change innovation and is close to the market break-
through. It represents a smart grid enabling technology necessary for a power infrastructure based on re-
newable energy infeed. Digital substations are critical for the efficient integration or renewable generation
by centralized monitoring of T&D networks and automated grid management. Digital substations are char-
acterized by providing a higher operational flexibility and allow operation closer to design limits. Automated
grid control and monitoring of T&D networks enable the further integration of renewables. For this reason, it
is advisable to include digital substation technologies in the grid planning as an option and to evaluate its
techno-economic performance from case to case.
3.6.3.2.4.3 Flexible AC Transmission Systems (FACTS)
Flexible AC Transmission System (FACTS) devices play a key role for integrating VRE in a power system.
They can provide fast continuous control of power flow in the transmission system by controlling voltages at
critical buses, by changing the impedance of transmission lines, or by controlling the phase angles between
the ends of transmission lines. With these characteristics, as well as given their technical maturity, they are
suitable to be employed in the Brazilian grid, wherever their functions are needed.
Referring to the influence on power system today, the following key points are listed:
· FACTS stabilize transmission systems with increased transfer capability and reduced risk of line
trips;
· They provide the required quality of supply by reducing voltage dips, frequency variations or the loss
of supply;
· FACTS installation provides the flexibility for future extensions;
· With FACTs a transmission line capacity requirement can be reduced by actively controlling the
power flows in a certain range.
3.6.3.2.4.4 Fault current limiters (FCLs)
The application of a fault current limiter is one of the promising solutions to the stability and security issues
of power systems. FCLs have been implemented for different purposes such as stability enhancement, pro-
tection improvement, fault current reduction and fault ride through capability enhancement. The cost reduc-
tion of power system components is observed to be significant as well.
Fault Current Limiters can improve the transient stability of the power system by suppressing the level of
fault currents in a fast and effective manner. The number of outages is reduced and again a better asset
utilization is provided. Because transmission and distribution lines are enabled to increase the level of power
transmission as well as their reliability, no new lines have to be built. This can be cost effective, especially
considered in a long-term context.
For the interconnection of fixed speed wind bridge fault current limiter (BFCL) or modified bridge type fault
current limiter (MBFCL) have been implemented. For the interconnection PV systems series dynamic brak-
ing resistor (SDBR) and superconducting FCL have been employed. Given the maturity of FCLs, it is advis-
able to consider them for utilisation in grid regions with large amounts of VRE generation.
November 2019 144

20-26-00173
Processing No.: 15.2126.9-001.00
3.6.3.2.5 Electricity storage systems (ESS)

The rapid growth of infeed by wind power and solar photovoltaic drives the need for high-capacity as well as
rapid-response energy storage technologies to smoothen out highly variable generation. Storage can help
to limit the need for investments in grid capacity, to reduce operation costs of generation facilities and to
increase the reliability of the supply. With its capability in providing a range of energy services such as grid-
stability ancillary services and long-term storage, hydropower’s role in renewable-based energy systems is
becoming increasingly important.
ESS technologies will play important but different roles towards the energy system transition in the future.
The analysis shows that particularly for battery electricity storage systems (BESS) technologies the cost will
reduce significantly and the performance will increase in the coming years. No single storage system can
meet all the criteria to become the ideal energy storage system. So, the relevance of each ESS technology
depends primarily on the type of application:
· For shorter response time (up to 1 s) and discharge duration (minutes to few hours), for power quality
improvement, for smoothing intermittency of VRE or voltage stabilization the promising ESS tech-
nologies are the batteries, SMES or fuel cells;
· For medium response time (minute) and discharge duration (hours-day), for energy management,
for time/peak load shifting the promising technologies are Large-scale batteries, CAES, fuel cells or
TES;
· For a response time (minutes) and discharge duration (hours-days), for emergency back-up or sea-
sonal energy storage the promising ESS technologies are PHS, CAES or TES.
In general, those ESS that having the following characteristics will likely be more predominant as costs fall;
quick charge time, high power density, high energy efficiency, high reliability, high cycle times, low carbon
footprint and low disposal issues.
Energy storage systems have the following specific applications for generation and grid side support:
· Provision of ancillary services: Battery technologies are playing a more significant role for reserve
markets providing ancillary services to maintain the balance between load and generation, providing
primary frequency response;
· Power generation output smoothing (ramp-rate control): ESS can be used to smooth out fluctuations
in frequency and voltage, caused by the inherently variable nature of VRE power output, to keep the
system stable;
· Seasonal electricity storage: ESS can also be used on a larger time level to cater for seasonal com-
pensation of electricity production/demand;
· BESS as grid forming unit in isolated hybrid systems with high share of VRE: One of the most eco-
nomical applications of BESS. In such hybrid systems, BESS take over the function of grid forming
unit, which monitors and controls the frequency as well as the voltage and correspondingly needed
reactive power supply.
3.6.3.3 Demand response for load management

In general, the principle behind Demand Response (DR) is enhanced flexibility of demand to follow genera-
tion. The demand response comprises a broad range of automated and manual initiatives to modify electric-
ity consumption in reaction to demand-supply imbalances or exceptionally high-power prices. When properly
implemented and integrated in the system, demand response incorporated in a system with a high share of
variable generation sources has the potential to provide services that are equivalent (or even superior) to
November 2019 145

20-26-00173
Processing No.: 15.2126.9-001.00
services provided by traditional resources. Besides that, demand response can provide reserve services,
dynamic system regulation and load-following capabilities delivering during any hour of the year.
An overview of the co-existence of generation, storage and demand response is provided in Error! Refer-
ence source not found..
Figure 3-66: Services provided by Generation, Storage and Demand Response [23]
The following advantages can be attributed to DR:

· Cost-efficiency / economic advantages;
· Increased grid stability;
· Improved market conditions;
· Harnessing beneficial power generation;

· Emission reduction;
On the other hand, the following existing disadvantages can be attributed to DR:
· Regulatory challenges: minimum load requirements, DR does not qualify as non-spin reserve in
some markets;
· Limited adherence due to unknown consequences of impact on (industrial) processes and potential
benefits;
· Challenge on the Information & Communication Technology (ICT) side;
· Challenge of availability between DR resources and electricity services;
· Functioning market environment providing equal opportunities for all participants is required;
· Market platforms and regulatory environment are mainly presented in field trials;
· Oversupply of flexibility in many markets, limited volatility in whole-sale market.
A Demand Response (DR) program has been launched in the North and North-East regions of Brazil in
January 2018. Its objective is to shift load of the previously qualified consumers to replace thermal power
generation outside the merit order in order to both achieve reliability of service while stabilizing the power
system, responding to the increase of fluctuating power sources and keeping cost of service at a reduced
November 2019 146

20-26-00173
Processing No.: 15.2126.9-001.00
level. The DR pilot program has set to go until June 2019. If until then, positive experiences will have been
made, similar programs may be established in other regions or on national level to foster the benefits of DR
in large scale.
In contrast to supply and transmission side flexibility options, for which clear rules can be set, the TSO has
only limited control on the availability of demand side flexibility options, because potential DR participants
cannot be forced to. This said, demand side flexibility options might be even more economical compared to
supply and transmission side flexibility options.
However, if the potential DR participants have an understanding of the character of their electricity demand
as well as of the flexibilities, it can provide to the operation of the grid and also on the financial benefits they
can have from the participation, then there will also be solid supply of demand side flexibility capacity, which
the TSO can rely on in its task to cater for a stable grid operation.
3.7 Capacity Building

Fostering learning and innovation at the organization and corporate levels is a key aspect of the technical
cooperation between GIZ and the Ministry of Mines and Energy (MME). Within the framework of this specific
assignment, this objective was accomplished through a set of initiatives summarised as follows:
· Active participation of the Beneficiaries (EPE and ONS) on the project execution through the partic-
ipation in the Working Groups;
· Knowledge transfer sessions during the Workshops;
· Virtual and classroom training sessions;
· Joint publication (EPE, ONS and the Consultant) of at least one scientific/technical paper cover-
ing the results of the study in an accredited conference (i.e. CIGRE, IEEE, etc.).
The capacity building also aims at providing structuring effects and extending the outreach of new and inno-
vative approaches. In this sense, the Consultant also proposes a set of initiatives to foster the culture of
knowledge sharing within the EPE and ONS organizations.
A more detailed description of the aforementioned initiatives is given in the sequel.
3.7.1 Working Groups

The success of this assignment is strongly linked to the Working Groups (WGs) proposed by the Consultant
for the project execution. The WGs are constituted by a group of experts, composed of representatives of
the Consultant, EPE and ONS, interacting together to achieve the specified goals of the project.
Within the framework of this project, the Consultant proposed the following WGs to be set up:
· WG 1: Technical Regulation Studies (Product 1);
· WG 2: Energy Studies (Product 2);
· WG 3: Power System Studies (Product 3);
· WG 4: Methodology Studies (Product 4);
· WG 5: Technology Studies (Product 5).
The Working Groups consisted of regular interactions between the Consultant and the Beneficiaries on the
execution of each of the five main products of the assignment. In the frame of these regular interactions, the
Beneficiaries learned from the Consultant on the methodologies applied, the ways of working, the results
interpretation and reflected about their current practices through these knowledge exchanges. On the other
November 2019 147

20-26-00173
Processing No.: 15.2126.9-001.00
side, the Consultant learned from the Beneficiaries on the current practices related to the operation and
planning of the Brazilian power system and reflected on their understanding of the available best practices.
These knowledge exchanges ensured that the methodologies employed by the Consultant for the execution
of the different products were in line with the vision and expectations of EPE and ONS in the frame of this
project.
The proposed modus operandi of the WGs was the following:
· Regular meetings between the WG members:
o Schedule: one meeting every 4 weeks;
o Format: physical or virtual meetings with duration of maximum 4 hours;
o Content: overview of the ongoing work and discussion about specific topics or blocking
points for the execution of the product.
· Ad-hoc meetings between WG members:
o Schedule: defined according to the needs;
o Format: physical or virtual meetings;
o Content: defined according to critical needs of the project.
3.7.2 Workshops
The project included a set of structured workshops making possible close interactions between the Benefi-
ciaries and the Consultant. The workshops were organised under the form of forum for presenting and dis-
cussing the work progress of the different products, the methodologies applied by the Consultant and the
obtained results and derived conclusions and recommendations. Each workshop included a time dedicated
for structured discussion on a given topic corresponding to the project activities.
3.7.2.1 Workshop 0
Day 1: Kick-off meeting
The objectives of WS0-D1 are the following:
· Introduction of the Project;
· Institutional presentation - EPE;
· Institutional presentation – ONS;
· Presentation of the Consultant;
· Presentation of the general overview of the Project;
· Presentation of the general overview of the methodology;
· Introduction of Product 1: Technical Regulation Studies;
· Introduction of Product 2: Energy Studies;
· Introduction of Product 3: Power System Studies;
· Introduction of Product 4: Methodology Studies;
· Introduction of Product 5: Technology Studies;
· Introduction of the capacity building;
· Presentation of the daily project routine;
· General Discussion.
November 2019 148

20-26-00173
Processing No.: 15.2126.9-001.00
3.7.2.2 Workshop 1
Day 1: Products 1 and 5 (intermediate)
· Presentation of the project management;
· Presentation of the capacity building;
· Presentation of Product 1: Technical Regulation Studies;
· Presentation of Product 5: Technology Studies;
· Presentation of Product 2: Energy Studies;
· Presentation of Product 3: Power System Studies;
· General Discussion.
3.7.2.3 Workshop 2
Day 1: Products 1 and 5 (intermediate)
· Presentation of intermediate results by the Consultant;
· Presentation of current challenges from the perspective of EPE and ONS with respect to the subject
of the product;
· Collection of feedback from EPE and ONS;
· Agreement on the guidelines for the finalization of the product.
Day 2: Products 2 and 3 (kick-off) & plenary session

· Presentation of the methodology to be applied for Products 2 and 3;
· Discussion of the input data and main assumptions for the study;
· Setting up of working group activities for Products 2 and 3;
· Plenary session with the Steering Committee members to discuss project management aspects.
3.7.2.4 Workshop 3
Day 1: Products 1 and 5 (final)
· Technical Regulation Studies;
· Presentation of draft final results by the Consultant;
· Discussion on the results;
· Guidelines for the finalization of the product.
Day 2: Product 2 (intermediate)

· Presentation of Product 2 (intermediate);
· Introduction;
· Candidate projects for the expansion planning;
November 2019 149

20-26-00173
Processing No.: 15.2126.9-001.00
· Economic and reliability criteria;

· Generation expansion: methodology and preliminary results;
· Operation policy: methodology and preliminary results;
· Probabilistic dynamic reserves;
· Transmission network reinforcements;
· Detailed simulation of the energetic operation (hourly resolution, with full network model);
· Discussion on next steps for the finalization of the Product.
Day 3: Product 3 (intermediate)

· Presentation of Product 3 (intermediate);
· Introduction;
· Model conversion overview;
· Dependencies with Product 2;
· Criteria and methodology for the power system studies;
· Modelling of VRE: wind, solar and DG;
· Discussion on next steps for the finalization of the Product.
3.7.2.5 Workshop 4
Day 1: Product 2
· Presentation of the methodology;
· Presentation of the main assumptions;
· Presentation of intermediate results;
· Discussion of the next steps to finalize the product.
Day 2: Products 3 and 4

· Presentation of the methodology of Product 3;
· Presentation of the main assumptions of Product 3;
· Presentation of intermediate results of Product 3;
· Discussion of the next steps to finalize the Product 3;
· Discussion of the objectives and guidelines for the execution of the Product 4;
· Steering Committee meeting.
3.7.2.6 Workshop 5
Day 1: Products 1, 2, 3 and 5
· Presentation of the key messages, conclusions and recommendations of Products 1 and 5;
November 2019 150

20-26-00173
Processing No.: 15.2126.9-001.00
· Presentation of final results of Product 2 - Expansion planning;

· Presentation of final results of Product 2 - Hourly operation;
· Presentation of the methodology of Product 3;
· Presentation of the draft results of Product 3.
Day 2: Product 4
· Presentation of the methodology of Product 4 - Capacity expansion and energy studies;
· Presentation of the methodology of Product 4 - Power system studies and database;
· Presentation of the concluding remarks;
· Steering Committee meeting.
3.7.3 Training Sessions

Sharing of technical knowledge between the Consultant and the Beneficiaries also took place during dedi-
cated training sessions, organized in the form of classroom or virtual training sessions.
The list of the training sessions topics is given in Table 3-27.
Table 3-27: List of proposed topics for the training sessions
Training Topic
1 - The Role of Grid Codes for VRE Integration
2 - Technological Aspects of VRE Integration
2.a - Wind and solar PV technology overview (emphasis on "grid-friendly" control functions)
2.b - Energy Storage Technologies (BESS, PSP, Hydrogen)
2.c - The economics of VRE technologies (LCOE, cost evolution, etc.)
2.d - Technologies for Enhancing Transmission System Flexibility (FACTS, HVDC, PST, BESS, DLR,
WAMPAC)
3 - Operating Reserves and their Impacts on VRE Integration

3.a - Introduction to Operating Reserves
3.b - Impact of VRE on Primary Reserves
3.c - Impact or VRE on Secondary Reserves
3.d - Dynamic Reserves
3.e - Methodologies for Operating Reserve Sizing
4 - Selection of Operating Conditions for Power System Studies
4.a - Selection of Operating Conditions for Reactive Power Sizing
November 2019 151

20-26-00173
Processing No.: 15.2126.9-001.00
4.b - Selection of Operating Conditions for Transient Stability
4.c - Selection of Operating Conditions for Frequency Stability
4.d - Selection of Operating Conditions for Small-Signal Stability
5 - Wind, Solar PV and BESS Modelling for Static and Dynamic Analyses
5.1 - Modelling for Load Flow and Contingency Analysis
5.2 - Modelling for Short-Circuit Analysis
5.3 - Modelling for Dynamic Analysis
6 - Generation and Transmission Expansion Planning

6.a - Modelling the operation of the components of a power system: hydro, thermal, wind, solar, bio-
mass, etc.
6.b - Modelling uncertainties in the energy production (correlations between inflows and wind, load and
solar, etc.)
6.c - Operating reserves: probabilistic reserves and its modelling in generation scheduling problems
6.d - Expansion planning criteria
6.e - Transmission network performance analysis
6.f - Candidates for generation and transmission expansion
6.g - Modelling of candidates for transmission expansion
6.h - Optimal transmission expansion models
November 2019 152

20-26-00173
Processing No.: 15.2126.9-001.00
4 REFERENCES
[1] International Electrotechnical Commission - IEC, IEC 61400-27-1 Electrical simulation models – Wind
turbines, Edition 1.0 2015-02 ed., Geneva, 2015.
[2] Western Electricity Coordinating Council (WECC) Renewable Energy Modeling Task Force, Generic
Solar Photovoltaic System Dynamic Simulation Model Specification, 2012.
[3] D. R. A. G. J. C. B. I. ALVAREZ-FERNANDEZ, “Parameterization of aggregated Distributed Energy
Resources (DER_A) model for transmission planning studies,” CIGRE Science and Engineering, vol.
15, pp. 158-168, 2019.
[4] International Renewable Energy Agency (IRENA), “Scaling up Variable Renewable Power: The Role
of Grid Codes,” IRENA, Abu Dhabi, UAE, 2016.
[5] Operador Nacional do Sistema (ONS), “Procedimentos de Rede - Vigentes,” ONS, [Online].
Available: http://ons.org.br/paginas/sobre-o-ons/procedimentos-de-rede/vigentes. [Accessed 10 June
2019].
[6] [Online]. Available: www.iea.org/statistic.
[7] E. W. Kimbark, Power System Stability - Volume III - Synchronous Machines, Wiley, 1956.
[8] P. Kundur, Power System Stability and Control, New York: McGraw-Hill, 1994.
[9] North American Electric Reliability Corporation (NERC), “Integrating Inverter-Based Resources into
Low Short Circuit Strength Systems,” NERC, 2017.
[10] Operador Nacional do Sistema Elétrico (ONS), “Submódulo 23.3 - Diretrizes e Critérios para Estudos
Elétricos,” ONS, Rio de Janeiro, 2018.
[11] Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ), “Analysis of System Stability in
Developing and Emerging Countries: Impact of Variable Renewable Energies on Power System
Reliability and System Security,” GIZ, 2013.
[12] IHS, “Global Wind Turbine Installation Overview,” 2015.
[13] New Zealand Wind Energy Association, “How Wind Energy Works,” [Online]. Available:
http://www.windenergy.org.nz/improvements-in-technology. [Accessed July 2018].
[14] R. Wiser and M. Bolinger, “2016 Wind Technologies Market Report, US DoE”.
[15] IEA Wind Energy Systems, “Task 26: Wind Technology, Cost, and Performance Trends in Denmark,
Germany, Ireland, Norway, the European Union, and the United States: 2007 - 2012”.
[16] IRENA, “Renewable Power Generation Costs in 2017”.
[17] IRENA, “The power to change: Solar and wind cost reduction potential to 2025,” 2016.
[18] NREL, “U.S. Solar Photovoltaic System Benchmark: Q1 2017. Benchmark Study,” 2017.
[19] Open Energi, “Demand Response Market Overview – Glossary of Demand Response Services,”
2017.
[20] Siemens, “Battery Energy Storage Solution; Enhancing the operational flexibility of flexible com-bined
cycle industrial gas turbines,” in 12th International Renewable Energy Storage (IRES) Conference,
Essen, Germany, 2018.
[21] CIGRE, “Joint Working Group C2/C6.36. System operation emphasizing DSO/TSO interaction and
coordination. Technical Brochure,” 2018.
[22] R. Bucher, “Smart grid functionality for high-voltage transmission grid: on the market readiness of
digital sub-station 2.0 technology,” Bad Vilbel, Germany, 2017.
[23] D. Hurley, P. Peterson and M. Whited, “Demand Response as a Power System Re-source –
Program Designs, Performance, and Lessons Learned in the United States, Synapse Energy
Economics,” 2013.
[24] P. Tielens, P. Henneaux and S. Cole, “Penetration of renewables and reduction of synchronous
inertia in the European power system – Analysis and solutions (ASSET Project),” Tractebel,
Brussels, Belgium, 2018.
[25] R. Serway, R. Beichner and J. Jewett, Physics for Scientists and Engineers, Holt Rinehart & Winston,
2000.
November 2019 153

20-26-00173
Processing No.: 15.2126.9-001.00
[26] P. Kundur, J. Paserba, V. Ajjarapu, G. Andersson, A. Bose, C. Canizares, N. Hatziargyriou, D. Hill, A.

Stankovic, A. Taylor, T. Van Cutsem and V. Vittal, “Definition and classification of power system
stability IEEE/CIGRE joint task force on stability terms and definitions,” IEEE transaction on Power
Systems, vol. 19, no. 3, pp. 1387-1401, 2004.
[27] International Electrotechnical Commission (IEC), “IEC 60909-0: Short-circuit currents in three-phase
a.c. systems,” International Electrotechnical Commission (IEC), 2016.
November 2019 154

20-26-00173
ENERGY SYSTEMS OF THE FUTURE:

GIZ BR ESOF Product 6

Uploaded by

Copyright:

Available Formats

GIZ BR ESOF Product 6

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

GIZ BR ESOF Product 6

Uploaded by

Copyright:

Available Formats

ENERGY SYSTEMS

Processing No.: 15.2126.9-001.00

Energy Systems of the Future: Integrating Variable Renewable Energy Sources

Document Title: Product 6 – Final Report

Prepared by: Lahmeyer International GmbH

Authors: Leonardo Rese, Rebecca Sinder (Tractebel)

Date: November 2019

Coordination: Florian Geyer (GIZ)

Este estudo foi elaborado no âmbito da Cooperação Alemã para o Desenvolvimento

Processing No.: 15.2126.9-001.00

1. Todas as indicações, dados e resultados deste estudo foram compilados e

Processing No.: 15.2126.9-001.00

List of Participants / Lista de Participantes

Steering Committee / Comitê Gestor

Technical Committee / Comitê Técnico – EPE

Technical Committee / Comitê Técnico – ONS

Technical Committee / Comitê Técnico – Consultant / Consultor

Processing No.: 15.2126.9-001.00

Enjoy the reading.

Processing No.: 15.2126.9-001.00

O estudo sobre a Integração de Fontes Renováveis Variáveis na Matriz Elétrica

Processing No.: 15.2126.9-001.00

The Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ), thanks the

Many thanks to all.

Special thanks to the friend and colleague Juarez Castrillon Lopes

Thank you very much from the entire team.

Processing No.: 15.2126.9-001.00

A Deutsche Gesellschaft für Internationale Zusammenarbeit – GIZ, agradece aos

Um muito obrigado a todos.

Agradecimento especial ao amigo e colega Juarez Castrillon Lopes

Na conclusão deste estudo, a equipe não poderia deixar de registrar

Muito obrigado de toda a equipe.

Client Deutsche Gesellschaft für Internationale Zusammenarbeit

Project title Energy systems of the future:

Consultant: Lahmeyer International GmbH

PSR Soluções e Consultoria em Energia Ltda

Date November 2019

Revision Date Status Author Verified Approved

Processing No.: 15.2126.9-001.00

2.1 Brazilian power system and future prospect 12

2.2 Technical regulations for VRE integration in Brazil 13

2.3 VRE technology landscape 14

2.4 Energy planning studies 19

2.5 Power System Studies 25

2.6 Methodology Studies 36

3.2 Product 1: Technical Regulation Studies 51

3.3 Product 2: Energy Studies 64

3.4 Product 3: Power System Studies 85

3.5 Product 4: Methodology Studies 116

3.6 Product 5: Technology Studies 129

3.7 Capacity Building 147

Processing No.: 15.2126.9-001.00

Processing No.: 15.2126.9-001.00

Figure 2-1: Overview of the energy study methodology 20

Figure 2-2: The evolution of percentage of total installed capacity 24

Figure 2-3: Total hourly production for a week in August 24

Figure 2-4: Coupling of energy and power system simulation models 26

Figure 2-5: Methodology for power system studies (overview) 43

Figure 3-1 Product structure and flow 50