ThesisPDF Available

Thermal, energy and daylight performance of office buildings with balconies in subtropical climate

Authors:

Abstract

The use of balconies in tropical climates holds the potential to block direct solar radiation, thereby reducing energy consumption for cooling and enhancing visual comfort. However, balconies may also diminish daylight availability within the room, affecting occupants’ satisfaction and well-being. From a building performance perspective, balcony design is not always trivial and can affect building performance in multiple domains. The complexity in trade-offs of balcony design is evident in the literature, as scientific papers have scarcely explored the effects of the use of balconies on the building performance considering a multi-objective approach. Therefore, the main objective of this thesis was to provide balcony design recommendations to improve daylight, thermal, and energy performance of mixed-mode office buildings. A reference model representing a high-rise mixed-mode office building and possible variations of geometric parameters of balconies were defined based on a database of office buildings located in the city of São Paulo, Brazil. A parametric analysis was used to evaluate the effects of balcony and building design parameters on the reference model’s performance. A multi-objective assessment was performed through building performance simulations (BPS) for daylight, thermal and energy performance, using reliable software tools to assess different balcony design scenarios. To improve accuracy, computer fluid dynamics (CFD) simulations validated with wind tunnel experiments were used to generate wind pressure coefficient data, and daylight experiments were developed to validate the daylight simulations. The trade-offs of balcony design were assessed for daylight availability and visual comfort, as well as for natural ventilation, thermal and energy performance. The results were cross-analysed for daylight, thermal, and energy performance and recommendations for balcony design were tailored to each façade orientation and thoughtfully combined with the glazed door width. An optimal combination, featuring a 3-meter-wide glazed door and a 2-meter-deep balcony, proved beneficial for all façade orientations across all floor levels. However, the choice of parapet type should align with the balcony's location, which is directly correlated with the room's depth. This balcony design combination ensures optimal levels of daylight availability and improves visual comfort by up to 14%, enhancing thermal and energy performance within the room by up to 40%. The results of this research study pioneer in providing balcony design recommendations and offer valuable information for building designers, highlighting that balconies should not be designed solely as decorative façade elements or spaces for building services. Additionally, the methods developed in this research can be applied by researchers/designers to their own case studies.
UNIVERSIDADE ESTADUAL DE CAMPINAS
Faculdade de Engenharia Civil, Arquitetura e Urbanismo
ÍRIS MARIA COSTA FAJARDO WERNECK LOCHE
DESEMPENHO TERMOENERGÉTICO E LUMINOSO
DE EDIFÍCIOS DE ESCRITÓRIOS COM VARANDAS
EM CLIMA SUBTROPICAL
THERMAL, ENERGY AND DAYLIGHT
PERFORMANCE OF OFFICE BUILDINGS WITH
BALCONIES IN SUBTROPICAL CLIMATE
CAMPINAS
2024
ÍRIS MARIA COSTA FAJARDO WERNECK LOCHE
THERMAL, ENERGY AND DAYLIGHT
PERFORMANCE OF OFFICE BUILDINGS WITH
BALCONIES IN SUBTROPICAL CLIMATE
DESEMPENHO TERMONERGÉTICO E LUMINOSO
DE EDIFÍCIOS DE ESCRITÓRIOS COM VARANDAS
EM CLIMA SUBTROPICAL
Tese de Doutorado apresentada à Faculdade de
Engenharia Civil, Arquitetura e Urbanismo da
Universidade Estadual de Campinas para a
obtenção do título de Doutora em Arquitetura,
Tecnologia e Cidade, na área de Arquitetura,
Tecnologia e Cidade.
Thesis presented to the Faculty of Civil
Engineering, Architecture and Urban Design, of
Universidade Estadual de Campinas in partial
fulfilment of the requirements for the degree of
Doctor in Architecture, Technology and Cities, in
the area of Architecture, Technology and Cities.
Orientadora: Profa. Dra. Leticia de Oliveira Neves
ESTE EXEMPLAR CORRESPONDE À VERSÃO FINAL DA TESE
DEFENDIDA PELA ALUNA ÍRIS MARIA COSTA FAJARDO
WERNECK LOCHE E ORIENTADA PELA PROFA. DRA.
LETÍCIA DE OLIVEIRA NEVES.
CAMPINAS
2024
Ficha catalográfica
Universidade Estadual de Campinas
Biblioteca da Área de Engenharia e Arquitetura
Rose Meire da Silva - CRB 8/5974
Loche, Íris Maria Costa Fajardo Werneck, 1994-
L787t LocThermal, energy and daylight performance of office buildings with balconies
in subtropical climate / Íris Maria Costa Fajardo Werneck Loche. – Campinas,
SP : [s.n.], 2024.
LocOrientador: Leticia de Oliveira Neves.
LocTese (doutorado) – Universidade Estadual de Campinas, Faculdade de
Engenharia Civil, Arquitetura e Urbanismo.
Loc1. Iluminação natural. 2. Ventilação natural. 3. Consumo energético. 4. Ar
condicionado. 5. Edifícios comerciais. I. Neves, Leticia de Oliveira, 1980-. II.
Universidade Estadual de Campinas. Faculdade de Engenharia Civil,
Arquitetura e Urbanismo. III. Título.
Informações Complementares
Título em outro idioma: Desempenho termoenergético e luminoso de edifícios de
escritórios com varandas em clima subtropical
Palavras-chave em inglês:
Daylight
Natural ventilation
Energy consumption
Air conditioning
Office buildings
Área de concentração: Arquitetura, Tecnologia e Cidade
Titulação: Doutora em Arquitetura, Tecnologia e Cidade
Banca examinadora:
Leticia de Oliveira Neves [Orientador]
Roberta Vieira Gonçalves de Souza
Ricardo Forgiarini Rupp
Cláudia Naves David Amorim
João Roberto Gomes de Faria
Data de defesa: 20-06-2024
Programa de Pós-Graduação: Arquitetura, Tecnologia e Cidade
Identificação e informações acadêmicas do(a) aluno(a)
- ORCID do autor: https://orcid.org/0000-0003-4117-0367
- Currículo Lattes do autor: http://lattes.cnpq.br/6821873005410221
Powered by TCPDF (www.tcpdf.org)
UNIVERSIDADE ESTADUAL DE CAMPINAS
FACULDADE DE ENGENHARIA CIVIL, ARQUITETURA E
URBANISMO
THERMAL, ENERGY AND DAYLIGHT PERFORMANCE OF
OFFICE BUILDINGS WITH BALCONIES IN SUBTROPICAL
CLIMATE
Íris Maria Costa Fajardo Werneck Loche
Tese de Doutorado aprovada pela Banca Examinadora, constituída por:
Profa. Dra. Leticia de Oliveira Neves
Presidente e Orientadora / Universidade Estadual de Campinas
Profa. Dra. Roberta Vieira Gonçalves de Souza
Universidade Federal de Minas Gerais
Prof. Dr. Ricardo Forgiarini Rupp
Universidade Federal de Santa Catarina
Profa. Dra. Cláudia Naves David Amorim
Universidade de Brasília
Prof. Dr. João Roberto Gomes de Faria
Universidade Estadual Paulista “Júlio de Mesquita Filho”
A Ata da defesa com as respectivas assinaturas dos membros encontra-se no SIGA/Sistema de
Fluxo de Dissertação/Tese e na Secretaria do Programa da Unidade.
Campinas, 20 de junho de 2024
À minha mãe, Sônia, e à minha avó, Maria Aparecida,
por sempre me incentivarem a buscar o conhecimento.
Acknowledgments / Agradecimentos
A trajetória pelo doutorado foi repleta de surpresas, desafios e conquistas. Iniciou-se como mestrado,
evoluiu para doutorado direto e abarcou duas experiências de intercâmbio, percorrendo por três
universidades até culminar na conclusão desta tese. Trilhar esse caminho foi possível somente pelo apoio
das pessoas que caminharam ao meu lado durante todo o percurso ou que estiveram comigo durante
parte dessa jornada.
Agradeço à minha orientadora Letícia de Oliveira Neves por caminhar ao meu lado por essa jornada, de
mãos dadas, me orientando pelo caminho certo e me ajudando a voltar ao rumo correto quando me
perdia. Foi uma honra e um privilégio ser orientada por uma pessoa tão comprometida, dedicada e
humana, que nunca mediu esforços para me ensinar, apoiar meus desejos e aspirações, e me fazer evoluir
como pesquisadora.
Agradeço à minha mãe, que tem estado ao meu lado desde antes e durante toda esta jornada, sempre me
permitindo fazer minhas próprias escolhas, oferecendo suporte incondicional para que eu possa seguir
adiante e acreditando em meu potencial.
Agradeço à minha avó Maria Aparecida, que sempre me inspirou com sua inteligência, incentivando-
me a dedicar-me aos estudos.
Agradeço aos meus irmãos, pais, tios, tias, primos e avós (presentes e em memória), por sempre estarem
do meu lado, apoiando e torcendo por mim.
Obrigada aos amigos que andam ao meu lado desde o colégio, faculdade e intercâmbio, e pelos amigos
que fiz durante o trajeto. Agradeço pelos momentos de apoio, alegria e descontração, fundamentais para
conseguir chegar até o fim da jornada.
I am grateful to Jesper for his love and the moments of joy throughout this journey, as well as for his
companionship during our weekend work sessions, where each of us was immersed in our respective
theses. I appreciate his encouragement during moments of doubt and for keeping me motivated.
I thank my colleagues from UNICAMP, Cardiff University and Eindhoven University of Technology for
their willingness to help and assist me during my struggles.
I am grateful to my supervisors, Dr. Clarice Bleil de Souza, Dr. Benjamin Spaeth and Dr. Roel Loonen,
for welcoming me as a visiting researcher and for dedicating their time to educate and guide me through
my PhD. Your invaluable collaboration has not only enhanced this research but has also enriched my
journey as a researcher.
Aos professores da FECFAU, expresso minha gratidão pelos ensinamentos oferecidos ao longo desta
jornada. E às secretárias da FECFAU, agradeço o apoio e esclarecimento fornecidos durante todo o
processo.
Agradeço à Faculdade de Arquitetura Civil, Arquitetura e Urbanismo da Universidade Estadual de
Campinas, pelo apoio administrativo.
I acknowledge the Welsh School of Architecture at Cardiff University and the Department of the Built
Environment, Eindhoven University of Technology, for hosting me as a visiting researcher and for their
administrative support.
This research was funded by the São Paulo Research Foundation (FAPESP) - GRANT# 2019/01579-9;
2019/22496-4; 2020/16495-2 and 2021/14680-0.
Resumo
O uso de varandas em climas tropicais tem o potencial de bloquear a radiação solar direta, reduzindo o
consumo de energia com refrigeração e melhorando o conforto visual. As varandas, no entanto, também
podem diminuir a disponibilidade de luz natural dentro do ambiente, afetando a satisfação e o bem-estar
dos ocupantes. Do ponto de vista do desempenho da edificação, o projeto de varandas nem sempre é
trivial e pode afetar o desempenho do edifício em múltiplos domínios. A complexidade de uma análise
multiobjetivo para o projeto de varandas é evidenciada na literatura pela escassez de artigos científicos
que exploram os efeitos do uso de varandas no desempenho do edifício considerando uma abordagem
multiobjetivo. Em vista disso, o principal objetivo desta tese é fornecer recomendações projetuais de
varandas para melhorar o desempenho de luminoso, térmico e energético de edifícios de escritórios
multipavimentos que operam em modo-misto de ventilação. Um modelo de referência representando
um edifício de escritórios e variações de parâmetros geométricos de varandas foram definidos com base
em um banco de dados de edifícios de escritórios localizados na cidade de São Paulo, Brasil. Uma
análise paramétrica foi utilizada para avaliar os efeitos dos parâmetros projetuais de varandas
combinados com parâmetros geométricos da edificação no desempenho de edifícios de escritórios que
operam em modo misto. Uma avaliação multiobjetivo foi realizada por meio de simulações
computacionais de desempenho luminoso, térmico e energético, usando programas computacionais
confiáveis para avaliar diferentes casos de projeto de varandas. Para melhor precisão dos resultados,
simulações de Dinâmica de Fluidos Computacional (CFD), validadas com experimentos de túnel de
vento, foram usadas para gerar dados de coeficiente de pressão; e ensaios de iluminação natural foram
desenvolvidos para validar as simulações de desempenho luminoso. Os resultados das simulações,
considerando as variações projetuais de varandas, foram avaliados para disponibilidade de luz natural e
conforto visual, bem como para o desempenho térmico, energético e da ventilação natural. Os resultados
foram cruzados para desempenho de luminoso, térmico e energético, e recomendações projetuais de
varandas foram desenvolvidas para cada orientação solar de fachada e combinadas cuidadosamente com
a largura da porta envidraçada. Uma combinação ótima, apresentando uma porta envidraçada de 3
metros de largura e uma varanda de 2 metros de profundidade, mostrou-se benéfica para todas as
orientações de fachada e para todos os pavimentos do edifício. No entanto, a escolha do tipo de parapeito
deve estar alinhada com a localização da varanda, que está diretamente correlacionada com a
profundidade do ambiente. Esta combinação projetual de varanda garante níveis ótimos de
disponibilidade de luz natural e melhora o conforto visual em até 14%, aumentando o desempenho
termoenergético da sala de escritório em até 40%. Os resultados desta pesquisa são pioneiros em
fornecer recomendações de projeto de varandas e oferecem informações valiosas para os projetistas de
edifícios, destacando que as varandas não devem ser projetadas apenas como elementos decorativos de
fachada ou espaços para serviços. Além disso, os métodos desenvolvidos nesta pesquisa podem ser
aplicados por pesquisadores/projetistas em seus próprios estudos de caso.
Palavras-chave: varanda; modo misto de ventilação; escritório, desempenho luminoso; conforto
visual; desempenho térmico; desempenho energético; otimização; recomendações projetuais.
Abstract
The use of balconies in tropical climates holds the potential to block direct solar radiation, thereby
reducing energy consumption for cooling and enhancing visual comfort. However, balconies may also
diminish daylight availability within the room, affecting occupants’ satisfaction and well-being. From a
building performance perspective, balcony design is not always trivial and can affect building
performance in multiple domains. The complexity in trade-offs of balcony design is evident in the
literature, as scientific papers have scarcely explored the effects of the use of balconies on the building
performance considering a multi-objective approach. Therefore, the main objective of this thesis was to
provide balcony design recommendations to improve daylight, thermal, and energy performance of
mixed-mode office buildings. A reference model representing a high-rise mixed-mode office building
and possible variations of geometric parameters of balconies were defined based on a database of office
buildings located in the city of São Paulo, Brazil. A parametric analysis was used to evaluate the effects
of balcony and building design parameters on the reference model’s performance. A multi-objective
assessment was performed through building performance simulations (BPS) for daylight, thermal and
energy performance, using reliable software tools to assess different balcony design scenarios. To
improve accuracy, computer fluid dynamics (CFD) simulations validated with wind tunnel experiments
were used to generate wind pressure coefficient data, and daylight experiments were developed to
validate the daylight simulations. The trade-offs of balcony design were assessed for daylight
availability and visual comfort, as well as for natural ventilation, thermal and energy performance. The
results were cross-analysed for daylight, thermal, and energy performance and recommendations for
balcony design were tailored to each façade orientation and thoughtfully combined with the glazed door
width. An optimal combination, featuring a 3-meter-wide glazed door and a 2-meter-deep balcony,
proved beneficial for all façade orientations across all floor levels. However, the choice of parapet type
should align with the balcony's location, which is directly correlated with the room's depth. This balcony
design combination ensures optimal levels of daylight availability and improves visual comfort by up
to 14%, enhancing thermal and energy performance within the room by up to 40%. The results of this
research study pioneer in providing balcony design recommendations and offer valuable information for
building designers, highlighting that balconies should not be designed solely as decorative façade
elements or spaces for building services. Additionally, the methods developed in this research can be
applied by researchers/designers to their own case studies.
Keywords: balcony; mixed-mode ventilation; office; daylight performance; visual comfort; thermal
performance; energy performance; optimisation; design recommendation
List of publications during the doctoral program
I. LOCHE, I.; OLIVEIRA, K. P.; DURANTE, M.; FRACALANZA, B. C.; NEVES, L. O.
Effects of balconies on the wind pressure coefficients of naturally ventilated high-rise office
buildings. In: Symposium on simulation in architecture + urban design, 2020, Viena. Anais
[…]. SimAUD: [S. l.], 2020.
II. SILVA, V. G.; LOCHE, I.; SAADE, M. R. M.; PULGROSSI, L. M.; FRANCESCHINI, P.
B.; RODRIGUES, L. L.; PIMENTA, R. G.; NEVES, L. O.; KOWALTOWSKI, D.
Operational and embodied impact assessment as retrofit decision-making support in a
changing climate. In: Windsor Conference, 11., 2020, Windsor. Anais […]. The Windsor
Conference: Windsor, 2020.
III. LOCHE, I.; NEVES, L. O. Efeitos do uso de varandas no desempenho térmico de salas de
escritório em edifícios verticalizados. In: Encontro Nacional de Tecnologia do Ambiente
Construído, 18., 2020, Porto Alegre. Anais […]. Porto Alegre: ANTAC, 2020.
IV. LOCHE, I.; NEVES, L.O. Efeitos das varandas no desempenho térmico, energético e
luminoso de edificações: revisão sistemática da literatura. In: Encontro Nacional de Conforto
no Ambiente Construído, 16., 2020, Palmas. Anais […]. Porto Alegre: ANTAC, 2021.
V. LOCHE, I.; SOUZA, B. C.; SPAETH, A. B.; NEVES, L. O. Decision-making pathways to
daylight efficiency for office buildings with balconies in the tropics. Journal of Building
Engineering, [S. l.], v. 43, p. 1-24, 2021. DOI: 10.1016/j.jobe.2021.102596.
VI. SILVA, V. G.; SAADE, M.; NEVES, L. O.; LOCHE, I.; PULGROSSI, L. M.; SILVA, M. G.
Sight beyond reach: Dynamic life cycle assessment to support resilient retrofit decision-
making in a changing climate. In: NICOL, F.; RIJAL, H. B.; ROAF, S. (org.). Routledge
Handbook of Resilient Thermal Comfort. [S. l.]: Routlege, 2022. cap. 25, p. 417-432.
VII. LOCHE, I.; BRE, F.; GIMENEZ, J. M.; LOONEN, R.; NEVES, L. O. Balcony design to
improve natural ventilation and energy performance in high-rise mixed-mode office buildings.
Building and Environment, [S. l.], v. 258, p. 1-16, 2024. DOI:
10.1016/j.buildenv.2024.111636.
VIII. LOCHE, I.; LOONEN, R; NEVES, L. O. Balcony design recommendations to enhance
daylight, thermal and energy performance of mixed-mode office buildings. Energy and
Buildings. (Submitted - Under review)
List of figures
Figure 1.1: Examples of mixed-mode office buildings with balconies in the city of São Paulo, Brazil
available in the database. ........................................................................................................................19
Figure 1.2: Thesis workflow ..................................................................................................................23
Figura 2.1: Processo de condução da RSL .............................................................................................28
Figura 2.2: Classificação dos artigos selecionados na revisão sistemática da literatura ........................30
Figura 2.3: Classificação dos artigos por localização do estudo e tipo de estratégia passiva utilizada..30
Figure 3.1: Research design workflow ...................................................................................................43
Figure 3.2: Base case model ...................................................................................................................44
Figure 3.3: Examples of office buildings with balconies .......................................................................45
Figure 3.4: Office room dimensions ......................................................................................................46
Figure 3.5: Results analysis diagram ......................................................................................................48
Figure 3.6: Decision tree configuration ..................................................................................................50
Figure 3.7: Dataset results for the window width ..................................................................................51
Figure 3.8: Dataset results for the glazing visible transmittance ...........................................................52
Figure 3.9: Dataset results for the balcony depth ...................................................................................52
Figure 3.10: Dataset results for the room depth .....................................................................................53
Figure 3.11: Dataset results for the solar orientation .............................................................................54
Figure 3.12: Dataset results for the Ratio of balcony width to window width .......................................54
Figure 3.13: Scatter plots for sDA per window width ...........................................................................55
Figure 3.14: Scatter plots for sDA per window width ...........................................................................56
Figure 3.15: Scatter plots for sDA per window width ...........................................................................57
Figure 3.16: Scatter plots for sDA per window width ...........................................................................58
Figure 3.17: Decision tree ......................................................................................................................60
Figure 4.1: Examples of mixed-mode office buildings with balconies in the city of São Paulo, Brazil 66
Figure 4.2: Methodology workflow .......................................................................................................69
Figure 4.3: Representative building model. ...........................................................................................70
Figure 4.4: Example of design cases for each variable parameter. ........................................................71
Figure 4.5: Computational domain for the CFD simulations .................................................................72
Figure 4.6: The modelling approach employed in the Energy Management System (EMS) of
EnergyPlus .............................................................................................................................................75
Figure 4.7: Building parameters selected to analyse the design cases. ..................................................76
Figure 4.8: São Paulo weather conditions during occupied hours (8 a.m. – 6 p.m.) according to the
weather file used in the simulations .......................................................................................................76
Figure 4.9: Sensitivity analysis results for natural ventilation ...............................................................78
Figure 4.10: Impacts of the use of balconies considering the wind speed, wind incidence, and balcony
location. In the margins, density curves show the frequency of occurrence in the outdoor wind speed
(axis x) and in the air change rate (axis y). ............................................................................................78
Figure 4.11: Wind pressure coefficients results. ....................................................................................80
Figure 4.12: Contours for U/UH and details of the velocity fields around balconies for levels 1, 5, and
9. .............................................................................................................................................................81
Figure 4.13: Impacts of balcony design on the air change rate when located in the short-axis façade. .82
Figure 4.14: Impacts of balcony design on the air change rate when located in the long-axis façade. ..82
Figure 4.15: Sensitivity analysis for thermal and energy performance (cooling thermal loads). ...........83
Figure 4.16: Impacts of balcony design on the office located in the north façade. ................................84
Figure 4.17: Impacts of balcony design on the south façade. ................................................................84
Figure 5.1: Mixed-mode office buildings with balconies in the city of São Paulo, Brazil. ...................89
Figure 5.2: Methods workflow ...............................................................................................................91
Figure 5.3: Representative mixed-mode office building ........................................................................92
Figure 5.4: Office rooms analysed for each façade orientation..............................................................93
Figure 5.5: São Paulo weather conditions during occupied hours (8am - 6pm) ....................................93
Figure 5.6: First round of sensitivity analysis: Screening study ............................................................98
Figure 5.7: Second round of sensitivity analysis ....................................................................................99
Figure 5.8: Second round of sensitivity analysis separated by key parameters .....................................99
Figure 5.9: Results for daylight availability (sDA) and visual comfort (sGA) classified according to
results for thermal and energy performance (ACU). In the margins, density curves show the frequency
of occurrence of design cases for sDA values (x-axis) and sGA values (y-axis). ................................100
Figure 5.10: Classification of design cases with balconies located on the North façade for sDA and
ACU. ....................................................................................................................................................102
Figure 5.11: Percentage change for optimal design cases with balconies located on the North façade
..............................................................................................................................................................103
Figure 5.12: Classification of design cases with balconies located on the South façade for sDA and
ACU .....................................................................................................................................................103
Figure 5.13: Percentage of change for optimal design cases with balconies located on the South façade
..............................................................................................................................................................104
Figure 5.14: Classification of design cases with balconies located on the East façade for sDA and
ACU .....................................................................................................................................................105
Figure 5.15: Percentage of change for optimal design cases with balconies located in the East façade
..............................................................................................................................................................105
Figure 5.16: Classification of design cases with balconies located on the West façade for sDA and
ACU .....................................................................................................................................................106
Figure 5.17: Percentage of change for optimal design cases with balconies located in the West façade
..............................................................................................................................................................106
Figure 5.18: Optimal combination of glazed door width and balcony depth for all façade orientations,
varying the parapet type according to the balcony location .................................................................109
Figure 6.1: Design cases with cross-ventilation and single-sided ventilation ......................................114
Figure 7.1: Surrounding scenarios used in the ongoing analysis .........................................................119
Figure A.1: Graphic structure to improve legibility of results .............................................................132
Figure A.2: Results for sDA and ASE - Scenarios with front balconies and clear glass (GTvis 0.88) 133
Figure A.3: Results for UDI - Scenarios with front balconies and clear glass (GTvis 0.88) ...............134
Figure A.4: Results for sDA and ASE - Scenarios with front balconies and laminated glass (GTvis
0.48) .....................................................................................................................................................135
Figure A.5: Results for UDI - Scenarios with front balconies and laminated glass (GTvis 0.48) .......136
Figure A.6: Results for sDA and ASE - Scenarios with side balconies and clear glass (GTvis 0.88) .137
Figure A.7: Results for UDI - Scenarios with side balconies and clear glass (GTvis 0.88) ................138
Figure A.8: Results for sDA and ASE - Scenarios with side balconies and laminated glass (GTvis
0.48) .....................................................................................................................................................139
Figure A.9: Results for UDI - Scenarios with side balconies and laminated glass (GTvis 0.48) .......140
Figure B.1: Base case floor plan ..........................................................................................................144
Figure B.2 Surrounding buildings ........................................................................................................145
Figure B.3: Scale models used in the wind tunnel experiments ...........................................................146
Figure B.4: Sensors position for scenarios with side and front balconies ............................................147
Figure B.5: Velocity profile obtained through wind tunnel tests .........................................................147
Figure B.6: Wind tunnel interior: roughness blocks and spires. ..........................................................148
Figure B.7: Floor plans, location of Cps in analysis and selected floors. ............................................149
Figure B.8: Wind direction 90º - Wind pressure coefficient values for windward and leeward façades
..............................................................................................................................................................150
Figure B.9: ΔCp for scenarios with balconies located at side façades and wind direction of 90º ........150
Figure B.10: Wind direction 0º - Wind pressure coefficient values for windward and leeward façades
..............................................................................................................................................................151
Figure B.11: ΔCp for scenarios with balconies located at frontal and back façades and wind direction
of 0o ......................................................................................................................................................152
Figure B.12: Wind direction 0o - Wind pressure coefficient values for windward and leeward façades
..............................................................................................................................................................153
Figure B.13: ΔCp for scenarios with surrounding buildings and wind direction of 0º ........................153
Figure C.1: Building with balconies on the long axis. .........................................................................158
Figure C.2: Comparison of ABL inlet wind profiles used in experiments and CFD simulations. .......158
Figure C.3: Grid details on ground and building surfaces for the three scenarios analysed in the present
appendix. These correspond to the Fine configuration selected for the automatic meshing with the
CpSimulator platform. ..........................................................................................................................159
Figure C.4: Comparison between measured and simulated Cp data. ...................................................160
Figure D.1: Location of the experiments and reduced-scale model .....................................................161
Figure D.2: Solar radiation measured during the experiments .............................................................162
Figure D.3: Comparison between data measured in the experiments and simulated ...........................162
List of tables
Table 1.1: Thesis chapters summary ......................................................................................................22
Table 1.2: Correspondence of terminologies used across chapters ........................................................25
Table 3.1: Base case model characteristics ............................................................................................44
Table 3.2: Base case variable parameters ...............................................................................................45
Table 3.3: Ratio of balcony width to window width ..............................................................................46
Table 3.4: Radiance input parameters ....................................................................................................47
Table 3.5: Data classification .................................................................................................................49
Table 3.6: Rules to achieve preferred sDA for any balcony configuration ............................................59
Table 3.7: Rules to achieve the same UDI for any balcony configuration and to improve UDI using
deeper balconies. ....................................................................................................................................59
Table 3.8: Rules extracted from the decision tree ..................................................................................61
Table 4.1: Summary of the boundary conditions employed...................................................................73
Table 4.2: Building performance simulation input parameters. .............................................................74
Table 5.1: First selection of variable design parameters ........................................................................94
Table 5.2: Daylight simulation input parameters ...................................................................................94
Table 5.3: Building performance simulation input parameters. .............................................................95
Table B.1: Base case - fixed parameters ..............................................................................................144
Table B.2 Scenarios - Isolated building ...............................................................................................145
Table B.3: Surrounding buildings height for context 1 and 2 ..............................................................145
Table B.4: Scenarios Building with surrounding constructions ........................................................146
Table C.1: Summary of the mesh configurations evaluated for each case study and R2. A value of one
in this indicator means a perfect agreement. ........................................................................................159
Table D.1: Reflectance values used for the building performance simulations ...................................162
Table D.2: Comparison of measured and simulated data for models with and without balconies.......163
Table of contents
1. Introduction ......................................................................................................................................18
1.1 Objectives .....................................................................................................................................21
1.2 Hypothesis and research questions ...............................................................................................22
1.3 Thesis structure ............................................................................................................................22
1.3 Consistency of terms across chapters ...........................................................................................24
2. Literature review ..............................................................................................................................26
2.2 Objetivo ........................................................................................................................................28
2.3 Método .........................................................................................................................................28
2.4. Resultados ...................................................................................................................................29
2.5 Conclusões ...................................................................................................................................37
Agradecimentos ..................................................................................................................................38
3. Effects of the use of balconies on daylight performance ...............................................................39
3.1 Introduction ..................................................................................................................................40
3.2 Background ..................................................................................................................................40
3.3 Methodology ................................................................................................................................42
3.4 Results and discussion ..................................................................................................................50
3.5 Conclusion ....................................................................................................................................62
Acknowledgments ..............................................................................................................................64
4. Effects of the use of balconies on natural ventilation, thermal and energy performance .........65
4.1 Introduction ..................................................................................................................................66
4.2 Methods ........................................................................................................................................69
4.3 Results and discussion ..................................................................................................................77
4.4 Limitations and future studies ......................................................................................................86
4.5 Conclusions ..................................................................................................................................86
Acknowledgments ..............................................................................................................................87
5. Effects of the use of balconies on daylight, thermal and energy performance ...........................88
5.1 Introduction ..................................................................................................................................89
5.2 Methods ........................................................................................................................................91
5.3 Results and discussion ..................................................................................................................98
5.4 Limitations and prospects for future studies ..............................................................................109
5.5 Conclusions ................................................................................................................................110
Acknowledgments ............................................................................................................................110
6. General discussions ........................................................................................................................111
7. Conclusions .....................................................................................................................................116
7.1 Main contributions to science and society .................................................................................117
7.2 Limitations and future studies ....................................................................................................118
Appendices ..........................................................................................................................................132
Appendix A: Daylight simulations dataset (Appendix of Chapter 3) ..............................................132
Appendix B. Wind tunnel experiments (Appendix of Chapter 3) ....................................................141
Appendix C: Validation of CFD simulations (Appendix of Chapter 4) ...........................................157
Appendix D: Validation of daylight simulations (Appendix of Chapter 5) .....................................161
Appendix E: Permission for published journal papers .....................................................................164
18
1. Introduction
In office buildings located in subtropical climates, the use of climate control systems is often necessary
to ensure occupants’ thermal comfort and productivity. In Brazil, office buildings accounted for 10% of
the total energy consumption in 2022, representing a 2% increase when compared to 2010 (EPE, 2023).
One of the reasons for it is the increased use of air-conditioning systems, driven by global warming and
the occurrence of more frequent heatwaves (EPE, 2023), which highlights the importance of using
passive design strategies, such as solar shading devices and natural ventilation, to enhance thermal
performance and decrease energy consumption with cooling.
In hot climates of tropical and subtropical regions, combining mixed-mode ventilation (MMV) systems
with solar shading strategies in office buildings can effectively reduce energy use for cooling
(Mohammed, 2021). MMV systems enable indoor spaces to use natural ventilation during favourable
outdoor conditions to maintain thermal comfort, thereby reducing reliance on air-conditioning for
supplementary cooling (Brager, 2006); also improving indoor air quality (IAQ) and enhancing
occupants' satisfaction and productivity (Arata; Kawakubo, 2022; Brager; Baker, 2009; De Oliveira;
Rupp; Ghisi, 2021).
Balconies are protruding structures that can work as solar shading devices for buildings situated in
tropical and subtropical climates. When wide and deep, they show a good potential in reducing
undesirable direct solar radiation and diminishing cooling loads (Elgohary; Abdin; Mohamed, 2023;
Ribeiro et al., 2024). From a building physics perspective, their design presents a complex challenge
due to trade-offs between daylighting, thermal and energy performance. Regarding daylight
performance, balconies can reduce glare, improving visual comfort (Al-Sallal; AbouElhamd; Dalmouk,
2018), but they may also hinder daylight availability within the room, leading to higher energy
consumption with electric lighting (Gabrova, 2014; Ribeiro et al., 2024).
Since balconies are, usually, prominent elements on building façades, they influence airflow patterns
and pressure distribution, affecting natural ventilation inside the building. This, in turn, directly
influences building thermal and energy performance. Depending on the balcony geometry and the
direction of prevailing winds, they can either improve or obstruct airflow, adding complexity to their
design. (Bamdad et al., 2022; Cochran, 2020; Cui; Mak; Niu, 2014; Izadyar et al., 2020; Montazeri;
Blocken, 2013; Omrani et al., 2017, 2015; Zheng; Montazeri; Blocken, 2020, 2021).
Commonly used in residential buildings, balconies are often desired by occupants, since they allow the
development of multifunctional activities such as leisure, entertainment, and gardening (Peters;
Halleran, 2021; Peters; Masoudinejad, 2022; Ribeiro; Ramos; Flores-Colen, 2020). Also in offices, the
use of balconies is gaining popularity. A study in São Paulo, Brazil, for example, revealed that the
incorporation of balconies in mixed-mode office buildings increased by 85% from 1995 to 2016 (Neves;
19
Melo; Rodrigues, 2019). This study, which compiled a comprehensive database of design parameters
for mixed-mode office buildings in São Paulo, reveals that 23% of these buildings have exterior shading
devices, of which 92% are balconies (Neves; Melo; Rodrigues, 2019).
The increased use of balconies in mixed-mode office buildings is potentially linked to the popularization
of split air-conditioning units, which require an outdoor space for housing the condenser unit. Balconies
are a common choice for placing the condenser unit in high-rise buildings located in hot climates,
following recommendations to place the condenser in well-ventilated areas for optimal heat dissipation
and system efficiency (Abdullah; Barwari, 2022; Chow; Lin; Wang, 2000; Xue et al., 2007) (Figure 1).
This trend presents an opportunity to advocate for the use of balconies in office buildings, as they could
serve multiple purposes, such as accommodating building services, acting as shading control systems,
and providing a pleasant transitional space between indoor and outdoor environments for occupants.
a) Balconies located on the short-axis
façade
b) Balconies located on the long-
axis façade
c) Balconies located on the short-
axis façade
d) Office room with balcony e) Balcony used as a service area
to house the condenser unit in an
office room
f) Office building façade with
condenser unit in the balcony
Figure 1.1: Examples of mixed-mode office buildings with balconies in the city of São Paulo, Brazil available in
the database.
Source: (Pereira, 2019)
Balcony design can also determine their utility and advantages for building occupants. A wide and deep
balcony is more likely to be occupied, while a narrow and shallow balcony may be relegated to a non-
occupied service area, leading to occupants’ dissatisfaction (Li et al., 2023; Song et al., 2024). In São
Paulo, the use of balconies is encouraged by the building code, as it exempts open balconies from being
20
classified as built area if their dimensions per floor are less than 5% of the total site area (São Paulo,
2017). However, depending on the site area’s size, it can lead to shallow and narrow balconies that are
used as service spaces solely.
Balcony design is not always trivial and may affect building performance in multi-domains. Due to this
complexity, there is a significant lack of research exploring the effects of balcony considering multi-
objectives. A systematic literature mapping showed that the exploration of the simultaneous effects of
balconies on natural ventilation and thermal comfort was conducted by Ai et al. (2011), Izadyar et al.
(2020) and Omrani et al. (2017), employing CFD simulations and thermal comfort calculations. Ai et
al. (2011) demonstrated that adding a balcony to the façade enhanced indoor air distribution, leading to
improved thermal comfort. Aligned with these findings, Izadyar et al. (2020) showed that balconies
between 2.5 and 3 m deep produced the best results in terms of indoor air distribution and thermal
comfort in adjacent living rooms of residential buildings. Omrani et al. (2017) concluded that increasing
the balcony depth led to a decrease in air speed. Moreover, the impacts of the balcony depth on thermal
comfort were more pronounced in rooms relying on single-sided ventilation, compared to those
employing cross-ventilation.
The impacts of balconies on daylight, thermal and energy performance were analysed concurrently in
six studies (Dahlan et al., 2009a; Elgohary; Abdin; Mohamed, 2023; Li et al., 2023; Liu; Chen, 2017;
Ribeiro et al., 2024; Yang; Li, 2022). The effects of balconies on thermal and daylight performance of
residential rooms during summer were evaluated through surveys and field measurements by Dahlan et
al. (2009), Yang and Li (2022) and Ribeiro et al (2024). Balconies were identified as effective elements
for enhancing thermal comfort, although they reduced indoor daylight availability. Nevertheless, the
results remained compliant with minimum daylight requirements (Dahlan et al., 2009a; Yang; Li, 2022).
Similarly, results from Ribeiro et al. (2024) demonstrated that rooms without balconies exhibited higher
indoor temperatures during summer, with temperatures 4 °C higher than outdoor temperatures, whereas
indoor spaces provided with open balconies on the South façade were 3 °C cooler. While apartments
without balconies achieved higher illuminance levels compared to those with balconies, these values
exceeded thresholds considered visually comfortable, and the presence of balconies helped regulate
excessive daylight levels. Building performance simulations were conducted by Liu and Chen (2017),
Li et al. (2023), and Elgohary, Abdin and Mohamed (2023) to optimise the balcony design for daylight
and thermal performance. The three studies showed that balcony design should be aligned with the
window-to-wall ratio (WWR). Liu and Chen (2017) suggested increasing the WWR from 50% to 75%
and 100%, to allow a deeper balcony (3 m) without compromising daylight performance, enhancing the
energy savings with cooling in residential buildings located in Taiwan. Li et al. (2023) recommended
adjusting the WWR during balcony renovations to enhance both thermal and daylight performance in
ancient residential buildings located in Beijing, China. A WWR of 50-60% with a 1.2-meter-depth
balcony, for example, could improve indoor thermal comfort while meeting daylight requirements.
21
Elgohary, Abdin and Mohamed (2023) employed parametric design to determine optimal block
arrangements and balcony shapes for designing residential buildings in Cairo. The authors suggest that
the balcony depth should be designed in correlation with the WWR and based on the façade orientation.
The WWR should vary according to each façade, in order to balance daylight penetration and glare, with
the optimum percentage being 10% for the West, South, and East façades, and 60% for the North
orientation. There are no benefits to thermal and energy performance in adding balconies to the North
façade (considering Cairo's location above the equator line), except for architectural aesthetics or
occupants’ preferences. The design for the West façade could incorporate more vertical elements, while
the South façade would require deeper balconies.
Current scientific literature predominantly focuses the analysis of balconies’ performance on residential
buildings (Bayazit; Kisakurek, 2023; Dahlan et al., 2009a; Duarte et al., 2023; Izadyar et al., 2020; Liu
et al., 2021; Omrani et al., 2017; Peters; Masoudinejad, 2022; Ribeiro et al., 2024; Ribeiro; Ramos;
Flores-Colen, 2020; Tungnung, 2020). This focus is evident in the comprehensive literature review
conducted by Ribeiro et al. (2020) on the influence of balconies on the indoor environment of residential
buildings, and in the ongoing discussion about the role of balconies in dwellings during the COVID-19
pandemic (Bayazit; Kisakurek, 2023; Duarte et al., 2023; Peters; Halleran, 2021; Peters; Masoudinejad,
2022; Säumel; Sanft, 2022). However, findings from daylight, thermal and energy performance of
residential studies may have limited applicability to office buildings due to different occupancy
schedules and internal loads, underscoring the necessity for research studies specific to the office
typology. Moreover, balcony performance-based design is complex and correlated to other building
design parameters. Possible consequences of design changes are not always trivial and could affect its
performance in multiple domains.
1.1 Objectives
This research aimed to develop balcony design recommendations to optimise daylight, thermal and
energy performance of mixed-mode office buildings in subtropical climates.
The specific objectives (SO) represent the objectives of the papers that compose this thesis, as follows:
SO1: Assessing both the existing knowledge and gaps in the literature regarding the influence of balcony
design parameters on daylight, thermal and energy performance of buildings. (Chapter 2)
SO2. Evaluating the effects of balcony design parameters on the daylight performance of mixed-mode
office buildings in subtropical climates. (Chapter 3)
SO3. Evaluating the effects of balcony design parameters on the natural ventilation, thermal and energy
performance of mixed-mode office buildings in subtropical climates. (Chapter 4)
Chapters 2, 3, and 4 contribute to accomplishing the main objective, which is presented in Chapter 5.
22
1.2 Hypothesis and research questions
This research considered the hypothesis that balconies can enhance daylight, thermal and energy
performance of mixed-mode office buildings.
In order to better clarify the research problem, the following research questions were developed:
What are the effects of balcony design on daylight availability and visual comfort of mixed-
mode office rooms in subtropical climates?
What are the effects of balcony design on the natural ventilation, thermal and energy
performance of mixed-mode office rooms in subtropical climates?
Which combinations of balcony design parameters, combined with building design parameters,
optimise daylighting, thermal efficiency, and energy performance simultaneously in office
buildings in subtropical climates?
1.3 Thesis structure
This thesis comprises seven chapters and five appendices, as shown in Table 1.1. Chapters 2 to 5 and
Appendix B present papers published or submitted to publication during the doctoral program. The
papers were incorporated into this thesis using the same wording as originally published or submitted,
with adjustments made to their layout for consistency within this document. As some references are
redundant across the chapters, the references were compiled at the end of this thesis for conciseness,
adopting NBR 10520 for citations (ABNT, 2023) and NBR 6023 for references (ABNT, 2018).
Table 1.1: Thesis chapters summary
References
Status
Chapters
1
Introduction
-
-
2
Literature review
LOCHE, I.; NEVES, L. O. Efeitos do uso de
varandas no desempenho térmico de salas de
escritório em edifícios verticalizados. In: Encontro
Nacional de Tecnologia do Ambiente Construído,
18., 2020, Porto Alegre. Anais […]. Porto Alegre:
ANTAC, 2020.
Published
3
Effects of the use of balconies
on daylight performance
LOCHE, I.; SOUZA, B. C.; SPAETH, A. B.;
NEVES, L. O. Decision-making pathways to
daylight efficiency for office buildings with
balconies in the tropics. Journal of Building
Engineering, [S. l.], v. 43, p. 1-24, 2021. DOI:
10.1016/j.jobe.2021.102596.
Published
4
Effects of the use of balconies
on natural ventilation, thermal
and energy performance
LOCHE, I.; BRE, F.; GIMENEZ, J. M.; LOONEN,
R.; NEVES, L.O. Balcony design to improve
natural ventilation and energy performance in high-
rise mixed-mode office buildings. Building and
Environment, [S. l.], v. 258, p. 1-16, 2024.
DOI: 10.1016/j.buildenv.2024.111636.
Published
5
Effects of the use of balconies
on daylight, thermal and
energy performance
LOCHE, I.; LOONEN, R; NEVES, L. O. Balcony
design recommendations to enhance daylight,
Submitted in
April 2024
(Under review)
23
thermal and energy performance of mixed-mode
office buildings. Energy and Buildings.
6
General discussions
-
-
7
Conclusions
-
-
Appendices
A
Daylight simulations dataset
Appendix of Chapter 3
-
B
Wind tunnel experiments
LOCHE, I.; OLIVEIRA, K. P.; DURANTE, M.;
FRACALANZA, B. C.; NEVES, L. O. Effects of
balconies on the wind pressure coefficients of
naturally ventilated high-rise office buildings. In:
Symposium on simulation in architecture + urban
design, 2020, Viena. Anais […]. SimAUD: [S. l.],
2020.
Published
C
Validation of CFD simulations
Appendix of Chapter 4
-
D
Validation of daylight
simulations
Appendix of Chapter 5
-
The thesis structure is shown in Figure 1.2 and is explained as follows.
Figure 1.2: Thesis workflow
Chapter 2 presents a systematic literature review about the effects of balconies on daylight, thermal and
energy performance of buildings. This chapter addresses the first specific objective and identifies trends
and gaps in the literature concerning this subject. The review also provides information regarding
influential balcony design parameters and building design parameters that may affect daylight, thermal
and energy performance of buildings, used to approach specific objectives 2 and 3, as well as this thesis’
main objective.
Chapter 3 presents an investigation of the effects of balcony design in the daylight performance of
mixed-mode office buildings. This chapter approaches the second specific objective by creating
24
successful combinations of building design parameters as well as important cut-off points for decision-
making design, to achieve daylight efficiency in mixed-mode office buildings in São Paulo, Brazil. This
chapter was developed in collaboration with Dr. Clarice Bleil de Souza and Dr. Benjamin Spaeth during
a 5-month research internship at the Welsh School of Architecture, Cardiff University, United Kingdom.
Chapter 4 presents an investigation of the effects of balcony design in the natural ventilation, thermal
and energy performance of mixed-mode office buildings. This chapter approaches the third specific
objective through an integrated method that involves computer fluid dynamics (CFD) and building
performance simulations (BPS) to identify optimal balcony design solutions to enhance natural
ventilation and reduce energy consumption for cooling in mixed-mode office buildings in São Paulo,
Brazil. This chapter was developed in collaboration with Dr. Facundo Bre, Dr. Juan Marcelo Gimenez
and Dr. Roel Loonen during an 18-month research internship at the Department of the Built
Environment, Eindhoven University of Technology, Netherlands.
The main objective of the thesis is addressed in Chapter 5. Balcony design recommendations are
developed to enhance daylight, thermal and energy performance of mixed-mode office buildings. This
chapter was also developed in collaboration with Dr. Roel Loonen during the research internship at the
Department of the Built Environment, Eindhoven University of Technology, Netherlands.
Chapter 6 presents a discussion section that establishes connections between the results of Chapters 2,
3, 4, and 5. Chapter 7 provides a comprehensive conclusion for all preceding chapters, including
limitations, prospects for future studies, and highlighting the main contributions this thesis brings to the
state-of-art.
Appendix A presents the dataset of daylight simulations developed in Chapter 3. Appendix B presents
the wind tunnel experiments and Appendix C provides the validation of CFD simulations with the wind
tunnel experiments, used to support the thermal and energy performance simulations conducted in
Chapters 4 and 5. Appendix D presents the validation of the daylight simulations conducted in Chapter
5 through daylight experiments using scale models. Lastly, Appendix E presents the permissions to
fully incorporate journal papers published and submitted to Elsevier into this thesis.
1.3 Consistency of terms across chapters
Throughout the period during which this study was conducted, we made adjustments to the terminology
used in the chapters comprising this thesis. Consequently, Chapter 3 and Appendix B, published in 2021
and 2020, respectively, uses terminologies that differ from that used in Chapters 4 and 5, which were
later submitted for publication in 2023 and 2024, respectively. These variations are outlined in Table
1.2 for clarity.
25
Table 1.2: Correspondence of terminologies used across chapters
Terminology
(Chapter 3 and Appendix B)
Correspondence
(Chapters 4 and 5)
Meaning
Base case Reference case
Scenarios Design cases
Front-façade / Back façade Short-axis façade
study is a rectangular building, composed by a short-
Side façade
Long-axis façade
Window width
Glazed door width
26
2. Literature review
This chapter is the transcription of the following paper:
Efeitos das varandas no desempenho térmico, energético e luminoso de edificações: Revisão
sistemática de literatura
Authored by: Íris Loche1 and Letícia Oliveira Neves1
1 Universidade Estadual de Campinas, Faculdade de Engenharia Civil, Arquitetura e Urbanismo, Campinas,
Brasil.
Published and presented at XVIII Encontro Nacional de Tecnologia do Ambiente Construído
(ENCAC), Porto Alegre, 2020.
Resumo
O sombreamento e a ventilação natural são estratégias de arquitetura passiva eficazes para garantir bom
desempenho térmico de edificações em grande parte do território brasileiro. A varanda é um tipo de
elemento de proteção solar que atua como um beiral para o pavimento inferior, reduzindo a incidência
de luz solar direta, mas permitindo a entrada de luz refletida e difusa e a abertura de janelas para
ventilação. Uma revisão sistemática da literatura foi desenvolvida com o objetivo de levantar o estado
da arte sobre os efeitos das varandas no desempenho térmico, energético e luminoso de edificações e
identificar tendências e lacunas de pesquisa sobre a temática em questão. Identificou-se, no estudo, um
crescimento sobre a abordagem do tema nos últimos cinco anos, tendo como principal enfoque os efeitos
da presença/ausência de varandas no desempenho térmico de habitações residenciais utilizando, como
método, simulações computacionais. A revisão da literatura mostra a complexidade de cruzamento dos
resultados para desempenho térmico, luminoso e energético em edificações providas de varandas,
apontando a existência de uma interdependência entre as variáveis preditoras de maior influência nos
resultados e, muitas vezes, relações negativas entre elas. Como trabalhos futuros evidencia-se,
principalmente, a necessidade de estudos no tema para a tipologia de escritórios e da elaboração de
diretrizes projetuais que colaborem nas decisões iniciais de projeto, de forma que a varanda possa ser
utilizada em seu máximo potencial para promover desempenho térmico, energético e luminoso,
concomitantemente.
Palavras-chave: varanda; desempenho térmico; desempenho luminoso; desempenho energético;
revisão sistemática da literatura.
Abstract
Shading and natural ventilation are effective passive strategies to assure thermal comfort in most part of
the Brazilian territory. Balconies can be used as a solar shading device that behave as an eave to the
lower floor, reducing the incidence of direct solar radiation, while allowing reflected and diffuse daylight
and natural ventilation through windows openings. A systematic literature review was conducted aiming
to present the state of the art about the effect of balconies on thermal, energy and daylight performance
27
of buildings and identify trends and gaps in this subject. The systematic review indicated an increase in
the number of publications in the last five years, showing, as a main focus, the analysis of the effect of
balconies on the thermal performance of residential buildings, through computational simulations. This
study shows the complexity in crossing the results obtained for thermal, daylight and energy
performance in buildings provided with balconies, showing the interdependence of the independent
variables and, often, negative relation between them. As future work, we identify the need of studies
specifically for the office typology and to develop design guidelines for the early design stages, so
balconies can behave to their full potential in providing thermal, daylight and energy performance
concurrently.
Keywords: balcony; thermal performance; daylight performance; energy performance; systematic
review.
2.1 Introdução
O consumo energético médio nacional de edificações dos setores residencial, comercial e público
evidencia a crescente participação dos sistemas de ar-condicionado e de iluminação artificial na
demanda energética (EPE, 2018). Diante disso, destaca-se a necessidade de projetos arquitetônicos
adequados ao clima em que estão inseridos, visando a otimização do desempenho térmico, energético e
luminoso. Estratégias passivas, como o sombreamento e a ventilação natural, são indicadas para a
melhoria do desempenho térmico de edificações em grande parte do território brasileiro.
A varanda é uma estratégia passiva de arquitetura que, se adequadamente projetada e dimensionada,
pode contribuir na melhoria do desempenho térmico, energético e luminoso de edificações e no bem-
estar de seus usuários. Em edifícios multipavimentos, pode funcionar como um elemento de proteção
solar para o pavimento inferior, reduzindo a incidência de luz solar direta enquanto permite a entrada de
luz refletida e difusa e o uso de grandes aberturas para ventilação. Em relação à ventilação natural, pode
interferir na taxa de renovação do ar, na direção do fluxo de ar e, dependendo do seu projeto e
localização, pode agir como captadora ou como barreira do vento (Wong; Istiadji, 2003).
Por contribuir para o desempenho térmico da edificação, a varanda pode auxiliar na redução do consumo
de energia com soluções artificiais de condicionamento (Brandão; Martins, 2008). Com relação à
iluminação natural, pode reduzir o ofuscamento, melhorando o conforto visual, mas também reduzir a
disponibilidade de iluminação natural no interior dos ambientes, podendo acarretar um maior consumo
de energia pela utilização de iluminação artificial (Wong; Istiadji, 2003). A complexidade de
cruzamento dos resultados para desempenho térmico, energético e luminoso reforça a necessidade de
uma melhor compreensão de como as varandas e seus parâmetros geométricos afetam o desempenho
das edificações, de modo a aproveitar ao máximo o seu potencial.
28
O papel da varanda na sensação de bem-estar dos usuários vem sendo evidenciado durante a pandemia
de COVID-19, emergindo o debate sobre a importância desse elemento durante o confinamento global
(Bournas, 2021; Peters; Halleran, 2021; Ribeiro; Ramos; Flores-Colen, 2020). De acordo com Peters e
Halleran (2021), as varandas tornaram-se elementos ainda mais desejáveis e relevantes durante a
pandemia, sendo utilizadas como ambientes de socialização, protestos, celebrações e promovendo uma
janela para a vida pública. Estudos de avaliação pós-ocupação mostram a varanda como um espaço
desejado pelos usuários que a associam à promoção do seu bem-estar físico e mental (Dahlan et al.,
2009b; Wågø; Hauge; Støa, 2016; Xue et al., 2016a). Os usuários apreciam a experiência sensorial que
a varanda oferece ao possibilitar a expansão das vistas para o exterior e o aumento no tamanho das
aberturas, permitindo a criação de um espaço privativo de conexão com o ambiente externo (Wågø;
Hauge; Støa, 2016).
2.2 Objetivo
Tendo em vista a crescente relevância do assunto em questão, este trabalho tem como objetivo levantar
o estado da arte sobre os efeitos das varandas no desempenho térmico, energético e luminoso de
edificações, assim como identificar tendências e lacunas de pesquisa sobre o tema.
2.3 todo
Utilizou-se como método de pesquisa a Revisão Sistemática da Literatura (RSL), que combina
informações de estudos relevantes para responder a determinada pergunta de pesquisa a partir de um
método científico. Para a realização da RSL utilizou-se o processo proposto por Jesson, Matheson e
Lacey (2011), que propõe sua divisão em seis fases principais (Figura 2.1).
Figura 2.1: Processo de condução da RSL
Fonte: Os autores, adaptado de Jesson, Matheson e Lacey (2011)
Uma busca por artigos relevantes ao tema foi feita nas bases de dados Scopus e Web of Science,
utilizando as seguintes strings de busca: “daylight” OR “visual comfort” AND veranda OR balcony, e
“thermal performance” OR “thermal comfort” OR “natural ventilation” AND veranda OR balcony.
Definiu-se que deveriam ser incluídos nos resultados da busca apenas os artigos que: a) tivessem as
strings de busca no título, resumo ou palavras-chave; b) fossem redigidos em língua inglesa; c) fossem
publicados nos últimos 20 anos (2001 a 2021). Como resultado deste primeiro filtro, foram encontrados
295 artigos. Uma segunda seleção foi feita excluindo os artigos por duplicidade e, a partir da leitura dos
resumos e resultados, excluindo aqueles que não se enquadravam ao objetivo da pesquisa, como estudos
29
com abordagem no desempenho acústico e estrutural das varandas. Após a seleção, 28 artigos foram
selecionados para análise e desenvolvimento do presente trabalho.
2.4. Resultados
2.4.1 Análise bibliométrica da amostra
A categorização dos artigos por ano de publicação permitiu identificar que, no último quinquênio (2016-
2021), a frequência de publicação aumentou 150% se comparado à soma das publicações dos anos
anteriores, representando 61% das publicações (Figura 2.2a), o que indica o aumento do interesse em
pesquisas sobre o tema nos últimos anos, tendo em vista o aumento da preocupação com as questões
climáticas na última década (Fernandes et al., 2020). Presume-se a tendência de um número crescente
de publicações sobre a temática nos próximos anos, fomentado pelas discussões levantadas sobre a
qualidade dos ambientes durante o confinamento global provocado pela pandemia de COVID-19, visto
que o assunto foi abordado nas produções mais recentes sobre o tema (Bournas, 2021; Peters; Halleran,
2021; Ribeiro; Ramos; Flores-Colen, 2020).
A tipologia residencial foi a mais abordada nos estudos aqui analisados, representando 65% da amostra.
A predominância dos estudos nessa tipologia deve-se ao fato de as varandas serem amplamente adotadas
em residências unifamiliares e multifamiliares (Brandão; Martins, 2008; Peters; Halleran, 2021), embora
o uso de varandas na tipologia de escritórios tenha crescido, como mostra o estudo de Manoel e Neves
(2017), realizado para a cidade de São Paulo, que contabilizou um aumento de 85% no uso de varandas
em edifícios de escritórios nos últimos 20 anos. Destaca-se, entretanto, a ausência de estudos nessa
tipologia com enfoque específico em varandas (Figura 2.2b).
A categorização dos artigos por variável de saída analisada permitiu identificar que a maior parte dos
artigos (55%) avalia apenas os efeitos da presença/ausência de varandas no desempenho do ambiente
em estudo, sem incluir análises de aspectos relativos ao projeto das varandas. O segundo parâmetro mais
avaliado, presente em 27% dos estudos, é a profundidade da varanda, seguido pelo tipo de varanda (tipo
de peitoril e formato) e a altura do pavimento onde o ambiente com varandas está localizado,
representando 12% e 6% dos artigos da amostra, respectivamente (Figura 2.2c). Com relação ao método
adotado, destaca-se o uso de simulações (58% da amostra), seguido pela avaliação pós-ocupação por
meio de questionários (21%) e por meio de medições em campo (21%) (Figura 2.2d).
Ao categorizar os artigos selecionados na RSL em objetivo do estudo, identificou-se uma maior
quantidade de estudos sobre os efeitos das varandas no desempenho térmico (45% da amostra), seguido
pelo enfoque na ventilação natural (26% da amostra), o que tem relação direta com o desempenho
térmico da edificação. Isso deve-se ao fato de as varandas serem conhecidamente utilizadas como
estratégia de sombreamento, contribuindo positivamente para o conforto térmico e diminuindo o
consumo de energia com resfriamento (Omrani et al., 2017). Uma menor quantia de estudos avaliou os
efeitos das varandas no desempenho luminoso e energético de edificações, correspondente a 20% e 9%
30
da amostra, respectivamente. Uma parcela ainda menor, correspondendo a apenas dois artigos da
amostra (7%), avaliaram os efeitos das varandas sobre o desempenho térmico e luminoso em conjunto,
uma vez que a integração dessas duas variáveis aumenta a complexidade de análise de resultados (Ochoa
et al., 2012) (Figura 2.2e).
a) Datas publicação b) Tipologia c) Variáveis de análise d) Método e) Objetivo do estudo
Figura 2.2: Classificação dos artigos selecionados na revisão sistemática da literatura
2.4.2 Efeitos das varandas no desempenho térmico e ventilação natural
O uso das varandas foi abordado na literatura como estratégia passiva para edificações situadas em
diferentes localizações geográficas e climas, podendo ser utilizada como estratégia de sombreamento ou
para aquecimento passivo, dependendo de sua configuração (Figura 2.3).
Figura 2.3: Classificação dos artigos por localização do estudo e tipo de estratégia passiva utilizada
A influência das varandas como elementos de sombreamento foi amplamente investigada na literatura
devido ao seu conhecido potencial em reduzir a incidência de radiação solar direta e possibilitar amplas
31
aberturas para ventilação natural. Ai et al. (2011a), Bhikhoo, Hashemi e Cruickshank (2017), Hashemi
(2018), Kisnarini, Krisdianto e Indrawan (2018), Tungnung (2020) confirmaram, por meio de
simulações computacionais, a varanda como potencial estratégia para sombreamento de edificações e,
consequentemente, redução do sobreaquecimento de ambientes internos. Tungnung (2020) ao analisar
uma edificação residencial localizada na Índia, identificou as varandas como espaços agradáveis de
conexão entre o interior e o exterior, por apresentarem temperaturas mais amenas que o interior.
Kisnarini, Krisdianto e Indrawan (2018), Hashemi (2018) e Bhikhoo, Hashemi e Cruickshank (2017)
indicaram a contribuição da varanda para o decréscimo da temperatura de seu cômodo adjacente.
Kisnarini, Krisdianto e Indrawan (2018) mostraram que o uso de varanda em uma edificação residencial
localizada na Indonésia diminuiu em 1 ºC a temperatura interna do ambiente e, com a adição de
dispositivos externos de sombreamento, a redução aumentou para 2 ºC. Hashemi (2018) e Bhikhoo,
Hashemi e Cruickshank (2017) avaliaram o efeito das varandas na redução da probabilidade de
superaquecimento dos ambientes internos. Ao analisar uma residência localizada na Uganda, Hashemi
(2018) concluiu que a varanda reduziu em 50% o risco de superaquecimento. Bhikhoo, Hashemi e
Cruickshank (2017) mostraram, para uma residência localizada na Tailândia, que a eliminação da
varanda aumentou em 19,94% os dias anuais de superaquecimento. Os autores ressaltaram também a
contribuição da ventilação natural na diminuição da probabilidade de superaquecimento, em especial
pelo fato de as varandas possibilitarem o uso de grandes aberturas para ventilação natural.
O método dos estudos aqui analisados compreende avaliação pós-ocupação (medições de campo e
questionários) e simulação computacional. Al-Absi, Abas e Baharum (2018), Arab, Hassan e Qanaa
(2018), Dahlan et al. (2009a), Dahlan, Jones e Alexander (2011) avaliaram as influências das varandas
na redução da temperatura de ambientes internos e na percepção de conforto térmico dos usuários de
edificações localizadas na Malásia. Arab, Hassan e Qanaa (2018) realizaram medições nas fachadas de
duas edificações residenciais e concluíram que, ao obstruir a incidência de radiação solar, as varandas
reduziram a temperatura na superfície da fachada, contribuindo para o desempenho térmico do ambiente
interno. Em concordância, as medições de campo realizadas por Dahlan, Jones e Alexander (2011)
mostraram que os dormitórios sombreados por varandas apresentaram temperatura operativa 1,3 ºC
abaixo dos valores medidos em dormitórios sem varandas. Em adição, Al-Absi, Abas e Baharum (2018)
concluíram que, durante o pico de temperatura no interior da edificação, às 17 h, o ambiente sombreado
por varandas apresentou temperatura 1,5 ºC inferior ao ambiente sem varandas quando as janelas de
ambos os ambientes estavam abertas e 2,5 ºC inferior quando fechadas, mostrando a importância da
ventilação natural em dissipar o calor do ambiente. Os estudos de Dahlan et al. (2009b) não
identificaram mudanças significativas na medição de temperatura operativa dos ambientes com e sem
varandas. No entanto, os autores identificaram, por meio de questionários, que os ocupantes de
dormitórios sombreados por varandas mostraram-se mais satisfeitos com o conforto térmico do
ambiente que os ocupantes de dormitórios não sombreados. Resultados semelhantes foram obtidos por
32
Dahlan, Jones e Alexander (2011). Ai et al. (2011b) combinaram análises em CFD e cálculos de voto
médio estimado (Predicted Mean Vote PMV) e porcentagem de pessoas insatisfeitas (Predicted
Percentage of DissatisfiedPPD) para investigar os efeitos das varandas no conforto térmico em um
edifício de cinco pavimentos naturalmente ventilado localizado em região de clima tropical. Os
resultados mostraram que, embora as varandas tenham reduzido a velocidade do ar no interior da
edificação, sua presença melhorou o conforto térmico por aumentar a uniformidade da distribuição
interna do ar.
Embora o uso das varandas tenha sido abordado majoritariamente como estratégia de sombreamento do
ambiente, os estudos de Fernandes et al. (2015, 2020) e Grudzinska (2016) identificaram que o
fechamento da varanda com elementos envidraçados permite sua atuação como estratégia de
aquecimento passivo em edificações residenciais localizadas em climas frios. Fernandes et al. (2015,
2020) mostraram, por meio de medições em campo e questionários, que as varandas envidraçadas, ao
serem estrategicamente posicionadas na orientação Sul, contribuíram para a entrada de radiação solar,
aumentando o ganho de calor no ambiente interno em edificações residenciais localizadas à Norte de
Portugal. Grudzinska (2016) comparou, por meio de simulação computacional, a eficácia do uso de
varandas constituídas por materiais de alto e baixo isolamento térmico para promover a elevação da
temperatura no ambiente interno de edificações residenciais situadas na Polônia. A varanda de alto
isolamento mostrou-se mais eficaz, ao reduzir em 64,9% o número de dias com consumo energético
para aquecimento, contra uma redução 32,1% resultante da varanda de baixo isolamento térmico,
quando comparadas a um modelo sem varandas. Fernandes et al. (2015, 2020) e Grudzinska (2016)
evidenciaram a importância de os elementos envidraçados de fechamento das varandas serem operáveis,
para evitar o superaquecimento dos ambientes nos períodos quentes do ano, por meio do uso da
ventilação natural. Adicionalmente, Fernandes et al. (2015, 2020) ressaltaram a importância de aliar a
ventilação natural ao sombreamento das áreas envidraçadas da varanda. Os autores demonstraram a
necessidade de uso de elementos fixos de sombreamento externo de modo a suprir, nos períodos sem
ocupação, a falta de uma adequada operação de persianas e janelas.
Além dos fatores aqui destacados, a avaliação do desempenho térmico de edificações naturalmente
ventiladas depende, em grande parte, do desempenho da estratégia de ventilação natural adotada. A
existência de varandas nas fachadas das edificações modifica a distribuição da pressão dos ventos na
envoltória, modificando o fluxo de distribuição de ar no interior da edificação (Ghadikolaei; Ossen;
Mohamed, 2013). Entender os efeitos da presença de varandas na ventilação natural auxilia no
desenvolvimento de projetos com maior desempenho térmico e energético, já que a ventilação natural
interfere diretamente nesses aspectos (Ghadikolaei; Ossen; Mohamed, 2013; Izadyar et al., 2020).
Alguns artigos selecionados na RSL abordaram os efeitos das varandas na ventilação natural, avaliando
aspectos como o tipo de estratégia de ventilação natural, a direção de incidência dos ventos, a altura do
33
pavimento onde a varanda está localizada, a presença/ausência de varandas na fachada e parâmetros
geométricos da varanda como largura, profundidade e projeto de fachada, conforme detalhado a seguir.
a) Dimensionamento das varandas
O dimensionamento adequado das varandas em acordo com o tipo de estratégia de ventilação adotado
pode potencializar o desempenho da ventilação natural em edificações (Ai et al., 2011b; Izadyar et al.,
2020; Omrani et al., 2017). Omrani et al. (2017) investigaram os impactos da geometria das varandas
comparando as estratégias de ventilação unilateral e cruzada em edifícios residenciais. Os resultados
demonstraram que, para ambas as estratégias de ventilação, o aumento da profundidade das varandas
(variações de 1 m a 4 m) ocasionou a redução da velocidade do ar no ambiente interno. Izadyar et al.
(2020) estudaram os efeitos da profundidade das varandas na ventilação natural e no desempenho
térmico de dormitórios residenciais ventilados unilateralmente. Os autores constataram que varandas de
2 m de profundidade provocaram uma distribuição interna do fluxo de ar heterogênea e instável,
enquanto varandas mais profundas (2,5 m e 3 m) apresentaram maiores taxas de renovação de ar no
ambiente interno, colaborando positivamente para o desempenho térmico. Em relação ao comprimento
das varandas, Ai et al. (2011b) mostraram que a variação deste parâmetro trouxe variações
insignificantes no desempenho da ventilação natural da edificação.
b) Inserção das varandas na fachada
Omrani et al (2017) identificaram que, para ambas as estratégias de ventilação natural (unilateral ou
cruzada), uma varanda aberta e protuberante na fachada, oclusa apenas por um parapeito, proporciona
melhor desempenho na ventilação natural do que uma varanda semifechada (oclusa por um parapeito e
por paredes laterais). A adição de uma varanda aberta aumentou a velocidade do ar interno em até 80%
em ambientes com ventilação unilateral e reduziu a velocidade do ar interno em ambientes com
ventilação cruzada. Contudo, os melhores resultados obtidos para os ambientes com ventilação
unilateral continuaram sendo inferiores aos resultados obtidos com ventilação cruzada, apresentando a
velocidade do ar no interior do ambiente até duas vezes menor. Em contradição, Mirabi, Nasrollahi e
Dadkhak (2020) concluíram que varandas semifechadas (oclusas por um parapeito e por paredes
laterais) aumentaram a diferença de pressão entre as paredes opostas, aumentando a eficácia da
ventilação natural no interior do ambiente quando comparadas às varandas abertas (protuberantes na
fachada, oclusa apenas por um parapeito). Em relação ao tipo de peitoril das varandas, Kotani e
Yamanaka (2007) identificaram que a distribuição da pressão do vento nas fachadas de um edifício de
cinco pavimentos não apresentou modificações significativas na comparação entre varandas providas
de peitoril opaco com gradil, sendo a distribuição de pressão mais impactada pela direção incidente dos
ventos na fachada (0o, 90o e 180o).
34
c) Ângulo de incidência do vento nas aberturas
Os estudos disponíveis na literatura destacam a importância de orientar as aberturas da edificação em
relação aos ventos predominantes para a potencialização da ventilação natural nos ambientes internos.
Omrani et al. (2017), concluíram que o ângulo de incidência do vento nas aberturas de edificações
providas de varandas é um parâmetro de maior influência no desempenho da ventilação natural do que
as características geométricas das varandas. Os autores mostraram que, tanto para os ambientes com
ventilação cruzada quanto para os ventilados unilateralmente, a velocidade do ar no ambiente interno
foi superior para os casos com incidência do vento perpendicular às aberturas (0o) e inferior para os
casos com incidência do vento paralela às aberturas (90º). No entanto, Mohamed (2017) identificou que,
quando a incidência do vento ocorre perpendicular às aberturas (0°), o uso de varandas provocou o
crescimento das taxas de ventilação natural em 99% para ambientes com ventilação unilateral contra
uma redução de 44% para ambientes com ventilação cruzada. De maneira oposta, a incidência do vento
oblíqua às aberturas (45°) provocou um aumento de 38% nos ambientes com ventilação cruzada contra
uma redução de 39% para os ambientes ventilados unilateralmente.
d) Altura do pavimento
Cui, Mak e Niu (2014) identificaram que, em uma edificação de dez pavimentos, a presença de varandas
melhorou o desempenho da ventilação natural nos pavimentos intermediários (4o ao 6o). Ai et al. (2011a)
chegaram à mesma conclusão ao analisarem ambientes com ventilação cruzada localizados em uma
edificação de cinco pavimentos, mostrando que a presença de varandas aumentou o desempenho da
ventilação natural nos pavimentos intermediários e diminuiu sua eficácia nos pavimentos superior e
térreo. No entanto, para ambientes ventilados unilateralmente, a adição de varandas reduziu o
desempenho da ventilação natural nos pavimentos intermediários.
2.4.3 Efeitos das varandas no desempenho luminoso
Os estudos selecionados na RSL avaliaram os efeitos da presença de varandas e de sua profundidade na
disponibilidade de luz natural e probabilidade de ofuscamento nos ambientes, utilizando como método
de estudo simulações de desempenho luminoso e avaliação pós-ocupação (questionários e medições em
campo).
Kim e Kim (2010a) e Liu e Chen (2017) mostraram, por meio de simulações computacionais, que o
aumento da profundidade das varandas pode diminuir a disponibilidade de iluminação natural dentro do
ambiente. Adicionalmente, os estudos de Kim e Kim (2010a) e de Liu e Chen (2017) identificaram a
profundidade das varandas como o parâmetro de maior impacto no desempenho luminoso do ambiente
interno. Kim e Kim (2010a) mostraram que varandas mais profundas (1,5 m, 3 m e 6 m de profundidade)
reduziram o valor médio do fator de luz do dia em 46%, 70% e 90%, respectivamente, em relação a um
modelo sem varandas, em análise realizada em um edifício de tipologia indefinida com varandas
localizadas no átrio central. Em consonância, Liu e Chen (2017), ao analisarem uma edificação
35
residencial localizada em Taiwan, sugeriram que aberturas menores deveriam ser associadas a varandas
pouco profundas, de modo a evitar um desempenho luminoso insatisfatório do ambiente interno.
Com relação ao conforto visual e à probabilidade de ofuscamento, Al-Sallal, AbouElhamd e Dalmouk
(2018) mostraram que a presença de uma varanda profunda (3 m) foi capaz de reduzir o ofuscamento
no ambiente interno de 81% para 0%, quando comparado ao mesmo ambiente sem varanda. A análise
foi realizada via simulação computacional, tendo como base a métrica dinâmica de análise Exposição
Solar Anual (Annual Sunlight Exposure ASE). Kim e Kim (2010b) também demonstraram a
importância das varandas na promoção do conforto visual do ambiente interno ao investigarem uma
prática comum nas edificações residenciais coreanas, que consiste na eliminação das varandas para
incorporação de sua área ao cômodo adjacente. Ao remover o sombreamento proporcionado pela
varanda, os autores indicaram que a transmitância do vidro deveria ser reduzida para valores abaixo de
0,54, de forma a promover o mesmo nível de conforto visual oferecido por uma varanda de 1,8 m de
profundidade. Liu e Chen (2017) identificaram que, para uma edificação isolada (sem interferência do
entorno), a altura do pavimento em que o ambiente provido de varandas se localiza apresenta impactos
insignificantes no desempenho luminoso do ambiente interno.
Bournas (2021), Dahlan et al. (2009b) e Xue et al. (2016) basearam-se em avaliações pós-ocupação para
avaliar os efeitos das varandas no desempenho luminoso de edificações por meio de questionários e
medições em campo. Os resultados foram complacentes com os resultados das simulações de
desempenho luminoso encontrados na literatura, confirmando que as varandas reduzem a incidência de
iluminação natural no ambiente interno, mas melhoram o conforto visual. Bournas (2021) comparou o
comportamento dos moradores de apartamentos com e sem varandas de seis edificações residenciais
localizadas na Suécia. O autor identificou que a presença de varandas não exerceu influência no
acionamento da iluminação artificial durante o dia. O estudo de Dahlan et al. (2009b), realizado em
albergues estudantis localizados na Malásia, identificou que 78% dos ocupantes de dormitórios sem
varandas mostraram-se satisfeitos com a iluminação natural, contra 60% dos ocupantes de dormitórios
com varandas. As medições em campo, realizadas em dormitórios localizados no primeiro pavimento e
orientados a Norte, ecoaram os resultados dos questionários, indicando que a presença de varandas
reduziu em três vezes o nível de iluminação natural, medido através da métrica razão de iluminação
natural (daylight ratio). Não obstante, os dormitórios com varandas continuaram apresentando daylight
ratio complacente com os níveis mínimos de iluminação exigidos pela legislação local, o que explica a
baixa diferença entre o número de usuários satisfeitos com os níveis de iluminância dos dormitórios com
e sem varandas, obtido pelos questionários aplicados.
Xue et al. (2016a) e Dahlan et al. (2009b) analisaram o comportamento dos usuários na operação de
elementos internos de sombreamento em apartamentos residenciais com e sem varandas. De acordo com
Xue et al. (2016a), os usuários consideram que as varandas proporcionam pouca privacidade, induzindo-
36
os a fechar as cortinas com mais frequência do que usuários de dormitórios sem varandas, aumentando,
por consequência, a necessidade do uso de iluminação artificial. Em contradição, de acordo com Dahlan
et al. (2009b) os moradores de dormitórios com varandas raramente fecham as cortinas, devido à
percepção de que seus quartos não possuem iluminação natural suficiente, enquanto os moradores de
dormitórios sem varandas disseram fechar as cortinas com mais frequência em função do ofuscamento
provocado pela falta de proteção solar.
2.4.4 Efeitos das varandas no desempenho energético
O desempenho energético de uma edificação está diretamente relacionado ao seu desempenho térmico
e luminoso, que irão definir a demanda energética por aquecimento, resfriamento e iluminação artificial
dos ambientes. Os estudos de Liu e Chen (2017), Grudzińska (2016) e Nikolic et al. (2020) investigaram,
por meio de simulações computacionais, os efeitos das varandas no desempenho energético de
edificações residenciais.
Liu e Chen (2017) analisaram o efeito da profundidade das varandas no desempenho termoenergético
de uma edificação localizada em Taiwan e concluíram que, quanto mais profunda a varanda, maior a
energia economizada no uso de ar-condicionado. Ao relacionarem a profundidade das varandas com o
tamanho das aberturas das janelas, os autores indicaram que varandas mais profundas devem ser
associadas com aberturas maiores. Para um cenário de percentual de abertura na fachada de 50%, os
autores sugeriram o uso de varandas de 2,5 m de profundidade para reduzir o consumo energético anual
em 22% em comparação a uma edificação sem varandas. Para um percentual de abertura na fachada de
75% ou 100%, os autores sugeriram o uso de varandas com 3 m de profundidade, de forma a reduzir o
consumo energético anual em 30%. Grudzińska (2016) comparou a eficácia do uso de varandas de alto
e baixo isolamento térmico na redução da demanda energética para aquecimento de edificações
residenciais situadas na Polônia. Quando comparada a uma residência sem varandas, a varanda de alto
isolamento mostrou-se mais eficaz em reduzir a demanda energética, atingindo até 30% de redução no
ambiente da própria varanda e até 90% no cômodo adjacente. Já a varanda de baixo isolamento atingiu
até 15% de redução no ambiente da varanda e até 70% de redução no cômodo adjacente. Nikolic et al.
(2020) investigaram, por meio de simulações computacionais, como a profundidade das varandas
impacta no consumo energético para aquecimento, resfriamento e iluminação artificial em edificações
residenciais localizadas na Sérvia. Os autores indicaram a profundidade da varanda correspondente ao
melhor cenário de redução do consumo energético, para cada orientação solar: 2,6 m para a fachada
leste, 0,7 m para a fachada sul, 2,4 m para a fachada oeste e 0,4 para a fachada norte. A redução da
incidência direta de radiação solar diminuiu a demanda energética para resfriamento em 44,15%
enquanto o consumo energético com aquecimento e iluminação artificial aumentou em 16,33% e 4,98%,
respectivamente, em comparação a uma edificação sem varandas. O consumo total de energia da
edificação reduziu em 7,12%.
37
2.4.5 Lacunas de pesquisa e indicação de trabalhos futuros
Dentre os trabalhos levantados que trataram dos efeitos das varandas no desempenho luminoso, 50%
utilizaram simulações computacionais como método de estudo. No entanto, apenas o estudo de Al-
Sallal, AbouElhamd and Dalmouk (2018)utilizou métricas dinâmicas de análise dos resultados. As
métricas dinâmicas são consideradas mais eficientes do que as métricas estáticas por considerarem as
condições reais do céu presentes nos arquivos climáticos. Dessa forma, destaca-se a necessidade de mais
estudos que avaliem os efeitos das varandas no desempenho luminoso por meio de métricas dinâmicas
de análise, como Daylight Autonomy (DA), Annual Sunlight Exposure (ASE), Spatial Daylight
Autonomy (sDA) e Useful Daylight Illuminance (UDI). Além disso, outros parâmetros de projeto da
edificação devem ser considerados na abordagem do desempenho luminoso de edifícios providos de
varandas, como a transmitância visível do vidro, o tamanho da abertura das janelas e a profundidade do
ambiente, sendo que, o último não foi abordado em nenhum dos estudos da amostra selecionada.
Não foram identificados estudos que abordem os efeitos das varandas no desempenho térmico, luminoso
e/ou energético para a tipologia de edifícios de escritórios. A necessidade de estudos específicos para
esta tipologia é evidenciada tendo em vista características específicas da tipologia, como o período de
ocupação, por exemplo, que difere significativamente entre edifícios residenciais e edifícios de
escritórios. Outro exemplo é o nível mínimo de iluminância exigido pela NBR ISO/CIE 8995 (ISO,
2002), que é mais elevado para salas de escritórios (mínimo de 300 lux) do que para ambientes
residenciais (mínimo de 100 lux). Outros fatores que diferem entre ambas as tipologias e que
influenciam na avaliação do desempenho térmico e energético do ambienteo a densidade de ocupação,
a carga térmica de equipamentos internos e o isolamento térmico da vestimenta dos usuários. Assim,
evidencia-se a necessidade de estudos que abordem os efeitos do desempenho térmico, energético e
luminoso para a tipologia específica de escritórios.
A maior parte dos estudos aqui analisados compararam a presença e a ausência de varandas no
desempenho de edificações habitacionais, em especial com enfoque em desempenho térmico. Dessa
forma, evidencia-se a importância de estudos que avaliem, simultaneamente, o desempenho térmico,
energético e luminoso de varandas em edificações, bem como os impactos de sua geometria. É
importante ressaltar também a necessidade de estudos nesta vertente que indiquem diretrizes de projeto
que auxiliem os projetistas nas tomadas de decisão nas etapas iniciais de projeto, visando a inserção das
varandas nas edificações tendo em vista seu máximo potencial para contribuir no desempenho térmico,
energético e luminoso.
2.5 Conclusões
Este artigo apresentou uma revisão sistemática da literatura sobre os efeitos das varandas no desempenho
térmico, energético e luminoso de edificações, com o intuito de identificar tendências e lacunas de
pesquisa sobre o tema. Como tendências, evidencia-se um aumento do interesse pelo tema, indicado
38
pelo crescimento do número de publicações nos últimos cinco anos. Evidencia-se também a abordagem
majoritária dos efeitos das varandas no desempenho térmico de edificações residenciais, realizadas por
meio de simulações computacionais. As principais variáveis preditoras analisadas consistem na
presença/ ausência de varandas, seguida pela sua profundidade.
Evidencia-se, na revisão da literatura aqui empreendida, a complexidade do cruzamento de análises dos
efeitos das varandas sob o enfoque em desempenho térmico, energético e luminoso, devido ao fato de
que cada abordagem apresenta parâmetros específicos, que podem ser interdependentes e apresentar
relações positivas ou negativas. De fato, as pesquisas aqui analisadas confirmam a existência de
potenciais conflitos que precisam ser melhor explorados. O aumento da profundidade das varandas, por
exemplo, pode aumentar o conforto visual ao reduzir a probabilidade de ofuscamento, mas também
diminuir a disponibilidade de iluminação natural no ambiente interno, aumentando o consumo
energético com iluminação artificial. Pode também ocasionar a redução da velocidade do ar no ambiente
interno, causando interferência direta no seu desempenho térmico. Por outro lado, pode reduzir a
incidência direta de radiação solar, reduzindo o sobreaquecimento do ambiente interno.
Agradecimentos
Os autores agradecem à Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP), processo
2019/01579-9, pelo financiamento da pesquisa.
39
3. Effects of the use of balconies on daylight performance
This chapter is the transcription of the following paper:
Decision-making pathways to daylight efficiency for office buildings with balconies in
the tropics
Authored by: Iris Loche1,2, Clarice Bleil de Souza2, Benjamin Spaeth3 and Letícia Oliveira Neves1
1School of Civil Engineering, Architecture and Urban Design, Universidade Estadual de Campinas, Brazil
2 Welsh School of Architecture, Cardiff University, United Kingdom
3Technical University of Applied Sciences Lübeck, Germany
Published in: Journal of Building Engineering, vol. 43, 2021
DOI: https://doi.org/10.1016/j.jobe.2021.102596
Abstract
Daylight in the indoor environment is directly influenced by the building surroundings, envelope and its
shading devices, such as balconies. Despite their potential in contributing to increase shaded periods
and, at the same time, act as a daylight distribution system, balconies are not designed to their full
potential when used in office buildings and the literature lack studies that investigate the effect of
balconies on their luminous performance. This study aims to explore this niche: the integration of
balconies to the design of office buildings in the tropics, in order to improve their daylight performance.
The research method was based on a parametric design approach in combination with daylight
simulations, while combining a systematic analysis with a data mining algorithm. The study revealed
successful combinations of building design parameters as well as important cut-off points for design
decision-making to achieve daylight efficiency in typical mixed-mode office buildings in the city of São
Paulo, Brazil. Results provided multiple design routes to achieve successful performance targets
showing that, if properly dimensioned, balconies could be an efficient shading device and daylight
diffuser. As a key contribution, successful combinations of design parameters that allow deeper
balconies to yield better Useful Daylight Illuminance levels were identified. Further details about when
balconies stop influencing daylight performance results as well as when an increase in balcony depth
becomes beneficial to performance were reported in attempt to develop design guidelines for the early
design stages for office buildings in São Paulo.
Keywords: Office building, Balcony, Visual comfort, daylight performance, decision-making, data
mining
40
3.1 Introduction
This study aimed to identify decision-making pathways to design efficient balconies for office buildings
in the tropics, considering improving indoor daylight conditions in the climate of the city of São Paulo
(Brazil). Balconies are common architectural features in residential and hospitality buildings used for
several purposes which span from having a semi-private outdoor space, which can afford social
activities, over urban greenery, and farming in multi-storey buildings, up to providing ‘immersive’
experiences to unique outdoor sceneries. From a building physics perspective, balconies are powerful
shading devices. With larger depths and widths, they demonstrate great potential in increasing diffuse
daylight into office spaces while reducing undesirable direct solar radiation with the effect of
diminishing cooling loads commonly present in the tropics.
However, despite their large potential in contributing to well-being, sustainability and comfort,
balconies are not used to their potential in office spaces in the tropics. In São Paulo, for instance, they
tend to be shallow (normally 0.5 m depth) and primarily used to accommodate air conditioning
condensers from split unit systems, rather than thought as a multipurpose and multi-functional façade
element. More than half (56%) of the office buildings in São Paulo are supplied with split air
conditioning systems, locally controlled, and metered (i.e. per room) with a mixed-mode regime (Acesse
Buildings, 2016). They are office rooms rented by small companies, in a market scenario of medium-
rise buildings with small office floor plans and mixed-mode operation systems, in opposition to wide
open plan office buildings, which usually consists of non-operating curtain-wall façade systems.
This study aimed to explore the integration of balconies to the design of office buildings using a
parametric design approach in combination with daylighting simulations. It did not see balconies in
isolation but as an element of façade design. Therefore, it focuses on unfolding the most important
combination of design parameters to be used in the early design stages for balconies to become efficient
shading devices and daylight diffusers in a ‘typical’ mixed-mode office building in São Paulo. The
richness of parametric design possibilities was explored via an extensive systematic analysis of
simulation results in combination with the application of a datamining technique to show pathways of
successful parametric combinations towards achieving different sets of daylight targets for comfort and
well-being.
3.2 Background
Daylight has a positive impact on human well-being, particularly in workplaces, leading to better
productivity and higher employee satisfaction (Knoop et al., 2020; Turan et al., 2020). Besides its human
health benefits, the daylight is monetarily valued, providing energy savings with electric lighting, and
impacting the real estate market, mainly in dense urban environments. According to Turan et al. (2020)
in Manhattan, NY, tenants are willing to pay up to 6% more for office rooms with high daylight access
over those with low daylight access. Daylight is therefore a fundamental component to achieve indoor
41
wellbeing, energy efficiency and sustainability, stated through building regulations, urban zoning
policies and green building certifications, which determine targets for building daylight performance.
Daylight performance can be evaluated through different daylight metrics. Static daylight measures,
widely criticised throughout the literature (Mardaljevic et al., 2012; Reinhart; Mardaljevic; Rogers,
2006; Tregenza; Mardaljevic, 2018), are still being adopted as a performance measure for daylighting
in building regulations such as the Brazilian to provide guidelines for lighting in residential buildings
(ABNT, 2013) and workplaces (ISO, 2002). Despite the fact that there is still no consensus on a metric
that should replace the daylight factor, it is commonly agreed in the literature that Climate-Based
Daylight Modelling (CBDM) (IES LM 83-12, 2012; Mardaljevic et al., 2012), which uses real sun and
sky conditions from standard weather files to quantify daylight and visual performance, could be the
most suitable approach to assess daylight availability and distribution.
Daylight in an indoor environment is directly influenced by the building envelope and its shading
devices, which are used to avoid glare and improve visual comfort, but also decrease the incidence of
direct daylight in interior spaces (Glassman, 2015; Kim; Kim, 2010a). Balconies are horizontal
overhanging structures enclosed by walls or parapets that behave as an eave to the lower floor, reducing
the incidence of direct solar radiation, while allowing the penetration of reflected and diffuse light (Xue
et al., 2016). Investigations about the effects of balconies on the luminous performance of office
buildings have not been sufficiently explored yet. A systematic literature review identified that 62% of
the papers on the subject investigated residential and hotel buildings. Moreover, also 62% used computer
simulation tools as part of their methods, but only 12% used CBDM daylight metrics. The parameters
analysed include the façade’s Window-to-Wall Ratio (WWR), the glazing’s visible transmittance, the
balcony’s configuration (width, depth, parapet material and solar orientation), the room’s depth and the
floor height.
Results presented by Al-Sallal, AbouElhamd and Dalmouk (2018), Kim and Kim (2010a) and Gábrova
(2014) through daylight simulations showed that the presence of balconies increased daylight uniformity
and decreased glare and illuminance levels in the indoor environment. Gábrova (2014) stated that the
use of balconies increased daylight uniformity up to 55% inside the room. Al-Sallal, AbouElhamd and
Dalmouk (2018) showed that a 3-meters wide balcony was able to eliminate glare inside the room. Kim
and Kim (2010a) stated that balconies 3 and 6-meters deep decreased illuminance levels by 18% and
46%, respectively. Gábrova (2014) showed that balconies 0.75, 1.0, 1.25 and 1.5-meters deep decreased
illuminance levels by 20%, 25%, 30% and 35%, respectively.
Xue et al. (2016) and Dahlan et al. (2009b, 2009a) investigated the impact of balconies on the luminous
performance of residential buildings through field measurements and questionnaires applied to the
occupants. Their results complied with daylight simulation results found in the literature, confirming
that balconies reduce direct daylight incidence and increase visual comfort in indoor spaces. Regarding
42
the balcony’s parapet material, Dahlan et al. (2009a) showed that balconies with an opaque parapet
provided higher levels of visual comfort than balconies with glazing parapet. Liu and Chen (2017)
pointed out the WWR as an outstanding parameter in daylight performance for buildings with balconies,
indicating that the smaller the WWR the shallower the balcony should be in order to avoid a negative
impact on the indoor daylight availability. As to the window glazing properties, Kim and Kim (2010b)
stated that, in order to provide the same visual comfort as a balcony does, the glazing visible
transmittance should be lower than 0.54. Liu and Chen (2017) pointed out the floor level as the parameter
with lower impact on the luminous performance of indoor environments, when considering an isolated
building with balconies.
To the best of authors’ knowledge, there are no studies that investigate the effect of balconies on the
luminous performance of office buildings, perhaps because balconies are seen as a space of no use in
commercial environments. Yet, the use of balconies in office buildings has been growing in recent years.
In the city of São Paulo, Brazil, 23% of the mixed-mode office buildings are provided with exterior
shading devices, of which 92% are balconies (Manoel; Neves, 2017). Between 1995 and 2015 alone,
the use of balconies has increased by 85%, potentially related to the increasing use of split air-
conditioning units, which demand an outdoor area to allocate the condenser (Neves; Melo; Rodrigues,
2019). This can be seen as an opportunity to promote the use of balconies in office buildings as they
could be justified as a space to accommodate building services as well as act as a daylight and shading
control system.
Already overheated by internal gains, office buildings need to minimise incident solar radiation and
annual sunlight exposure to reduce cooling energy consumption, particularly in the tropics (CBCS,
2014). Balconies can offer possibilities to increase shaded periods and, at the same time, act as a daylight
distribution system due to their special configuration, which can block direct sunlight but potentially
contribute to increase reflected daylight. Seeing balconies as an architectural element to reduce
overheating and at the same time act as a potential daylight system distributor, this study focused on
exploring what balcony configuration together with window design parameters could improve daylight
performance and contribute to reduce overheating in office buildings in São Paulo. The study was
undertaken in a ‘typical’ mixed mode office building in São Paulo and simulation results were classified
using systematic analysis in combination with a decision tree algorithm.
3.3 Methodology
The research design used in this study was threefold: It started by defining and parameterizing a ‘typical’
mixed-mode office building for the city of São Paulo based on a dataset of surveyed buildings developed
by Neves, Melo and Rodrigues (2019) and Pereira and Neves (2018). Daylight simulations were then
undertaken for a dataset of 6,360 combinations of parameters using specific weather data for São Paulo
(latitude: 23°32'56'' South, longitude: 46°38'20'' West, altitude 800 m). CBDM metrics and relevant
43
performance thresholds were specified to enable comparability and classification of daylight results. In
the second stage, results were systematically analysed to identify effective combinations of parameters,
with specific attention to the role of balconies, to reach the prescribed thresholds for daylight
performance. In the third stage, results were mined and grouped using a decision tree algorithm to
increase the number of successful combinations of design parameters to achieve daylight performance
thresholds. Combining stages two and three would provide enough breath for designers to reach daylight
performance targets in the early design stages, when simulation is potentially expensive and unavailable.
The proposed research design was depicted in Figure 3.1 and further detailed in the following
subsections.
Figure 3.1: Research design workflow
3.3.1 Parametric study
According to Neves, Melo and Rodrigues (2019) and Pereira and Neves (2018) , mixed-mode office
buildings located in the city of São Paulo (Brazil) are mostly medium-rise buildings with multiple office
units per floor (four to five units, in most cases), served by operable windows and individual air-
conditioning systems, both manually operated in a concurrent mode. Each unit is normally a different
tenancy, and most office buildings have no building facilities manager.
A database with a sample of 153 surveyed case studies of mixed-mode office buildings in São Paulo
(Neves; Melo; Rodrigues, 2019) was used as a reference to create a representative of ‘typical’ mixed-
mode office room model, which was used as a base case to develop the parametric study proposed in
this research. This corresponds to 10% of offices of this type in São Paulo and was considered a
representative sample to extract typical features. The selection of the sample took into consideration the
following criteria: small office rooms, excluding wide open plan office buildings (which usually consists
of non-operating curtain wall façade systems); individual air-conditioning systems, consisting of
independent outdoor air units per office room (excluding central systems, which usually corresponds to
fully air-conditioned buildings); office buildings built between 1995 and 2016.
44
The geometry and envelope design parameters were chosen according to the highest frequency values,
for categorical variables, and the mean values, for continuous variables, from the database. The indoor
surfaces’ reflectance was defined according to ISO/CIE 8995-1 (ISO, 2002) and the ground surface
reflectance according to IES LM 83-12 approved method (IES LM 83-12, 2012). The base case
characteristics are shown in Table 3.1 and illustrated in Figure 3.2.
Table 3.1: Base case model characteristics
Parameter
Value
Building orientation (longitudinal axis)
North - South
Number of floors
11
Office room shape
Normally rectangular
Office room area
38.5 m2 (5.5 m x 7 m)
Office room height (floor-to-ceiling)
2.75 m
Wall thickness
0.25 m
Balcony’s parapet type / height
Opaque parapet / 1.1 m
External wall and balcony’s external surface reflectance
0.30 - dark colour
Internal wall and balcony’s internal surface reflectance
0.5
Room’s and balcony’s floor reflectance
0.2
Ceiling reflectance
0.7
Ground surface reflectance (albedo)
0.1
a) Floor plan b) Building axonometric
Figure 3.2: Base case model
The literature suggests that balconies, window features and their configurations impact the luminous
performance of indoor environment (Al-Sallal; AbouElhamd; Dalmouk, 2018; Liu; Chen, 2017).
Specifically, balcony’s geometry (depth, width, and location), window characteristics (window width
and glazing visible transmittance), building’s solar orientation and floor height were also considered as
important variables in daylight performance of buildings with balconies. The abovementioned
parameters were therefore selected as the main parameters to be used in a sensitivity analysis in the base
case building. Table 3.2 illustrates the range of variation used for these parameters based on the database
from Pereira and Neves (Neves; Melo; Rodrigues, 2019).
45
Table 3.2: Base case variable parameters
Parameter
Value
Balcony depth
0.0 m (no balcony) / 0.5 m / 1.0 m / 1.5 m / 2.0 m
Room’s width and depth
Side façade (5.5 m width and 7 m depth) / front façade (7 m
width and 5.5 m depth)
Balcony width
Ratio of balcony width to window width (0.5, 1.0, 2.0)
Glazing visible transmittance
0.88 (clear glass) / 0.48 (laminated glass)
Window width
1.0 m to 6.5 m (increments of 0.5 m) for the front façade
1.0 m to 5.0 m (increments of 0.5 m) for the side façade
Office room solar orientation
North / South / West / East
Floor number
Upper floor (10th floor 30.1 m height) / middle floor (6th floor
18.6 m height) / lower floor (1
st
floor 3.1 m height)
Balcony and room’s geometry were selected based on the most common dimensions found in the
database of Pereira and Neves (2018). According to Pereira and Neves (2018), the 0.5 m depth balconies
are designed to house the outdoor air conditioning units, while balconies used as a liveable area,
connecting indoor and outdoor spaces, are usually 1 m to 2 m deep (Figure 3.3). The room configuration
shows that the office spaces from the selected sample tend to be small and normally used by 2 to 3
people. Parametric variations were applied to the four basic orientations to illustrate their effect in
daylight conditions for a lower, middle and top floor. The window width, one of the most impacting
parameters on daylight performance, was varied from 1 m up to 5 m (representing a fully glazed façade).
Thus, increments of 0.5 m were used to iterate between the minimum and maximum scenarios. Windows
and balconies were always considered as central to the room, increasing its width symmetrically for both
sides (Figure 3.4).
(a) Office building with balconies (b) Office room with balcony (c)
Balcony used to house the
condenser unit
Figure 3.3: Examples of office buildings with balconies
Source: Pereira (2019)
46
a) Office room floor plan: side
balcony
b) Office room floor plan:
front balcony
c) Office room section
Figure 3.4: Office room dimensions
The ratio of balcony width to window is proposed to evaluate the impact of the balcony’s width and the
corresponding window width, as illustrated in Table 3.3. This variable represents a façade compositional
rule based on a central axis symmetry, enabling the investigation of apertures in connection with their
corresponding daylight systems. This parameter allowed to simplify the number of cases when varying
balcony’s width and facilitated results comparison. Even though ratios 2 and 1 are considered more
usual in the building façades, the 0.5 ratio was also selected to have its performance evaluated as a
possible design scenario, despite not being commonly found in practice.
Table 3.3: Ratio of balcony width to window width
Window width (m)
Ratio of balcony width to window width
2
1
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
5.5
6
6.5
3.3.2 Daylight simulations
The plug-in Grasshopper (Mcneel, 2020) was used to model the base case geometry, which consists of
the entire building, as shown in Figure 3.2. Daylight simulations were run in Honeybee+ (Roudsari; Pak,
47
2013), a plug-in for Grasshopper that connects the Rhinoceros’ shape parameterization with Radiance
and Daysim for daylight simulation. To perform the simulations, the 2-phase Radiance-based method
was used. The Radiance input parameters were chosen through Honeybee+ based on the simulation
complexity, which was set as “medium complexity” due to the number of simulated cases (Table 3.4).
The grid configuration and sky density were set according to the IES LM 83-12 approved method (IES
LM 83-12, 2012). Simulations were run without any urban surrounding to assess daylight scenarios with
the worst condition for direct incident solar radiation.
Table 3.4: Radiance input parameters
Parameter
Value
Ambient bounces (ab)
5
Ambient divisions (ad)
15,000
Ambient resolution (ar)
64
Ambient super-samples (as)
2,048
Ambient accuracy
0.2
Work plane
Grid size: 0.5 m
Height: 0.8 m
Offset from the walls: 0.5 m
Sky density
Reinhart sky
Three climate-based metrics were used to assess the daylight performance of the office room and its
design variants: Useful Daylight Illuminance (UDI), Spatial Daylight Autonomy (sDA) and Annual
Sunlight Exposure (ASE). The UDI is defined as the annual occurrence of illuminances across the work
plane that are within a range considered “useful” by occupants, when artificial lighting is not necessary.
For office buildings, the UDI useful range was identified by Mardaljevic et al. (2012) as being 300-3000
lux, with the upper value considered a good proxy for excessive illuminance. The sDA is a measure of
daylight illuminance sufficiency for a given area, reporting a percentage of floor area that exceeds a
300-lux illuminance level for more than 50% of annual working hours (8 am to 6 pm). The sDA was
ranked in the following daylight sufficiency levels, according to IES LM-83 (2012) : preferred daylight
sufficiency (must meet or exceed 75% of the analysis area), nominally accepted daylight sufficiency
(must meet or exceed 55% of the analysis area), not accepted (sDA does not meet the minimum required
value of 55% of the analysis area). The ASE describes the potential for visual discomfort and determines
unsatisfactory visual comfort when its result is over 10% for daylit spaces (IES LM 83-12, 2012). This
metric can also be used as a proxy for overheating, as it means the percentage of the year each point in
space receives direct solar radiation.
The ASE and sDA metrics were reported together to provide a meaningful first-level understanding on
how a space is expected to perform, since sDA sets a minimum value for daylight sufficiency but not
any indication of an excess thereof whereas ASE sets a maximum value to prevent visual discomfort.
Daylight simulations were performed considering the occupancy period from 8 am to 6 pm, as suggested
by IES LM 83 (2012), with no user interference, i.e. considering the worst-case scenario with no blinds.
This setting is the same as the one from the Brazilian energy efficiency regulation (INMETRO, 2020)
48
which also considers an occupancy period 10 h per day, without a lunchtime break. The weather file
used to perform the simulations was a Typical Meteorological Year (TMY), based on weather data from
the years 2000-2010, available in an EnergyPlus weather file (epw) format for the city of São Paulo,
Brazil (LABEEE, 2018). A cross combination of all the seven parametric variations described in Table
2 were combined into 6,360 simulations, meaning all design solutions were explored. Thus, simulation
outputs for the three aforementioned daylight metrics comprised a total of 19,080 results.
3.3.3 Post-processing and data mining
Simulation post-processing was divided into two large steps (Figure 3.5). Initially a systematic analysis
in the extensive dataset presented in the Appendix A was undertaken, starting with a sensitivity analysis
of each parameter in the three CBDM (Section 3.4.1), followed by unfolding interesting pairwise
combinations for design decision-making (Section 3.4.2). Further explorations specifically on
identifying the role of balconies in daylight performance were undertaken via a combination of results
from the sensitivity analysis, the pairwise comparisons and the information contained in the Appendix
A, from which general rules were extracted and discussed (Section 3.4.3).
Figure 3.5: Results analysis diagram
Simulation results were grouped into six classes, for sDA and ASE only, using a data mining process so
full patterns of successful combinations of design parameters could be quickly retrieved to aid design
decision-making (section 4.4). Simulation results contained one nominal (solar orientation) and six
numeric attributes (floor level, room depth, glazing visible transmittance, ratio of balcony width to
window width, window width and balcony depth). Classes for nominal and numeric variables were
described in Table 3.5, with a respective traffic light system indicating results practical suitability. The
UDI metric was used specifically to further qualify cases in which the sDA was convergent above the
threshold.
49
Table 3.5: Data classification
sDA
ASE
Class
numeric
nominal
numeric
nominal
Higher or equal to
75%
Preferred
0%
No glare
sDA preferred no glare
(green)
Equal or lower than 10%
Tolerable glare
sDA preferred tolerable
glare (yellow)
Higher to 10%
Glare
Glare not accepted (red)
Lower than 75%
and higher or
equal to 55%
Accepted
0%
No glare
sDA accepted no glare
(yellow)
Equal or lower than 10%
Tolerable glare
sDA accepted tolerable
glare (orange)
Higher to 10%
Glare
Glare not accepted (red)
Lower than 55%
Not accepted
0%
No glare
sDA not accepted (red)
Equal or lower than 10%
Tolerable glare
Higher to 10%
Glare
Decision tree was considered the best data mining option to illustrate successful routes through
combinations of parameters which would lead to sDA and ASE respectively above and below thresholds
established in section 3.2 and further detailed in Table 3.5. Decision trees popularly known as ‘recursive
divide and conquer’ data mining methods, select the best attribute among a set of alternatives to produce
routes with maximum information gain. They start by statistically selecting the attribute for a root node,
to then creating branches for each possible value and splitting instances into sub-sets, recursively
repeating this process until each instance belongs to a class. The widely applied J48 algorithm (Witten;
Frank; Hall, 2014) was used as a classifier for the decision tree and could hierarchically organise 6,360
instances creating clear paths to achieve the desirable classes. The algorithm is based on a top-down
strategy and uses information gain to measure the amount of information provided by each attribute as
a basis to determine which one best splits the dataset at each step (Witten; Frank; Hall, 2014) The data
mining process was undertaken in WEKA (Witten; Frank; Hall, 2014), and the decision tree which
provided simultaneously a satisfactory level of accuracy and complexity, achieving an 85% correctly
classified instances was discussed in section 4.4, with its configuration synthesised in Figure 3.6.
Successful end nodes of the decision tree are highlighted using the traffic light system proposed in Table
3.5, so designers can visualize the best routes to achieve performance.
50
Figure 3.6: Decision tree configuration
Source: the authors, based on information provided by Witten; Frank and Hall (2014)
3.4 Results and discussion
3.4.1 Sensitivity analysis
Box plots were used to evaluate the sensitivity of each design parameter in sDA and ASE considering
the thresholds illustrated in Table 3.5. UDI graphs were plotted when necessary to verify sDA results,
especially when compressed around the upper threshold (sDA = 100%). Box plots for each parametric
variation were presented with different shades of grey, with the cross illustrating the mean and the line
within the box illustrating the median. The discussion attempted to extract relevant information for
design decision-making, i.e. to gauge the impact of specific design parameters in daylight performance
as well as to identify relevant dimensions to achieve specific performance thresholds.
a) Window width
Window width seems to be the determinant parameter on office rooms’ daylight performance. These
findings echo the results shown by Al-Sallal et al. (2018) and Liu and Chen (2017). All cases with
windows width between 5 m and 6.5 m reached the sDA preferred class together with more than 50%
of the cases with windows widths between 3.5 m and 4.5 m (Figure 3.7a). Cases with narrow windows
(1 m and 1.5 m width) presented the best potential to prevent glare (Figure 3.7b) by keeping ASE results
close to null but all windows up to 3 m width still complied with ASE below 10% as well as 75% of
windows with up to 6 m width. UDI results confirmed that the room’s daylight performance was directly
proportional to the window width (Figure 3.7c) and indicated that windows with widths above 5.5 m
provided nearly the same performance results (see three light grey box plots from Figure 3.7c). This was
partially confirmed by the ASE figures which showed almost 3/4 of results falling within the 10%
51
threshold, meaning windows above 6.5 m width would require more careful attention with regards to
shading design.
(a) sDA
(b) ASE
(c) UDI
Figure 3.7: Dataset results for the window width
b) Glazing visible transmittance
While the majority of the cases (62%) with clear glass (Glazing visible transmittance =0.88) achieved
the sDA preferred class, this class was reached by only 24% of the cases with laminated glass (glazing
visible transmittance = 0.48) (Figure 3.8a). However, the glazing visible transmittance showed less
impact on ASE than on sDA results. While 80% of the cases with laminated glass were classified as
ASE=0%, 68% of the cases with clear glass achieved this threshold (Figure 3.8b), with more than 3/4
of cases with this type of glass falling within the ASE 10% threshold. UDI results confirmed sDA ones
but showed that results for clear glass have an even higher impact on performance when compared to
results for laminated glass (Figure 3.8c).
52
(a) sDA
(b) ASE
(c) UDI
Figure 3.8: Dataset results for the glazing visible transmittance
c) Balcony depth
Figure 3.9 illustrates the decrease in daylight illuminance levels resulting from the increase in balcony
depth, as confirmed by Liu and Chen (2017), Gábrova (2014) and Kim and Kim (2010b). Figure 3.9a,
however, illustrated an interesting cut-off point for design decision-making as balcony depths below 1.0
m had, in the majority of cases, sDA values falling within the preferred threshold whereas balcony
depths above 1.0 m would tend to have sDA values falling, on average, within the accepted threshold.
As expected from Al-Sallal, AbouElhamd and Dalmouk (2018), Kim and Kim (2010a), Gábrova (2014),
Xue et al. (2016) and Dahlan et al. (2009a, 2009b), the addition of balconies was determinant in reducing
the ASE (Figure 3.9b). However, this study identified that balconies deeper than 1.5 m will achieve null
ASE and therefore are optimum to avoid glare and overheating due to direct solar radiation. UDI results
(Figure 3.9c) confirmed sDA ones, also showing that the upper UDI limit is not affected by the balcony
depth, since the 3rd quartile is roughly the same for all cases.
(a) sDA
(b) ASE
(c) UDI
Figure 3.9: Dataset results for the balcony depth
53
d) Room’s width and depth
Room width and depth were investigated simultaneously by changing balcony and window positions
from the front and back façades to side façades, as displayed in Figure 3.4, reflecting the most common
floor plan proportions for mixed-mode office spaces in São Paulo (Pereira; Neves, 2018). As confirmed
by Gábrova (2014), results clearly showed that the shallower the room, the better the daylight
performance. More than 50% of the front balcony cases (room depth 5.5 m) were classified as sDA
preferred, while only 29% of the side balcony cases (room depth 7 m) achieved this threshold. UDI
levels confirmed sDA values for shallower rooms showing however, that room depth does not really
affect the shape of the distribution curve as average cases will have a UDI of 55% for shallower rooms
and 45% for deeper rooms (Figure 3.10c). ASE results were however not significantly different as the
vast majority of cases for both configurations fell within the 10% threshold (Figure 3.10b), none of them
with means in the null category.
a) sDA
b) ASE
c) UDI
Figure 3.10: Dataset results for the room depth
e) Solar orientation
In the Southern tropical climate, the North and the South façades receive, respectively, the highest and
the lowest amount of direct and diffuse solar radiation during the year. Thus, the office room facing
North showed the highest daylight levels (Figure 3.11a) but also the highest probability of glare (Figure
3.11b). More than 50% of the North-oriented rooms were classified as having preferred sDA level and
tolerable glare from ASE. As to the UDI level, the North-oriented office rooms presented higher results
for the 1st quartile, the median and the mean values, if compared to the other solar orientations, showing
the potential for the North orientation to achieve the best daylight performance (Figure 3.11c). In
opposition, the South-oriented rooms showed the best results for ASE and the lowest mean and median
values for sDA and UDI, with mean and median values for the former barely achieving the acceptable
threshold. The West and East façades exposed similar daylight performance to each other. The mean
and median values for both solar orientations were classified as sDA acceptable and ASE tolerable,
54
although the West façade resulted in higher levels of ASE, possibly due to the fact that the number of
occupied hours in the afternoons is higher than in the mornings.
a) sDA
b) ASE
c) UDI
Figure 3.11: Dataset results for the solar orientation
f) Ratio of balcony width to window width
The use of balconies wider than the windows (ratio of balcony width to window width = 2) decreased
the daylight performance (Figure 3.12a) but also prevented glare (Figure 3.12b), echoing results found
by Kim and Kim (2010a), Gábrova (2014), Al-Sallal, Abouelhamd and Dalmouk (2018). The ratio of
balcony width to window width of 1 and 2 showed similar performance for ASE, with most part of the
cases classified as null, reinforcing the importance of having balconies with full window width. UDI
results again confirmed sDA ones in terms of how ratio of balcony width to window width affect not
only the average figures but also their distribution (Figure 3.12c).
(a) sDA
(b) ASE
(c) UDI
Figure 3.12: Dataset results for the Ratio of balcony width to window width
55
g) Floor level
Results from daylight simulations indicated that the floor level was the parameter with least impact on
the room’s daylight performance (Appendix A), findings echoed by Liu and Chen (2017), likely to be
related to the fact that the building was simulated without any urban surroundings. Higher floors
indicated a small increase in sDA and UDI in relation to lower floors possibly due to the albedo setting.
The presentation of box plots for this parameter was therefore deemed unnecessary.
Nevertheless, in the case of densely built urban neighbourhoods, results would differ between higher,
intermediary and lower floors. However, the development of suitable urban environments to undertake
these experiments is still open to discussion as cities, especially in Brazil, have a very heterogeneous
urban context meaning multiple types of geometric combinations for building height and lower floor
configurations can be expected, making it difficult to extract typical cases to standardise the simulation
of surrounding buildings.
3.4.2 Unfolding interesting pairwise combinations
To further extract relevant information for design decision-making, parameters were also analysed in
pairs. Scatterplots were used to depict the most relevant pairwise comparisons, i.e. the ones from which
it was possible to extract cut-off points for both parameters in relation to different performance
thresholds. This section explored pairwise comparisons for the sDA metric only as ASE and UDI did
not reveal any information different than the one received from the box plots. The analysis is focused
on daylight illuminance sufficiency and does not include excessive illuminance (glare probability) or
overheating probability issues.
a) Window width and glazing visible transmittance
When plotting window width against glazing visible transmittance it is possible to see that all windows
wider than 4 m with clear glass achieved the sDA preferred threshold, whereas only windows wider than
6 m with laminated glass achieved this same threshold (Figure 3.13).
(a) Glazing Tvis 0.48
(b) Glazing Tvis 0.88
Figure 3.13: Scatter plots for sDA per window width
56
b) Window width and room depth
All the 5.5 m-deep rooms achieved preferred sDA thresholds when having windows wider than 5.5 m,
whereas the 7 m-deep rooms did not achieve the preferred sDA threshold 100% of the time, even with
a fully glazed facade (Figure 3.14).
(a) Room depth = 5.5 m
(b) Room depth= 7 m
Figure 3.14: Scatter plots for sDA per window width
c) Window width and balcony depth
This pairwise combination showed that the deeper the balcony, the narrower the difference between
sDA values falling within the preferred and acceptable thresholds. Figure 3.15 illustrated that when no
balconies are present, the preferred threshold was achieved for all the scenarios with a 5 m width window
whereas only a 3.5 m width window was necessary to achieve the acceptable one. Adding a balcony of
0.5 m and 1 m depth did not affect the window width needed to achieve the preferred threshold but did
push the minimum width to achieve the accepted threshold to 4 m and 4.5 m, respectively. For balconies
deeper than 1 m, for every increment of 0.5 m in depth, the distance between the preferred and acceptable
thresholds seemed to remain constant.
(a) Balcony depth = 0 m (without balconies)
(a) Balcony depth = 0.5 m
57
(c) Balcony depth = 1 m
(d) Balcony depth = 1.5 m
(e) Balcony depth = 2 m
Figure 3.15: Scatter plots for sDA per window width
d) Window width and solar orientation
This pairwise combination, illustrated in Figure 3.16, showed that for the North orientation, the preferred
sDA threshold was achieved for all scenarios with a 5 m window width whereas a 4 m window width
was enough for all scenarios to achieve at least the acceptable sDA threshold. For the South façade, the
preferred and acceptable thresholds were achieved with window widths of respectively 6 m and 5 m for
all scenarios. The East and West orientations exhibited window width differences 0.5 m apart for all
scenarios to achieve the preferred and acceptable thresholds.
58
a) North
b) South
c) East
d) West
Figure 3.16: Scatter plots for sDA per window width
3.4.3 Extracting general rules
Results from section 3.4.1 showed the sensitivity of CBDM to each design parameter explored in this
study when analysed in isolation, whereas results from section 3.4.2 attempted to unfold interesting
pairwise combinations of parameters to examine how they, together, influenced sDA in particular.
Whereas section 3.4.1 indicated that window width was potentially the most important parameter to
achieve the targets, this was confirmed by the pairwise comparisons which showed that interesting
thresholds could be identified when combining this parameter with the others. This section examined
the summaries from section 3.4.1 and 3.4.2 in conjunction with Appendix A and attempted to provide
general rules useful to designing medium rise mixed-mode office buildings in São Paulo, with or without
balconies.
Table 3.6 shows general rules for window widths to achieve the sDA preferred threshold, for any
balcony depth (from 0 m to 2 m) and ratio of balcony width to window width (0.5, 1 and 2) considered
in this study. Window widths were established as a function of solar orientation, room depth and glass
visible transmittance. The North façade contained the path to success with minimum dimensions,
59
whereas the South orientation presented the worst-case scenario, needing the largest window widths for
balconies to be used without affecting daylight performance. East and West orientations with shallow
rooms and clear glass needed both 3 m window widths to achieve the preferred threshold but behaved
differently when the room depth increased, and the glass visible transmittance decreased. No
configuration achieved the preferred threshold when the room depth was 7 m and laminated glass was
used.
Table 3.6: Rules to achieve preferred sDA for any balcony configuration
Solar
orientation
Room
depth (m)
Glass Tvis Window width to achieve sDA preferred threshold (m)
North
5.5
0.88
>= 2.5
0.48
>= 4.5
7.0
0.88
>= 3.0
0.48
-
South
5.5
0.88
>= 3.5
0.48
>= 6.0
7.0
0.88
>= 4.0
0.48
-
East
5.5
0.88
>= 3.0
0.48
>= 6.0
7.0
0.88
>= 3.5
0.48
-
West
5.5
0.88
>= 3.0
0.48
>= 5.0
7.0
0.88
>= 4.0
0.48
-
Table 3.7 suggests the window width above which using deeper balconies improved UDI figures, i.e.
when balconies were considered effective daylight diffusers. Deeper balconies could be particularly
difficult to be used in the South façade and would improve performance only when used in shallow
rooms with clear glass. It is important to notice that deeper balconies would never improve daylight
performance of any configuration using laminated glass (Tvis = 0.48), therefore these values were not
added to Table 3.7.
Table 3.7: Rules to achieve the same UDI for any balcony configuration and to improve UDI using deeper
balconies.
Solar
Orientation
Room depth
(m)
Glass Tvis
Window width above which UDIs improve with
deeper balconies (m)
North
5.5
0.88
>= 3.5
7.0
>= 4.0
South
5.5
>= 5.0
7.0
-
East
5.5
>= 4.5
7.0
>= 4.5
West
5.5
>= 4.0
7.0
5.0
60
3.4.4 Data mining: Relevant combinations of design parameters to improve daylight performance.
Sections 3.4.1 to 3.4.3 showed limitations in further detailing causal relationships and extracting more
specific rules when undertaking a systematic analysis. Therefore, this section focused on expanding this
analysis to improve the search for successful routes towards sDA and ASE respectively above and below
thresholds, through the use of data mining. As previously detailed in the methodology, Figure 3.17
depicted successful end nodes for a decision tree, produced by the J48 algorithm, using a traffic light
system (discussed previously in Table 3.5), so designers could visualize the best routes to achieve any
desired performance.
Figure 3.17: Decision tree
The decision tree from Figure 3.17 was achieved after multiple experiments in WEKA. Different
decision tree settings were investigated altering the number of objects, to control complexity. A
minimum number of objects of 3, yielded a confidence factor as high as 97%. However, these settings
resulted in highly complex trees, with 298 end nodes, difficult to parse through and complex to be
analysed. The tree displayed in Figure 3.17 provided simultaneously a satisfactory level of correctly
classified instances (85%) and complexity and was achieved with the following settings: pruned, number
of folds set to 3 and minimum number of objects set to 30. Decision tree paths led to 57 end nodes, from
which 28% of them were highly successful (sDA preferred no glare), whereas 23% of them led to
unacceptable results (both glare not accepted and sDA unacceptable) and should therefore be avoided.
Thus, 77% of nodes were within the acceptable and/or tolerable thresholds and therefore yielded valid
decision-making paths to be pursued.
61
Results depicted by the decision tree and summarised in Table 3.8, show general rules for which
combination of parameters are likely to yield preferred sDA with no glare as well as which combinations
of parameters should be avoided as they lead to either sDA or glare at not acceptable levels. Design
parameters which do not belong to a specific rule marked as ‘-‘. This table contains 29 rules from which
12 display combinations which should be avoided with the remaining ones listed deemed as highly
successful. However, it is important to notice that window widths smaller than 3 m are not allowed by
the São Paulo Building Regulation (1978), despite showing acceptable results. This means, in principle,
non-listed combinations should lead to the achievement of acceptable targets, bearing in mind the
classification correctness of 85%.
Table 3.8: Rules extracted from the decision tree
Class
Window
width (m)
Glazing
Tvis
Solar
orientation
Balcony
Depth (m)
Room
depth (m)
Ratio of balcony
width to window
width
sDa
preferred
no glare
> 2 to <= 3
0.88
East or
West
<=1.0
5.5
-
>2 to <= 3
0.88
South
=1.0
-
-
>2 to <=3
0.88
-
>1.0
5.5
-
= 3
0.88
-
>1.0
7.0
-
> 3
0.88
South
<=1.0
-
-
> 3
0.88
West and
East
=1.0
-
>0.5
> 3
0.88
-
>1.0
-
>0.5
> 3
0.88
South or
East
>1.0
-
-
>3 to <= 4.5
0.88
West
>1.0
-
-
> 4
0.48
East
-
7.0
<=0.5
> 3 to <=4
0.48
West
-
5.5
<=0.5
>4 to <=5
0.48
South
<=1.0
5.5
-
>4
0.48
East
>0.5
5.5
-
>4
0.48
West
-
5.5
-
>4
0.48
North
<=1.0
5.5
>0.5
>5
0.48
South
-
5.5
-
>5
0.48
South
-
5.5
-
sDA not
accepted
<=1.5
0.88
-
-
-
-
<=2.5
0.48
-