Modelling and simulation of virtual natural lighting solutions in buildings

Modelling and simulation of virtual natural lighting solutions inbuildingsCitation for published version (APA):Mangkuto, R. A. (2014). Modelling and simulation of virtual natural lighting solutions in buildings. Eindhoven:Technische Universiteit Eindhoven. https://doi.org/10.6100/IR772754

DOI:10.6100/IR772754

Document status and date:Published: 01/01/2014

Document Version:Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can beimportant differences between the submitted version and the official published version of record. Peopleinterested in the research are advised to contact the author for the final version of the publication, or visit theDOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and pagenumbers.Link to publication

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, pleasefollow below link for the End User Agreement:www.tue.nl/taverne

Take down policyIf you believe that this document breaches copyright please contact us at:[email protected] details and we will investigate your claim.

Download date: 05. Jun. 2020

https://doi.org/10.6100/IR772754

https://doi.org/10.6100/IR772754

https://research.tue.nl/en/publications/modelling-and-simulation-of-virtual-natural-lighting-solutions-in-buildings(4323f211-2b74-4416-98e8-379dfd58f8aa).html

Modelling and Simulation of Virtual

Natural Lighting Solutions in Buildings

PROEFSCHRIFT

ter verkrijging van de graad van doctor

aan de Technische Universiteit Eindhoven, op gezag van de rector magnificus, prof.dr.ir. C.J. van Duijn,

voor een commissie aangewezen door het College voor Promoties in het openbaar te verdedigen

op woensdag 14 mei 2014 om 16.00 uur door

Rizki Armanto Mangkuto

geboren te Bogor, Indonesië

Dit proefschrift is goedgekeurd door de promotoren en de samenstelling van de promotiecommissie is als volgt:

voorzitter: prof.dr.ir. J.J.N. Lichtenberg 1e promotor: prof.dr.ir. J.L.M. Hensen 2e promotor: prof.dr.ir. E.J. van Loenen copromotor: dr.ir. M.B.C. Aries leden: prof.dr. J. Mardaljevic (Loughborough University)

prof.dr. M. Andersen (École Polytechnique Fédérale de Lausanne) prof.dr. E.H.L. Aarts

dr.ir. Y.A.W. de Kort

Modelling and Simulation of Virtual Natural Lighting Solutions in Buildings

Rizki A. Mangkuto

The work described in this thesis has been carried out in the Unit Building Physics and Services at the Department of the Built Environment, Eindhoven University of Technology. This research was supported by the Sound Lighting research line of the Intelligent Lighting Institute (ILI) at Eindhoven University of Technology. Copyright © 2014 by Rizki A. Mangkuto Eindhoven University of Technology, the Netherlands All rights reserved. No part of this document may be photocopied, reproduced, stored, in a retrieval system, or transmitted, in any form or by any means whether, electronic, mechanical, or otherwise without the prior written permission of the author. A catalogue record is available from the Eindhoven University of Technology Library ISBN: 978-90-386-3604-7 NUR: 955 Bouwstenen 194 Cover design provided by A. Davie (Modul8) and adapted by P. Verspaget. Printed by Gildeprint Drukkerijen, Enschede, the Netherlands. Modelling and Simulation of Virtual Natural Lighting Solutions in Buildings / by Rizki A. Mangkuto – Eindhoven University of Technology – proefschrift – Subject headings: Virtual Natural Lighting Solutions / virtual window prototype / computational modelling / lighting simulation / visual comfort / building performance

“… But it is possible that you hate a thing that is good for you,

and that you love a thing that is bad for you.

Allah knows, while you know not.”

Al-Qur’an, 2: 216

To the memory of my dear mother

and to my dear wife,

for their endless and invaluable love,

just like the sun shining over the world

…

“Jiwaku tetap mengabdi pada Ibunda.

Dan aku pun tetap Timur, Adinda.”

Nur St. Iskandar (1922)

vii

Table of Contents

Acknowledgements xi

Summary xv

Samenvatting xvii

Ikhtisar xix

Nomenclature xxi

Chapter 1 – Introduction 1

1.1. Natural Lighting Demand 1

1.2. Shortcomings of Natural Lighting 3

1.3. Proposed Solution 4

1.4. Aim and Objectives 8

1.5. Research Methodology 8

1.6. Thesis Outline 10

Chapter 2 – General Concept of Virtual Natural Lighting Solutions 13

2.1. Definition of VNLS 13

2.2. Classification of VNLS 13

2.2.1. Prototypes with a simplified view 14

2.2.2. Prototypes with a complex view 16

2.3. Expectation of VNLS 19

2.3.1. Light quality 19

2.3.2. View quality 22

2.3.3. Target range 24

2.3.4. Comparison of prototypes 25

2.4. Concluding Remarks 29

Chapter 3 – Measurement and Simulation of a First Generation Virtual Natural Lighting Solutions Prototype 31

3.1. Introduction 31

3.1.1. Modelling concept in Radiance 31

Table of Contents

viii

3.2. Case Description 34

3.3. Measurement Protocol 37

3.4. Simulation Protocol 38

3.5. Results and Discussion 42

3.5.1. Measurement of prototype 42

3.5.2. Simulation of prototype 44

3.5.3. Simulation of real windows 46


Chapter 4 – Discomfort Glare Evaluation and Simulation of a First Generation Virtual Natural Lighting Solutions Prototype 51


4.2. Method 56

4.2.1. Model description 56

4.2.2. Glare rating correlation 57


4.3.1. Rendering and glare source detection 58

4.3.2. Unadjusted rating 60

4.3.3. Polynomial regression 60

4.3.4. Adjusted rating 62

4.3.5. Percentage of disturbed subjects 66


Chapter 5 – Design, Measurement, and Simulation of a Second Generation Virtual Natural Lighting Solutions Prototype 71


5.2. Design Method 72

5.2.1. Test environment 72

5.2.2. Light sources 73

5.2.3. Control circuit 75

5.2.4. Display and structure 77

5.2.5. Programming and setting 79

Table of Contents

ix

5.3. Measurement Protocol 80

5.4. Simulation Protocol 82

5.4.1. Model description 82

5.4.2. Validation 83

5.5. Analysis of Various Configurations 84

5.6. Analysis of Various Operating Scenarios 85

5.6.1. Settings and data collection 85

5.6.2. Daily profiles and annual modes 86


5.7.1. Measurement of actual test room 91

5.7.2. Simulation of actual test room 95

5.7.3. Comparison of various configurations 97

5.7.4. Comparison of various operating scenarios 99


Chapter 6 – Modelling and Simulation of a Virtual Natural Lighting Solutions with a Simplified View and Directional Light 107


6.2. Methods 108

6.2.1. Modelling 108

6.2.2. Settings 113

6.2.3. Assessment 114


6.3.1. Sensitivity analysis 121

6.3.2. Comparison with real windows 123


Chapter 7 – Modelling and Simulation of a Virtual Natural Lighting Solutions with Complex Views and Directional Light 131


7.2. Methods 134

7.2.1. Modelling 134

Table of Contents

x

7.2.2. Settings 136

7.2.3. Assessment 136


7.3.1. Transmissive approach 139

7.3.2. Comparison of transmissive and emissive approaches 145


Chapter 8 – Conclusions and Recommendations 151

8.1. Conclusions 151

8.2. Recommendations 154

References 157

Appendices 169

Curriculum Vitae 187

Publication List 188

xi

Acknowledgements

All praise is due to Allah, the Light upon light, the Lord of the heaven and earth and everything in between; for without Him, no single particle of light will ever exist in the universe, and nor will this small thesis.

It was a little bit more than four years ago, when I considered submitting an online application for a doctoral research position on the topic of modelling and simulation of something I had never heard about: Virtual Natural Lighting Solutions. Knowing that such a position in the topic of lighting can be relatively rare, and that no external scholarships were required from my side to run the project, I had no doubt to proceed with applying. It was then followed with a short, phone interview at almost midnight (western Indonesia time). The outcome was destiny, a move to the City of Light in the ‘Land of Windmills’.

I would therefore like to express my gratitude to my first promotor, Prof. Dr. Ir. Jan Hensen, who actually spoke on the other side of the line in that crucial phone interview, for providing the invaluable opportunity to conduct the doctoral research, and for giving the trust on me to execute the work in his leading, internationally diverse and recognised group of Building Performance (simulation). Many of his (former) doctoral candidates, including myself, appreciate the monthly progress meetings. Not only because the meetings give plenty of opportunity to practice presentation skill and to learn from each other, but also because they ensure us to stay on track in the progress; since as one says, it is very easy to get lost in the jungle of research.

I would also express my gratitude to my second promotor, Prof. Dr. Ir. Evert van Loenen, for his ideas, suggestions, and most of all his detailed comments on the experiments, simulations, and any writings that were related with this doctoral research. Among others, he also provided opportunity and support to conduct the experimental setup and measurements of prototypes in the (old and new) ExperienceLab of Philips Research at the High Tech Campus Eindhoven. The work there has proven to be an integral part of the project and significantly contributed to the build-up of this thesis.

My gratitude and thanks also go to my copromotor, Dr. Ir. Myriam Aries, for her advice, suggestions, and recommendations during the course of this long-term project. Many of her inputs were so crucial in determining the direction of the project, which I appreciate very much. Outside the office, I also had some nice chance to get in touch with her family and I really enjoyed that.

Furthermore, I would like to express my appreciation to Prof. Dr. John Mardaljevic (Loughborough University, the United Kingdom), Prof. Dr. Marilyne Andersen (École Polytechnique Fédérale de Lausanne, Switzerland), Prof. Dr. Emile Aarts (Eindhoven Uni-versity of Technology), and Dr. Ir. Yvonne de Kort (Eindhoven University of Technology)

Acknowledgements

xii

for willing to sit in the Doctorate Committee and for giving their invaluable comments on this thesis.

This project was supported by the Sound Lighting research line of the Intelligent Lighting Institute (ILI) at Eindhoven University of Technology. The works with prototypes were mostly conducted with the help of Bernt Meerbeek, MSc, PDEng and other colleagues in Philips Research at the High Tech Campus Eindhoven.

In the Unit of Building Physics and Services, my thanks are given to the past and present colleagues in the group of Building Performance, in alphabetical order of the first name: Ana Paula Melo, Azzedine Yahiaoui, Bruno Lee, Chul-sung Lee, Daniel Cóstola, Giovanni Pernigotto, Hamid Montazeri, John Bynum, Katarína Košútová, Marcel Loomans, Massimo Chiappini, Meng Liu, Mike van der Heijden, Mohammad Mirsadeghi, Mohammed Hamdy, Petr Zelensky, Pieter-Jan Hoes, Qimiao Xie, Rajesh Kotireddy, Rebeca Barbosa, Roel Loonen, Song Pan, Vojta Zavrel, and Wiebe Zoon. Among them, I particularly acknowledge the suggestions from and brief discussion with Hamid (also thanks for the coffee-break chats, proof-reading the thesis, and giving tutorial in creating high quality figures), Daniel, Roel, and Wiebe, at certain points at some time within my stay in the group. My thanks are also given to the colleagues in the group of Building Lighting: Mariëlle Aarts, who together with Myriam had kindly provided the samenvatting of this thesis; Carlos Ochoa Morales, who was partially involved in giving directions during the first two years of my project; and Parisa Khademagha, who just started her doctoral research in the group, also under the umbrella of ILI.

My next thanks are given to the colleagues, particularly the ‘senior members’ of the group of Building Material, with whom I shared the open-plan working space in almost the first two years of my projects: Qingliang Yu, Przemek Spiesz, Miruna Florea, Alberto Lazaro Garcia, George Quercia Bianchi, Azee Taher, and Štěpán Lorenčik. I thank Przemek in particular for his advice on my trip to Kraków and Oświęcim; and Qingliang for the rather unique research collaboration on the application of modelling and simulation in Radiance for photocatalytic material. My appreciations also go to the past and present secretaries of the unit: Renée van Geene, Yeliz Varol, Janet Smolders, and Ginny Vissers; and to the staff in HR service and International Office, particularly Carol van Iperen and Kara de Rooy. My special gratitude is expressed to the head of BPS Laboratory, Jan Diepens, for providing Radiance in Oracle VM VirtualBox and for providing the measurement devices; and to Harry Smulders for setting up the connection to the UNIX server.

During my project, I had the honour to supervise three graduation projects of master students: Chris van Dronkelaar, Ruben Pelzers, and Shen Wang; two pre-graduation projects of master students: Richard Claessen and Annelous Bossers; and an honour project of two bachelor students: Bas Kil and Martijn Gootzen. I thank all of them for the cooperation. In particular, I express my gratitude to Wang, whose passion in lighting design and technology had significantly contributed to two chapters of this thesis.

Acknowledgements

xiii

Living in a country faraway from home can be dull without a nice and welcoming social environment. I hereby would like to appreciate all of the Indonesian families and friends in Eindhoven and surrounding, for their support to my family and myself in many ways. To all of them, I would like to thank; terima kasih banyak dan sampai berjumpa lagi di lain kesempatan.

In the Laboratory of Building Physics and Acoustics at Engineering Physics Research Group, Institut Teknologi Bandung, I would like to acknowledge my respected gurus that have shaped my academic experience, among others: Prof. Dr. Ir. R.M. Soegijanto, Dr. Ir. F.X. Nugroho Soelami, Dr. Ir. Joko Sarwono, and Dr. Iwan Prasetiyo. I am very grateful for have been learning from and working with them in the past, and will always look forward to doing so again in the future.

As the Prophet said, heaven is beneath the sole of one’s mother. I am greatly indebted to my mother, Saptarina (Allahu yarham), no mountains of gold can ever pay her unconditional love. It is my biggest regret not to have her here to witness the completion of this thesis; and I pray to Allah for keeping her in the best place in His heaven. To my father, Harry Mangkuto, I am also indebted for his love and support. To my sister, Puti Kemalasari, and my brother-in-law, Hario Rahadanto, I thank them for keeping on contact while separated in time and space. To my parents-in-law, Asmiati dan Arman Nur, I give my greatest respect and gratitude. To my aunt and uncle, Solita dan Santo Koesoebjono, I respectfully thank them for their support during my stay in the Netherlands, and for warmly welcoming us whenever we visited them in Wassenaar. To my grandmother, Sophie Sarwono, in Bogor and the rest of my big family in Padang and Jakarta and other various places in Indonesia, I thank and appreciate their support, prayers, and wishes during my period of stay here.

To my two boys, Mufid Raqillasyah dan Mirzavan Raskha, it is a great pleasure and indescribable experience to witness their birth (both in Eindhoven) and watch them constantly growing day by day. They are indeed the living time indicators: Mufid was born only a year after I started my project, whereas Raskha was born when I was struggling with writing and refining all of the messy chapters. I give my prayers for their health, safety, and well-being; and hope that their period of stay in Eindhoven becomes an important and somehow memorable part of their life.

Finally, to my dear and beloved wife, Asmelia, who always gives her infinite and endless love, care, support, advice, and patience, despite so little I could only give her in return: no words of thanks could describe my gratefulness. This thesis is dedicated to her, who has dedicated herself to the sake of her family. Bunda, tesis ini Ayah persembahkan untuk Bunda!

Eindhoven, March 2014 Rizki A. Mangkuto

Acknowledgements

xiv

xv

Summary


In situations where daylight is not or insufficiently available, the concept of Virtual Natural Lighting Solutions (VNLS), which are systems that can artificially provide natural lighting as well as a realistic outside view, can be promising. The benefit of installing VNLS in a building is the possibility to use more floor area that currently has very limited or no access to daylight, with the additional possibility to control the light and view quality.

The aim of this research is predicting the impact of various VNLS applications on lighting performance and visual comfort in buildings. Design methods including rapid prototyping and multiple design concepts using computational modelling and building performance simulation are proposed to obtain a better design of VNLS. The development process is based on existing VNLS prototypes, which are further improved by applying the relevant simulation tool.

Throughout the process, the following steps are carried out: (1) literature review of current development of VNLS prototypes; (2) measurement of VNLS prototypes, which are presented in case studies of a prototype with diffuse light (the so-called ‘first generation’ prototype) and a prototype with directional light (the so-called ‘second generation’ prototype); (3) simulation of VNLS prototypes and models, which are performed to predict the lighting performance, measurements of the prototypes are incorporated to validate the VNLS model; (4) computational modelling of future VNLS, which involves arrays of small light sources with various tilt angles to deliver the light in various directions; (5) sensitivity analysis, to understand the influence of the relevant input parameters to the relevant performance indicators of the VNLS model; (6) performance comparison with real windows scenes in simulation, to understand the benefit of installing VNLS in a given space; and (7) theoretical calculation, to estimate the annual space availability and electrical energy consumption of a VNLS prototype under various operating scenarios.

The most important results are summarised as follows:

• Based on the conducted measurement of a first generation VNLS prototype displaying simplified sky scenes in a test room with no façades, it is known that the provided settings do not satisfy the minimum lighting criteria. Based on Radiance lighting simulation, the investigated prototype performs better in terms of light distribution uniformity than a corresponding, hypothetical real window under overcast or partly cloudy scenes. Under the clear sky scene, the difference between the real and virtual windows is less, due to the influence of direct sunlight.

• A method is proposed to correlate the commonly applied glare metrics, which can be predicted using simulations, to the glare rating that was used in an experiment conducted

Summary

xvi

elsewhere, to assess discomfort glare from a first generation VNLS prototype with various complex views. The results suggests the simulated, normalised values of DGI, UGR, and CGI using Evalglare are all overestimated relative to the values converted from the experiment data, while the simulated values of DGP are in a better agreement with the converted DGP values.

• A second generation VNLS prototype has been designed and built. This prototype was assessed to validate the computational model that can be extended for further development of non-existing VNLS. Arrays of LED tiles were installed to provide diffuse light and a simplified view, whereas arrays of linear LED fixtures were installed to provide direct light that yielded some visible patches on the side walls of the test room. Simulation and measurement values of horizontal illuminance at certain distances were evaluated and showed a good agreement. The space availability can be optimised by placing a prototype on each short wall facing each other, or by placing two prototypes on a long wall. All of the investigated operating scenarios yield a relatively similar impact on the average annual space availability and electrical energy consumption of the prototypes.

• More complex VNLS configurations composed of small light emitting sources have been developed in computational model. The model has a simplified view, delivering the light from the ‘ground’ to the ceiling and from the ‘sky’ to the floor. Sensitivity analysis shows that total luminous flux of the ‘sky’ largely influences the space availability of the test room, whereas the source beam angle largely influences other output variables, including discomfort glare. Most of the modelled VNLS with a beam angle of 76° perform the closest to the corresponding real windows. The performance can be optimised by increasing the beam angle to 114°, yielding a space availability of around two times larger than the corresponding real windows.

• A model of VNLS configurations with complex views has been created, where the light was provided by arrays of white-coloured directional light emitting sources, while the view was provided by pasting a two-dimensional image on a transparent glass in front of the light sources. Comparisons are shown between 10 image scenes. The use of the transmissive approach in this point offers more flexibility to apply complex views on the display, but also requires more light to satisfy the illuminance criteria. On the other hand, the use of emissive approach in the previous point may introduce more light, but the view complexity is limited by the number of pixels.

xvii

Samenvatting

Modellering en Simulatie van Virtuele Natuurlijke Verlichtingsoplossingen in Gebouwen

In situaties waarbij geen of onvoldoende daglicht aanwezig is, is het concept van Virtuele Natuurlijke Verlichtingsoplossingen (VNLS) veelbelovend. VNLS zijn systemen die op kunstmatige wijze zowel natuurlijk licht als een realistisch uitzicht bieden. Het voordeel van het toepassen van VNLS in een gebouw is dat een groter vloeroppervlak met daglicht-kwaliteit gebruikt kan worden dat eerder een beperkte of geen toegang tot daglicht had. Het biedt tevens de mogelijkheid om het licht te regelen als ook de kwaliteit van het uitzicht.

Het doel van dit onderzoek is het voorspellen van de invloed van VNLS toepassingen op de lichttechnische prestatie en het visuele comfort in gebouwen. Ontwerpmethodes waaronder ‘snelle prototypering’ en meerdere ontwerpconcepten, waarbij gebruik gemaakt wordt van computermodellering en -simulatie van gebouwprestatie, zijn toegepast om het ontwerp van VNLS te verbeteren. Het ontwikkeltraject is gestart op basis van bestaande VNLS prototypes en is middels toepassing van relevante simulatiemiddelen ontwikkeld.

Gedurende het proces zijn de volgende stappen genomen: (1) literatuuronderzoek naar de huidige ontwikkeling van VNLS prototypes; (2) metingen aan bestaande VNLS prototypes, weergegeven in case studies van een prototype met diffuus licht (eerste generatie prototype) en een prototype met gericht licht (tweede generatie prototype); (3) simulatie van niet-bestaande VNLS prototypes en modellen, uitgevoerd om de lichtprestatie te voorspellen naast metingen van de bestaande prototypes welke een onderdeel vormen van de validatie van het VNLS model; (4) computermodellering van de VNLS van de toekomst bestaande uit de simulatie van een verzameling van kleine lichtbronnen met diverse richtingshoeken om licht vanuit verschillende richtingen te creëren; (5) gevoeligheidsanalyse om de invloed van de inputparameters op de relevante prestatie-indicatoren van het VNLS model te kwantificeren; (6) prestatievergelijking van VNLS met een eveneens gesimuleerd echt raam om inzicht te krijgen in het voordeel van het installeren van VNLS in een ruimte; en (7) theoretische berekeningen om de jaarlijkse ruimte-beschikbaarheid en het elektrische energieverbruik van een VNLS prototype onder verschillende gebruiksscenario’s in te schatten.

Een opsomming van de belangrijkste resultaten is de volgende:

• Uit metingen aan een eerste generatie VNLS prototype met vereenvoudigde hemelscènes in een testruimte zonder gevels is gebleken dat de hierbij beschikbare instellingen niet aan de minimale lichtcriteria voldoen. Lichtsimulaties in Radiance laten zien dat het onderzochte prototype beter presteert wat betreft lichtverdeling dan een vergelijkbaar, gesimuleerd echt raam onder een bewolkte of gedeeltelijk bewolkte hemelscènes. Met een heldere hemelscène is het verschil tussen een echt en een virtueel raam kleiner. Dit wordt veroorzaakt door de invloed van het directe zonlicht.

Samenvatting

xviii

• Een methode is voorgesteld om vier algemeen toegepaste verblindingindexen, die middels simulaties kunnen worden voorspeld, te correleren met de resultaten van het helderheid-oordeel in een elders uitgevoerd experiment met proefpersonen. Op die manier kan het visueel discomfort van een eerste generatie VNLS prototype met verschillende complexe mogelijkheden tot uitzicht worden beoordeeld. De resultaten suggereren dat de gesimu-leerde, genormaliseerde indexwaarden DGI, UGR en CGI, berekend met behulp van Evalglare allemaal zijn overschat in vergelijking tot waarden verkregen uit de experimentele data. De gesimuleerde waarden van DGP komen beter overeen met de geconverteerde DGP waardes uit het experiment.

• Een tweede generatie VNLS prototype is ontworpen en gebouwd. Dit prototype werd gebruikt voor ter validatie van het computermodel en kan worden uitgebreid voor verdere ontwikkeling van niet-bestaande VNLS. Een raster van LED-tegels werd geïnstalleerd om een diffuse lichtbron en een vereenvoudigd uitzicht. Lineaire LED-armaturen zijn geïnstalleerd om te voorzien in direct licht, dat zichtbaar wordt als enkele heldere vlakken op de zijwanden van de testruimte. Simulatieresultaten en meetwaarden van de horizontale verlichtingssterkte op vergelijkbare meetpunten werden geëvalueerd en toonden een goede overeenkomst. De ruimte-beschikbaarheid in een standaard kantoorruimte kan worden geoptimaliseerd door een prototype op elke korte wand tegenover elkaar of door twee prototypes op een lange wand naast elkaar te plaatsen. Alle onderzochte operationele scenario’s laten een relatief vergelijkbaar effect zien op de gemiddelde jaarlijkse ruimte-beschikbaarheid en het elektrische energieverbruik van de prototypes.

• Meer complexe VNLS configuraties, bestaande uit kleine licht-uitstralende bronnen, zijn ontwikkeld in een simulatiemodel. Het model heeft een vereenvoudigd uitzicht en voorziet in het gereflecteerde licht van de ‘grond’ op het plafond en van de ‘hemel’ op de vloer. Uit een gevoeligheidsanalyse bleek dat de totale lichtstroom van de ‘hemel’ voornamelijk de ruimte-beschikbaarheid van de testruimte beïnvloedt, terwijl de stralingshoek van de lichtbron voornamelijk van invloed is op de andere outputvariabelen, waaronder verblinding. Het merendeel van de gemodelleerde VNLS met een stralingshoek van 76° presteerden nagenoeg identiek aan echte ramen. De prestaties kunnen worden geoptima-liseerd door de stralingshoek te vergroten naar 114°, waardoor een ruimte-beschikbaarheid ontstaat die ongeveer twee keer groter is dan die van corresponderende echte ramen.

• Een model van VNLS configuraties met een complex uitzicht is gesimuleerd, waarbij het licht afkomstig was van een raster van gerichte, witgekleurde lichtbronnen. Het uitzicht werd gerealiseerd door een tweedimensionale afbeelding aan te brengen op een transparante glasplaat welke geplaatst is voor de lichtbronnen. Vergelijkingen tussen 10 afbeeldingsscènes zijn getoond door het gebruik van de transmissieve aanpak. Deze aanpak biedt weliswaar meer flexibiliteit om een complex uitzicht op het scherm te projecteren maar behoeft meer licht om aan de vereiste verlichtingssterkte te voldoen. Anderzijds kan het gebruik van een emissieve aanpak, zoals gedemonstreerd in het vorige punt, het licht efficiënter in de ruimte brengen, maar de complexiteit van het uitzicht wordt beperkt door het aantal pixels.

xix

Ikhtisar

Pemodelan dan Simulasi Solusi Pencahayaan Alami Virtual dalam Bangunan

Dalam situasi di mana pencahayaan alami tidak tersedia secara mencukupi, konsep Solusi Pencahayaan Alami Virtual (VNLS), yaitu sistem yang dapat menyediakan cahaya alami dan pemandangan ke luar secara realistis, adalah sangat menjanjikan. Manfaat dari pemasangan VNLS dalam bangunan yaitu adanya kemungkinan untuk menggunakan lebih banyak luas lantai yang sebelumnya hanya memiliki sedikit akses kepada cahaya alami. Hal ini menciptakan kemungkinan untuk mengendalikan kualitas dari cahaya dan pemandangan yang tersedia.

Penelitian ini bertujuan untuk memprediksi pengaruh dari berbagai aplikasi VNLS pada kinerja pencahayaan dan kenyamanan visual dalam bangunan. Dalam penelitian ini, diajukan metode perancangan menggunakan kreasi purwarupa secara cepat serta konsep perancangan berdasarkan pemodelan komputasional dan simulasi kinerja bangunan, untuk mendapatkan desain VNLS yang lebih baik. Proses pengembangan tersebut didasarkan pada purwarupa VNLS yang telah tersedia, kemudian disempurnakan lebih lanjut menggunakan perangkat simulasi yang relevan.

Dalam penelitian ini dilakukan langkah-langkah sebagai berikut: (1) tinjauan literatur yang terkait dengan perkembangan purwarupa VNLS terkini; (2) pengukuran purwarupa VNLS, berupa studi kasus dari suatu purwarupa bercahaya difus (disebut juga purwarupa ‘generasi pertama’) serta purwarupa bercahaya terarah (disebut juga purwarupa ‘generasi kedua’); (3) simulasi purwarupa dan model VNLS, menggunakan perangkat simulasi komputasi untuk memprediksi kinerja pencahayaan dalam bangunan; yang mana pengukuran dari purwarupa ‘generasi kedua’ digunakan untuk memvalidasi model VNLS; (4) pemodelan komputasi dari VNLS, melibatkan rangkaian sumber cahaya berukuran kecil dengan berbagai sudut kemiringan, guna menghantarkan cahaya ke berbagai arah; (5) analisis sensitivitas, untuk mengetahui pengaruh dari parameter-parameter masukan terhadap indikator kinerja dari model VNLS yang terkait; (6) perbandingan kinerja dengan jendela sejati untuk mengetahui seberapa besar manfaat dari pemasangan VNLS dalam suatu ruangan; dan (7) perhitungan teoretis, untuk memperkirakan konsumsi energi listrik tahunan dari suatu purwarupa VNLS dengan berbagai skenario pengoperasian.

Hasil yang terpenting dari penelitian ini dapat dirangkum sebagai berikut:

• Berdasarkan pengukuran dari suatu purwarupa ‘generasi pertama’ yang menampilkan pemandangan langit yang disederhanakan, didapatkan bahwa kriteria pencahayaan minimum yang dikehendaki tidaklah terpenuhi. Berdasarkan simulasi pencahayaan dengan Radiance, purwarupa tersebut menghasilkan kemerataan yang lebih baik daripada jendela sejati, di bawah kondisi langit mendung dan berawan sebagian. Di bawah kondisi langit

Ikhtisar

xx

cerah, perbedaan antara jendela sejati dan virtual menjadi lebih rendah karena pengaruh dari cahaya matahari langsung.

• Sebuah metode diajukan untuk menghubungkan empat metrik kesilauan yang umum diterapkan, yang dapat diprediksi menggunakan simulasi, dengan nilai kesilauan subjektif dari suatu purwarupa VNLS generasi pertama yang didapatkan dalam suatu eksperimen menggunakan subjek manusia. Hasil simulasi dengan Evalglare menunjukkan bahwa nilai-nilai DGI, UGR, dan CGI yang dinormalisasi adalah lebih tinggi daripada nilai-nilai yang dikonversi dari data eksperimen, sedangkan nilai-nilai DGP berdasarkan simulasi lebih sesuai dengan nilai-nilai DGP yang dikonversi.

• Sebuah purwarupa VNLS generasi kedua telah dirancang dan dibangun. Purwarupa ini dievaluasi untuk memvalidasi model komputasi yang dapat digunakan lebih lanjut untuk memodelkan VNLS yang belum ada. Suatu rangkaian ubin LED secara khusus digunakan untuk menghasilkan cahaya difus dan pemandangan, sedangkan rangkaian armatur LED linear digunakan untuk memberikan cahaya terarah yang menghasilkan berkas-berkas yang dapat terlihat pada dinding samping dari ruang uji. Simulasi dan pengukuran iluminansi horisontal pada lokasi yang bersesuaian menunjukkan hasil yang serupa satu sama lain. Ketersediaan ruang dalam sebuah ruang kantor standar dapat dioptimalkan dengan cara menempatkan satu purwarupa pada setiap dinding pendek, atau dengan menempatkan dua purwarupa pada sebuah dinding panjang. Seluruh skenario pengope-rasian menghasilkan pengaruh yang relatif sama pada rata-rata ketersediaan ruang dan konsumsi energi listrik tahunan dari purwarupa.

• Konfigurasi VNLS yang lebih kompleks menggunakan sumber-sumber cahaya berukuran kecil telah dirancang dalam model komputasi. Model tersebut menampilkan pemandangan yang disederhanakan, menghantarkan cahaya dari ‘tanah’ ke langit-langit dan dari ‘langit’ ke lantai. Analisis sensitivitas menunjukkan bahwa fluks cahaya total dari ‘langit’ sangat mempengaruhi ketersediaan ruang, sedangkan sudut pancaran dari sumber sangat mempengaruhi variabel-variabel keluaran yang lain, termasuk tingkat kesilauan. Sebagian besar model VNLS dengan sudut pancaran 76° menghasilkan kinerja yang paling serupa dengan jendela sejati. Dengan meningkatkan sudut pancaran menjadi 114°, ketersediaan ruang dapat dioptimalkan menjadi sekitar dua kali lebih besar dibandingkan dengan ketersediaan ruang yang dihasilkan oleh jendela sejati.

• Sebuah model dari konfigurasi VNLS dengan pemandangan kompleks telah dirancang, di mana cahaya dihasilkan dari susunan sumber-sumber cahaya terarah berwarna putih, sedangkan pemandangan dihasilkan dengan cara menempelkan gambar dua dimensi pada permukaan kaca transparan di depan sumber-sumber cahaya. Perbandingan antara 10 jenis gambar ditunjukkan dengan menggunakan metode transmisif yang menawarkan fleksibilitas dalam menampilkan pemandangan yang kompleks, namun juga membutuhkan lebih banyak cahaya untuk memenuhi kriteria pencahayaan dalam ruang. Pada sisi lain, penggunaan metode emisif pada poin sebelumnya dapat menghasilkan lebih banyak cahaya, namun kompleksitas dari pemandangan yang ditampilkan menjadi terbatas oleh banyaknya piksel yang digunakan.

xxi

Nomenclature

Roman symbols

%A space availability [%]

%A R space availability under real windows scene [-]

%A V space availability under VNLS scene [-]

%G ground contribution on the ceiling [%]

%Gav average ground contribution on the ceiling [%]

Ai projected light source surface [m2]

BA beam angle of the ‘sky’ element [°]

C1 Einhorn’s weighting coefficient, i.e. 8 [-]

C2 Einhorn’s weighting coefficient, i.e. 2 [-]

CGI CIE glare index [-]

CGIn normalised CIE glare index [-]

CGIn conv converted normalised CIE glare index [-]

CGIn sim simulated normalised CIE glare index [-]

d distance between individual windows [m]

DF daylight factor [%] DGI daylight glare index [-]

DGIn normalised daylight glare index [-]

DGIn conv converted normalised daylight glare index [-]

DGIn sim simulated normalised daylight glare index [-]

DGP daylight glare probability [-]

DGP conv converted daylight glare probability [-]

DGP sim simulated daylight glare probability [-]

Eav average illuminance [lx]

Ecrit criterion illuminance [lx]

Ed direct vertical illuminance [lx]

Eground illuminance contribution from the ‘ground’ element on the ceiling [lx]

Ei diffuse vertical illuminance [lx]

Emea measured illuminance [lx]

Emin minimum illuminance [lx]

Esim simulated illuminance [lx]

Etotal total illuminance [lx]

Ev vertical illuminance on the observer’s eye [lx]

ERC externally reflected component [%] IA interval of tilt angle of the ‘sky’ element [°]

Nomenclature

xxii

IR,G,B total spectral irradiance [W/m2]

IR red spectral irradiance values [W/m2]

IG green spectral irradiance values [W/m2]

IB blue spectral irradiance values [W/m2]

IRC internally reflected component [%] Lav average surface luminance [cd/m2]

Lb background luminance [cd/m2]

Le emitted radiance [W/(sr·m2)]

Li incoming radiance [W/(sr·m2)]

Lmax maximum surface luminance [cd/m2] Lmin minimum surface luminance [cd/m2]

Lo outgoing radiance [W/(sr·m2)]

Lr reflected radiance [W/(sr·m2)]

LR,G,B total spectral radiance [W/(sr·m2)]

LR red spectral radiance values [W/(sr·m2)]

LG green spectral radiance values [W/(sr·m2)]

LB blue spectral radiance values [W/(sr·m2)] Ls surface or glare source luminance [cd/m2]

N total number of points [-]

n(E ≥ 500 lx) number of points with illuminance ≥ 500 lx [%]

n(E ≥ Ecrit) number of points with illuminance exceeding the criterion illuminance [%]

P position index [-]

PDGav average probability of discomfort glare [-]

PDGav R average probability of discomfort glare under real windows scene [-]

PDGav V average probability of discomfort glare under VNLS scene [-]

SC sky component [%] U0 uniformity [-]

U0 R uniformity under real windows scene [-]

U0 V uniformity under VNLS scene [-]

UGR unified glare rating [-]

UGRn normalised unified glare rating [-]

UGRn conv converted normalised unified glare rating [-]

UGRn sim simulated normalised unified glare rating [-]

Wreal real-time electrical power consumption [W]

Nomenclature

xxiii

Greek symbols β regression coefficient [-]

β’ standard regression coefficient [-]

θi incoming angle [rad]

θo outgoing angle [rad] ρ weighted average spectral reflectance [-]

ρR spectral reflectance in red [-]

ρG spectral reflectance in green [-]

ρB spectral reflectance in blue [-] τ weighted average spectral reflectance [-]

τR spectral transmittance in red [-]

τG spectral transmittance in green [-]

τB spectral transmittance in blue [-]

Φ total luminous flux of the ‘sky’ element [lm]

Φi total radiative flux of light source [W]

Φv total luminous flux of light source [lm]

ψi incoming angle [rad]

ψo outgoing angle [rad]

Ωi solid angle of the incoming radiance [sr]

ωpos modified solid angle of the glare source [sr]

ωs solid angle of the glare source [sr]

Abbreviations

CCT correlated colour temperature

CEN Comité Européen de Normalisation

CIE Commission Internationale de l'Eclairage

CRI colour rendering index

CQS colour quality scale

DALI digital addressable lighting interface

DIR directionality

DML distant mixed land

DMM distant man-made

DMR distant mixed river

DMX digital multiplex

DNL distant near land

DNR distant near river

Nomenclature

xxiv

DoE Department of Energy

DPC depth perception cues

EN European standard

GSD green, sky, and distant objects

HD high-definition

HDR high dynamic range

INF information

ISO international standard (for measuring film speed)

LED light emitting diode

MRI magnetic resonance imaging

NML near mixed land

NMM near man-made

NMR near mixed river

NNL near natural land

NNR near natural river

ORG organisation

PAR parabolic aluminium reflector

RGB red, green, blue

RMS root mean square error

RW real windows

SPD spectral power distribution

TL tubular fluorescent lamp

VNLS virtual natural lighting solutions

1

Chapter 1 Introduction

This chapter discusses the use of natural light in buildings, its benefits and shortcomings, particularly the limitation in time and space. This chapter introduces the concept of Virtual Natural Lighting Solutions (VNLS), and the use of computational modelling and simulation to steer the development of the solutions. The aim and objectives of the research, as well as the thesis outline, are further described.

1.1. Natural Lighting Demand

Human beings have a strong preference for natural light. Many researchers have shown that natural light is highly preferred over electrical lighting in the built environment for its positive effects on user satisfaction and health (e.g. Farley & Veitch, 2001; Boyce, 2003; Chang & Chen, 2005; Galasiu & Veitch, 2006; Aries et al., 2010). Natural light provides various stimulations throughout the day, and it is believed that access to natural light can reduce stress and increase productivity (e.g. Boyce et al., 2003; Heschong, 2003a; Heschong, 2003b). A relationship between stress, depression, and little exposure to natural light has been discovered, on which thorough overviews are provided by, e.g. Edwards & Torcellini (2002); Boyce et al. (2003); Boubekri (2008); and Beute & de Kort (2014). A thorough literature review of studies on the effects of daylight exposure on human health since 1989 until 2013 is presented by Aries et al. (2013), in which a statistically significant and well-documented evidence for the relationship between daylight and its potential effect on health was found to be limited. Nonetheless, some first practical implementations for building design can already be shown (Aries et al., 2013).

In buildings, the admission of natural light also provides a view with information about the outside situation, such as time of the day and weather condition. In general, it is found that weather is influential on people’s health and mood (e.g. Eagles, 1994; Keller et al., 2005; Denissen et al., 2008). Several studies have reported on beneficial and restorative effects of views onto a natural scene (e.g. Ulrich, 1984; Ulrich et al., 1991; Tennessen & Cimprich, 1995) whereas views onto human-built environments yield effects which are similar to having no window at all (Kaplan, 1993). Kim & Wineman (2005) showed in an internal report that empirically, views and windows have psychological and economic values. In the first part of their study, they showed that availability of view from a building was positively related to assigned property values. In the second part, they recorded seating selection occupancy rates in a cafeteria and a library, and found that people were more likely to choose a seat near windows and views.

In general, a window is an opening in the wall that allows the admittance or flows of air, light, and sound, which mostly influence the indoor environment (Tregenza & Loe, 1998). In its very basic function, a window provides the possibility for having light and air from

Chapter 1

2

outside into the inside space. Nevertheless, Collins (1975) found that windows provided many more functions for people than just sources of light and air. In her study with 88 window-related cases conducted in a variety of settings, she found that windows provided a view to the outside, knowledge of the weather and time of day, feeling of connection to the outside environment, relief from feelings of claustrophobia, monotony or boredom.

Many researchers show that the view is an important aspect provided by a real window, and even cannot be separated from the natural light itself (e.g. Tuaycharoen & Tregenza, 2007). The findings of Markus (1967) and Keighley (1973a, 1973b) showed that views should have three specific layers: a layer of sky, a layer of city or landscape, and a layer of ground. Each layer has its own specific function.

Depending on the position of the observer inside the building, as well as the location of the building itself, Keighley (1973b) pointed out that for a typical office building in an urban area, the view can be classified into three types. Figure 1.1 illustrates these three types: the first one represents a cityscape scene with a natural horizon, such as normally seen from the uppermost floors of a tall building; the second represents a panorama of mid-ground buildings, giving an elevated skyline seen from approximately ground floor position; and the third is entirely occupied by the façade of a nearby building. The results showed that the type of view, together with other factors, can influence the satisfaction of the observer.

(a) (b)

(c)

Figure 1.1. Three types of view investigated by, and taken from Keighley (1973b). View (a) is a cityscape scene with a natural horizon, (b) is a panorama of mid-ground buildings, and (c) is entirely obstructed by the façade of a nearby building.

Introduction

3

1.2. Shortcomings of Natural Lighting

Despite all of its advantages, the quality and quantity of natural light is highly variable. Its availability is limited by space and time. For instance, there is not enough or no daylight at all during nighttimes, buildings can be too deep to supply sufficient daylight throughout the space, and some rooms are simply not provided with windows, skylights, or any form of daylight transporting systems. The latter situation can be found, among others, in operating rooms in hospitals and in control rooms in industrial plants; due to hygienic or safety considerations. Or, as suggested in Figure 1.1c, people may be working nearby a window whose view is obstructed by the façade of a neighbouring/adjacent building, and find there is not much natural light penetrating into their work stations, for instance because the sky is obstructed in such a way that there is no functional daylighting possible.

In general, useful natural light in a sidelit space for typical office and educational activities can be roughly estimated by the so-called ‘window-head-height’ rule of thumb (Reinhart & Weismann, 2012). This rule relates how far ‘adequate, useful and balanced daylight enters the spaces for most of the year’, as a function of the distance from any point on the floor to the top of the window (Reinhart, 2005). A simulation-based validation study of this rule of thumb for unobstructed facades yielded that the depth of the daylit area usually lies between 1 and 2 times the size of the window-head-height, in a typical sidelit office space with Venetian blinds. For spaces that are not equipped with movable shading devices, such as atria or circulation areas, the ratio can increase up to 2.5 times the size of the window-head-height.

Admission of natural light into work places is recommended in most national building standards, even though the legislation differs from country to country (Boubekri, 2004; Boubekri, 2008). In commercial buildings such as offices, the layout design is however very much dependent on the national context, as thoroughly discussed by van Meel (2000). For example, according to Saxon (1994), employees in the United States tend to sit within 14 ~ 16 m from a window, British employees are used to sitting within 8 ~ 10 m, and German employees within 4 ~ 6 m. According to Duffy et al. (1993), British employees tend to work in open-plan offices while their counterparts in North European countries are used to work in cellular offices. A clear example of this tendency is illustrated in Figure 1.2, adapted from van Meel (2000), which shows the work place layout of the main headquarters of a commercial bank in Amsterdam and London.

As shown, most employees in Amsterdam (representing most of the North European countries) sit next to a window, in a building with corridors and long, narrow floor plan. Such layout can ensure sufficient access to natural light for everyone in the building; however, the design will also require more land area. In London (representing the United Kingdom and the United States), most employees are working in relatively small cubicles in large open areas. This concept can efficiently reduce the use of total land area, but on the other hand, it will provide very little access of natural light to the employees who work in the cubicles.

Chapter 1

4

(a) (b)

Figure 1.2. Work place layout of the main headquarters of a same commercial bank in (a) Amsterdam and (b) London, taken and adapted from van Meel (2000)

Another important fact is that significant fractions of the working population in the world do their work during nighttime. In the European Union, approximately 15% of women and 29% of men of the working population (< 45 years old) do night shift work (Härmä & Ilmarinen, 1999). For the elder working population (≥ 45 years old), the figures were 12% for women and 24% for men (Härmä & Ilmarinen, 1999). Night shift workers experience various discomfort issues, such as sleep problems, fatigue, and poor performance, and even increased long-term risk of some types of cancer due to a lack of synchronisation between the shift work schedule and the worker’s light-dark cycle (Stevens, 2009; Blask, 2009).

Moreover, many studies have reported that the value of increased productivity due to an improved indoor climate can be much greater than the costs of energy it consumes (e.g. Woods, 1989; Skåret, 1992; Kosonen & Tan, 2004; CABA, 2008; EC, 2013). The quality of the working environment in offices, schools, and factories could therefore outweigh any savings in energy and should become one of the major drivers in research on buildings (EC, 2013). All of these considerations lead to a demand for having an artificial solution that can bring natural light with all of its qualities to the inside space.

1.3. Proposed Solution

In situations where daylight is not or insufficiently available, Virtual Natural Lighting Solutions (VNLS) concept can be promising. VNLS are systems that can artificially provide

Introduction

5

natural lighting as well as a realistic outside view, with properties comparable to those of real windows and skylights. The benefit of installing VNLS in a building is the possibility to use more floor area that currently has very limited or no access to daylight, with the additional possibility to control the light and view quality.

The real, ideal product that gives both light and view in a very high quality does not yet exist at the moment. Nonetheless, it is known that in its intense appearance without a sufficient view, bright light can have a positive effect on human well-being (e.g. Glass et al., 1985; Badia et al., 1991; Avery et al., 1992; Eastman et al., 1998; Lingjærde et al., 1998; Avery et al., 2001; Mottram et al., 2011; Smolders et al., 2012; Smolders, 2013; Smolders et al. 2013; Beute & de Kort, 2014). The inverse is also true, that artificial views which emit no light themselves can also have a positive effect on humans (e.g. Heerwagen & Orians, 1986; Heerwagen, 1990; Ulrich et al., 1993). The concept of VNLS is to combine both light and view together, to provide even more positive effects on the users.

Investigations of the psychological effects of existing VNLS prototypes are still an ongoing process. For example, de Vries et al. (2009) conducted experiments to study the work performance of test subjects in a standard office room with two units of ‘emulated windows’, obstructed with a diffuse screen. Prototypes of the same type were used in the experiments of Smolders et al. (2012) to investigate the effects of brightness, focusing on the effect of eye illuminance on subjective measures, task performance, and heart rate variability. Experiments on glare sensation from another prototype with a simplified view were conducted by Rodriguez & Pattini (2014), observing its effects on glare-sensitive and glare-insensitive subjects when performing a computer task. A number of short- (one day) and long-term (3 and 10 years) acceptance studies of a VNLS prototype with a simplified sky scene and sunlight were conducted by Enrech Xena (1999), as also shown in (Fontoynont, 2011a, 2011b), which results showed a high acceptability in windowless space for long-term use, under some certain settings.

Some user perception studies on view and light (quality) aspects of VNLS prototypes have been reported. For example, Tuaycharoen & Tregenza (2005) studied subjective discomfort glare from screen projected images, and concluded that a good view (also described as a view with high interest), which mainly consists of the natural scenes, tends to reduce discomfort glare perception. IJsselsteijn et al. (2008) focused on depth perception cues from screen projected images, and concluded that motion parallax, occlusion, and blur had a significant effect on the viewer’s see-through experience, with motion parallax yielding the greatest effect size. Shin et al. (2012) investigated subjective discomfort glare from a backlit, transparent printed image, and concluded that the tolerance of discomfort glare sensation for the distant views including skyline was greater than the near views. In all of those experiments, the prototypes/displays were assumed to be a representation of what the subjects normally see through a real window.

Nevertheless, the existing prototypes are considered not suitable for meeting the whole expectation, since they are only able to meet part of the natural light and view expectation (Mangkuto et al., 2011). For instance, the light produced is much less than that coming from

Chapter 1

6

a real window, there is no directional light component, the view displayed is limited; and so on. These limitations indicate that the ideal VNLS concept should go beyond its predecessors.

In this thesis, the light and view qualities are considered two key aspects in assessing existing and future solutions. In particular, the presence of a directional light component and the complexity of the view are taken as the general descriptors for the qualities. Therefore, VNLS prototypes (that exist) and models (that do not yet exist) can be generally classified into four types, as illustrated in Figure 1.3, which are:

1. Type that provides a simplified view and (mainly) diffuse light 2. Type that provides a complex view and (mainly) diffuse light 3. Type that provides a simplified view and (mainly) directional light 4. Type that provides a complex view and (mainly) directional light

Note that the first two types are existing prototypes, which can alternatively be considered as the ‘first generation’ prototypes, which are available for research and/or commercial purpose. The last two types are future generations, which are proposed and built in this thesis using computational modelling and simulation, and for which the physical models are not yet available.

Figure 1.3. Classification of VNLS based on light and view qualities

Introduction

7

While the relationship between the first generation VNLS prototype and user perception has been investigated elsewhere (e.g. Tuaycharoen & Tregenza, 2005; IJsselsteijn et al., 2008; Shin et al., 2012), there is very little discussion about the impact of VNLS application on building performance. This thesis focuses on the latter, in which the influence of such solutions on the lighting performance and visual comfort inside the relevant spaces is evaluated. Since the ideal VNLS are future, not-yet-existing systems with lots of possible input variables, computational modelling and building performance simulation are applied to predict the performance of the solutions and to accelerate the development process.

In his thesis, Crawley (2008) described computational modelling as an approach that allows evaluation of alternative designs or technologies without having to create the artifacts. It is generally cheaper to create a model and to test alternative designs configurations than to build a real prototype and revise it later based on trial and error. Furthermore, in view of application in buildings, Crawley described building performance simulation as a powerful tool that emulates the dynamic interaction of natural, physical phenomena such as heat, light, and sound within the building to predict its energy and various environmental performances, available today for use by policy setters and decision makers. In this thesis, design methods including rapid prototyping and providing multiple design concepts are proposed as means to obtain better design solutions.

Within the building design context, a number of design stages can be distinguished, as suggested e.g. by Stoelinga (2005). These stages mainly consist of decision, programme of requirements, preliminary design, final or detailed design, and the contract document. The objectives and requirements are defined in the programme of requirements or project brief. The main systems are selected and a number of concepts are developed in the preliminary or conceptual design. Next, the development and integration of design elements to operate design solutions takes place in the final design stage, which is finally closed with the contract document in which the production drawings, specification, and construction resource documentation are finalised. In particular, there is a demand to use building performance simulation for design support of the generation and selection of alternative design concepts during early phases in the design process, where decisions often have to be made with limited resources and based on limited knowledge (Stoelinga, 2005; Hopfe, 2009).

In the context of this thesis, computational modelling and building performance simulation are applied to predict the performance of both existing and non-existing solution, in terms of lighting performance and visual comfort in buildings. The validated lighting simulation and rendering software Radiance (Ward, 1994; Ward & Shakespeare, 1998) is employed as the main tool to (re-)create the model of existing prototypes and non-existing solutions, as well as to predict the relevant performance indicators.

Chapter 1

8

1.4. Aim and Objectives

The aim of the research in this thesis is predicting the impact of various VNLS applications on lighting performance and visual comfort in buildings, by means of computational modelling and building performance simulation.

The main objectives of the research are investigating and enabling innovative application of modelling approaches for VNLS, which involve: • Determining the relevant properties and performance indicators for VNLS. • Finding the appropriate modelling approach for VNLS. • Evaluating the lighting performance and visual comfort of various VNLS model. • Finding the potential of applying VNLS under various configurations by predicting their

performance, and under various operating scenarios by estimating total annual electrical energy consumption.

1.5. Research Methodology

In this thesis, design methods including rapid prototyping and providing multiple design concepts using computational modelling and building performance simulation are proposed for obtaining a better design of VNLS. The research deals with two main parts, i.e. existing VNLS prototypes and non-existing VNLS models. The existing VNLS prototypes are classified into the ‘first’ and ‘second’ generations, which correspond to those with low and high directionality. Measurement and simulation of the first generation are performed under various display settings to validate the lighting performance calculation, whereas simulation of real window is performed to understand the difference between the prototype and its corresponding real window. Measurement and simulation of the second generation are also performed under various display settings to validate the lighting performance calculation, and to analyse various configurations of the prototypes. In addition, theoretical calculation is conducted based on the measurement data, to estimate the annual lighting performance and electrical energy consumption under various operating scenarios.

In turn, results from the existing VNLS prototypes are used as inputs to develop non-existing VNLS models, which are classified based on view complexity. Simulation is performed for both models with simplified and complex views, taking various input variables into account to investigate the influence of each variable on lighting performance, whereas simulation of real windows is also performed to analyse the difference between models and their corresponding real windows. For the model with a complex view, additional simulation is conducted using two modelling approaches, i.e. ‘transmissive’ and ‘emissive’ approaches, to understand the impact of employing both approaches on the lighting performance of the VNLS models. Based on these results, conclusions and recommendations for further research are drawn.

The research framework diagram is illustrated in Figure 1.4.

Introduction

9

Figure 1.4. Research framework diagram

Throughout the process, the following steps are carried out:

• Literature review of current development of VNLS prototypes (existing virtual windows and skylights).

• Measurement of VNLS prototypes, which are presented in case studies of a prototype with diffuse light (the first generation prototype) and a prototype with directional light (the so-called second generation prototype).

• Simulation of VNLS prototypes, which are performed using computational building performance simulation tools to predict the lighting performance. The measurement and simulation results of the prototypes are incorporated as calibration of the future VNLS model.

• Computational modelling of future VNLS, which involves arrays of small light sources with various tilt angles to deliver the light in various directions into the space.

• Sensitivity analysis, to understand the influence of the relevant input parameters on the relevant performance indicators of the VNLS model.

• Performance comparison with real windows scenes, to understand the benefit of installing VNLS in a given space, relative to the similar scene with real windows, by comparing the performance indicators of interest.

Chapter 1

10

• Theoretical calculation, to estimate the annual space availability and electrical energy consumption of a VNLS prototype under various operating scenarios.

1.6. Thesis Outline

The thesis focuses on the impact of various types of VNLS on lighting performance and visual comfort in buildings. In addition, to estimate the total electrical energy consumption on an annual basis, a theoretical study based on measured power consumption data of an existing VNLS prototype is applied. Figures 1.3 and 1.4 can be referred back to describe the framework of this thesis. Chapter 2 discusses the left-hand side of the chart in Figure 1.3; by giving a literature review of existing first generation prototypes. The general concept of VNLS is discussed, explaining the definition and ambitions of having VNLS installed in a building.

Chapter 3 presents an example of using Radiance to reproduce the scenes and to evaluate the lighting performance of a first generation prototype with a simplified view and diffuse light (lower-left quadrant in Figure 1.3). The performance of the measured prototype was compared to a hypothetical real window under the same settings in simulation.

Chapter 4 provides an evaluation of discomfort glare from a first generation VNLS prototype with complex views and diffuse light (upper-left quadrant in Figure 1.3), correlated to the results of an experiment conducted elsewhere (Shin et al., 2012) on subjective glare perception from the same prototype. Radiance and Evalglare were applied to recreate the scenes and evaluate the glare metrics. The results from both the subjective evaluation and simulation were compared to each other.

Chapter 5 discusses the lighting performance of a second generation VNLS prototype (between the lower-left and lower-right quadrant in Figure 1.3), in which more directional light is installed. The lighting performance obtained from measurements in a reference space was compared to simulation using Radiance, as a basis to calibrate the model that can be extended for future VNLS. Various possibilities of placing the prototype inside the room were investigated in Radiance to determine the effect on space availability and visual comfort. Various operating scenarios were introduced and calculated to determine the effect on the average space availability and total annual electrical energy consumption that are produced and consumed by the prototype.

Chapter 6 introduces a new VNLS model providing a simplified view and directional light (lower-right quadrant in Figure 1.3), and discusses the calculated lighting performance of the model. The model was created and simulated using Radiance to understand the effect of varying input variables of the VNLS model on lighting performance of a reference space, and to compare the lighting performance of the simulated VNLS, relative to that of real windows under the standard CIE overcast sky.

Chapter 7 discusses the lighting performance of a VNLS model with complex views and directional light (upper-right quadrant in Figure 1.3). The model was created and simulated

Introduction

11

using Radiance to understand the effect of varying input variables of the more comprehensive VNLS model on the lighting performance of a reference space. Comparisons of various image views, as well as two modelling approaches, i.e. the transmissive and emissive approaches, are shown.

Chapter 8 contains the main conclusions of the thesis and recommendations for further research.

Chapter 1

12

13

Chapter 2 General Concept of Virtual Natural Lighting Solutions

This chapter discusses the general concept of Virtual Natural Lighting Solutions (VNLS), including the definition, expectation, classification, as well as state of the art of development. The VNLS prototypes and models can be generally classified in terms of light directionality and view complexity. Comparisons of light and view qualities of the existing prototypes are shown, based on their observed features.

2.1. Definition of VNLS

In most buildings, a real natural lighting solution can be thought of as any opening in the façade or ceiling of a building, which can bring natural light to the space inside. Two common examples of elementary natural light openings are the (vertical) window and the (horizontal) skylight. A window is a transparent opening in the building façade, door or wall, which allows the passage of light and, if not closed or sealed, air and sound. Windows are usually glazed or fitted with some other translucent or transparent material like glass. A skylight is an opening located in the roof, covered with translucent or transparent material like glass or plastic, which is designed to admit daylight. An example of a complex natural light opening is a light reflecting tubular device (light pipe).

In the cases where a real natural lighting solution is absent or ineffective, for instance due to space and time limitation, the concept of Virtual Natural Lighting Solutions (VNLS) can be promising to overcome the problem of lack of daylight. VNLS are defined here as ‘systems that can artificially provide natural lighting as well as a realistic outside view, with properties comparable to those of real windows and skylights’.

2.2. Classification of VNLS

A number of efforts have been made to imitate one or more elements of natural light inside buildings, in the form of artificial solutions. Originally, the efforts were more focused on bringing the ‘view’ of an outside condition into the room. Attempts to create a realistic artificial view have been under development for centuries. For example, in art history, trompe l'oeil is known as an art technique involving realistic imagery to create the optical illusion that the depicted objects appear in three dimensions, while actually being a two-dimensional painting. This technique can be traced back to the ancient Greek era around the year 400 BC, and was developed further mostly by Italian artists between the 15th and 17th century. Despite being very inspiring, this example is not discussed further in detail, since it is not an actual light source, nor a device that can transmit light from the outside environment. Nevertheless,

Chapter 2

14

the concept of displaying artificial sceneries of nature is still used in the later form of VNLS prototypes. Some researchers have shown that artificial views, which do not emit light themselves, can actually have a positive effect on human health (e.g. Heerwagen & Orians, 1986; Heerwagen, 1990; Ulrich et al., 1993).

Interestingly, the inverse is also true. In its intense appearance without a sufficient view, artificial bright light can create a positive effect on human well-being and healing (e.g. Glass et al., 1985; Badia et al., 1991; Avery et al., 1992; Eastman et al., 1998; Lingjærde et al., 1998; Avery et al., 2001). Specific lighting products have been manufactured to generate large amounts of light with a particular spectral power distribution for this application. In general, the idea behind this type of VNLS prototype is to recreate the situation with natural light and its qualities inside a space, and to harvest the benefit it may offer.

Directionality of the light is another important property that typically distinguishes a real window or skylight from an artificial version. In fact, directional light is something that rarely appears in existing VNLS prototypes; most of them only generate light in a nearly diffuse direction. Therefore, a non-diffuse, or directional, light is considered a key feature that should appear in an ideal VNLS prototype.

Based on these considerations, any VNLS prototypes (that exist) and models (that do not yet exist) can be classified based on their light and view qualities, as previously illustrated in Figure 1.3.

2.2.1. Prototypes with a simplified view One of the simplest versions of VNLS prototype is the ‘light box’, which is generally

constructed of a series of artificial light sources behind a translucent diffuse surface. In view of health application, it is known that human bodies use natural (sun-) light to regulate a variety of functions that affect mood and energy level, cure skin disorders, and make vitamin D (e.g. Cajochen, 2007; Vandewalle, 2009). Without enough (sun-) light, humans often feel down, lack energy, and sometimes even suffer physical disorders. To help reduce these symptoms, specific light boxes have been designed to provide illuminances up to 10000 lx at a distance of approximately 50 cm, where the individual sits for a specified duration. It has been shown that the so-called bright light therapy can have a positive effect on human well-being and healing (e.g. Glass et al., 1985; Badia et al., 1991; Avery et al., 1992; Eastman et al., 1998; Lingjærde et al., 1998; Avery et al., 2001; Mottram et al., 2011; van Hoof et al., 2012). A similar solution uses sets of blue light emitting diodes (LEDs) in a light box, designed with an enhanced blue spectrum component, based on independent clinical research showing that blue light from the summer sky can regulate mood and can trigger human bodies to become active and energetic (e.g. Webb, 2006; Glickman et al., 2006; Viola et al., 2008; Iskra-Golec et al., 2012). Another new application to create the effect of natural light uses gradually increasing levels of brightness, to wake up people in the morning in a natural way.

General Concept of VNLS

15

A number of studies have been performed using such prototype as a method to study various effects of light and view on subjects. For instance, in their experiments, de Vries et al. (2009) installed two units of ‘emulated windows’, each measuring 1.20 m × 1.20 m with 12 rows of tubular fluorescent lamps, covered with a diffuse screen (see Figure 2.1). The experiments were conducted to evaluate the work performance of the subjects, which results showed that the performance of the test subjects increased when the view was removed and when daylight was replaced by an artificial light source. It should be noticed however that the study was only a pilot with a relatively small number of samples (N = 10).

Figure 2.1. Interior view of the room with obstructed windows in the experiments of de Vries et al. (2009)

Prototypes of the same type were used in the experiments of Smolders et al. (2012), focusing on the effect of eye illuminance on subjective measures, task performance, and heart rate variability. The results showed that a higher eye illuminance could improve not only subjective feelings of alertness and vitality, but also objectively measured performance. The performance measures suggested that white light could improve performance and yield faster responses and higher accuracy on simple cognitive tasks. In addition, the exposure to a higher illuminance can also increase physiological arousal.

Experiments on glare sensation from another prototype with a simplified view were conducted by Rodriguez & Pattini (2014), observing its effects on glare-sensitive and glare-insensitive subjects when performing a computer task. The results showed that luminance and size of the window had the same, statistically significant effect on glare sensation for both groups. However, when occasionally looking directly at the glare source, glare-sensitive people had a higher relative risk of being disturbed. In all of those mentioned studies, the prototype was installed to provide the intended light qualities such as vertical illuminance and view luminance.

A prototype that provided not only light with a simplified sky scene but also sunlight has been developed by Philips (van Loenen et al., 2007). The prototype was a 1.20 m × 1.20 m luminaire with 12 rows of tubular colour fluorescent lamps. Each lamp could be tuned to mimic the colour gradients of, for example, the sunrise, noon, or sunset. A high intensity discharge (HID) spot light was added and could be controlled to mimic direct sunlight. The view variation of this prototype was higher compared to those mentioned in the previous

Chapter 2

16

paragraph, since there was a possibility to control the colour gradient and to create the impression of having a spot of sunlight inside the space.

Another prototype with a simplified sky scene and sunlight has been also developed by ENTPE-EDF Lyon (Enrech Xena, 1999; Fontoynont, 2011a, 2011b). Short- (one day) and long-term (3 and 10 years) acceptance studies were performed under various colour modes (Figure 2.2). The results showed a high acceptability in windowless space for long-term use, individual control was indispensable, and non-natural light spectra were sometimes preferred in the end of the day.

Figure 2.2. Example of prototype with a simplified view and diffuse light, and a possibility of adding sunlight, taken from Fontoynont (2011a, 2011b)

To briefly summarise, the type of solutions with a simplified view and mainly diffuse light can be classified as shown in Table 2.1 as follows.

Table 2.1. Classification of solutions with a simplified view and mainly diffuse light

Source Features References

Fluorescent/LED Large brightness, static view

de Vries et al. (2009); Smolders et al. (2012); Rodriguez & Pattini (2014)

Varying brightness, static view Enrech Xena (1999); van Loenen et al. (2007); Fontoynont (2011a, 2011b)

2.2.2. Prototypes with a complex view

While light from a window is beneficial for building occupants, view is another important feature of a window (e.g. Collins, 1975; Collins, 1976; Kaplan & Kaplan, 1989; Kaplan, 1993; Farley & Veitch, 2001; Boyce, 2003; Galasiu & Veitch, 2006; Aries et al. 2010). A number of commercial efforts have been developed to provide a view from a VNLS prototype, using static, semi-transparent photographs in front of a group of light sources, mostly fluorescent or LED lamps, an example of which is shown in Figure 2.3. Application of this prototype can be found in windowless healthcare environments such as critical care units and magnetic resonance imaging (MRI), particularly to reduce the anxiety of the patient.


17

Figure 2.3. Example of prototype with a complex view and diffuse light using backlit, transparent photos showing static image: round skylight for healthcare environment, taken from TESS (2012)

Research regarding subjective discomfort glare from such prototypes has been performed, for example by Shin et al. (2012) and Kim et al. (2012), using backlit, transparent printed photographs on top of a light box constructed of incandescent lamp arrays. Experiments on subjective discomfort glare were also performed by Tuaycharoen & Tregenza (2007), using a number of screen projected images displaying natural and man-made sceneries. A similar technique of using projected image on a screen was applied by IJsselsteijn et al. (2008), in their investigation on subjective depth perception cues. In all of those experiments, the prototypes/displays were assumed to be the representation of what the subjects normally see through a real window.

Next to the backlit and projection image technique, other researchers and manufacturers have utilised electronic large, high-definition (HD) monitor displays for the purpose of simulating window-views in a more flexible manner. A number of commercial products of virtual window have been developed (Figure 2.4), consisting of a number of LCD screens, displaying a recorded, realistic moving image that could be chosen by the users, e.g. Rational Craft (2013) and Windauga (2013).

(a) (b)

Figure 2.4. Examples of solutions with a complex view and diffuse light using HD monitor display showing dynamic image: (a) Winscape ®, taken from Rational Craft (2013) and (b) Windauga ®, taken from Windauga (2013)

Chapter 2

18

A specific type of VNLS prototype with motion parallax has been developed by Gaver et al. (1995), for the purpose of remote communication. The motion parallax can be simulated if the location of the viewer’s head in relation to the display is known. Local head locations were detected by a tracking camera and were used to control a moving camera in the remote office. The effect was that the image on the local monitor changed as if it were a window. In particular, the prototype might offer an expanded field of view and reduced visual discontinuities. The prototype was however still relatively large, slow, and inaccurate for an extended application. The number of people who can correctly experience the motion parallax was also limited to one.

Another research on a VNLS prototype with a complex view was performed by Radikovic et al. (2005). They presented a system using a head-coupled display and image rendering to simulate a photorealistic view of nature with motion parallax. A pan-tilt-zoom camera tracked the observer as long as the face was visible to the camera. Below the camera was a large display showing the window view that should be seen from the observer’s position. Evaluation data obtained from test subjects suggested the prototype was a better window substitute than a static image, and had significantly more positive effects on the observers’ arousal, positive affects, and interest. The test subjects judged the system prototype as an acceptable replacement for a real window, and gave it higher ratings for realism and preference than a static image.

Research on HD monitor displays without motion parallax was conducted, for example by Friedman et al. (2008) and Kahn et al. (2008). The monitors were installed on the walls of seven inside offices of faculty and staff at a university, and displayed, as the default image, real-time views of the immediate outside scene (Figure 2.5). Data were collected over a 16-week period to explore the user experience with these large display windows. The results showed that users deeply appreciated many aspects of the experience. One of the benefits was the reported increase in users’ connection to the wider social community, connection to the natural world, psychological well-being, and cognitive functioning.

Figure 2.5. Examples of solution with a complex view and diffuse light, in the forms of HD monitor display showing real-time, dynamic image, taken from Friedman et al. (2008)


19

To summarise, the classification of solutions with complex view and mainly diffuse light is shown in Table 2.2 as follows.

Table 2.2. Classification of solutions with a complex view and mainly diffuse light

Source Features References

Incandescent/ fluorescent/LED

Transparent printed photograph, static view

Shin et al. (2012); Kim et al. (2012); TESS (2012)

Projection Recorded, static and dynamic view Tuaycharoen & Tregenza (2007); IJsselsteijn et al. (2008)

HD monitor display

Recorded, dynamic view, no parallax

Rational Craft (2013); Windauga (2013)

Recorded, dynamic view, with parallax

Gaver et al. (1995); Radikovic et al. (2005)

Real-time, dynamic view, no parallax

Friedman et al. (2008); Kahn et al. (2008)

2.3. Expectation of VNLS

A virtual natural lighting solution should give effects comparable to or even better effects than the real natural lighting solutions. The latter has a number of essential properties, which can be generally classified into two categories, i.e. light and view qualities. The light quality mainly indicates the photometric output of the solution, and normally can be described in numerical values. The view quality is mostly related to the process of seeing and perceiving the viewed objects, and may not always be reported in numbers. Nonetheless, they do have a significant contribution in developing a virtual natural lighting solution.

2.3.1. Light quality

The properties of light quality are derived from characteristics of the real natural light opening, in its role as a light source. Since natural light is highly variable, the values range of the properties is also variable, depending on the time and weather condition. Those properties and their general range are listed as follows.

1. Surface luminance. Luminance of a surface is the luminous flux emitted in a given direction divided by the projected area of the surface element in that direction. Surface luminance of a natural light opening (viewed from inside) mostly depends on the sky condition and the transmittance of the glazing material. For example, under the CIE overcast sky, the window surface luminance can range from zero until around 3000 cd/m2. Under the CIE clear sky, the values can be four or five times higher, or even more if direct sunlight is present. However, based on the findings of Shin et al. (2012), it is preferable to

Chapter 2

20

have an average absolute surface luminance of not more than 3200 cd/m2, that is the value at which on average people perceive the glare from simulated windows as ‘acceptable’, i.e. scored as 2.5 out of 4.5 on their discomfort glare scale. On the other hand, it should be noted that it is usually not the absolute luminance that causes the discomfort, but rather the difference of luminance values between two adjacent surfaces, i.e. the contrast.

2. Colour temperature. The (correlated) colour temperature is the temperature of the Planckian radiator whose perceived colour most closely resembles that of a given stimulus at the same brightness, under a specified viewing condition. Natural light has a wide range of CCT (Chain et al., 1999, 2001), from warm colours (red to yellowish white, during sunrise/sunset: 2700 ~ 3000 K) until cool colours (bluish white, during sunny day/around noon: 5000 ~ 6500 K; and blue, during overcast day or very blue sky condition: 6500 ~ 20000 K), which varies over time in a day. Note that for electric lighting, the colour temperature is predominantly constant, as opposed to natural lighting.

3. Colour quality. Natural light generally shows a ‘complete’ spectral power distribution (SPD), while artificial light sources have an ‘incomplete’ SPD, e.g., incandescent lamps usually lack on blue light component, fluorescent lamps have mercury spikes at some wavelengths which outnumber the quantity of other wavelengths. Bouma (1948) described natural (day-) light as the ideal source of illumination for good colour rendering, because ‘it displays a great variety of colours, makes it easy to distinguish slight shades of colour, and the colours of objects around us obviously look natural’.

For artificial light sources, the CIE Colour Rendering Index (CRI) has been widely used for many years to measure their ability to reproduce the colours of various objects, in comparison with an ideal (blackbody radiator) or natural light source. It ranges from 0 to 100. However, the CRI is 40 years old and various problems with the CRI when used for light-emitting diode (LED) sources have been identified, as reported in many publications (e.g. Rea & Freyssinier-Nova, 2008; Zukauskas et al., 2009; Davis & Ohno, 2009; Davis & Ohno, 2009).

A new alternative to describe colour rendering is the Colour Quality Scale (CQS) developed by The National Institute of Standards and Technology (NIST) (Ohno & Davis, 2010). It addresses the problems of the CRI for solid state lighting sources, also has 0 ~ 100 scale, and yet maintains good consistency of scores with the CRI for traditional sources. Nevertheless, more concepts have been proposed earlier and later, and yet there is no consensus found on this issue.

4. Directionality. The directionality of light is a balance between the directional and diffuse components within the luminous environment. In the design stage, these properties are normally described with the sky component (SC [%]), externally reflected component (ERC [%]), and internally reflected component (IRC [%]); which are illustrated in Figure 2.6. This approach is known as the split-flux method, and is applied to predict the daylight factor (DF [%]) at any given point inside a building, according to Equation 2.1.


21

DF = SC + ERC + IRC (2.1)

(a)

(b)

(c)

Figure 2.6. Section view of a building with a calculation point (U) and the presence of (a) sky component, (b) externally reflected component, and (c) internally reflected component

Chapter 2

22

Under the CIE clear or intermediate sky, light comes from a certain direction and may create shadow. Under the CIE overcast sky, which is the assumption in daylight factor calculation, the sky provides diffuse light coming in uniform direction (circle-shaped luminous intensity distribution), but the shape and surface of the glazing may still direct the incoming light.

Related to the built environment, Inanici & Navvab (2006) defined directionality as a ratio between directional and diffuse luminance components in a given space. It gives an indication about the spatial distribution of light flow onto an element or into a space. They suggested an image subtraction method for obtaining the diffuse and directional components of light, by analysing two images of a given luminous environment scene; one image includes the directional and diffuse components of the luminous environment, while the other excludes the diffuse component. The ratio of the directional and diffuse components is then obtained from the average luminance values of both images. The target range is between 1.4 : 1 (typical CIE overcast sky) and 2.5 : 1 (typical CIE clear sky).

However, that indicator depends largely on the environment setting (e.g. room dimension, indoor surface reflectance) where the solution is applied. For describing the characteristic of the existing prototypes, parameters with more qualitative levels are proposed, classified as follows: A (the best choice; generates non-uniform, directional light and possibility to control or vary the directionality), B (generates non-uniform, directional light but no possibility to control or vary the directionality), C (generates a uniform pattern of diffuse and directional light), and D (the worst choice; generates a uniform pattern of only diffuse light).

2.3.2. View quality

The view quality is defined as quality of the outside view (image) experienced by the viewer in the room. Studies on what kind of component should be present in the viewed image have been done previously by many researchers (e.g. Ulrich, 1984; Ulrich et al., 1991; Tennessen & Cimprich, 1995; Chang & Chen, 2005; de Kort et al., 2006; Aries et al., 2010; Beute & de Kort, 2013). In their experimental studies, Tuaycharoen & Tregenza (2007) stated that “view is simply one of the ways we interpret or perceive the light that flows in through the window”, and therefore cannot be separated from the natural (day-) light itself. They concluded that a good view (also referred as a view with high interest), which mainly consists of natural scenes, tends to reduce glare perception.

Related to daylight and view, Hellinga & de Bruijn-Hordijk (2009) proposed certain quality levels for themes that influence visual comfort. They proposed parameters with qualitative levels, which are classified as: A (the absolute best choice for that parameter), B (good), C (sufficient), and D (insufficient); as shown in Table 2.3. The values for the


23

different quality levels are based on values found in the literature (e.g. Kaplan & Kaplan, 1989; Tregenza & Loe, 1998).

Table 2.3. Quality levels for parameters of view that influence visual comfort, adapted from Hellinga & de Bruijn-Hordijk (2009)

Parameter A B C D

- Green, sky, and distant objects

The view contains all 3 elements

The view contains 2 of the 3 elements

The view contains 1 of the 3 elements

The view contains none of the 3 elements

- Information The view gives maximum information about outside environment: weather, sea-son, time of day, activities

The view gives information about weather, season, time of day, and activities

The view gives information about weather, season, and time of day

The view gives no information

- Organisation The view is highly complex and coherent

Medium complexity and coherence

Low complexity and coherence

The view is simple and incoherent

The parameters proposed by Hellinga & de Bruijn-Hordijk (2009) were defined specifically for windows. In their analyses of windows, ‘green’ has a specific meaning and importance as natural elements, irrespective of being nearby or distant. In this thesis, VNLS covers not only windows, but also skylights. For example, ‘green’ in a virtual skylight display can be the leaves and branches of a tree (see Figure 2.3), whereas there are more examples for those in a virtual window display. In terms of information, activity through a skylight is rarely human, but can for example be flying birds.

In addition to the view quality, IJsselsteijn et al. (2008) performed experiments on the efficacy of three ‘depth perception cues’, i.e. motion parallax, occlusion, and blur, using projected photorealistic scenes to create a window-like ‘see-through experience’, as illustrated in Figure 2.7. The three cues can be briefly explained as follows:

1. Motion parallax is the phenomenon that occurs as one moves his/her body (or only head) from side to side, more distant objects will traverse smaller angles across the retina than do objects closer by (Markus, 1967; IJsselsteijn et al., 2008). Technically, it is an apparent displacement in the position of an object viewed along two different lines of sight, and can be measured by the angle or semi-angle of inclination between those two lines.

Chapter 2

24

2. Occlusion or window framing is the addition of a frame, which is expected to provide additional depth information regarding the position (depth layer) of the frame related to the outside view, via the occlusion or interposition cue, particularly in the case where motion parallax is present. For more details, see Cutting & Vishton (1995).

3. Blur is the addition of something that is hazy and indistinct to the boundaries of the frame, which would give a signal to the visual system that the frame is located at a different depth layer than the view being displayed (IJsselsteijn et al. 2008).

Figure 2.7. Schematic representation of the experimental 2 × 2 × 2 design by, and taken from IJsselsteijn et al. (2008), varying in blurring of the frame, the presence of an occluding cross-shaped frame, and the presence of movement parallax

The results of IJsselsteijn et al. (2008) indicated all of the three cues have a significant

main effect on the viewer’s see-through experience, with motion parallax yielding the greatest effect size (F(1,19) = 24.86, p < 0.001). Following the classification of three aforementioned properties, the depth perception cues therefore can also be evaluated in four levels as follows: • Level A: the view contains all of the three depth perception cues • Level B: the view contains two depth perception cues • Level C: the view contains one depth perception cues • Level D: the view contains none of the three depth perception cues

2.3.3. Target range

Based on the discussed expectation in Sections 2.3.1 and 2.3.2, the target range of the VNLS properties can be summarised in terms of light and view qualities. In this section, the

light quality is described by the surface luminance (Ls), colour temperature (CCT), colour

quality scale (CQS), and directionality (DIR), whereas the view quality is described by


25

presence of green, sky, and distant objects (GSD), information (INF), organisation or complexity and coherence (ORG), and depth perception cues (DPC). Table 2.4 summarises the possible and target range of VNLS properties as follows.

Table 2.4. Target range of VNLS properties

Properties Symbol Unit Possible range

Target range

Light quality

Surface luminance Ls cd/m2 0 ~ ∞ 1000 ~ 3200

Colour temperature CCT K 2700 ~ 17000 2700 ~ 6500

Colour quality scale CQS - 0 ~ 100 71 ~ 93

Directionality DIR - A, B, C, D A or B

View quality

Presence of green, sky, and distant objects

GSD - A, B, C, D A or B

Information INF - A, B, C, D A or B

Organisation ORG - A, B, C, D A or B

Depth perception cues DPC - A, B, C, D A or B

2.3.4. Comparison of prototypes In order to compare the light and view qualities, the existing prototypes that have been

discussed in Section 2.2 are classified based on their specific features. The classification is intended to be generic, and does not specifically point to the products taken as examples. All values, particularly the ones related with light qualities, are roughly estimated based on given specification from which the properties in question are derived. For instance, the luminous

flux of the light source Φv [lm] may be given in the product specification, or can be estimated by multiplying the light source’s electrical power with its typical luminous efficacy. For diffuse (or very low directionality) light source, the light is transmitted approximately evenly

over the entire hemisphere, i.e. the solid angle Ωi is equal to 2π sr. The luminous intensity

therefore can be predicted, and hence also the surface luminance Lv [cd/m2], expressed as follows:

Lv = ii

vA

(2.2)

where Ai is the projected light source surface [m2].

Chapter 2

26

Regarding the view quality, it is obvious that the prototypes with a simplified view obtain the lowest score (level D), since most of them display very little or even no view at all. For the prototypes with a complex view, the view quality is estimated by observation. The summary of the comparisons is given in Table 2.5.

Table 2.5. Comparisons of the properties of existing prototypes with a simplified view + mainly diffuse light and a complex view + mainly diffuse light

Properties Light quality View quality

Ls CCT CQS DIR GSD INF ORG DPC

Features [cd/m2] [K] [-] [-] [-] [-] [-] [-]

Simplified view, diffuse light

Fluorescent/LED, large brightness, static view

≤ 100002700

~6500 64~93 D D D D B

Fluorescent/LED, varying brightness, static view

≤ 10000 2700 ~6500

64~93 C C C C B

Complex view, diffuse light

Incandescent/ fluorescent/LED, transparent photograph, static view

≤ 100002700

~6500 64~93 D A~B C A~B C

Projection, recorded, static view

≤ 500 6500 71~93 D A~B C A~B A

HD monitor, recorded, dynamic view, no parallax

≤ 1000 5500

~1050071~93 D A~B B A~B C

HD monitor, recorded, dynamic view, with parallax

≤ 1000 5500

~1050071~93 D A~B B A~B B

HD monitor, real-time, dynamic view, no parallax

≤ 1000 5500

~1050071~93 D A~B A A~B C

Ideal VNLS 1000 ~3200

2700 ~6500

71~93 A~B A~B A~B A~B A~B

Referring to the description in Sections 2.3.1 and 2.3.2, the following explanation can be given:


27

1. Surface luminance: A number of artificial light sources can be combined to create a prototype generating up to 10000 cd/m2 of luminance, for instance in the case of the light box display of Shin et al. (2012). A single HD monitor display normally gives less than 10% of that amount. An image projection normally gives a very low luminance (less than 500 cd/m2), considering the reduction of light after reflection from the (diffuse) screen.

2. Colour temperature: Most prototypes with transparent printed photographs generate light with colour temperatures between 2700 K and 6500 K, depending on the type of the light source. The high intensity lamps for healing purpose can give a very bright light, which is required to cure diseases such as winter blues. Colour temperature of this type of lamp is usually very high; in some cases can be up to 17000 K, to enhance the blue light properties. Most HD monitor displays generate light with high colour temperatures, somewhere between 5500 K and 10500 K. Reducing the colour temperature gives the entire screen an increasingly reddish cast, while increasing it makes the colour cast increasingly blue. For ordinary personal computer use, a colour temperature of 6500 K is standard. For television broadcasting, the value is defined based on national standards; most countries use a colour temperature standard of 6500 K according to the US broadcasting standard (NTSC), whereas 9300 K is used under the Japanese standard (NTSC-J); see more details in EIZO (2013).

3. Colour quality: In their experiment on a number of samples, Ohno & Davis (2010) showed that most recent phosphor LED type products had a CQS of 71 ~ 93. Their tested fluorescent type products had a CQS of 64 ~ 80, whereas an incandescent type had 98. An HD monitor display is assumed to have a similar colour quality as LED type products.

4. Directionality: Most prototypes do not feature any directional light at all. The prototype of Philips (van Loenen et al., 2007) included an HID lamp for simulating the sun position. This lamp created the impression of a sun ‘spot’ on the blurred display and in the room, although smaller in size and weaker in luminous intensity than real sun patches.

5. Presence of green, sky, and distance objects: Prototypes with a simplified view mostly display none of the three layers. The prototype of Philips (van Loenen et al., 2007) displayed a blurred sky scene by varying the intensity of each coloured light source behind the diffuser, hence essentially showed one layer. Most of the prototypes with a complex view display at least two layers, i.e. sky and distance objects or sky and green objects.

6. Information: Obviously there is very little to see from the prototypes with a simplified view, except vague information about the weather, depicted on the sky display. Most of the prototypes with a complex view display sceneries from which the weather, season, and time of day can be deduced. In the case of a static view, the information is unchanged, whereas in the case of recorded, dynamic view, the information may be repeated after some time. Table 2.3 describes human activities as important information to display; since human activities are constantly changing, a dynamic view is required.

Chapter 2

28

The maximum information about the outside environment is obtained when a real-time view is presented, ensuring no pattern repetition.

7. Organisation: In line with the information property, there is very little or no complexity on most of the prototypes with a simplified view. Prototypes with a complex view mostly display sceneries of medium or high complexity (see Figures 2.4 and 2.5).

8. Depth perception cues: All of the prototypes with a simplified view do not feature any motion parallax, but mostly have a blurred display. The occlusion effect can be present when the prototype is put behind window frames. Most of the prototypes with a complex view do not come with motion parallax either, but do have occlusions or window framing. In the prototype with a transparent photograph and a static view, the frame boundaries are very often not enough to give an impression that there is a different depth layer between the frame and the view. In the prototype with projected image, the three depth perception cues can be created (IJsselsteijn et al. 2008). In the prototype with a recorded dynamic image, the occlusion effect is present, but the blur effect is missing, as shown for instance in Figure 2.4.

This overview shows that an ideal VNLS does not yet exist at the moment. The existing prototypes have their own limitations, and each prototype addresses only a subset of all aspects required for an ideal VNLS. For instance, directionality of the incoming light is an issue to be addressed, since most displays will provide only diffuse light, and it is very challenging to imitate the constantly moving direction of natural light. In some prototypes, the light direction may be varied by aiming the lamps behind the display into different angles, but this also needs a sophisticated control system.

Dynamics of the view and motion parallax are even harder to imitate. A video display may create a moving image, but the given motion parallax effect is mostly limited to one position in front of the display, depending on the number of sensors installed. High detailed images may not be required for skylights, but it is required for the vertical windows. A very complex view can be provided by an HD monitor display, either recorded or real-time, but the light quality criteria are yet to be satisfied. One of the most interesting challenges in this direction is probably to create an HD monitor display with real-time image at a sufficient depth to the frame, sufficient light output, and also with the appearance of motion parallax.

Looking at the discussed aspects, direction of further development should be steered toward improving the light directionality and view dynamics. In order to approach the ideal condition, a number of evaluation stages must be performed, including theoretical analysis, initial design, numerical testing of the design, prototype construction, physical testing, subjective laboratory testing, field trials, and so on. In the early design stage, computational modelling and simulation is a powerful tool to predict the system performance, with regards to the relevant physical phenomena. By using computational modelling and simulation, one is able to rapidly test multiple design concepts for better solutions, in an efficient way in terms of time and cost.


29

2.4. Concluding Remarks

A number of efforts have been made to recreate the elements of natural light inside buildings, in the form of artificial solutions. Such solutions, the so-called VNLS, can be generally classified based on their light and view qualities into four types, which are those providing: (1) simplified view and diffuse light, (2) complex view and diffuse light, (3) simplified view and directional light, and (4) complex view and directional light. The first two types already exist in reality as prototypes, while the last two are still under development in conceptual models. Building performance simulation tools have their role to predict the performance. Most prototypes with a simplified view appear in the shape of ‘light box’, which is generally constructed of a series of artificial light sources behind a diffuse surface. Despite the insufficient view on their display, research in health-related contexts has shown that installation of this particular prototype, under a large brightness setting, can be useful for the purpose of healing or therapy. Adding a complex view on VNLS prototypes is also beneficial, for example to reduce the anxiety of patients in windowless healthcare environments. Prototypes with a complex view are also available for entertainment purpose as well as for application in office environments, to provide the feeling of being connected to the outside world.

Based on the given overview of relevant properties, further research and development should be directed toward improving the light directionality and view dynamics of the prototype. At the current situation, all of the existing prototypes only address a part of the ideal VNLS properties. Particularly in the early design stage, computational modelling and building performance simulation can help steering the process of VNLS design development, with regards to the relevant physical phenomena. In this context, the use of computational modelling and simulation can help to rapidly test multiple design concepts in an efficient way in terms of time and cost, even though the results still need to be validated by real user experiments.

Finally, it is learned that a number of subject-based experiments have been performed using various simple forms of VNLS prototype, to gain knowledge on how people perceive it, and/or to investigate which aspects of natural light people appraise in reality. Nonetheless, very little is known about how the prototypes physically influence the indoor lighting condition where the prototypes are installed. This suggests another research direction, which is to evaluate the objective performance of the prototypes, and to propose better design solutions to improve it.

Chapter 2

30

31

Chapter 3 Measurement and Simulation of a First Generation Virtual Natural Lighting Solutions Prototype

This chapter discusses an example of application of Radiance as a simulation tool to reproduce the scenes and to evaluate the lighting performance of a first generation VNLS prototype with a simplified view and diffuse light displaying various sky scenes, located in a test room with no façades. The performance of the measured prototype was compared to a hypothetical real window under the same settings in simulation.

3.1. Introduction

The overview of various examples of prototypes in Chapter 2 shows that an ideal VNLS does not yet exist at the moment. In order to approach the ideal condition, a number of evaluation stages must be performed, such as theoretical analysis, initial design, numerical testing of the design, prototype construction, physical testing, subjective laboratory testing, field trials, and so on. In the early design stage, computational modelling and simulation is a powerful tool to predict the system performance, with regards to the relevant physical phenomena. Computational modelling and simulation are subsets of design methods that can be used to rapidly analyse multiple design concepts, and to predict their performance.

In the context of this thesis, it is intended to know how a certain VNLS prototype will influence the indoor lighting condition and visual comfort under certain scenarios. Moreover, it is also intended to further improve the performance towards the ideal condition, by introducing better design options. Therefore, there is a need to create a representative model of the prototypes, and to predict their performance by mean of simulations. For a lighting- and view-related system such as VNLS, a sophisticated lighting modelling and simulation technique is therefore needed.

3.1.1. Modelling concept in Radiance In general, lighting simulation technique can be divided into two main types: the first one

is photorealistic rendering, which produces images that are mostly used for artistic impressions or perception studies, and the second one is physically-based visualisation, which produces a physically accurate representation and predicts reality under given conditions (Ward & Shakespeare, 1998; Moeck & Selkowitz, 1996; Ochoa et al., 2012). Each technique has algorithms and their supporting calculation methods, which also come with their own specific applications and limitations (Ochoa et al., 2012). The most generally used algorithms are:

Chapter 3

32

1. Direct calculations used for artificial lighting; these are specific physical formulas and simplifications, often delineated in national standards to cover most usual illumination situations. The algorithms are simple and often used as rules-of-thumbs, but can lack accuracy in the real situation.

2. View-dependent algorithms; these are classified based on direction from which tracing rays are computed; i.e. from the light source (forward raytracing), from the observer’s eyes (backward raytracing), or from light source and observer (bidirectional raytracing). They are used for lighting calculations and renderings, and require a specific observer position.

3. Scene-dependent algorithms; these are mainly radiosity calculations, adapted from heat transfer techniques. They are mainly used for calculations but not for rendering due to complex formulas.

As reported by Maamari et al. (2005) and Maamari et al. (2006), the CIE technical committee 3.33 has defined a set of simple test cases, based on analytical or experimental references, with the objective of assessing lighting computer programmes. It was shown that the use of the CIE test cases allows verifying the accuracy level of the tested programmes, with respect to physical laws in lighting. They observed that for Radiance (Ward & Shakespeare, 1998), good accuracy was observed in general, except for the indirect lighting test with reflectance values of 0.8 and above. However, they suggested that a single ideal lighting programme does not exist, but some are adapted to given tasks and constraints. For the purpose of daylighting modelling and simulation in particular, Radiance has been validated elsewhere (e.g. Mardaljevic, 1995, 1997; Reinhart & Herkel, 2000; Reinhart & Walkenhorst, 2001; Reinhart & Andersen, 2006).

In general, Radiance calculates the outgoing radiance (Lo) as the sum of total reflected

radiance (Lr) and emitted radiance (Le) from a surface (dA) to another surface, defined by the radiance equation as follows:

Lo (ψo ,θo) = Le (ψo ,θo) + Lr (ψo ,θo ,ψi ,θi) (3.1)

where Lo (ψo, θo) [W/(sr·m2)] is the outgoing radiance in a direction given by the angles ψo and

θo expressed in [rad]; which is composed of Le (ψo, θo) [W/(sr·m2)], the emitted radiance, and

Lr (ψo, θo, ψi, θi) [W/(sr·m2)], the reflected radiance as function of the incoming angles ψi and

θi and outgoing angles ψo and θo, all given in [rad]. In turn, the reflection is further defined as

the function of the incoming radiation (Li):

Lr (ψo, θo, ψi, θi) = iiiiiiiiioor ddLfi i

sincos),(),,,( (3.2)

where fr (ψo, θo, ψi, θi) [sr-1] is the reflection as a function of outgoing angles ψo and θo and the

incoming angles ψi and θi [rad], and Li [W/(sr·m2)] is the incoming radiance from a specific projection, as a function of the incoming angles of the incoming ray representing that

Measurement and Simulation of a First Generation VNLS Prototype

33

projection. The incoming radiance is integrated over the incoming angles or solid angle Ωi [sr] of the incoming radiance, as illustrated in Figure 1.

Figure 3.1. Schematic illustration of radiance equation and the corresponding variables in Equations (3.1) and (3.2), adapted from Pelzers et al. (2014)

Emission of a given light source can be modelled with a material type or obtained from photometric measurements. While experimental data for visible light is normally obtained with a moving-cell photometer, the emission data may also be approached with the ‘light’ material type. Normally, it is applied as general light source material which emits radiation

(Li [W/(sr·m2)]) on which the material type is applied, defined by the three radiance values:

LR, LG, and LB, for the red, blue and green component respectively [W/(sr·m2)]. The radiation

is then related to the total radiative flux of light source (Φi [W]) and the projected light source

surface (Ai [m2]) by the following equation:

Li = ii

iA

(3.3)

In principle, Radiance solves the radiance equation for the red, green, and blue (RGB)

values separately to obtain the radiance Li [W/(sr·m2)], or the irradiance IR,G,B [W/m2], if

integrated over the solid angle Ωi [sr]. When a picture is rendered, the spectral irradiance

values in red, green, and blue (IR, IG, IB, respectively) are summed and weighted to obtain the

single value of IR,G,B, according to Ward & Shakespeare (1998):

I R,G,B = 0.265 IR + 0.670 IG + 0.0648 IB (3.4)

Assuming a conversion factor of 179 lm/W (Ward & Shakespeare, 1998), the irradiance

values can be easily converted to illuminance (E [lx]) as follows:

E = 179 I R,G,B = 179 (0.265 IR + 0.670 IG + 0.0648 IB) (3.5)

Chapter 3

34

Recent development within the Radiance community has led to application of spectral rendering with Radiance for other purposes (Geisler-Moroder & Dur, 2010). In fact, Ruppertsberg & Bloj (2006) showed that spectral rendering could actually be more accurate than the original Radiance RGB model. Nevertheless, Equations 3.4 and 3.5 remain as the governing equations, which in most cases are applicable to determine illuminance values on individual sampling points.

Another important feature in Radiance is the ambient calculation, which consists of all diffuse and indirect specular reflection, diffuse transmission, and emitted light from the secondary light sources. The main ambient parameters are ambient bounces, ambient divisions, ambient super samples, ambient accuracy, and ambient resolution. A more detailed explanation on these parameters can be found in Ward & Shakespeare (1998).

In the context of this thesis, Radiance is employed to address indoor lighting and visual comfort aspects of VNLS. Very little is known about how these solutions influence the indoor condition in various scenarios; for instance how the image variation affects the lighting performance on the workplane. Another important question is how a VNLS prototype actually correlates to a real window or skylight; could a VNLS prototype perform as good as, or even better than the real one? A comparison to the real window is then required on that aspect; such a comparison will be useful for designing a better solution in the future.

Therefore, the study in this chapter aims to address lighting measurements and simulation of a ‘first generation’ VNLS prototype. The objective is to evaluate the lighting performance of the prototype under various settings, to analyse if the results can be accurately reproduced in simulation, and to compare the performance with the corresponding simulated real windows.

3.2. Case Description

An example of the so-called first generation prototype is the one developed by Philips (van Loenen et al., 2007), which is briefly discussed in Section 2.2.1. Due to the possibility to vary the view display and to add directional light using a spot lamp for sun simulation, this prototype was selected as the first case study to demonstrate how Radiance can be employed to reproduce the scenes and obtain the lighting performance of the space, validated by actual measurements.

The prototype was installed in a kitchen laboratory setting, located in the ExperienceLab of Philips Research in Eindhoven, the Netherlands. The prototype was constructed of 12 colour tubular fluorescent (TL5) lamps of 54 W each, put in an array of 12 rows, and covered with a diffuse panel of 1.20 m × 1.20 m. A halogen, parabolic aluminised reflector (PAR) spot lamp of 70 W was installed in the upper right corner to simulate the sunlight.

The construction was put vertically in an adjacent control room behind a transparent, clear glass window which was a part of the kitchen room interior. During the experiment, the general lighting in the kitchen was switched off all the time. The room had no façades and


35

real windows, ensuring no daylight admission. No motion parallax was associated with this prototype.

The TL5 lamp array was covered by a white, diffuse panel, installed 0.35 m behind the window glass plane. The dimension of the diffuse panel was 1.20 m × 1.20 m, while the window opening was 0.65 m × 0.65 m. The 12 TL5 lamps were divided into four groups; each group consisted of three lamps emitting red, green, and blue light, respectively. Every lamp had its own ballast so that it could be dimmed independently, using the Digital Addressable Lighting Interface (DALI) system. The overcast, clear, and partly cloudy sky scenes were realised by adjusting the intensity of each lamp and were subjectively evaluated to imitate the real sky scenes. Table 3.1 shows the type of colour emitted by each lamp, the electrical power rating, and the intensity level settings for the three scenes.

Table 3.1. Intensity level settings for the three sky scenes of the prototype

Lamp’s row (from

top) Type

Power rating [W]

Overcast Clear Partly cloudy Intensity level

[%] Intensity level

[%] Intensity level

[%] 1 Red 54 30 15 15

2 Green 54 30 3 15

3 Blue 54 30 100 100

4 Red 54 30 80 80

5 Green 54 30 15 80

6 Blue 54 30 100 100

7 Red 54 20 0 0

8 Green 54 45 100 100

9 Blue 54 30 100 100

10 Red 54 20 80 80

11 Green 54 20 0 80

12 Blue 54 20 3 0

n/a PAR 70 0 100 100

Three scenes were defined for the lamp setting to simulate the colour gradients of typical sky conditions, i.e. ‘overcast’, ‘clear’, and ‘partly cloudy’. The interior, equiangular fisheye views of the kitchen room under the three scenes of the prototype are shown in Figure 3.2. In general, the overcast scene (the PAR spot lamp was turned off) produced a white-grey appearance on the window, while the other two scenes (the PAR spot lamp was turned on) produced a combination of light blue and red appearance with a bright spot in the upper right corner, suggesting the sun’s position, which was not moved during the experiment.

Chapter 3

36

(a) (b) (c)

Figure 3.2. Interior view of the kitchen room with the prototype under the (a) overcast, (b) clear, and (c) partly cloudy sky scenes

The actual lighting performance was measured and obtained by collecting the following data at certain lighting conditions: • Horizontal illuminance on the workplane; data were collected for 55 horizontal points on

the workplane height, i.e. the countertop in the kitchen (0.95 m from the floor). • Vertical illuminance on the observer’s eye plane; data were collected for two vertical

points on the typical observer height (1.20 m from the floor). • Minimum, maximum, and average luminance perceived by the observer; data were

collected for two vertical points on the typical observer height (1.20 m from the floor). • Reflectance of interior surface materials; data were collected for the relevant interior

surface, such as floor, walls, ceiling, and furniture.

Furthermore, the horizontal illuminance data were post-processed to obtain the average

illuminance values (Eav [lx]), the uniformity (U0), and the space availability (%A [%]). The latter is defined as the percentage of the measuring points satisfying a minimum illuminance value of 500 lx, which is the required indoor illuminance criterion for kitchens (CEN, 2002). These three indicators can be expressed in equations as follows:

Eav =

N

EN

ii

1 (3.6)

U0 = avE

Emin (3.7)

%A = N

N lxE 500 × 100% (3.8)


37

where Ei [lx] is the horizontal illuminance on each measuring point, Emin [lx] is the minimum

horizontal illuminance, lxEN 500 is the number of measuring points satisfying the criterion

of minimum illuminance value of 500 lx, and N is the total number of measuring points.

To evaluate the visual comfort in this case, the Daylight Glare Probability (DGP) (Wienold & Christoffersen, 2006) was used as an indicator, which can be expressed as follows:

DGP = 5.87 × 10–5 Ev + 9.18 × 10–5 log 2

n

i iv

isis

PE

L

1 287.1,

2,1

(3.9)

where Ev is the total vertical eye illuminance [lx], ωs is the solid angle of the glare source [sr],

Ls is the glare source luminance [cd/m2], and P is the position index, i.e. a weight factor based on position in a viewing hemisphere.

During the measurement, the following instruments were used: • SpectraDuo PR-680 photometer; for measuring luminance and illuminance values, as well

as spectral power distribution. • Canon EOS50D digital single-lens reflex camera + Sigma 4.5mm fisheye lens + Photolux

3.1 software; for taking multiple (20 in this case) photographs in equiangular 180° view with various exposure values, which in turn were post-processed to obtain the luminance pictures. The luminance values were calibrated with the SpectraDuo photometer.

• Konica Minolta CM-2600D spectrophotometer; for measuring reflectance values of the interior surface materials.

3.3. Measurement Protocol

Horizontal illuminance data were collected on 55 points at a height of 0.95 m (countertop level) as displayed in Figure 3.3. Vertical illuminance and luminance perceived by the observer were measured by taking 20 photographs (ISO 400, f/5.6, shutter time varied from 4 s to 1/8000 s) each at positions 1 and 2, at a height of 1.20 m, with the view direction specified by the arrows in Figure 3.3.

To determine the glare index value at both observer’s positions, the obtained photographs were exported to Radiance, combined into High Dynamic Range (HDR) images using the Hdrgen programme, and analysed using Evalglare (Wienold & Christoffersen, 2006).

Chapter 3

38

Figure 3.3. Floor plan of the kitchen with the measuring points for horizontal illuminance

3.4. Simulation Protocol

Since the test room was not connected to the building’s façade, the condition under a real window could not be observed. Therefore, the real window scene was modelled and simulated in Radiance. In addition, the actual conditions under all scenes of the prototype were also modelled and simulated, to give an insight in the difference between simulation and actual measurement. Comparisons were made between the values of horizontal illuminance at the central line, where points P1 and 1 were located (i.e. the blue-coloured points on Figure 3.3). The difference between the average illuminance, uniformity, and space availability was also evaluated.

The front, top, and perspective views of the modelled prototype are displayed in Figure 3.4. The 12 TL5 lamps were modelled as 12 rows of cylinders, with a length of 1.20 m and a diameter of 0.016 m, constructed with a ‘light’ material. Assuming a total luminous flux of 4250 lm for each lamp (Philips, 2013a), a conversion factor of 179 lm/W between photometric and radiometric units (Ward & Shakespeare, 1998), and a solid angle of the incoming radiation of π sr (Ward & Shakespeare, 1998), Equation 3.3 was applied to obtain the total radiance value of each lamp, i.e. 394 W/(sr·m2) at the maximum setting.

Equation 3.4 was applied to obtain the red, green, and blue radiance components for the ‘light’ material. For the red-coloured lamps, the green and blue radiance components were assumed to be zero; for the green-coloured lamps, the red and blue were assumed to be zero; and for the blue-coloured lamps, the red and green were assumed to be zero. Hence, at the maximum setting, the red-coloured lamps were set to have a red component of 1487 W/(sr·m2), the green-coloured lamps have a green component of 588 W/(sr·m2), and the blue-coloured lamps have a blue component of 6059 W/(sr·m2). For other settings, the values were adjusted proportionally.


39

The PAR lamp was modelled as a thin cylinder with a diameter of 0.12 m, aimed at an angle of 45°, and constructed with a ‘light’ material. Assuming a total luminous flux of 1415 lm (Philips, 2013b), and by applying Equation 3.3, a total radiance value of 223 W/(sr·m2) is obtained. The red, green, and blue components were assumed to be equal.

Table 3.2 displays the assigned values for the light sources in the prototype.

(a) (b) (c)

Figure 3.4. (a) Front, (b) top, and (c) perspective views of the modelled prototype. The TL5 lamps and the diffuse panel are coloured in red, the PAR lamp is coloured in green.

Table 3.2. Red, green, and blue irradiance components of ‘light’ material defined in Radiance for the 12 TL5 lamps in the prototype

Lamp’s row

(from top)

Overcast Clear Partly cloudy

Red Green Blue Red Green Blue Red Green Blue

1 446 0 0 223 0 0 223 0 0

2 0 176 0 0 18 0 0 88 0

3 0 0 1818 0 0 6059 0 0 6059

4 446 0 0 1189 0 0 1189 0 0

5 0 176 0 0 88 0 0 470 0

6 0 0 1818 0 0 6059 0 0 6059

7 297 0 0 0 0 0 0 0 0

8 0 265 0 0 588 0 0 588 0

9 0 0 1818 0 0 6059 0 0 6059

10 297 0 0 1189 0 0 1189 0 0

11 0 118 0 0 0 0 0 470 0

12 0 0 1212 0 0 182 0 0 0

PAR - - - 223 223 223 223 223 223

Chapter 3

40

The detailed values assigned for the window construction properties are specified in Table 3.3, together with the room’s interior surfaces reflectance as obtained from the measurement. The properties of the diffuse panel were estimated based on the ‘trans’ model of the translucent panel in Reinhart & Andersen (2006), by fine-tuning the diffuse transmissivity to 0.35.

Table 3.3. Material definitions in Radiance for the window construction and room’s interior

Material Red Green Blue Specula-

rity Rough-

ness Diffuse

transmiss.Transmit. specularity

Diffuse panel 0.21 0.21 0.21 0.08 0 0.35 0

Window glass 0.88 0.88 0.88 - - - -

Window frame 1.00 0.78 0.60 - - - -

Ceiling 1.00 1.00 0.95 0 0 - -

Walls 0.90 0.90 0.90 0 0 - -

Floor 0.56 0.55 0.48 0 0 - -

Door 0.56 0.48 0.56 0 0 - -

Countertop 1.00 1.00 0.97 0 0 - -

Simulations were run for the three sky scenes, i.e. overcast, clear, and partly cloudy. Calculation was performed for the 55 measuring points on the workplane. One-to-one comparison between measurement and simulation was done for all values of horizontal illuminance at the line where the points P1 and 1 were located. This line, at which there were seven measuring points, was located directly in the central projection of the window.

In addition, the prototype scenes were compared to real window scenes. The latter were modelled in Radiance by replacing the entire construction of artificial light sources, with the corresponding sky models, i.e. overcast, clear, and partly cloudy. These sky models were generated in Radiance using the Gensky programme by using the options –c, –s, and +i, respectively. Site location was set for Eindhoven, the Netherlands (51.45°N, 5.47°E), with south-facing window orientation, on 21 June at 12.20 hrs local time.

The zenith radiance [W/(sr·m2)] of each sky was defined so that the illuminance values at the nearest point to the window (P1) were the same under the corresponding real and virtual window scenes. The relevant zenith radiance was respectively 7.5, 5.5, and 11 W/(sr·m2) for the overcast, clear, and partly cloudy skies. Illuminance values on the rest of the points at the central column were determined for the comparison. DGP values at positions 1 and 2 (see Figure 3.3) were also analysed using Evalglare.

Furthermore, simulation parameters in Radiance were set as shown in Table 3.4.


41

Table 3.4. Radiance simulation parameters

Parameter Description Value-ab Ambient bounces 4 -aa Ambient accuracy 0.08 -ar Ambient resolution 128 -ad Ambient divisions 1024 -as Ambient super-samples 256

In order to assess whether the simulation results are fit for the purpose of recreating the measured scene, several criteria can be applied. There is no definitive agreement on an acceptable degree of accuracy (Ochoa et al., 2012). For example, in their report on testing accuracy of various lighting simulation programmes, Maamari et al. (2006) suggested a criterion of two times the global error, based on the estimated error in the measurements and in the scenario description, e.g. sensor cosine and colour corrections, sensor calibration, lumen output fluctuation, luminaire position and flux output distribution, room dimensions, and surface reflectance. These were approximately ±21% from the true value, see also Slater & Graves (2002) and CIE TC-3-33 (2005). According to Fisher (1992), an acceptable criteria range would be 10% for average illuminance calculations and 20% for measured point values. The criterion of 20% for use in real cases has been validated by Reinhart & Andersen (2006), as appeared in studies replicating built realities.

In view of subjective lighting perception, the European Standard EN 12464-1 (CEN, 2002) mentions that “a factor of approximately 1.5 represents the smallest significant difference in subjective effect of illuminance”, as given in the recommended scale of illuminance [lx] for various conditions in work places. This is approximately in line with the findings of Slater et al. (1993) in their subjective study, where illuminance ratios between two work stations of at least 0.7 (or 1.4 if the ratio is inversed) were ‘generally acceptable’. They mentioned that even though there was a trend of decreasing acceptability at lower illuminance ratios, there were indications that lower illuminance ratios may also be acceptable under some conditions.

Taking this recommendation into account, the criterion for which difference between

simulation (Esim [lx]) and measurement (Emea [lx]) values does not lead to a significant difference in subjective effect is:

0.67 < mea

simEE

< 1.50 (3.10)

In other words, the ratio of simulation and measurement values at any measuring point should not be less than 2 : 3 (or approximately 0.67) and not more than 3 : 2 (or 1.50), so that the values do not lead to a significant difference in their subjective effect. This criterion is applied in the following sections to evaluate the simulation results.

Chapter 3

42

3.5. Results and Discussion

Section 3.5.1 presents the measurement results of the prototype. Section 3.5.2 presents simulation results of the prototype, whereas Section 3.5.3 presents those of the corresponding, hypothetical real window in the same positions, under the same sky scenes.

3.5.1. Measurement of prototype

Measurement results of the average illuminance values (Eav [lx]), the uniformity (U0),

and the space availability (%A [%]) under the three sky scenes of the prototype are summarised in Table 3.5.

Table 3.5. Measurement results of the average illuminance, uniformity, and space availability under the overcast, clear, and partly cloudy sky scenes

Eav [lx] U0 [-] %A [%]

Overcast 52 0.28 0

Clear 70 0.28 0

Partly cloudy 102 0.27 0

The measurement results show that at the nearest point to the window, the horizontal illuminance value is found to be 400 lx under the partly cloudy scene, compared to 180 lx under the overcast one. Despite this large variation, the uniformity in the three scenes are relatively similar (0.27 ~ 0.28), which means the influence of the HID spot lamp on uniformity is limited, mainly increasing the total light output.

Moreover, none of the points receives a horizontal illuminance larger than 500 lx, under all sky scenes. As a result, the space availability (taking 500 lx as the minimum criterion) in all cases is zero. It should be noted that the general lighting in the room was completely switched off, to ensure that only the prototype contributed to the light inside the room.

Vertical illuminance on the observer’s eye plane (Ev [lx]), together with minimum (Lmin

[cd/m2]), maximum (Lmax [cd/m2]), and average luminance (Lav [cd/m2]) perceived by the observer at positions 1 and 2 (referring to Figure 3.3) are displayed in Table 3.6. These values were extracted from the post-processing software Photolux 3.1. In addition, the DGP values obtained from Evalglare are also given.

In line with the measurement results of horizontal illuminance, the lowest measured vertical illuminance is also found under the overcast sky scene, while the highest is found under the partly cloudy one. This is also true for the minimum, maximum, and average luminance, as well as DGP perceived by the observer. While the vertical illuminance at position 1 under the partly cloudy scene is around 1.5 times the value under the overcast


43

scene, the maximum luminance under the former is 4 times higher than that under the latter (6000 to 1550 cd/m2). The maximum luminance is actually found on the location of the ‘sun spot’, whereas the vertical illuminance at position 1 is determined by the total window surface area.

Table 3.6. Vertical illuminance on the observer’s eye, minimum, maximum, average luminance, and DGP perceived by the observer at positions 1 and 2 under the three sky scenes of the prototype

Position – scene Ev [lx]Lmin

[cd/m2]

Lmax

[cd/m2]

Lav

[cd/m2]

DGP [-]

1 – overcast 403 0.23 1550 73 0.24

2 – overcast 208 0.18 1540 39 0.21

1 – clear 427 0.28 5500 80 0.32

2 – clear 246 0.22 3400 47 0.28

1 – partly cloudy 600 0.40 6000 111 0.34

2 – partly cloudy 348 0.30 4200 66 0.30

According to a discomfort glare classification of Jakubiec & Reinhart (2012), DGP values of < 0.30 are considered ‘imperceptible’, a DGP range of 0.30 ~ 0.35 corresponds to a ‘perceptible’ category, DGP values between 0.35 and 0.45 represent disturbing glare, and DGP values over 0.45 represent intolerable glare. Hence, only the observers at position 1 under the partly cloudy and the clear sky scenes are expected to experience perceptible discomfort glare from the prototype.

Figure 3.5 displays the luminance false colour pictures of the prototype as seen from position 1; note there are different scales used in the three pictures. The window surface under the overcast scene obviously appears more uniform, whereas a bright spot of the PAR lamp in the upper right corner of the window is revealed under the other two scenes. Combined high dynamic range (HDR) images of the same views are displayed in Figure 3.2, in which the directional light from the PAR lamp leaves its pattern on the countertop (Figures 3.2b and 3.2c).

Chapter 3

44

(a) (b)

(c)

Figure 3.5. Luminance false colour pictures of the prototype observed at position 1, under (a) overcast, (b) clear, and (c) partly cloudy sky scene

From the pictures in Figure 3.5, one can conclude that, as the mean view luminance of the prototype is more than 1800 cd/m2, the display is capable of creating discomfort glare (Shin et al., 2012; Kim et al., 2012). This level is present in the clear and partly cloudy sky scenes. The contrast between the surrounding wall and the window is very often over 1 : 20 or 1 : 40, which is another sign of potential discomfort glare.

3.5.2. Simulation of prototype

Table 3.7 summarises the simulation results of the horizontal illuminance point at the central line on the workplane, together with the overall average illuminance values (Eav [lx]), uniformity (U0), and space availability (%A [%]) under the three sky scenes of the prototype.


45

For comparison, the measurement results, and the ratio between simulation and measurement values are also shown.

The lighting simulation and measurement results of the prototype generally show similar trends with a maximum relative difference of 26%, found on the farthest point from the window, under the overcast sky scene. The maximum relative difference for the average illuminance is 18%, also found under the overcast sky scene. However, the ratio of the simulated value to the measured one at all points is always in the range of 0.67 ~ 1.50, which represents the smallest significant difference in subjective effect of illuminance (CEN, 2002). Looking at the criterion, the models are therefore considered sufficient for the purpose of reproducing the scenes without giving a significant subjective difference, even though more care should be taken when interpreting the modelling results of scenes with relatively low lighting levels, as shown here in the overcast sky scene.

Table 3.7. Simulation (sim.) and measurement (meas.) results of horizontal illuminance point at the central column, together with the average illuminance values (Eav [lx]), uniformity (U0), and space availability (%A [%]) under the three sky scenes of the prototype


Distance to window

[m]

Sim. [lx]

Meas. [lx]

Ratio [-]

Sim. [lx]

Meas. [lx]

Ratio[-]

Sim. [lx]

Meas. [lx]

Ratio [-]

0.4 245 204 1.15 432 354 1.22 533 491 1.09

0.9 155 155 1.00 253 236 1.07 346 355 0.97

1.4 79 88 0.89 139 129 1.07 187 190 0.99

1.9 52 61 0.85 104 102 1.01 132 149 0.89

2.4 38 44 0.86 65 56 1.16 81 80 1.01

2.9 28 36 0.77 48 47 1.02 65 65 1.00

3.4 25 34 0.74 43 44 0.98 61 65 0.94

Eav [lx] 42 52 0.82 79 71 1.11 99 102 0.96

U0 [-] 0.19 0.28 0.68 0.20 0.27 0.71 0.20 0.27 0.74

%A [%] 0 0 n/a 0 0 n/a 2 0 n/a

Chapter 3

46

3.5.3. Simulation of real windows

Figure 3.6 displays the graphs showing the relationship between horizontal illuminance and the distance to the window under the three sky scenes, based on the measurement and simulation of the prototype (VW) and simulation of real windows (RW).

(a) (b)

(c) Figure 3.6. Graphs showing the relationship between horizontal illuminance and distance to window under the (a) overcast, (b) clear, and (c) partly cloudy sky scene

Compared to the corresponding real window scenes, it is seen that all of the modelled sky scenes, the prototype gives a more uniform illuminance distribution throughout the space. Figure 3.6 shows how the light from the real window rapidly drops at the distance of more than 1 m from the window, while the decreases are less dramatic under the virtual window scenes. Under the clear sky scene, the difference between the real and virtual windows is less, due to the influence of direct sunlight, which reduces the illuminance decrease throughout the space with the real window.


47

(a) (b)

(c) (d)

(e) (f)

Figure 3.7. False colour maps of the workplane illuminance [lx] under the (a) overcast, (c) clear, and (e) partly cloudy sky scenes of the measured prototype; and under the (b) overcast, (d) clear, and (f) partly cloudy sky scenes of the simulated real window

Chapter 3

48

Figure 3.7 displays false colour maps of workplane illuminance values under the three sky scenes, from both the measured prototype and the simulated real window. Comparison of the corresponding illuminance contour maps reveals that the prototype yields a wider illuminance distribution throughout the workplane. This is mainly due to the fact that the light sources of the prototype are placed at a certain distance from the window glass; whereas under the real windows scenes, the sun and sky are at infinity, therefore the light distribution rapidly drops throughout the space. Under the clear sky scene, the real window gives a wider distribution at the left-hand side of the workplane, as shown in Figure 3.7d, since the direct sunlight comes from the upper-right corner of the window. Under the overcast and partly cloudy scenes, the diffuse panel plays a role not only in creating the blur and cloudy display of the window, but also in spreading the generated light onto the back of the room.

Table 3.8 displays the maximum luminance and DGP perceived at positions 1 and 2, from the simulated real window under all sky scenes, as compared to those in the measured prototype. It is interesting to see that even though both scenes have the same illuminance value at the nearest point (P1 in Figure 3.3), the luminance values perceived by the observers greatly differ. Under nearly all sky scenes and observer’s positions, the maximum luminances of the real window are lower, and so are the DGP values, than those of the virtual one.

Table 3.8. Maximum luminance and DGP perceived by the observer at positions 1 and 2, under the three sky scenes in the simulated real window and the measured prototype

Real window VNLS prototype

Position – scene Lmax

[cd/m2]

DGP [-]

Lmax

[cd/m2]

DGP [-]

1 – overcast 601 0.23 1550 0.24

2 – overcast 556 0.21 1540 0.21

1 – clear 3596 0.28 5500 0.32

2 – clear 3523 0.27 3400 0.28

1 – partly cloudy 2587 0.25 6000 0.34

2 – partly cloudy 2111 0.24 4200 0.30

Despite the relatively high luminance a real window can produce, particularly in appearance of the sun, the discomfort glare perception is still relatively low compared to the situation with the prototype. This is also due to the fact that the light sources of the prototype are placed at a certain distance from the window glass, instead of at infinity. Under the real window scenes, the light is scattered in a more diffuse way, therefore the discomfort glare under the real window scenes is less than that under the prototype scenes. It is also noticed that the placement of the PAR spot lamp as a virtual sun at the upper corner of the prototype cannot always represent the real sun’s position at the site location of the test room, particularly for low solar elevation angles; it should be placed at a sufficient distance behind the window glass in order to do so.


49

In general, the measurement and simulation results give an idea of how a VNLS prototype with various sky scenes compares to a real window under a similar sky scene, in terms of physical lighting phenomena. The VNLS prototype analysed here had a limited complexity level of the view. Additional features such as motion parallax and sound transmission could also improve the degree of similarity between the virtual and real windows, even though it may not be directly related to the lighting performance on the workplane.

It can also be argued that while the investigated prototype lacked some features that are usually associated with a real window, this prototype was designed and constructed to create a subjective, rather than accurately measured, impression or feeling of being connected to the outside world, without necessarily reproducing all of the details. For instance, the addition of curtains or Venetian blinds on the window frame makes it less visible, which in some cases can remove the impression that the window is artificial (van Loenen et al., 2007). Compared to other prototypes with a simplified view discussed in Section 2.2.1, this particular prototype scores better in terms of visual appearance, due to the possibility to vary the sky view, colour gradients, and directional light. The future work on this subject will be to investigate how building occupants actually appraise such artificial solutions in reality. Therefore, thorough user’s performance and perception studies are required.


Computational modelling and building performance simulation can help steer the process of VNLS design development. An example of the influence of simulation in VNLS development is shown in this chapter, where Radiance was applied to reproduce the scenes and to evaluate the lighting performance of a first generation VNLS prototype displaying a view of overcast, clear, and partly cloudy skies. Based on the performed measurements, it is observed that for the selected three sky scenes, the overcast scene produced the lowest average horizontal illuminance, while the partly cloudy one produced the highest. Using the designed setting, none of the measuring points received a horizontal illuminance of 500 lx or larger, suggesting the need of a higher intensity setting for each scene, to ensure sufficient amount of light for typical working activities.

The key point of this chapter is to show that simulations can be used to compare an actual VNLS prototype with a hypothetical real window under the same settings, which was not possible physically, since the test room was not located at the building’s façade. Based on the lighting simulation in Radiance, the investigated prototype performed better in terms of light distribution uniformity than a corresponding, hypothetical real window under the overcast and partly cloudy scenes. Under the clear sky scene, the difference between the real and virtual windows is less, due to the influence of direct sunlight.

Further works should be focused on improving the sun mimicking under the clear sky scene. Moreover, the greatest next challenge possibly is to understand how people will

Chapter 3

50

actually appraise VNLS in reality. Therefore, thorough user’s performance and perception studies are required in the future.

51

Chapter 4 Discomfort Glare Evaluation and Simulation of a First Generation Virtual Natural Lighting Solutions Prototype

This chapter provides an evaluation of discomfort glare from a first generation VNLS prototype with complex views and diffuse light, correlated to the results of experiment conducted by Shin et al. (2012) on subjective glare perception from the same prototype. Correlation was made between four normalised glare metrics and the reported glare perception ratings. Radiance and Evalglare were employed to model and recreate the scenes, and to evaluate the glare metrics.

4.1. Introduction

Many researchers have shown the significant role of windows in buildings. Windows are important in controlling the amount of natural light admitted from the exterior environment into the buildings. A proper use of natural light would potentially save considerable amount of energy from artificial lighting use (e.g. Assem & Al-Mumin, 2010; Hammad & Abu-Hijleh, 2010; Yun et al., 2010). Moreover, it has been shown that building occupants feel windows are important due to their preference for having natural light over electric light (e.g. Markus, 1967; Ulrich, 1984; Farley & Veitch, 2001; Hartig et al., 2003; Chang & Chen, 2005; Aries et al., 2010). Several studies have reported beneficial and restorative effects of views on a natural scene (e.g. Kaplan, 1993; Kaplan, 1995; Tennessen & Cimprich, 1995; Berman et al., 2008), whereas views on human-built environments yield effects, which are similar to having no window at all (Kaplan, 1993). Kim & Wineman (2005) showed empirically that views and windows have psychological and economic values. In general terms, the correct application of a daylighting strategy in buildings can increase visual comfort and energy efficiency (Galasiu & Veitch, 2006).

Nonetheless, natural light is known to have a certain limitation, mostly on its availability in time and space. As mentioned earlier in Chapter 1, there are situations in which natural light is absent; for instance during nighttimes, in the inner part of buildings where access to the façade is limited, and in working spaces where having a real daylight opening is not possible due to hygienic or safety reasons.

Another well-known disadvantage of natural light is discomfort glare, which very often reduces the effective use of natural light inside buildings. This leads to discomfort glare being one of the most important topics to address in daylighting research. Many researchers have conducted experiments to study the human response to discomfort glare from windows. However, due to the limited possibility of controlling natural light, some researchers opted to apply prototypes of ‘simulated’ or ‘virtual’ windows to create a controlled daylit scene in their experiments, which mostly include discomfort glare (e.g. Tuaycharoen & Tregenza,

Chapter 4

52

2005; Tuaycharoen & Tregenza, 2007; Kim et al., 2008; Kim & Kim, 2011; Yun et al., 2011; Kim et al., 2012). Such prototypes have a complex view display but generate a mainly diffuse light output, hence are still classified as the first generation VNLS prototypes.

Note that the aforementioned studies were not conducted to measure the difference in perception of views from real and virtual windows. Instead, those studies focused on the variation of the view itself, with regard to subjective discomfort glare perception. As mentioned by Shin et al. (2012), despite the fact that some luminous characteristics of virtual windows can be different from those of real windows, the luminance ratios within the view area can be set equal to the real scenes, to reproduce same luminance conditions to the subjects. Another advantage of using virtual windows is the possibility to repeat exactly the same experiment condition for other subjects, which will be difficult if using real windows.

In the experiment of Shin et al. (2012), the prototype had a dimension of 1.2 m × 1.2 m × 0.25 m, of which 0.9 m × 0.9 m was viewable. Incandescent lamps were installed behind the prototype in rows of 14 × 14 at an interval of 80 mm, which in total could generate a luminance level of approximately 15000 cd/m2. In the experiment, mean view luminance values of 1000, 1800, 3200, 5600, and 10000 cd/m2 were used, which had equal increment factors of approximately 1.8.

The view image was digitally photographed and printed on a transparent film, and was pasted in front of the prototype surface. A total of 10 view images were used in the experiment, which represented some distant or near views, natural or man-made, as well as land or river landscapes. The view images were named as follows: ‘Distant Mixed Land’, ‘Near Mixed Land’, ‘Distant Natural River’, ‘Near Natural River’, ‘Distant Man-made’, ‘Near Man-made’, ‘Distant Natural Land’, ‘Near Natural Land’, ‘Distant Mixed River’, and ‘Near Mixed River’; as displayed in Figure 4.1.

Figure 4.1. View images on the prototype, from (Shin et al., 2012): (a) ‘Distant Mixed Land’, (b) ‘Near Mixed Land’, (c) ‘Distant Natural River’, (d) ‘Near Natural River’, (e) ‘Distant Man-made’, (f) ‘Near Man-made’, (g) ‘Distant Natural Land’, (h) ‘Near Natural Land’, (i) ‘Distant Mixed River’, and (j) ‘Near Mixed River’.

Discomfort Glare Evaluation and Simulation of a First Generation VNLS Prototype

53

The experiment was conducted in a room with dimensions of 6.25 m × 4.56 m × 2.5 m. The prototype was located 0.6 m above the floor, and the subjects were seated facing the prototype at a distance of 1.5 m, as illustrated in Figure 4.2. The subjects were then asked to look at the centre of the view for 5 seconds and to rate the discomfort glare that they sensed on a rating questionnaire, presented in both English and Korean (native) languages. The discomfort glare rating was based on Hopkinson (1972), in which the scale of the discomfort glare was divided into eight stages, using semantic scales proposed by Flynn et al. (1979) as shown in Table 4.1. A total of 10 views with five luminance mean values were shown to each subject. A total of 48 subjects (24 men and 24 women, ages 23 ± 2.4 yr) participated in the experiment.

(a) (b)

Figure 4.2. (a) Floor plan of the test room (taken from Kim et al. (2012)), (b) section plan of setup of the subject and the prototype (taken from Shin et al. (2012))

Table 4.1. Discomfort glare rating used in the experiments of Shin et al. (2012)

Glare perception Rating

Just perceptible 1

Noticeable 1.5

Just acceptable 2

Acceptable 2.5

Just uncomfortable 3

Uncomfortable 3.5

Just intolerable 4

Intolerable 4.5

In Shin et al.’s study, it was not discussed that people generally accept higher luminance values (more glare) from daylight compared to electric lighting. It is uncertain if this also

Chapter 4

54

applies to VNLS, since they are supposed to be a category in between daylight and electric lighting, and therefore the question of the study on this chapter.

Meanwhile, in (day-) lighting simulation, the use of existing glare metrics such as the Daylight Glare Probability (DGP) (Wienold & Christoffersen, 2006), the Daylight Glare Index (DGI) (Hopkinson, 1972), the CIE Glare Index (CGI) (Einhorn, 1979), or the Unified Glare Rating (UGR) (CIE TC-3-13, 1995) is practical, since they are based on objective, physical phenomena leading to glare itself, which is mainly indicated with the size, location, and luminance of the sources and background in the field of view.

In historical order of the development, the four glare metrics are expressed as follows:

DGI = 10 × log

n

i isisb

iposis

LL

L

1 ,5.0

,

8.0,

6.1,

)07.0(

(4.1)

where ωs is the solid angle of the glare source [sr], ωpos is the solid angle of the glare source

modified for its position in the field of view [sr], Ls is the glare source luminance [cd/m2], and

Lb is the background luminance, i.e. the average luminance of areas not indentified as glare sources [cd/m2].

CGI = C1 × log C2

n

i i

isis

id

dP

LEE

E1 2

2,

2,)500/1(

(4.2)

where Ed is the direct vertical illuminance [lx], Ei is the diffuse vertical illuminance [lx], P is

the position index, i.e. a weight factor based on position in a viewing hemisphere, and C1 and

C2 are the weighting coefficients defined as 8 and 2 by Einhorn (1979).

UGR = 8 × log

n

i i

isis

b P

LL 1 2

2,

2,25.0

(4.3)

DGP = 5.87 × 10–5 Ev + 9.18 × 10–5 log 2

n

i iv

isis

PE

L

1 287.1,

2,1

(4.4)

where Ev is the total vertical eye illuminance [lx].

To evaluate the aforementioned glare metrics, Jakubiec & Reinhart (2012) performed simulations of three indoor spaces under 144 clear sky conditions. In order to directly compare the results from the various glare metrics, they proposed a normalisation procedure


55

for DGI, UGR, and CGI results, so that the ranges were between 0 (no likelihood of discomfort) and 1 (100% probability of discomfort). The metrics were multiplied by a certain multiplier, chosen to correlate the intolerable value ranges with those of DGP (Jakubiec & Reinhart, 2012). The normalisation equations read as follows:

DGIn = 0.01452 × DGI (4.5)

UGRn = 0.01607 × UGR (4.6)

CGIn = 0.01607 × CGI (4.7)

where DGIn, UGRn, CGIn are the normalised DGI, UGR, and CGI values, respectively. Next to that, they also proposed the resulting value ranges in which glare was considered to be ‘imperceptible’, ‘perceptible’, ‘disturbing’, and ‘intolerable’ for the corresponding metrics. These ranges are given in Table 4.2.

Table 4.2. Value ranges of DGP, DGI, UGR, and CGI, after Jakubiec & Reinhart (2012)

Glare perception DGP DGI UGR CGI

Imperceptible < 0.30 < 18 < 13 < 13 Perceptible 0.30~0.35 18~24 13~22 13~22 Disturbing 0.35~0.45 24~31 22~28 22~28 Intolerable > 0.45 > 31 > 28 > 28

In their article, Shin et al. (2012) did not report a direct relationship between their glare ratings and any of the four glare metrics. In another article, Kim et al. (2012) briefly discussed correlations between their glare study findings and DGI, but a one-to-one relationship between the glare ratings and DGI itself was not determined. In this thesis, the glare metrics are incorporated to indicate visual discomfort from VNLS, performed in simulation. Therefore, it is intended to have a comparison between the simulation-based glare metrics and the reported subjective glare ratings, so that the results can be used to indicate which glare metrics are the most suitable for VNLS application, and what the effect of various VNLS display views on discomfort glare perception will be.

The objective of this study is to correlate simulated glare metrics from a VNLS prototype to the experimental results of Shin et al. (2012). As a method, Radiance and Evalglare were applied to model the test room and to evaluate the relevant glare metrics, i.e. DGP, DGI, UGR, and CIE from the prototype. Polynomial regression was applied to correlate the glare perception ratings of Shin et al. and the normalised glare metrics of Jakubiec and Reinhart. The polynomial equation was used to convert the ratings of Shin et al. to the four glare metrics. Finally, the values obtained from simulation were compared with the converted values, for each image view.

Chapter 4

56

4.2. Method 4.2.1. Model description

The VNLS prototype was modelled in Radiance, by using the standard ‘light’ material to form a relatively large light emitting area, with a uniform surface luminance. To add the view image on top of the light emitting surface in Radiance, the technique of two-dimensional image mapping was applied. It is basically a technique of pasting a picture on top of a plain surface, for example to display a painting or photograph inside a scene (Jacobs, 2012a). By assigning a ‘light’ material for the plain surface, a luminous display is created.

The test room (see Figure 4.2) was modelled in Radiance, assuming typical values for the interior reflectance, as given in Table 4.3.

Table 4.3. Material reflectance definitions in Radiance for the room’s interior

Material Red Green Blue Specularity Roughness

Window frame 0.600 0.430 0.210 0 0

Ceiling 0.800 0.800 0.800 0 0

Walls 0.700 0.700 0.700 0 0

Floor 0.200 0.200 0.200 0 0

Door 0.600 0.430 0.210 0 0

The relevant two-dimensional image was mapped on a flat, infinitely thin surface constructed with standard ‘light’ material of a very small thickness. The luminance of this thin surface can be obtained with applying the following equation, which converts red, green, and blue radiance components of a surface to its corresponding luminance value (Ward & Shakespeare, 1998).

Ls = 179 LR,G,B = 179 (0.265 LR + 0.670 LG + 0.0648 LB) (4.8)

where Ls is the surface luminance [cd/m2], LR,G,B is the surface radiance [W/(sr·m2)], and LR,

LG, LB are the spectral radiance values [W/(sr·m2)] in red, green, and blue components,

respectively.

In this case, for a certain image and luminance setting, a same value was assigned for the

three spectral radiance values (i.e. LR = LG = LB) to produce only white light behind the pasted image. Different values were assigned to different images and luminance settings, so that the mean window luminance as seen by the observer became equal to 1000, 1800, 3200, 5600, or 10000 cd/m2. Table 4.4 gives the assigned spectral radiance values for each image and luminance setting.


57

Table 4.4. Spectral radiance values assigned for the painting mat at each image and luminance setting

Image Mean window luminance [cd/m2]

1000 1800 3200 5600 10000

Distant Man-Made 9.80 17.64 31.36 54.89 98.01 Near Man-Made 13.30 23.94 42.57 74.49 133.01 Distant Natural Land 10.95 19.72 35.05 61.34 109.54 Near Natural Land 18.02 32.44 57.67 100.92 180.21 Distant Mixed Land 9.16 16.49 29.31 51.29 91.58 Near Mixed Land 14.70 26.46 47.05 82.33 147.02 Distant Natural River 11.89 21.40 38.04 66.56 118.86 Near Natural River 13.97 25.14 44.69 78.21 139.66 Distant Mixed River 13.30 23.94 42.57 74.49 133.01 Near Mixed River 11.64 20.95 37.24 65.18 116.39

Simulations were run individually in Radiance and Evalglare for the 10 view images and five mean luminance values, to calculate DGP, DGI, UGR, and CGI at the observer’s position. The ambient parameters used in Radiance for all variations are set as shown in Table 4.5.

Table 4.5. Radiance ambient parameters used in the simulation of all variations

Parameter Description Value-ab Ambient bounces 4 -aa Ambient accuracy 0.08 -ar Ambient resolution 128 -ad Ambient divisions 1024 -as Ambient super-samples 256

4.2.2. Glare rating correlation

To correlate the glare rating of Shin et al. and the four normalised glare metrics, values with approximately similar glare perception were selected. As a different researcher can use a different terminology for the glare perception, some assumptions were made. Those were as follows:

• The lowest end of each scale is assumed to be zero for all metrics, and is perceived as ‘imperceptible’, i.e. not perceptible at all.

• ‘Just perceptible’ value (1 in Shin et al.’s rating) is assumed equal to the minimum value defined in the original DGP experiment (Wienold & Christoffersen, 2006), which is 0.20. This is approximately 2/3 of the noticeable (perceptible) value.

Chapter 4

58

• ‘Noticeable’ in Shin et al.’s rating is assumed equal in perception to ‘perceptible’ in other metrics.

• ‘Uncomfortable’ in Shin et al.’s rating is assumed equal in perception to ‘disturbing’ in other metrics.

• ‘Just intolerable’ value in all metrics is assumed equal to the mid-point between ‘uncomfortable’ and ‘intolerable’.

Using the aforementioned assumptions, the corresponding values can be listed for all metrics, as shown in Table 4.6.

Table 4.6. Discomfort glare rating used by Shin et al. and the corresponding values of DGP, DGIn, UGRn, and CGIn as correlated to the glare perception

Glare perception Shin DGP DGIn UGR

n CGI

n

Imperceptible 0 0 0 0 0

Just perceptible 1 0.20 0.17 0.14 0.14

Noticeable (perceptible) 1.5 0.30 0.26 0.21 0.21

Uncomfortable (disturbing) 3.5 0.35 0.35 0.35 0.35

Just intolerable 4 0.40 0.40 0.40 0.40

Intolerable 4.5 0.45 0.45 0.45 0.45

Applying the technique of curve fitting with polynomial regression, one can derive an equation correlating the glare metric (DGP, DGIn, UGRn, or CGIn) as a function of Shin et al.’s rating. Knowing this equation, the reported mean glare rating of Shin et al. can be converted to estimate the glare metric, symbolised with DGPconv, DGIn conv, UGRn conv, or CGIn conv. These values were then compared with the simulated values from Evalglare, symbolised with DGPsim, DGIn sim, UGRn sim, or CGIn sim, by putting them in a scatter plot for all 50 variations. The converted and simulated values should be the same in the most ideal case. By examining mean square error between those values, one can observe which of the four glare metrics is the most reliable for describing the glare perception in Shin et al.’s experiment.

4.3. Results and Discussion 4.3.1. Rendering and glare source detection

Two examples of a 180° equiangular view of the rendered model of the test room are shown in Figures 4.3a and 4.3b, displaying the prototype with the ‘Distant Man-Made’ (DMM) and ‘Near Man-made’ (NMM) image scenes at mean luminance value of 3200 cd/m2, observed from the subject’s position. The false colour luminance pictures of both scenes are displayed in Figures 4.3c and 4.3d.


59

(a) (b)

(c) (d)

(e) (f)

Figure 4.3. Impressions of the test room as observed from the subject’s position under the ‘Distant Man-Made’ (DMM) scene and ‘Near Man-made’ (NMM) image scenes. (a) and (b) display the rendered images, (c) and (d) display the false colour luminance pictures, (e) and (f) display the glare source detection pictures of the DMM and NMM scenes respectively. The mean luminance value of the window is 3200 cd/m2 under both scenes.

Chapter 4

60

Evalglare displays detected glare sources in different colours where the rest of the image is set to grey. The colour is randomly chosen by the programme, and does not indicate a specific range of luminance values. The glare source detection pictures of the same scenes are displayed in Figures 4.3e and 4.3f.

From these two examples, it is seen that different view images produce a different glare source, which in turn will result in a different glare perception. In the case of distant view such as the DMM image scene, the ‘sky’ element is mostly the main glare source. The sky element also delivers some light to the ceiling, as seen in Figure 4.3c, even though that part of the ceiling is not considered as a glare source (Figure 4.3e). In the case of near view such as the NMM image scene, a more uniform luminance distribution is observed on the window surface. The white part of the building in the image becomes the glare source.

4.3.2. Unadjusted rating

The results of DGP, DGIn, UGRn, and CGIn under the DMM and NMM image scenes are plotted in Figure 4.4. The mean glare rating of Shin et al. (value range of 1 ~ 4.5) is plotted without any adjustment in a linear scale together with the normalised glare metric (value range of 0 ~ 1). The resulting curve shows a significantly higher gradient, compared to the four metrics. Note that the mean values have a relatively large standard deviation, showed by the error bars. Similar graphs and trends are also found under the other image scenes.

Figure 4.4. Results of DGP, DGIn, UGRn, CGIn, and uncorrected Shin et al.’s rating under the DMM and NMM image scenes. Error bars indicate standard deviation from the mean values of Shin et al.’s rating.

4.3.4. Polynomial regression

The values of glare rating used by Shin et al. and the glare metrics as displayed in Table 4.4 are plotted in Figure 4.5. By applying curve fitting with polynomial regression, the equation relating the metric and the rating can be derived, as also displayed in the charts. For


61

every glare metric, a third-order polynomial equation is obtained, with a coefficient of

determination R2 > 0.99.

Figure 4.5. Graphs showing the selected glare rating of Shin et al. and the corresponding values of DGP, DGIn, UGRn, and CGIn

Based on the obtained polynomial equations, the reported mean glare rating of Shin et al. were converted to yield the glare metrics. These converted values and the simulated values from Evalglare for all 50 variations are plotted in Figure 4.6. By applying curve fitting with linear regression, the equation relating the simulated and converted metric can be derived, as also displayed in the charts.

Chapter 4

62

Figure 4.6. Graphs showing the simulated and converted values of DGP, DGIn, UGRn, and CGIn for all 50 variations

4.3.4. Adjusted rating

In the most ideal case, both simulated (y) and converted (x) values are equal, hence the root mean square error (RMSE) will be zero, and the linear equation will have a gradient of 1

and an offset of 0, i.e. y = x. It is found that DGP has the smallest RMSE (0.04) among the

four metrics, indicating the smallest difference between simulated and converted values. The CGIn linear equation has a gradient of approximately 1, but the RMSE is large (0.29). Moreover, the linear equations of DGIn, UGRn, and CGIn have relatively large offset (0.22 ~ 0.39), while DGP has a relatively small offset (–0.13). This suggests that the simulated values of DGIn, UGRn, and CGIn are all overestimated, compared to converted values from the


63

experiment data of Shin et al. In turn, the simulated values of DGP show a good agreement with the converted ones, even though for the higher range (around 0.40), the simulated values are overestimated.

The high accuracy of DGP compared to other glare metrics was also reported elsewhere, e.g. Kleindienst & Andersen (2009) and Jakubiec & Reinhart (2012), mainly due to the fact that DGP uses vertical illuminance on the observer’s eye as the main input, while this type of illuminance has a strong, positive relationship with the glare metric (Kleindienst & Andersen, 2009). However, all of the earlier findings on DGP were based on real daylight scenes, as obviously implied in the metric’s name. The finding in this chapter gives a new insight on the applicability of DGP for VNLS prototype, together with its correlation to the subjective glare perception.

From the relationship between DGP and Shin et al.’s rating in Figure 4.5, the third-order polynomial regression equation reads as follows:

y = 0.013x3 – 0.107x2 + 0.3193x – 0.038 (4.9)

where y is DGP and x is Shin et al.’s rating. Taking the derivative function of Equation 4.9 yields the following equation:

xy

≈dxdy

= 0.039x2 – 0.214x + 0.3193 (4.10)

where Δy and Δx are the deviations of y and x, respectively. Since the actual standard deviations from Shin et al.’s rating are known from the article (see Figure 4.4), the estimated deviations of the converted values can be obtained.

To give a better illustration, the simulated and converted values of DGP under the DMM and NMM image scenes are plotted in Figure 4.7.

Chapter 4

64

Figure 4.7. Graphs showing the simulated and predicted values of DGP under the DMM and NMM scenes. Error bars indicate the estimated deviation of the converted values.

The resulting curve shows similar values of the simulated and converted DGP,

particularly at the low and medium luminance. It is noticed that the largest differences (approximately 0.1) are found at the 10000 cd/m2 luminance, possibly due to the lower tolerance that Evalglare gives for scenes with a relatively high vertical illuminance on the subject’s eye (Wienold & Christoffersen, 2006). All predicted values have standard deviations of less than 0.05, shown by the error bars. Under the other image scenes, the differences between the simulated and predicted values are also found to be less than 0.1. The graphs are displayed in Figure 4.8.

Figure 4.8. Graphs showing the simulated and predicted values of DGP under all image scenes, excluding DMM and NMM scenes. Error bars indicate the estimated deviation of the predicted values.


65

Figure 4.8. (continued)

Chapter 4

66

4.3.5. Percentage of disturbed subjects

In their article, next to reporting the mean values of the glare rating as a function of mean window view luminance, Shin et al. also reported the distribution of their subjects who voted for a certain glare rating, which corresponded to a certain glare perception. In this study, it has been demonstrated that the mean values of the rating can be converted into normalised glare metrics (i.e. DGP, which is the most accurate one) and compared them with the values from simulation. However, it is unknown whether these values can be correlated with the actual percentage of subjects who felt disturbed from glare in Shin et al.’s experiment.

To investigate this, the percentage of subjects in Shin et al.’s experiment who voted for each corresponding glare perception, under all image scenes, are shown in Table 4.7. From there, the total percentage of the subjects who voted for ‘just uncomfortable’ or worse (i.e. higher discomfort glare rating) was calculated for each mean window view luminance. Similarly, total percentage of the subjects who voted for ‘uncomfortable’ or worse, and those who voted for ‘just intolerable’ or worse, were also calculated. These values represent the actual percentage of disturbed subjects in the experiment, and are drawn in graphs in Figure 4.9.

Table 4.7. Percentage of subjects in Shin et al.’s experiment who voted for each corresponding glare perception, under all image scenes

Rating Glare perception Window luminance [cd/m2]

1000 1800 3200 5600 10000 1 Just perceptible 71.3 24.3 11.7 0.7 0.1

1.5 Noticeable 16.5 35.3 18.2 4.2 0.0 2 Just acceptable 10.3 30.1 32.9 14.5 0.7

2.5 Acceptable 1.8 7.4 25.4 28.3 4.4 3 Just uncomfortable 0.2 4.4 9.8 29.5 20.1

3.5 Uncomfortable 0.0 0.2 1.8 14.5 25.3 4 Just intolerable 0.0 0.2 0.2 8.2 32.7

4.5 Intolerable 0.0 0.0 0.0 0.0 17.7

Total 100 100 100 100 100


67

Figure 4.9. Graphs showing the percentage of subjects in Shin et al.’s experiment who perceived glare equal to or worse than ‘just uncomfortable’, ‘uncomfortable’, or ‘just intolerable’ under all image scenes

Based on Table 4.7, if the ‘just uncomfortable’ category is taken as the lower threshold of the disturbed feeling from discomfort glare, there are approximately 0% subjects who felt disturbed at 1000 cd/m2, 4% at 1800 cd/m2, 12% at 3200 cd/m2, 53% at 5600 cd/m2, and 96% at 10000 cd/m2. If the threshold is raised to ‘uncomfortable’, the figures will be 0%, 0%, 2%, 23%, and 76% at the same luminance values. If ‘just intolerable’ is the threshold, the figures will be 0%, 0%, 0%, 8%, and 51%.

Moreover, when the data are analysed separately for each image scene, a roughly similar trend is observed; two examples are illustrated in Figure 4.10 for the DMM and NMM image scenes.

Figure 4.10. Graphs showing the percentage of subjects in Shin et al.’s experiment who perceived glare equal to or worse than ‘just uncomfortable’, ‘uncomfortable’, or ‘just intolerable’ under the DMM and NMM image scenes

Chapter 4

68

Whichever glare perception rating is chosen as the lower threshold, the curves in Figures 4.9 and 4.10 do not correspond well to the curves of DGP in Figures 4.7 and 4.8. This may be explained by drawing the distribution of subjects in the experiment who voted for each glare perception, under all image scenes, grouped based on the mean window view luminance values as shown in Figure 4.11. From the charts, it is observed that there is a ‘floor effect’ phenomenon on the lowest window value (1000 cd/m2), where more than 70% of the subjects voted for ‘just perceptible’. Such a dominant vote is not found elsewhere. Note that the minimum rating given was 1; if there was a rating of 0 (imperceptible), many would probably have chosen it.

The subjects agreed less at the higher luminance settings; at 1800 cd/m2 most subjects voted for ‘noticeable’ (i.e. 1.5 on the rating), but there were a large difference between the percentage of subjects who voted ‘just acceptable’ (rating of 2) and ‘acceptable’ (rating of 2.5). The votes at 3200 cd/m2 and 5600 cd/m2 were more normally distributed, but were still relatively widespread. At the highest luminance of 10000 cd/m2, the subjects were still divided between perceiving the glare as ‘uncomfortable’ and ‘just intolerable’. A possible explanation can be that the meaning of the words are very close to each other and can be understood differently between the subjects. Another explanation would be that the scale was agreed on, but different people simply perceive glare differently.

Figure 4.11. Graphs showing the distribution of subjects in Shin et al.’s experiment who voted for each glare perception under all image scenes, grouped based on the mean window view luminance

Due to the interpersonal variance in subjective glare perception ratings, it is relatively

hard to predict the actual percentage of the subjects who felt disturbed. A comparison with the experiment of Wienold & Christoffersen (2006) in developing the DGP suggests the use of fewer rating scales. In that experiment, the subjects were asked to rate the glare on a four-point scale, i.e. ‘imperceptible’, ‘noticeable’, ‘disturbing’, and ‘intolerable’. They were also asked to rate the lighting condition as ‘comfortable’ or ‘uncomfortable’, if they had to


69

perform daily work at the test location. The subjects would then have fewer criteria on assessing the glare, which presumably led to a sharper distinction between disturbed and undisturbed feelings from the glare. Furthermore, there were in total 76 subjects (48 men and 28 women, ages ranged from 20 to 59 yr), slightly more than the number of subjects in Shin et al.’s experiment.

While it has been shown that DGP yields the most accurate results compared to other glare metrics, as also found in the study of Jakubiec & Reinhart (2012), it is noticed that DGP also has some limitations. Next to the low availability of user acceptance study at lower luminance ranges (Wienold & Christoffersen, 2006), it is also mentioned that DGP is less accurate for luminance contrast-based glare (Kleindienst & Andersen, 2009). A typical example of this kind of glare can be found in scenes with a generally low ambient illuminance, and with a high contrast point such as bright window glass with a dark frame, which is similar to the situation in the Shin et al.’s experiment. To address this problem, Kleindienst & Andersen (2009) proposed the so-called model-based method to approximate the glare. In a recent study, Suk et al. (2013) proposed the use of absolute and relative glare factors to develop a more practical analysis method. In this method, there is a different analysis for glare from extremely high luminance light sources, and that from glare sources people can still adapt to. Further development on this specific topic of discomfort glare metrics is still expected or ongoing.

Based on the discussed observation, it is concluded that the mathematical models correlating the glare metrics and the glare rating of Shin et al. can accurately predict the reported mean values in that particular experiment. The reported mean values can be converted to DGP, which values are comparable to those obtained from simulation. The simulation values however are not representations of the actual number of subjects who felt disturbed in the experiments, even though the simulation values correspond very well to the mean values, with certain standard deviations, as voted by the subjects.

Furthermore, this simulation study confirms the general finding from the experiment; that at mean view luminance of 3200 cd/m2, glare from the prototype display is perceived as noticeable or perceptible, i.e. corresponds to a DGP value of 0.30. This suggests an upper limit, where any exceeding mean values will only create visual discomfort and inefficient use of electrical energy.

4.4. Concluding Remarks A method is proposed to correlate the commonly used glare metrics, i.e. DGP, DGI,

UGR, and CGI, which can be obtained from simulation, to subjective glare ratings from a first generation VNLS prototype with complex views (Shin et al., 2012). Based on the scatter plots comparing the simulated and converted values, it was found that the linear equations of DGIn, UGRn, and CGIn have a relatively large root mean square error (RMSE), while DGP has a relatively small one. This suggests that the simulated values of DGP are in a better

Chapter 4

70

agreement with the converted ones. Hence, the DGP yields the most accurate results in this particular case. Even though the accuracy of DGP has been widely reported, all of the earlier findings were based on real daylight scenes. The finding in this chapter demonstrates the applicability of DGP for VNLS prototype, and how it correlates to the subjective glare perception.

Neither the simulated values nor the predicted values of the glare metrics can be correlated with the actual percentage of subjects who felt disturbed from glare in the experiment. This is mainly due to different glare perception, particularly at the higher luminance values. There is also an observed ‘floor effect’ at the lowest mean luminance value. Meanwhile, it has been demonstrated that the mean values of the Shin et al.’s rating can be converted into normalised glare metrics and then compared with the values from simulation.

71

Chapter 5 Design, Measurement, and Simulation of a Second Generation Virtual Natural Lighting Solutions Prototype

This chapter discusses the design and evaluation of a second generation VNLS prototype. The design features a more sophisticated direct light component and more energy-efficient light sources. The prototype was realised, and was evaluated by measuring workplane illuminance in a test room. Simulation using Radiance was performed and validated with the measurement results. Various possibilities of placing the prototypes inside the room were investigated in Radiance to determine the effect on space availability and visual comfort. Various operating scenarios were introduced and calculated to determine the effect on the average space availability and total annual electrical energy consumption that are produced and consumed by the prototypes.

5.1. Introduction

In Chapter 3 of this thesis, an existing first generation VNLS prototype as described in van Loenen et al. (2007) has been investigated and modelled, to give an idea on how such a prototype compares to a (simulated) real window. This prototype featured a blurred, diffuse view of sky scenes, i.e. overcast, clear, and partly cloudy, with an additional PAR spot lamp to create the impression of direct sunlight. The partly cloudy sky scene used in this study produced the highest average horizontal illuminance, whereas the overcast sky scene produced the lowest. None of the measuring points received a horizontal illuminance of 500 lx or larger, suggesting the need for a larger area or a higher intensity setting for each scene, particularly the overcast one, to ensure a sufficient amount of light for typical working activities.

While the findings from that particular prototype give an insight in the actual lighting performance, it gives direction to design improvements needed to overcome its current limitations, particularly its relatively small window-to-wall ratio, the missing of ground elements and horizon in the view, and the use of fluorescent light sources instead of more energy-efficient ones. The spot lamp also had limited ability to create a realistic impression of sun patches in the space. A new generation prototype would then be required to observe whether the performance can be improved by adding the missing features, and to validate a computational model that can be extended for further development of future VNLS.

Another issue that has not been addressed so far is the influence of operating scenarios on the lighting performance and energy demand of a certain VNLS prototype. Next to the average space availability, it is intended to predict the amount of electrical energy the entire prototype system will consume on an annual basis, provided the prototype is operated on a routine schedule on each working day.

Chapter 5

72

Therefore, the study in this chapter aims to design and build a second generation VNLS prototype with an improved performance compared to the first one. Three objectives are defined: the first one is to validate the illuminance distribution results obtained from Radiance simulation with the ones obtained from measurement, by evaluating the interior lighting condition inside a test room. The second objective is to determine the effect of various configurations of the prototypes inside the test room on the space availability, uniformity, and visual comfort. Finally, the third objective is to estimate the influence of various operating scenarios on the average space availability and total annual electrical energy consumption that would respectively be produced and consumed when applying such prototypes.

5.2. Design Steps

5.2.1. Test environment

The test environment was built in the new ExperienceLab of Philips Research at the High Tech Campus in Eindhoven, the Netherlands. The dimension of the test room was 6.81 m × 3.63 m × 2.70 m (L×W×H), slightly longer than the standard reference office room (van Dijk & Platzer, 2003). Two original real window openings were placed on a short wall (south) to enable daylight admission. The VNLS prototype was installed on the opposite wall (north).

In order to avoid daylight entering from outside during the measurements, in the entire experiments, the two real windows were blocked with two white covers of the same colour and reflectance as the surrounding wall finishing. Figure 5.1 illustrates the floor plan and section view of the test room.

(a)

Figure 5.1. (a) Floor plan and (b) cross section view of the standard test environment

N

Prototype

Design, Measurement, and Simulation of a Second Generation VNLS Prototype

73

(b)


There were two openings for the prototype; each had a dimension of 0.90 m × 1.20 m

(W×H) excluding the window frames, while the height of the window bottom was 0.93 m from the floor. The distance between the frames of the two openings was 0.14 m. Figure 5.2 shows the prototype openings on the wall’s front section.

Figure 5.2. Front view of the prototype openings

5.2.2. Light sources

The light source of this prototype was an array of light emitting diode (LED) tiles. The decision to use LEDs was taken mainly due to its relatively long life time, high efficacy, high flexibility, and possibility to individually control and to display multiple colours. Moreover, LED technology has a great potential for saving energy consumption in buildings (Jenkins & Newborough, 2007; Pandharipande & Caicedo, 2011). In this study, a total of eight Philips Origami BPG762 luminaires were incorporated to provide the light as well as to construct the view. An individual Origami BPG762 is actually a combination of four Origami BPG732 tiles. Each luminaire houses 108 LUXEON RGB power LEDs, and can display a virtually limitless range of dynamic changing colours, given in red, green, and blue (RGB)

Chapter 5

74

components. Because it is designed to be a uniform edge-free lighting tile, it can be combined to form an array and perform as a ‘pixel’ unit.

An individual Origami BPG762 has a maximum power consumption of 128 W and is driven by a 24 V DC source. Each tile can be independently controlled by a Digital Address-able Lighting Interface (DALI) or Digital Multiplex (DMX) protocol with a proper programming. The dimension of an individual Origami BPG762 tile is shown in Figure 5.3.

Figure 5.3. Dimensions of Origami BPG762 tile

Inside each individual Origami BPG732, 27 LUXEON power LEDs are distributed. These LEDs are covered by multiple layers of diffusers with different properties to ensure a uniform light output distribution. Red, green and blue channels can be controlled independently.

To fit the size of the prototype window openings, a total of eight (2 rows × 4 columns) Origami BPG762 tiles were applied. Since each individual Origami BPG762 is actually a combination of four Origami BPG732 tiles, the entire prototype consisted of 32 pixels in total. Each Origami BPG762 tile could be controlled by 12 channels. The conceptual structure of the display lighting panel together with window opening is illustrated in Figure 5.4.

Figure 5.4. Rear view illustration of the 2 × 4 Origami BPG762 array


75

Beside the main display elements, additional direct light sources were also required to realise another feature of a (real) window, which is providing direct sunlight into the indoor space. For this purpose, the use of the LED tiles alone was not sufficient. To select the direct light sources, some requirements were taken into account, i.e.: • The shape should be linear • The emitted light should be directional • The array should be independently dimmable • The colour temperature should be controllable • The beam angle of the light source should be wide (100° ~ 110°)

The Philips iW Cove MX Powercore (wide beam version) module was selected for the purpose. It is a high-performance, white light LED fixture with adjustable colour temperatures. It has independent channels of warm, neutral, and cool white LEDs to produce colour temperatures in a range from 2700 K to 6500 K. Moreover, an individual iW Cove is only 0.30 m long (see Figure 5.5) and can be easily connected in a line with other lamps of the same type to reach the total length needed.

Figure 5.5. Dimensions of an individual iW Cove

5.2.3. Control circuit

In this study, the DMX512 protocol was used to realise the control function. DMX512 is a standard for digital communication networks that is commonly used to control lighting dimmers. Under this protocol, a digital dimming lighting network can be set up by using DMX512 controllers and control software. Each colour channel can be dimmed from the full RGB value of 255 to 0. Compared to the DALI protocol, the advantage of using DMX512 is that each can control 512 channels at the same time, whereas each DALI controller can only control 63 channels. DMX also has a high data rate, as needed for dynamic views. Therefore the DMX512 is a better choice for the control circuit in this case.

Chapter 5

76

The DMX512 modules are already integrated with the Origami and iW Cove arrays. The Origami accepts the so-called regular DMX ESTA signals, whereas iW Cove uses DMX Kinetics signals. To control both types of lamps, a DMX Splitter was applied to split the signal from the computer, convert it to the respective DMX formats, and transfer the signal to the luminaires. By connecting them with a DMX controller, the basic network was formed. Commands could be sent by the computer to control the light emitting level of each ‘pixel’ or display unit.

The Origami array was powered by 24 V DC supplies, provided by rectifiers. The iW Cove array, connected in series, was powered by a Philips Data Enabler Pro, providing integrated DMX signals. The conceptual control circuit structure illustration is shown in Figure 5.6, in which gray lines symbolise the signal circuit and black lines symbolise the power circuit.

Figure 5.6. Control and power supply circuits

ORIGAMI ORIGAMI

iW Cove

iW Cove

iW Cove

iW Cove

iW Cove

iW Cove

iW Cove

iW Cove

ORIGAMI ORIGAMI

ORIGAMI ORIGAMI

ORIGAMI ORIGAMI

Rectifier

Rectifier

Rectifier

Rectifier

DMX CONTROLLER

DMX SPLITTER

COMPUTER

DATA ENABLER PRO

SIGNAL:

POWER:


77

5.2.4. Display and structure

The prototype was designed to provide a blurred suggestion of a view, rather than a high resolution view. A single diffusive panel was mounted in front of the Origami array, replacing the original separate outer diffusers, to eliminate the visibility of the tile boundaries. To achieve that, the diffusive panel should also have a high transmittance, and be placed in a certain distance from the Origami array. A 5 mm thick PLEXIGLASS Satinice colourless diffuser (type 0F00 SC) was selected for this purpose. It had a one-sided matte surface and was made of frosted surface material, which means that its thickness does not significantly influence the amount of light loss.

The layers of Origami array – diffuser – window glass formed the basic structure of the prototype display, as shown in Figure 5.7.

Figure 5.7. Basic three-layer structure of the prototype display

In this prototype, all of the Origami tiles were fixed to a MDF back plate holding the 2.40 m × 1.20 m array. During system operation, heat is generated by the LEDs, as discussed for example by Ahn et al. (2014). Therefore, openings were made in the back plate (see Figure 5.8), allowing cooling by natural convection.

Figure 5.8. Front view of the MDF frame

Chapter 5

78

The prototype appearance from outside and inside the test room (at ‘off’ and ‘on’ conditions) are displayed in Figure 5.9.

(a) (b)

(c)

Figure 5.9. The prototype appearance from (a) outside and (b) inside the test room at ‘off’ condition, (c) inside the test room at ‘on’ condition

The current prototype could display a diffuse, 32 pixel view. Therefore, any given view image should be converted or filtered into 32 pixels, and should comply with some general rules as follows: • The aspect ratio is 2 : 1 (width : height), as is the cased for the prototype source. • The horizon is at eye height, which is 1.20 m for sitting viewers. Using the defined space

configuration and given the display is four pixels high, the ‘ground’ element should be displayed only at the lowest row.

• The ‘sky’ element is composed of blue sky and white clouds. To increase the white light output, the clouds should preferably take more than half of the sky area.

The general steps to simulate a view on the VNLS prototype are described in Table 5.1.


79

Table 5.1. General steps to simulate a view on the prototype

No Step View example 1 Select a view image with an aspect ratio of 2 : 1,

adjust the height of the horizon into approximately 0.25 of the total view height.

2 Use image processing software to apply a mosaic filter to turn the image into 32 pixels. Use colour picker to obtain the RGB values of each pixel, as input for the DMX.

3 Apply filter to simulate the diffuser and glass plate in front of the view image.

4 Apply the direct light source.

5 Add the window frame.

5.2.5. Programming and setting

To define the colour display settings, each pixel was given a code name and was assigned with a DMX value. For the Origami array, each pixel was given a code name of A, B, C, or D, followed by a number. Code name ‘A’ referred to the lowest row, while ‘D’ was the highest row. The number was named 1 to 8, indicating the position from left to right, seen from inside the test room. Each pixel had three channels corresponding to the red, green, and blue (R, G, and B) components.

Chapter 5

80

For the iW Cove array, the codes started with ‘P’ followed by number from 1 to 8. Each pixel had three channels corresponding to the colour temperature of warm (2700 K), neutral (4000 K), and cool (6500 K).

Referring to the view image shown in Table 5.1, the DMX values addressed for each pixel of the display at the maximum setting (i.e. the highest luminance) are given in Table 5.2. The values range from 0 (entirely off) to 255 (100% on), where RGB values of [255, 255, 255] correspond to a full-white colour display.

Table 5.2. DMX values addressed for each pixel on the display at the maximum setting

Position D1 D2 D3 D4 D5 D6 D7 D8

Red 20 220 255 26 25 220 220 25 Green 255 255 230 233 255 255 255 248 Blue 230 255 255 255 255 255 255 227

Position C1 C2 C3 C4 C5 C6 C7 C8

Red 230 220 230 220 220 230 200 220 Green 255 255 255 255 255 255 255 255 Blue 255 255 255 255 255 255 255 255

Position B1 B2 B3 B4 B5 B6 B7 B8 Red 255 255 225 205 200 180 230 200 Green 255 255 255 255 255 255 255 255 Blue 255 255 255 255 255 255 255 255

Position A1 A2 A3 A4 A5 A6 A7 A8 Red 0 0 0 0 0 0 0 0 Green 200 216 221 215 227 224 225 229

Blue 100 10 10 10 10 10 10 10

For further analysis, any daily profile scenario can be defined by proportionally scaling the DMX values of the maximum setting. Three daily profile scenarios were defined, depending on the climate type, to represent ‘summer’, ‘spring’, and ‘winter’ conditions, which are discussed later in detail in Section 5.6.

5.3. Measurement Protocol To evaluate the actual lighting performance of the prototype, a number of data were

collected at three settings, i.e. 100% (see Table 5.2), 62.5% (representing a medium setting), and 25% (representing a low setting) of the maximum. The DMX values were proportionally scaled, while the display showed the view image in Table 5.1. The collected data were as follows: • Horizontal illuminance on the workplane; data were collected for 91 points (see Figure

5.10) on the workplane height, i.e. 0.75 m from the floor.


81

• Vertical illuminance on the observer’s eye plane; data were collected for four positions (namely 1 until 4, see Figure 5.10) on the typical observer height when seated (1.20 m from the floor).

• Minimum, maximum, and average luminance perceived by the observer at the same four vertical points on the typical observer height (1.20 m from the floor).

• Reflectance of interior surface materials; data were collected for the relevant interior surface, such as floor, walls, ceiling, and furniture.

Figure 5.10. Horizontal and vertical illuminance measurement points in the test room

Furthermore, the horizontal illuminance data were post-processed to obtain the average

illuminance values (Eav [lx]), uniformity (U0), and space availability (%A [%]), referring to Equations 3.6 until 3.8 in Chapter 3.

Vertical illuminance and luminance perceived by the observer were measured by taking 20 photographs (ISO 400, f/5.6, shutter time varied from 4 s to 1/8000 s) each at positions 1 until 4, at a height of 1.20 m, with the view direction specified by the arrows in Figure 5.10.

To determine glare index, i.e. the Daylight Glare Probability (DGP) (Wienold & Christoffersen, 2006), at the observer’s position, the obtained photographs were exported to Radiance, combined into HDR images using the Hdrgen programme, and then analysed using Evalglare (Wienold & Christoffersen, 2006).

During the measurement, the following instruments were used: • SpectraDuo PR-680 photometer; for measuring luminance values as well as spectral power

distribution. • Lutron LX-1118 light meter; for measuring illuminance values on the workplane.

1

2

3

4Prototype

Chapter 5

82

• Canon EOS50D digital single-lens reflex camera + Sigma 5.5mm fisheye lens + Photolux 3.1 software; for taking multiple (20 in this case) photographs in equiangular 180° view with various exposure values, which in turn were post-processed to obtain the luminance pictures. The luminance values were calibrated with the SpectraDuo photometer.

• Konica Minolta CM-2600D spectrophotometer; for measuring reflectance values of the interior surface materials.

5.4. Simulation Protocol

5.4.1. Model description

The first objective of this chapter is to validate the illuminance distribution results obtained from simulation using Radiance with the ones obtained from measurement, since the computational model is to be extended for development of future (not-yet existing) VNLS. Therefore, the actual conditions under the three lighting scenes were also modelled and simulated, to give an insight in the difference between simulation and actual measurement. Comparison was made between the simulated and measured values of horizontal illuminance at the middle row in Figure 5.10, as well as between the simulated and measured values of the space availability and uniformity.

The front, top, and perspective views of the modelled prototype are displayed in Figure 5.11. The 32 Origami tiles were modelled as boxes of 0.30 m × 0.30 m × 0.05 m (L×W×D), constructed with a ‘light’ material. The eight iW Cove lamps were modelled as a continuous row of eight cylinders, with a length of 0.30 m each and a diameter of 0.016 m, also constructed with a ‘light’ material. Each lamp had various red, green, and blue radiance components [W/m2/sr], depending on the row position and the sky scene. The assigned values for each component under the maximum setting in Radiance are defined in Table 5.3, which is fine-tuned proportionally to the actual DMX values (Table 5.2). For other settings, the assigned values are proportionally scaled.

(a) (b) (c)

Figure 5.11. (a) Front, (b) top, and (c) perspective views of the modelled prototype


83

Table 5.3. Red, green, and blue irradiance components of ‘light’ material defined in Radiance for each pixel on the display at the maximum setting

Position D1 D2 D3 D4 D5 D6 D7 D8

Red 4.70 51.8 60.0 6.12 5.88 51.8 51.8 5.88

Green 60.0 60.0 55.1 46.6 54.8 60.0 60.0 58.3

Blue 55.1 60.0 60.0 60.0 60.0 60.0 60.0 53.4

Position C1 C2 C3 C4 C5 C6 C7 C8

Red 55.1 51.8 55.1 51.8 51.8 55.1 47.1 51.8

Green 60.0 60.0 60.0 60.0 60.0 60.0 60.0 60.0

Blue 60.0 60.0 60.0 60.0 60.0 60.0 60.0 60.0

Position B1 B2 B3 B4 B5 B6 B7 B8

Red 60.0 60.0 60.0 48.2 47.1 42.3 55.1 47.1

Green 60.0 60.0 60.0 60.0 60.0 60.0 60.0 60.0

Blue 60.0 60.0 60.0 60.0 60.0 60.0 60.0 60.0

Position A1 A2 A3 A4 A5 A6 A7 A8

Red 0 0 0 0 0 0 0 0

Green 47.1 50.8 52.0 50.8 53.4 52.7 52.9 53.9

Blue 23.5 2.35 2.35 2.35 2.35 2.35 2.35 2.35

The detailed values assigned for the window construction and the room’s interior surfaces reflectance as obtained from the measurement are specified in Table 5.4.

Table 5.4. Material definitions in Radiance for the window construction and room’s interior

Material Red Green Blue Specula-

rityRough-

nessDiffuse

transmiss.Transmit. specularity

Diffuse panel 0.25 0.25 0.25 0 0 0.55 0

Window glass 0.90 0.90 0.90 - - - -

Window frame 0.79 0.79 0.79 0 0 - -

Ceiling 0.91 0.91 0.91 0 0 - -

Walls 0.79 0.79 0.79 0 0 - -

Floor 0.13 0.08 0.03 0 0 - -

Door 0.79 0.79 0.79 0 0 - -

5.4.2. Validation

The first objective of this chapter is to validate the simulation of workplane illuminance with the measurement results. Simulations were run for the three settings, i.e. 100%, 62.5%, and 25%; by addressing the input defined in Table 5.3. Calculation was performed for the 91 measuring points on the workplane, referring to Figure 5.10. One-to-one comparison between measurement and simulation was done for all values of horizontal illuminance at the middle

Chapter 5

84

row where point 4 was located. This row, at which there were 13 measuring points, was located directly in the central projection of the windows. Furthermore, simulation parameters in Radiance were set as previously shown in Table 3.4 in Chapter 3.

As previously discussed in Section 3.4 in Chapter 3, the European Standard EN 12464-1 (CEN, 2002) mentions that “a factor of approximately 1.5 represents the smallest significant difference in subjective effect of illuminance”, as given in their recommended scale of illuminance. This criterion is applied to evaluate the simulation results, in which the ratio of simulation and measurement values at any measuring point should not be less than 2 : 3 (or approximately 0.67) and not more than 3 : 2 (or 1.50).

5.5. Analysis of Various Configurations

The second objective of this chapter is to determine the effect of various configurations of the prototype inside the test room, on the space availability, uniformity, and visual comfort. In order to maximise the light distribution on the workplane, two prototypes were modelled inside the test room, and were placed either on the short walls or on the long walls. Seven configurations were introduced, named Configuration 1 until 7, which floor plans are illustrated in Figure 5.12. Note that in most configurations, the prototype was split into two equal parts; each consisted of 4 × 4 tiles. Each opening had a dimension of 0.90 m × 1.20 m (W×H), and the height of the window bottom in all configurations was 0.93 m from the floor, that is the same as in the tested configuration (Figure 5.2).

(a) (b)

(c) (d)

Figure 5.12. Floor plan of the test room with the prototypes in configurations (a) 1, (b) 2, (c) 3, (d) 4, (e) 5, (f) 6, and (g) 7


85

(e) (f)

(g)


Using the properties in Tables 5.2 and 5.3, simulations in Radiance and Evalglare were run to obtain the space availability, uniformity, and DGP at the defined four observer’s positions. In Section 5.3, the test room was assumed to be used for typical office work that requires workplane illuminance of 500 lx, but in practice it can also be used for other purposes, for instance waiting rooms in healthcare facilities, that require a lower workplane illuminance. Therefore, for this analysis, the space availability was evaluated not only for the criterion of 500 lx, but also for 300 and 200 lx.

5.6. Analysis of Various Operating Scenarios

5.6.1. Settings and data collection

The third objective of this chapter is to get an overview of the space availability and electrical energy consumption of a VNLS prototype under various operating scenarios in a given year. To estimate the space availability, the workplane illuminance measurement data of the prototype were revisited. With the same reason as in Section 5.5, to maximise the light distribution on the workplane, it was assumed that there were two prototypes inside the test room, placed facing each other on the short walls; i.e. according to Configuration 2 in Figure 5.12. Linear regression was performed for each point to estimate the illuminance values under any given DMX setting.

To estimate the energy consumption, the real-time power consumption of the same prototype was measured, as a function of the corresponding DMX values, i.e. 0%, 10%, 20%, 40%, 50%, 62.5%, 80%, 90%, and 100%. The measurement was conducted under the defined

Chapter 5

86

displayed setting, i.e. the view image was according to the values in Table 5.2 and was proportionally scaled. During the measurement, the ELV EM800 energy monitor was used for measuring the real-time power and current of the entire system.

5.6.2. Daily profiles and annual modes

Assuming an operating schedule from 09.00 to 18.00 hrs local time on every working day, it is intended to predict the amount of average space availability and total electrical energy the two prototypes will produce and consume in a given year. To obtain this, certain schedules based on daily profiles and annual modes should be defined beforehand.

Given the maximum DMX value settings in Table 5.2, three daily profiles were defined to represent ‘spring’, ‘summer’, and ‘winter’ conditions. In this study, the daily profiles were obtained by calculating vertical illuminance on a point facing south in an exterior field, on 21 March (spring), 21 June (summer), and 21 December (winter); at every hour from 09.00 to 18.00 hrs local time, using Gensky programme of Radiance. All values were normalised to the maximum value among those three days, so that they could be presented in a scale of 0 ~ 1.

In addition, five locations were selected to represent various Köppen-Geiger climate types (Peel et al., 2007). The representative cities for each selected climate type were Singapore (tropical rainforest), Cairo (arid, hot desert), Amsterdam (temperate, oceanic), Sevilla (temperate, Mediterranean), and Chicago (cold, continental). For each climate type, a sky model for Gensky input was selected by considering the typical weather data provided by the Department of Energy of the United States (US Department of Energy, 2011). Table 5.5 gives the description of those five locations.

Table 5.5. Five climate types selected for total electrical energy consumption estimation

Climate type

General description Typical

city Latitude, longitude

Sky model

Af Tropical, rainforest Singapore 1.37°N, 103.97°E Standard CIE overcast

(–c)

Bwh Arid, hot desert Cairo 30.12°N, 31.38°E Partly cloudy with sun

(+i)

Cfb Temperate, warm summer/

oceanic Amsterdam 52.28°N, 4.77°E

Partly cloudy without sun (–i)

Csa Temperate, hot summer/

Mediterranean Sevilla 38.42°N, 5.90°E

Partly cloudy with sun (+i)

Dfb Cold, warm summer/

continental Chicago 41.97°N, 87.92°W

Partly cloudy with sun (+i)


87

(a) (b)

(c) (d)

(e)

Figure 5.13. Daily profile settings as defined for (a) Singapore, (b) Cairo, (c) Amsterdam, (d) Sevilla, and (e) Chicago

Chapter 5

88

(a) (b)

(c) (d)

(e)

Figure 5.14. Monthly average solar radiation of (a) Singapore, (b) Cairo, (c) Amsterdam, (d) Sevilla, and (e) Chicago, based on the weather data (US Department of Energy, 2011)


89

As a direct consequence of using various climate types, the daily profiles for the selected locations are different from each other, see Figure 5.13. It should be also noted that the profiles were based on vertical illuminance at each hour on the three selected days, which makes it possible that the values on ‘spring’ and ‘winter’ are larger than those on ‘summer’, because the sun’s altitude is actually lower on spring and winter days, hence more direct sunlight.

For estimating the total electrical energy consumption in a year, two annual operating modes were defined. These modes were based on the order of assigning the daily pattern, either following the real seasons, or the opposite. In detail, they are as follows:

• Mimicking mode: the assigned daily pattern of the prototypes follows the real seasons, i.e. the prototype displays the ‘summer’ setting during summer months, and displays the ‘winter’ setting during winter months.

• Compensating mode: the assigned daily pattern of the prototypes follows the opposite of the real seasons, i.e. the prototype displays the ‘winter’ setting during summer months, and displays the ‘summer’ setting during winter months.

In both modes, ‘spring’ is considered representing also the autumn months, therefore they are interchangeable. To simplify the calculation, it is assumed that every day in a given month has exactly the same daily profile. For instance, if June is assumed to be a summer month, then the prototypes will display the ‘summer’ setting for the entire month of June under the mimicking mode, or the ‘winter’ setting for the entire month of June under the compensating mode.

Note that in the calculation, both prototypes are assumed to display exactly the same setting at all time, i.e. there is no difference between light output and view from both of them. The terms mimicking and compensating only refer to the difference between the prototypes display and the real seasons, not to the difference between the two prototypes display.

The choice of which months to be considered as ‘spring’, ‘summer’, and ‘winter’ was made by considering the variation in monthly average global horizontal solar radiation of the particular location (US Department of Energy, 2011), as illustrated in Figure 5.14. For example, the months with the highest solar radiation in Cairo (i.e. May, June, and July) can be regarded as summer months; however, one can also argue that April, August, and September should be included in the ‘summer’, given the relatively high solar radiation in those months as compared to the other climates. To test the sensitivity of the results for such choices, three annual scenarios were defined for each location. Under the mimicking mode, Scenario 1 (Sc1) has the shortest summer and the longest winter months, Scenario 2 (Sc2) has roughly equal length of summer and winter, and Scenario 3 (Sc3) has the longest summer and the shortest winter months.

Table 5.6 gives the selected scenarios for the five climate types, where ‘Sp’, ‘Su’, and ‘Wi’ stand for spring, summer, and winter, respectively.

Chapter 5

90

Table 5.6. Annual profile scenarios for each selected climate type under mimicking and compensating modes

Location Singapore Cairo Mode Mimicking Compensating Mimicking Compensating Month Sc1 Sc2 Sc3 Sc1 Sc2 Sc3 Sc1 Sc2 Sc3 Sc1 Sc2 Sc3

Jan Sp Sp Sp Sp Sp Sp Wi Wi Wi Su Su Su Feb Su Su Su Wi Wi Wi Sp Wi Sp Sp Su Sp Mar Su Su Su Wi Wi Wi Sp Sp Sp Sp Sp Sp Apr Su Su Su Wi Wi Wi Sp Sp Su Sp Sp Wi May Sp Su Su Sp Wi Wi Su Su Su Wi Wi Wi Jun Sp Sp Sp Sp Sp Sp Su Su Su Wi Wi Wi Jul Sp Sp Su Sp Sp Wi Su Su Su Wi Wi Wi

Aug Sp Sp Sp Sp Sp Sp Sp Su Su Sp Wi Wi Sep Sp Sp Sp Sp Sp Sp Sp Sp Su Sp Sp Wi Oct Sp Sp Sp Sp Sp Sp Sp Sp Sp Sp Sp Sp Nov Wi Wi Wi Su Su Su Wi Wi Wi Su Su Su Dec Wi Wi Wi Su Su Su Wi Wi Wi Su Su Su

Location Amsterdam Sevilla Mode Mimicking Compensating Mimicking Compensating Month Sc1 Sc2 Sc3 Sc1 Sc2 Sc3 Sc1 Sc2 Sc3 Sc1 Sc2 Sc3

Jan Wi Wi Wi Su Su Su Wi Wi Wi Su Su Su Feb Wi Sp Sp Su Sp Sp Sp Wi Sp Sp Su Sp Mar Sp Sp Sp Sp Sp Sp Sp Sp Sp Sp Sp Sp Apr Sp Sp Sp Sp Sp Sp Sp Sp Sp Sp Sp Sp May Su Su Su Wi Wi Wi Su Su Su Wi Wi Wi Jun Su Su Su Wi Wi Wi Su Su Su Wi Wi Wi Jul Su Su Su Wi Wi Wi Su Su Su Wi Wi Wi

Aug Sp Sp Su Sp Sp Wi Sp Su Su Sp Wi Wi Sep Sp Sp Sp Sp Sp Sp Sp Sp Sp Sp Sp Sp Oct Sp Sp Sp Sp Sp Sp Sp Sp Sp Sp Sp Sp Nov Wi Wi Wi Su Su Su Wi Wi Wi Su Su Su Dec Wi Wi Wi Su Su Su Wi Wi Wi Su Su Su


91

Table 5.6. (continued)

Location Chicago Mode Mimicking Compensating Month Sc1 Sc2 Sc3 Sc1 Sc2 Sc3

Jan Wi Wi Wi Su Su Su Feb Sp Wi Sp Sp Su Sp Mar Sp Sp Sp Sp Sp Sp Apr Sp Sp Sp Sp Sp Sp May Su Su Su Wi Wi Wi Jun Su Su Su Wi Wi Wi Jul Su Su Su Wi Wi Wi

Aug Sp Su Su Sp Wi Wi Sep Sp Sp Sp Sp Sp Sp Oct Sp Sp Sp Sp Sp Sp Nov Wi Wi Wi Su Su Su Dec Wi Wi Wi Su Su Su


Results of the measurement and simulation of the actual test room and prototype are given in Sections 5.7.1 and 5.7.2. Results of analyses of various configurations and operating scenarios are respectively given in Sections 5.7.3 and 5.7.4.

5.7.1. Measurement of actual test room

Measurement results of the average workplane illuminance, uniformity, and space availability (%A [%]) at 25%, 62.5%, and 100% of the maximum setting in the test room are summarised in Table 5.7. It is found that under the maximum setting of the defined configuration, the average workplane illuminance was 239 lx and 11% of the workplane met the target illuminance of 500 lx. This suggests an addition of local or task lighting on the workspace is still required to satisfy the criterion for typical working activity (CEN, 2002). Under the lower settings, the space availability becomes zero; i.e. no spots on the workplane satisfy the illuminance criterion. The uniformity is about the same under all settings, within the range of 0.30 ~ 0.37.

Table 5.7. Measurement results of the average illuminance (Eav), uniformity (U0), and space availability (%A) at 25%, 62.5%, and 100% of the maximum setting in the test room

Setting Eav [lx] U0 [-] %A [%]

25% 49 0.37 0

62.5% 136 0.33 0

100% 239 0.30 11

Chapter 5

92

Figure 5.15 displays false colour maps of horizontal illuminance values on the workplane under the three settings.

(a) (b)

(c)

Figure 5.15. False colour maps of the measured workplane illuminance [lx] under (a) 25%, (b) 62.5%, and (c) 100% of the maximum setting

The measurement results show that at the nearest point to the prototype, the workplane illuminance value is 800 lx under the maximum setting, with space availability (taking 500 lx as the minimum criterion) of around 11%. Under 62.5% and 25% settings, none of the points receives illuminance larger than 500 lx, which is required for typical office work. However, when the test room is used for other activities which require lower illuminance criteria, the space availability will be higher. For instance, taking 300 lx as the criterion, approximately 30% space availability can be achieved under the maximum setting, as observed from Figure 5.15c. The light distribution is symmetrical along the prototype’s central axis.

Vertical illuminance on the observer’s eye plane (Ev [lx]), together with minimum (Lmin

[cd/m2]), maximum (Lmax [cd/m2]), and average luminance (Lav [cd/m2]) perceived by the

observer at all positions (referring to Figure 5.10) are displayed in Table 5.8. In addition, the DGP values obtained from Evalglare are also given.


93

Table 5.8. Vertical illuminance on the observer’s eye, minimum, maximum, average luminance, and DGP perceived by the observer at all positions at 25%, 62.5%, and 100% of the maximum setting

Position Ev [lx] Lmin

[cd/m2] Lmax

[cd/m2] Lav

[cd/m2] DGP

[-]

1 – at 25% 124 1.09 57000 60 0.13*

2 – at 25% 37 0.69 3060 18 0.01*

3 – at 25% 55 0.44 480 13 0.03*

4 – at 25% 57 0.50 540 13 0.02*

1 – at 62.5% 316 2.44 143000 155 0.21

2 – at 62.5% 94 1.88 17500 49 0.10*

3 – at 62.5% 144 1.13 1740 34 0.14*

4 – at 62.5% 148 1.48 1310 34 0.15*

1 – at 100% 521 2.94 320000 255 0.26

2 – at 100% 177 1.57 36000 89 0.14*

3 – at 100% 233 1.80 2270 55 0.21

4 – at 100% 239 2.58 2130 55 0.21

*values are lower than 0.20, i.e. the minimum defined for DGP

(a) (b)

(c) (d)

Figure 5.16. HDR view of the test room with the prototype under the maximum (100%) setting, observed at position (a) 1, (b) 2, (c) 3, and (d) 4

Chapter 5

94

Figure 5.16 displays combined high dynamic range (HDR) images of the test room under the maximum (100%) setting, observed from position 1 until 4. Figure 5.17 displays the luminance false colour pictures of Figure 5.16, generated by Photolux 3.1 software.

(a) (b)

(c) (d)

Figure 5.17. Luminance false colour pictures of the prototype under the maximum (100%) setting, observed at position (a) 1, (b) 2, (c) 3, and (d) 4

From Figures 5.16a and 5.16b, one can observe visible patches of light on the two side walls. These patches are in fact the intended result from installing the direct lamps arrays. The effect can be varied or extended if necessary, for example by controlling the light output of each lamp, or adding Venetian blinds to create structured patches on the walls.

Based on the measurement data in Table 5.8 and view images in Figures 5.16 and 5.17, it is revealed that position 1 experienced the worst glare perception, due to the large source area observed. The luminance of the observed window area was in the range of 1000 ~ 5000 cd/m2 (see Figure 5.17a). The maximum average luminance was 255 cd/m2; this is approximately two times larger than the maximum average luminance measured from the first generation prototype in Chapter 3 (see Table 3.6).

Based on the criteria of Jakubiec & Reinhart (2012), a DGP range of 0.30 ~ 0.35 corresponds to a ‘perceptible’ category, while DGP values of < 0.30 are considered


95

‘imperceptible’; therefore the largest DGP in this case (i.e. 0.26) was still classified as imperceptible. At the other positions and settings, the perceived glare was below the minimum defined value of DGP (i.e. 0.20).

5.7.2. Simulation of actual test room

Table 5.9 summarises the simulation results of the horizontal illuminance point at the

middle row on the workplane, together with the average illuminance values (Eav [lx]),

uniformity (U0), and space availability (%A [%]) under the three settings. For comparison, the measurement results and the relative differences are also shown.

Table 5.9. Simulation (sim.) and measurement (meas.) results of horizontal illuminance point at the middle row, together with the average illuminance values (Eav [lx]), uniformity (U0), and space availability (%A [%]) under the three settings

25% 62.5% 100%

Distance to window [m]

Sim. [lx]

Meas. [lx]

Ratio [-]

Sim. [lx]

Meas. [lx]

Ratio [-]

Sim. [lx]

Meas. [lx]

Ratio [-]

0.5 171 165 1.03 430 459 0.94 807 858 0.94

1.0 147 137 1.07 365 382 0.95 684 677 1.01

1.5 106 97 1.10 273 272 1.01 511 480 1.07

2.0 78 71 1.10 198 203 0.98 364 349 1.04

2.5 58 52 1.12 147 155 0.95 270 254 1.06

3.0 43 41 1.07 108 118 0.91 208 204 1.02

3.4 36 36 1.02 93 94 0.98 172 163 1.06

3.9 28 29 0.97 69 79 0.88 127 135 0.94

5.4 22 25 0.88 56 66 0.84 105 111 0.94

4.9 19 22 0.86 46 57 0.81 89 98 0.90

5.3 17 20 0.85 41 51 0.81 74 88 0.84

5.8 15 20 0.74 38 49 0.77 68 81 0.84

6.3 14 19 0.72 36 47 0.76 65 78 0.83

Eav [lx] 50 49 1.01 124 136 0.91 236 239 0.99

U0 [-] 0.26 0.37 0.69 0.26 0.33 0.79 0.27 0.30 0.88

%A [%] 0 0 n/a 0 0 n/a 12 11 1.10

Chapter 5

96

The lighting simulation and measurement results of the prototype generally show a good agreement, with a maximum relative difference of 28% at the farthest point under the 25% setting, possibly dominated by measurement accuracy limits, since the absolute difference is only 5 lx. However, ratio of the simulated value to the measured one (or measured to simulated, whichever is greater) at all points is actually always less than 1.5, which represents the smallest significant difference in subjective effect of illuminance (CEN, 2002). Therefore, the computational model is considered sufficient for the purpose of reproducing the scenes without a significant subjective difference, and can be further extended for non-existing solutions.

Figure 5.18 displays the graphs showing the relationship between horizontal illuminance and the distance to the windows under the three defined settings, based on the measurement and simulation.

(a) (b)

(c) Figure 5.18. Graphs showing the relationship between horizontal illuminance and distance to windows under the (a) 25%, (b) 62.5%, and (c) 100% of the maximum setting


97

5.7.3. Comparison of various configurations

The relationship between space availability, uniformity, and maximum DGP in the test room under the simulated configurations (Figure 5.12) is illustrated in Figure 5.19. It is observed that when 200 lx is taken as the criterion for workplane illuminance, nearly all configurations yield a space availability of 100% or very close to it, except Configuration 1 in which all of the four openings are placed on a short wall. Consequently, the far side of the room is left without sufficient light. When 300 lx is taken as the criterion, only Configurations 2 (two openings on each short wall facing each other) and 5 (four openings on a long wall) yield space availabilities of more than 90%. When 500 lx is taken as the criterion, all configurations yield space availabilities of less than 50%, the highest being Configuration 3 (two openings on each long wall facing each other, 0.14 m distance between openings on the same wall).

Figure 5.19. Graphs showing the relationship between space availability, uniformity, and maximum DGP in the test room, under the simulated configurations scenes

The highest uniformity is achieved under Configuration 2 (0.59), and second to that are Configurations 5 and 7 (0.55). The maximum DGP values under all configurations range are very similar, within the range of 0.25 ~ 0.30, mostly found at the observer’s positions that are the closest to the openings, or those that are able to view the entire four openings. The highest value is found at position 1 under Configuration 1, since the position is located near the wall

Chapter 5

98

where both of the prototypes are placed. Under Configuration 2, the observer at the same position would experience the second highest DGP value, since each prototype is placed on the short walls, one of which is at a distance of 1.0 m to the observer’s position.

Figure 5.20 displays false colour maps of horizontal illuminance values on the workplane, under Configurations 2, 5, 3, and 7. The false colour maps of those under Configurations 1, 4, and 6 are displayed in Appendix E.

(a) (b)

(c) (d)

Figure 5.20. False colour maps of the simulated horizontal illuminance [lx] under Configurations (a) 2, (b) 5, (c) 3, and (d) 7

The simulation results show that at the nearest points to the openings, the workplane illuminance values are approximately 800 lx under Configuration 2 (Figure 5.20a), whereas the values are between 600 ~ 650 lx under Configuration 5 (Figure 5.20b). The distance between two adjacent openings is 0.14 m in Configuration 2, and 0.80 m in Configuration 5. In the latter, the individual prototype (4 × 8 tiles) is split into two parts; each consisted of 4 × 4 tiles. Therefore, the maximum illuminance in Configuration 5 is less than that in Configuration 2, since the light is spread more evenly and is not concentrated as much as in Configuration 2. As shown in Figure 5.19, both configurations have a space availability of 100% for 200 lx criterion, and approximately 90% for 300 lx criterion. For 500 lx criterion, the satisfying workplane area in Configuration 5 is slightly larger than that in Configuration 2.


99

Configuration 3 (Figure 5.20c) yields the largest space availability for 500 lx criterion, but not the largest for 300 lx criterion. Under this configuration, the prototypes are placed in the centre of the long walls; therefore the light is more concentrated in the middle part of the room, and drops towards the edges. Configuration 7 (Figure 5.20d) is close in terms of performance to Configuration 5, only the light is less concentrated at the nearest points to the openings, and there is less workplane area with illuminance of higher than 300 lx.

Under Configuration 5, the highest DGP value (0.26) is observed at position 2 (see Figure 5.12), which directly faces four openings; whereas under Configuration 2, the highest DGP value (0.29) is observed at position 1, as mentioned earlier. This finding leads to a suggestion of placing workstations and viewing directions that will give the least discomfort glare perception. For instance, under Configuration 2, the viewing direction at position 2 is recommended; whereas position 3 or 4 is recommended under Configuration 5.

5.7.4. Comparison of various operating scenarios

5.7.4.1. Space availability

Relationship between the normalised DMX settings and the space availability produced by two prototypes (four openings) in Configuration 2 is illustrated in Figure 5.21. It is observed that the space availability is highly sensitive to the choice of illuminance criterion. Note that the graphs are stepwise due to the discrete number of points used to calculate the space availability. The higher the illuminance criterion, the fewer points receive illuminance above that level, and the higher the minimum setting required to reach non-zero space availability.

Figure 5.21. Space availability and normalised DMX setting relationship, using the criteria of 200, 300, and 500 lx

Using the daily profile in Figure 5.13, one can estimate the space availability at each hour. In turn, the daily and annual average values can also be estimated for each defined climate type. Table 5.10 gives the results for each climate type and daily profile setting.

Chapter 5

100

Table 5.10. Estimated daily average space availability [%] in Configuration 2 for each climate type and daily profile, i.e. ‘spring’ (Sp), ‘summer’ (Su), and ‘winter’ (Wi)

Space availability – 500 lx

[%] Space availability – 300 lx

[%] Space availability – 200 lx

[%] Location Sp Su Wi Sp Su Wi Sp Su Wi

Singapore 17 13 13 61 51 49 88 86 83

Cairo 11 0 15 40 3 52 80 17 73

Amsterdam 15 15 0 53 52 2 79 95 8

Sevilla 18 0 13 63 15 47 88 39 70

Chicago 18 6 3 61 31 17 82 82 45

Table 5.11 gives the average annual space availability for each climate type, annual mode, and scenario. The results suggest that variation within the same climate type is very small. The difference between mimicking and compensating modes is insignificant, since the annual space availability is averaged out of the values in a year. As discussed earlier, the values are indeed sensitive to the choice of required illuminance. The values in Singapore are found to be the highest, while those in Cairo are the lowest. Note that the profiles are based on exterior vertical illuminance on the defined day and sky condition. Consequently, the daily values for ‘spring’ can be larger than those for ‘summer’ in non-tropical locations, since the sun is in a lower altitude at the spring time, resulting in a larger vertical illuminance.

Table 5.11. Estimated average annual space availability [%] in Configuration 2 for each climate type, annual mode, and scenario

Space availability – 500 lx [%] Space availability – 300 lx [%]

Location Mimicking Compensating Mimicking Compensating Sc1 Sc2 Sc3 Sc1 Sc2 Sc3 Sc1 Sc2 Sc3 Sc1 Sc2 Sc3

Singapore 15 15 15 15 15 15 57 56 55 57 56 54 Cairo 9 9 7 9 9 10 33 31 25 34 32 37

Amsterdam 10 11 11 11 11 10 36 40 40 39 39 35 Sevilla 12 10 10 12 10 12 46 41 42 47 42 45

Chicago 11 9 10 11 9 10 42 36 39 41 36 38

Space availability – 200 lx [%]

Location Mimicking Compensating Sc1 Sc2 Sc3 Sc1 Sc2 Sc3

Singapore 87 87 86 87 86 86 Cairo 61 56 47 62 57 61

Amsterdam 60 66 67 65 64 59 Sevilla 70 65 66 71 66 69

Chicago 73 70 73 72 69 69


101

5.7.4.2. Electrical energy consumption

The relationship between the normalised DMX settings and the real-time power of the actual, individual prototype is given in Figure 5.22. Linear regression curves are drawn based on the relationship. One can observe a highly linear relation (R2 > 0.99) between the real-time power and the normalised DMX settings. Moreover, the linear regression equation can be applied to predict the real-time power of the system, which can be expressed as follows:

Wreal = 528.12x + 117.46 (5.1)

where Wreal [W] is the real-time power consumed by the individual prototype, and x is the normalised DMX setting that ranges from 0 to 1. Equation 5.1 can be applied in particular for estimating the total electrical energy consumption of the system in a day, given a certain daily profile as illustrated in Figure 5.13.

Figure 5.22. Relationship between total real-time power and normalised DMX setting of the actual prototype

Given the obtained Equation 5.1 and assuming that a working day consists of nine bins of DMX value, i.e. 09.00 ~ 10.00 hrs, 10.00 ~ 11.00 hrs, and so forth until 17.00 ~ 18.00 hrs, then the total daily electrical energy consumption for two prototypes can be estimated. The daily profile settings in Figure 5.13 are reflected in the estimated total daily electrical energy consumption. In the tropical climate region such as Singapore, where the difference between seasons is minimal, the estimated total daily energy consumption is relatively similar within the three designated days. Since the daily profiles were determined based on the simulated vertical illuminance on an exterior point, the values for ‘spring’ are larger than those for ‘summer’ in the non-tropical locations, due to the lower sun altitude at the spring time. The values for ‘winter’ are generally the smallest because the day is shorter, except for the cases of Cairo and Sevilla, where the total daily energy consumption in ‘winter’ is larger than that in ‘summer’.

Chapter 5

102

Assuming that every month has either 20 or 25 working days (i.e. 4 or 5 working weeks), then the monthly electrical energy consumption can be estimated for each location, annual mode, and scenario. Summing up the monthly electrical energy consumption gives the total annual value [kWh] for 260 working days in a year. Table 5.12 shows the results for two prototypes (Configurations 1 to 7), indicating that in a given climate type, the total annual electrical energy consumption is relatively similar for the two annual modes and three scenarios. Applying the compensating mode instead of the mimicking mode essentially means switching the ‘summer’ and ‘winter’ months, which does not significantly influence the annual sum. Differences between scenarios 1, 2, and 3 are also small, since most of these scenarios were created by adding or removing only one or two ‘summer’ (and consequently, ‘winter’) months, based on the actual profile of monthly average solar radiation in the relevant location.

Table 5.12. Estimated total daily and annual electrical energy consumption [kWh] of two prototypes for each climate type, annual mode, and scenario

Total daily energy consumption [kWh]

Total annual energy consumption [kWh]

Location Spring Summer Winter Mimicking Compensating Sc1 Sc2 Sc3 Sc1 Sc2 Sc3

Singapore 9.6 9.1 8.9 2380 2367 2354 2434 2417 2400 Cairo 8.2 5.2 8.5 1945 1891 1767 1961 1907 1977

Amsterdam 9.1 9.3 4.1 1952 2051 2054 2029 2026 1926 Sevilla 9.7 6.5 8.4 2218 2127 2153 2228 2136 2201

Chicago 9.4 8.1 6.0 2131 2037 2104 2121 2027 2054

If real windows are also present in the same room, then under the mimicking mode, one can expect to experience similar weather conditions shown by the real and virtual windows, but the interior ambient lighting level may be amplified too much, hence a concern on the energy use. Under the compensating mode, the ambient lighting level is more stable, but one may experience a contradicting view and perception between the real and virtual windows. Nevertheless, the total annual electrical energy consumption under both annual modes is found very similar within all selected locations, as shown by the results. The value for all climate types is approximately 2000 kWh, the highest being 2434 kWh in Singapore and the lowest being 1767 kWh in Cairo.

It should be noticed that the settings of daily profile for all selected climate types are normalised to the same maximum DMX setting. This means at the maximum setting, for example, the prototype set for Cairo would give the same luminance as the one set for Amsterdam. Meanwhile, based on the simulation and weather data, the peak vertical illuminance and horizontal solar radiation values in both locations would differ by a ratio of around 1.6. However, it is unknown whether the difference in vertical illuminance and/or solar radiation has a one-to-one relationship with the expected maximum luminance of the


103

prototype. Given the nature of human perception and preference in visual comfort, most probably there is a certain upper limit for a ‘tolerable’ window luminance (Kim et al., 2012; Shin et al., 2012), where any values exceeding that limit would only create discomfort and energy spill. With this consideration, next to the technical limitation of the applied luminaires, setting the same maximum value in the daily profile for all selected climate types is considered appropriate.

A direct consequence of setting the same maximum value for all climate types is that the one with the least variation between seasons, hence the most consistent daily profile values, requires the most energy. Figure 5.23 illustrates the average of total annual electrical energy consumption in the selected climate types, together with their upper and lower limits, which are taken from the maximum and minimum values when the same daily profile is applied the entire year. To present the results more generally, all values are normalised to the maximum value that can be achieved in the defined settings, i.e. the total electrical energy consumed by

the prototypes when they constantly display the maximum setting (x = 1) in Table 5.2 at each working hour and on each working day in a year.

Figure 5.23. Estimated normalised total annual electrical energy consumption with the upper and lower limits in the five selected climate types

It is observed from Figure 5.23, that within the selected climate types, the annual profile

of Amsterdam yields the most variation in electrical energy consumption, whereas that of Singapore yields the least. The highest value is achieved when displaying ‘spring’ in Sevilla, even though the highest on average is found for Singapore. The lowest is for ‘winter’ in Amsterdam. Displaying ‘summer’ in Cairo for the entire year gives the lowest annual electrical energy consumption, while displaying ‘winter’ in the same location yields a relatively high value, due to the sun’s appearance and position at that time. For Sevilla and Chicago, the highest value is achieved when displaying ‘spring’ for the entire year. Note that the values in Figure 5.23 are normalised, therefore it is still applicable for other light sources with a different efficacy.

Chapter 5

104

Even though the presented results have given a general idea on the influence of operating scenarios on the performance of VNLS, it is noticed that the findings are bounded to the setup of the particular prototype. There are in fact many more factors that may contribute to the resulting space availability, for instance, the windows configuration and placement, the source beam angle, and the display resolution. Nevertheless, the obvious advantage of applying VNLS in a space is the possibility to put the displays on any wall or ceiling surface, independent of where the actual façade is located. In this manner, a high uniformity within the workplane can reasonably be achieved, whereas the same thing is practically impossible in a sidelit room having only real windows on one façade.


A second generation VNLS prototype has been designed and built by installing an array of 32 LED tiles and a line of LED linear fixtures with adjustable colour temperatures to provide direct light into the test room. This particular prototype has an important role in validating the computational model that can be extended for further development of not-yet-existing VNLS. It is found that under the tested maximum (100%) setting with a view, the average workplane illuminance of the test room was 239 lx and 11% of its total workplane area met the target illuminance of 500 lx. Patches of direct light on the side walls could be created as an intended result from the installation of the LED line.

Simulation and measurement values of horizontal illuminance at certain distances from the prototype were compared under three settings. The maximum relative difference is 28%, found at the farthest point under the tested minimum (25%) setting. The ratio of the simulated value to the measured one at all points is always between 0.67 and 1.50, ensuring no significant difference in subjective effect of illuminance. Therefore, the computational model is considered sufficient for the purpose of reproducing the scene without creating a significant subjective difference.

Based on the comparison of seven configurations of two prototypes with equal total opening size in the test room, it is found that nearly all configurations, except Configuration 1 (four openings on a short wall), yield a space availability of 100%, taking 200 lx as the criterion. When 300 lx is taken as the criterion, Configurations 2 (two openings on each short wall facing each other) and 5 (four openings on a long wall) yield space availabilities of more than 90%. When 500 lx is taken as the criterion, the configurations yield space availabilities between 25% and 50%. The highest uniformity is achieved under Configurations 2 (0.59) and 5 and 7 (0.55). The maximum DGP values under all configurations range between 0.25 and 0.30.

Based on the comparison of various operating scenarios with regards to annual space availability and total electrical energy consumption, calculated for a room with two proto-types in Configuration 2, it is found that:

• Variation of average annual space availability within a given climate type is found to be very small; the highest being under the annual profile of Singapore, whereas the lowest


105

being under that of Cairo. The values are however sensitive to the chosen criterion of workplane illuminance.

• Based on the designated daily profiles and annual modes, the normalised, total annual electrical energy consumption in all climate types is on average within the range of 0.63 ~ 0.79, relative to the total electrical energy consumed by the prototype when it constantly displays the maximum setting. There is no significant difference found between applying the mimicking and compensating modes. Within the selected climate types, the annual profile of Amsterdam yields the most variation in electrical energy consumption, whereas Singapore yields the least.

This prototype could be further improved in terms of display resolution and luminous efficacy. By applying the latest developed LED technology, it is possible in the future to display more detailed image views with higher light output and/or lower energy use. Another feature not yet tested in this prototype is view dynamics. This specific feature can be applied by further programming, so that the displayed view is constantly changing.

Chapter 5

106

107

Chapter 6 Modelling and Simulation of Virtual Natural Lighting Solutions with a Simplified View and Directional Light

This chapter discusses the modelling and simulation of future (not-yet-existing) VNLS with a simplified view and directional light, using Radiance to predict the lighting performance. The model features an array of small light sources with various direct light components, resembling a simplified view of the sky and ground. Four input variables, i.e. distance between windows, interval of tilt angle, beam angle, and total luminous flux, were varied to observe their effect on the lighting performance of a reference office space. The VNLS models were also compared to corresponding real windows under the standard CIE overcast sky.

6.1. Introduction

The concept of VNLS is relatively new and the real, ideal product does not exist yet. The currently available prototypes can not meet all the expectations yet; they are only able to meet part of the natural light expectation (Mangkuto et al., 2011, 2012). The previous chapters of this thesis have demonstrated the role of using modelling and simulation of existing prototypes; Chapter 3 shows comparisons of a first generation prototype (van Loenen et al., 2007) to a hypothetical real window of the same sky scenes, Chapter 4 shows comparisons of subjective discomfort glare perception from a first generation prototype of Shin et al. (2012) to glare metrics obtained by simulation, while Chapter 5 shows the possibility to recreate the scenes of a second generation prototype in simulation, of which the model was validated with the measurement results.

In this chapter, development of not-yet-existing next generation VNLS is discussed, by introducing a new VNLS model, investigating various possible input variables, and evaluating their lighting performance. The VNLS model in this case is an improvement of the ‘second generation’ prototype as discussed in Chapter 4. The model was created in Radiance, in the form of arrays of small directional light sources, providing a ‘simplified view’, resembling the sky, the horizon, and the ground. The light directionality was optimised by varying the tilt angle of the light sources in different rows.

In particular, the objectives of this study are twofold. The first objective is to understand the effect of changing input variables of the VNLS, which in this case are: the window’s configuration, tilt angles of the sources, beam angle of the sources, and total luminous flux of the sources; on the lighting performance of a reference office space. The second objective is to compare the lighting performance of the simulated VNLS in a reference office space, relative to that of real windows under the standard CIE overcast sky.

Chapter 6

108

Real windows VNLS

The lighting performance is described in terms of the ability to meet the space availability demand, the illuminance uniformity on the workplane, the illuminance contribution from the ground elements on the ceiling, and the ability to produce minimal glare at predefined observer positions in the given space. The space availability is defined here as the percentage of workplane (at a height of 0.75 m from the floor) meeting a certain minimum illuminance criteria. VNLS ideally provide space availability comparable to or better than real windows with the same configuration.

6.2. Methods

6.2.1. Modelling

While all detailed characteristics of the view from a window are considered very important for developing the requirement of VNLS, the focus in this study is only on modelling the characteristics of direct and diffuse light from the sky and reflected light from the exterior ground. One of the reasons that people distinguish the difference between a real and an existing VNLS prototype with a view, either static or dynamic, may be because the directionality of the light coming out of the surfaces is different. In general, most of the existing prototypes behave like a diffuse light source, without the possibility of seeing the impression of direct and reflected light components on the interior surfaces.

Therefore, in this chapter, a model of VNLS in the form of an array of small light emitting areas is proposed, displaying a simplified view of green ground, (horizon) and blue sky, as illustrated in Figure 6.1. The third layer (distant objects such as built landscape) was not yet represented. The bottom array acted as the ‘ground’ which was tilted upward to mimic reflected light, directed to the ceiling. The rest of the light sources acted as the ‘sky’, and were tilted downward to direct the light to the workplane area.

Figure 6.1. Schematic overview of the VNLS as used in the simulation. The light sources are constructed in arrays, such that the light from the ‘ground’ is delivered to the ceiling and light from the ‘sky’ is delivered to the floor

Modelling and Simulation of VNLS with a Simplified View and Directional Light

109

The light emitting areas were modelled to fit two individual vertical windows, each with the size of 0.80 m × 1.17 m (W×H), which corresponds to a window-to-wall ratio of 20% of a short wall with the size of 3.6 m × 2.7 m. Each light emitting area in each individual window had a size of 0.05 m × 0.05 m and resembled a blue sky. In the lowest row, there were four light emitting areas (0.20 m × 0.20 m each) resembling a green ground surface.

In order to model the directionality of the entering light, the ‘sky’ sources were tilted downward with a certain interval of tilt angle. Three variations of this interval were introduced, i.e. 1.0°, 1.5°, and 2.0°. The sources in the row directly above the ‘ground’ were never tilted (i.e. 0°), while the sources in the second row above the ‘ground’ were tilted downward by 1.0°, or 1.5°, or 2.0°. The sources in the third row above the ‘ground’ were tilted downward by 2.0°, or 3.0°, or 4.0°, and so on. As a result of using the defined window height, the sources in the top row were tilted downward by 20°, 30°, and 40°. The ‘ground’ sources were tilted at a 40° angle pointing upward. Figure 6.2 displays the views of an individual VNLS with an interval of tilt angle of 2.0°.

Figure 6.2. Front and side views of the individual VNLS with an interval of tilt angle of 2.0°

The sources had a certain beam angle, i.e. the angle between the two directions at which the luminous intensity is half that of the maximum luminous intensity. To see the effects of varying beam angle, three values of beam angle for the ‘sky’ were introduced, i.e. 38° (relatively narrow spread), 76° (medium), and 114° (wide).

The luminous intensity distribution of each light source was written in an IES format file, based on the character of downlights with a certain beam angle. The distributions in every row had similar patterns. The luminous intensity values for the sources with a 114°

Chapter 6

110

beam angle were set in such a way that the combination of these sources gives an average surface luminance (L [cd/m2]) of either 1000 cd/m2 (low luminance setting), 1800 cd/m2

(medium luminance setting), or 3200 cd/m2 (high luminance setting). These are the first three values used in the experiments with an ‘emulated window’ by Shin et al. (2012).

The luminous intensity values for the ‘sky’ sources with 38° and 76° beam angles were adjusted accordingly, so that the total luminous flux coming from the ‘sky’ sources altogether remained the same. The technique for calculating the total luminous flux from the source was based on the zonal cavity method described by Lindsey (1997). Given the luminous intensity values at various angles of a luminaire, and assuming that the luminous intensity distribution is direct (no values for angles more than 90°) and symmetrical around the luminaire’s axis, the area surrounding the luminaire can be divided into nine zonal cavities, which are the volumes of conic solid angles with a width of 10°, starting from 0° ~ 10° up to 80° ~ 90°.

The total luminous flux produced by the luminaire (Φ [lm]) can be determined as follows:

Φ = ))cos(cos2( maxmin

9

1 NNNNI

(6.1)

The minimum and maximum angles in each zonal cavity (θminN, θmaxN [°]) determine the

zonal constant which is multiplied by the average luminous intensity (IN [cd]) to yield the luminous flux of that particular zonal cavity. The total luminous flux is then the sum of luminous flux of all zonal cavities.

For this case, the calculated total luminous fluxes of all ‘sky’ sources (two windows) were approximately 6200 lm, 11100 lm, and 19900 lm, respectively for the low, medium, and high luminance settings. Figure 6.3 displays the nine evaluated luminous intensity distributions of the ‘sky’ sources. Note that the luminous intensity values of the polar diagrams changed with varying beam angle.

Each ‘ground’ source had a maximum luminous intensity of 110 cd at the low luminance setting, 199 cd at the medium one, and 354 cd at the high one; all had a similar pattern of luminous intensity distribution. The beam angle of the ‘ground’ source remained constant at 76° for all variations. A tilt angle of 40° upward was chosen so that the ‘ground’ sources do not stand completely vertical, which could possibly create too much glare; and that they were not tilted too much, which could reduce the visibility of the ‘ground’ itself.


111

(a) (b)

(c) (d)

(e) (f)

Figure 6.3. Polar diagram of luminous intensity (in candela) of the light sources resembling the sky, with beam angles of 38°, 76°, 114° and total luminous flux of 6200, 11100, and 19900 lm

Chapter 6

112

(g) (h)

(i)


The variation in all input variables is summarised in Table 6.1. In total, there were 54 different combinations possible based on the input variables.

Table 6.1. Input variables and their values

Variable Symbol Unit Values

Distance between windows d m 0, 0.75

Interval of tilt angle IA deg 1.0, 1.5, 2.0

Beam angle BA deg 38, 76, 114

Total luminous flux Φ lm 6200, 11100, 19900


113

6.2.2. Settings

The space discussed in this study was a reference office space with dimensions of 5.4 m × 3.6 m × 2.7 m (L×W×H). There were two vertical window configurations chosen from the earlier studies of Diepens et al. (2000) and Lawrence Berkeley National Laboratory (LBL) (2010), see Figure 6.4. Each VNLS was modelled with a simplified view image on its surface, which is explained in the section 6.2.1. No real windows were present together with the VNLS in the modelled spaces.

(a) (b)

Figure 6.4. Elevation views of the VNLS window configurations on the wall

In the given space, VNLS were put on one side of the wall (W 3.6 m × H 2.7 m). Frames of 5 cm wide were defined at the perimeters of the windows. Reflectance values of the room’s interior were: ceiling: 85%, walls: 50%, floor: 20%, door: 50%, window and door frames: 50%; all of which based on the IEA Task 27 reference office (van Dijk & Platzer, 2003).

Three different observer positions, namely A, B, and C, were defined at the eye height of 1.2 m above the floor. Position A was near the window and viewing parallel to the window plane, B was in the middle of the room and viewing parallel to the window plane opposite to the viewing position A, while C was near the rear wall and directly facing the window plane, as shown in Figure 6.5.

Chapter 6

114

Figure 6.4. (a) Plan view and (b) section view of the simulated space

For all simulations, the simulation parameters in Radiance simulations were set as shown in Table 6.2.

Table 6.2. Radiance simulation parameters

Parameter Description Value

-ab Ambient bounces 4

-aa Ambient accuracy 0.08

-ar Ambient resolution 128

-ad Ambient divisions 1024

-as Ambient super-samples 256

-ds Direct sub-sampling 0.2

6.2.3. Assessment

6.2.3.1. Performance indicators

The assessment for this study is based on the performance indicators of interest, which are:

- Space availability (%A): The percentage of workplane area (at a height of 0.75 m, with a size equal to the total floor area) with illuminance ≥ 500 lx (typical criterion for office work). Calculation was performed at 1944 (= 54 × 36) points which were evenly distributed

on the workplane. The %A is the percentage of the number of points with illuminance ≥ 500

lx (n(E ≥ 500 lx)), compared to the total number of points (N).

%A = N

En )lx500( × 100% (6.2)


115

- Uniformity (U0): The ratio between the minimum illuminance (Emin [lx]) to the average (Eav [lx]); based on the defined calculation points.

U0 =

avEEmin (6.3)

- Ground contribution (%G): The percentage ratio of illuminance contribution from the

‘ground’ element sources (Eground [lx]) to the total illuminance (Etotal [lx]) received at a certain point on the ceiling, with the surface normal facing downward (z- axis). Calculation

was performed for N = 10 points on the ceiling, located as displayed in Figure 6.6. The

average value is reported as %Gav.

%G = total

groundE

E× 100% (6.4)

%Gav = N

GN

ii

1%

(6.5)

Figure 6.5. (a) Plan view and (b) section view of calculation points for ground contribution

- Probability of discomfort glare: The normalised values of all potentially relevant glare indices, i.e. Daylight Glare Probability (DGP), Daylight Glare Index (DGI), Unified Glare Rating (UGR), and CIE Glare Index (CGI) were calculated with the Evalglare programme (Wienold & Christoffersen, 2006). Those four indices were taken into account since to the best of the author’s knowledge, very little is known about which glare indices are most suited for the case of not-yet-existing VNLS models. Another often-used index, the Visual Comfort Probability (VCP) was not considered, since it was specially developed for

Chapter 6

116

typically sized, ceiling-mounted, artificial lighting installations with uniform luminances (Harrold, 2003). Because the calculated glare indices have different ranges of values, it is intended to normalise the values for the purpose of comparison. The normalisation factors were suggested by Jakubiec and Reinhart to determine the ‘probability of discomfort glare’, by multiplying the DGI value with 0.01452, and multiplying UGR and CGI values with 0.01607 (Jakubiec & Reinhart, 2012). No normalisation is required for DGP, since it is already defined within the range of 0 ~ 1. Thus, the relationships can be expressed as follows:

DGIn = 0.01452 × DGI (6.6)

UGRn = 0.01607 × UGR (6.7)

CGIn = 0.01607 × CGI (6.8)

where DGIn, UGRn, CGIn are the normalised DGI, UGR, and CGI values, respectively. Next, the four normalised glare indices are averaged, and then the value is reported as the

average probability of discomfort glare (PDGav).

PDGav = (DGP + DGIn + UGRn + CGIn) / 4 (6.9)

6.2.3.2. Sensitivity analysis

To evaluate the effect of the current input variables on the defined performance indicators, a sensitivity analysis using multiple linear regressions was performed. This

regression model assumes a linear relationship between the output variable yi and the p-vector

of input variables xi. This relationship is modelled through an error variable εi, which is an unobserved random variable that adds noise to the linear relationship between the output and input variables. The mathematical model takes the general form as follows:

yi = iiiii xxxx 44332211 , i = 1, 2, …, n (6.10)

where βi is a p-dimensional regression coefficient. In this case: p = 4, n = 2 × 3 × 3 × 3 = 54,

x1 is the distance between windows (d, in metres), x2 is the interval of tilt angle (IA, in

degrees), x3 is the beam angle (BA, in degrees), and x4 is the total luminous flux (Φ, in

lumens); while y is evaluated for %A, U0, %Gav, and PDGav, individually. Since the variables have different units, it is intended to standardise the values for each of the output and input variables, that is:


117

'iy =

y

i yy

; 'nix =nx

nni xx

(6.11)

where 'iy and 'nix are the standardised output and input variables, iy and nix are the actual

output and input variables, y and nx are the arithmetic mean of the output and input

variables, and y and nx are the standard deviation of the output and input variables. The

standardised values were then put in the regression model, which can be expressed in matrix form as follows:

'...

'1

ny

y =

''''

''''

............''''

4

3

2

1

4321

14131211

nnnn xxxx

xxxx+

'...

'1

n

, n = 54 (6.12)

The built equations were then solved using a MATLAB toolbox in order to deter-mine '1 , '2 , '3 , and '4 ; which are the standard regression coefficients that determine the

sensitivity of the output as function of the input. The standard regression coefficients range between 1 (strong positive correlation) and –1 (strong negative correlation). The interaction between input variables was not investigated.

6.2.3.3. Comparison with real windows

As a means of comparison, the VNLS in all configurations were replaced with real windows (double clear glass 6 mm, transmittance 88.5%) under a CIE overcast sky condition. The comparison with real windows is considered necessary, since the general concept of this VNLS type is to increase the possibility of seeing the impression of direct light components from the sky and reflected light components from the ground, on the interior surfaces. This impression often appears in a space with real windows, but typically not in a space with a conventional electrical lighting installation. Since the typical general lighting installation is ceiling-mounted, a fair comparison with wall-mounted VNLS will be difficult to achieve. Moreover, the display of a simplified view of blue sky and green ground is also an important feature of the proposed VNLS model, which should also be compared with a relatively simple view of overcast sky and plain ground outside the real windows.

The sky condition of the real windows scenes was defined in the Gensky programme in Radiance, by inserting the zenith radiance value [W/(sr·m2)]. This value was chosen so that the interior surface of the window will give approximately the same average luminance as the corresponding VNLS. It should be noted that VNLS with the same total luminous flux can have a different average surface luminance, particularly when the beam angles are different. Therefore, each VNLS scene was compared only to the real window scene with

Chapter 6

118

approximately the same average surface luminance. The assessments for determining the performance indicators were then performed for the real window scenes.

Since the three observer positions in the room were located in such a way that they were facing different directions, one can presume that position C, which was directly facing the window, may experience the most severe glare amongst the three positions. Therefore, the glare assessment may be reduced to focus only on position C, as it will be sufficient to reflect the worst situation. To demonstrate this, the glare assessment for positions A, B, and C were performed on the following sample of variations:

• VNLS: configuration 1a (d = 0), IA = 2.0°, BA = 76°, Φ = 11100 lm, Lav = 3200 cd/m2

• VNLS: configuration 2a (d = 0.75 m), IA = 2.0°, BA = 76°, Φ = 11100 lm, Lav = 3200 cd/m2

• Real window: configuration 1a (d = 0), Lav = 3200 cd/m2

• Real window: configuration 2a (d = 0.75 m), Lav = 3200 cd/m2

To evaluate the performance of all VNLS variations, four performance criteria were applied on the relative comparison between performance indicators of the VNLS and the real windows with the same average surface luminance. These were based on the expected benefit of having VNLS, i.e. gaining more well-lit and uniform space; while maintaining the ground contribution on the ceiling and the probability of discomfort glare comparable to those in real windows scenes. The criteria are defined in terms of a ratio, which was evaluated up to one significant digit after the decimal point. The criteria are as follows:

• The VNLS should create equal or larger space availability, compared to the real windows. • The VNLS should create equal or better illuminance uniformity, compared to the real

windows. • The VNLS should create average ground contribution on the ceiling that is within ±0.1

(10%) of that in the real windows scene. • The VNLS should create equal or smaller average probability of discomfort glare as

observed in position C, compared to the real windows.

The criteria are expressed in mathematical forms as follows.

R

V %%

AA

≥ 1.0 (6.13)

R0

V 0UU

≥ 1.0 (6.14)

0.9 ≤

R

V %%

av

avGG

≤ 1.1 (6.15)

V

R PDGPDG

av

av ≥ 1.0 (6.16)


119

where the subscripts of V and R correspond to the VNLS and real windows scene with the same average surface luminance.

Moreover, it is preferable to have an average surface luminance which does not exceed 3200 cd/m2. This was the value given in the experiments of Shin et al. (2012), where the subjects on average perceived glare from the simulated windows as ‘acceptable’, i.e. scored as 2.5 out of 4.5 on their discomfort glare rating scale.


The results of glare assessment for positions A, B, and C performed on the four aforementioned variations are shown in Table 6.3. The average probabilities of discomfort glare are shown together with their standard deviations.

Table 6.3. Results of glare assessment for positions A, B, and C performed on the four variations in both VNLS and real windows (RW) scenes

Type Conf. IA [°] BA[°] Φ [lm] Pos. DGP DGIn UGRn CGIn PDGav SD

RW 1a Lav = 3200 cd/m2 A 0.24 0.21 0.35 0.39 0.30 0.08

B 0.21 0.19 0.31 0.33 0.26 0.07

C 0.26 0.31 0.43 0.45 0.36 0.09

VNLS 1a 2.0 76 11100A 0.24 0.21 0.36 0.39 0.30 0.09

B 0.21 0.20 0.32 0.35 0.27 0.08

C 0.27 0.33 0.46 0.48 0.38 0.10

RW 2a Lav = 3200 cd/m2 A 0.22 0.26 0.36 0.39 0.31 0.08

B 0.21 0.22 0.32 0.34 0.27 0.07

C 0.26 0.33 0.43 0.45 0.37 0.09

VNLS 2a 2.0 76 11100A 0.21 0.17 0.32 0.35 0.26 0.09

B 0.21 0.21 0.31 0.34 0.27 0.07

C 0.27 0.34 0.45 0.47 0.38 0.09

Since the standard deviations in VNLS scenes are found to be very similar and never differing more than 0.01 from their real windows counterpart, the average probability of discomfort glare can be taken as an indicator for both the VNLS and the real windows scene. The results also show that the probability of discomfort glare at position C is always found to be the largest; hence it is considered sufficient to take into account only this position in the complete glare assessment for the entire variations.

Table 6.4 summarises the space availability (%A), uniformity (U0), ground contribution

(%Gav) and average probability of discomfort glare (PDGav) for all window variations/

Chapter 6

120

configurations with total luminous flux of 11100 lm. Note that configurations with the same distance between windows, beam angle, and total luminous flux are compared to the same reference real window, of which the performance indicators are shown directly above them in

the table. For instance, configurations (1a, IA = 2.0°, BA = 38°, Φ = 11100 lm), (1a, IA =

1.5°, BA = 38°, Φ = 11100 lm), and (1a, IA = 1.0°, BA = 38°, Φ = 11100 lm), are all compared to real windows with an average surface luminance of 10000 cd/m2.

Table 6.4. Summary of space availability, uniformity, average ground contribution, and probability of discomfort glare for all variations and position C in both VNLS and real windows (RW) scenes with Φ = 11100 lm

Conf. IA [°] BA [°] Φ [lm] %A [%]

U0

[-] %Gav

[%] PDGav

[-] 1a RW – 10000 cd/m2 70 0.19 51 0.42 1a 2.0 38 11100 32 0.21 61 0.43 1a 1.5 38 11100 33 0.23 59 0.45 1a 1.0 38 11100 34 0.26 55 0.46

1a RW – 3200 cd/m2 27 0.16 51 0.36 1a 2.0 76 11100 31 0.28 50 0.38 1a 1.5 76 11100 32 0.30 47 0.39 1a 1.0 76 11100 32 0.32 44 0.39

1a RW – 1800 cd/m2 14 0.18 50 0.34 1a 2.0 114 11100 28 0.37 49 0.35 1a 1.5 114 11100 29 0.37 47 0.35 1a 1.0 114 11100 30 0.37 45 0.35

2a RW – 10000 cd/m2 63 0.17 48 0.43 2a 2.0 38 11100 35 0.23 59 0.44 2a 1.5 38 11100 35 0.25 57 0.46 2a 1.0 38 11100 32 0.28 55 0.47

2a RW – 3200 cd/m2 30 0.15 50 0.37 2a 2.0 76 11100 34 0.32 49 0.38 2a 1.5 76 11100 34 0.33 47 0.39 2a 1.0 76 11100 33 0.35 44 0.39

2a RW – 1800 cd/m2 15 0.16 48 0.35 2a 2.0 114 11100 29 0.36 48 0.35 2a 1.5 114 11100 31 0.36 46 0.35 2a 1.0 114 11100 31 0.38 45 0.35


121

6.3.1. Sensitivity analysis

Figure 6.7 displays the standard regression coefficient (β’) of all input variables (i.e.,

distance between windows (d), interval of tilt angle (IA), beam angle (BA), and total

luminous flux (Φ)), evaluated for the four performance indicators, i.e. %A, U0, %Gav, and

PDGav. The coefficients of determination values R2 are respectively 0.97, 0.95, 0.80, and

0.96.

Figure 6.6. Standard regression coefficient of all input variables (i.e. distance between windows, interval of tilt angle, beam angle, and total luminous flux), evaluated for the four performance indicators (i.e. %A, U0, %Gav, and PDGav)

As can be seen in the graph, luminous flux and beam angle are the most influential input variables. The graphs showing the relationship between arithmetic mean of the output and the most influential input variable(s) with a 95% confidence level are displayed in Figure 6.8. Table 6.5 gives the summary of arithmetic mean, minimum, maximum, standard deviation, and 95% confidence level of the output.

Table 6.5. Summary of mean, minimum, maximum, standard deviation, and 95% confidence level of the output and the most influential input variable(s)

Input Output Mean Min. Max. SD Confd. 95%

Φ = 6200 lm

%A [%]

10 1 16 3 2

Φ = 11100 lm 32 28 35 2 1

Φ = 19900 lm 70 56 84 9 4

Chapter 6

122


Input Output Mean Min. Max. SD Confd. 95%

BA = 38° U0 [-]

0.24 0.21 0.28 0.02 0.01 BA = 76° 0.32 0.28 0.35 0.02 0.01 BA = 114° 0.37 0.36 0.38 0.00 0.00

BA = 38° %Gav [%]

57 54 61 2 1 BA = 76° 47 44 50 2 1 BA = 114° 46 44 49 2 1

BA = 38° PDGav [-]

0.45 0.40 0.51 0.03 0.01 BA = 76° 0.39 0.36 0.42 0.02 0.01 BA = 114° 0.35 0.32 0.38 0.02 0.01

(a) (b)

(c) (d)

Figure 6.7. Graphs showing the relationship between arithmetic mean of the output and the most influential input variable(s), with a 95% confidence level


123


The space availability is highly, positively influenced (β’ = 0.98) by the total luminous flux of VNLS. Figure 6.8a shows that on average, a total luminous flux of 6200 lm, 11100 lm, and 19900 lm will create space availability of around 10%, 32%, and 70%, respectively. Note that the total luminous flux values are set on a logarithmic scale, with an increment factor of around 1.8. The mean space availability values increase with a larger factor; that is 3.2 when increased from 6200 lm to 11100 lm, and 2.2 when increased from 11000 lm to 19900 lm.

6.3.1.2. Uniformity

The uniformity is highly, positively influenced (β’ = 0.94) by the beam angle of VNLS. On average, a beam angle of 38°, 76°, and 114° will create uniformity of around 0.24, 0.32, and 0.37, respectively (see Figure 6.8b). The relationship between these input and output variables is almost perfectly linear.

6.3.1.3. Ground contribution on the ceiling

The average ground contribution on the ceiling is highly, negatively influenced (β’ = –0.82) by the beam angle of the VNLS. On average, a beam angle of 38°, 76°, and 114° will create an average ground contribution of around 57%, 47%, and 46%, respectively. The mean output values are decreased by around 10% (absolute difference), when the input is increased from 38° to 76°; but they are only decreased by around 0.4% when the input is increased from 76° to 114°, see Figure 6.8c.

6.3.1.4. Probability of discomfort glare The average probability of discomfort glare is highly, negatively influenced (β’ = –0.85)

by the beam angle of the VNLS. Figure 6.8d shows that on average, a beam angle of 38°, 76°, and 114° will create an average probability of discomfort glare of around 0.45, 0.39, and 0.35, respectively, as observed at position C. The relationship between these input and output variables is almost perfectly linear.

6.3.2. Comparison with real windows

Table 6.6 summarises the ratio of space availability, uniformity, and average ground contribution of each VNLS configuration to those of real windows with the same average surface luminance; and the ratio of the average probability of discomfort glare at position C in the real windows scene to that in a VNLS scene with the same average surface luminance; with total luminous flux of 11100 lm. The bold-typed values are those satisfying the criteria, given that the average surface luminance should not exceed 3200 cd/m2.

Chapter 6

124

Table 6.6. Ratio of %A, U0, %Gav,of each VNLS configuration to those of real windows with the same average surface luminance; and ratio of PDGav at position C in real windows scene to that in VNLS scene with the same average surface luminance, with Φ = 11100 lm

Conf. IA [°] BA [°] Φ [lm]R

V %%

AA

R0

V 0UU

R

V %%

av

avGG

V

R PDGPDG

av

av

1a 2.0 38 11100 0.5 1.1 1.2 1.0

1a 1.5 38 11100 0.5 1.2 1.1 0.9

1a 1.0 38 11100 0.5 1.4 1.1 0.9

1a 2.0 76 11100 1.1 1.7 1.0 0.9

1a 1.5 76 11100 1.2 1.8 0.9 0.9

1a 1.0 76 11100 1.2 2.0 0.9 0.9

1a 2.0 114 11100 2.0 2.1 1.0 1.0

1a 1.5 114 11100 2.0 2.1 0.9 1.0

1a 1.0 114 11100 2.1 2.1 0.9 1.0

2a 2.0 38 11100 0.6 1.3 1.2 1.0

2a 1.5 38 11100 0.6 1.4 1.2 0.9

2a 1.0 38 11100 0.5 1.6 1.1 0.9

2a 2.0 76 11100 1.1 2.1 1.0 1.0

2a 1.5 76 11100 1.1 2.1 0.9 0.9

2a 1.0 76 11100 1.1 2.3 0.9 0.9

2a 2.0 114 11100 2.0 2.2 1.0 1.0

2a 1.5 114 11100 2.1 2.3 1.0 1.0

2a 1.0 114 11100 2.1 2.3 0.9 1.0

The results show that most of the VNLS with a beam angle of 38° (narrow spread) fail to create larger space availability relative to the real windows. In terms of space availability, most of the VNLS with a beam angle of 76° (medium spread) yield a ratio of around 1.0, relative to the corresponding real window, which mean they perform the closest to real windows.

Meanwhile, all of the VNLS with a beam angle of 114° (wide spread) satisfy all criteria and yield a ratio of space availability of around 2.0, relative to the corresponding real window. This means the VNLS with a beam angle of 114° can outperform real windows, by giving a larger space availability. A luminous intensity from a VNLS with a 114° beam angle is more evenly distributed throughout the space; hence more space can satisfy the illuminance criterion on the workplane. The appearance of the CIE overcast sky model for the real windows, which is typically characterised by an almost diffuse luminous intensity distribution pattern over the workplane, can be best approached by using a wide spread beam angle for the VNLS model. Note that all of the modelled VNLS, regardless their beam angles,


125

yield larger uniformity than the corresponding real windows, due to the variation of tilt angles of the ‘sky’ sources, which in turn deliver more light to the rear side of the room.

6.3.2.1. Space availability and uniformity Within the variations satisfying all criteria, the gain of space availability is between 1.1 ~

2.3, and the gain of uniformity is between 1.4 ~ 2.5, compared to real windows. For example, for an office of 19.4 m2 floor area as in this case, the real windows with an average surface luminance of 1800 cd/m2 produce a daylit area of approximately 2.9 m2. If VNLS with 114° beam angle and the same average surface luminance are used instead of the real windows, they can produce a daylit area of approximately 5.7 m2 ~ 6.0 m2, which is a 100% increase. The uniformity is also increased from 0.16 in the real windows scene to 0.36 in VNLS scene.

Figure 6.9 displays two sets examples of images with isolux contour lines on the workplane of the following configurations which satisfy all criteria:

• Real window: configuration 1a (d = 0), Lav = 1800 cd/m2

• VNLS: configuration 1a (d = 0), IA = 2.0°, BA = 114°, Φ = 11100 lm, Lav = 1800 cd/m2

• Real window: configuration 2a (d = 0.75 m), Lav = 1800 cd/m2

• VNLS: configuration 2a (d = 0.75 m), IA = 2.0°, BA = 114°, Φ = 11100 lm, Lav = 1800 cd/m2

(a) (b)

(c) (d)

Figure 6.8. Isolux contour lines on the workplane of configurations (a) real window 1a (d = 0), Lav = 1800 cd/m2; (b) VNLS 1a (d = 0), IA = 2.0°, BA = 114°, Φ = 11100 lm; (c) real window 2a (d = 0.75 m), Lav = 1800 cd/m2; (d) VNLS 2a (d = 0.75 m), IA = 2.0°, BA = 114°, Φ = 11100 lm

Chapter 6

126

From these shown examples, it can be seen that the VNLS have an isolux pattern similar to the corresponding real windows. The contour lines for 500 lx values are however located at different distances from the window. The area covered by the 500 lx contour line, which is the space availability, in the VNLS scene is approximately double the size of that in the real windows scene. The uniformity under the VNLS is also larger compared to the real window, by approximately the same factor of 2.

6.3.2.2. Ground contribution on the ceiling Within the variations satisfying all criteria, the ratio of average ground contribution on the

ceiling in the VNLS and real windows scene is within the range of 0.9 ~ 1.1. However, a relatively large difference is found between the pattern of ground contribution propagation in the VNLS and real windows scene. Figure 6.10 displays graphs showing the ground contribution propagation for a selected number of VNLS variations; all with Φ = 11100 lm, together with the reference real window case.

(a) (b)

(c) (d)

Figure 6.9. Graphs showing propagation of ground contribution on the ceiling for a selected number of variations of VNLS (solid lines), displayed together with the reference real windows (dotted line)


127

From Figure 6.10, it can be seen that the ground contribution on the ceiling propagates differently under real windows and under VNLS with different beam angles. In the real windows scene, the ground contribution values are always within 30% to 70% (configuration 1a) or within 40% to 70% (configuration 2a). The propagation also shows more fluctuations, since the received light on the ceiling mostly come from reflections from the ground outside or from other interior surfaces. In the VNLS scene, the values can reach up to 90% near the window, but then rapidly decline. The received light on the ceiling mostly come from the ‘ground’ light sources, therefore the propagation is smoother. The row of light sources that represents the ground in the VNLS sources have their role here, where the beam angle is set at 76° (medium spread, constant in all variations) and tilted with a 40° angle pointing upward.

In general, given a constant beam angle and tilt angle of the ‘ground’ source, the lower the beam angle of the ‘sky’ source, the higher the average ‘ground’ contribution on the ceiling; since the ‘sky’ will contribute less to the ceiling. The results show that there is an inverse correlation between the beam angles chosen for the light sources that represent the ‘sky’ and the average ‘ground’ contribution on the ceiling. However, a low beam angle of the ‘sky’ source creates a rapid decline in propagation, making it less similar to the situation with real windows (with an overcast sky condition).

6.3.2.3. Probability of discomfort glare Within the variations satisfying all criteria, the ratio of average probability of discomfort

glare (real windows to VNLS) is found to be 1.0. In the other variations, this ratio is 0.9, thus in no cases are these ratios found to be larger than 1.0. It means that relative to the corresponding real windows, the VNLS generally create similar or slightly worse average probability of discomfort glare. Figure 6.11 displays the impression of the selected configurations whose isolux contour lines are displayed in Figure 6.9.

From these shown examples, it can be seen that the VNLS with 114° beam angle have some similarities and differences compared to the corresponding real windows. While the average window surface luminance and the average probability of discomfort glare are approximately similar, a few differences are still recognisable. For instance, the luminance from the ‘ground’ element sources of VNLS are significantly larger than that from the real ground element, if viewed from position C. This high luminance is needed for the VNLS to be able to resemble the impression of ground reflection on the ceiling. Despite the high luminance of the ‘ground’ sources in VNLS, the average probability of discomfort glare viewed from the observer positions is still comparable to that in real window scenes. The wide spread beam angle of the ‘sky’ sources reduces the green-coloured impression on the ceiling, but on the other hand also distributes the light to a wider area of the space, hence creating a more uniformly lit space.

Chapter 6

128

(a) (b)

(c) (d)

Figure 6.10. Impression of configurations (a) real windows 1a (d = 0), Lav = 1800 cd/m2; (b) VNLS 1a (d = 0), IA = 2.0°, BA = 114°, Φ = 11100 lm; (c) real windows 2a (d = 0.75 m), Lav = 1800 cd/m2; (d) VNLS 2a (d = 0.75 m), IA = 2.0°, BA = 114°, Φ = 11100 lm


One of the key research objectives in this thesis is to investigate if VNLS could be designed with a similar lighting and view performance to real windows. In this simulation study, VNLS configurations composed of light emitting sources with a size of 0.05 m × 0.05 m have been developed. It shows the possibility to model the direction of light from the ‘ground’ to the ceiling and from the ‘sky’ to the floor. It is concluded that:

• The total luminous flux of VNLS greatly influences the space availability (standard regression coefficient β’ = 0.98).

• The beam angle of VNLS greatly influences the uniformity (β’ = 0.94), average ground contribution on the ceiling (β’ = –0.82), and average probability of discomfort glare (β’ = –0.85).


129

• Most of the VNLS with a beam angle of 76° (medium spread) yield a ratio of space availability of around 1.0, relative to the corresponding real window, which mean they perform the closest to real windows.

• All of the VNLS with a beam angle of 114° (wide spread) satisfy all criteria and yield a ratio of space availability of around 2.0, relative to the corresponding real window, which mean they outperform real windows.

Compared to real windows (under the CIE overcast sky), the gain of space availability is between 1.1 ~ 2.3, and the gain of uniformity is between 1.4 ~ 2.6. For example, real windows with an average surface luminance of 1800 cd/m2 produce a daylit area of approximately 2.9 m2 in an office of 19.4 m2 floor area. The VNLS with 114° beam angle and the same average surface luminance can produce a daylit area of 5.7 m2 ~ 6.0 m2 in the same space without real windows. The uniformity is also increased from 0.16 to 0.36. To some extent, it shows the benefit of VNLS compared to the real windows.

As mentioned in the introduction, the work presented in this chapter is a report on progress in the VNLS development. The results of this study are based on simulation of VNLS with a rather simple image and the light sources in the same row are all set with a same horizontal angle. The rendered images are very different from actual window views. Therefore, further studies involving more detailed images on the window view, as well as more features of real daylight that influence visual comfort as previously mentioned in Table 2.3 in Chapter 2, are required to improve the similarity to real windows. More sophisticated configurations and source parameters will be studied to further improve the visual comfort characteristics. Nevertheless, the results presented in this chapter show clear examples on how building performance simulation contributes in the research and development of non-existing solutions, by demonstrating that the numerical model of VNLS can perform better in some ways than that of real windows.

Chapter 6

130

131

Chapter 7 Modelling and Simulation of Virtual Natural Lighting Solutions with Complex Views and Directional Light

This chapter discusses the modelling and simulation of future (not-yet-existing) VNLS with complex views and directional light, using Radiance to predict the lighting performance. The model features an array of small light sources with various direct light components, a transparent glass surface in front of the light sources array and a two-dimensional image pasted on this surface. Two input variables were varied to observe their effect on the lighting performance of a reference office. Ten view images were introduced and compared to each other.

7.1. Introduction

In Chapter 6 of this thesis, a model of VNLS with a small light source array constructing a simplified view that resembles the blue sky and green ground has been described. The blue-coloured ‘sky’ elements were tilted downward to deliver light to the workplane, while the green-coloured ‘ground’ elements were tilted upward to deliver light to the ceiling (Mangkuto et al., 2014), referring to the ideal CIE overcast sky where the split-flux method applies (Tregenza, 1989). The results show that the total luminous flux greatly influences the space availability, while the beam angle is highly influential on the uniformity, average ground contribution on the ceiling, and average probability of discomfort glare. Most of the VNLS satisfying all criteria (in terms of ratio, compared to the real windows scene) are those having a beam angle of 114° (wide spread).

While the findings may give an illustration on how VNLS will perform in a space, it is noticed that VNLS ideally should generate directional (non-diffuse) light as well as a relatively complex view, which is not yet the case for the VNLS model in the previous chapter. In this particular chapter, more complex views are incorporated into the model, while maintaining the directional light component.

Many researchers have conducted investigation on components that should be present in viewed images (e.g. Ulrich, 1984; Ulrich et al., 1991; Tennessen & Cimprich, 1995; Chang & Chen, 2005; de Kort et al., 2006; Aries et al., 2010; Beute & de Kort, 2013). Tuaycharoen & Tregenza (2007) suggested that view cannot be separated from the natural (day-) light itself. Beute & de Kort (2013) found consistent preferences for natural over urban scenes, sunny over overcast scenes, and bright over dim scenes.

Hellinga & de Bruijn-Hordijk (2009) proposed that for the view elements of windows, the highest quality level will be achieved if the view contains the following: 1. Green, sky, and distant objects

Chapter 7

132

2. Maximum information about outside environment, such as weather, season, time of day, and (human) activities

3. Highly complex and coherent image scene

The use of complex view on virtual windows in a laboratory environment to investigate the psychological effects has been explored by some researchers. For example, IJsselsteijn et al. (2008), who focused on depth perception cues from screen projected images, used five image scenes, showing the presence of: (1) trees, ground, and three people standing, (2) trees and ground without people, (3) creek, (4) desert, and (5) city skyline on a river at night. All scenes other than the last one display a daytime sky.

In their experiments on discomfort glare from projected images, Tuaycharoen & Tregenza (2005) used 10 pairs of image scenes displaying either ‘natural’ (showing the presence of mountains, river, and/or trees) or ‘urban’ (showing the presence of buildings, i.e. houses, skyscrapers, castle, or school) views. A daytime sky (with or without clouds) is visible in every image scene. They concluded that a good view (also described as a view with high interest), which mainly consists of the natural scenes, tends to reduce discomfort glare perception.

Shin et al. (2012), who focused on subjective discomfort glare evaluation from a backlit, transparent printed image, used five pairs of image scene displaying either ‘distant’ (i.e. the viewed objects are relatively faraway from the window) or ‘near’ (i.e. the viewed objects are relatively close to the window). The scene displayed a ‘mixed land’ (skyscrapers and trees), ‘man-made’ (skyscrapers), ‘mixed river’ (city skyline on a river), ‘natural land’ (trees and green ground), or ‘natural river’ (mountains or plants on a river). All pictures in the scenes were taken during daytime, but the sky is only (partly) visible on the ‘distant’ scenes, and not at all on the ‘near’ scenes. They concluded that the tolerance of discomfort glare sensation for the distant views including skyline was greater than the near views.

Considering the variation and clear distinction of the objects’ distance, the 10 image scenes of Shin et al. were incorporated to model the VNLS with complex views in this chapter. Some adaptations were made, including stretching, mirroring, and cropping the upper and lower part of the original image to get the same height of horizon (i.e. the border between the ground and the sky), and to get the same image size in every scene. The adapted image scenes are displayed in Figure 7.1.

Modelling and Simulation of VNLS with Complex Views and Directional Light

133

(a) (b) (c) (d)

(e) (f) (g) (h)

(i) (j)

Figure 7.1. Image scenes adapted from Shin et al. (2012): (a) ‘Distant Mixed Land’ (DML), (b) ‘Near Mixed Land’ (NML), (c) ‘Distant Man-made’ (DMM), (d) ‘Near Man-made’ (NMM), (e) ‘Distant Mixed River’ (DMR), (f) ‘Near Mixed River’ (NMR), (g) ‘Distant Natural Land’ (DNL), (h) ‘Near Natural Land’ (NNL), (i) ‘Distant Natural River’ (DNR), (j) ‘Near Natural River’ (NNR)

The work described in this chapter aims to demonstrate the role of building performance simulation in the research and development of VNLS, by predicting the performance of numerical model of VNLS with complex views and directional light on the display. In particular, the objectives of this study are twofold. The first objective is to understand the effect of changing input variables of the VNLS with complex views, which in this case are: beam angle and total luminous flux of the ‘non-ground’ elements, as well as the variation on the image scene itself; on the lighting performance of a reference office space. The second objective is to compare two techniques of modelling the view: using the ‘emissive’ approach, i.e. where the light sources are coloured and constructing the view itself, and using the ‘transmissive’ approach, i.e. where the light sources are all white and the view is made by pasting an image on a transparent surface in front of the sources.

Similar to the case of VNLS with a simplified view, the lighting performance is hereby described in terms of the ability to meet the space availability demand, the illuminance uniformity on the workplane, the illuminance contribution from the ground elements on the ceiling, and the ability to produce minimal glare at predefined observer’s positions in the space.

Chapter 7

134

7.2. Methods

7.2.1. Modelling

While all detailed characteristics of view from a window are considered very important for developing the requirement of VNLS, this study focuses on modelling the characteristics of direct light from the sky and reflected light from the exterior ground. In general, most of the existing VNLS prototypes behave like a diffuse light source, reducing the possibility of seeing the impression of direct and reflected light components on the interior surface.

Therefore, a more complex model of VNLS is proposed, in the form of arrays of small directional light emitting areas that are tilted, as specified in Chapter 6 of this thesis. To realise a complex view in the Radiance simulation tool, a two-dimensional image scene was imported and pasted/mapped on a very thin, vertically flat, transparent glass (τ = 0.90) material. The glass itself did not emit light, and it was put in front of the light source arrays. Ten image scenes adapted from Shin et al. (2012) were used, as displayed in Figure 7.1.

All of the light emitting areas were white, since the actual colour display was given by the mapped image. The bottom array acted as the ‘ground’ which was tilted upward to deliver the light to the ceiling. The rest of the sources acted as the ‘sky’ which was tilted downward to deliver the light to the workplane.

The light emitting areas were modelled to fit two individual vertical openings, each with the size of 0.80 m × 1.17 m (W×H), corresponding to a window-to-wall ratio of 20%. Each light emitting area in each individual window had a size of 0.05 m × 0.05 m and had the role of lighting the ‘non-ground’ part of the image. At the lowest row, there were four light emitting areas (0.20 m × 0.20 m each) to light the ‘ground’ part of the image.

To model the directionality of light entering through a window, the sources at the row directly above the ‘ground’ were at all times not tilted (i.e. 0°), while the sources at the second row above the ‘ground’ were tilted downward by 2.0°. The sources at the third row above the ‘ground’ were tilted downward by 4.0° and so forth. The ‘ground’ sources were always tilted with a 40° angle pointing upward.

Figure 7.2 displays the front and side views of the VNLS model.


135

Figure 7.2. Front and side views of the VNLS model

The sources had a certain beam angle, i.e. the angle between the two directions opposed to each other over the beam axis for which the luminous intensity is half that of the maximum luminous intensity. To observe the effects of varying beam angle, three values of beam angle for the ‘non-ground’ sources were used, i.e. 38° (relatively narrow spread), 76° (medium), and 114° (wide).

The luminous intensity distribution of each light source was written in an IES format file, based on the character of downlights with a certain beam angle. For the sources with a 114° beam angle, the distributions were set so that the combination of these sources, without

the addition of the transparent glass, gave an average surface luminance (L [cd/m2]) of 1000 cd/m2 (low luminance setting), 1800 cd/m2 (medium), or 3200 cd/m2 (high). These were the first three values used in the experiments of Shin et al. (2012). The intensity values for the ‘non-ground’ sources with 38° and 76° beam angles were adjusted accordingly, so that the total luminous flux coming from the ‘non-ground’ sources altogether remained the same. The total luminous flux from the source was then calculated based on the zonal cavity method described by Lindsey (1997).

In this case, the calculated total luminous fluxes of all ‘non-ground’ sources, without the addition of the transparent glass, are approximately 6200 lm, 11100 lm, and 19900 lm. Figure 6.3 is referred to display the luminous intensity distributions of the ‘non-ground’ sources. The settings for each ‘ground’ source in this case were also identical with the case for VNLS with a simplified view, having a fixed beam angle of 76 degrees, and maximum intensity of 110 cd, 199 cd, and 354 cd for the three conditions respectively.

The variation in all input variables, including the image scene is summarised in Table 7.1. It should be noticed that the interval of tilt angle and the distance between windows are not considered as input variables that vary, since the finding in the case of VNLS with a simplified view suggests that both of them are less influential than the total luminous flux and beam angle (Mangkuto et al., 2014). In total, there are 90 possible combinations based on the input variables, taking the image scene variations into account.

Chapter 7

136

Table 7.1. Input variables and their values variation

Variable Symbol Unit Values

Beam angle BA deg 38, 76, 114

Total luminous flux Φ lm 6200, 11100, 19900

Image scene - - DML, NML, DMR, NMR, DMM, NMM, DNL, NNL, DNR, NNR

7.2.2. Settings The space discussed in this study was the same reference office space as discussed in

Chapter 6. Since the findings in Chapter 6 show no significant influence from the distance between windows to the lighting performance, there was only one window configuration, chosen for the study in this chapter, illustrated in Figure 6.4a. Reflectance values of the room’s interior were: ceiling: 85%, walls: 50%, floor: 20%, door: 50%, window and door frames: 50%; which are based on the IEA Task 27 reference office (van Dijk & Platzer, 2003).

Three observer positions were defined at the eye height of 1.2 m above the floor, as previously illustrated in Figure 6.5. According to the finding in the case of VNLS with a simplified view, position C that directly faced the window plane receives the most severe discomfort glare. Therefore, the glare analysis was performed only for the observer at position C. For all simulations, the parameters in Radiance were set as previously shown in Table 6.2.

7.2.3. Assessment 7.2.3.1. Performance indicators

As for the case of VNLS with a simplified view, the assessment for this study is also based on the relevant performance indicators as discussed in Chapter 6, which are the space availability, uniformity, average ground contribution on the ceiling, and average probability of discomfort glare.

7.2.3.2. Sensitivity analysis

Sensitivity analysis using multiple linear regressions was performed to evaluate the influencing effect of the current input variables on the defined performance indicators. This

regression model assumes a linear relationship between the output variable yi and the p-vector

of input variables xi. The mathematical model reads as follows:


137

iii xxy 22111 , i = 1, 2, …, 9 (7.1)

where βi is a p-dimensional regression coefficient. In this case: p = 2, n = 3 × 3 = 9, x1 is the

beam angle (BA, in degrees), and x2 is the total luminous flux (Φ, in lumens); while y is

evaluated individually for %A, U0, %Gav, and PDGav. The calculation is performed individually for the 10 image scenes, i.e. no mix between different image scenes.

The values were standardised, and then were put in the regression model expressed in the matrix form as follows:

'...

'1

ny

y =

''

'...

'

'...

'

2

1

2

12

1

11

nn x

x

x

x+

'...

'1

n

, n = 9 (7.2)

The equations were then solved using a MATLAB toolbox to determine '1 and '2 ,

which are the standard regression coefficients that determine the sensitivity of the output, i.e.

BA and Φ respectively, as function of the input. The values range from 1 (strong, positive influence) to –1 (strong, negative influence).

7.2.3.3. Comparison of transmissive and emissive approaches

As mentioned in Section 7.1, the second objective of this study is to compare the approach of modelling the view on VNLS, by observing the effect on the lighting performance. In Chapter 6 of this thesis, the so-called ‘emissive’ approach was employed to model the VNLS with a simplified view. In that case, the light sources were either blue or green-coloured in order to construct the view of the blue sky and green ground. While in this chapter, the so-called ‘transmissive’ approach is introduced, where the light sources are all white and do not build the view themselves. The view is realised by pasting a two-dimensional image on a transparent, glass surface in front of the sources. It is then intended to know the impact of using these two approaches on the lighting performance of the given space.

To make the comparison, two characteristically different image scenes were introduced, i.e.: (1) a blue sky and green ground (referred as BG), and (2) a green ‘sky’ (entirely obstructed) and blue ‘ground’ (referred as GB), as in a simplified view of green trees on a river, seen from a relatively near distance. The image scenes are illustrated in Figure 7.3. While the first scene is more common, the second scene can be considered located on the other side of the spectrum. In the GB scene, the general composition of blue and green colours is inverted; the sky is entirely covered and the ‘ground’ appears brighter.

Chapter 7

138

(a) (b)

Figure 7.3. Image scenes of (a) ‘blue sky and green ground’ (BG) and (b) ‘green obstructed sky and blue ground’ (GB)

The image scenes were drawn using the Colour Picker from JALOXA website (Jacobs, 2012b), where colour properties are given in red, green, and blue values between 0 ~ 1 as used in Radiance. If these values are seen as the spectral reflectance of a given material, then the weighted average reflectance (ρ) can be obtained using the formula described by Jacobs (2012a):

ρ = 0.265 ρR + 0.670 ρG + 0.065 ρB (7.3)

where ρR, ρG, ρB are respectively the spectral reflectance in red, green, and blue. By analogy, a similar expression can be applied for relating the weighted average transmittance and spectral transmittance of a given material, which is the case for the transmissive approach. The equation then reads as follows:

τ = 0.265 τ R + 0.670 τ G + 0.065 τ B (7.4)

The properties of each element of both image scenes are described in Table 7.2. As shown there, using the transmissive approach, the shade of green in the GB scene is equal to that in the BG scene, as displayed in Figure 7.3. To display a darker colour using the emissive approach, one needs to reduce the luminous intensity of the source, hence also reduce the total luminous flux. In this comparison, it was chosen to maintain the same total luminous flux of the light sources in both approaches. Therefore, the colour displays in the emissive approach are practically different with that in the transmissive approach.

Table 7.2. Colour properties of each element of the BG and GB image scenes

Scene Element Red Green Blue Average

BG ‘Sky’ 0.500 0.800 1.000 0.734

BG ‘Ground’ 0.300 0.500 0.200 0.428

GB ‘Sky’ 0.300 0.500 0.200 0.428

GB ‘Ground’ 0.500 0.800 1.000 0.734


139

In the emissive approach, the light emitting areas were created from standard IES files describing the luminous intensities at all relevant angles, then converted into standard Radiance object files using the programme ies2rad. By default, the light emitting areas are defined as ‘glow’ materials with colour properties of (1, 1, 1), i.e. white. To change the colour, one may use the colour properties as suggested by JALOXA, but the values should be normalised with their weighted average value, before inserting in Radiance. This is necessary to ensure the final weighted average value stays equal to 1, thus maintaining correct values for the resulting luminances (Jacobs, 2012a). For example, to display a light-blue colour of (0.5, 0.8, 1), one has to calculate the weighted average transmittance value using Equation 7.4, which gives 0.734. The normalised values are then (0.5/0.734, 0.8/0.734, 1/0.734), or (0.682, 1.091, 1.363). These values should be filled in the Radiance object file of the light source, replacing the default values of (1, 1, 1).

The input variables were beam angle of 38°, 76°, and 114°, interval of tilt angle of 2.0°, and total luminous flux of the ‘non-ground’ elements of 11100 and 19900 lm. In the emissive approach, the models were built by assigning the relevant colours to the light sources; while in the transmissive approach, the relevant image in Figure 7.7 was pasted on the glass surface in front of the white light sources. Note that for the latter approach, the total luminous flux of the ‘non-ground’ elements belongs to the light sources only, without considering the transparent glass surface.

For all variations, the simulations were performed to evaluate the four performance indicators mentioned in Section 6.3.1.


The results are divided into two main parts, i.e. the sensitivity analysis for the ten image scenes (Figure 7.1) modelled using the transmissive approach (Section 7.2.1), and analysis of the BG and GB scenes (Figure 7.3) modelled using the transmissive and emissive approach (Section 7.2.3.3).

7.3.1. Transmissive approach

The standard regression coefficients of all input variables, i.e. beam angle (BA) and total

luminous flux (Φ), are shown in Figure 7.4. They were evaluated for the four performance

indicators, i.e. %A, U0, %Gav, and PDGav, under the 10 image scenes. The coefficients of

determination values R2 are also shown for each performance indicator and each image scene.

Chapter 7

140

(a)

(b)

(c)

Figure 7.4. Standard regression coefficient of all input variables (i.e. BA and Φ), evaluated for the four performance indicators, i.e. (a) %A, (b)U0, (c) %Gav, and (d) PDGav under the 10 image scenes


141

(d)


As illustrated in Figure 7.4, total luminous flux is the most influential input variable to the space availability, while beam angle is the most influential input variable to the other output variables, independent of the image scene. It is however noticed that under the scenes of ‘Near Natural Land’ and ‘Near Natural River’, the standard regression coefficients (β’) for total luminous flux related to space availability are respectively 0.56 and 0.59, while the value under other scenes is always larger than 0.80, which is discussed later on. Figure 7.5 displays the graphs showing the relationship between arithmetic mean of the output and the most influential input variable(s) with a 95% confidence level, under the 10 image scenes.

(a) (b)

Figure 7.5. Graphs showing the relationship between arithmetic mean of the output (i.e. %A, U0, %Gav, and PDGav) and the most influential input variable(s), with a 95% confidence level

Chapter 7

142

(c) (d)



The space availability is highly, positively influenced (β’ = 0.56 ~ 0.96, depends on the image scene) by the total luminous flux of the ‘non-ground’ elements. The largest values are achieved under the ‘Near Mixed River’ and ‘Distant Natural River’ scene, where a total luminous flux of 19900 lm will create a space availability of around 27% and 24%, respectively. The mean space availability values increase with a factor of 4 (under NMR scene) and 5 (under DNR scene) when increased from 6200 lm to 11100 lm, then 2.2 (NMR) and 2.4 (DNR) when increased from 11000 lm to 19900 lm. Both image scenes provide either a large white-coloured area (i.e. the skyscraper buildings’ façade in NMR) or a bright, bluish-coloured area (i.e. the sky in DNR). Hence, more light is transmitted through the image plan, generating a larger workplane area with illuminance exceeding 500 lx.

On the other side of the scale, the ‘Near Natural Land’ and ‘Near Natural River’ scenes generate the smallest space availability, which is only 1% and 3% when the total luminous flux is 19900 lm, and zero when the total luminous flux is lower. Both scenes have the sky entirely covered with either green trees or red plants on a hill. Those elements block most of the light transmission, resulting in a low value of space availability. Under these two scenes, increasing total luminous flux does not necessarily increase the space availability, which in turn makes the standard regression coefficient not as large as under the other eight scenes.


143

7.3.1.2. Uniformity

The uniformity is highly, positively influenced (β’ = 0.75 ~ 1.00) by the beam angle of the ‘non-ground’ elements. Under a given scene and with a given beam angle, the uniformity practically stays the same when the total luminous flux is changed. The minimum and average illuminance values on the workplane are directly proportional to the total light output of the source, hence the constant value of uniformity. On average, the ‘Near Natural Land’ and ‘Near Natural River’ scenes generate the largest uniformity, which are 0.44 and 0.48 for beam angle of 114°. While both scenes have the sky entirely covered with relatively uniform, dim objects (i.e. green trees or red plants), the resulting minimum and average illuminance values on the workplane are also relatively close to each other.

Under the other eight scenes, the uniformity values range from 0.12 to 0.19 for beam angle of 38°, 0.22 to 0.28 (76°), and 0.32 to 0.38 (114°). The relationship between the beam angle and uniformity is almost perfectly linear under all image scenes.

7.3.1.3. Ground contribution on the ceiling

The average ground contribution on the ceiling is highly, negatively influenced (β’ = –0.91 ~ –0.97) by the beam angle, quite independent of the image scene. As found for the uniformity, under a given scene and with a given beam angle, the average ground contribution on the ceiling stays constant when the total luminous flux is changed. A similar

trend is also observed in terms of the image scenes creating the largest %G values. The ‘Near

Natural Land’ and ‘Near Natural River’ scenes generate the largest average ground contribution on the ceiling, which are respectively 73% and 78% for beam angle of 38°, 65% and 70% (76°), and 64% and 67% (114°). In both scenes, the view is almost entirely filled with dim objects (i.e. green trees or red plants), but the ‘ground’ part (i.e. light green grass or blue river) actually appears brighter than the ‘non-ground’ part. This results in a larger contribution of illuminance values on the ceiling, compared to the contribution of the ‘non-ground’ part.

Meanwhile, in the other scenes, the ‘ground’ part is generally darker than the ‘non-ground’ part, creating a lower value of average ground contribution. The scenes of ‘Near Mixed Land’ and ‘Distant Natural River’ generate the smallest average ground contribution on the ceiling, which are respectively 30% and 34% for beam angle of 38°, 21% and 22% (76°), and 17% and 19% (114°).

7.3.1.4. Probability of discomfort glare The average probability of discomfort glare is highly, negatively influenced (β’ = –0.78 ~

–0.89) by the beam angle. Figure 7.5d shows that at a given beam angle, the values under various image scenes are nearly the same, except again under the ‘Near Natural Land’ and ‘Near Natural River’ scenes, which give lower values. For example, the two scenes give

Chapter 7

144

average probability of discomfort glare of 0.33 and 0.32 (beam angle of 38°), while the figures are between 0.35 and 0.40 under the other scenes with the same beam angle. As explained earlier, both scenes have the sky entirely covered with relatively dim objects (i.e. green trees or red plants), which effectively reduce the transmitted light, and in turn also reduce the space availability and discomfort glare perception. The relationship between the beam angle and probability of discomfort glare is almost perfectly linear under all image scenes. Figure 7.6 displays the rendered impression, observed from position C, of some selected image scenes, i.e. ‘Near Mixed River’, ‘Near Natural Land’, and ‘Near Natural River’, with beam angle of 38°, 76°, 114°, and total luminous flux of 19900 lm.

(a) (b) (c)

(d) (e) (f)

(g) (h) (i)

Figure 7.6. Impression of some image scenes, all with Φ = 19900 lm: (a) NMR, BA = 38°; (b) NMR, BA = 76°; (c) NMR, BA = 114°; (d) NNL, BA = 38°; (e) NNL, BA = 76°; (f) NNL, BA = 114°; (g) NNR, BA = 38° (h) NNR, BA = 76°; (i) NNR, BA = 114°


145

While the results of the sensitivity analysis seem to be obvious, the quantitative influence of each input parameter on each output variables under various image scenes can be obtained only by simulation. An interesting example is the probability of discomfort glare; theoretically, as well as empirically, one can assume that the more light output coming from the display, the more discomfort glare it will create. The reverse can be assumed for beam angle; the largest the beam angle, the more discomfort glare. These are indeed also found in the simulation. However, it turns out from the sensitivity analysis that the beam angle is actually the most influential factor on discomfort glare, and not the total luminous flux. One potential reason is that the probability of discomfort glare is mostly dominated by the amount of contrast, and increasing beam angle will result in reducing contrast. Based on the performed simulation, one can decide better on which input variables to focus on, with regard to a certain output variables; rather than merely guessing based on visual observation.

7.3.2. Comparison of transmissive and emissive approaches

Table 7.3 summarises the space availability, uniformity, average ground contribution, and average probability of discomfort glare of the two image scenes, i.e. ‘blue sky and green ground’ (BG) and ‘green obstructed sky and blue ground’ (GB); and the modelling approaches, i.e. transmissive and emissive; all with total luminous flux of 11100 and 19900 lm.

Table 7.3. Summary of %A, U0, %Gav, and PDGav of each VNLS configuration with the BG and GB image scenes, using emissive and transmissive approaches

Approach Scene BA [°]

Φ [lm]

%A [%]

U0 [-]

%Gav

[%] PDGav

[-] Transmissive BG 38 11100 17 0.18 49 0.37

BG 76 11100 13 0.27 38 0.28

BG 114 11100 9 0.37 35 0.25

GB 38 11100 0 0.26 64 0.33

GB 76 11100 0 0.35 55 0.26

GB 114 11100 0 0.42 54 0.23

Emissive BG 38 11100 32 0.21 61 0.43

BG 76 11100 31 0.28 50 0.38

BG 114 11100 28 0.37 49 0.35

GB 38 11100 32 0.21 61 0.43

GB 76 11100 31 0.28 50 0.38

GB 114 11100 28 0.37 49 0.35

Chapter 7

146


Approach Scene BA [°]

Φ [lm]

%A [%]

U0 [-]

%Gav

[%] PDGav

[-] Transmissive BG 38 19900 32 0.18 49 0.41

BG 76 19900 31 0.26 38 0.30

BG 114 19900 28 0.37 35 0.27

GB 38 19900 7 0.25 64 0.36

GB 76 19900 4 0.35 55 0.28

GB 114 19900 2 0.44 54 0.25

Emissive BG 38 19900 72 0.21 60 0.46

BG 76 19900 70 0.28 50 0.41

BG 114 19900 59 0.37 49 0.37

GB 38 19900 72 0.21 60 0.46

GB 76 19900 70 0.28 50 0.41

GB 114 19900 59 0.37 49 0.37

Table 7.3 shows that with the emissive approach, given the same beam angle and total luminous flux, the performance indicators do not change when a different image scene is displayed. To display various colours of a ‘glow’ material, one can edit the red, green, and blue radiance components of the sources, but they have to be normalised using the procedure given by Jacobs (2012a), so that two light sources with the same light output will produce the same illuminance values on the same point, even though the colours are different in display. The consequence of using this technique of producing colours from light is that darker colours can only be realised by reducing the light intensity. In practice, one may not get the intended dark colours; for example a completely black colour will be difficult to display, instead it may just become a dark grey.

Using the transmissive approach, the light output is reduced by the transparent glass on which the image scene is pasted. It is clearly shown that the GB scene reduces a significant amount of light, compared to the BG scene. As a result, the space availability under the GB scene is much smaller than that under the BG scene. Nevertheless, the ground contribution under the GB scene is relatively larger, due to the fact that the ‘ground’ part appears brighter than the rest of the display. The probability of discomfort glare under the GB scene is also smaller than that under the BG scene, obviously due to the smaller amount of light transmitted from the window display.

To understand the difference between the two approaches, Table 7.3 can be further simplified by showing the ratio of each performance indicator obtained using the transmissive approach, compared to that obtained using the emissive approach, for a given beam angle and total luminous flux. These ratio values are displayed in Table 7.4.


147

Table 7.4. Ratio of %A, U0, %Gav, and PDGav of each VNLS configuration with the BG and GB image scenes using the transmissive approach (subscript t), compared to those using the emissive approach (subscript e)

Scene BA [°]

Φ [lm] e

t %%

AA

e0

t 0UU

e

t %%

av

avGG

e

t PDGPDG

av

av

BG 38 11100 0.5 0.9 0.8 0.9

BG 76 11100 0.4 1.0 0.8 0.7

BG 114 11100 0.3 1.0 0.7 0.7

GB 38 11100 0.0 1.2 1.1 0.8

GB 76 11100 0.0 1.2 1.1 0.7

GB 114 11100 0.0 1.1 1.1 0.7

BG 38 19900 0.4 0.9 0.8 0.9

BG 76 19900 0.4 0.9 0.8 0.7

BG 114 19900 0.5 1.0 0.7 0.7

GB 38 19900 0.1 1.2 1.1 0.8

GB 76 19900 0.1 1.2 1.1 0.7

GB 114 19900 0.0 1.2 1.1 0.7

It is seen from Table 7.4, that under the BG scene, the space availability using the transmissive approach is around 30% ~ 50% of the corresponding values using the emissive approach. Under the GB scene, the space availability using the transmissive approach is very near or equal to zero. It should be noticed however that the space availability is calculated based on 500 lx as minimum criterion; hence a value of zero does not necessarily imply that the entire workplane has zero illuminance, but the maximum illuminance is certainly smaller than 500 lx. Moreover, even though the glass transmittance for all cases is fixed at τ = 0.90, the actual light transmitted by the display is apparently much less than this proportion, due to the additional reduction which largely depends on the view elements of the image scene.

The uniformity and ground contribution under the GB scene is slightly larger when using the transmissive approach, compared to the emissive one; while the opposite is true under the BG scene. Under both scenes, the average probability of discomfort glare using the transmissive approach is around 70% ~ 90% of the values obtained using the emissive approach.

To give a clearer illustration, Figure 7.7 displays some rendered impression of the scenes with total luminous flux of 19900 lm and BA = 114°, under both the BG and GB image scenes, using the emissive and transmissive approaches. In the scenes using the emissive approach, the side walls apparently show a strong impression of the colour of the ‘non-ground’ elements, i.e. blue under the BG scene and green under the GB scene. The floor areas near the window also appear less bright and less bluish/greenish in the scenes using the

Chapter 7

148

transmissive approach, due to the light reduction by the display. While the emissive approach generally creates larger space availability, it also generates more contrast between the display and its immediate surrounding, which leads to a larger probability of discomfort glare.

Scene Emissive approach Transmissive approach

BG

GB

Figure 7.7. Impression of the image scenes, all with Φ = 19900 lm and BA = 114°: BG, emissive approach; BG, transmissive approach; GB, emissive approach; and GB, transmissive approach

7.4. Concluding Remarks A model of VNLS configurations with complex views has been created, where the light

was provided by arrays of white-coloured directional light emitting sources with specific tilt angles, while the view was provided by mapping a two-dimensional image on a transparent glass in front of the light sources. It is concluded that:

• Under every image scene, the total luminous flux of the ‘non-ground’ element has a large influence on the space availability (standard regression coefficient β’ = 0.56 ~ 0.96).

• Under every image scene, the beam angle of the ‘non-ground’ element has a large influence on the uniformity (β’ = 0.75 ~ 1.00), average ground contribution on the ceiling (β’ = –0.91 ~ –0.97), and average probability of discomfort glare (β’ = –0.78 ~ –0.89).


149

• The largest space availabilities are achieved under the ‘Near Mixed River’ and ‘Distant Natural River’ scene, while the smallest are under the ‘Near Natural Land’ and ‘Near Natural River’ scenes. In turn, the ‘Near Natural Land’ and ‘Near Natural River’ scenes generate the largest uniformity and ground contribution on the ceiling, as well as the smallest probability of discomfort glare.

Comparison of the emissive and transmissive approaches shows that the transmissive approach generally results in smaller values of space availability, relative to the emissive one. The actual light transmitted is smaller than the transmittance value of the glass; it largely depends on the colour of the view elements of the image scene. In turn, the average probability of discomfort glare using the transmissive approach is also smaller than that using the emissive approach.

These findings can give a first indication on what kind of performance future VNLS would bring. The use of emissive approach may introduce more light inside the space, but the view complexity is limited by the number of pixels or individual elements. If this limitation can be addressed in the future, then this approach will be the reasonable choice. On the other hand, the transmissive approach offers more flexibility to apply complex views on the display, but also requires more light, and thus consumes more energy to satisfy the lighting criteria.

Lastly, the use of computational modelling and simulation has the advantage of saving time in evaluating various input alternatives, which may include more complex situations where interactions can become hard to visually observe and theoretically forecast. As mentioned earlier, there is a long process in developing future solutions such as VNLS. The use of modelling and simulation is important in influencing the design decision; therefore it deserves to be discussed on its own.

Chapter 7

150

151

Chapter 8 Conclusions and Recommendations

This chapter summarises the general conclusions of every chapter of this thesis, highlights the main contribution, and gives recommendations for further research.

8.1. Conclusions

Despite the benefit it gives to human well-being, natural light entrance in buildings in practice has a limited availability, particularly in space and time. Examples of this situation can be found, among others, in cubicle workspaces in open-plan offices, operating rooms in hospitals, and control rooms in industrial plants. To solve the problem in building spaces without sufficient access to natural light, a number of efforts have been made to recreate the elements of natural light, in the form of artificial solutions. Such solutions, the so-called VNLS, can be generally classified based on their light and view qualities into four types, which are those providing: (1) simplified view and mainly diffuse light, (2) complex view and mainly diffuse light, (3) simplified view and mainly directional light, and (4) complex view and mainly directional light.

Subject-based experiments have been performed elsewhere using various form of VNLS prototypes, to gain knowledge on how people perceive it, and/or to investigate which aspects of natural light people appraise in reality. Previous research in the health-related area has suggested that artificial light and view from VNLS prototypes can have positive effects on subjects. Nonetheless, there is very little exploration on how the prototypes physically influence the indoor lighting condition of the space where they are installed, as this will also be related to the total performance of the building where they are located. This suggests a research direction on how to evaluate the objective performance of the prototypes, and how to propose better design solutions to improve it. To answer this challenge, this thesis aims to predict the impact of various VNLS applications on lighting performance and visual comfort in buildings, by means of computational modelling and simulation.

Based on a comparison of existing prototypes in Chapter 2, it is found that an ideal VNLS prototype does not yet exist at the moment. Most of the existing prototypes generate only diffuse light, and each prototype addresses only a subset of the required properties of an ideal VNLS. There are several properties that need to be improved, for instance, the information content on the view and the depth perception cues such as blur and motion parallax. Direction of further development should be therefore steered toward improving the light directionality and view dynamics, including dynamic elements, e.g. rustling leaves, running water, flying birds, and moving clouds.

Chapter 8

152

For the purpose of lighting modelling and simulation, Radiance has been validated elsewhere for daylight and electric lighting scenes, but not yet for VNLS scenes in particular. An example of the influence of simulation in VNLS development is shown in Chapter 3, where Radiance was applied to reproduce the scenes and to evaluate the lighting performance of a first generation VNLS prototype displaying a simplified view of overcast, clear, and partly cloudy skies, located in a test room with no façades. The key point of this chapter is to show that simulations can be used to compare an actual VNLS prototype with a hypothetical real window under the same sky settings, which was physically not possible. Based on the lighting simulation in Radiance, the investigated prototype performs better in terms of light distribution uniformity than a corresponding, hypothetical real window under the overcast and partly cloudy scenes. Under the clear sky scene, the difference between the real and virtual windows is less, due to the influence of direct sunlight. The findings demonstrate how a VNLS prototype compares to its real counterpart, in terms of illuminance distribution.

In terms of glare perception, it is generally unknown which glare metrics or ratings are most suited for the case of VNLS prototypes. In Chapter 4, a method is proposed to correlate the commonly applied glare metrics, i.e. DGP, DGI, UGR, and CGI, which can be predicted using simulations in Radiance and Evalglare, to the glare rating that was used in the experiment of Shin et al. (2012), in order to assess discomfort glare from a first generation VNLS prototype with various complex views. It was found the simulated values of normalised DGI, UGR, and CGI are all overestimated relative to the values converted from the experiment data of Shin et al., while the simulated values of DGP are in a better agreement with the converted values of DGP. Even though the accuracy of DGP has been widely reported, all of the earlier findings were based on real daylight scenes. The findings in this chapter demonstrate the applicability of DGP for the investigated VNLS prototype, and how it correlates to the subjective glare perception.

A second generation VNLS prototype has been designed and built by installing an array of LED tiles providing diffuse light and a view, and a line of LED linear fixtures with adjustable colour temperatures to provide direct light into the test room. This particular prototype has an important role in validating the computational model that can be extended for further development of future (not-yet-existing) VNLS. As intended, patches of direct light could be created and were visible on the side walls. Simulation and measurement values of horizontal illuminance at certain distances were evaluated and showed a good agreement.

Based on simulation of seven configurations of the second generation prototype with equal total opening size in the test room, it was found that nearly all configurations yield a space availability of 100%, taking a workplane illuminance of 200 lx as the criterion. When 300 lx is taken as the criterion, Configurations 2 (two openings on each short wall facing each other) and 5 (four openings on a long wall) yield space availabilities of more than 90%. When 500 lx is taken as the criterion, the configurations yield space availabilities between 25% and 50%. The maximum DGP values under all configurations range between 0.25 and 0.30.

Conclusions and Recommendations

153

Various operating schedules were defined and calculated for the second generation prototype, from 09.00 to 18.00 hrs local time, with one-hour time step on every working day. Based on the designated daily profiles (i.e. ‘spring’, ‘summer’, and ‘winter’), annual modes (i.e. mimicking and compensating the real seasons), and climate type scenarios (i.e. Singapore, Cairo, Sevilla, Amsterdam, and Chicago), the variation of average annual space availability within a given climate type was found to be very small. The values are however sensitive to the chosen criterion of workplane illuminance. The normalised, average annual electrical energy consumption in all climate types is within the range of 0.63 ~ 0.79, relative to the total electrical energy consumed by the prototype when it constantly displays the maximum setting at each working hour and on each working day in a year.

To further improve the performance of the solution, more complex VNLS configurations composed of small, light emitting sources have been developed and investigated in a computational model in Chapter 6. The model comprises of more than 600 small light emitting sources placed in rows, each of which has a different tilt angle. In this way, the model can display a simplified view of blue sky and green ground, while also delivering the light from the ‘ground’ to the ceiling and from the ‘sky’ to the floor. Sensitivity analysis shows that total luminous flux of the ‘sky’ significantly influences the space availability of the test room, whereas beam angle of the source largely influences the uniformity, ground contribution on the ceiling, and probability of discomfort glare.

To increase the view quality, a model of VNLS configurations with complex views has been created in Chapter 7, by pasting a two-dimensional image on a transparent glass in front of the light sources. The light was provided by the same configuration of light sources as used in Chapter 6, but all sources are white-coloured. Comparisons were shown between 10 image scenes. The use of the so-called transmissive approach offers more flexibility to apply complex display views, but also requires more light, and thus consumes more energy to satisfy the lighting criteria. On the other hand, the use of the emissive approach in Chapter 6 may introduce more light, but the view complexity is limited by the number of pixels. The more complex the view, the more individual pixels are required.

To conclude, the main contribution of this thesis is demonstrating the application of

computational modelling and building performance simulation in providing multiple

design concepts to improve the objective performance of VNLS. The defined objectives

have been addressed, and can be summarised as follows:

• Determining the relevant properties and performance indicators for VNLS. Chapter 2 is dedicated to give an overview of various properties of existing prototypes.

Light directionality and view dynamics are two properties that are the most complicated to feature. To indicate the benefit of VNLS in term of gaining more ‘daylit’ space, the space availability is determined as the main light quality performance indicator, and is used in evaluating all types of VNLS that are discussed in Chapters 3 until 7.

• Finding the appropriate modelling approach in order to model VNLS.

Chapter 8

154

Chapter 3 is opened by giving a brief introduction on lighting modelling and simulation, and Radiance is selected as the most appropriate tool. Chapter 4 demonstrates how Evalglare can be employed to predict glare metrics, in comparison to subjective glare ratings based on an experiment with subjects. Chapter 7 shows a comparison between two modelling approaches, i.e. an emissive and a transmissive approach for modelling a more complex VNLS. The transmissive approach has the advantage of creating more possibility to display complex views, whereas the emissive approach has the advantage of delivering more light into the space.

• Evaluating the lighting performance and visual comfort of various VNLS models. Chapters 3 and 4 discuss lighting performance and visual comfort evaluations of the first generation VNLS prototypes, whereas Chapter 5 discusses those of the second generation prototype that also serves for validation purpose. Chapters 6 and 7 describe the development and assessment of the next generation VNLS in computational models. Most of the modelled, future VNLS with a beam angle of 76° (medium spread) perform the closest to the corresponding real windows. The performance can be optimised by increasing the beam angle to 114° (wide spread), yielding a space availability of around two times larger than the corresponding real windows. The probability of discomfort glare from most of the models is relatively close to that from the corresponding real windows.

• Finding the potential of applying VNLS under various configurations by predicting their performance, and under various operating scenarios by estimating total annual electrical energy consumption.

Chapter 5 particularly addresses these objectives. Space availability under various configurations of the location of the prototype is evaluated using simulation in a rectangular test room, whereas the total annual electrical energy consumption is estimated based on the measured power of the prototype that has been built. The space availability can be optimised by placing the prototypes facing each other on both short walls, or placing all prototypes on a long wall. All of the investigated operating scenarios yield a relatively similar impact on the annual space availability and electrical energy consumption. However, the values are sensitive to the chosen criterion of workplane illuminance.

8.2. Recommendations

This thesis focuses on better understanding of how VNLS influence the indoor lighting performance and visual comfort, and how to design better solutions to improve the light and view qualities. For that purpose, most of the results in this thesis are based on physical measurement and/or computational modelling and simulation of VNLS prototypes and models. Room for improvement is mostly open in the topic of assessing subjective user’s perception on the particular prototypes and models. Even though there is already first supporting evidence on the positive effects of some existing (i.e. first generation) VNLS prototypes in Chapter 2, as the solutions become more complex and sophisticated, further

Conclusions and Recommendations

155

evaluation and analysis on how people actually appraise VNLS in reality is still necessary, before proceeding to the design implementation.

Recommendations for future research can be summarised as follows:

• In general, the first generation prototype with a simplified view is limited not only in its view complexity and information content, but also in the possibility to switch between views. To improve the degree of similarity to the real window, additional features such as motion parallax and sound transmission can be included. While these features are not directly influential to the indoor lighting performance, they can be helpful in giving the impression of having a real window inside the space.

• With regard to the subjective glare experiment of Shin et al. (2012), neither the simulated values nor the converted values of the glare metrics can be correlated with the actual percentage of subjects who felt disturbed. This finding suggests two things to improve; on the subjective experiment, giving fewer semantic scales for glare perception may be more efficient as it will reduce the chance of confusion and disagreement on the meaning itself. On the discomfort glare model development, a more appropriate method to approximate glare in a high contrast environment is required, for instance as proposed by Kleindienst & Andersen (2009) and Suk et al. (2013).

• It is noticed that the second generation prototype in Chapter 5 was limited in its display view resolution and luminous efficacy. By applying the latest lighting technology, it is possible to create a more detailed view on the display with higher light output, while consuming less energy. To introduce view dynamics, digital programming can be applied to automate the display variation at every given time step. The application of such dynamic lighting solutions is also in line with the roadmap of the European Commission (EC, 2013), which has put healthy and comfortable indoor environment (including air quality, ventilation, lighting, and acoustic) as one of its cross-platform target areas for 2020. Under this target, future research and innovation topics should be aimed at, among others, efficient and comfortable indoor lighting; for example by developing flexible lighting based on LEDs.

• The findings in Chapter 5 give a rough idea on the impact of varying operating schedules of the second generation prototype, but the results are based on the defined settings that are rather simplified, with one-hour time step in the daily profiles. Further research should be directed towards introducing more realistic display scenarios, for instance by introducing constantly changing weather conditions with a small time step, to closely follow the variation of real natural light.

• The next generation VNLS has been developed in computational models, using arrays of small light emitting areas. As the size of the individual light source in the model is only 0.05 m × 0.05 m, further models can incorporate more dimensions and shapes to optimise the output variables, as well as to allow possibility of displaying more detailed views.

• The use of an emissive approach to model VNLS with a complex view in Chapter 6 has the advantage of introducing more light inside the space, but the view complexity is

Chapter 8

156

limited by the number of pixels or individual elements. In line with the previous point, if future research can address this challenge, then the emissive approach will become the most reasonable option in creating efficient solutions. On the other hand, the transmissive approach offers more flexibility to apply complex views, but requires more light output to satisfy the lighting criteria. Further research on this approach can be directed towards finding the most suitable transmissive material for applying the two-dimensional image, while not losing too much light in the process. Alternatively, further research can be directed towards combining the two approaches, for example using the emissive one for the ‘ground’ and the transmissive one for the ‘sky’, or the emissive one for the lower layer and the transmissive one for the details.

• Finally, to predict the impact of VNLS on total building energy consumption, a detailed study on thermal properties of a VNLS prototype is required. Such information will be useful particularly as an input for building energy simulation tools, in determining the total amount of heating and cooling energy demand of the relevant space.

157

References

Ahn, B.-L., Jang, C.-Y., Leigh, S.-B., Yoo, S., & Jeong, H. (2014). Effect of LED lighting on the cooling and heating loads in office buildings. Applied Energy, 113, 1484-1489.

Aries, M. B. C., Veitch, J. A., & Newsham, G. (2010). Windows, view, and office characteristics predict physical and psychological discomfort. Journal of Environmental Psychology, 30, 533-541.

Aries, M. B. C., Aarts, M. P. J., & van Hoof, J. (2013). Daylight and health: A review of the evidence and consequences for the built environment. Lighting Research and Technology, 1-22, doi: 10.1177/1477153513509258.

Assem, E. O. & Al-Mumin, A. A. (2010). Code compliance of fully glazed tall office buildings in hot climate. Energy and Buildings, 42, 1100-1105.

Avery, D. H., Bolte, M. A., & Millet, M. (1992). Bright dawn simulation compared with bright morning light in the treatment of winter depression. Acta Psychiatrica Scandinavica, 85, 430-434.

Avery, D. H., Eder, D. N., Bolte, M. A., Hellekson, C. J., Dunner, D. L., Vitiello, M. V. et al. (2001). Dawn simulation and bright light in the treatment of SAD: a controlled study. Biological Psychiatry, 50, 205-216.

Badia, P., Myers, B., Boecker, M., Culpepper, J., & Harsh, J. R. (1991). Bright light effects on body temperature, alertness, EEG and behavior. Physiology and Behavior, 50, 583-588.

Berman, M. G., Jonides, J., & Kaplan, S. (2008). The cognitive benefits of interacting with nature. Psychological Science, 19, 1207-1212.

Blask, D. E. (2009). Melatonin, sleep disturbance and cancer risk. Sleep Medicine Reviews, 13, 257-264.

Beute, F. & de Kort, Y. A. W. (2013). Let the sun shine! Measuring explicit and implicit preference for environments differing in naturalness, weather type and lightness. Journal of Environmental Psychology, 36, 162-178.

Beute, F. & de Kort, Y. A. W. (2014). Salutogenic effects of the environment: Review of health protective effects of nature and daylight. Applied Psychology: Health And Well-Being, 6, 67–95.

Boubekri, M. (2004). An overview of the current state of daylight legislation. Journal of the Human-Environment System, 7, 57-63.

Boubekri, M. (2008). Daylighting, Architecture and Health: Building Design Strategies. Oxford: Architectural Press.

References

158

Bouma, P. J. (1948). Physical Aspects of Colour: An Introduction to the Scientific Study of Colour Stimuli and Colour Sensations. Eindhoven: Philips Gloeilampenfabrieken (Philips Industries) Technical and Scientific Literature Dept.

Boyce, P. R. (2003). Human Factors in Lighting. (2nd ed.) London: Taylor & Francis.

Boyce, P. R., Hunter, C., & Howlett, O. (2003). The Benefits of Daylight Through Windows Capturing the Daylight Dividend Program.

CABA (2008). Bright Green Buildings: Convergence of Green and Intelligent Buildings. Retrieved 27-1-2013, from http://www.caba.org/brightgreen

Cajochen, C. (2007). Alerting effects of light. Sleep Medicine Reviews, 11, 453-464.

CEN (2002). EN 12464-1: Light and Lighting - Lighting of work places, Part 1: Indoor work places. Brussels: Comité Européen de Normalisation.

Chain, C., Dumortier, D., & Fontoynont, M. (1999). A comprehensive model of luminance, colour temperature and spectral distribution of skylight – comparison with experimental data. Solar Energy, 65, 285-295.

Chain, C., Dumortier, D., & Fontoynont, M. (2001). Consideration of daylight's colour. Energy and Buildings, 33, 193-198.

Chang, C.-Y. & Chen, P.-K. (2005). Human response to window views and indoor plants in the workplace. HortScience, 40, 1354-1359.

CIE TC-3-33 (2005). Test cases to assess the accuracy of lighting computer programs Vienna: Commission Internationale de l'Eclairage.

Collins, B. L. (1975). Windows and people: a literature survey. Psychological reaction to environments with and without windows (Rep. No. NBS-BSS-70). National Bureau of Standards, Washington, D.C. (USA); U.S. Dept. of Commerce, National Bureau of Standards, Institute for Applied Technology, Washington, DC.

Collins, B. L. (1976). Review of the psychological reaction to windows. Lighting Research and Technology, 8, 80-88.

Crawley, D. B. (2008). Building Performance Simulation: A Tool for Policymaking. PhD thesis, University of Strathclyde, Glasgow.

Cutting, J. E. & Vishton, P. M. (1995). Perceiving layout: The integration, relative dominance, and contextual use of different information about depth. In W. Epstein & S. Rogers (Eds.), Handbook of Perception and Cognition: Vol. 5. Perception of Space and Motion (pp. 69-117). New York: Academic Press.

Davis, W. & Ohno, Y. (2009). Approaches to color rendering measurement. Journal of Modern Optics, 56, 1412-1419.

References

159

de Kort, Y. A. W., Meijnders, A. L., Sponselee, A.-M., & IJsselsteijn, W. A. (2006). What’s wrong with virtual trees? Restoring from stress in a mediated environment. Journal of Environmental Psychology, 26, 309-320.

de Vries, H. J. A., Aarts, M. P. J., Knoop, M., & Cornelissen, H. (2009). Beneficial non-visual effects of daylight, Research into the influential parameters. In Proceedings of Lux Europa 2009 - the 11th European Lighting Conference. Istanbul: Aydinlatma Türk Milli Komitesi.

Denissen, J. J. A., Butalid, L., Penke, L., & van Aken, M. A. G. (2008). The effects of weather on daily mood: A multilevel approach. Emotion, 8, 662-667.

Diepens, J. F. L., Bakker, F., and Zonneveldt, L. (2000). Daylight Design Variations Book. Retrieved 12-3-2012, from http://sts.bwk.tue.nl/daylight/varbook/index.htm

Duffy, F., Laing, A., & Crisp, V. (1993). The responsible workplace: The redesign of work and offices. Oxford: Butterworth Architecture.

Eagles, J. M. (1994). The relationship between mood and daily hours of sunlight in rapid cycling bipolar illness. Biological Psychiatry, 36, 422-424.

Eastman, C. I., Young, M. A., Fogg, L. F., Liu, L., & Meaden, P. M. (1998). Bright light treatment of winter depression – A placebo-controlled trial. Archive of General Psychiatry, 55, 883-889.

EC (2013). Energy-Efficient Buildings: Multi-annual roadmap for the contractual PPP under Horizon 2020. Luxembourg: European Commission.

Edwards, L. & Torcellini, P. (2002). A Literature Review of the Effects of Natural Light on Building Occupants (Rep. No. NREL/TP-550-30769). Golden: National Renewable Energy Laboratory.

Einhorn, H. D. (1979). Discomfort glare: a formula to bridge differences. Lighting Research and Technology, 11, 90-94.

EIZO (2013). Altering color dramatically with a single setting: Examining color temperature on an LCD monitor. Retrieved 06-01-2014, from http://www.eizo.com/global/library/basics/color_temperature_on_an_LCD_monitor/

Enrech Xena, C. (1999). Simulation de la lumiere naturelle par des sources artificielles: Enjeux architecturaux. PhD thesis, Université de Nantes, Nantes.

Farley, K. M. J. & Veitch, J. A. (2001). A Room with a View: A Review of the Effects of Window on Work and Well-being (Rep. No. RR-136). IRC Research.

Fisher, A. (1992). Tolerances in lighting design. In Proceedings of CIE Seminar on Computer Programs for Light and Lighting. Vienna: Commission Internationale de l'Eclairage.

Flynn, J. E., Hendrick, C., Spencer, T., & Martyniuk, O. (1979). A guide to methodology procedures for measuring subjective impressions in lighting. Journal of the Illuminating Engineering Society, 8, 95-110.

References

160

Fontoynont, M. (2011a). Benefits of quality lighting in hospitals. In Konference og Workshop - Lys som Forbreder Sundhed & Trivsel i Hospitalssektoren. Copenhagen.

Fontoynont, M. (2011b). Solid State Lighting and changed situation for lighting conditions. In The Future of Energy Efficient Lighting. Katrineholm.

Friedman, B., Freier, N. G., Kahn, J. P., Lin, P., & Sodeman, R. (2008). Office window of the future? – Field-based analyses of a new use of a large display. International Journal of Human-Computer Studies, 66, 452-465.

Galasiu, A. D. & Veitch, J. A. (2006). Occupant preferences and satisfaction with the luminous environment and control systems in daylit offices: a literature review. Energy and Buildings, 38, 728-736.

Gaver, W. W., Smets, G., & Overbeeke, K. (1995). A virtual window on media space. In Proceedings of CHI '95 - SIGCHI Conference on Human Factors in Computing Systems (pp. 257-264). New York: ACM Press/Addison-Wesley Publishing.

Geisler-Moroder, D. & Dur, A. (2010). Estimating Melatonin Suppression and Photosynthesis Activity in Real-World Scenes from Computer Generated Images Society for Imaging Science and Technology.

Glass, P., Avery, G. B., Subramanian, K. N. S., Keys, M. P., Sostek, A. M., & Friendly, D. S. (1985). Effect of Bright Light in the Hospital Nursery on the Incidence of Retinopathy of Prematurity. New England Journal of Medicine, 313, 401-404.

Glickman, G., Byrne, B., Pineda, C., Hauck, W. W., & Brainard, G. C. (2006). Light therapy for seasonal affective disorder with blue narrow-band light-emitting diodes (LEDs). Biological Psychiatry, 59, 502-507.

Hammad, F. & Abu-Hijleh, B. (2010). The energy savings potential of using dynamic external louvers in an office building. Energy and Buildings, 42, 1888-1895.

Härmä, M. I. & Ilmarinen, J. E. (1999). Towards the 24-hour society – new approaches for aging shift workers? Scandinavian Journal of Work, Environment & Health, 25, 610-615.

Harrold, R. (2003). IESNA Lighting Ready Reference: A Compendium of Materials from the IESNA. New York: Illuminating Engineering Society of North America.

Hartig, T., Evans, G. W., Jamner, L. D., Davis, D. S., & Gärling, T. (2003). Tracking restoration in natural and urban field settings. Journal of Environmental Psychology, 23, 109-123.

Heerwagen, J. H. (1990). The psychological aspects of windows and window design. In Annual Conference of the Environmental Design Research Association (pp. 269-280).

Heerwagen, J. H. & Orians, G. (1986). Adaptation to windowlessness: A study on the use of visual decor in windowed and windowless offices. Environment and Behavior, 18, 623-639.

References

161

Hellinga, H. Y. & de Bruijn-Hordijk, G. J. (2009). A new method for the analysis of daylight access and view out. In Proceedings of Lux Europa 2009 - the 11th European Lighting Conference. Istanbul: Aydinlatma Türk Milli Komitesi.

Heschong, L. (2003a). Window and Classrooms: A Study of Student Performance and the Indoor Environment (Rep. No. P500-03-082-A-7). California, USA: Heschong Mahone Group.

Heschong, L. (2003b). Window and Offices: A Study of Office Worker Performance and the Indoor Environment (Rep. No. P500-03-082-A-9). California, USA: Heschong Mahone Group.

Hopfe, C. J. (2009). Uncertainty and sensitivity analysis in building performance simulation for decision support and design optimization. PhD thesis, Eindhoven University of Technology, Eindhoven.

Hopkinson, R. G. (1972). Glare from daylighting in buildings. Applied Ergonomics, 3, 206-215.

IJsselsteijn, W. A., Oosting, W., Vogels, I., de Kort, Y. A. W., & van Loenen, E. J. (2008). A room with a cue: The efficacy of movement parallax, occlusion, and blur in creating a virtual window. PRESENCE, 17, 269-282.

Inanici, M. N. & Navvab, M. (2006). The Virtual Lighting Laboratory: Per-Pixel Luminance Data Anaysis. LEUKOS, 3, 89-104.

Iskra-Golec, I. M., Wazna, A., & Smith, L. (2012). Effects of blue-enriched light on the daily course of mood, sleepiness and light perception: A field experiment. Lighting Research and Technology, 44, 506-513.

Jacobs, A. (2012a). RADIANCE Cookbook. Retrieved 12-3-2012, from http://www.jaloxa.eu/resources/radiance/documentation/docs/radiance_cookbook.pdf

Jacobs, A. (2012b). Colour Picker for Radiance. Retrieved 2-4-2013, from http://www.jaloxa.eu/resources/radiance/colour_picker/index.shtml

Jakubiec, J. A. & Reinhart, C. F. (2012). The 'adaptive zone' – A concept for assessing discomfort glare throughout daylit spaces. Lighting Research and Technology, 44, 149-170.

Jenkins, D. & Newborough, M. (2007). An approach for estimating the carbon emissions associated with office lighting with a daylight contribution. Applied Energy, 84, 608-622.

Kahn, J. P., Friedman, B., Gill, B., Hagman, J., Severson, R. L., Freier, N. G. et al. (2008). A plasma display window? -- The shifting baseline problem in a technologically mediated natural world. Journal of Environmental Psychology, 28, 192-199.

Kaplan, R. (1993). The role of nature in the context of the workplace. Landscape and Urban Planning, 26, 193-201.

References

162

Kaplan, S. (1995). The restorative benefits of nature: Toward an integrative framework. Journal of Environmental Psychology, 15, 169-182.

Kaplan, R. & Kaplan, S. (1989). The Experience of Nature: A Psychological Perspective. Cambridge: Cambridge University Press.

Keighley, E. C. (1973a). Visual requirements and reduced fenestration in office buildings - A study of window shape. Building Science, 8, 311-320.

Keighley, E. C. (1973b). Visual requirements and reduced fenestration in offices - A study of multiple apertures and window area. Building Science, 8, 321-331.

Keller, M. C., Fredrickson, B. L., Côté, S., Johnson, K., Mikels, J., Conway, A. et al. (2005). A warm heart and a clear head: The contingent effects of weather on mood and cognition. Psychological Science, 16, 724-731.

Kim, J. & Wineman, K. (2005). Are Windows and View Really Better? A Quantitative Analysis of the Economic and Psychological Value of Views. New York: Lighting Research Center, Rensselaer Polytechnic Institute.

Kim, J. T., Shin, J. Y., & Yun, G. Y. (2012). Prediction of discomfort glares from windows: Influence of the subjective evaluation of window views. Indoor and Built Environment, 21, 92-97.

Kim, W., Ahn, H. T., & Kim, J. T. (2008). A first approach to discomfort glare in the presence of non-uniform luminance. Building and Environment, 43, 1953-1960.

Kim, W. & Kim, J. T. (2011). The scope of the glare light source of the window with non-uniform luminance distribution. Indoor and Built Environment, 20, 54-64.

Kleindienst, S. A. & Andersen, M. (2009). The adaptation of daylight glare probability to dynamic metrics in a computational setting. In Proceedings of Lux Europa 2009 - the 11th European Lighting Conference. Istanbul: Aydinlatma Türk Milli Komitesi.

Kosonen, R. & Tan, F. (2004). Assessment of productivity loss in air-conditioned buildings using PMV index. Energy and Buildings, 36, 987-993.

Lawrence Berkeley National Laboratory (LBL) (2010). Virtual Lighting Simulator. Retrieved 12-3-2012, from http://gaia.lbl.gov/vls/

Lindsey, J. L. (1997). Applied Illumination Engineering (2nd ed.). Lilburn: The Fairmont Press, Inc.

Lingjærde, O., Føreland, O., & Dankertsen, J. (1998). Dawn simulation vs. lightbox treatment in winter depression: a comparative study. Acta Psychiatrica Scandinavica, 98, 73-80.

Maamari, F., Fontoynont, M., & Adra, N. (2006). Application of the CIE test cases to assess the accuracy of lighting computer programs. Energy and Buildings, 38, 869-877.

References

163

Maamari, F., Fontoynont, M., Tsangrassoulis, A., Marty, C., Kopylov, E., & Sytnik, G. (2005). Reliable datasets for lighting programs validation--benchmark results. Solar Energy, 79, 213-215.

Mangkuto, R. A., Aries, M. B. C., van Loenen, E. J., & Hensen, J. L. M. (2011). Properties and performance indicators of Virtual Natural Lighting Solutions. In Proceedings of CISBAT 2011 (pp. 379-384). Lausanne: LESO-PB, École Polytechnique Fédérale de Lausanne.

Mangkuto, R. A., Aries, M. B. C., van Loenen, E. J., & Hensen, J. L. M. (2012). Lighting performance of Virtual Natural Lighting Solutions with a simplified image in a reference office space. In Proceedings of Experiencing Light 2012 (pp. 1-5). Eindhoven: Eindhoven University of Technology.

Mangkuto, R. A., Aries, M. B. C., van Loenen, E. J., & Hensen, J. L. M. (2014). Simulation of virtual natural lighting solutions with a simplified view. Lighting Research and Technology, 46, 198-218.

Mardaljevic, J. (1995). Validation of a lighting simulation program under real sky conditions. Lighting Research and Technology, 27, 181-188.

Mardaljevic, J. (1997). Validation of a lighting simulation program: A study using measured sky brightness distributions. In Proceedings of Lux Europa 1997 - the 8th European Lighting Conference (pp. 555-569). Amsterdam: Nederlandse Stichting voor Verlichtingskunde.

Markus, T. A. (1967). The function of windows – A reappraisal. Building Science, 2, 97-121.

Moeck, M. & Selkowitz, S. E. (1996). A computer-based daylight systems design tool. Automation in Construction, 5, 193-209.

Mottram, V., Middleton, B., Williams, P., & Arendt, J. (2011). The impact of bright artificial white and 'blue-enriched' light on sleep and circadian phase during the polar winter. Journal of Sleep Research, 20, 154-161.

Ochoa, C. E., Aries, M. B. C., & Hensen, J. L. M. (2012). State of the art in lighting simulation for building science: a literature review. Journal of Building Performance Simulation, 5, 209-233.

Ohno, Y. & Davis, W. (2010). Rationale of color quality scale. National Institute of Standards and Technology (NIST).

Pandharipande, A. & Caicedo, D. (2011). Daylight integrated illumination control of LED systems based on enhanced presence sensing. Energy and Buildings, 43, 944-950.

Peel, M. C., Finlayson, B. L., & McMahon, T. A. (2007). Updated world map of the Köppen-Geiger climate classification. Hydrology and Earth System Sciences, 4, 439-473.

References

164

Pelzers, R. S., Yu, Q. L., & Mangkuto, R. A. (2014). Radiation modeling of a photo-reactor using a backward ray-tracing method: an insight into indoor photocatalytic oxidation. Environmental Science and Pollution Research, 1-12, doi: 10.1007/s11356-014-2552-1.

Philips (2013a). Master TL5 HO 54W/850 UNP - Product catalog. Retrieved 13-1-2014, from http://www.ecat.lighting.philips.com/l/lamps/fluorescent-lamps/tl5/master-tl5-high-output/927929085018_eu/

Philips (2013b). Long Life IR PAR38 - Product catalog. Retrieved 15-1-2014, from http://www.ecat.lighting.philips.com/l/lamps/halogen-lamps/mv-halogen-with-reflector/long-life-ir-par38/924757119802_na/

Radikovic, A. S., Leggett, J. J., Keyser, J., & Ulrich, R. S. (2005). Artificial window view of nature. In Proceedings of CHI '05 (pp. 1993-1996). New York: ACM.

Rational Craft (2013). Winscape. Retrieved 18-11-2013, from http://rationalcraft.com/Winscape.html

Rea, M. S. & Freyssinier-Nova, J. P. (2008). Color rendering: A tale of two metrics. Color Research & Application, 33, 192-202.

Reinhart, C. F. & Herkel, S. (2000). The simulation of annual daylight illuminance distributions – a state-of-the-art comparison of six RADIANCE-based methods. Energy and Buildings, 32, 167-187.

Reinhart, C. F. & Walkenhorst, O. (2001). Validation of dynamic RADIANCE-based daylight simulations for a test office with external blinds. Energy and Buildings, 33, 683-697.

Reinhart, C. F. & Andersen, M. (2006). Development and validation of a Radiance model for a translucent panel. Energy and Buildings, 38, 890-904.

Reinhart, C. F. & Weismann, D. A. (2012). The daylit area – Correlating architectural student assessments with current and emerging daylight availability metrics. Building and Environment, 50, 155-164.

Rodriguez, R. G. & Pattini, A. (2014). Tolerance of discomfort glare from a large area source for work on a visual display. Lighting Research and Technology, 46, 157-170.

Ruppertsberg, A. & Bloj, M. (2006). Rendering complex scene for psychophysics using RADIANCE: How accurate can you get? Journal of the Optical Society of America A, 23, 759-768.

Saxon, R. (1994). The atrium comes of age. Harlow: Longman.

Shin, J. Y., Yun, G. Y., & Kim, J. T. (2012). View types and luminance effects on discomfort glare assessment from windows. Energy and Buildings, 46, 139-145.

Skåret, J. E. (1992). Indoor environment and economics (Project No. N 6405), Oslo: NBI-Byggforsk.

References

165

Slater, A. I. & Graves, H. (2002). Benchmarking Lighting Design Software (Rep. No. TM 28/00). London: CIBSE.

Slater, A. I., Perry, M. J., & Carter, D. J. (1993). Illuminance differences between desks: Limits of acceptability. Lighting Research and Technology, 25, 91-103.

Smolders, K. C. H. J. (2013). Daytime light exposure: Effects and preferences. PhD thesis, Eindhoven University of Technology, Eindhoven.

Smolders, K. C. H. J., de Kort, Y. A. W., & Cluitmans, P. J. M. (2012). A higher illuminance induces alertness even during office hours: Findings on subjective measures, task performance and heart rate measures. Physiology and Behavior, 107, 7-16.

Smolders, K. C. H. J., de Kort, Y. A. W., & van den Berg, S. (2013). Diurnal light exposure and feelings of vitality: Results of a field study during regular weekdays. Journal of Environmental Psychology, 36, 270-279.

Stevens, R. G. (2009). Working against our endogenous circadian clock: Breast cancer and electric lighting in the modern world. Journal of Mutation Research/Genetic Toxicology and Environmental Mutagenesis, 680, 106-108.

Stoelinga, P. (2005). Current practice of simulation in the building design process: a matter of communication. Delft: IBPSA NVL.

Suk, J. Y., Schiler, M., & Kensek, K. (2013). Development of new daylight glare analysis methodology using absolute glare factor and relative glare factor. Energy and Buildings, 64, 113-122.

Tennessen, C. M. & Cimprich, B. (1995). Views to nature: Effects on attention. Journal of Environmental Psychology, 15, 77-85.

TESS (2012). Therapeutic Environmental Solutions - Creating Healing Environments. Retrieved 18-11-2013, from http://www.tesspatientexperience.com/

Tregenza, P. R. (1989). Modification of the split-flux formulae for mean daylight factor and internal reflected component with large external obstructions. Lighting Research and Technology, 21, 125-128.

Tregenza, P. R. & Loe, D. (1998). The Design of Lighting. London: E & FN Spon.

Tuaycharoen, N. & Tregenza, P. R. (2005). Discomfort glare from interesting images. Lighting Research and Technology, 37, 329-341.

Tuaycharoen, N. & Tregenza, P. R. (2007). View and discomfort glare from windows. Lighting Research and Technology, 39, 185-200.

Ulrich, R. S. (1984). View through a window may influence recovery from surgery. Science, 224, 420-421.

Ulrich, R. S., Lundén, O., & Eltinge, J. L. (1993). Effects of exposure to nature and abstract pictures on patients recovering from heart surgery. Psychophysiology, 30, 7.

References

166

Ulrich, R. S., Simons, R. F., Losito, B. D., Fiorito, E., Miles, M. A., & Zelson, M. (1991). Stress recovery during exposure to natural and urban environments. Journal of Environmental Psychology, 11, 201-230.

US Department of Energy (2011). Weather Data - All Regions: Southwest Pacific WMO Region 5. Retrieved 7-3-2013, from http://apps1.eere.energy.gov/buildings/energyplus/cfm/weather_data2.cfm/region=5_southwest_pacific_wmo_region_5

van Dijk, D. & Platzer, W. J. (2003). Reference office for thermal, solar and lighting calculations (Rep. No. swift-wp3-tno-dvd-030416).

van Hoof, J., Westerlaken, A. C., Aarts, M. P. J., Wouters, E. J. M., Schoutens, A. M. C., Sinoo, M. M. et al. (2012). Light therapy: Methodological issues from an engineering perspective. Technology and Health Care, 20, 11-23.

van Loenen, E. J., Oosting, W., Vogels, I., Lashina, T., van Hasen, R., Staats, C. et al. (2007). Virtual Windows. In E. Aarts & E. Diederiks (Eds.), Ambient Lifestyle: From Concept to Experience (pp. 148-161). Amsterdam: BIS Publishers.

van Meel, J. (2000). The European Office: Office Design and National Context. Rotterdam: 010 Publishers.

Vandewalle, G., Maquet, P., & Dijk, D. J. (2009). Light as a modulator of cognitive brain function. Trends in Cognitive Sciences, 13, 429-438.

Viola, A. U., James, L. M., Schlangen, L. J. M., & Dijk, D. J. (2008). Blue-enriched white light in the workplace improves self-reported alertness, performance and sleep quality. Scandinavian Journal of Work, Environment & Health, 34, 297-306.

Ward, G. (1994). The RADIANCE lighting simulation and rendering system. In SIGGRAPH (pp. 459-472). Orlando: ACM SIGGRAPH.

Ward, G. & Shakespeare, R. A. (1998). Rendering with Radiance: The Art and Science of Lighting Visualization. San Francisco: Morgan Kaufman Publishers.

Webb, A. R. (2006). Considerations for lighting in the built environment: Non-visual effects of light. Energy and Buildings, 38, 721-727.

Wienold, J. & Christoffersen, J. (2006). Evaluation methods and development of a new glare prediction model for daylight environments with the use of CCD cameras. Energy and Buildings, 38, 743-757.

Windauga (2013). Windauga® - Products and services. Retrieved 18-11-2013, from http://www.windauga.com/english/products.htm

Woods, J. E. (1989). Cost avoidance and productivity in owning and operating buildings. Occupational Medicine, 4, 753-770.

References

167

Yun, G. Y., Hwang, T., & Kim, J. T. (2010). Performance prediction by modelling of a light-pipe system used under the climate conditions of Korea. Indoor and Built Environment, 19, 137-144.

Yun, G. Y., Shin, J. Y., & Kim, J. T. (2011). Influence of window views on the subjective evaluation of discomfort glare. Indoor and Built Environment, 20, 65-74.

Zukauskas, A., Vaicekauskas, R., Ivanauskas, F., Vaitkevicius, H., Vitta, P., & Shur, M. S. (2009). Statistical approach to color quality of solid-state lamps. IEEE Journal of Selected Topics in Quantum Electronics, 15, 1189-1198.

References

168

169

Appendix A

This appendix summarises the complete measurement and simulation results of horizontal illuminance values on the workplane in the kitchen test room as discussed in Chapter 3. Figure 3.3 is referred for the orientation; the lower left corner on the floor plan is defined as

the origin. Position x and y are both measured in metres, the workplane height is 0.92 m. Table A.1 shows the results under the overcast clear, and partly cloudy sky scenes of the VNLS prototype as well as the simulated real window.

Table A.1. Measurement and simulation results of horizontal illuminance values on the workplane in the kitchen test room with VNLS prototype and real window (RW)

x [m]

y [m]


VNLSEmea [lx]

VNLS Esim [lx]

RW Esim [lx]

VNLS Emea [lx]

VNLS Esim [lx]

RW Esim [lx]

VNLS Emea [lx]

VNLS Esim [lx]

RW Esim [lx]

0.2 0.5 34 22 10 43 45 57 60 54 33 0.7 0.5 34 24 10 44 37 64 65 59 35 1.2 0.5 35 25 8 45 47 68 66 60 35 1.7 0.5 34 25 10 44 43 66 65 61 36 2.2 0.5 32 24 10 41 47 66 63 60 32 2.7 0.5 31 23 8 36 44 61 55 54 38 3.2 0.5 28 19 9 31 37 58 50 48 33 3.7 0.5 24 12 8 26 31 50 43 30 37 0.2 1.0 30 22 10 38 40 61 52 50 36 0.7 1.0 35 22 11 42 54 67 62 61 39 1.2 1.0 38 26 12 45 50 76 72 62 43 1.7 1.0 36 28 12 47 48 71 65 65 43 2.2 1.0 37 25 12 47 52 72 68 67 39 2.7 1.0 33 21 12 41 49 63 57 63 38 3.2 1.0 29 18 11 32 40 60 51 46 37 3.7 1.0 25 15 13 27 30 57 41 37 41 0.2 1.5 28 26 11 40 49 66 51 62 42 0.7 1.5 38 33 13 47 59 80 68 75 50 1.2 1.5 44 38 15 51 68 88 85 80 51 1.7 1.5 44 38 15 56 65 87 80 81 55 2.2 1.5 48 31 15 55 59 82 79 80 49 2.7 1.5 37 26 13 48 52 74 65 66 45 3.2 1.5 31 21 12 33 42 62 54 53 39 3.7 1.5 25 16 11 28 28 58 41 38 37

Appendices

170

Table A.1. (continued)

x [m]

y [m]


VNLSEmea [lx]

VNLS Esim [lx]

RW Esim [lx]

VNLS Emea [lx]

VNLS Esim [lx]

RW Esim [lx]

VNLS Emea [lx]

VNLS Esim [lx]

RW Esim [lx]

0.2 2.0 39 31 21 54 60 106 70 80 72 0.7 2.0 47 43 27 59 81 121 86 104 86 1.2 2.0 57 51 41 74 102 132 130 109 108 1.7 2.0 61 52 45 102 104 137 149 132 116 2.2 2.0 76 46 36 67 82 106 107 94 86 2.7 2.0 45 33 24 57 66 84 87 71 61 3.2 2.0 32 18 18 27 40 69 56 49 47 3.7 2.0 23 15 11 26 29 52 40 34 44 0.2 2.5 42 36 13 61 72 74 81 85 51 0.7 2.5 61 50 40 85 99 136 110 119 126 1.2 2.5 103 76 81 109 136 207 174 166 237 1.7 2.5 88 79 105 129 139 193 190 187 227 2.2 2.5 76 68 64 129 125 137 191 146 139 2.7 2.5 52 40 31 73 76 88 116 91 74 3.2 2.5 32 23 16 38 42 62 58 52 47 3.7 2.5 20 13 14 23 18 52 35 22 40 0.2 3.0 23 19 39 39 37 153 53 46 153 0.7 3.0 63 61 125 94 94 367 119 125 407 1.2 3.0 135 119 230 156 209 376 241 256 500 1.7 3.0 155 155 93 236 253 165 355 346 165 2.2 3.0 142 104 23 176 181 89 256 220 69 2.7 3.0 62 43 11 76 76 60 109 90 44 3.2 3.0 30 20 10 39 37 46 53 44 36 3.7 3.0 17 9 21 23 15 106 31 21 72 0.7 3.5 48 43 27 86 87 121 120 115 86 1.2 3.5 115 127 41 215 260 132 287 316 108 1.7 3.5 204 234 45 354 432 137 491 533 116 2.2 3.5 118 103 36 151 200 106 237 242 86 2.7 3.5 40 26 24 52 42 84 72 69 61 3.2 3.5 21 12 18 36 23 69 39 29 47 3.7 3.5 15 8 11 20 16 52 27 19 44

171

Appendix B

This appendix gives the high dynamic range (HDR) and luminance false colour pictures of the VNLS prototype in the kitchen test room under the three sky scenes in Chapter 3, as seen from position 2, shown in Figures B.1 and B.2. Figure 3.3 is referred for the orientation. Note there are different scales used in the three pictures in Figure B.2.

(a) (b)

(c)

Figure B.1. High dynamic range pictures of the prototype under the (a) overcast, (b) clear, and (c) partly cloudy sky scenes; as seen from position 2

Appendices

172

(a) (b)

(c)

Figure B.2. Luminance false colour pictures of the prototype observed at position 2, under (a) overcast, (b) clear, and (c) partly cloudy sky scene

173

Appendix C

This appendix summarises the mean subjective glare ratings, converted and normalised simulation values of glare metrics from the first generation VNLS prototype with complex views (Shin et al., 2012) as discussed in Chapter 4. Table C.1 shows the results under the five mean view luminance and 10 image scenes.

Table C.1. Mean subjective glare ratings, converted and normalised simulated values of glare metrics from the VNLS prototype in the experiment of Shin et al. (2012)

Image Lav [cd/m2]

Shin’s rating

[-]

DGP conv[-]

DGPsim [-]

DGIn conv [-]

DGIn

sim [-]

CGIn conv [-]

CGIn

sim

[-]

UGR conv

[-]

UGRn

sim [-]

DMM 1000 1.1 0.24 0.26 0.20 0.34 0.16 0.45 0.16 0.48 1800 1.3 0.26 0.28 0.22 0.36 0.18 0.50 0.18 0.51 3200 1.6 0.29 0.31 0.25 0.38 0.21 0.55 0.21 0.54 5600 2.5 0.33 0.36 0.31 0.40 0.28 0.60 0.28 0.57

10000 3.4 0.36 0.46 0.35 0.43 0.35 0.66 0.35 0.61

NMM 1000 1.3 0.26 0.26 0.22 0.35 0.18 0.46 0.18 0.50 1800 1.6 0.29 0.28 0.25 0.37 0.21 0.51 0.21 0.53 3200 2.1 0.32 0.31 0.29 0.39 0.25 0.56 0.25 0.56 5600 2.8 0.34 0.37 0.32 0.41 0.30 0.61 0.30 0.59

10000 3.5 0.36 0.46 0.36 0.43 0.35 0.67 0.35 0.63

DNL 1000 1.2 0.25 0.26 0.21 0.35 0.17 0.46 0.17 0.51 1800 1.6 0.29 0.28 0.25 0.37 0.21 0.50 0.21 0.54 3200 2.3 0.32 0.31 0.30 0.39 0.27 0.56 0.27 0.57 5600 2.9 0.34 0.36 0.33 0.41 0.31 0.61 0.31 0.60

10000 3.5 0.36 0.46 0.36 0.44 0.35 0.67 0.35 0.63

NNL 1000 1.3 0.26 0.26 0.22 0.35 0.18 0.46 0.18 0.49 1800 1.8 0.30 0.28 0.27 0.37 0.23 0.50 0.23 0.53 3200 2.3 0.32 0.31 0.30 0.39 0.27 0.55 0.27 0.56 5600 3.1 0.35 0.36 0.33 0.41 0.32 0.61 0.32 0.59

10000 3.6 0.37 0.46 0.36 0.43 0.36 0.67 0.36 0.62

DML 1000 1.2 0.25 0.25 0.21 0.33 0.17 0.45 0.17 0.47 1800 1.4 0.27 0.27 0.23 0.36 0.19 0.49 0.19 0.51 3200 1.7 0.29 0.31 0.26 0.38 0.22 0.54 0.22 0.54 5600 2.7 0.33 0.36 0.32 0.40 0.30 0.60 0.30 0.57

10000 3.5 0.36 0.46 0.36 0.42 0.35 0.66 0.35 0.60

NML 1000 1.2 0.25 0.26 0.21 0.33 0.17 0.45 0.17 0.48 1800 1.5 0.28 0.28 0.24 0.36 0.20 0.50 0.20 0.51 3200 2.4 0.33 0.31 0.30 0.38 0.27 0.55 0.27 0.55 5600 3.0 0.34 0.36 0.33 0.40 0.32 0.60 0.32 0.58

10000 3.7 0.37 0.45 0.37 0.42 0.37 0.66 0.37 0.61

Appendices

174

Table C.1. (continued)

Image Lav [cd/m2]

Shin’s rating

[-]

DGP conv

[-]

DGPsim [-]

DGIn conv [-]

DGIn

sim [-]

CGIn conv [-]

CGIn

sim

[-]

UGR conv [-]

UGRn

sim [-]

DNR 1000 1.1 0.24 0.25 0.20 0.33 0.16 0.44 0.16 0.47 1800 1.5 0.28 0.27 0.24 0.36 0.20 0.49 0.20 0.50 3200 2.0 0.31 0.31 0.28 0.38 0.24 0.54 0.24 0.53 5600 2.8 0.34 0.36 0.32 0.40 0.30 0.60 0.30 0.56 10000 3.8 0.38 0.45 0.38 0.42 0.38 0.66 0.38 0.60

NNR 1000 1.2 0.25 0.26 0.21 0.34 0.17 0.46 0.17 0.48 1800 2.0 0.31 0.28 0.28 0.36 0.24 0.50 0.24 0.51 3200 2.5 0.33 0.31 0.31 0.38 0.28 0.55 0.28 0.54 5600 3.3 0.35 0.37 0.34 0.40 0.34 0.61 0.34 0.57 10000 3.9 0.39 0.46 0.38 0.43 0.39 0.67 0.39 0.61

DMR 1000 1.2 0.25 0.25 0.21 0.34 0.17 0.44 0.17 0.48 1800 2.0 0.31 0.27 0.28 0.36 0.24 0.49 0.24 0.51 3200 2.2 0.32 0.30 0.29 0.38 0.26 0.54 0.26 0.54 5600 3.0 0.34 0.35 0.33 0.40 0.32 0.59 0.32 0.57 10000 3.7 0.37 0.45 0.37 0.42 0.37 0.65 0.37 0.61

NMR 1000 1.2 0.25 0.26 0.21 0.35 0.17 0.45 0.17 0.52 1800 1.4 0.27 0.28 0.23 0.37 0.19 0.50 0.19 0.55 3200 1.8 0.30 0.31 0.27 0.40 0.23 0.55 0.23 0.59 5600 2.7 0.33 0.36 0.32 0.42 0.30 0.61 0.30 0.62 10000 3.6 0.37 0.45 0.36 0.44 0.36 0.66 0.36 0.65

175

Appendix D

This appendix summarises the complete measurement and simulation results of horizontal illuminance values on the workplane in the test room as discussed in Chapter 5. Figure 5.10 is referred for the orientation; the upper right corner on the floor plan is defined as the origin.

Position x and y are both measured in metres, the workplane height is 0.75 m. Table D.1 shows the results under the 25%, 62.5%, and 100% of the maximum setting of the VNLS prototype.

Table D.1. Measurement and simulation results of horizontal illuminance values on the workplane in the test room with the prototype under 25%, 62.5% and 100% maximum setting

x [m]

y [m]

25% 62.5% 100%

Emea

[lx]Esim

[lx]Emea

[lx]Esim

[lx]Emea

[lx]Esim

[lx] 2.3 0.375 66 70 193 172 343 336 2.8 0.375 80 82 223 212 382 379 3.3 0.375 69 72 197 186 343 347 3.8 0.375 56 59 161 149 284 280 4.3 0.375 45 46 129 118 227 216 4.8 0.375 36 37 104 91 187 169 5.2 0.375 33 32 86 77 154 140 5.7 0.375 27 25 72 61 127 120 6.2 0.375 24 20 61 53 107 91 6.7 0.375 21 17 58 40 95 80 7.1 0.375 19 15 47 36 85 73 7.6 0.375 19 14 47 34 76 64 8.1 0.375 18 13 47 32 73 63 2.3 0.875 127 133 351 331 632 617 2.8 0.875 106 117 302 295 524 548 3.3 0.875 82 90 230 227 404 432 3.8 0.875 63 69 182 174 317 319 4.3 0.875 48 51 142 129 241 241 4.8 0.875 39 40 111 105 196 190 5.2 0.875 34 33 90 84 160 160 5.7 0.875 29 27 76 66 132 123 6.2 0.875 25 22 64 53 111 105 6.7 0.875 22 16 56 44 98 82 7.1 0.875 19 16 50 40 87 67 7.6 0.875 19 14 48 36 81 69 8.1 0.875 18 14 48 34 78 67

Appendices

176

Table D.1. (continued)

x [m]

y [m]

25% 62.5% 100%

Emea

[lx]Esim

[lx]Emea

[lx]Esim

[lx]Emea

[lx]Esim

[lx] 2.3 1.375 171 180 470 444 866 834 2.8 1.375 130 145 362 354 637 677 3.3 1.375 93 104 261 258 460 496 3.8 1.375 69 77 196 188 340 366 4.3 1.375 51 57 152 139 249 276 4.8 1.375 40 43 116 108 203 203 5.2 1.375 35 35 93 84 163 168 5.7 1.375 29 27 78 70 135 127 6.2 1.375 25 22 65 54 111 104 6.7 1.375 22 19 57 43 100 84 7.1 1.375 20 16 50 40 87 78 7.6 1.375 20 14 49 37 81 71 8.1 1.375 19 14 48 35 78 68 2.3 1.875 165 174 459 441 858 842 2.8 1.875 137 146 382 364 677 701 3.3 1.875 97 108 272 268 480 516 3.8 1.875 71 80 203 197 350 373 4.3 1.875 52 58 155 145 254 267 4.8 1.875 41 45 119 112 204 205 5.2 1.875 36 36 94 91 163 170 5.7 1.875 29 28 79 69 135 136 6.2 1.875 25 22 67 57 112 109 6.7 1.875 22 18 57 46 98 86 7.1 1.875 20 15 51 41 88 76 7.6 1.875 20 15 49 35 82 70 8.1 1.875 19 14 47 35 78 67 2.3 2.375 175 170 484 431 903 820 2.8 2.375 132 140 366 349 650 649 3.3 2.375 92 102 258 257 458 478 3.8 2.375 68 75 196 179 337 349 4.3 2.375 51 58 150 135 248 261 4.8 2.375 40 44 118 103 203 200 5.2 2.375 35 34 93 89 159 172 5.7 2.375 29 27 79 67 134 132 6.2 2.375 25 23 66 53 110 90 6.7 2.375 22 17 57 47 98 87 7.1 2.375 20 16 51 41 86 82 7.6 2.375 20 14 47 36 82 70 8.1 2.375 19 14 47 35 79 68

Appendices

177

Table D.1. (continued)

x [m]

y [m]

25% 62.5% 100%

Emea

[lx]Esim

[lx]Emea

[lx]Esim

[lx]Emea

[lx]Esim

[lx]

2.3 2.875 132 131 364 334 670 628 2.8 2.875 105 116 304 284 518 552 3.3 2.875 81 88 229 223 398 411 3.8 2.875 62 69 181 173 304 330 4.3 2.875 48 54 143 134 234 247 4.8 2.875 39 41 113 102 291 191 5.2 2.875 35 34 91 86 153 153 5.7 2.875 29 26 77 65 129 125 6.2 2.875 25 22 66 55 109 102 6.7 2.875 22 17 57 45 95 86 7.1 2.875 20 17 50 40 85 76 7.6 2.875 20 14 49 34 81 64 8.1 2.875 19 14 46 36 77 67 2.3 3.375 66 73 191 182 337 344 2.8 3.375 77 82 214 204 378 384 3.3 3.375 68 74 194 179 329 347 3.8 3.375 56 60 162 148 273 282 4.3 3.375 46 47 133 116 219 223 4.8 3.375 37 37 107 95 178 179 5.2 3.375 34 32 88 79 146 145 5.7 3.375 28 25 75 62 124 119 6.2 3.375 24 20 63 50 105 96 6.7 3.375 21 16 55 42 91 79 7.1 3.375 19 15 49 37 83 71 7.6 3.375 19 14 45 33 76 66 8.1 3.375 18 13 47 33 73 65

178

Appendix E

This appendix gives the false colour maps of horizontal illuminance values on the workplane, under Configurations 1, 4, and 6 of the prototype in the test room as discussed in Chapter 5, shown in Figure E.1.

(a) (b)

(c)

Figure E.1. False colour maps of the simulated horizontal illuminance [lx] under Configurations (a) 1, (b) 4, and (c) 6

179

Appendix F

This appendix summarises the monthly average space availability for 300 lx criterion and total electrical energy consumption for each climate type, annual mode, and scenario of the display of the second generation prototype in Configuration 2; as described in Chapter 5. The monthly values of average space availability are averaged for the entire year, whereas those of total electrical energy consumption are summed up. Table F.1 gives the complete values. The light gray-, white-, and dark gray-coloured cells respectively correspond to the ‘spring’, ‘summer’, and ‘winter’ months.

Table F.1. Estimated average monthly space availability (%A) for 300 lx criterion and total monthly electrical energy consumption (Wmonth) in Configuration 2 for each climate type, annual mode, and scenario

Location: Mimicking Compensating Singapore %A [%] Wmonth [kWh] %A [%] Wmonth [kWh]

Month Days Sc1 Sc2 Sc3 Sc1 Sc2 Sc3 Sc1 Sc2 Sc3 Sc1 Sc2 Sc3

Jan 25 61 61 61 241 241 241 61 61 61 241 241 241

Feb 20 51 51 51 182 182 182 49 49 49 179 179 179

Mar 20 51 51 51 182 182 182 49 49 49 179 179 179

Apr 20 51 51 51 182 182 182 49 49 49 179 179 179

May 25 61 51 51 241 228 228 61 49 49 241 223 223

Jun 20 61 61 61 192 192 192 61 61 61 192 192 192

Jul 25 61 61 51 241 241 228 61 61 49 241 241 223

Aug 20 61 61 61 192 192 192 61 61 61 192 192 192

Sep 20 61 61 61 192 192 192 61 61 61 192 192 192

Oct 25 61 61 61 235 235 235 61 61 61 235 235 235

Nov 20 49 49 49 179 179 179 51 51 51 182 182 182

Dec 20 49 49 49 121 121 121 51 51 51 182 182 182

260 57 56 55 2380 2367 2354 57 56 55 2434 2417 2400

Appendices

180

Table F.1. (continued)

Location: Mimicking Compensating Cairo %A [%] Wmonth [kWh] %A [%] Wmonth [kWh]


Jan 25 52 52 52 212 212 212 3 3 3 131 131 131

Feb 20 40 52 40 164 169 164 40 3 40 164 105 164

Mar 20 40 40 40 164 164 164 40 40 40 164 164 164

Apr 20 40 40 3 164 164 105 40 40 52 164 164 169

May 25 3 3 3 131 131 131 52 52 52 212 212 212

Jun 20 3 3 3 105 105 105 52 52 52 169 169 169

Jul 25 3 3 3 131 131 131 52 52 52 212 212 212

Aug 20 40 3 3 164 105 105 40 52 52 164 169 169

Sep 20 40 40 3 164 164 105 40 40 52 164 164 169

Oct 25 40 40 40 205 205 205 40 40 40 205 205 205

Nov 20 52 52 52 169 169 169 3 3 3 105 105 105

Dec 20 52 52 52 169 169 169 3 3 3 105 105 105

260 33 31 25 1945 1891 1767 34 32 37 1961 1907 1977

Location: Mimicking Compensating Amsterdam %A [%] Wmonth [kWh] %A [%] Wmonth [kWh]


Jan 25 1 1 1 103 103 103 52 52 52 232 232 232

Feb 20 1 53 53 82 182 182 52 53 53 185 182 182

Mar 20 53 53 53 182 182 182 53 53 53 182 182 182

Apr 20 53 53 53 182 182 182 53 53 53 182 182 182

May 25 52 52 52 232 232 232 1 1 1 103 103 103

Jun 20 52 52 52 185 185 185 1 1 1 82 82 82

Jul 25 52 52 52 232 232 232 1 1 1 103 103 103

Aug 20 53 53 52 182 182 185 53 53 1 182 182 82

Sep 20 53 53 53 182 182 182 53 53 53 182 182 182

Oct 25 53 53 53 227 227 227 53 53 53 227 227 227

Nov 20 1 1 1 82 82 82 52 52 52 185 185 185

Dec 20 1 1 1 82 82 82 52 52 52 185 185 185

260 36 40 40 1952 2051 2055 39 39 35 2029 2026 1926

Appendices

181

Table F.1. (continued)

Location: Mimicking Compensating Sevilla %A [%] Wmonth [kWh] %A [%] Wmonth [kWh]


Jan 25 47 47 47 210 210 210 15 15 15 162 162 162

Feb 20 63 47 63 195 168 195 63 15 63 195 130 195

Mar 20 63 63 63 195 195 195 63 63 63 195 195 195

Apr 20 63 63 63 195 195 195 63 63 63 195 195 195

May 25 15 15 15 162 162 162 47 47 47 210 210 210

Jun 20 15 15 15 130 130 130 47 47 47 168 168 168

Jul 25 15 15 15 162 162 162 47 47 47 210 210 210

Aug 20 63 15 15 195 130 130 63 47 47 195 168 168

Sep 20 63 63 63 195 195 195 63 63 63 195 195 195

Oct 25 63 63 63 243 243 243 63 63 63 243 243 243

Nov 20 47 47 47 168 168 168 15 15 15 130 130 130

Dec 20 47 47 47 168 168 168 15 15 15 130 130 130

260 46 41 42 2218 2127 2153 47 42 45 2228 2136 2201

Location: Mimicking Compensating Chicago %A [%] Wmonth [kWh] %A [%] Wmonth [kWh]


Jan 25 17 17 17 151 151 151 31 31 31 201 201 201

Feb 20 61 17 61 188 121 188 61 31 61 188 161 188

Mar 20 61 61 61 188 188 188 61 61 61 188 188 188

Apr 20 61 61 61 188 188 188 61 61 61 188 188 188

May 25 31 31 31 201 201 201 17 17 17 151 151 151

Jun 20 31 31 31 161 161 161 17 17 17 121 121 121

Jul 25 31 31 31 201 201 201 17 17 17 151 151 151

Aug 20 61 31 31 188 161 161 61 17 17 188 121 121

Sep 20 61 61 61 188 188 188 61 61 61 188 188 188

Oct 25 61 61 61 235 235 235 61 61 61 235 235 235

Nov 20 17 17 17 121 121 121 31 31 31 161 161 161

Dec 20 17 17 17 121 121 121 31 31 31 161 161 161

260 42 36 39 2131 2037 2104 42 36 38 2121 2027 2054

182

Appendix G

This appendix summarises the space availability (%A), uniformity (U0), ground contribution


configurations with total luminous flux (Φ) of 6200 and 19900 lm; as described in Chapter 6. Note that configurations with the same distance between windows, beam angle, and total luminous flux are compared to the same reference real window, of which the performance indicators are shown directly above them in Table G.1.

Table G.1. Summary of space availability, uniformity, average ground contribution, and probability of discomfort glare for all variations and position C in both VNLS and real windows (RW) scenes with total luminous flux of 6200 and 19900 lm

Conf. IA [°] BA [°] Φ [lm] %A [%]

U0

[-] %Gav

[%] PDGav

[-] 1a RW – 5600 cd/m2 48 0.19 50 0.39 1a 2.0 38 6200 16 0.21 60 0.40 1a 1.5 38 6200 13 0.23 58 0.42 1a 1.0 38 6200 8 0.26 55 0.43

1a RW – 1800 cd/m2 14 0.18 50 0.34 1a 2.0 76 6200 12 0.28 50 0.36 1a 1.5 76 6200 11 0.29 47 0.36 1a 1.0 76 6200 9 0.32 44 0.36

1a RW – 1000 cd/m2 6 0.18 50 0.31 1a 2.0 114 6200 9 0.37 49 0.32

1a 1.5 114 6200 9 0.37 47 0.32 1a 1.0 114 6200 9 0.37 44 0.32

2a RW – 5600 cd/m2 43 0.16 49 0.39 2a 2.0 38 6200 15 0.23 59 0.41 2a 1.5 38 6200 9 0.24 57 0.43 2a 1.0 38 6200 1 0.28 54 0.44

2a RW – 1800 cd/m2 15 0.16 49 0.35 2a 2.0 76 6200 12 0.32 49 0.36 2a 1.5 76 6200 9 0.33 47 0.36 2a 1.0 76 6200 5 0.35 44 0.36

2a RW – 1000 cd/m2 3 0.15 48 0.32 2a 2.0 114 6200 8 0.36 48 0.32 2a 1.5 114 6200 8 0.37 46 0.32 2a 1.0 114 6200 8 0.37 45 0.32

Appendices

183

Table G.1. (continued)

Conf. IA [°] BA [°] Φ [lm] %A [%]

U0

[-] %Gav

[%] PDGav

[-] 1a RW – 18000 cd/m2 100 0.18 51 0.46 1a 2.0 38 19900 72 0.21 60 0.46 1a 1.5 38 19900 79 0.23 58 0.49 1a 1.0 38 19900 76 0.26 55 0.50

1a RW – 5600 cd/m2 48 0.19 50 0.39 1a 2.0 76 19900 70 0.28 50 0.41 1a 1.5 76 19900 74 0.30 47 0.42 1a 1.0 76 19900 80 0.32 44 0.42

1a RW – 3200 cd/m2 27 0.16 51 0.36 1a 2.0 114 19900 60 0.37 49 0.37

1a 1.5 114 19900 62 0.37 47 0.38 1a 1.0 114 19900 63 0.38 44 0.37

2a RW – 18000 cd/m2 89 0.15 49 0.46 2a 2.0 38 19900 79 0.23 59 0.48 2a 1.5 38 19900 84 0.24 57 0.50 2a 1.0 38 19900 81 0.28 54 0.51

2a RW – 5600 cd/m2 43 0.16 49 0.39 2a 2.0 76 19900 66 0.32 49 0.42 2a 1.5 76 19900 71 0.33 47 0.42 2a 1.0 76 19900 74 0.35 44 0.42

2a RW – 3200 cd/m2 30 0.15 50 0.37 2a 2.0 114 19900 56 0.36 48 0.38 2a 1.5 114 19900 58 0.36 46 0.38 2a 1.0 114 19900 60 0.37 45 0.38

184

Appendix H

This appendix summarises the space availability (%A), uniformity (U0), ground contribution


configurations with total luminous flux (Φ) of 6200, 11100, and 19900 lm, and with 10 image scenes; as described in Chapter 7. Table H.1 gives the complete values.

Table H.1. Summary of space availability, uniformity, average ground contribution, and probability of discomfort glare for all variations and position C in VNLS scenes with total luminous flux of 6200, 11100, and 19900 lm, and with 10 image scenes

Scene BA [°] Φ [lm] %A [%]

U0

[-] %Gav

[%] PDGav

[-] DML 38 6200 6 0.14 41 0.35

38 11100 15 0.14 40 0.37 38 19900 27 0.14 40 0.40 76 6200 0 0.23 30 0.31 76 11100 9 0.23 30 0.33 76 19900 23 0.23 30 0.36 114 6200 0 0.34 26 0.27 114 11100 4 0.34 26 0.30 114 19900 18 0.34 26 0.32

NML 38 6200 2 0.12 30 0.33 38 11100 10 0.12 29 0.35 38 19900 21 0.12 30 0.38 76 6200 0 0.22 21 0.30 76 11100 3 0.22 21 0.32 76 19900 14 0.22 21 0.35 114 6200 0 0.32 17 0.26 114 11100 0 0.32 17 0.28 114 19900 8 0.32 17 0.31

DMM 38 6200 4 0.15 44 0.35 38 11100 11 0.15 44 0.38 38 19900 22 0.15 44 0.41 76 6200 0 0.25 33 0.30 76 11100 6 0.25 33 0.33 76 19900 18 0.25 33 0.35 114 6200 0 0.35 30 0.27 114 11100 3 0.35 30 0.29 114 19900 14 0.34 30 0.31

Appendices

185

Table H.1. (continued)


U0

[-] %Gav

[%] PDGav

[-] NMM 38 6200 1 0.19 55 0.35

38 11100 9 0.19 55 0.38 38 19900 20 0.19 55 0.40 76 6200 0 0.28 44 0.30 76 11100 5 0.28 44 0.32 76 19900 16 0.28 44 0.35 114 6200 0 0.38 42 0.26 114 11100 2 0.37 42 0.28 114 19900 12 0.37 42 0.31

DMR 38 6200 3 0.16 43 0.37 38 11100 12 0.16 43 0.40 38 19900 25 0.16 43 0.43 76 6200 1 0.25 31 0.31 76 11100 8 0.25 31 0.34 76 19900 22 0.25 31 0.36 114 6200 0 0.34 29 0.27 114 11100 6 0.34 29 0.29 114 19900 18 0.35 28 0.32

NMR 38 6200 6 0.15 44 0.37 38 11100 17 0.15 44 0.40 38 19900 31 0.15 44 0.42 76 6200 2 0.25 33 0.32 76 11100 11 0.24 33 0.35 76 19900 27 0.25 33 0.38 114 6200 0 0.35 29 0.28 114 11100 7 0.35 29 0.31 114 19900 23 0.34 29 0.33

DNL 38 6200 4 0.19 60 0.33 38 11100 14 0.18 60 0.36 38 19900 27 0.19 60 0.38 76 6200 0 0.28 49 0.30 76 11100 7 0.28 49 0.33 76 19900 23 0.29 49 0.35 114 6200 0 0.39 45 0.27 114 11100 2 0.39 45 0.29 114 19900 18 0.38 45 0.32

Appendices

186

Table H.1. (continued)


U0

[-] %Gav

[%] PDGav

[-] NNL 38 6200 0 0.29 73 0.30

38 11100 0 0.29 73 0.33 38 19900 3 0.29 73 0.35 76 6200 0 0.38 65 0.24 76 11100 0 0.38 65 0.27 76 19900 1 0.38 65 0.29 114 6200 0 0.44 64 0.22 114 11100 0 0.44 64 0.24 114 19900 0 0.44 64 0.26

DNR 38 6200 5 0.14 34 0.37 38 11100 14 0.14 34 0.40 38 19900 28 0.13 34 0.43 76 6200 1 0.23 22 0.32 76 11100 10 0.23 22 0.34 76 19900 24 0.22 22 0.37 114 6200 0 0.33 19 0.28 114 11100 7 0.33 19 0.30 114 19900 20 0.33 19 0.33

NNR 38 6200 0 0.32 78 0.30 38 11100 0 0.32 78 0.32 38 19900 6 0.32 78 0.34 76 6200 1 0.23 70 0.24 76 11100 0 0.41 70 0.26 76 19900 3 0.42 69 0.29 114 6200 0 0.48 67 0.22 114 11100 0 0.47 67 0.24 114 19900 0 0.48 67 0.27

187

Curriculum Vitae

Rizki A. Mangkuto was born on 17 April 1984, in Bogor, West Java, Indonesia. In August 2002, he started his bachelor study in the Department of Engineering Physics at Institut Teknologi Bandung in Bandung, West Java, Indonesia. Since his third year in the university, he had been involved in assisting courses and practicums. In November 2006, he obtained the degree of Sarjana Teknik (equivalent to Bachelor of Science) with honour (cum laude). His bachelor final project was entitled ‘Effect of Luminaire Spacing on the Lighting Condition at Several Roadways in Bandung City’. Soon after that and until early 2010, he was involved in various research, education, and consultancy activities in the Laboratory of Building Physics and Acoustics, under Engineering Physics Research Group of the same university, within the subject of building and/or environmental acoustics and lighting.

In August 2007, he was granted a’voucher’ (full) scholarship to pursue a master degree within two years in the Graduate Programme of Engineering Physics at the same university, with concentration in building physics. In July 2009, he obtained the degree of Magister Teknik (equivalent to Master of Science) with honour. His master thesis was entitled ‘Study of Daylight Effect on Building Occupants Based on Electroencephalograph Signal’.

In the first semester of 2010/2011 academic year, he was a teaching assistant in the subject of Elementary Physics for first-year students of Diploma-3 (pre-bachelor) programme of Metrology and Instrumentation, held within the Undergraduate Programme of Engineering Physics. In April 2010, he was appointed as ‘academic assistant’ in Engineering Physics Research Group at Institut Teknologi Bandung. In the end of April 2010, he moved to the Netherlands to start his doctoral research in the Unit of Building Physics and Services at Eindhoven University of Technology. His project was supported by the Sound Lighting research line of the Intelligent Lighting Institute at Eindhoven University of Technology. The doctoral research was on ‘Modelling and Simulation of Virtual Natural Lighting Solutions in Buildings’, which resulted in this thesis.

188

List of Publications

Refereed academic journals

Pelzers, R. S., Yu, Q. L., & Mangkuto, R. A. (2014). Radiation modeling of a photo-reactor using a backward ray-tracing method: An insight into indoor photocatalytic oxidation. Environmental Science and Pollution Research, 1-14, doi: 10.1007/s11356-014-2552-1.

van Dronkelaar, C., Cóstola, D., Mangkuto, R. A., & Hensen, J. L. M. (2014). Heating and cooling energy demand in underground buildings: Potential for saving in various climates and functions. Energy and Buildings, 71, 129-136.

Mangkuto, R. A., Aries, M. B. C., van Loenen, E. J., & Hensen, J. L. M. (2014). Simulation of virtual natural lighting solutions with a simplified view. Lighting Research and Technology, 46, 198-218.

Conference proceedings

Pelzers, R. S., Yu, Q. L., Mangkuto, R. A., & Brouwers, H. J. H. (2013). Employing RADIANCE to refine indoor photocatalytic oxidation modeling. In Proceedings of JEP 2013 – the 3rd European Symposium on Photocatalysis, 25-27 September 2013, Portoroz, Slovenia (pp. P3-30). Portoroz: European Photocatalysis Federation (EPF) & University of Nova Gorica.

Mangkuto, R. A., Claessen, R. N. H., Aries, M. B. C., van Loenen, E. J., & Hensen, J. L. M. (2013). Space availability of buildings with virtual natural lighting solutions. In Proceedings of Lux Europa 2013 – the 12th European Lighting Conference, 17-19 September 2013, Kraków, Poland (pp. 275-280). Kraków: Polski Komitet Oswietleniowy.

Mangkuto, R. A., Aries, M. B. C., van Loenen, E. J., & Hensen, J. L. M. (2013). Development of virtual natural lighting solutions with a simplified view using lighting simulation. In Proceedings of Building Simulation 2013 – the 13th International Conference of the International Building Performance Simulation Association, 26-28 August 2013, Chambery, France, (pp. 3383-3390). Chambéry: IBPSA & Institute Nationale de l'Energie Solair (INES).

Mangkuto, R. A., Aries, M. B. C., van Loenen, E. J., & Hensen, J. L. M. (2012). Lighting performance of virtual natural lighting solutions with a simplified image in a reference office space. In Proceedings of Experiencing Light 2012, 12-13 November 2012, (pp 1-4), Eindhoven, the Netherlands. Eindhoven: Eindhoven University of Technology.

Mangkuto, R. A., Ochoa Morales, C. E., Aries, M. B. C., van Loenen, E. J., & Hensen, J. L. M. (2011). Review of modelling approaches for developing virtual natural lighting solutions. In Proceedings of Building Simulation 2011 – the 12th International


189

Conference of the International Building Performance Simulation Association, Sydney, Australia, 14-16 November 2011, (pp. 2643-2650). Sydney: IBPSA Australasia & AIRAH.

Mangkuto, R. A., Aries, M. B. C., van Loenen, E. J., & Hensen, J. L. M. (2011). Properties and performance indicators of virtual natural lighting solutions. In Proceedings of CISBAT 2011, Lausanne, Switzerland, 14-16 September 2011, (pp. 379-384). Lausanne: Solar Energy and Building Physics Laboratory (LESO-PB), École Polytechnique Fédérale de Lausanne (EPFL).

Harahap, Y., Mangkuto, R. A., & Soelami, F. X. N. (2009). Lighting design for axis of Gedung Sate and Monument of West Java People's Struggle. In Proceedings of ITB International Conference on Regional Development, Environment and Infrastructures, Bandung, Indonesia, 18-19 June 2009. Bandung: Institut Teknologi Bandung.

Mangkuto, R. A., Soelami, F. X. N., & Suprijanto, S. (2009). Study of effect of daylight on building User's performance based on electroencephalograph signal. In Proceedings of the 10th SENVAR / 1st CONVEESH, Manado, Indonesia, 26-27 October 2009. Manado: Universitas Sam Ratulangi.

Mangkuto, R. A., Paripurna, A., & Soelami, F. X. N. (2009). Evaluation on glare from vehicle lamps and effectiveness of road components as glare barriers. In Proceedings of the 3rd SEATUC Symposium, Johor Bahru, Malaysia, 25-26 February 2009. Johor Bahru: Universiti Teknologi Malaysia.

Mangkuto, R. A., Soelami, F. X. N., & Soegijanto, R. M. (2008). Evaluation on lighting condition and visual legibility of road surfaces and traffic signs in Bandung City. In Proceeding of the 9th SENVAR / 2nd ISESEE, Shah Alam, Malaysia, 1-3 December 2008. Shah Alam: Universiti Teknologi Mara.

Professional journals

van Dronkelaar, C., Cóstola, D., Mangkuto, R. A., & Hensen, J. L. M. (2013). Ondergronds als alternatief voor bovengronds bouwen. TVVL Magazine, 42(10), 44-47.

Mangkuto, R. A., Ochoa Morales, C. E., Aries, M. B. C., van Loenen, E. J., & Hensen, J. L. M. (2012). Simulaties voor R&D: ‘Virtual Natural Lighting’ systemen. TVVL Magazine, 41(5), 38-40.


190

Bouwstenen is een publikatiereeksvan de Faculteit Bouwkunde,Technische Universiteit Eindhoven.Zij presenteert resultaten vanonderzoek en andere aktiviteiten ophet vakgebied der Bouwkunde,uitgevoerd in het kader van dezeFaculteit.

Bouwstenen zijn telefonisch tebestellen op nummer040 - 2472383

KernredaktieMTOZ

Reeds verschenen in de serieBouwstenen

nr 1Elan: A Computer Model for Building Energy Design: Theory and ValidationMartin H. de WitH.H. DriessenR.M.M. van der Velden

nr 2Kwaliteit, Keuzevrijheid en Kosten: Evaluatie van Experiment Klarendal, ArnhemJ. SmeetsC. le NobelM. Broos J. FrenkenA. v.d. Sanden

nr 3Crooswijk: Van ‘Bijzonder’ naar ‘Gewoon’Vincent SmitKees Noort

nr 4Staal in de WoningbouwEdwin J.F. Delsing

nr 5Mathematical Theory of Stressed Skin Action in Profiled Sheeting with Various Edge ConditionsAndre W.A.M.J. van den Bogaard

nr 6Hoe Berekenbaar en Betrouwbaar is de Coëfficiënt k in x-ksigma en x-ks? K.B. LubA.J. Bosch

nr 7Het Typologisch Gereedschap: Een Verkennende Studie Omtrent Typologie en Omtrent de Aanpak van Typologisch OnderzoekJ.H. Luiten nr 8Informatievoorziening en BeheerprocessenA. NautaJos Smeets (red.)Helga Fassbinder (projectleider)Adrie ProveniersJ. v.d. Moosdijk

nr 9Strukturering en Verwerking van Tijdgegevens voor de Uitvoering van Bouwwerkenir. W.F. SchaeferP.A. Erkelens

nr 10Stedebouw en de Vorming van een Speciale WetenschapK. Doevendans

nr 11Informatica en Ondersteuning van Ruimtelijke BesluitvormingG.G. van der Meulen

nr 12Staal in de Woningbouw, Korrosie-Bescherming van de Begane GrondvloerEdwin J.F. Delsing

nr 13Een Thermisch Model voor de Berekening van Staalplaatbetonvloeren onder BrandomstandighedenA.F. Hamerlinck

nr 14De Wijkgedachte in Nederland: Gemeenschapsstreven in een Stedebouwkundige ContextK. DoevendansR. Stolzenburg

nr 15Diaphragm Effect of Trapezoidally Profiled Steel Sheets: Experimental Research into the Influence of Force ApplicationAndre W.A.M.J. van den Bogaard

nr 16Versterken met Spuit-Ferrocement: Het Mechanische Gedrag van met Spuit-Ferrocement Versterkte Gewapend BetonbalkenK.B. LubirM.C.G. van Wanroy

nr 17De Tractaten van Jean Nicolas Louis DurandG. van Zeyl

nr 18Wonen onder een Plat Dak: Drie Opstellen over Enkele Vooronderstellingen van de StedebouwK. Doevendans

nr 19Supporting Decision Making Processes: A Graphical and Interactive Analysis of Multivariate DataW. Adams

nr 20Self-Help Building Productivity: A Method for Improving House Building by Low-Income Groups Applied to Kenya 1990-2000P. A. Erkelens

nr 21De Verdeling van Woningen: Een Kwestie van OnderhandelenVincent Smit

nr 22Flexibiliteit en Kosten in het Ontwerpproces: Een Besluitvormingondersteunend ModelM. Prins

nr 23Spontane Nederzettingen Begeleid: Voorwaarden en Criteria in Sri LankaPo Hin Thung

nr 24Fundamentals of the Design of Bamboo StructuresOscar Arce-Villalobos

nr 25Concepten van de BouwkundeM.F.Th. Bax (red.)H.M.G.J. Trum (red.)

nr 26Meaning of the SiteXiaodong Li

nr 27Het Woonmilieu op Begrip Gebracht: Een Speurtocht naar de Betekenis van het Begrip 'Woonmilieu'Jaap Ketelaar

nr 28Urban Environment in Developing Countrieseditors: Peter A. Erkelens George G. van der Meulen (red.)

nr 29Stategische Plannen voor de Stad: Onderzoek en Planning in Drie Stedenprof.dr. H. Fassbinder (red.)H. Rikhof (red.)

nr 30Stedebouwkunde en StadsbestuurPiet Beekman

nr 31De Architectuur van Djenné: Een Onderzoek naar de Historische StadP.C.M. Maas

nr 32Conjoint Experiments and Retail PlanningHarmen Oppewal

nr 33Strukturformen Indonesischer Bautechnik: Entwicklung Methodischer Grundlagen für eine ‘Konstruktive Pattern Language’ in IndonesienHeinz Frick arch. SIA

nr 34Styles of Architectural Designing: Empirical Research on Working Styles and Personality DispositionsAnton P.M. van Bakel

nr 35Conjoint Choice Models for Urban Tourism Planning and MarketingBenedict Dellaert

nr 36Stedelijke Planvorming als Co-ProduktieHelga Fassbinder (red.)

nr 37 Design Research in the Netherlandseditors: R.M. Oxman M.F.Th. Bax H.H. Achten

nr 38 Communication in the Building IndustryBauke de Vries

nr 39 Optimaal Dimensioneren van Gelaste PlaatliggersJ.B.W. StarkF. van PeltL.F.M. van GorpB.W.E.M. van Hove

nr 40 Huisvesting en Overwinning van ArmoedeP.H. Thung P. Beekman (red.)

nr 41 Urban Habitat: The Environment of TomorrowGeorge G. van der Meulen Peter A. Erkelens

nr 42A Typology of JointsJohn C.M. Olie

nr 43Modeling Constraints-Based Choices for Leisure Mobility PlanningMarcus P. Stemerding

nr 44Activity-Based Travel Demand ModelingDick Ettema

nr 45Wind-Induced Pressure Fluctuations on Building FacadesChris Geurts

nr 46Generic RepresentationsHenri Achten

nr 47Johann Santini Aichel: Architectuur en AmbiguiteitDirk De Meyer

nr 48Concrete Behaviour in Multiaxial CompressionErik van Geel

nr 49Modelling Site SelectionFrank Witlox

nr 50Ecolemma ModelFerdinand Beetstra

nr 51Conjoint Approaches to Developing Activity-Based ModelsDonggen Wang

nr 52On the Effectiveness of VentilationAd Roos

nr 53Conjoint Modeling Approaches for Residential Group preferencesEric Molin

nr 54Modelling Architectural Design Information by FeaturesJos van Leeuwen

nr 55A Spatial Decision Support System for the Planning of Retail and Service FacilitiesTheo Arentze

nr 56Integrated Lighting System AssistantEllie de Groot

nr 57Ontwerpend Leren, Leren OntwerpenJ.T. Boekholt

nr 58Temporal Aspects of Theme Park Choice BehaviorAstrid Kemperman

nr 59Ontwerp van een Geïndustrialiseerde FunderingswijzeFaas Moonen

nr 60Merlin: A Decision Support System for Outdoor Leisure PlanningManon van Middelkoop

nr 61The Aura of ModernityJos Bosman

nr 62Urban Form and Activity-Travel PatternsDaniëlle Snellen

nr 63Design Research in the Netherlands 2000Henri Achten

nr 64Computer Aided Dimensional Control in Building ConstructionRui Wu

nr 65Beyond Sustainable Buildingeditors: Peter A. Erkelens Sander de Jonge August A.M. van Vlietco-editor: Ruth J.G. Verhagen

nr 66Das Globalrecyclingfähige HausHans Löfflad

nr 67Cool Schools for Hot SuburbsRené J. Dierkx

nr 68A Bamboo Building Design Decision Support ToolFitri Mardjono

nr 69Driving Rain on Building EnvelopesFabien van Mook

nr 70Heating Monumental ChurchesHenk Schellen

nr 71Van Woningverhuurder naar Aanbieder van WoongenotPatrick Dogge

nr 72Moisture Transfer Properties of Coated GypsumEmile Goossens

nr 73Plybamboo Wall-Panels for HousingGuillermo E. González-Beltrán

nr 74The Future Site-ProceedingsGer MaasFrans van Gassel

nr 75Radon transport in Autoclaved Aerated ConcreteMichel van der Pal

nr 76The Reliability and Validity of Interactive Virtual Reality Computer ExperimentsAmy Tan

nr 77Measuring Housing Preferences Using Virtual Reality and Belief NetworksMaciej A. Orzechowski

nr 78Computational Representations of Words and Associations in Architectural DesignNicole Segers

nr 79Measuring and Predicting Adaptation in Multidimensional Activity-Travel PatternsChang-Hyeon Joh

nr 80Strategic BriefingFayez Al Hassan

nr 81Well Being in HospitalsSimona Di Cicco

nr 82Solares Bauen:Implementierungs- und Umsetzungs-Aspekte in der Hochschulausbildung in ÖsterreichGerhard Schuster

nr 83Supporting Strategic Design of Workplace Environments with Case-Based ReasoningShauna Mallory-Hill

nr 84ACCEL: A Tool for Supporting Concept Generation in the Early Design Phase Maxim Ivashkov

nr 85Brick-Mortar Interaction in Masonry under CompressionAd Vermeltfoort

nr 86 Zelfredzaam WonenGuus van Vliet

nr 87Een Ensemble met Grootstedelijke AllureJos BosmanHans Schippers

nr 88On the Computation of Well-Structured Graphic Representations in Architectural Design Henri Achten

nr 89De Evolutie van een West-Afrikaanse Vernaculaire ArchitectuurWolf Schijns

nr 90ROMBO TactiekChristoph Maria Ravesloot

nr 91External Coupling between Building Energy Simulation and Computational Fluid DynamicsEry Djunaedy

nr 92Design Research in the Netherlands 2005editors: Henri Achten Kees Dorst Pieter Jan Stappers Bauke de Vries

nr 93Ein Modell zur Baulichen TransformationJalil H. Saber Zaimian

nr 94Human Lighting Demands: Healthy Lighting in an Office EnvironmentMyriam Aries

nr 95A Spatial Decision Support System for the Provision and Monitoring of Urban GreenspaceClaudia Pelizaro

nr 96Leren CreërenAdri Proveniers

nr 97SimlandscapeRob de Waard

nr 98Design Team CommunicationAd den Otter

nr 99Humaan-Ecologisch Georiënteerde WoningbouwJuri Czabanowski

nr 100HambaseMartin de Wit

nr 101Sound Transmission through Pipe Systems and into Building StructuresSusanne Bron-van der Jagt

nr 102Het Bouwkundig ContrapuntJan Francis Boelen

nr 103A Framework for a Multi-Agent Planning Support SystemDick Saarloos

nr 104Bracing Steel Frames with Calcium Silicate Element WallsBright Mweene Ng’andu

nr 105Naar een Nieuwe HoutskeletbouwF.N.G. De Medts

nr 108GeborgenheidT.E.L. van Pinxteren

nr 109Modelling Strategic Behaviour in Anticipation of CongestionQi Han

nr 110Reflecties op het WoondomeinFred Sanders

nr 111On Assessment of Wind Comfort by Sand ErosionGábor Dezsö

nr 112Bench Heating in Monumental Churches Dionne Limpens-Neilen

nr 113RE. ArchitectureAna Pereira Roders

nr 114Toward Applicable Green ArchitectureUsama El Fiky

nr 115Knowledge Representation under Inherent Uncertainty in a Multi-Agent System for Land Use PlanningLiying Ma

nr 116Integrated Heat Air and Moisture Modeling and SimulationJos van Schijndel

nr 117Concrete Behaviour in Multiaxial CompressionJ.P.W. Bongers

nr 118The Image of the Urban LandscapeAna Moya Pellitero

nr 119The Self-Organizing City in VietnamStephanie Geertman

nr 120A Multi-Agent Planning Support System for Assessing Externalities of Urban Form ScenariosRachel Katoshevski-Cavari

nr 121Den Schulbau Neu Denken, Fühlen und WollenUrs Christian Maurer-Dietrich

nr 122Peter Eisenman Theories and PracticesBernhard Kormoss

nr 123User Simulation of Space UtilisationVincent Tabak

nr 125In Search of a Complex System ModelOswald Devisch

nr 126Lighting at Work:Environmental Study of Direct Effects of Lighting Level and Spectrum onPsycho-Physiological VariablesGrazyna Górnicka

nr 127Flanking Sound Transmission through Lightweight Framed Double Leaf WallsStefan Schoenwald

nr 128Bounded Rationality and Spatio-Temporal Pedestrian Shopping BehaviorWei Zhu

nr 129Travel Information:Impact on Activity Travel PatternZhongwei Sun

nr 130Co-Simulation for Performance Prediction of Innovative Integrated Mechanical Energy Systems in BuildingsMarija Trcka

nr 131Allemaal WinnenM.J. Bakker

˙

�

nr 132Architectural Cue Model in Evacuation Simulation for Underground Space DesignChengyu Sun

nr 133Uncertainty and Sensitivity Analysis in Building Performance Simulation for Decision Support and Design OptimizationChristina Hopfe

nr 134Facilitating Distributed Collaboration in the AEC/FM Sector Using Semantic Web TechnologiesJacob Beetz

nr 135Circumferentially Adhesive Bonded Glass Panes for Bracing Steel Frame in FaçadesEdwin Huveners

nr 136Influence of Temperature on Concrete Beams Strengthened in Flexure with CFRPErnst-Lucas Klamer

nr 137Sturen op KlantwaardeJos Smeets

nr 139Lateral Behavior of Steel Frames with Discretely Connected Precast Concrete Infill PanelsPaul Teewen

nr 140Integral Design Method in the Context of Sustainable Building DesignPerica Savanovic

nr 141Household Activity-Travel Behavior: Implementation of Within-Household InteractionsRenni Anggraini

nr 142Design Research in the Netherlands 2010Henri Achten

nr 143Modelling Life Trajectories and Transport Mode Choice Using Bayesian Belief NetworksMarloes Verhoeven

nr 144Assessing Construction Project Performance in GhanaWilliam Gyadu-Asiedu

nr 145Empowering Seniors through Domotic HomesMasi Mohammadi

nr 146An Integral Design Concept forEcological Self-Compacting ConcreteMartin Hunger

nr 147Governing Multi-Actor Decision Processes in Dutch Industrial Area RedevelopmentErik Blokhuis

nr 148A Multifunctional Design Approach for Sustainable ConcreteGötz Hüsken

nr 149Quality Monitoring in Infrastructural Design-Build ProjectsRuben Favié

nr 150Assessment Matrix for Conservation of Valuable Timber StructuresMichael Abels

nr 151Co-simulation of Building Energy Simulation and Computational Fluid Dynamics for Whole-Building Heat, Air and Moisture EngineeringMohammad Mirsadeghi

nr 152External Coupling of Building Energy Simulation and Building Element Heat, Air and Moisture SimulationDaniel Cóstola

´

nr 153Adaptive Decision Making In Multi-Stakeholder Retail Planning Ingrid Janssen

nr 154Landscape GeneratorKymo Slager

nr 155Constraint Specification in ArchitectureRemco Niemeijer

nr 156A Need-Based Approach to Dynamic Activity GenerationLinda Nijland

nr 157Modeling Office Firm Dynamics in an Agent-Based Micro Simulation FrameworkGustavo Garcia Manzato

nr 158Lightweight Floor System for Vibration ComfortSander Zegers

nr 159Aanpasbaarheid van de DraagstructuurRoel Gijsbers

nr 160'Village in the City' in Guangzhou, ChinaYanliu Lin

nr 161Climate Risk Assessment in MuseumsMarco Martens

nr 162Social Activity-Travel PatternsPauline van den Berg

nr 163Sound Concentration Caused by Curved SurfacesMartijn Vercammen

nr 164Design of Environmentally Friendly Calcium Sulfate-Based Building Materials: Towards an Improved Indoor Air QualityQingliang Yu

nr 165Beyond Uniform Thermal Comfort on the Effects of Non-Uniformity and Individual PhysiologyLisje Schellen

nr 166Sustainable Residential DistrictsGaby Abdalla

nr 167Towards a Performance Assessment Methodology using Computational Simulation for Air Distribution System Designs in Operating RoomsMônica do Amaral Melhado

nr 168Strategic Decision Modeling in Brownfield RedevelopmentBrano Glumac

nr 169Pamela: A Parking Analysis Model for Predicting Effects in Local AreasPeter van der Waerden

nr 170A Vision Driven Wayfinding Simulation-System Based on the Architectural Features Perceived in the Office EnvironmentQunli Chen

nr 171Measuring Mental Representations Underlying Activity-Travel ChoicesOliver Horeni

nr 172Modelling the Effects of Social Networks on Activity and Travel BehaviourNicole Ronald

nr 173Uncertainty Propagation and Sensitivity Analysis Techniques in Building Performance Simulation to Support Conceptual Building and System DesignChristian Struck

nr 174Numerical Modeling of Micro-Scale Wind-Induced Pollutant Dispersion in the Built EnvironmentPierre Gousseau

nr 175Modeling Recreation Choices over the Family LifecycleAnna Beatriz Grigolon

nr 176Experimental and Numerical Analysis of Mixing Ventilation at Laminar, Transitional and Turbulent Slot Reynolds NumbersTwan van Hooff

nr 177Collaborative Design Support:Workshops to Stimulate Interaction and Knowledge Exchange Between PractitionersEmile M.C.J. Quanjel

nr 178Future-Proof Platforms for Aging-in-PlaceMichiel Brink

nr 179Motivate: A Context-Aware Mobile Application forPhysical Activity PromotionYuzhong Lin

nr 180Experience the City:Analysis of Space-Time Behaviour and Spatial Learning Anastasia Moiseeva

nr 181Unbonded Post-Tensioned Shear Walls of Calcium Silicate Element MasonryLex van der Meer

nr 182Construction and Demolition Waste Recycling into Innovative Building Materials for Sustainable Construction in TanzaniaMwita M. Sabai

nr 183Durability of Concretewith Emphasis on Chloride Migration��

nr 184Computational Modeling of Urban Wind Flow and Natural Ventilation Potential of Buildings Rubina Ramponi

nr 185A Distributed Dynamic Simulation Mechanism for Buildings Automation and Control SystemsAzzedine Yahiaoui

nr 186Modeling Cognitive Learning of UrbanNetworks in Daily Activity-Travel BehaviorSehnaz Cenani Durmazoglu

nr 187Functionality and Adaptability of Design Solutions for Public Apartment Buildingsin GhanaStephen Agyefi-Mensah

nr 188A Construction Waste Generation Model for Developing CountriesLilliana Abarca-Guerrero

nr 189Synchronizing Networks:The Modeling of Supernetworks for Activity-Travel BehaviorFeixiong Liao

nr 190Time and Money Allocation Decisions in Out-of-Home Leisure Activity Choices Gamze Zeynep Dane

nr 191How to Measure Added Value of CRE and Building Design Rianne Appel-Meulenbroek

nr 192Secondary Materials in Cement-Based Products:Treatment, Modeling and Environmental InteractionMiruna Florea

nr 193Concepts for the Robustness Improvement of Self-Compacting Concrete: Effects of Admixtures and Mixture Components on the Rheology and Early Hydration at Varying TemperaturesWolfram Schmidt

�¸

Propositions

Propositions associated with the thesis


1. Applying the ideal VNLS, which combines artificial light and view together, will have more positive effects on human well-being, compared to only applying artificial bright light without a view, or applying an artificial view without emitting light.

This thesis, Chapter 1.

2. Directionality of the light is an important property that typically distinguishes a real window or skylight from an artificial version.


3. Developing future solutions such as VNLS is a long process. The use of modelling and simulation is important in influencing the design decision; therefore it deserves to be discussed on its own.


4. In climbing towards the goal of making robots appear human, our affinity for them increases until a certain point where we realise that the human-like robot may at the first instance look real, but is in fact artificial, so that we experience an eerie sensation. Such a sensation occurs since we at that point perceive the robot as a proximal threat – it would hardly occur if the ideal object was, for instance, a window.

Mori, M. (1970). The uncanny valley. Energy, 7, 33-35 [in Japanese].

5. It must not be believed that because almost all problems can be solved with computers, there is no need to examine the properties of the solutions. It is always essential to choose reasonable approximations to solve the problem; it is a nonsense to develop or run a big routine to compute results that can be obtained simply through analytical methods.

Filippi, F., Habault, D., Lefebvre, J-P., Bergassoli, A. (1999). Acoustics: Basic Physics, Theory and Methods, London: Academic Press.

Propositions

6. Architects and the building industry have started moving in the direction of sustainable practice. The importance of having a critical attitude is even greater now than it was. Unfortunately there are many who use the label of ‘sustainable’ without the substance. Few dare to say to them that ‘the emperor has no clothes’.

Szokolay, S. V. (2008). Introduction to Architectural Science: The Basis of Sustainable Design (2nd ed.). Oxford: Architectural Press.

7. In the long journey of a scientific adventure, of which a doctoral project is a typical example, it is very easy to get lost in the jungle of research.

8. Lecture notes can only give methods, but we must define the boundary condition ourselves. We should go to the streets, out to the villages, take notes on every symptom, and understand the real problems. What is the meaning of art, if separated from the anguish of environment. What is the meaning of thinking, if separated from the problems of life.

Rendra, W. S. (1977). Sajak Sebatang Lisong (Poem of a Cigar) [in Indonesian].

9. Today individuals can detach themselves from the village unit in order to move to a nearby city, another island, or even to an urban setting halfway around the world. Those who have lived away from their village, in the rantau, for much of their lives, do return to the village because they still consider it as a comfortable place to retire in their old age.

Tanner, N. M. (1982). The nuclear family in Minangkabau matriliny: A mirror of disputes. Bijdragen tot de Taal-, Land- en Volkenkunde, 138, 129-151.

10. The most succesful person is the one who manages to give the most benefits to his family, society, and environment. As the Prophet said, “When a son of Adam dies, his actions are cut off except for three: a continuing charity, knowledge which gives benefit, and a virtuous child who prays for him”.

Prophet Muhammad S.A.W. Shahih Muslim, hadits narrated by Imam Muslim, from Abu Hurairah [in Arabic].

Documents

Modelling and simulation of virtual natural lighting solutions in buildings