Climate Adaptive Building Shells

What can we simulate?

Roel Loonen 21 June 2010 MSc-Thesis Technische Universiteit Eindhoven

Master of Science Thesis

Climate Adaptive Building Shells

What can we simulate?

21 June 2010

Student: R.C.G.M. Loonen ID: Master: Faculty: 0570677

Advisors prof. dr. ir. J.L.M. Hensen

Building Services Architecture, Building & Planning Eindhoven University of Technology

dr. dipl.-ing. M. Trka D. Cstola, MSc

S UMMARY

Building shells are at the interface between the building interior and ambient climate, and therefore fulll a number of vital functions that dictate most of the buildings energy consumption. The buildings environment is changing over time (short-term weather conditions, diurnal and seasonal cycles), and this also applies for occupancy and comfort preferences. In spite of these changing conditions, traditional building envelopes are rigid systems with xed thermophysical and optical properties, and hence cannot adapt to this variability. Climate adaptive building shells (CABS) do have the ability to repeatedly and reversibly change their functions, features or behavior over time in response to changes in performance requirements and variable boundary conditions. Therefore, these faade can seize the opportunity to save energy by adapting to prevailing weather conditions, and support comfort levels by immediately responding to occupants wishes. A total number of one hundred CABS concepts has been analyzed in this study which reveals that application of CABS is despite of its promises still rather limited. Successful design and operation of CABS is a complex task. CABS typically are complex systems with their relevant phenomena taking place across various physical domains. In addition, CABS have to resolve conicts and trade-offs between energy consumption and thermal and visual comfort requirements. Traditional design methods are deemed inadequate to facilitate these insights. This study explores the role that building performance simulation (BPS) can play in designing buildings with CABS. The tool is well-suited for this purpose because it promotes further insights in the dynamic interactions between climate, the building and its occupants. By offering the possibility to answer what-if questions, it enables us to predict building performance in different scenarios. An analysis of current capabilities and requirements however reveals some space for improvements. Both exibility of features to model adaptive behavior, and integration of impacts from the various physical domains are open to further development. The work also resulted in a strategy for effective modeling and simulation of CABS. The applicability of BPS in relation to CABS was exemplied in the case study of Smart Energy Glass. This innovative concept uses a unique switchable coating that makes it possible to effectively control the optical properties of a window. By switching transparency in one of three possible states (bright, dark and translucent), part of the incoming sunlight is captured and converted into electricity. Based on a description of underlying physics, an integrated BPS model was developed that uses TRNSYS in combination with DAYSIM. The model was rened and validated in a reduced-scale experimental study. Subsequently, energy consumption and thermal and visual comfort were evaluated under various control strategies and compared to the performance of a conventional building shell. Apparently, the challenge is to control adaptive behavior in such a way that energy demand is reduced to a minimum while good indoor environmental quality is retained. In the end, the results helped to identify directions for future developments that can contribute to make the concept of SEG a success. This study concludes that CABS have the ability to become a contributor to satisfy the increasingly stricter energy performance targets without being constrained by the need for concessions in terms of comfort. Building performance simulation is a valuable tool that can prevent this latent potential from continuing to remain largely untouched.

S AMENVATTING

Gevels bevinden zich op het grensvlak tussen binnen en buiten, en spelen daarom een belangrijke rol in de energiehuishouding van een gebouw. De conventionele gevel heeft echter een voornamelijk beschermende functie met vaste eigenschappen en kan daardoor niet inspelen op de constant veranderende gebouwomgeving (weer, dag/nacht cycli, seizoenen) en comfortbehoeften. Adaptieve gevels daarentegen hebben wel de mogelijkheid om hun gedrag of eigenschappen te veranderen in de tijd. Door op slimme wijze in te spelen op de buitencondities kan de mogelijkheid worden aangegrepen om energie te besparen. Door daarnaast ook te anticiperen op de wisselende gebruikersbehoeften kan tegelijkertijd een gunstig binnenklimaat gewaarborgd blijven. Dit onderzoek heeft onder meer geleid tot een overzicht van een honderdtal adaptieve gevelconcepten, er kan echter geconstateerd worden dat ondanks het veelbelovende concept de toepassing in de praktijk vooralsnog beperkt is. Het succesvol ontwerpen en regelen van adaptieve gevels blijkt een complexe opgave. Prestaties worden bepaald door dynamische invloeden uit verschillende fysische domeinen. Daarnaast dient een adaptieve gevel op verstandige wijze om te kunnen gaan met de concurrerende of mogelijk zelfs conicterende belangen tussen thermisch en visueel comfort enerzijds, en de bijbehorende energievraag anderzijds. Traditionele ontwerpmethoden blijken onvoldoende geschikt om dit inzicht te verschaffen. In dit onderzoek is ingegaan op de rol die gebouwsimulatie kan spelen in de verdere ontwikkeling van adaptieve faades. De tools lenen zich uitstekend voor dit doel doordat ze inzicht verschaffen in de dynamische interacties tussen buitenklimaat, het gebouw en zijn gebruikers. Door de mogelijkheid van het beantwoorden van wat-als vragen is het mogelijk gebouwprestaties te voorspellen onder verschillende scenarios. Op basis van een analyse van de huidige mogelijkheden en behoeften kan worden opgemaakt dat er nog wel ruimte voor verbetering bestaat. Zowel de exibiliteit die de gebruiker heeft voor het modelleren van adaptief gedrag als de koppeling tussen invloeden uit verschillende fysische domeinen verdienen nog verdere aandacht. Dit onderzoek heeft onder andere geleid tot een algemene aanpak voor modelleren en simuleren van gebouwen met adaptieve gevels. De mogelijkheden van simulatie zijn daarna gellustreerd aan de hand van Smart Energy Glass. Dit in ontwikkeling zijnde, innovatieve type beglazing kan geschakeld worden in drie standen licht, donker en diffuus waarbij een deel van het invallende zonlicht wordt afgevangen en omgezet in elektriciteit. Op basis van een beschrijving van het fysische gedrag is een simulatiemodel ontwikkeld waarin TRNSYS wordt gecombineerd met DAYSIM. Dit model is verder verjnd en vervolgens gevalideerd met behulp van een experimentele studie waarbij een prototype was genstalleerd in een testgevel en blootgesteld aan het buitenklimaat. De prestaties van het concept zijn in kaart gebracht op basis van integrale simulaties voor verwarming, koeling, daglicht en elektriciteit in verschillende situaties. Hieruit blijkt dat de crux is gelegen in de uitdaging om adaptief gedrag zodanig te regelen dat energiegebruik tot een minimum beperkt blijft maar tegelijkertijd behoud van acceptabel comfort gewaarborgd is. Daarnaast zijn richtingen geformuleerd voor verdere doorontwikkeling die het welslagen van het concept Smart Energy Glass kunnen bevorderen. De algemene conclusie is dat adaptieve gevels kunnen bijdragen aan het bereiken van de steeds scherpere energiedoelstellingen, zonder dat daarbij getornd hoeft te worden aan comfortbeleving. Gebouwsimulatie vormt een onmisbaar gereedschap om dit latente potentieel niet langer onbenut te laten.

iii

A CKNOWLEDGEMENTS

I want to thank and acknowledge several people whose input and inuence helped shape this thesis. First of all, I wish to thank prof.dr.ir. Jan Hensen, dr.dipl.-ing. Marija Tr ka and Daniel Cstola, MSc from c Eindhoven University of Technology, unit Building Physics and Systems for their continuous advice and suggestions in the supervision of the project. The cooperation with Peer+ denitely enriched the work in this thesis. It not only made the rather abstract concept of CABS much more tangible, but also positively contributed to the multidisciplinary character of the research. I would like to thank Casper van Oosten and Teun Wagenaar for giving me this chance to get an inside look at the innovation process of a real product. I also thank Marcel Demandt for his assistance and pleasant collaboration during the course of the project. I gratefully acknowledge the practical support I received in setting up the experimental part of the graduation project. I want to thank Paul Verbunt for his constructive thoughts during initiation of, and kind assistance with conducting the experiments in the foreign environment of the laboratory at the department of Chemical Engineering. Finally, I acknowledge the support from the staff at the BPS laboratory. Without their expertise, the measurements in the test-faade could not have been accomplished.

C ONTENTS

Summary Samenvatting Acknowledgements 1 Introduction 1.1 Background . . . . . . . . . . . . . 1.2 Problem statement . . . . . . . . 1.3 Objective and research questions 1.4 Methodology . . . . . . . . . . . . 1.5 Outline . . . . . . . . . . . . . . .

i iii v 1 1 2 3 3 4 7 7 8 9 10 11 12 12 16 20 22 27 27 27 29 33 33 34 37 38 42 43 44 45 48

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

Climate adaptive building shells 2.1 Introduction . . . . . . . . . . . . . . . 2.2 CABS performance . . . . . . . . . . . 2.2.1 Robustness vs Flexibility . . . 2.2.2 CABS as exible systems . . . 2.2.3 Performance benets of CABS 2.3 CABS Analysis . . . . . . . . . . . . . . 2.3.1 Adaptive behavior . . . . . . . 2.3.2 Concepts . . . . . . . . . . . . . 2.3.3 Control . . . . . . . . . . . . . . 2.3.4 Barriers and challenges . . . . 2.4 CABS Classication . . . . . . . . . . . 2.4.1 Introduction . . . . . . . . . . 2.4.2 Methodology . . . . . . . . . . 2.4.3 Results . . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

Building Performance Simulation 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . 3.2 Requirements for performance simulation of CABS 3.3 Capabilities of BPS tools . . . . . . . . . . . . . . . . 3.4 Physical domains . . . . . . . . . . . . . . . . . . . . 3.5 Coupling of physical domains . . . . . . . . . . . . . 3.6 Adaptive features . . . . . . . . . . . . . . . . . . . . 3.6.1 Application-oriented . . . . . . . . . . . . . 3.6.2 General purpose . . . . . . . . . . . . . . . . 3.7 Discussion . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

Modeling and simulation of climate adaptive building shells 51 4.1 Model abstraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 4.2 Simulation requirement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 vii

C ONTENTS 4.3 4.4 4.5 5 Relevant physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Simulation strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Verication and Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 59 59 61 62 63 63 64 69 69 70 71 73 73 74 74 76 76 76 78 81 83 83 83 84 86 87 89 90 91 92

Modeling Smart Energy Glass 5.1 Introduction . . . . . . . . . 5.2 Real world system . . . . . . 5.3 Conceptual model . . . . . . 5.3.1 Optical Model . . . . 5.3.2 Thermal Model . . . 5.3.3 Electrical Model . . 5.4 Simulation model . . . . . . 5.4.1 Simulation strategy 5.4.2 DAYSIM . . . . . . . 5.4.3 TRNSYS . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

Smart Energy Glass model validation 6.1 Introduction . . . . . . . . . . . . 6.2 Methodology . . . . . . . . . . . . 6.2.1 Experimental setup . . . 6.2.2 Simulations . . . . . . . . 6.3 Comparing results . . . . . . . . . 6.3.1 Daylight model . . . . . . 6.3.2 Electrical model . . . . . 6.4 Conclusion . . . . . . . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

Performance simulation of Smart Energy Glass 7.1 Methodology . . . . . . . . . . . . . . . . . . 7.1.1 Model setup . . . . . . . . . . . . . . 7.1.2 Performance indicators . . . . . . . 7.2 Case 1 - Energy efcient ofce . . . . . . . . 7.3 Case 2 - Advanced renovation . . . . . . . . 7.4 Case 3 - SEG in a continental climate . . . . 7.5 Discussion . . . . . . . . . . . . . . . . . . . 7.5.1 Dynamic daylight performance . . 7.5.2 Radiant asymmetry . . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

Conclusions 95 8.1 Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 8.2 Recommendations for future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 99 111 113

References A Overview of 100 Climate Adaptive Building Shells B Smart Energy Glass product information

viii

CHAPTER
1.1 Background
1

I NTRODUCTION

Most people spend up to 90% of their time indoors (Bougdah and Sharples, 2010), which justies our strive for buildings that are as safe, healthy and comfortable as possible. Historically, it has always been the building shell that provided shelter and protection against the negative inuences caused by erratic external conditions. With the advent of articial lighting and HVAC systems in the twentieth century however, this important task has gradually been delegated to the building services. As a consequence, the building shell lost its role as moderator for energy and comfort, and it is nowadays more or less accepted that buildings pose a major load on our environment. Energy consumption in buildings has increased in such a way that it now dominates over the industrial en transportation sectors (PrezLombard et al., 2008). In addition, the built environment contributes as much as one third of total global greenhouse gas emissions. Buildings are thus seen as one of the main instigators of global warming with all the corresponding implications, but this threat also introduces high opportunities for the sector. The International Panel on Climate Change (IPCC) labeled Buildings as the sector with the highest potential to mitigate the detrimental effects in a cost-effective way. However, if the desired targets for greenhouse gas emissions reduction are to be met, the emissions from the building sector have to be suppressed with much greater seriousness and vigor than has been the case to date (UNEP, 2009). Continuation of current practice will likely turn out to be inadequate (Figure 1.1). What is needed is a transformation; a paradigm shift in the way we envision energy efciency. Low-energy buildings must become the norm, rather than the novelty project, in both new and existing buildings (WBCSD, 2009). Innovations and technological development are supposed to be the catalyst that provides the necessary impulses for this major leap forward. In this respect, it is worthwhile to reconsider the role of the building shell. Unlike only being a wrapper, the building shell can be entrusted with multiple vital functions that dictate the buildings energy consumption and perception of indoor environmental quality. Traditionally, most building shells are static, whereas the climatological boundary conditions and user preferences are constantly changing. As a result, traditional faades cannot adapt to the changes it is exposed to, and this results in a loss opportunity for energy saving and increasing thermal and visual occupant comfort. Climate adaptive building shells (CABS) on the other hand consider the building shell in a fundamentally different way. CABS can actively adapt their behavior over time in response to changing environmental conditions and performance requirements. Therefore, CABS have the ability to harness the natural energy avail-

1. I NTRODUCTION

Total energy consumption

Sleepwalking

Too little too late

Transformation

Time

Figure 1.1: Three scenarios for energy in buildings (WBCSD, 2009).

able in our environment much more effectively. These adaptive building shells not only offer a high potential to reduce the energy demand for lighting and space conditioning, but are at the same time also expected to introduce positive contributions to indoor air quality and thermal and visual comfort levels. The relevance of adaptivity in construction elements was also recognized by the recently completed IEA - ECBCS Annex 44, entitled Integrating Environmentally Responsive Elements in Buildings (Aschehoug and Andresen, 2008). Annex 44 concludes that the development, application and implementation of responsive building elements is considered to be a necessary step towards further energy efciency improvements in the built environment

1.2 Problem statement

By exploiting advances in material sciences, and due to the widespread availability of sensors and actuators, there are hardly any technological obstacles for making the building shell adaptive. Despite of the anticipated performance benets, application of CABS in current architectural practice is still very limited. Those examples that have been built are mainly restricted to either small-scale experimental projects or the high-prole, high-budget projects. VMRG, the Dutch faade builders association, every year publishes the high-performance faade book (van Maarsseveen and van Duijnen, 2008, 2009): An annual overview of the most complex, aesthetic, functional and promising building envelopes, the branch is proud of. Among the 150 featured projects, there is no example that can be classied as being CABS. Making building shells adaptive and more sustainable not only requires technological development but also faces some design process-related, nancial, legislative and social barriers (Hoffman and Henn, 2008; Williams and Dair, 2007; Zerkin, 2006). In principle, all these barriers can be traced back to a lack of awareness and fundamental understanding about CABS performance. Consequently, there is a great demand for effective tools and instruments that can be used in the design process of CABS. These tools
Within Annex 44, responsive building elements have been dened as building construction elements that assist to maintain

an appropriate balance between optimum interior conditions and environmental performance by reacting in a controlled and holistic manner to changes in external or internal conditions and to occupant intervention.

1.3. Objective and research questions should be able to unambiguously quantify the performance of CABS. Which in turn increases transparency about how performance benets during operation might outweigh rst cost arguments. In addition, these tools should be able to rationalize the uncertainty and associated risks that are introduced with the implementation of CABS. This allows higher ambition levels to be aspired with condence and helps to overcome the cautious attitude introduced by the anchoring bias towards prescriptive and standard design solutions (Klotz et al., 2010). The logic behind CABS seems quite straightforward: buildings that can adapt in response to changing boundary conditions are likely to achieve higher performance throughout the year. Designing and operating buildings with climate adaptive building shells is however far from trivial. CABS are inherently complex systems, consisting of several interrelated components that are working across various physical domains. These components together have to deal with trade-offs and resolve conictive performance objectives in real-time. Introduction of CABS moreover changes the traditional way in which buildings are designed; CABS are intrinsically dynamic, and therefore it becomes the design of a process rather than an artifact (Moloney, 2007). For these reasons, traditional design methods are likely deemed inappropriate and it will no longer be adequate to rely on past experience or rules of thumb. Using building performance simulation (BPS), makes it possible to predict why, how, and when a building will consume energy. BPS-tools are able to map the dynamics between the building and its systems, the ambient environment and occupants; and consequently can predict building performance. This instrument enables us to model the complex interactions of a large number of components across the boundaries of multiple physical domains. By doing this, the software tools create virtual building models that can be used to predict building performance for the following purposes: Choose correctly, understand why, explore possibilities, diagnose problems, identify constraints, develop understanding, build consensus, etc (Sokolowski and Banks, 2009). Because of these qualities, BPS can also be a valuable tool for performance prediction in the design stage of CABS, provided that the tool is used in a well-considered way. BPS already proved to be effective for supporting decision-making and code compliance checking in the building design process. However the application of BPS in relation to CABS is still largely unexplored.

1.3 Objective and research questions

This study is embedded within the strive for developing climate adaptive building shells, and focuses on the building performance simulation area. The primary objective of this study is to determine the role that building performance simulation can play in designing buildings with climate adaptive building shells. In order to fulll this objective, the following ve research questions are answered: 1. What are the characteristics of CABS and why is BPS needed? 2. What are the possibilities and restrictions of BPS related to CABS? 3. How to model and simulate CABS? 4. What is the performance of CABS? 5. What role can BPS play in CABS development?

1.4 Methodology
Various research methods can be applied in the search for answers to the formulated research questions. Within the scope of this thesis, literature survey, experiments, modeling and simulation are used. The research methodology can be subdivided in three distinct parts, which are described below in more detail. 3

1. I NTRODUCTION Analysis of CABS Before the role of BPS in the development of CABS can be indicated, it is essential to rst clarify what CABS really are. Starting with establishing an appropriate denition for CABS, the various concepts and underlying principles are investigated. For this purpose, an extensive literature review is conducted which covers the history of CABS, but also considers cutting-edge and even visionary adaptive building shell concepts. This resulted in a list of 100 examples of CABS that have been built, or are in research, design or development stage. A systematic classication method is then employed to indicate the particularities of individual examples but also to extract the overall trends and observations of CABS in general. Subsequently, this information is used to identify the critical elements that are essential for successful performance prediction in CABS. BPS in relation to CABS Building performance simulation is used more and more in design of buildings, however, its potential for design and operation of CABS is still largely unexplored. Therefore, this part starts with indicating the intrinsic qualities that computational building simulation offers for performance prediction of CABS, but also highlights the deciencies and possible ways to work around these problems. Based on the results of comparative software reviews and a literature survey, the capabilities of specic BPS-tools in relation to CABS are illustrated. This approach both includes the relevant physical domains and the available mechanisms for modeling the adaptive behavior. The knowledge obtained via this process is then generalized in a universal framework for modeling and simulation of CABS. Case Study In order to demonstrate the potential strengths of BPS in relation to CABS in more detail, the case study of Smart Energy Glass (SEG) is used. Smart Energy Glass is an innovative, energy generating window-concept that is switchable in three states, and therefore eligible as a representative example of CABS. First the concept of Smart Energy Glass is introduced and placed in the context of other innovative/adaptive window technologies. Following the methodology for modeling and simulation of CABS, a conceptual model is then derived that describes the adaptive behavior of SEG. The abstraction of the governing physical principles is complemented by conducting laboratory experiments on small-scale test samples. The assumptions of this model are validated in a reduced-scale experimental set-up using prototypes that are implemented in a real faade and exposed to atmospheric conditions. After some renements, this model is used in a simulation study that is employed to illustrate the capabilities of BPS for performance prediction of CABS. These simulations identify SEGs energy saving potential and also demonstrate the role that BPS can play in dealing with trade-offs between energy savings and thermal and visual comfort levels.

1.5 Outline
Figure 1.2 shows a graphical representation of the way this thesis is structured. The approach and ndings of this study are structured in two distinct parts; the main body of the thesis (numbered chapters), and the CABS overview in the form of Appendix A. In this respect, the term appendix might understate the relevance of the CABS overview. The booklet is believed to be most valuable when it is read in conjunction with the main text, rather than being studied afterwards. The body of the text is also further subdivided in two main parts. Chapters 2, 3 and 4 deal with CABS in general, whereas Chapters 5, 6 and 7 present the case study of Smart Energy Glass. The outline of the chapters is as follows: Chapter 2 starts with analyzing the characteristics of CABS and in this way determines the essential elements that are to be considered in performance prediction of CABS. In Chapter 3, the qualities of BPS in general and the capabilities of the available tools in particular with regards to performance prediction in CABS are specied. Chapter 4 relates the CABS overview to the analysis of BPS capabilities 4

1.5. Outline

1 Introduction

2 CABS

3 BPS

4 Modeling and simulation of CABS A CABS Overview

Casestudy

5 Modeling SEG

6 Validation

7 Performance of SEG

8 Conclusions
Figure 1.2: Structure of the thesis.

as it presents a general methodology to be used for performance prediction in CABS. Chapter 5 starts with building sufcient domain knowledge about Smart Energy Glass and subsequently describes the modeling approach that was used. The credibility of this models assumptions is assured in Chapter 6 which describes the methodology and outcomes and of the reduced-scale empirical validation study exposed to atmospheric conditions. Chapter 7 then uses this validated model in an application study that quanties CABS/SEGs performance benets in terms of energy saving potential and thermal and visual comfort. Finally, Chapter 8 concludes with summarizing the ndings of the entire study and provides recommendations for future work. This chapter also gives an answer to the question what role BPS can play in designing buildings with CABS.

CHAPTER
2.1 Introduction
Figure 2.1: Variations on the theme adaptive.

C LIMATE ADAPTIVE BUILDING SHELLS

Climate adaptive building shells, or CABS, is only one designation for a concept that is described by a multitude of different terms. Figure 2.1 summarizes the variations on the term adaptive in this context found in literature. Although all these expressions have a somewhat different meaning, they are often used interchangeably and in an ad hoc manner. In order to avoid ambiguity, the term CABS is adopted in this study, and is dened as follows: A climate adaptive building shell has the ability to repeatedly and reversibly change its functions, features or behavior over time in response to changing performance requirements and variable boundary conditions. By doing this, the building shell effectively seeks to improve overall building performance in terms of primary energy consumption while maintaining acceptable thermal and visual comfort conditions. The building shell is interpreted here as those construction elements that form the division between inside and outside air. This includes both the opaque and transparent wall elements as well as the roof.

2. C LIMATE ADAPTIVE BUILDING SHELLS Adaptive behavior in internal partitions, oors and foundation is not taken into account. In common architectural practice, the design of the roof frequently plays a subordinate role. Obviously prompted by the fact that for most buildings the roof lies outside the perceptible eld of view. Nevertheless, their surface area is often substantial compared to faade surface, and in terms of energy exchange, neglecting the roofs impact is unjustiable. For these reasons, this study is not limited to adaptive faades only, but also includes adaptive roof concepts; Essential in the denition of CABS is the word effectively. This term implies that performance aspects should be deliberately considered in the design of adaptive behavior and not be based on coincidences. Current faades tend to be developed in layers, resulting in a subdivision of the problems and also a subdivision of only partial solutions (Knaack et al., 2008). As Lichtenberg (2005) puts it, construction has evolved in a way called innovation by addition; an approach that leaves room for substantial improvements. CABS attempt to reconcile the multiple performance aspects in a more holistic way. Resolving occurring problems via the layered approach is not what the concept of CABS aims for. Furthermore, the term effectively also implies some kind of of automatic but not necessarily automated control. Too many projects have demonstrated that for example adding blinds that are irregularly controlled by occupants will not fundamentally change the performance picture (Reinhart and Voss, 2003; Selkowitz et al., 2003; Mahdavi et al., 2008). This chapter continues with a discussion of the performance of the building shell in general, and the performance benets of CABS in particular. Next, the characteristics of various CABS concepts are analyzed, which leads to a methodology for classifying CABS, described in Section 2.4. Finally, the chapter concludes with the results of a more systematic analysis of CABS, which reveals the general trends and patterns of this broad concept.

2.2 CABS performance

Before the performance of CABS can be specied, it is important to rst discuss the performance aspects of the building shell in general. Hutcheon (1963) presented the following, widely accepted list of eleven principle functional requirements to be considered in the design process of building envelopes: control heat ow; control air ow; control water vapor ow; control rain penetration; control light, solar and other radiation; control noise; control re; provide strength and rigidity; be durable; be aesthetically pleasing; be economical. In order to obtain adequate building designs, all these functional requirements need to be considered and satised simultaneously. Because the requirements are strongly interrelated and sometimes even conicting, this is not always a simple task (Rivard et al., 1995). The relative importance of the various functional requirements is different for every project and is dictated by the context of the project and 8

2.2. CABS performance the clients wishes. Nevertheless, all functional aspects are bounded by certain minimum requirements, typically prescribed in the form of building standards. The performance of a technology refers to its technical effectiveness in a specic evaluation or a set of applications (Kolarevic and Malkawi, 2005). When this denition is projected to building performance, it implies consideration of both a certain environmental context, and the interests of the various stakeholders involved in the building process. The performance of contemporary building shells is best exemplied via the Triple Bottom Line philosophy (Figure 2.2). This business concept stresses that for every successful technology a right balance needs to be achieved between performance attributes for people, planet and prot.

Figure 2.2: Triple Bottom Line principle (Elkington, 1998).

In the ideal case, the building shell is designed in such a way that performance of the whole building, in terms of environmental, economical and social attributes is maximized. The buildings environment however changes over time, and this also applies for occupants preferences. By denition, changes thus also occur in building performance, in ways that tend to divert from what was intended. In order to still be able to provide satisfactory performance all year long, buildings need to have some way to cope with these changing conditions.

2.2.1 Robustness vs Flexibility

Robustness and exibility are the two design strategies which refer to a systems ability to handle change (Saleh et al., 2009). Robustness is often described as a systems resistance or immunity to change. It is about de-sensitizing a systems performance or quality characteristics to changes in the systems environment, without eliminating the cause of these changes. Flexible systems on the other hand are dened to be systems designed to maintain a high level of performance through real-time changes in conguration when operating conditions or functional requirements change in a predictable or unpredictable way (Olewnik et al., 2004). Once a building has been constructed, its thermophysical and optical properties are xed, preventing changes in the building shells state. Therefore these faades are rigid systems that will not be able to give optimal performance except for a few time periods during the year. The only way in which conventional buildings can be exibly adapted to changes in functional requirements is via adaptive reuse, or via renovation projects like lateral or vertical extensions, structural alterations and refurbishment (Brand, 1994). For coping with real-time changes in its context, conventional buildings have to rely on the concept of robustness. Unfortunately, maintaining high performance in robust building designs comes at a cost. When buildings are left in free-running state, this will lead to uncomfortable or even
Application of exible design does not remove the demand for robust control strategies.

2. C LIMATE ADAPTIVE BUILDING SHELLS unhabitable circumstances during the largest part of the year. The only way to mitigate these unwanted conditions is via energy intensive HVAC and lighting systems, which transform the indoor space in a manufactured rather than a mediated environment (Addington, 2009). And in turn, this leads to high economical and ecological loads, thereby disturbing the balance in the Triple Bottom Line approach. As opposed to the conventional building shell, CABS do have the ability to change their functions, features or behavior over time, and hence they can take advantage of the concept of exibility. The fundamental need for adaptive or exible systems stems from the presence of multiple performance requirements, operating conditions or users preferences. Systems engineering has developed a thorough foundation with terminology that can be used for understanding the performance benets of exible systems, including CABS. The qualities of CABS and the associated performance benets of being exible can be described with three different terms: adaptability, multi-ability and evolvability; these concepts are discussed in more detail in the following subsections.

2.2.2 CABS as exible systems

Adaptability In systems engineering literature, adaptability is dened as the ability of a system to deliver intended functionality under variable conditions through the design variables changing their physical values over time (Ferguson et al., 2007). As a consequence, CABS can immediately act in response to changes in ambient conditions. This offers huge energy saving potential compared to conventional buildings because the valuable energy resources in our environment can be actively exploited, but only at times when their effects are deemed benecial. A second energy-related aspect of CABS adaptability is the fact that it allows for effectively taking advantage of the inertia of thermal energy stored in the buildings construction. In addition to providing energy savings, adaptability is also a useful means for improving occupants comfort. In pursuit of fully utilizing CABS energy saving potential, the operational control strategy has to take care that indoor environmental quality does not get compromised. In conventional buildings this can only be done by consuming energy in articial lighting- and HVAC systems. However, because of their adaptivity, CABS can act as climate mediator, negotiating between comfort needs and the ambient environment. Consequently, building services and the building shell can work in symbiosis to resolve the occurring trade-offs in the best possible way. Multi-ability The concept of multi-ability originates from the existence non-simultaneous performance requirements, or the need to fulll new roles over time. A key aspect of multi-ability is that the modiable functions are part of the systems initial functional requirements (Ferguson et al., 2007), and in this way, it allows for a plurality of optimized states to address change. Multi-ability in CABS is best explained via an example: Bloomframe balcony can dynamically change its function between window and balcony on-demand, depending on ambient conditions and users wishes (Figure 2.3). By doing this, multi-ability facilitates resource efciency, which adds to the list of benets already mentioned in the previous paragraph. A second application of multi-ability is spatial versatility. At the same point in time, the properties of the envelope can be different for various positions of the building shell. In this way, the building shell can react to dissimilar conditions at the different orientations of the building, or distinct comfort preferences requested by individual users in separate zones. Evolvability Future building requirements and circumstances are not always known in advance, or may even be unpredictable. Building shells that can evolve over time are a means of overcoming the uncertainty of these unforeseen events. Evolvability, sometimes called survivability (Siddiqi and de Weck, 2008) can be employed to deal with changes in (i) system objectives, (ii) operating environment or (iii) the system itself. In CABS, changing system objectives come into play when for example ofce buildings turn into schools or apartments into shops. In case of these organizational function changes, existing building 10

2.2. CABS performance

Figure 2.3: Bloomframe dynamic balcony (Bloomframe, 2010).

services often turn out to be inadequate, and need to be replaced or extended. An example of changing operating environment is the climate change we are likely to face in the (near) future. Although intensity of the effects is still under debate, in traditional buildings it will probably lead to either increased energy consumption for cooling, or in case cooling capacity is limited or unavailable, to severe risks for overheating. Not only requirements and ambient conditions, but also the building and its systems itself naturally evolve over time. Two typical examples are weathering of the building envelope and soiling of ducts and lters. Both effects may have negative impacts for the buildings energy balance: the rst can lead to increased absorption of solar energy, and the latter is likely to reduce operational efciency of the HVAC system. In all of the three instances mentioned, application of CABS increases the chances that the building can continue operation as intended by changing functions, features or behavior, without suffering from the potential negative impacts of unforeseen future conditions.

2.2.3 Performance benets of CABS

Now the performance aspects related to CABS exibility are highlighted, it is worthwhile to return again to the Triple Bottom Line principles. As the previous sections pointed out, CABS can offer positive contributions to all the three aspects: people, planet and prot. When CABS are operated effectively, their application holds the potential to result in energy savings compared to conventional building shells. As this suppresses the need for import of fossil fuel based energy, it relieves the load on the earths environment (i.e. planet) and also directly inuences the endusers energy bills (i.e. prot). In addition, a large number of CABS is equipped with building integrated renewable energy generation components. Environment and economy also benet from CABS material resource efciency and the fact that CABS facilitate evolvability which will probably lead to increased building lifetimes. CABS performance benets for people are typically expressed in terms of improved indoor environmental quality. This not only includes thermal comfort, but also occupants satisfaction with the luminous environment (visual comfort), acoustics and air quality. Increasing evidence furthermore shows that comfortable working environments are associated with improved health and productivity levels (Heerwagen, 2000; Haynes, 2008; Wargocki, 2009). Focusing on providing high comfort levels thus directly inuences the economic aspect as well. It should however also be noted that CABS do not merely introduce positive contributions to the Triple Bottom Line. Instead of only saving energy, several CABS actually consume electricity for facilitating the adaptive behavior. In addition, CABS typically require additional components and subsystems for making the building shell adaptive. This requires the use of more raw material and also asks for higher investment costs. However, when operated in an intelligent way, these drawbacks of CABS are expected to be off-set by the benets. These, and several other challenges and barriers that need to be overcome for successful application of CABS will be discussed further in Section 2.3.4. Besides the direct performance attributes, CABS can also have impact on more indirect performance aspects. These aspects are sometimes summarized as being cultural or socio-economic performance. Adaptivity in the building shells offers architects and designers an attractive additional design variable: 11

2. C LIMATE ADAPTIVE BUILDING SHELLS time. This can have dramatic impacts compared to the traditional aesthetics of a faade, which makes that building equipped with CABS often results in exceptional or iconic architecture. Eventually, this might also lead to what is called the Bilbao effect; after Frank Gehrys Guggenheim museum in Bilbao, Spain (Jencks, 2005). This expressive landmark transformed a whole city from being in a dilapidated state to a town with inviting allure, mainly because of its architecture. In the rst year after opening, the building attracted 1,3 million extra visitors to Bilbao which brought the city huge nancial growth and international prestige (Plaza, 2000).

2.3 CABS Analysis

Methodology For the analysis of CABS, an inductive research approach was used in this thesis. This type of research tries to retrieve the general characteristics of CABS by studying and observing a large number of individual examples. A structured scientic literature survey was conducted with the goal of exploring the full diversity of climate adaptive building shells. For this purpose, numerous pairs of search terms were used. The rst keyword was selected from the terms in Figure 2.1, while the second keyword was always picked from one of the following words: architecture, building, cladding, enclosure, envelope, exterior, faade, glazing, roof, shell, skin, wall or window. After an initial round of investigations, the conclusion was drawn that CABS is still an immature eld of research. The number of existing products, built examples, design proposals and patents published in books, journal- and conference papers is limited, and extensive review papers have not yet been published. Furthermore, it was observed that these references tend to run in circles, with Jean Nouvels shutters and the switchable glazings as the most often cited examples. The application area of CABS however ranges from built examples, successfully operating for many years, to the wildest utopian concepts. But because the outcomes of these developments emanate from creative processes, they are far from always published in scientic literature. In order to still be able to obtain a representative number of CABS examples, the area of attention was extended to other sources of information. The second round of the literature survey also included both architecture and design magazines (e.g. Architectural Record, Architecture Week, Design Intelligence and Metropolis Magazine) and architecture, design and sustainability weblogs like ArchDaily, DesignBoom, Dezeen, Greenlineblog, Inhabitat and Wired. Because of CABS dynamic nature, CABS also ideally lend themselves for video clips. Therefore YouTube also turned out to be a valuable source of information. The ndings of this literature survey are reported in the following sub-sections. This starts with discussing the four critical elements that constitute the adaptive behavior in CABS: climate, people, timescales and mechanisms. After that, the various concepts, inspirations and backgrounds of the adaptive principles are described. The existing topologies for controlling the adaptive mechanisms and behavior is the subject of Section 2.3.3. Finally, the challenges and barriers that need to be overcome for general adoption of CABS are highlighted. An approach to the more systematic classication of CABS, and the outcomes of this process are described in Section 2.4. The investigation of CABS also resulted in the CABS overview in the form of Appendix A. Cross-references to this overview are often given in the main text, and are indicated with superscripts, referring to the particular CABS-number.

2.3.1 Adaptive behavior

Climate The outdoor climate forms a main constituent of CABS adaptive behavior, as its name already suggests. Climate (from the Greek word klima meaning inclination) describes the long term atmospheric conditions observed at a site. Weather is dened as the individual short term (e.g. hourly or daily) 12

2.3. CABS Analysis observations of climatic features (Bougdah and Sharples, 2010). Climate has been regarded by architectural practice both as the enemy or as a companion. The rst design philosophy tries to eliminate the impact of the climate on buildings. This is often manifested via highly conditioned buildings with sealed envelopes which makes the building insensitive to its environment, and uncouples the building shell from its role as environmental moderator (Addington, 2009). The latter works in close conjunction with its environment and takes full advantage of the positive inuences found in nature. In this respect, a lot of things can still be learned from vernacular architecture. With its adaptive capabilities, CABS decidedly belong to the second category. The most important weather elements of interest in architecture and buildings, together with their corresponding implications are: dry bulb air temperature - thermal comfort, heating and cooling; relative humidity - thermal comfort, condensation, mould growth; precipitation - drainage, loading, damage; wind speed and direction - energy, ventilation, comfort, loading; solar radiation - daylight availability and useful solar heat gains; cloud cover - diffuse daylight and radiation to the sky. Variabilities in all of these elements are also actively exploited by CABS. Often these impacts are hidden in the control algorithms that drive the adaptive behavior. However, there are also some CABS that directly react in response to the changes in for example ambient temperature23,27,28,77,92 , intensity of solar radiation47,65 , rain intensity64 , wind speed72,88 or relative humidity84 . Additionally, some CABS use the position and movement of the sun through the sky to optimize the collection or rejection of solar energy29,35,49,86,90 . Because the sun follows a prescribed path, no complex control algorithms or reactive behavior is necessary. Therefore, the adaptive behavior in these CABS rather transforms into a choreography that can be determined in advance. The perceived weather conditions on a building site are closely related to its geographical location. Numerous climate classications are in use to map the global climate in different zones. In architecture, typically four different zones are distinguished: cool, moderate, hot-dry and warm-humid (Szokolay, 2004; Cook, 1996). Because of their adaptability, most CABS can be applied universally. However, several CABS are specically tailored to a certain climate-zone; some CABS are suited for cool climates33,41 while others are more prosperous in warm and sunny weather40,43,81 . The orientation of the building determines to a large extent the actual climatological impacts acting on the building. Some buildings with CABS actively manipulate these inuences by changing their orientation. Weather forecasts are capable of predicting short-term weather conditions reasonably well. However, these forecasts are always subject to certain condence bands and occasions of unexpected behavior. That is the reason why anticipating to future weather variables will always retain a certain sense of uncertainty. Even larger unpredictability is associated with expected long-term climate conditions. The built environment is not only one of the main instigators of the expected climate change; future buildings also have to withstand the effects of this global warming (Roaf et al., 2005). Because CABS can draw upon the concept of evolvability, the detrimental effects for CABS are probably less severe. People The adaptive behavior of CABS is not limited to reaction of changes in exterior climate only. Tuning behavior in response to variable requirements imposed by people is at least equally important. Providing comfortable spaces is after all the reason why most buildings are being built. Some of the CABS examined are able to respond to the mere presence of persons05,08 . However, in terms of fullling comfort requirements, it makes more sense to adapt in response to occupants performance preferences. 13

2. C LIMATE ADAPTIVE BUILDING SHELLS The presence of people inuences building performance in two different ways. In the rst place it introduces contaminants, respiratory products and internal heat gains via metabolic activity and the use of equipment. In the second place, occupancy dictates the need for comfort. Daylighting needs witnesses to be appreciated and indoor air quality is of secondary concern when nobody inhales the indoor air. Roughly spoken, buildings are uninhabited for about half of the time; dwellings during the day, schools, ofces and factories at night. Arrival and departure, and hence occupancy is typically a stochastic process (Page et al., 2008), although periodic patterns can be discerned. Unexpected behavior and occasional events further complicate the prediction of occupancy. Consequently, occupancy and thus comfort requirements are subject to both high variability and uncertainty which clearly supports the strive for adaptive rather than rigid building shells. It is important to take occupants requirements into account already in the design-stage of CABS, especially because low-energy architecture is relatively sensitive to the impacts of tenants. Some authors prefer to use the word participant instead of user to describe the way people should interact with architecture (Fox and Kemp, 2009). Occupants preferences are typically expressed in terms of setpoints and constraints, preferably in the form of some measurable quantities. Together this results in aggregate comfort performance indicators (e.g. Predicted Mean Vote (PMV) or Daylight Glare Probability (DGP)) or comfort classes. The actual perception of comfort is however a much more complex process. Appreciation of indoor environmental quality in the widest sense involves all our sensory perceptions; these processes are interrelated and take place across various physical domains. The fact that comfort perception is not merely physically-based, but also invisible mental processes play a role, further complicates the task of anticipating comfort perception. Adaptation is not only observed in the building shell, but also plays a role in peoples perception of thermal comfort. People respond to discomfort in one or both of two ways: either by trying to adjust their environment to suit their requirements or by adjusting their requirements to match what the building provides (Nicol and Humphreys, 2009). The fundamental assumption of this approach is expressed by the adaptive principle: If a change occurs such as to produce discomfort, people always react in ways which tend to restore their comfort (Roaf et al., 2005). Three types of adaptive behavior are distinguished: behavioral adaptation (adjustment), physiological adaptation (acclimatization) and psychological adaptation (habituation) (Brager and de Dear, 1998). Because of this adaptivity, comfort requirements can be less tight; which also increases the option space for saving energy while at the same time supporting favorable comfort conditions. Timescales As demonstrated before, buildings are subject to various environmental impacts. All these inuences typically occur at a characteristic temporal resolution, ranging from the order of (sub)seconds to impacts that are only perceptible during the buildings whole lifetime. The adaptive actions observed in CABS are either continuous or discrete. Discrete adaptations are suddenly triggered after a certain threshold-value or period of time has been exceeded, whereas the continuous type of adaptation typically follows gradual transition paths. The following seven characteristic timescales were discerned in the analysis of CABS. seconds These short-term uctuations typically have a stochastic nature. The swift variations in wind speed and direction for example, enable movement in the Wind shaped kinetic pavilion98 . CABS that behave in response to peoples vicinity, like Aperture08 and Living Glass54 are also capable of changing their conguration on this short time-horizon. minutes Cloud cover and daylight availability have a characteristic time in the order of minutes. Thats the reason why all CABS that aim to optimize daylight utilization and solar shading for reducing energy demand and increasing visual comfort should be able to alter their degree of transparency at a rate of change in the order of minutes. Additionally, changes in occupants thermal and visual comfort preferences also typically vary at this temporal resolution. 14

2.3. CABS Analysis hours The angular movement of the sun through the sky is in principle a continuous process. However, CABS that track the path of the sun35,38 , typically adjust in the order of hours. Fluctuations in air temperature, both internal and external can appropriately be discretized in hourly values as well. CABS that directly adapt in response to temperature stimuli27,28,40 are consequently also classied in this category. diurnal Presence of occupants in buildings normally follows diurnal patterns. In addition, these daynight cycles are also noticeable in meteorological boundary conditions like intensity of solar radiation, and ambient air temperature. Several CABS are specically designed to take advantage of this xed 24 hour pattern. Examples are: Blight16 , Nocturnal ventilator61 and Solar barrel wall79 . seasonal Adapting to changing conditions occurring across the seasons is perhaps the most elegant application area of CABS. Winter, spring, summer and autumn all impose widely different boundary conditions, and buildings that can adapt to these changes are expected to provide substantial performance benets. Allwetterdach07 and Smart Shade77 are examples of CABS that t in this category. annual Buildings are also subject to impacts varying from year to year. One can for example think of hot summer heat waves, occasional rainy periods or extremely cold winters. Anticipating to these annual variations is not the major concern of any of the 100 CABS analyzed. In conventional buildings however, these unexpected environmental conditions almost always lead to increased energy consumption. Because CABS have the ability to adapt to these changing conditions, higher energy consumption is likely to be avoided. decades Impacts that occur at these timescales are surrounded by uncertainty and can impossibly be incorporated in the buildings initial functional requirements. This type of changing long-term conditions can come from the outside (e.g. climate change, changing urban environment) or from the inside (e.g. organizational function changes of the building). But changes in the faade itself are also possible. An example is the weathering or soiling of the envelope which may give rise to a change in the absorption coefcient of the outer surface (Heaberlin, 2004; Berdahl et al., 2008). Because CABS can draw upon the concept of evolvability, they have a higher change to withstand these long-term changes without performance losses. Mechanisms The mechanisms that foster adaptation in CABS can be categorized in two classes. The adaptive behavior is either based on a change in properties or behavior at the micro-level, or at the macro-level; although combinations are also possible. The distinction between both mechanisms is raised by the spatial resolution at which the adaptive actions take place. Both concepts are described below in more detail. Micro-level Adaptive mechanisms on the micro-level can be further subdivided in two groups, following the same approach as proposed by Addington and Schodek (2005): 1. If the mechanism affects the internal energy of the material by altering either the materials molecular structure or microstructure then the input results in a property change of the material. Adaptation in CABS may alter either thermophysical41,78,89,95 or optical properties02,23,32,37,65,75 . 2. If the adaptive mechanism changes the energy state of the material composition, but does not alter the material, then the input results in an exchange of energy from one form to another. In CABS, this feature typically applies to photovoltaic cells which are used in no less than twenty of the CABS analyzed. Other examples of energy exchange at the micro-level include: electricity to light16,21,42,47 , and electricity to thermal radiation 01,33 .
An energy state is any identiable collection of matter that can be described by a single temperature, pressure, density and internal energy

15

2. C LIMATE ADAPTIVE BUILDING SHELLS The conditions which force the changing behavior may either be imposed via a direct energy input emanating from the ambient environment, or a control signal from occupants preferences or supervisory operational strategy. Other features of CABS acting on the micro-level are the fact that the number of components is limited and no moving parts are involved. Macro-level The second class of adaptive mechanisms consists of adaptations where movement can be observed with the naked eye. This type of building shells is often referred to as kinetic envelopes, which implies that there is a certain kind of motion. Adaptation on the macro-level usually results in a change in the building shells conguration via moving parts, which can be (i) individual components supplemented to the building shell20,29,35,63 , (ii) subsystems of the building shell itself08,22,73 as well as (iii) movement of the faade as a whole69,86,96 . The types of motion observed vary widely and are typically described by one of the following gerunds: folding, sliding, expanding, creasing, hinging, rolling, inating, fanning, rotating, owing, curling, etc. The most frequently encountered driving principle for the mechanical transformation is still the electromotor. However, the CABS overview also shows some examples of more innovative14,28,54 and sustainable27,40,76 actuation principles. Moreover, the CABS overview includes some examples where not only the building shell, but rather the whole building is dynamic25,45,98 . Randl (2008) provides an extensive overview of the history of these rotating buildings.

2.3.2 Concepts
The study of CABS is not an isolated concept without background. In fact, it draws its inspiration from many disciplines in- and outside the world of buildings and architecture. The following sections briey review these concepts and put them in the right perspective in relation to CABS. Polyvalent wall What is needed is an environmental diode, a progressive thermal and spectral switching device, a dynamic interactive multi-capability processor acting as a building skin. The diode is logically based on the remarkable physical properties of glass, but will have to incorporate a greater range of thermal and visual adaptive performance capabilities in one polyvalent product. This environmental diode, a polyvalent wall as the envelope of a building, will remove the distinction between solid and transparent. Davies (1981) Back in 1981, Mike Davies, a British architect working for Richard Rogers company proposed the idea for The polyvalent wall as depicted in Figure 2.4. This article is typically renowned for the integration of several functions into one layer and can be seen as the keynote for research and imagination in integrated adaptive building envelopes. Although his imaginary concept for a wall for all seasons was way ahead of his time it has served as source of inspiration for many. Smart materials Not all materials in the built environment are used as construction elements only. The performativity of some new materials goes further than just being ordinary bricks, concrete, steel or glass. Smart materials is a relatively new term for materials and products that have changeable properties and are able to reversibly change their thermophysical or optical properties in response to some physical or chemical inuence (Ritter, 2007). Apart from only bothering about structural properties, durability and aesthetics, application of smart materials also demand to focus on what we want the material to do. Smart materials are sometimes used as bulk material, but more often nds its application in coatings; the boundary layer of a material which can drastically change total performance (Klooster, 2009). The ve characteristics that distinguish a smart material from the more traditional materials used in architecture are given in Table 2.1. 16

2.3. CABS Analysis

Figure 2.4: Schematic representation of the polyvalent wall (Davies, 1981).

Table 2.1: The ve characteristics of smart materials (Addington and Schodek, 2005).

transiency selectivity immediacy self-actuation directness

the response is instigated by more than one environmental state the response is discrete and predictable the response happens in real-time the response is controlled inside the material or system the response acts locally

Because of their dynamic and adaptive nature, smart materials are well-suited for application in CABS. The following ten different smart materials were encountered in the 100 examined CABS-concepts: electrochromic21,32,66 , electroluminescent16,78 , hygroactive84,97 , photochromic65 , piezoelectric14,43,87 , thermochromic23,92 , thermoelectric01 and thermostrictive27,28,40,61,77,97 materials, shape memory alloys (sma)54,67,76,87 and phase changing materials (pcm)23,41,78,95 . Discussing the ins and outs of all these smart materials in detail lies outside the scope of this thesis. However, it can be observed that most of the smart materials have names that are combinations of adjectives, mainly derived from the Greek language. The working mechanisms behind each type can therefore easily be deduced from their respective names. The concept of smart materials shows great promise, however, their application is still rare. Now the issue arises of how to bypass the gadget-isation stage of smart materials (Kula and Ternaux, 2009). As with all other building materials, the following conditions absolutely need to be fullled: (i) the cost/benet ratio needs to be realistic, (ii) the technology has to comply with the existing labor and assembly practices of the building industry, and (iii) the concepts need to be accepted from an aesthetic point of view (Knaack et al., 2007). Biomimicry Human subtlety will never devise an invention more beautiful, more simple or more direct than nature, because in her inventions, nothing is lacking and nothing is superuous Leonardo da Vinci Some buildings are designed to look like natural organisms (Aldersey-Williams, 2004). Most of time time, this results in funny or iconic buildings, but imitation of organisms appearance can also be functional. Structural engineers have successfully studied phyllotaxis (i.e. the arrangement of leaves 17

2. C LIMATE ADAPTIVE BUILDING SHELLS in plants) for optimization of their constructions. But what if buildings not only look like their natural counterpart, but also behave like living organisms? Biomimicry from the Greek words bios (life) and mimesis (imitation) is a design philosophy that seeks for sustainable solutions by emulating natures time-tested ideas (Benyus, 1997). Citing John et al. (2005), the natural world has an immense amount to tell us about how to achieve sustainability. It uses energy far more efciently and effectively and is capable of producing materials and structures that are far more benign than anything we have achieved in industry. Indeed, natural adaptation found in ora and fauna has proven to be a valuable source of inspiration for the design of CABS (Gruber, 2008; Pedersen Zari, 2007): The hibiscus ower opens during the day but closes up at night; A chameleon can change the color of its skin under various environmental circumstances; Gas exchange through stomata is a function of relative humidity; A distinct type of biomimicry related to CABS is the concept of tropisms. A tropism is dened as a movement found in nature fostering adaptability to change and occurs over single life-spans. Both phototropism (i.e. changing in response to light) and heliotropism (i.e. changing in response to the sun), have been effectively transformed to buildings in CABS-concepts enabling to actively collect or reject solar energy (Vermillion, 2002). A clear parallel exists between CABS and the evolution principles of Darwin: the capacity to survive depends on the ability to adapt to a changing environment. However, the superposition of biological paradigms to architecture is not straightforward, and despite lots of efforts in this cross-disciplinary research area the open question still remains: How to go beyond the metaphor? CABS versus Passive design Last decades, the design of low-energy buildings has diverged into two different directions: active technologies and passive design strategies. The term passive refers to buildings where the design of the construction and form of the building itself, as opposed to its servicing, is mainly responsible for capturing, storage and distribution of solar energy, normally with the aim of displacing other fossil fuels for space heating (Porteous and MacGregor, 2005). Often, this results in very airtight buildings with a high insulation standard and balanced ventilation with heat recovery. In warm climates on the other hand, passive cooling strategies try to utilize the building design in combination with natural energy available in the environmental context with the aim of avoiding overheating without using electricity-powered cooling systems. In this context, the term active is often considered as passives counterpart, as it represents the energy intensive HVAC-systems and articial lighting the passive design tries to circumvent. By denition, CABS are also associated with active technology. This paradox may lead to the misleading interpretation that CABS and passive design are conicting concepts. In fact, both CABS and passive buildings are designed with the same goals in mind; reducing energy demand while maintaining good comfort levels. Compared to CABS, passive design offers some distinct advantages. Passive solutions are often derived form traditional technologies, and therefore they are reliable and easy to implement, which also makes them relatively cost-effective. In addition, they have no moving parts, there is no need for complex control systems, and no energy is dissipated for facilitating adaptive behavior. On the other hand, potential drawbacks are also associated with the passive approach. Passive houses are sometimes accused of providing poor indoor air quality and the sealed building envelopes have been linked to negative health issues. In addition, there is the risk of not always meeting performance requirements, often expressed in terms of overheating. Because the passive approach heavily relies on the concept of robustness, passive designs typically have difculties to cope with unexpected conditions. Eventually, this has been the 18

2.3. CABS Analysis motive underlying the search for means of making buildings responsive and dynamic (Wigginton and Harris, 2002). To conclude, the concepts active and passive are not contradictory, and both strategies can mutually enhance each other. This symbiosis is encountered several times41,70,77,79 in the CABS overview. Passive design urges the need to carefully design and tailor the building to local conditions (e.g. site, climate, surroundings), this is also a very good point of departure for the design of CABS. Furthermore, it should be noted that it is not fair to expect CABS to compensate for the deciencies of poor building designs. Intelligent buildings Intelligent buildings are dened as buildings that should be sustainable, healthy, technologically aware, meet the needs of occupants and business and should be exible and adaptable to deal with change (Clements-Croome, 2009). Because of this last argument, CABS can be part of an intelligent building. However, the realm of intelligent buildings is much wider than that of CABS as it also includes surveillance cameras, security and access control, elevators, re safety alarms, etc. Nonetheless, several authors have embraced the term intelligent building envelope and alike as a substitute for CABS (Aschehoug et al., 2005; Kuo et al., 2009; Ochoa and Capuleto, 2009; Wyckmans, 2005). The word intelligent implies some human characteristics that give the building shell the ability to learn, adjust and respond instinctively to its immediate environment (Wigginton and Harris, 2002). To achieve this target, all components and subsystems constituting the building shell have to be coupled to a supervisory control system, which may be the building energy management system. The third skin People face the same difculties as the building envelope in the sense they both have to deal with variable conditions in their environment (e.g. temperature differences and the impacts of wind, rain and solar radiation). Human skin has developed several ingenious functionalities to deal with these intricacies. Depending on a specic situation, the skin is capable of autonomously changing its (thermoregulatory) functions by e.g. releasing sweat, creating goosebumps, darkening color (tanning), shivering, changing porosity, controlling blood ow (vasoconstriction, vasodilation) or raising hairs (Wigginton and Harris, 2002). Furthermore the skin also communicates information about physical and emotional states. Our second skin is often conceived as the clothes we wear. By adjusting type and amount of clothing, we actively assist our physiology to cope with the negative impacts coming from our environment in order to sustain the feeling of comfort. Carrying this analogy further, it will probably be no surprise that the building shell can be regarded as the third skin (Drake, 2007). The metaphor has been used in attempts to transform the building shell in a living membrane. Thereby mimicking humans dynamic exchange of energy and matter with the environment, and especially focusing on the skins sensitivity, multi-functionality and ability of self-control. Despite various honorable efforts91,97 , human skins outstanding adaptive qualities are however still unsurpassed. Nanotechnology Nanotechnology is a novel area in multidisciplinary materials science, which promises to open up new horizons in the way our buildings are designed and constructed. The area of nanotechnology is concerned with the study, creation, application and manipulation of structures, molecular materials and systems that are between 1 and 100 nanometer in size (Klooster, 2009). On this molecular scale, forces and effects that do not play a role on the large scale will become dominant. Recent advances in scanning electron microscopes have made it possible to see and manipulate matter at this small scale. Hence, we are now able to explore the possibilities of material concepts and composites that cannot be found in nature. Some of the envisioned concepts include construction materials that are: self-diagnostic, able to communicate, can grow or evolve over time. Leydecker et al. (2008) presents a comprehensive 19

2. C LIMATE ADAPTIVE BUILDING SHELLS overview of nano-applications in architecture and design. In general, the nano-scale products under development can be divided in two different groups. The rst type currently forms the majority of nano-applications in design and architecture and involves enhancing the functionalities of (transparent) surfaces (Yeadon, 2008). Currently, several products with protective nano-coatings are on the market. These coating are for example: anti-bacterial, antigrafti, re-proof, scratch-proof, self-cleaning and abrasion resistant. But also some CABS are based on materials with enhanced functionalities on the nano-scale; examples are CABS with phase changing materials95 , and all switchable glazings32,65,92 . The second type of nano-materials aims at developing an entirely new class of materials and products that did not exist before. Nanotechnology offers the conceptual possibility to build materials bottom up. This opens up the possibility to assemble materials or structures on the molecular level, which enables us to program the overall macro-scale properties of a material (Kula and Ternaux, 2009). Eventually, nanotechnology may afford us to create revolutionary, ultra-performing materials which exhibit unprecedented behavior. Recent advances of this bottom-up approach of interest in architecture are super adhesives, carbon nanotubes and the extremely insulating aerogel. Active window02 , a CABSconcept currently under development at the University of Kassel, Germany is also a technology that ts into this category. This innovative light control systems integrates nanoscale structures in a macro object; a window (Klooster, 2009).

2.3.3 Control
Adding adaptive features to the building envelope does not directly guarantee effective or successful operation. In order to achieve the desired goals, CABS have to respond to changing conditions in a sound way. In addition, some kind of concerted activities have to take place; the various subsystems in the faade have to cooperate, together and with other building services, to resolve conicts, handle trade-offs and satisfy performance requirements synergetically and in the best possible way. Two types of control concepts are distinguished: closed-loop and open-loop control (Figure 2.5). The subtle semantic difference between the words automated and automatic marks the distinction between these control concepts. As the following sub-sections show, both types of control can work automatically, but only the rst type is automated.
CABS
input

Sensor

data

Processor

control signal

Actuator

action

input

CABS

action

Controller

Figure 2.5: The two control concepts for CABS: (a) closed-loop control (b) open-loop control.

Closed-loop control The structure of closed-loop controlled CABS consists of three basic elements: sensors, processors and actuators (Teuffel, 2004). The structure is illustrated in Figure 2.5(a), which also shows a fourth element; the controller which may or may not be present for operation of CABS. A sensor is a technical component that can register specic physical or chemical ambient conditions (e.g. temperature, moisture, pressure, sound, brightness or acceleration). These parameters are 20

2.3. CABS Analysis recorded by means of physical or chemical effects and invariably transformed into quantities (e.g. electrical signals or digital data) that can be interpreted by a processor (Klooster, 2009). Sensors are typically classied according to the kind of stimulus they can detect. The processor is the component that collects signals and data coming from all sensors and controllers. This computational logic is interpreted and subsequently results in a univocal control signal which is passed on the the actuator(s). The third component, the actuator, is a device that converts input in the form of a control signal into a mechanical, chemical or physical action (Addington and Schodek, 2005). In CABS this can both result in a property change on the micro-level or a kinetic action on the macro-level. Adaptive actions instigating components to move are actuated via various mechanisms, including magnetic switches69 , pneumatic pistons04,05,36,57 , shape memory alloys54,67 , piezo-electrics devices14 , hydraulic cylinders70,86 and electromotorse.g. 07,17,93,94 . Closed-loop controlled adaptations on the micro-level are typically actuated via electrical signals. The distinguishing quality of closed-loop controlled CABS is the ability to take advantage of feedback. Feedback implies that effects of the current conguration (action) can be compared to the desired effect, and if necessary the behavior of the building shell can be adjusted actively. Closed-loop adaptive behavior can be controlled at two different levels: distributed, via embedded computation and local control rules in the CABS itself; driven by a central, supervisory control unit possibly linked to the building energy management system (BEMS) or enhanced by interaction with users . While the rst concept can be vital for successful operation of CABS, the second option is the most sophisticated and complex, but also offers the highest potential benets. In recent years, a myriad of advanced and intelligent control strategies have been developed for controlling buildings environmental systems (Dounis and Caraiscos, 2009). Examples are, multi-agent systems, predictive controllers, and reinforcement learning. Very likely, advances made in these research-areas can also be exploited for controlling adaptive behavior in CABS. One of the most promising concepts for CABS is model-based predictive control. Because of its ability to think ahead, intelligent strategies can be devised rather than only impulsively responding in real-time to occurring disturbances. Open-loop control Open-loop controlled CABS are characterized by the fact that the adaptive capacity is an intrinsic feature of the sub-systems comprising the building shell. The adaptive behavior is automatically triggered by a certain environmental stimulus which makes that CABS of this type are self-adjusting. This kind of autonomous control is sometimes also called direct control (Fox and Yeh, 1999) because environmental impacts are directly transformed into actions without external decision making component. One could also state that the elements in intrinsic CABS-concepts are at the same time both sensor and actuator. The adaptive behavior typically is reversible and repeatable and acts in a predictable fashion which makes the concepts reliable. An advantage compared to CABS with closed-loop control is that these types of systems can immediately inuence a buildings climate without expending any fuel or electricity for its operation. In addition, the subtlety of the technology is a main asset: the number of components is limited since no control units, processors or wires are necessary. The main drawback of CABS with open-loop control is the fact that it can only adapt to variations that were expected in the design stage. The adaptive features are typically designed by tuning material properties or other system variables to a certain range of conditions. As soon as real-occurring conditions
Manual control by occupants is often triggered to counteract discomfort and can be seen as a third entry to this list. It

also includes sensors, actuators, processors and feedback; not in the form of technical components but by exploiting human intelligence.

21

2. C LIMATE ADAPTIVE BUILDING SHELLS deviate from design conditions, the intended performance is not guaranteed anymore. A second disadvantage is related to the self-controllability of open-loop systems. After the building is commissioned, there is no possibility for (manual) intervention. This may pose problems for the way conicts and trade-offs are handled and can also cause difculties for user interaction and the cooperation with other environmental systems in the building.

2.3.4 Barriers and challenges

The principle technologies that facilitate adaptive behavior in CABS are available for some years now, and prospects for the near future are auspicious. However, as e.g. Hoffman and Henn (2008) pointed out, it will require more that just technological development before adaptive and sustainable building technologies will become standard practice. The inherent economical, social and psychological challenges and barriers that can hamper widespread application of CABS are the main subject of this section. Several of the hurdles that are listed here can be tackled by using building performance simulation (BPS). The application of BPS in relation to CABS is the main focus of Chapters 3 and 4. Conicts and trade-offs The idea of CABS can not be seen as the panacea that cures all struggles of the building envelope with regards to people, planet and prot. Like every building shell, CABS have to cope with the competitive or even conicting requirements imposed by changing user requirements and environmental conditions. But fortunately, CABS have the capacity to behave in response to these changes during their operation. The open-loop CABS-concepts demand that the strategy for dealing with trade-offs is already deliberately considered in the design phase. Whereas in the closed-loop type of CABS, the sensors and actuators can even effectively adapt their behavior in the operation phase and hence ensure that conicts are resolved in the best possible way. In the analysis of CABS, seven different types of conicts were distinguished. Each of them is highlighted below, and where appropriate illustrated with a typical example: comfort comfort Occupants comfort is a wide concept which has many facets that need to be considered simultaneously. Opening the slats of Allwetterdach07 presumably improves indoor air quality and reduces the temperature in the zone. But opening the roof is also accompanied with increased chances for draft problems, noise pollution, irritating smells and insects, privacy losses, or water damage. Before deciding upon opening, all these performance attributes should be carefully weighed. energy energy Buildings are surrounded by various energy elds. The impacts of adaptive actions on each of these energy eld can be different. Since the shades in Dr. Jokisch Building29 track the sun, generation of electricity is maximized. Because this adaptive action also blocks penetration of valuable solar gains, it is perfectly conceivable that the surplus in electricity is offset by the required additional energy for space heating and articial lighting. comfort energy The traditional trade-off in operation of building systems is the balance between comfort and energy demand. High comfort levels can be achieved, but typically this also requires high energy consumption. When the roof of The Sliding House93 opens on a sunny day, energy demand can decrease because of passive solar gains and utilization of daylight. On the contrary, this also introduces a reasonable chance for discomfort in terms of overheating and glare. among occupants Perception of comfort is inuenced by many factors including age, gender, clothing, mood, position in the zone and type of activity. As a consequence, it is quite unlikely that optimal comfort conditions for the average person (whatever that may be) results in satisfaction of all comfort preferences for all occupants. In the worst case, this individualism can lead to different sets of requirements that may be misaligned. Thats the reason why CABS control systems have to nd the right trade-off between different individual comfort preferences on the one hand, and other performance aspects associated with these conditions on the other hand. 22

2.3. CABS Analysis short term vs long term Successfully operating CABS guard the balance between short-term interests and long-term behavior. Yielding to momentary temptations can disturb well-devised energy balances, while temporarily and slightly sacricing performance may be easily surpassed by benets on the long run. An illustrative example in this respect is the roof pond of the Hammond house70 : Closing the roofs lids during the day might relieve thermal comfort conditions in intermediate seasons. However, no heat is being accumulated in that way, which eventually may lead to disproportional additional heating energy consumption at night. frequency of adaptation Although the exact relationships are not yet quantied, it is believed that a certain inverse relation exists between switching frequency and users acceptation. A building shell that keeps on oscillating between two states is bound to cause great nuisance. A good solution is to constrain the rate-of-change to a certain maximum, notwithstanding the fact that in some cases instantaneous action is absolutely necessary . costs of adaptation The largest share of CABS consume some energy when they are adjusted from one state to another. The amount of energy required to drive the actuating mechanism is sometimes larger than the gains associated with the adaptive action. In these situations, it is advisable not to change and keep the system in its current state. In order to avoid a decrease in performance, the aforementioned trade-offs and conicts need to be reconciled as good as possible, by making fair compromises in an intelligent way. The various performance aspects spread across multiple physical domains with each domain having their own performance indicators. The optimization of a single factor is unlikely to coincide with the desired response in other domains (Addington and Schodek, 2005). Furthermore, the trade-offs that have to be made are difcult to compare in terms of a single metric. Therefore, this requires multi-attribute decision making, which is not always straightforward and may even be counterintuitive. While pursuing best possible overall building performance, special attention should be paid for attaining well-balanced trade-offs. The chosen control strategy should avoid that a slight improvement in one performance aspect (e.g. thermal comfort) evokes disproportional losses for other performance aspects (e.g. energy consumption). As a general rule, all performance aspects should always be kept within certain acceptable limits or constraints. Building process The building industry has the reputation of being a conservative and inert sector. This conception gradually arose from the long experience with existing practices, which have proven to work out reasonably well. However the high amounts of energy consumed for providing comfort to the built environment are more or less taken for granted. Moreover, it seems that the peak of optimization has almost been reached when the building sector keeps persisting in the customary way (Knaack et al., 2008). The current building sector can be characterized with the following phrases: driven by standardized procedures focusing on selection of components and materials, long lifespans which demand high reliability, strong predilection to historical values, prescriptive design methods instead of aiming at performance, dominated by small to medium-sized companies with fragmented responsibilities, thinking in terms of cubic shapes with vertical walls and straight corners, and nally contracts with a partiality towards rst costs instead of operational costs (Becker, 2008; Douglass, 2008; Sakamoto et al., 2008; Woudhuysen and Abley, 2004). Summarizing these characteristics, it may be stated that none of them is in favor of introducing new technologies like CABS. To enable adoption, the building sector needs to change in a more or less revolutionary way. A thorough understanding of the associated performance benets of CABS seems inevitable for this purpose. Despite of this conservative approach towards innovation, new building designs every time start from scratch as a one-off process (Woudhuysen and Abley, 2004). A quite remarkable approach, especially when compared with other design-oriented practices like product manufacturing, automotive and aeronautical industry. Although it is considered as inescapable step towards really sustainable building 23

2. C LIMATE ADAPTIVE BUILDING SHELLS practices, iterative learning via post occupancy evaluations (POE) is not routinely applied (Meir et al., 2009). This leads to a situation where concepts can be overrated based on false assumptions, whereas successful designs can remain unnoticed. Consequently, the building sector hardly uses mechanisms by which new advances can be explored, tested and validated. Some good steps towards this direction are the growing interest in design competitions and the role of prototypes: These so-called what-if projects facilitate involvement, help to communicate the unknown, and may make people think about the ways in which architecture can evolve in the future (Von Stamm, 2008). It follows from the CABS overview that these experimental projects form a large contribution in the current state of CABS development. When talking about CABS, people almost automatically start thinking of new buildings. In addition to new buildings, CABS also offer a huge advantages in retrotting projects. However, the construction sector has to become more aware of this potential. Especially good perspectives exist for the adaptive technologies that are embedded in window frames. This makes them easily replaceable and relatively cost-competitive. Maintenance issues Repetitive motions are critical elements in CABS; actuators in kinetic CABS systems must be able to move many times, and often in the same way from hour to hour, day after day. Without proper servicing, chances for dysfunctional behavior or breakdown drastically increase. This not only ruins performance but also reinforces the risk of high reparation costs. Regular maintenance programs can help to mitigate these problems. For successful operation of CABS, these maintenance issues should already be considered during the early stages of the design process. This allows the designer to realize lowmaintenance systems that can easily be cleaned, accessed and replaced. A second important aspect related to maintenance is the awareness that sufcient budget and resources should be allocated for the life-cycle implications of maintenance issues. With continuous servicing, the expected lifetime of the building is likely to increase. This not only improves return on investment, but is also a signicant indicator for sustainability. In spite of these benets, the signicance of maintenance issues is not always acknowledged in design and operation of CABS. The following two examples demonstrate that problems with maintenance can even devastate CABS performance. Arab Institute du Monde09 in Paris, designed by Jean Nouvel is a classical example of CABS. With its photoreactive shutters, the kinetic faade provides an exceptional means of controlling solar gains and daylight. Not long after opening however, mechanical defects forced the kinetic apertures to cease moving one after the other (Figure 2.6). And the overhead expenses for xing these problems were not accounted for in the budget. At the moment, the shutters still cover the building with an impressive enclosure, but the intended dynamic features are now completely lost (Meagher et al., 2009). Another example of a CABS that did not turn out to become a success story because of maintenance issues is the Beadwall11 . The styrofoam beads apparently built up static electricity in their transit throughout the piping and window panes, and as a consequence, the beads tended to cling to the window. The only solution to this problem was to periodically rinse the beads with glycerin. Besides being time-intensive, the glycerine deposited a waxy, unsightly lm-coating to the window as an adverse side effect. That is the reason why maintenance issues eventually condemned this promising concept to oblivion (Chiras, 2001). Human aspects Providing better comfort conditions for occupants is one of CABS underlying principles. However, CABS also run the risk to bring about just the opposite effect when not operated properly. The interaction with occupants should be carefully considered in the design of every building, but this is particularly true for CABS. Since every individual is different, it is practically impossible to simultaneously satisfy comfort requirements of all occupants. It is essential that the occupants at least understand why 24

2.3. CABS Analysis

Figure 2.6: Mechanical failures at Arab Institute du Monde, Paris. (Left) broken driving rod, (right) disconnected linear actuator.

the faade is taking a specic action. But it would be even better if the faades control system offers the occupants some opportunity for adjustment or manual override. The inability to overrule the systems decisions is the most frequently encountered complaint with intelligent building envelopes (Aschehoug et al., 2005). Stevens (2001) even reports some cases of sabotage, where the occupants wilfully tried to beat the behavior of active faades. The best way to avoid counterproductive operation is by paying special attention to the design of user interfaces and feedback mechanisms. Although the importance of considering human factors in design and operation of active building technologies is now widely recognized (Cole and Brown, 2009), there is still a lack of knowledge in this interdisciplinary research area. This especially relates to the perception of comfort and human interaction with environmental controls. Some of the open research question include: (i) How do effects of group dynamics affect individual user behavior, and (ii) what are the transient effects of switching the building envelope; acceptable rates-of-change remain a matter of conjecture. Fortunately, research efforts are currently underway which attempt to gain better understanding of this matter. Added complexity Knaack et al. (2007) reports that increasing complexity is the emerging trend in faade construction. Adding technical solutions to the building envelope has become the indicator of state-of-the-art. The ght against climate change in the 21st century however is not about solving our problems by putting more and more machines into buildings (Roaf et al., 2005). Introducing adaptivity in the building shell undeniably increases complexity, especially when compared to e.g. an ordinary brick wall. CABS are typical examples of complex systems, which means that the interaction among components of the system, and the interaction between the system and its environment, are of such a nature that the system as a whole cannot be fully understood simply by analyzing its components. In essence, there is nothing wrong with complexity; in fact, truly smart architecture has to be complex (Senagala, 2005). However, we should avoid that things become complicated. In integrated CABS-concepts, several performance requirements have to be satised simultaneously and only with a single conguration, while the systems own state, its functional requirements and its
Following the denition that a complex system is a system composed of interconnected parts that as a whole exhibit one or more properties (behavior among the possible properties) not obvious from the properties of the individual parts.

25

2. C LIMATE ADAPTIVE BUILDING SHELLS environment are constantly changing. The fact that (i) unexpected events can take place, (ii) decisions have to be made in realtime, and (iii) future conditions are highly uncertain makes the process of control even more complex. This asks for deliberate control strategies which take all the interrelated aspects into account. When one succeeds, it leads to an elegant solution for both energy and comfort. This multi-functionality of integrating instead of layering can thus have lots of positive aspects, but also introduces serious disadvantages in case of malfunctioning this is illustrated with the following two examples: 1. If the dynamics in adaptive frittng03 fails, not only a particular layer (the shading device), but the whole glass faade has to be replaced; 2. If the pump in solaroof80 breaks down, than this does not only affect solar shading but is it also responsible for performance losses in other physical domains (thermal insulation). Increased complexity associated with CABS is sometimes also expressed in terms of buildability. This is directly related to the fact that CABS cannot fall back on many years of experience with proven technology. CABS-projects often involve innovative technologies, resulting in challenging projects with relatively high risks. The construction companies have often not yet mastered the cutting-edge technology or are completely unfamiliar with it. As a consequence, they are wary of taking risks because they foresee chances for high investment and/or failure costs. Indeed, some kind of trial-and-error seems to be inevitable. However, taking risks also offers opportunities, and as soon as the eld of CABS becomes more mature, these initial struggles are likely to fade away. A more persistent hurdle that needs to be overcome is overestimation of the unknown. Some designers tend to overlook difculties whether consciously or not while optimistically focusing on the advantageous aspects of technology. The potential drawbacks are often ignored in marketing discourses, where a more critical attitude would be more realistic. Some questions that may arise from the CABS overview are: what happens to a liquid faade37,57 when it starts freezing, or can Living Glass54 withstand the forces of a heavy storm? Another important topic in CABS design is the long-term behavior. New technologies can turn out to be very attractive in experimental settings, when they are only tested for a short period, with expected user behavior and under controlled conditions. The question remains whether the concept proves to be successful on the long run, when its exposed to the intricacies of the real world. This actual behavior can only be assessed after the project has actually been built. Unfortunately, in some cases, things do not always turn out as expected. An iconic example in this respect is the green wall55 of the Paradise Park Childrens Centre is Islington, London (Figure 2.7). Once, this pleasing envelope was an awardwinning design, but only three years after completion, its thriving glory is now completely perished (BBC, 2009).

Figure 2.7: Green wall of the Paradise Park Childrens Centre, just after completion (left), and three years later (right).

26

2.4. CABS Classication

2.4 CABS Classication

2.4.1 Introduction
The CABS-examples we encountered thus far, all originate from individual design efforts, and hence their characteristics seem rather isolated and fragmented. Following the arbitrary order of the overview in Appendix A makes it complicated to discern the common generalities and differences. This was the motive for developing an approach for classication of CABS in a more systematic way. Not with the goal to obtain statistical evidence, but the approach should (i) reveal general trends and patterns, (ii) trace current practice and methods, and (iii) identify lacunae and potential problems. This classication places the variety of CABS in context with each other and thereby discovers the factors that have been critical to the early applications of CABS. Its results may further also identify valuable directions for future building projects and developments in CABS research.

2.4.2 Methodology
The methodology for classication builds upon the results obtained in the CABS analysis described in Section 2.3. The classication is accomplished in three different ways. Classication A starts by simply observing where adaption takes place and in what stage the concept is. In classication B the focus is on the adaptive behavior and the way in which it is controlled. Finally, classication C is more physically based since it groups the CABS according to their relevant physical domains. Each classication step is described below in more detail. A - Application area of CABS Adaptive surface Adding adaptive features to the building shell can affect the buildings appearance in various degrees. Figure 2.8 graphically shows the different possibilities. Adaptation can take place in a) opaque wall elements, b) windows or other (semi) transparent wall elements, c) the roof, d) inside the surface of the faade as a whole, or e) the entire building in three dimensions.

Figure 2.8: Surfaces where adaptation takes place.

State of development Two global categories can be distinguished, CABS in the form of nal products (to be) used in architecture, and CABS that are still in the research, design & development (RD&D) phase. A schematic representation of this classication is given in Figure 2.9. The dashed arrow in this scheme indicates that concepts that are currently still in the RD&D phase may become more mature in time to eventually become part of real occupied buildings. In Appendix A, this subdivision of CABS is indicated with the respective corresponding color in the pictograms in the bottom-left corner. The rst category consists of buildings that are actually being occupied, or will be in the near future. This category can be further subdivided in two groups. The rst are CABS that are representative for or especially tailored to a specic building (red). In these cases, the architect, city and date of construction are also indicated. The group subsystems and components consists of CABS that have been used in real buildings, but can be applied universally rather than being attributed to one particular building (orange). Then there is a second category of CABS-concepts that did not yet nd their way to the market, but exist in the form of prototypes. Three groups were distinguished here, where (i) the working mechanism is being demonstrated on the full-scale (blue), (ii) the prototype is being tested in a 27

2. C LIMATE ADAPTIVE BUILDING SHELLS

Real buildings Subsystems and components Fullscale protoypes Research, design & development Reducedscale prototypes Virtual prototypes

Architecture

CABS

Figure 2.9: Classication of CABS in stages of development.

reduced-scale proof-of-concept phase (green), or (iii) the CABS is only an imaginary concept or design that did not yet left the drawing board (purple). B - Adaptive behavior and control Classication of adaptive behavior and control builds further upon the observations made in Section 2.3. The adaptive behavior can either take place on the micro-level (i.e. changes in material properties) or on the macro-level which means that the building shell changes its conguration via moving parts. A second categorization is the way in which the the adaptive actions are controlled. Some CABS directly transform an input into a certain action (open-loop control), while others exploit the possibilities of feedback (closed-loop control). In Appendix A, the right classication for each CABS is indicated in the top-right corner of the page. C - Relevant physics Building shells form the division between the ambient environment and indoor zones of a building. Consequently, the building shell is the interface where several physical interactions take place. Every climate adaptive building shell inuences this multi-physical behavior in its own typical way, by for example blocking, ltering, converting, collecting, storing or passing through the various energy elds. In order to characterize the differences and similarities in CABS, four domains are distinguished. The denitions of these domains are given in Table 2.2.
Table 2.2: Denitions of the physical domains.

Thermal Optical Air-ow Electrical

Adaptation causes changes in the energy balance of the building via conduction, convection, (long-wave) radiation and storage of thermal energy. The adaptive behavior inuences occupants visual perception via changes in the (semi) transparent surfaces of the building shell. A ow of air across the boundary of the faade is present. Or the adaptive behavior is inuenced by the direction and speed of the wind. All processes where some kind of energy is converted into electricity in the perimeter zone of the building. On the other hand, this domain is also used for CABS where electricity consumption is an essential part of its working principle

The impacts of adaptive actions in CABS do not end at the borders of a single domain, in fact most CABS inuence performance in more than one domain. The interdependencies are visualized via the four-ellipse Venn diagram in Figure 2.10. The relevant physical interactions for every CABS-example are indicated via one of the surfaces in this graph. The four different domains and all possible multiphysical overlaps together, results in a number of fteen different possible combinations. 28

2.4. CABS Classication

Figure 2.10: Physical domains.

Boundaries of the study A total number of 100 CABS-concepts was included in the analysis. The decision whether a building shell was included in the overview or not, was based on the denition of CABS in Section 2.1. As a corollary of this strict denition, a number of building shells with seemingly adaptive characteristics were not included in the analysis. Some of these building shell concepts are given in Table 2.3, together with their respective main reason for exclusion.
Table 2.3: Concepts not included in CABS analysis.
Concept Lotus effect coatings Intumiscent coatings Building Integrated PV Insulative shutters Retractable stadium roofs Media faades Wind motion faade Faade Vertigo Reference Leydecker et al. (2008) Bailey (2009) Prasad and Snow (2005) Bokel and van der Voorden (2005) Ishii (2000) Haeusler (2009) Datta et al. (2009) Cisar (2003) Reason for exclusion Does not affect buildings energy balance Not reversible nor repeatable Active but not adaptive Only manual, no automatic control Providing comfort is not major concern Does not effectively improve performance Based on climate, but not adaptive Layered and not always effective

2.4.3 Results
A - Application area of CABS The results of the classication according to adaptive surfaces and CABS state of development are shown in Table 2.4. This table reveals that the contribution of existing CABS-concepts in this analysis is approximately equal to the the number of CABS that are still in the RD&D phase. Moreover, CABS in the prototype-phase tend to focus more on a single surface of the building (either wall, window or roof). Whereas the CABS concepts that are already found in architecture show a trend towards affecting the faade or total appearance of the building as a whole. Another interesting conclusion is the fact that a signicant number of CABS use the roof of the building as the surface where adaptation takes place. The conception of the roof as fth faade is therefore legitimate.
Although they are not included in this study, these roofs offer lots of learning opportunities for CABS

In this context, the term aims at projection surfaces, display faades and window matrices
Acts here as an exponent for all kinds of multiple-skin faades

29

2. C LIMATE ADAPTIVE BUILDING SHELLS

Table 2.4: Adaptive surfaces and state of development.

Real buildings Subsystems and components Full-scale prototypes Reduced-scale prototypes Virtual prototypes

1 4 4 7 8 24

7 4 7 2 20

8 3

2 13

17 2 5 1 6 31

10 1 1 12

36 16 14 15 19

B - Adaptive behavior and control This section shows the categorization of adaptive mechanisms and the way in which they are controlled. When both groups are combined, this leads to four alternative possibilities. The number of CABS that ts in each of these categories in indicated in Table 2.5.
Table 2.5: Adaptive behavior and control.

Micro-level Closed-loop control Open-loop control 14 6 20

Macro-level 63 17 80 77 23

Architecture

Research, Design and Development

4 5

5 25

38

12

closed-loop | macro-level closed-loop | micro-level

open-loop | macro-level open-loop | micro-level

closed-loop | macro-level closed-loop | micro-level

open-loop | macro-level open-loop | micro-level

Figure 2.11: Adaptive behavior and control in current (left) and future CABS (right).

Both adaptive behavior and the way in which adaptation is controlled show a strong trend toward one of two possible alternatives. The minority is formed by open-loop control or micro-level CABS in the proportion of roughly one to four. When the adaptive behavior is split according to CABS state of development, an interesting trend can be discovered. In the existing CABS-concepts it can be observed that adaptations on the macro level, in combination with closed loop control is the prevailing method of adaptation. In the majority of examples, this is manifested via sensor networks that drive servomo30

2.4. CABS Classication tors as actuator. Future CABS which are now still under development show a tendency toward other adaptive mechanisms. As Figure 2.11 shows, there is a trend toward both more open-loop controlled CABS, and also CABS with adaptation on the micro-level. C - Relevant physics The dispersion of CABS across the physical domains is illustrated in Figure 2.12. These results show that highest number of CABS are classied in surface E, i.e. a combination of thermal and optical effects. Followed by thermo-optical supplemented with airow (surface H) and electrical (surface O).

Figure 2.12: Physical domains.

Furthermore, Figure 2.12 shows that the domain Thermal is relevant in all CABS-examples. This is a direct consequence of the fact that the buildings thermal environment is constantly changing. Optical effects are important in 79 instances. In principle, this are all cases where controlled daylighting is part of the climate adaptive strategy. Airow needs to be considered in one third of the CABS. And nally, the electrical domain, mainly in the form of photovoltaics, is encountered in 26 of the 100 adaptive building shells. Figure 2.13 shows the relevant physics for CABS, subdivided according to their state of development. In general, no large discrepancies are found between both charts. However, CABS in the RD&D phase more often incorporate the electrical domain. In practice, this is expressed via the increasing trend to include distributed energy generation systems in the building shell.

Figure 2.13: Physical domains for current (left) and future CABS (right).

31

CHAPTER
3.1 Introduction
Traditional design methods Mono-disciplinary Solution oriented Narrow scope Static Extreme conditions Building Performance Simulation Multi-disciplinary Problem oriented Wide(r) scope Dynamic All conditions 33

B UILDING P ERFORMANCE S IMULATION

Building performance is a product of many interrelated factors, including the buildings shape and orientation, building shell design, occupant behavior, weather conditions, building services and control actions. All these dynamic, non-linear and multi-physical interactions together make that predicting building performance might not be a straightforward task. Equipping buildings with climate adaptive building shells even amplies the level of complexity. Nevertheless, Chapter 2 introduces the persistent demand for methods that can quantify CABS performance aspects in the design stage. As Table 3.1 shows, traditional design tools (e.g. analytic calculations, empirical relationships, selection charts and nomograms) are not capable of satisfying this demand.
Table 3.1: Comparison between conventional design tools and simulation tools (Hensen and Lamberts, 2010).

Building performance simulation (BPS) on the other hand allows the designer to create a virtual model of a building. By applying different sets of boundary conditions and operational logic, it is possible to study the dynamic interactions that affect building performance. These analyses can provide answers to what-if questions and can also promote further insight in the cause-and-effect relationships of building performance. Because of these qualities, BPS holds the potential to also become a valuable tool for performance prediction in the design stage of CABS. Maver (1980) identied the four basic elements of computer-based building performance prediction as follows: (i) The designer generates a design hypothesis which is input into the computer (representation); (ii) the computer software calculates the behavior of the hypothesized design and outputs measures of cost and other performance indicators (measurements); (iii) the designer then examines

3. B UILDING P ERFORMANCE S IMULATION this prediction and exercises his value judgement (evaluation); nally, (iv) the designer may decide on appropriate changes to the design (modication). Three decades later, this four-step approach is still the predominant application of BPS in current practice. The potential of BPS is however much higher than only comparing the performance of proposed design alternatives. The goal of this chapter is to explore the possibilities of BPS in relation to CABS. Literature provides a brief number of examples were dynamic performance simulation studies have already been applied in the design and/or development stage of CABS, including: Active Building Envelope01 (Xu and Van Dessel, 2008), Beadwall11 (Engel, 1984), Breathing wall18 (Samuel et al., 2003), Switchable glazing32,65,92 (e.g. Lee and Tavil (2007); Jonsson and Roos (2010)), Heliotrop45 (Wigginton and Harris, 2002), Green roofs55 (Sailor, 2009), Roof ponds70,74 (Yannas et al., 2006), SolVent82 (Leal et al., 2004) and wall integrated PCM95 (e.g. Heim (2010); Kuznik et al. (2010)). This limited set of examples reveals that the application of BPS in CABS development is still marginal, but also gives reason for a more in-depth analysis of BPS potential. The remainder of this chapter continues with the critical elements that have to be considered in performance prediction of CABS and how BPS can play a role in this process. It starts with identifying the qualities of BPS in general related to CABS but also shows the deciencies and ways to work around these problems. Subsequently, the specic capabilities and limitations of the state-of-the-art BPS tools in relation to CABS are analyzed. This analysis covers (i) relevant physical domains and their integration and (ii) the available options to model adaptive behavior. Finally, the chapter discusses some alternative developments that may complement or enrich performance prediction of CABS, now and in the future.

3.2 Requirements for performance simulation of CABS

Chapter 2 identied four critical elements that constitute the adaptive behavior in CABS: climate, people, timescales and adaptive mechanisms. This section deals with the same requirements, now in relation to BPS, and extends the discussion with control of adaptive behavior. By doing this, it indicates the qualities and points out why BPS can become a valuable instrument in design and operation of CABS, but also shows some deciencies and points for improvement. Climate One of the prerequisites for successful modeling and simulation of CABS is using an appropriate set of weather data for the problem at hand. In most BPS applications, typical meteorological year (TMY) weather-les are used for supplying these boundary conditions. These les contain full-year time-series of mean hourly values for all relevant weather variables at a specic location. The original data is derived from long-term weather observations collected from nearby weather stations, typically at airports (Barnaby and Crawley, 2010). Traditional applications of BPS using TMY-les are typically employed for the purpose of comparative studies, and therefore the need for absolute accuracy is limited. In addition, the potential effects of false predictions are compensated by energy consumption in oversized HVAC systems. This contrasts with CABS, whose actual operation is closely related to real occurring weather conditions at a site. Because actual weather conditions are likely to deviate from the original design conditions, the risk exists that buildings with CABS are not able to achieve desired performance levels. The appropriateness of applying TMY-les for performance prediction of CABS is debatable because of the following reasons: Most buildings are not situated in direct vicinity to the respective meteorological station. However in some areas, a distance of say 100 km may introduce structural discrepancies in observed weather conditions (Cook, 1996). Moreover, longitude and latitude, and hence solar geometry are not the same, which particularly affects performance of sun tracking CABS-concepts. 34

3.2. Requirements for performance simulation of CABS CABS adaptive actions occur at various timescales. In order to accurately predict the performance of CABS there is a need for high-resolution weather data, where the length of one time-step is of the same order or smaller than the time constant of the relationships being studied. TMYles consist of averaged or integrated hourly approximations of real-weather conditions which can lead to inadequate results. Other implications of the different timescales in performance prediction of CABS are further discussed in the Section Timescales. Data for TMY-les is typically collected in open eld conditions (e.g. airports) whereas buildings usually are situated in (dense) urban contexts. Consequently, this urban environment creates a micro-climate around the building with implications for e.g. temperature, solar radiation and wind speed and direction that are not represented in TMY-les. Especially open-loop CABS directly adapt in response to the site-specic environmental conditions. This introduces errors in performance predictions due to discrepancies between boundary conditions in the model and the real world situation. One way to solve this problem is to perform on-site measurements that are directly supplied to the BPS tool. In cases where this detailed information is not available, it is possible to take advantage of model-based approaches. One possible technique uses simplied analytical models based on empirical relationships to create local weather data (Erell and Williamson, 2006), other approaches directly modify existing weather les, based on rstprinciple models (Bueno Unzeta et al., 2009) or detailed CFD-studies (Yi and Malkawi, 2008). The weather data constituting TMY-les is intended to be as average as possible. Real-world CABS however, also have to deliver adequate performance under non-typical conditions. Three solutions exist to work around this problem: (i) manually alter the values in the weather les to represent (more) extreme conditions, (ii) use design-day weather les to assess performance under the worst-case scenarios or (iii) run simulations with multiple-year weather les to increase the number of different situations. Although it might seem trivial, it is important that all essential weather variables are included in the analysis of CABS. However, not all standard weather les do include all the relevant aspects. Typical omissions are precipitation data (e.g. important in Phantom house64 , but also for the opportunity to use natural ventilation), effective sky temperature for modeling long-wave radiant heat exchange (e.g. important in roofponds70,74 and solaroof80 ) and snow cover (e.g. important for daylight reections) (Hensen, 1999). People In building energy simulations, the impacts of people are typically expressed via deterministic estimated occupancy schedules combined with average metabolic rates and typical use of equipment. This approach is a rather simplistic representation of reality, because actual occupancy and metabolism are subject to both uctuations and change (Mahdavi et al., 2008). Recent studies have demonstrated the validity of simple but effective Markov-chain based, stochastic models for occupant presence in ofces (Page et al., 2008) and synthetically derived probabilistic occupancy and electricity load proles in the domestic sector (Richardson et al., 2008; Widn and Wckelgrd, 2010). In addition, efforts have been going on to include the more sophisticated occupancy models into BPS tools (Bourgeois et al., 2006; Hoes et al., 2009). Presence of people also introduces the need for providing comfort. In current use of BPS, comfort considerations are mainly used for determination of heating and cooling systems capacity and energy demand. Combinations of setpoints and system layouts result in a set of state variables (e.g. air temperature, surface temperatures, luminance etc.) which are then transformed into aggregate thermal and visual comfort indicators (e.g. predicted percentage of dissatised (PPD) or useful daylight illuminance (UDI)) that have to stay within certain bandwidths. Although not yet routinely applied, it is well conceivable that these state variables and/or comfort indicators can actually be used to drive the adaptive behavior in building simulations. 35

3. B UILDING P ERFORMANCE S IMULATION In buildings with CABS, people may also manually inuence adaptive behavior in order to (i) achieve desired indoor environmental conditions, or (ii) counteract discomfort. Prediction of these actions requires deep fundamental knowledge of comfort perception and associated behavioral responses which can only be obtained via extensive eld surveys (Haldi and Robinson, 2008). More additional work is required for gaining better understanding of individual users adaptive actions and group dynamics in relation to CABS. But in this respect, the capabilities of current BPS tools will likely not be a restricting factor for modeling this behavior. Timescales The traditional application of BPS computes full-year datasets containing 8760 hourly values per state variable. This approach has turned out to provide sufcient accuracy for thermal building energy simulation and will probably also provide adequate results for CABS hourly and diurnal cycles. However, as Section 2.3.1 points out, adaptation in CABS can also take place at considerable shorter and longer timescales. Successful simulation of CABS has to acknowledge these time-cycles by taking them into account in the models. Performance simulations at short time-steps do in essence not pose new problems, as long as the appropriate sub-hourly boundary conditions (people and weather) are available. Most BPS tools tend to neglect the short-term dynamics of daylight, however the stochastic Skartveit-Olseth model implemented by Walkenhorst et al. (2000) provides a proper solution. In order to make fair comparisons of CABS that behave in response to seasonal variations, it is required to perform simulations on an annual basis. In case of integrated high-resolution simulation approaches (e.g. using CFD or ray-tracing), this may raise problems with regards to computation speed and can result in excessive run-times. Three solutions have been proposed to by-pass this problem: (i) sacricing level-of-detail (with or without loss of accuracy) for increasing computation speed (Reinhart and Walkenhorst, 2001; Zuo and Chen, 2010); (ii) using the full-edged simulation only for a certain number of extreme conditions, not for whole-year simulations; and (iii) via external coupling of simulators, i.e. only occasionally invoking the detailed simulator (e.g. every n t h time-step) during the simulation (Zhai et al., 2002). The last group of timescales encountered in the analysis of CABS are the long-term effects. The existing working methods in BPS can handle these implications fairly well. It typically requires simulation-runs of multiple scenarios with dissimilar boundary conditions. For assessing the impacts of future climate change on CABS, it is possible to take advantage of the work by e.g. Guan (2009) and Jentsch et al. (2008). Furthermore, the long-term effects that are subject to unexpected behavior are the place where uncertainty analysis can play an important role. Adaptive mechanisms In traditional BPS applications, the design of the building shell is represented by (i) size, geometry and orientation of the various faces and (ii) the associated material properties, via either selecting layers of materials from an existing database, or actually entering the appropriate thermophysical and optical properties. As opposed to the dynamics of boundary conditions, these material properties are hard-coded as constants that are xed throughout the simulation in the majority of BPS applications. Adaptivity in material properties has not been part of the initial fundamentals in the development of most BPS tools which makes that the available adaptive mechanisms for micro-level CABS are limited in off-the-shelf BPS tools. Additionally, the number of available actuating mechanisms that can aptly model macro-level CABS is also restricted. Two notable exceptions are formed by shading devices and window openings which in several tools can be modeled and controlled in a dynamic way. In general, the lack of capabilities to model these adaptive mechanisms may turn out to be the major obstacle for successful performance prediction of CABS. For this reason, Section 3.6 presents an in-depth review of the adaptive mechanisms available in state-of-the-art BPS tools. 36

3.3. Capabilities of BPS tools Control of adaptation As Section 2.3.3 already points out, there are two different concepts for controlling adaptive behavior in CABS; open-loop and closed-loop control. In practice however, this control can range from straightforward boolean logic to highly complex multiple-input multiple-output strategies possibly in combination with user interactions or weather forecasts. In terms of modeling this adaptive behavior in BPS, control algorithms of CABS must always based on the values and interpretation of certain quantiable conditions in the buildings internal or external environment. One of the common outputs of dynamic building simulations are time-series of all the buildings state-variables. Together with weather les and occupancy, this allows us to sense all possible relevant variables and accordingly decide upon the right action. The specic response to a certain environmental state is characteristic for each CABS, and here the type of control (i.e. open-loop or closed-loop) plays a leading role. BPS tools are typically developed with design applications in mind instead of the purpose of control. This imposes some restrictions on the possibilities to represent adaptive behavior in CABS during runtime. The situation where state-variables are substitute for actual sensor-output is for example a rather idealized representation of reality. Unfortunately, BPS tools are not (yet) equipped with features that model sensors in a more realistic way. In order to increase credibility, future BPS tools should also include e.g. delays, hysteresis, probability of detection, range, coverage and the actual position of the sensor.

3.3 Capabilities of BPS tools

Introduction The previous section presented the requirements for performance prediction of CABS in a general toolindependent way. In practice, a plethora of different BPS tools is being used for performance prediction in building design and operation with each tool having its own characteristics and possibilities. Users of BPS ought so select the tool that best suits their needs and manage the job with the capabilities that tool offers. The rational decision process may be inuenced by various other factors, including: previous experience with the tool, purchasing costs or availability within the organization, required domain knowledge, support facilities and the presence of a convenient user-interface. The U.S. department of energy (DOE) maintains a database with information about building software tools for evaluating energy efciency, renewable energy and sustainability in buildings (DOE, 2010). At the moment of writing this thesis, this Building Energy Software Tools Directory contains 382 entries. The nature of the tools listed in the directory varies widely. Usually, BPS tools are classied in terms of aspects, resolution and scope: Aspects By denition, all simulation tools make use of simplied models to represent reality. Most BPS tools are primarily oriented and specialized towards one single aspect of the buildings performance while neglecting the impacts in other physical domains. Examples include Radiance and DIALux for (day)lighting design; COMIS and CONTAM for airow simulation; and PVsyst and PV*SOL for photovoltaic systems. On the other end of the spectrum are the integrated wholebuilding performance simulation tools. These types of tools predict all aspects of buildings performance via a comprehensive integrated modeling approach. The physical impacts in buildings span across multiple physical domains, which often necessitates that the multi-physical behavior is analyzed simultaneously. The subject of how to deal with performance prediction of CABS in multiple physical domains is discussed in more detail in Section 3.5 Resolution The physical interactions in the built environment take place at various levels of detail, both in terms of space and time. In the current classication of BPS tools, usually ve different spatial levels are distinguished. From macro to micro, the tools are able to predict performance via (1) urban-level simulations, (2) whole building performance simulation, (3) componentbased simulations (e.g. fenestration and HVAC-systems), (4) simulation of local effects (e.g. thermal bridges) and (5) simulation of small-scale effects (e.g. heat, air and moisture transfer inside 37

3. B UILDING P ERFORMANCE S IMULATION building slabs). Buildings are thus subject to impacts at multiple spatial resolutions but these impacts also take place at various timescales. The BPS tools can assume (1) steady-state behavior, (ii) integrated (annual) evaluations or (iii) explicitly model dynamic behavior at successive timesteps. When choosing a simulation tool for a given task at hand, one should always check that the spatial and temporal resolution of the tool matches with the typical characteristics of the problem to be investigated. Scope The scope of a BPS tool is closely related to the objective of the simulation task (e.g. feasibility, design, analysis, visualization, optimization). In current practice, the majority of simulation tools is used in the buildings detailed design stage. These tools have facilities for code compliance checking, and are suitable for comparing alternative design variants or even for optimizing the building design. Several attempts have been made to increase the value of BPS by making it an active design tool that can be used in an interactive way. Specialized conceptual design tools have been developed that can deal with the abstract information available in the buildings initial design stages. In addition, some research projects are currently going on where BPS is used in the post-construction phase, e.g. for commissioning, fault detection & diagnosis (FDD) and modelbased control (Claridge, 2010; Henze and Neumann, 2010). Since the scope of the simulation can have such a large variance, it also inuences the tools capabilities to a large degree. Every BPS tool can thus be characterized by the place it takes in the three-dimensional space constrained by aspects, resolution and scope. All of these criteria however do not give explicit indications about a tools capability to predict the performance of adaptive behavior in CABS. The next section provides the methodology that was used to clarify these abilities. Methodology In a recently published report, Crawley et al. (2008) contrasted the capabilities of twenty different whole-building energy performance simulation tools. The publication contains comparative information about many aspects of BPS but does not disclose much about capabilities for simulation of CABS. Therefore, this chapter continues with analyzing the capabilities of the various BPS tools with respect to CABS. The information that was used in this analysis comes from various different sources. The process started with reviewing the user manuals, software-tutorials and help-facilities of commercial and academic BPS tools. These documents however typically deal with the basics of BPS and consequently do not provide many clues for simulation of CABS. More advanced uses of BPS are found in research settings. Results of these efforts are usually published in scientic journals and dissertations or presented at conferences like the International Building Performance Simulation Association (IBPSA) conferences. Furthermore, users of BPS tools form a worldwide community where ideas are shared and information is interchanged via mailing lists. The online archives of these mailing lists together form a wealth of information, including questions and answers to the most sophisticated modeling approaches and non-standard uses of the various tools. It would be impracticable and precipitate to assess the simulation possibilities for all CABS-concepts (> 100) in all available BPS tools (> 380). Appendix A shows that the CABS-concepts can be classied according to their relevant physical domain(s). The application areas of the BPS tools tends to follow the same logic, as indicated in Figure 3.1. Section 3.4 presents the classication of BPS tools, and in this way simplies the analysis because in general, tools in each physical domain share the same qualities and characteristics. Appendix A further shows that effects of CABS typically span across more than one physical domain. Therefore Section 3.5 pays attention to the strategies for performance evaluation of this multi-physical behavior in an integrated way. Subsequently, the analysis of adaptive features in the various tools is summarized in Section 3.6.

3.4 Physical domains

The work by Crawley et al. (2008) focuses on the capabilities of BPS tools for predicting energy performance of buildings, a process sometimes also called building energy simulation (BES). In this thesis, 38

3.4. Physical domains

Figure 3.1: Overview of BPS tools and their physical domains.

the scope is wider as it also includes simulation tools for performance prediction of daylighting/visual comfort, indoor air quality and issues related to electricity. Here, the tools are classied following the same logic as the classication of CABS in Table 2.2. However, during this analysis it was discovered that a signicant number of tools specically deals with the interplay between the opaque and transparent elements of the building shell. For this reason, the analysis in this section is extended with a fth category Thermal and Optical which is equivalent to surface E in Figure 3.1. A - Thermal Prediction of thermal effects in buildings is what has usually been regarded as the typical application of BPS. Some authors prefer to use BES, but here the word Thermal is retained in order to avoid confusion with other energy related aspects of CABS. These tools simultaneously evaluate the transient energy owpaths of conduction, convection and radiation across the building envelope and also take casual gains and HVAC systems into account. Thermal BPS tools do include the effects of visible shortwave solar radiation, but often do not consider the effects of visual perception. The same is true for inltration and ventilation; the air volumes associated with these processes and the impacts it has on the buildings energy balance are considered. But because buildings are typically divided into zones that are assumed to be well-mixed, it is not possible to assess the impacts of airow inside the zones. Typical outputs of thermal BPS tools are e.g. air temperature, surface temperatures, thermal comfort indicators, and heating and cooling energy demand. These results are used to prove compliance with comfort standards and energy regulations and for comparison of proposed design alternatives. Some of the most popular tools are: ApacheSim/IES-ve, DesignBuilder, DOE-2, EnergyPlus, eQUEST, ESP-r, IDA-ICE, TAS, TRACE, TRNSYS type 56 and VABI 114. But as will be demonstrated in the following subsections, the capabilities of some of these tools extend further than the thermal domain only. B - Optical For performance prediction of CABS with transparent parts, it is important that the collection, admission and distribution of daylight and the interplay with articial lighting are considered simultaneously. Too low daylight levels result in (i) gloomy spaces, (ii) high electricity demand and internal heat production via articial lighting and (iii) limited views to the outside. Too high levels on the other hand can result in comfort problems associated with glare and increased cooling loads caused by excessive solar heat gains. Both intensity and spatial distribution of daylight is extremely dynamic throughout the year, which has led to the conception that well-daylit spaces can only be designed by using computational simulation (Rogers, 2007). In addition, daylight simulation promotes further insight in operational per39

3. B UILDING P ERFORMANCE S IMULATION formance conicts, and may help to devise intelligent control strategies for adaptive behavior that can help to achieve balanced trade-offs. Daylight simulation engines rst combine a sky model with a scene (i.e three-dimensional geometric model with optical material descriptions for all surfaces) and then calculate illuminances and/or luminances proles at sensor-points within the scene (Reinhart, 2010). The development of simulation engines has diverged into two directions: Ray-tracing and Radiosity. Table 3.2 summarizes the characteristics and qualities of both approaches.
Table 3.2: Comparison between classical ray-tracing and classical radiosity methods (Geebelen, 2003).

Ray-tracing View-dependent Handles specular behavior best Handles any geometry Can handle transparency Does not compute overall light distribution Difculties with indirect lighting

Radiosity View-independent Handles diffuse behavior best Performs best with facetted shapes Performs best with opaque surface Does compute overall light distribution Indirect lighting is treated correctly

Daylight simulations have traditionally been used for performance prediction under static conditions an approach that only gives information about daylight at a single point in time. The most widely used tool is Radiance (Reinhart and Fitz, 2006), not only for performance studies, but also for the purpose of visualization. In recent years, there is a shift towards dynamic daylight simulations, based on the Daylight Coefcient (DC) method which is a favorable development for aiding design of CABS . This approach results in annual illuminance proles that can be used to predict daylight contributions and electric lighting energy use, and can be combined with a thermal simulation program to predict overall energy use for heating, lighting and cooling (Reinhart, 2010). When used in this way, dynamic daylight simulations offer good perspectives for performance evaluation of CABS. Several climate-based daylighting metrics that satisfy both visual comfort and energy demand are currently being tested. These performance indicators consider the quantity and character of daily and seasonal variations of daylight for a given building site. However, there is still a need for establishing the proper benchmarks (Mardaljevic et al., 2009). C - Airow Airow in buildings serves two purposes: it helps in maintaining acceptable indoor air quality and also inuences the energy balance to effect heating and cooling. For typical engineering approximations and initial design studies, it is often sufcient to use scheduled bulk ows. Some applications of CABS however demand for a more accurate prediction of airow that accounts for ow-effects across the envelope driven by temperature differences and wind characteristics. Several modeling approaches exist for this purpose, at various complexities and levels of detail (Chen, 2009). The most rigorous approach is to use computational uid dynamics (CFD) to numerically solve the governing Navier-Stokes equations. This results in detailed information about pressure, contaminant concentration, airspeed and temperature at every point in space. CFD-software like Flovent, Fluent/Airpak and Phoenics has been extensively used in many building applications such as (natural) ventilation, thermal comfort, indoor air quality, and re and smoke safety studies (Lu et al., 2009). Application of CFD however introduces some drawbacks. Since no generally accepted modeling guidelines are available users have to take care of cost- and time-intensive validation and verication studies especially with regards to turbulence modeling, wall-functions and grid sensitivity. In addition, signicant domain expertise is required in order to manage uncertainties in the simulation process. Finally, CFD is computationally very demanding which may prevent dynamic performance prediction of airows in buildings that is required to predict performance of CABS. 40

3.4. Physical domains A second approach that offers less detailed information but that is potentially more valuable for CABSdesign are airow networks (AFN). This method calculates airow and contaminant transport between the zones of a building and between the building and outside air on the basis of pressure differences. Since AFN is based on some simplications (e.g. neglecting momentum effects and assuming uniform air temperature in a zone) the method is fast enough to be used for whole year ventilation performance prediction. AFN tools may be either stand-alone applications (e.g. COMIS and CONTAM) or coupled to thermal solvers as is the case in EnergyPlus, ESP-r, IES-ve and TRNSYS. Unfortunately, AFN does not calculate air velocity inside zones which is an important parameter for evaluating local thermal comfort (van Treeck, 2010). A third approach makes use of semi-empirical relations and results of analytical models. The method establishes case-specic bulk ventilation airow rates through different types of openings as a function of temperature and wind conditions. The methods can be used in a stand-alone manner, but are also valuable for specifying boundary conditions in thermal BPS tools, e.g. via the use of equations-type in TRNSYS. Although this method incorporates many assumptions, it is an improvement with respect to scheduled bulk-airow, and is also effective for modeling inltration, and allows for easy integration of adaptive features.

D - Electric Performance simulation of electrical systems in the built environment requires consideration of generation, distribution, storage and consumption of electricity. Distributed electricity generation via solar cells is encountered numerous times in the CABS overview. Various software packages exist for planning and development of such photovoltaic systems. These tools can be used for e.g. pre-feasibility assessment, system sizing, simulation of operation and economic assessment studies. The capabilities and typical application areas of the various available tools are summarized by Turcotte et al. (2001) and ri (2005). Unfortunately, these tools are only used for stand-alone congurations, and do not capture the mutual interactions with the buildings energy balance. Most of the BPS tools do also claim to support facilities for simulation of building-related electrical systems. Yet, in most cases this capability is only restricted to scheduled building power loads (Crawley et al., 2008). However, some tools offer the possibility to perform predictions of building integrated photovoltaic (BIPV) system performance at various levels of detail. Successful integration of embedded energy systems in the built environment requires careful matching of both supply and demand. Integrated power ow modeling in BPS can capture the transient electrical effects of decentralized energy systems like BIPV, (ducted) wind turbines, micro-cogeneration, fuel cells, batteries, articial lighting, etc. In this respect, the models in TRNSYS and ESP-r are the best equipped to serve this purpose. The libraries in TRNSYS offer various possibilities to simulate and couple electrical components to the building model. In the case of ESP-r, the power ow modeling facilities are fully integrated with the building model and other subsystems (Kelly, 1998).

E - Thermal and optical Most BES tools emanate from attempts to help engineers to realistically size HVAC equipment in buildings. This has led to some simplications which in turn make that these tools experience difculties to assist designers in the actual building design process. In response to this, several tools have been developed in recent years which are specically attuned to allow for more insights in the intricacies of the design of building shells. Such tools are able to simultaneously address thermal and visual comfort requirements together with the associated energy demand for heating, cooling and lighting. These tools also mark the shift towards increased use of BPS for assistance in making informed design decisions in earlier stages of the design process. In this section, three of such BPS tools are discussed. Each of them claims to support facilities for simulation of advanced, intelligent or climate-responsive building shell concepts. Here their applicability with respect to climate adaptive building shells is highlighted. 41

3. B UILDING P ERFORMANCE S IMULATION The MIT Design Advisor (DesignAdvisor, 2008) is a fast and simple design-tool that is especially intended for improving energy performance in the rst hours of the building design process. This webbased tool is specically developed for architects and building designers, and allows comparing design alternatives side-by-side (Urban and Glicksman, 2006). MIT Design Advisor was introduced by Lehar and Glicksman (2003) as a simulation tool for the optimization of advanced faades". The webinterface enables users to choose from eight different pre-dened faade typologies. Among these are two types of double-skin faades with internal airows and movable shading devices. It is however not possible to inuence control of adaptive behavior or dynamic properties of these building shells which makes the tool inappropriate as a design aid for CABS. NewFacades is a simulation tool that helps pass from ideas to signicant concepts in the design of intelligent faades (Ochoa and Capuleto, 2009). It basically is a shell around EnergyPlus that automatically generates and compares the performance of design alternatives. In this way, the tool explicitly models the thermal and visual comfort levels and energy demand of various building shell concepts on the basis of generic, high-level input data. The intelligent features that are considered are however restricted to lightshelfs, traditional shading devices and controlled night ventilation. The main objective of DAYLIGHT 1-2-3 is to help design professionals with interest in but no expert knowledge of daylighting to develop climate-responsive daylighting design concepts. Integrated thermal and visual performance predictions are performed via a customized version of ESP-r linked to Radiance-based Dynamic Daylight Simulations (DDS) and the sub-hourly occupancy-based control model (Reinhart et al., 2007). The approach that DAYLIGHT 1-2-3 uses seems to hold great promise for CABS as it satises most of the requirements for performance simulation of CABS presented in Section 3.2. The tool relates an intuitive user interface via quick but reliable simulation runs to a comprehensive set of output result. Currently, the only way to introduce adaptivity to the building shell model is via predened sets of window-blind combinations that are controlled automatically or via latest generation occupancy control models (SHOCC) (Bourgeois et al., 2006). Nevertheless, it would be recommendable to investigate the possibilities of expanding the tool with more adaptive mechanisms.

3.5 Coupling of physical domains

In order to assess building performance as a whole, impacts form the various physical domains presented in the previous section have to be interpreted simultaneously. This issue is no novelty that comes with introduction of CABS, but needs to be considered in in every building design. Several methods for multiple domain performance simulation have been discussed elsewhere (Citherlet, 2001; Hensen et al., 2004; Malkawi, 2004; Wang and Beausoleil-Morrison, 2009). This section summarizes these concepts and places them in the context of CABS. Combination of stand-alone tools The relevant physics in CABS happens to take place in four different domains. Since each domain demands for specic algorithms to predict performance, this has led to the development of monolithic domain-specic simulation tools. With a combination of stand-alone tools, CABS full performance can be assessed in parallel, with the reservation that each domain is treated in an isolated way. This approach may be most pragmatic in some cases but also introduces several drawbacks. Because there is no coupling between the domains, the interrelated aspects are not captured which prevents a simultaneous, integrated performance assessment. Since there is no communication between tools, the isolated tools are not aware of each others adaptive behavior. An illustrative example showing this shortcoming is a stand-alone daylight simulation which has no means for changing degree of transparency in response to e.g. air temperature. The stand-alone approach will further also give rise to data redundancy and to potential inconsistencies between the different building models that are developed separately. This last problem can be obviated by using data model interoperation in which multiple tools either (i) share or (ii) exchange (parts of) the building model. Examples of the rst type are the Integrated Data Model (ModelIT) in IES-ve and the 42

3.6. Adaptive features Uniform Environment in VABI. The second type works with neutral le formats like IFC or XML. In some applications of CABS there is no reason for objection of this simple combination of separated tools, but in other cases integrated performance assessments are inevitable. Chapter 4 presents a methodology for selecting the appropriate modeling and simulation strategy. Integrated performance assessment "The aim of an integrated approach is to preserve the integrity of the entire building system by simultaneously processing all energy transport paths to a level of detail commensurate with the objectives of the problem to hand and the uncertainties inherent in the describing data" (Clarke, 2001). Integrated simulations allow for a holistic performance assessment via the use of a single simulation tool with only one user interface and one building model. This conation of domains also makes sure that each domain is informed about the state of adaptive behavior in response to other domains stimuli an approach that is in favor of reliable performance prediction of CABS. Three frequently used integrated BPS tools are EnergyPlus, ESP-r and TRNSYS. As we will see in the next section, each of them offers different possibilities for modeling and simulation of CABS. A disadvantage is that the user is restricted to the (adaptive) features the particular tool offers. In addition some subdomains are still missing or inadequately represented in the tools (Clarke, 2009). Co-simulation No single BPS tool appears to be able to model all relevant physical phenomena that determine performance of CABS. In addition, the available adaptive mechanisms are limited and dispersed across the various tools. Consequently, users of BPS will often be confronted with the problem that there is no tool available that satises their simulation requirement. Co-simulation enables coupling of BPS tools with complementary capabilities and may thus be the solution to this problem. In this way, added value can be generated by exploiting adaptive features in one tool together with the reuse of powerful modeling capabilities of another tool. Co-simulation further allows for generalization of fragmented development and accelerated infusion of innovations: once new adaptive mechanisms are implemented in one tool, they can be made available in other BPS tools as well. Depending on the type of problem, the workow in co-simulation can be either sequential or bidirectional (Tr ka, 2008). In the rst approach, the different simulators are invoked successively, while c in the latter approach simulators run in parallel and data ows between simulators during run-time. The latter approach requires setting-up the connections and communication protocols between the tools; a task that might be too arduous for performance prediction of just a single building design. An interesting development that might offer opportunities for design and operation of CABS is the Building Controls Virtual Test Bed (BCVTB) (Figure 3.2). This modular, extensible, open-source software platform allows designers, engineers and researchers of building energy and control systems to interface different simulation programs with each other and, in the future, with Building Automation Systems (BAS) (Wetter and Haves, 2008). Instead of linking two programs to each other, this co-simulation approach uses a middleware-tool that manages communication, data exchange and integrated output facilities for a number of different simulators. One of the envisioned application areas of BCVTB is the investigation and implementation of intelligent local-loop and supervisory control strategies for active faades.

3.6 Adaptive features

Most state-of-the-art BPS tools have a history of gradual development and extension of capabilities. At the time these tools were created, software-developers had no incentive to implement adaptive behavior into building shell models. This inheritance is now the reason that the state-of-the-art in BPS lacks 43

3. B UILDING P ERFORMANCE S IMULATION

building energy EnergyPlus HVAC & controls controls Modelica Simulink

building energy TRNSYS

airflow Fluent

controls & data analysis MATLAB

lighting Radiance

BCVTB

wireless networks Ptolemy II

building energy ESP-r

hardware in the loop

real-time data building automation www+xml BACnet

implemented funded in proposal in discussion

Figure 3.2: Building Controls Virtual Test Bed (Wetter, 2009b).

abundant capabilities for performance prediction of CABS. Moreover, attempts to introduce adaptivity in these tools are confronted with severe restrictions pertaining to the basics of energy modeling in buildings: Because of their computational efciency and accuracy, Response Factors and Conduction Transfer Functions are the most widely used methods for load calculations and building energy analysis (Crawley et al., 2008; Spitler, 2010). The coefcients in these equations are constants that are calculated only once, prior to the actual simulation. As a consequence, thermophysical properties of opaque constructions are coerced to be xed, preventing the simulation of adaptive behavior (Pedersen, 2007). An alternative approach that allows for more exibility uses numerical methods for calculation of transient heat transfer in buildings. This approach also has difculties with simulation of adaptive behavior since the way the equations are solved makes that attempts to reassign state variables values at run-time is incongruous (Clarke, 2001). The state variables are the output of the numerical solution processes, and therefore reassigning particular values will violate energy balances. Instead, it is necessary to adjust the source terms of equations in question to effect the required adaptive behavior. Nakhi (1995) argues however that introduction of discontinuity in material properties is not straightforward because it may affect the accuracy and numerical stability of the overall solution. A further complexity is that building energy analyses have to respect the time-history effects of thermal storage. Simply hopping from one state to another in separate building models without state-rewinding is therefore unacceptable . Instead, adaptive behavior in BPS has to be considered on-the-y, which introduces high burden on the tools control algorithms. Despite of these barriers, several possibilities to simulate CABS have found their way in the state-of-theart BPS tools. In this study, two different types of adaptive features are distinguished, prompted by the nature of the BPS tools. The rst group consists of tools that only offer selection of predened functions and in this way adaptivity always serves a specic application. The second group allows for more liberty in the simulation process and can therefore be used to model general purpose adaptive behavior. The results of the analysis for both categories is described in the following subsections.

3.6.1 Application-oriented
Most BPS tools under study are closed environments where building models are created by selecting predened functions via e.g. clicking buttons, ticking boxes and lling in mandatory elds. In addition,
In CABS with distinct winter-summer behavior it can be a feasible approach to perform semiannual simulations with two building models.

44

3.6. Adaptive features the majority of these off-the-shelf BPS tools are non-transparent and the source-code is often inaccessible and unmodiable. This makes it difcult to manually tweak or extend the tools capabilities which is also the reason that the tools lack active developer communities. As a consequence, users of these tools have to manage their task merely with the capabilities the tool is offering. Unfortunately, these commercial tools are attuned to straightforward building designs. Making prot is the core activity of the software developers, and capabilities are usually only extended in response to their clients wishes. During the course of the research, the list of available adaptive features in these tools expanded with the launch of new software-versions. This might give an indication that there is an increasing demand for performance prediction of CABS. The analysis resulted in ve different application-oriented adaptive features; each of them is discussed below. Electrochromic glazing is a mature product for some years now, and many research papers have been written about its application in buildings and architecture. As a consequence, switchable glazing technologies are embedded in several simulation tools, for example BuildingDesignAdvisor, DesignBuilder, EnergyPlus, ESP-r and eQUEST offer the possibility to switch the window state during runtime. The differences between the various models are the number of possible window states (e.g. on/off or gradual transitions) and the state variables that can be used for control of adaptation. Simulation of thermochromic windows is slightly more complicated because of its open-loop character; adaptation is directly triggered by window temperature rather than via a decision making component based on more general simulation variables. Nevertheless, a thermochromic window capability was found to be available in EnergyPlus. The input of this model consists of sets of window properties at various temperatures. During the simulation, the thermochromic layer temperature of the previous time-step is passed on to the control algorithm which then selects the window properties that best matches with the given temperature. IES-ve (ApacheSim) offers the possibility to give dynamic shading devices additional thermal resistance properties. This makes it possible to simulate the performance of insulating solar shading systems, including Beadwall. Unfortunately, the thermal models in IES-ve do not explicitly account for daylight levels, which prevents insights in the trade-offs between thermal and visual comfort. Phase change materials models are present in EnergyPlus, ESP-r and TRNSYS. These models inuence heat transfer in constructions via either the effective heat capacity or the additional heat source/enthalpy method. Implementation of PCM forced developers of EnergyPlus to abandon the Conduction Transfer Functions approach and introduce a numerical nite difference algorithm. This new algorithm also has a provision for including a temperature coefcient that makes thermal conductivity variable during the simulation. Both EnergyPlus and its user-friendly interface DesignBuilder offer the possibility to simulate green roofs. The models account for: (i) long wave and short wave radiative exchange within the plant canopy, (ii) plant canopy effects on convective heat transfer, (ii) evapotranspiration from the soil and plants, and (iv) heat conduction and storage in the soil layer. Implementation of moisture dependent thermal properties is planned for the next version of EnergyPlus.

3.6.2 General purpose

The application-oriented adaptive features presented in the previous subsection only allow for restricted exibility and by denition lag behind the latest state of adaptive building shell technologies. Fortunately, some BPS tools offer the freedom to simulate more general purpose adaptive features, which also gives users the ability to respond to emerging technologies. Using these features demands for advanced use of BPS and also asks for creativity on the side of the user of the tool. In this light it is also possible to make use of workarounds features of a BPS tool that can aptly simulate certain adaptive behavior, but that initially were not intended for that specic purpose. General purpose features are expected to have highest potential in the RD&D stage of CABS, it is however questionable to what extent these modeling approaches will turn out to be valuable in more mainstream design processes. 45

3. B UILDING P ERFORMANCE S IMULATION Airow network (AFN) tools are based on wind pressure coefcients (c p ) and in this way are inherently adaptive to prevailing wind conditions in a passive way. All AFN tools under study also have ow coefcients and/or opening areas that are controllable during runtime. This introduces the possibility to actively bring about the desired adaptive behavior for airows. Apart from retractable Venetian blinds, none of the examined daylight simulation tools offers the capability to change behavior dynamically. However, daylight simulations do not suffer from history effects. Therefore, adaptive behavior can be accomplished articially after computation of raw luminance and illuminance data for all possible states of the building shell. This data can subsequently be coupled to a control algorithm that selects right state and corresponding solar gains. Matrices with this data can furthermore be transformed into the desired dynamic daylight performance metrics. Because of their particular background and open make-up, the BPS tools TRNSYS and ESP-r both offer a number of attractive capabilities for modeling and simulation of adaptive behavior. A comprehensive overview of both tools adaptive features is the subject of the following subsections. TRNSYS The approach that TRNSYS takes towards managing complexity in the built environment is typied by breaking the problems down into a series of smaller components. One of these components is a multizone building model in TRNSYS called TYPE 56 that can be connected to a vast number of other components, including: weather data, HVAC systems, occupancy schedules, controllers, output functions, storage tanks, renewable (solar) energy systems, etc. This typical conguration allows the user to set-up and manipulate the connection rules and order between the building and all other subsystems/components in the simulation environment. TRNSYS TYPE 56 offers the possibility to change the thermal and optical window properties during runtime with a function called variable window ID. Additionally, it is also possible to control the ratio of window/frame area which inuences the degree of transparent faade elements. All the other adaptive mechanisms in TRNSYS can be found in the connections with other components in the Simulation Studio. Using equations enables the application of boolean logic and algebraic manipulations to any variable in the simulation. This ow of information can then be used to drive a control algorithm that is able to dynamically switch on, switch off or modulate e.g. overhangs and wingwalls (TYPE 34), shading masks (TYPE 64), attached sunspaces (with or without movable thermal insulation) (TYPE 37), windows with variable insulation properties (TYPE 35) and photovoltaic cells (TYPES 94, 180 and 194). In addition, it is also allowed to adjust the connections with weather les and radiation processors. In this way it is possible to mimic the effects of changing orientations or to reduce the impacts of solar gains on the construction as a way of modeling external shading devices/objects. The standard TRNSYS distribution already comes with an extensive number of components. Yet, one of the distinct benets of TRNSYS modular nature is the fact that it allows users to add content by introducing new components (McDowell et al., 2004). With some coding efforts it is possible to encapsulate the desired CABS behavior in a new TRNSYS TYPE which can then be linked to the building model. Unfortunately, coupling these new TYPES with the building model works in a rather indirect way via the slab-on-grade approach. In TRNSYS it is not possible to replace building shell constructions, instead developers must impose the desired behavior via manipulation of inside surface layer temperatures of boundary zones and the respective heat transfer coefcients. Kuznik et al. (2010) recently demonstrated this approach for CABS via the implementation a new PCM wallboard TYPE (Figure 3.3). Another attractive functionality of TRNSYS is the opportunity to call external programs at every iteration step during the simulation. This approach extends the capabilities of TRNSYS with airow calculations of COMIS, CONTAM, and Fluent; spreadsheet functions and Visual Basic macros of Microsoft Excel, M-les and S-functions of Matlab/Simulink and the algebraic equation solving methods of EES. Capa With the launch of TRNSYS 17 in early 2010, it is now also possible to embed TRNSYS TYPES as .mdl models in Simulink (Riederer et al., 2009). With this approach, it is possible to exploit Matlabs full functionality together with the distinct qualities of

46

3.6. Adaptive features

Figure 3.3: Coupling of new PCM model (TYPE 260) with TRNSYS multizone building model (TYPE 56) (Kuznik et al., 2010).

bilities can be further expanded by reading in (measured) data from generic external data les. This can serve various purposes, including the import of high-resolution casual gains and illuminance data. ESP-r ESP-r is an integrated research-oriented BPS tool with an active developers community and a source code that is accessible and modiable. The tool is used for prediction of thermal, visual, indoor air, electrical and acoustic performance of buildings. Users can extend these features in the source code as desired and in this way introduce adaptive behavior. In the course of years, several functionalities that can be used to model adaptive behavior in the building shell have already been implemented. The use of these capabilities however remained limited, possibly because the features are (i) not well-documented or (ii) concealed somewhere in the distributed menu-structure of ESP-r. This section summarizes six of such adaptive features. One of the control laws in ESP-r is called thermophysical property substitution mode. Instead of being used for controlling operation of the HVAC-system, this control strategy can replace the thermophysical properties (, c p , ) of a construction during the course of the simulation. In essence, this control works like any other control algorithm in ESP-r, in the way that actions are triggered based on tests applied to sensed variables during run-time. Unfortunately, this feature does not allow for full exibility since it only affects opaque wall elements and the only sensor variable is indoor air temperature. The previous feature dealt with opaque construction elements only, however ESP-r also has a similar functionality available for modeling dynamic behavior of windows; transparent multi-layer construction control. This functionality can for example be used for performance prediction of switchable glazing technologies. Currently it is possible to replace window properties (.tmc-les) based on time, temperature, solar radiation level or lux level. Restrictions are that no more than two states are supported without the possibility for gradual transitions. Recently, the capabilities of ESP-r have been further extended with the implementation of a facility for modeling and controlling complex fenestration constructions (CFC) (Lomanowski and Wright, 2009). In ESP-r the special materials facility was introduced to model active building elements (Evans and Kelly, 1996). This universal functionality may be applied to any node within a multi-layer construction.
TRNSYS types.

47

3. B UILDING P ERFORMANCE S IMULATION The special material subroutines can actively modify the matrix coefcients of these specic nodes at every time-step. By doing this, it directly changes basic thermophysical or optical properties and/or the associated energy ows based on the respective physical relationships. Currently, the following special materials are implemented: Building integrated photovoltaics , ducted wind turbines, solar thermal collectors, thermochromic glazing, evaporating surfaces and phase change materials. It is possible to add new user-dened special materials, however this may require laborious programming work. A further capability of ESP-r is to impose measured data on a time-step basis. This information can indeed be obtained via experiments, to realistically represent actual behavior. But this functionality also allows to manually bring about certain predetermined transient behavior, and in this way model some of CABS features. Data les of this type can be associated to various attributes of the simulation, including: elements of climate data, casual gains, inltration, air velocities, setpoints and control states. ESP-r offers the peculiar possibility to use roaming les. This facility is used to change the location of a building as a function of time, and was intended to be used for cruise ships. Because this roaming le not only includes coordinates but also orientation of the zone, it is very well t to simulate rotating buildings. Nakhi (1995) introduced variable thermophysical properties in ESP-r to more realistically model heat transfer in building slabs under a wide range of operating conditions. It takes into account the temperature and/or moisture content dependency of a materials thermophysical properties. This is implemented via transient thermophysical material properties (, c p , ) that are linear or polynomial functions of layer temperature or moisture content. In this way it is possible to model the continuous or discontinuous properties of smart materials and other micro-level CABS.

3.7 Discussion
Together with mere advancements in BPS, alternative developments are going on that may enrich or complement the performance assessment capabilities for CABS. This section rst discusses the need to assess the more generic environmental performance of buildings and how this applies to CABS. Then, ways to deal more explicitly with the risks and uncertainties associated with CABS performance predictions are presented. Finally, this section provides some perspectives about BPS future directions and what this implies for CABS. Environmental impact Driven by an increasingly stronger demand for sustainable development in the built environment there is a growing interest for methods able to evaluate a buildings overall environmental impact. This has led to many country-specic environmental assessment tools (including e.g. LEED and BREAAM) whose outcomes are typically expressed in mandatory performance thresholds or voluntary ambition targets. These kinds of labels or certicates have proven to be effective for communication purposes and also for promoting the commercial value of being eco-friendly (Bougdah and Sharples, 2010). The various tools have in common that they unite the broad range of environmental aspects of buildings into one single metric by using weighting functions. The assessment tools do evaluate energy consumption in the whole buildings life-cycle, and also include issues like transportation, land use, materials, pollution and water consumption (Gowri, 2004). The prediction of energy-related aspects varies among the tools; some use simplied (steady-state) calculations while others are based on the results of dynamic performance simulations. But in general, all rating tools share the characteristic that Energy Performance of the building is eventually cast into one single value. The appropriateness of this approach in assessing environmental impact with regards
The efciency of the solar cells is a function of layer temperature. In turn, power output is subtracted from the nodal

absorption of the incident beam which reects the fact that part of absorbed solar radiation is converted to electrical power rather than being manifested as a heat gain

48

3.7. Discussion to CABS is disputable because of the following reasons. CABS are inherently dynamic and its performance is closely related to the specic application (e.g. climate, surroundings, orientation, building function) and control strategies. Consequently, it is complicated to generalize CABS performance aspects in terms of universal points or credits. In addition, CABS specic merits might get obscured in the weighting process with the other sustainable assets. Some authors further argue that the value of environmental assessment tools in the design process of buildings is limited (Ding, 2008; Lstiburek, 2008). Despite of its good intentions, the tools are used as checklists for endorsement of already completed designs rather than being used for supporting informed decisions in the actual design process. Analogue to the green building rating schemes, there is also an increasing interest to establish systems for energy labeling or energy rating of windows. These ratings are typically used for communication and marketing purposes and should be understandable for consumers. The current window energy rating systems (WERS) are based on the following static fenestration properties: U-value, solar heat gain coefcient, visible transmittance and air leakage (Urbikain and Sala, 2009) which makes them unsuitable to capture the dynamics of CABS. Recently, Papaefthimiou et al. (2009) proposed a new environmental rating scheme for switchable windows. This assessment indicator includes dynamic energy saving potential calculated with optimal control strategies in a customized BPS tool. Uncertainty and risk For prediction of building performance, several assumptions have to be made regarding e.g. real occurring weather-data, occupants behavior, and control actions. The innovative materials and emerging technologies abound in CABS further increases the level of uncertainty inherent in the problem denition process and associated with the selection of the thermophysical properties of materials (Clarke, 2001). While being aware of these uncertainties, it is a rather awkward situation that BPS typically computes results for only one scenario, that is at best an educated, but always arbitrary guess. Consequently, design decisions informed by BPS do not acknowledge the probability distribution of possible outcomes which is closely correlated to an increased risk, with associated consequences, of not meeting performance targets. Traditionally serviced buildings use oversized HVAC-systems as a means of overcoming the effects of uncertainty. In buildings with CABS, a more risk-conscious design approach seems a necessary step forward (Hu, 2009). Various types of uncertainty analyses (UA) are available to elucidate the effects of (i) physical, (ii) design and (iii) scenario uncertainties on building performance (Hopfe, 2009). Currently, UA-methods still exist is the form of research prototypes that require laborious pre-processing and post-processing efforts. It is now up to the software developers to merge these methods in the tools existing workows to let become UA a valuable tool for guiding both buildings and CABS innovation processes. Look at the future The analysis in this chapter mainly focused on (commercial) off-the-shelf software tools, because it is believed that these tools have the highest potential to be valuable for architecture, engineering and construction (AEC) professionals in the design process of CABS. However, the availability of adaptive features in these tools appears to be limited, which connes the option space for simulation of CABS. In academia, alternative developments are going on that will allow us to simulate buildings with much greater exibility and extendibility. By exploiting advances in computer science and software engineering these tools are at the source of what might become the next generation of BPS tools. This new approach combines a higher abstraction level with object-oriented modeling and in this way decouples the mathematical models from the solver in the simulation program. This results in BPS tools with general-purpose symbolic programming languages (e.g. Modelica) that discard the xed input-output relationships, and hence gives users the liberty to tune or extend models in an intuitive way. Wetter (2009a) identies four applications of a new Modelica-based modeling and simulation environment for building energy and control systems. The simulation tool should be able to: offer a platform to stimulate innovations in energy efcient building systems; 49

3. B UILDING P ERFORMANCE S IMULATION enable virtual rapid prototyping to select most promising alternatives for further renement and product development; permit researchers to quickly add models for performance assessment of emerging technologies; allow for extraction of design models to be embedded within building control systems for modelbased controls, fault detection and diagnostics. All the envisioned applications seem both relevant and promising for performance prediction of CABS as well. A concurrent research project at the University of Ljubljana already demonstrated the preliminary application of Modelica for performance simulation of building shells with dynamic properties (Zupan i and Sodja, 2008). Before this approach can become more mature, some important steps are cc required: the essential validation studies still have to be undertaken, and more attention has to be paid to the design of convenient user interfaces.

50

CHAPTER
User
Simulation requirement
ab str ac tio n
le Re

M ODELING AND SIMULATION OF CLIMATE ADAPTIVE

BUILDING SHELLS

Thus far, this thesis presented the properties and characteristics of CABS in Chapter 2, which has also led to the CABS overview in Appendix A. An analysis of the capabilities of state-of-the-art building performance simulation tools with respect to CABS is given in Chapter 3. Successful simulation of building performance is a challenging task and has been phrased as the art of performing the right type of virtual experiment with the right model and tool (Augenbroe, 2010). In this chapter, the two lines converge as it presents a strategy for effective modeling and simulation of CABS. A graphical representation of the developed strategy is given in Figure 4.1. The remainder of this chapter follows the logic of this graph.

Conceptual Models

n va

M od el

s sic hy tp

CABS Overview

Verification

Validation

Real World System

Tool Selection

BPS-tools Overview

Figure 4.1: Strategy for modeling and simulation of CABS.

4.1 Model abstraction

Simulation projects usually originate from queries about real world systems, which for CABS in most cases is a design for a new building or building shell concept. Before the performance of these objects can be predicted, the systems rst need to be translated into conceptual models. A conceptual model is 51

Sim u

Simulation Model

lat io

ns tra teg

ion ut ec Ex

4. M ODELING AND SIMULATION OF CLIMATE ADAPTIVE BUILDING SHELLS dened as a non-software specic description of a computer simulation model, describing the objectives, inputs, outputs, content, assumptions and simplications of the model (Robinson, 2008). Effective conceptual models of CABS contain the critical entities and relationships at the right level of detail to predict actual behavior with sufcient accuracy. In order to substantiate the choices made during the model abstraction process, it is essential to have sufcient domain knowledge and insights in occurring physical processes. Therefore, the rst step in model abstraction is knowledge acquisition (Kotiadis and Robinson, 2008). Conceptual models are either (i) physically-based or (ii) data-driven. Models of the rst type are based on rst-principles and theoretical considerations like mass and energy balances, differential equations and conservation laws. The second type is also known as the black-box or input-output approach and constructs models entirely based on mapping and tting input-output relationships based on observations from experiments or data. Models of this kind do not necessarily reect the structure of the actual system nor contain physically meaningful parameters (Henze and Neumann, 2010). Nowadays, semi-physical or grey-box models are gaining popularity as a means of overcoming the deciencies of one of both individual approaches (Lu et al., 2009). Modeling and simulation by denition implies approximation, introduced in the abstraction process by means of assumptions and simplications. Assumptions are ways of of incorporating uncertainties and beliefs about the real world into the model (Robinson, 2008). In this way, it deals with (quasi-) randomness and unknowns about the system and despite of this tries to obtain reliable outcomes. Here it should be noted that it is not a model itself, but the models results that should be close to reality (Chwif et al., 2000). Simplication is concerned with deliberately reducing the complexity of a model with the aim of increasing the models utility without signicantly affecting its accuracy and validity (Barlow, 2009). This involves reducing either (i) the scope (i.e. boundaries of the study or number of components), (ii) the resolution (i.e. level of detail in the components) or (iii) the interconnections among components of the model. In general, the model abstraction process establishes which aspects of the real-world system must be modeled, and for those to be included, how accurate the model must be. Consequently, conceptual models of CABS can be developed at various levels of detail and complexities. Finding the right balance between model complexity, efciency and resulting accuracy should be a major consideration in every model development process (Lu et al., 2009). Increasing the complexity of conceptual models adversely affects ease-of-use, computational performance, exibility and cost (Henriksen, 2008), and therefore it is not always the most sophisticated modeling approach that offers the highest value. Depending on available system knowledge, increasing complexity beyond a certain level further also reduces the utility of a model in terms of accuracy due to higher uncertainties involved (Tr ka and Hensen, 2010). The c nal structure and contents of the model is always determined in consultation with the simulation requirement, as discussed in Section 4.2. Once implemented, the validity of this model is either approved or the model needs to be modied on the basis of the verication and validation process which is described in Section 4.5. Nevertheless, all conceptual models of CABS should consider the following two elements: adaptive mechanisms and control. Both elements are discussed in more detail rst.

Adaptive mechanisms A description of how adaptive mechanisms are represented has to be part of every CABS-model, since adaptivity is what discriminates CABS from the conventional building envelope. Conceptual modeling of micro-level CABS is relatively straightforward: it basically involves a direct change in thermophysical or optical material properties. Therefore these CABS-concepts can rely on existing BPS models, but now with exible, rather than the traditional rigid material properties. Macro-level CABS on the other hand may not only affect material properties but also cause changes in the conguration or composition of the building shell. Consequently, an additional interpretation step is required that translates these adaptive mechanisms to an abstraction level that is manageable for the available BPS tools. 52

4.2. Simulation requirement Control Another important issue in conceptual modeling of CABS is the identication of the way in which adaptive behavior is controlled. As Section 2.3.3 already points out, two different types for control of adaptation in CABS exist: open-loop and closed-loop control. In the rst type, adaptive behavior is an intrinsic characteristic of the design of the building shell and these specic input-response relations have to be included as such in the conceptual model. Closed-loop control does give designers and developers the freedom to devise appropriate control strategies that satisfy their requirements. It is important to take these control concepts into account in the model because the way in which the CABS is controlled will potentially also have implications for the software-side of the simulation process. For this purpose, the following three types of closed-loop control are distinguished: Schedule based: The adaptation follows a scheduled or scripted sequence that is dened in advance. Strictly speaking, building shells controlled in this way are dynamic, but do not belong to adaptive building shells, because they do not have the ability to react in response to varying conditions. The adaptation rate is in principle free to choose, and can follow for example diurnal cycles or seasonal patterns. The switching frequency is however restricted by the length of the time-step used in the simulations. External variable based: Adaptive actions of this type are directly a function of one of the inputs to the simulation model. Most likely, this is a variable in the weather-le, another possibility is to control adaptivity as a function of occupancy level. The control of CABS in this way is independent of the output of the simulations. In case of coupled simulations, there is no need for communication of adaptive states, as long as the boundary conditions supplied to the models are the same. Internal variable based: The last, and most complex type of control is based on internal variables, which means that control stimuli are calculated during runtime. This typically is the case when adaptive behavior is controlled in response to comfort considerations. The complicating factor is that the control stimulus and adaptive behavior are mutually inuencing each other. This prevents the use of stand-alone simulation approaches and requires bi-directional workow in the case of co-simulation. In addition to the three mentioned approaches, all types of hybrid control schemes are possible as well. One can think of CABS-examples where the type of control strategy differs in various time-slots. Other examples of hybrid control are e.g. time-scheduled envelopes with the possibility for manual override when certain comfort limits are exceeded, which is in principle a combination of schedule based and internal variable based control.

4.2 Simulation requirement

Selection of the right conceptual model depends on what goal the user wants to accomplish with the simulation process, and the type of information he or she wants to obtain. As this is an iterative process, the simulation objectives should be considered throughout the whole model development cycle. Performance simulations of CABS exist in many forms and can amongst others be used for gaining more insight in cause-and-effect relationships, demonstrating performance benets to convince investors, investigating the effects of control strategies, the purpose of model-based control etc. Each particular application demands for consideration of different conceptual models of the same system at various levels of explicitness and accuracy. Consequently, it is wise to focus on those parts of CABS that are relevant to the problem situation and the modeling objectives, and to develop a model that only addresses these objectives, rather than a general model of the system that includes all that is known about the real world system. As will be shown next, the simulation requirement consists of two parts: the performance indicators of interest at the right resolution. 53

4. M ODELING AND SIMULATION OF CLIMATE ADAPTIVE BUILDING SHELLS Performance indicators Chapter 1 demonstrated that performance of CABS is a very broad concept. Some of these performance aspects may be conceived as being isolated, whereas others demand for an integrated evaluation. The level of building performance is usually expressed via performance indicators (PIs) which adequately represent satisfaction with a particular performance requirement of a system (Augenbroe, 2005). These PIs must be quantiable, well understood and amenable to computational analysis (Becker, 2008), and in this way they allow for comparing and contrasting various design options and the effects of operational strategies. PIs are coupled to objectives or performance requirements of the building via e.g. target values, minimum requirements and acceptable discomfort levels. Every physical domain in CABS typically uses its own set of PIs. However, some PIs can be combined to result for example the buildings overall energy performance. The trade-offs between multiple performance aspects can be evaluated by using multi-criteria decision analyses. Sometimes the user is only interested in the performance of a single component or sub-domain of the overall system. In this way the system boundaries of the conceptual model can be conned, which reduces the overall scope and thus complexity of the problem. Resolution The type of performance indicator under examination is closely related to the minimum required resolution of the simulation problem, and this in turn denes the eligible BPS tools. CABS performance is by denition a dynamic phenomenon, and in order to represent the system dynamics in simulations, the continuous problem has to be discretized in time-steps. The length of time-steps, or temporal resolution, needs to be matched with the characteristic time-length of the simulation problem to capture all relevant processes. New for CABS is the increased relevance of transient comfort aspects, and therefore it is important to focus on the rate-of-change of PIs as well. The way in which PIs are quantied also relates to the spatial resolution of the simulation problem. In various CABS-concept, the adaptation and corresponding performance depends on location-specic physical conditions in immediate vicinity to the respective sensor, occupant or active material. In these cases it might be insufcient to assume uniform or homogeneous conditions in the whole zone, especially when large gradients occur. Additionally, assessment of some comfort PIs further demands for high-detailed analysis; examples are effects of cold downdraught and the risk of glare. The spatial and temporal resolution together determine the required minimum level-of-detail and thus complexity of the simulation model. Djunaedy (2005) distinguishes the minimum modelresolution required to calculate particular performance indicators for simulations in the thermal and airow domains. Analogously, Reinhart et al. (2006) and Mardaljevic et al. (2009) provide guidelines for assessing the performance of daylight systems. All the authors stress that conceptual models should be based on the nature of the problem; not led by the capabilities a specic (familiar) BPS tool offers.

4.3 Relevant physics

Figure 4.1 shows that the relevant physics for a simulation problem is a product of the model abstraction process and the simulation requirement, and in this way is application-specic. In addition the models are different and characteristic for every particular CABS-concept. This section presents an attempt to clarify the relevant physics for a simulation problem by means of a graphical representation. Graphs of this kind were assigned to all 100 CABS concepts in Appendix A. The methodology for identication of relevant physics consists of the following four steps: 1. Identication of all physical domains that play a role in performance assessment of the CABSconcept. Classication in domains is identical to the rules for classifying CABS, presented in Table 2.2: (A) thermal, (B) optical, (C) air-ow and (D) electrical. 54

4.3. Relevant physics 2. Determination of the occurring adaptive mechanisms by assigning adaptivity to the appropriate physical domain(s). This determines which adaptive mechanisms should be considered in the model and is indicated with lled circles in the graphs. 3. The way in which adaptivity is controlled is indicated in red and may be either specic for the given situation (open-loop) or free to choose (closed-loop). For open-loop CABS, the red arrow represents the way in which the environment inuences adaptive behavior. In schedule based and external variable based control, it is not the simulation outcome, but the boundary conditions that force the building shell to change. The last type is based on internal variables, here indicated via red circles to represent the physical domain(s) in which the stimulus is imposed. 4. Lastly, the appropriate physical relationships and directions of interaction are established. These interactions are prescribed either by the intrinsic physics of the problem and/or the way in which adaptivity is controlled. Arrows indicate the nature of the relations and can be one-way or twoway. The rst type is only concerned with unidirectional dependencies and thus can be modeled in a sequential way. In the second case, the effects are mutually interrelated and therefore need to be considered simultaneously. In order to make the concept more clear, two examples are given that show how these graphs can be used for transforming relevant physics into an appropriate simulation strategy. Figure 4.2 shows the graphs for thermochromic and photochromic windows respectively. In both CABS-concepts, only thermal and optical effects play a role, and they further share the characteristic that adaptivity is manifested via changes in the optical properties of the window, as indicated with the grey circles. The open-loop stimulus that controls this behavior is however different in both cases; for thermochromic windows this takes place in domain A and for photochromic windows in domain B. As the black arrows indicate, this difference also has implications for the physical interactions to be considered in the model. Simulations in the thermal domain require input from the optical domain in both instances, as is inherent for every CABS with transparent parts. In thermochromic windows, adaptation occurs as a function of window layer temperature. When the user is only interested in the visual comfort of this switchable window, he or she is compelled to simultaneously consider the thermal domain in order to include the changing window state. For photochromic windows on the other hand, visual comfort performance can be be assessed via single domain lighting simulation tools since thermal interactions are not relevant.

(a)

(b)

Figure 4.2: Relevant physics for (a) thermochromic and (b) photochromic windows. Where A indicates thermal and B the optical domain; the red arrows indicate the open-loop stimulus and the black arrows indicate the corresponding physical interactions.

The graphs for electrically heated glass are given in Figure 4.3. It shows that airow is not relevant in this case and that adaptivity takes place in the electrical domain. Moreover, no arrow is pointing to domain B which implies that queries about visual comfort or daylighting performance can be answered in a stand-alone manner with rigid material properties. As this is a closed-loop controlled CABS, adaptive behavior can be controlled in many different ways and therefore it is insufcient to use only one graph to represent relevant physics. Figure 4.3 (a) shows the relevant physics for schedule based and 55

4. M ODELING AND SIMULATION OF CLIMATE ADAPTIVE BUILDING SHELLS external variable based control, where adaptation happens without regard to simulation outcomes. in these cases, only the minimal physical interactions are required in the model. Figure 4.3 (b) and (c) both represent internal variable based control, as a function of inputs from the thermal and electrical domain respectively. The graphs reveal that in case adaptivity is controlled in the thermal domain (e.g. as a function indoor air temperature), an additional coupling from the thermal to electrical domain is mandatory.

(a)

(b)

(c)

Figure 4.3: Relevant physics for Electrically heated glass. Where A indicates the thermal, B the optical and D the electrical domain; the red lines represent the type of control, and the black arrows indicate the corresponding physical interactions.

4.4 Simulation strategy

Conceptual models are dened to be non-software specic representations of a real-world system. However, once the relevant physics of the problem is established, the building-specic conceptual model needs to be implemented in one of the available BPS tools. Selection of appropriate candidate tools has to consider four requirements: 1) the adaptive features in the conceptual model; 2) the desired (mix of ) performance indicators at the appropriate level-of-detail; 3) the way adaptive behavior is controlled; and 4) the applicable physical interactions. Chapter 3 presented the foundations for this tool selection process, including the available adaptive mechanisms in the state-of-the-art BPS tools. Like every other BPS study, the simulation strategy for performance evaluation of CABS further includes specication of scene and scenario and selection of appropriate boundary conditions. Depending on the type of simulation this includes development of a 3D building model and selection of a location and its surroundings, a climate le, occupancy proles, internal gains etc. The user should further consider the required simulation period and an appropriate length of time-steps. In addition, the output variables and the way how the results are going to be presented should be decided. Users may be confronted with the situation that no tool is eligible for implementing the desired behavior. In these cases, the user should rst reconsider the conceptual model and/or the simulation requirement. This perhaps requires more rigorous assumptions or a less detailed analysis, which may, or may not violate the utility of the simulation process. In case these changes are unacceptable, users can endeavor to extend the capabilities of the current range of BPS tools. It should however be noted that this is a cost- and time-intensive process an effort that might be unjustiable in a routine design process.

4.5 Verication and Validation

Performance simulation of CABS is often concerned with innovative technologies which in general lack well-accepted modeling strategies. The assumption that a model is reliable can however not be made purely on the basis of good intentions by the developers (Sokolowski and Banks, 2009). For this reason, verication and validation (V&V) studies should be part of every CABS model development process. 56

4.5. Verication and Validation The role of verication and validation in relation to the other components of modeling and simulation of CABS is illustrated with the dashed lines in Figure 4.1. Verication is concerned with transformational accuracy, i.e. that of transforming the models requirements into a conceptual model and the conceptual model into an executable simulation model. In this way, verication conrms or rejects that the model implemented as software, is consistent and complete with respect to the conceptual model. Validation is concerned with representational accuracy, i.e. that of representing the real-world system in the conceptual model and the results produced by the executable simulation model (Sokolowski and Banks, 2009). Validation of CABS-models is always to be regarded within the specic domain of applicability, and hence accuracy and error tolerances should be consistent with the specic simulation requirement (Balci, 1998). There is no general recommended process or standard procedure available for selecting the right type of verication or validation technique (Sargent, 2005). In verication, general software engineering methods that compare the implemented model with its design specication are often applicable. During the development of CABS models, it is not the intention to prove validity of the general simulation algorithms of the BPS tools, since these codes have gone through rigorous tests via inter-program comparative (e.g. Judkoff and Neymark (1995); Reinhart and Herkel (2000)) and empirical validation studies (e.g. Mardaljevic (1995); Reinhart and Walkenhorst (2001); Judkoff (2008)). Instead, V&V studies should be used to conrm that the assumptions and simplications about adaptive behavior are justied.

57

CHAPTER
5.1 Introduction
59

M ODELING S MART E NERGY G LASS

The present chapter is the rst of three chapters that deals with the case study of Smart Energy Glass. This chapter starts with an introduction of the concept and then provides a brief description of underlying physics. Subsequently, the conceptual model development process is discussed according to the strategy for modeling and simulation of CABS as presented in Chapter 4. The chapter concludes with a description of the simulation strategy and shows how the conceptual model was implemented in software tools. Validation of this model is the subject of Chapter 6 and is followed by an application study of the model for performance simulation of Smart Energy Glass in Chapter7.

Smart Energy Glass (SEG) is an innovative window technology, currently under development by Peer+; a spin-off company from the faculty of chemical engineering at Eindhoven University of Technology (Peerplus, 2009). The working mechanism behind SEG is based on a polymer coating sandwiched between two layers of glass. These layers together form the external pane of an insulated glazing unit (IGU). By changing an external voltage applied to the coating, it is possible to control the optical properties of the window within less than a second. SEG can be switched into three different states: a bright state, a dark state and a translucent state (Figure 5.1). The polymer coating in SEG also acts as planar waveguide in the same way as a luminescent solar concentrator (LSC) (Goetzberger and Greube, 1977). The dye molecules absorb part of the incoming sunlight and re-emit photons in approximately random direction at a longer wavelength. Via this mechanism part of the incoming light is captured and redirected to the edges of the window where photovoltaic cells are situated to convert the collected radiation into electricity. The generated electricity is used for switching the state of the window, the surplus can be fed back to the grid. Switching the window can exploit the qualities of multi-ability since the translucent state then offers privacy rather than transparency. A more important aspect of changing the window state is that the ow of solar gains and admission of daylight can be actively modulated. In this way, the window can provide passive heating in the winter while reducing cooling load and overheating risk in the summer. Exploiting this adaptability further allows for maximizing the positive effects of daylight in terms of energy and comfort, but also acts as active instrument to counteract the likely occurrence of glare. SEG is a unique concept that facilitates energy saving while generating energy by combining display technology and photovoltaic energy conversion into one device. The window tint can be produced in various different colors, and the electrical connectors are invisible. In addition, no external wiring is

5. M ODELING S MART E NERGY G LASS

closed-loop

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MICRO

Smart Energy Glass

Peer+ | Eindhoven, The Netherlands

An innovative coating allows the window to switch its optical properties in three different states: dark, bright and privacy. Part of the sunlight is captured and converted into electricity in PV-cells mounted at the edges of the window. When equipped with batteries, the system becomes autonomous and can be integrated in the building envelope without the need for external wires.

75

[www.peerplus.nl]

Peer+ Smart Energy Glass, Website: http://www peerplus.nl Kemper, A. (2008) Schone energie dankzij superslimme ruiten Volkskrant 14 oktober 2008

Figure 5.1: Smart Energy Glass info sheet in the CABS overview (Appendix A).

required when electricity is stored in batteries, which makes SEG very attractive for renovation projects (Benson and Branz, 1995). SEG is however not the rst technology that aims at adding adaptivity and/or electricity generation to the list of window functionalities. Table 5.1 contrasts the capabilities of SEG with a selection of competitive technologies.

Table 5.1: Comparison between various types of advanced window systems.

Energy generation Smart Energy Glass Semi-transparent PV LSC Switchable glazing Photovoltaic shading Sphelar (Peerplus, 2009) (Wong et al., 2008) (van Sark et al., 2008) (Granqvist, 1991) (Janak, 2009) (Biancardo et al., 2007) Views unobstructed Switchable properties Unobtrusive appearance

The concept of SEG is currently not a market-ready commercial product, but is still under continuous development. Research and development activities currently focus on optimizing absorption and emission spectra, thermal performance of the window, optical losses, electrical circuits, stability and longevity of the dye, etc. The work presented in this thesis complements and interacts with these activities by providing computational support for assisting the innovation processes. 60

5.2. Real world system

5.2 Real world system

The characteristics of SEG are already briey introduced in Section 5.1. To predict the performance of the system, this information rst needs to be abstracted into an appropriate conceptual model and afterwards into a simulation model. To allow for substantiated assumptions and simplications in these models, it is important to further elucidate the physical processes occurring in SEG. In its most elementary form, SEG can be described as a LSC with tunable optical properties (Debije, 2010). The working mechanism of LSCs basically consists of two parts: (a) collection and concentration of photons and (b) conversion of these photons into electrons. Figure 5.2 gives a schematic overview of working mechanisms of SEG.

(i) (ii) (v) (iv) (vi)

PV cell

(iii)

Figure 5.2: Schematic overview of the physical phenomena taking place in SEG.

A stream of photons coming from the sun hits the window pane, where the light is either reected (i) or refracted (ii) into the material. In the bright state, most photons that enter the window do not get absorbed and thus again hit the interface. Depending on angle of incidence and refractive index, this photon ux may either be internally reected or leave the window as transmitted radiation (iii). The dening characteristic of SEG is the luminescent dye that has been dispersed with the goal of absorbing and re-emitting (iv) a controllable part of incoming radiation. The direction of re-emission is approximately isotropic, and consequently, a large fraction of the re-emitted photons is trapped in the waveguide via internal reection (v). Successive reections carry the luminescent photons to the edges of the plate where they are absorbed, captured and converted to electricity in PV cells (vi). The bandgap of these solar cells is matched with the spectral emission wavelength of the dye with the aim of obtaining highest possible efciency. LSCs are able to collect both direct and diffuse radiation without any moving objects and are particularly interesting because they take advantage of the geometric concentration factor between front surface and edge surface of the window. The electrical performance of LSCs is however subject to several loss mechanisms and fundamental efciency limits, some are indicated in Figure 5.3. The majority of losses in SEG occur because: (i) not all absorbed light is re-emitted (quantum efciency of the dye); (ii) absorbed light is re-emitted at longer wavelengths and thus possesses less energy (Stokes shift); (iii) imperfections in substrate layers scatters part of the light; (iv) not all luminescent light is trapped in the concentrator (loss cone)(A); and (v) not all trapped luminescent light reaches the edges due to (self ) absorption in the dye (B) or bulk molecules (C). Switching window state upon application of external voltage affects the global alignment of molecules in the dye, and this in turn changes the relative probability that each of the events in Figures 5.2 and 5.3
Approximately 75 % for a plate with a refractive index of 1.5 (Goldschmidt, 2009)

PV cell

61

5. M ODELING S MART E NERGY G LASS

(A)

PV cell

(B)

Figure 5.3: Schematic overview of the loss mechanisms in SEG.

takes place. As this probability distribution inuences transmission, reection and absorption characteristics of the dye, it ultimately results in changes in optical properties of the window. Figure 5.4 shows for example the spectral transmission curves of a SEG-sample in the dark and bright state determined by performing measurements with an integrating sphere and spectrophotometer. Like every other optical device, SEG satises the condition that the sum of reectance, transmittance and absorptance equals one for every wavelength.
1

Transmission (-)

Bright state Dark state

0 200 400 600 800 1000 1200 1400 1600 1800 2000

Wavelength (nm)

Figure 5.4: Spectral transmission curves of a SEG sample.

Classication Based on the foregoing discussion of governing physics in SEG, it is now possible to classify the concept according to the categorization of Chapter 2. Adaptive behavior is instigated by changing the applied external voltage to the SEG layer. The concept thus includes a decision making component and therefore the control in this type of CABS is closed-loop based. Switching the window in either the bright, dark or translucent state causes the molecules of the polymer dye to rearrange. This adaptation takes place at molecular scale and hence SEGs adaptivity takes place at the micro-level.

5.3 Conceptual model

The relevant physics that determines the performance of SEG is identied to take place in three different domains: thermal, optical and electrical. As a consequence, the conceptual model also consists of 62

PV cell

(C)

5.3. Conceptual model three parts. The development of conceptual models for each component is discussed in the following subsections.

5.3.1 Optical Model

The goal of the optical model is to predict visual performance and associated energy demand of switching SEGs window state by considering both daylight and articial lighting. Visual comfort comprises many factors; on the one hand, the daylight system needs to ensure adequate illuminance levels whereas on the other hand excessive brightness or glare needs to be avoided. Adaptivity in SEG is manifested via changes in the optical properties of the window assembly (Figure 5.1). These optical properties of SEG-samples were determined in experiments with an integrating sphere and spectrophotometer. Values for reectance and transmittance were measured according to the protocols in ISO 9050. This data was then transformed into spectrally averaged properties with the aid of the software tool OPTICS, developed by Lawrence Berkeley National Laboratory (LBNL, 2010). The software reads in measured data as laminate interlayer construction, where the SEG-layer is sandwiched between two layers of glass with known properties (Figure 5.5).
Reference laminates substrate Interlayer

+
Interlayer

OPTICS algorithms

Figure 5.5: Schematic representation of the laminate interlayer construction method in OPTICS.

The algorithms in the tool (Rubin et al., 1998) then extract the optical properties of this interlayer to yield the properties of the SEG coating only. Subsequently, OPTICS serves as virtual glass laboratory since it allows this coating to be applied to an extensive number of glazing materials available in the international glazing database (IGDB). After constructing an appropriate fenestration system, the optical properties of the entire SEG window can be analyzed. These calculations include the photopic response curve that accounts for the relative spectral sensitivity of the human eye according to ISO/CIE 10527. One of the outputs of OPTICS is the visible transmittance (Tvis ) of the entire window system, which is different for each window state.

5.3.2 Thermal Model

Although SEG fullls the function of window in the building shell, it in fact is a LSC with controllable optical properties and corresponding physics. Nevertheless, thermal behavior of SEG is approximated in this study as a conventional fenestration product, and modeled accordingly. The objective of this thermal model is to facilitate insights in the effects of SEG on thermal comfort perception and to demonstrate its saving potential for heating and cooling energy demand. Heat transfer phenomena in windows are well understood and usually characterized by using validated modeling methods as documented in e.g. ISO 15099. The inuence of windows on the buildings overall energy balance depends on the location of the window, its size and construction (including cavity, frame, spacers etc.) as well as the optical properties of the glazing. The momentary energy ow through windows can be split in the following two contributions, and is frequently calculated with Equation (5.1) (ASHRAE, 2001). Temperature driven heat transfer When a temperature difference between the inside and outside is present, heat is gained or lost through the glazing via the combined effects of conduction, con63

5. M ODELING S MART E NERGY G LASS vection, and radiation. These effects are quantied in terms of U [Wm2 K1 ]; the U -value or total heat transfer coefcient of a fenestration assembly. Solar gain Regardless of the temperature difference across the glazing, heat can be gained through fenestration products by direct or indirect solar radiation. The amount of heat gain through products is measured in terms of the solar heat gain coefcient (SHGC ) [] of the glazing. q = U A T + SHGC A I tot (5.1)

Where q is instantaneous energy ow [W ], A is window area [m2 ] T is difference in air temperature between inside and outside [ C] and I tot is incident total irradiance [Wm2 ]. Although U -value and SHGC are effective parameters for comparing window technologies, their value in modeling thermal window behavior is limited since both terms disguise the contributions of particular heat transfer mechanisms. The window properties used to model the thermal behavior of SEG are: convective heat transfer coefcients, thickness and conductivity of individual window panes and gaps, emissivity at front and back surface, and transmittance, reectance and absorptance as a function of angle of incidence. Numerical values for these thermal properties are again established on the basis of the experimental data as described in Section 5.3.1. The datasets for the different window states were rst extended with properties for the infrared wavelength area and then read in by the software-tool WINDOW5 (LBNL, 2010). This program then calculates properties of interest for SEG integrated in an arbitrary window system in compliance with ISO 15099.

5.3.3 Electrical Model

The aforementioned optical and thermal models of SEG are relatively straightforward and conform to the customary practice in building performance assessments. The electrical model aims to predict the annual electrical energy yield of SEG and therefore requires a more in-depth analysis of the photon-toelectron conversion process. These phenomena have been modeled for various purposes in the past, however no study has been reported in BPS literature. Figure 5.6 presents an illustrative impression that shows the complexity of photon trajectories inside LSCs.

Figure 5.6: Visualization of three-dimensional photon trajectories in a LSC (Schler et al., 2007).

The photon-to-electron conversion in these devices basically comprises a chain of gain-factors and efciencies for each (loss) mechanism in the process. But as Goldschmidt (2009) argues, these efciencies are neither easy to calculate, nor directly accessible by measurement. For this reason, several approaches for modeling the behavior in LSCs have been proposed in literature. In general, two different modeling methods can be distinguished: the rst type is based on explicit thermodynamic energy 64

5.3. Conceptual model balances (Chatten et al., 2003; van Sark et al., 2008) while the second type employs (Monte-Carlo) raytracing techniques (Bendig et al., 2008; Schler et al., 2007). Both approaches have in common that large numbers of parameters are required which introduces a large variance in possible outcomes and thus a high level of uncertainty. These simulations have been used to investigate the relative contributions and to assess which mechanisms are important in the collection and conversion processes. The focus in these studies was on optimization of product design in the research stages, with less attention for the actual output in large-scale settings under real conditions. Consequently, the resolution of these LSC models does not accord with the simulation requirement in this study. Therefore the decision was made to develop a data-driven model, not based on rst-principles, but on empirical knowledge obtained after conducting experiments. Electrical performance of LSCs is typically characterized in laboratory settings under standard test conditions (STC). These controlled conditions assume constant irradiance of 1000 W/m2 (AM 1.5 spectrum) at normal angle of incidence and a module temperature of 25 C. Outcomes of these experiments are then converted in maximum efciencies and used for comparisons of peak performance rather than focusing on annual energy yield. When installed in buildings, solar radiation incident on LSCs virtually never impinges at normal angle of incidence nor at 1000 W/m2 . To illustrate this, Figure 5.7 shows the angles of incidence of direct solar radiation on a vertical plane facing the four cardinal directions in the Netherlands.
800

700

North South West East

600

500

Hours per year

400

300

200

100

0 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90

Angle of incidence ()

Figure 5.7: Direct sunlight incident on vertical faades with different orientations in the Netherlands.

This deviation from optimal environmental conditions is expected to have considerable inuence on the annual electrical output of SEG. The experiments that feed the data-driven model were designed to emphasize this year-round behavior. Based on examined LSC literature, the hypothesis was made that the electrical output can be described as a function of intensity of incident solar radiation (I in ), angle of incidence of solar radiation (), window state, and geometric gain (G). Where the latter factor is specic for a given situation and dened as the ratio between front surface (A f ) and edge surface (A e ) of the window (Batchelder et al., 1979). A description of the experimental setup and interpretation of the data is presented in the following subsections. Experimental setup Figure 5.8 shows two pictures of the experimental setup that was used to supply the data-driven model. The tests were carried out on a sample of SEG with an area of 5 5 cm2 and a thickness of 6 mm, with65

5. M ODELING S MART E NERGY G LASS

Figure 5.8: Overview of the experimental setup (left) and close-up of the sample (right).

out photovoltaic cells. The sample was illuminated by a 300 W solar simulator equipped with lters to approximate the spectral make-up of the sun (AM 1.5 spectrum). The output emission coming from the edge of the sample was determined by placing one end of the sample adjacent to the entry port of an integrating sphere that was equipped with a spectral light measurement system (SLMS) array to determine the spectral radiant power.

Figure 5.9: Schematic representation of the experimental setup showing angle of incidence (red) and the direction of output (blue).

Figure 5.10: Projected area of the sample as a function of angle of incidence.

Variation of incident angle was accomplished by rotating the samples supporting mechanism in the direction of the white arrow in Figure 5.8. This set-up is further illustrated in the schematic representation in Figure 5.9. The graph shows that the angle of incidence was adjusted in the xz-plane while output was measured in the y-direction only. This conguration ensures that merely the effects of waveguiding are considered, not disturbed by effects of direct radiation. Moreover, all other ambient light sources were switched off. The output emission was measured in consecutive steps, for one edge at a time while black tape was applied to the other edges of the sample in order to prevent light from scattering in at other sides than the window face only . The spectral radiant power from the edge of the sample was recorded in eight discrete steps of 10 degrees, ranging from normal (i.e. 0 ) to 70 . Higher incidence angles were not attainable due to the practical limitation that the rotation angle would become too steep for the sample to remain xed. Furthermore, the intensity of incoming light was modulated by changing the applied current to the light source. Three different congurations were investigated, corresponding to a solar irradiance of 700 W/m2 , 920 W/m2 and 1170 W/m2 respectively.
The black tape is not shown in Figure 5.8

66

5.3. Conceptual model Results Figure 5.11 shows the spectral radiant power leaving the exit aperture for the wavelength-interval from 400 to 815 nm when the sample was illuminated with an irradiance of 920 W/m2 . The output in the graph shows a peak around 700 nm which coincides with the emission spectrum of the dye. The gure also shows that output at the edge decreases as increases, and that the level of decay increases at more oblique angles of incidence. Part of this effect is explained by the projected area and thus incident radiation received by the sample that drops with the cosine of angle of incidence (Figure 5.10). In addition, a larger part of incoming light gets reected at the interface at higher incident angles. The signicance of these phenomena is however more comprehensible when compared on the basis of efciency, as is demonstrated later. Figure 5.11 further shows that the tail end of the emission spectrum is missing in the analysis. This is due to the limited detection range of the spectrophotometer used in the experiments. Consequently, this shortcoming introduces an underprediction in measured output of approximately 5%. Further inaccuracies penetrate into the measured output because (i) spectral makeup of the light source is only an approximation of a single sky conguration, and (ii) adjusting the angle is done manually, in a coarse way.
0,14

0,12 Spectral radiant power (mW/nm)

0,10

0,08

0,06

0,04

0,02

0,00 400

450

500

550

600 650 Wavelength (nm) 30 40 50

700

750

800

10

20

60

70

Figure 5.11: Spectral radiant power as a function of wavelength and angle of incidence.

The radiant output power as a function of incidence angle for the three levels of irradiance is given in Figure 5.12. Radiant power is obtained by integrating the spectral output power of Figure 5.11 over the wavelength-interval from 400 to 815 nm. On the basis of these results, it is possible to calculate the optical collection efciency ( opt ) of the SEG-sample per angle of incidence according to Equation (5.2).
850

opt () =

photons out = photons in

400 2500 280

E out (, ) d A e = E in () d A f,proj

I out () A e I in A f cos()

(5.2)

Where E out is spectral intensity of the output at one edge [W m2 nm1 ], A e the area of one edge of the sample [m2 ], wavelength [nm], the angle of incidence [ ], E in spectral intensity of the light source [W m2 nm1 ], A f the front surface area of the sample [m2 ], I out radiant intensity emitted at one edge [Wm2 ] and I in irradiance of incoming radiation [Wm2 ]. Figure 5.13 shows angular efciency of SEG displayed as normalized optical collection efciency which is dened as the optical collection efciency at particular angles ( opt ()) divided by the optical col67

5. M ODELING S MART E NERGY G LASS

25
700 W/m 920 W/m 1170 W/m

20

Normalized optical collection efficiency

1,0

0,8

Radiant power (mW)

15

0,6

10

0,4

0,2
700 W/m 920 W/m 1170 W/m Curve fit

0 0 10 20 30 40 50 60 70 80 90

0,0 0 10 20 30 40 50 60 70 80 90

Angle of incidence

Angle of incidence

Figure 5.12: Radiant power as a function of angle of incidence.

Figure 5.13: Normalized optical collection efciency, or K , as a function of angle of incidence.

lection efciency at normal angle of incidence ( opt, ). The overlapping triangles in Figure 5.13 indicate that the optical collection efciency of SEG is independent of incident radiation intensity, which is inconsistent with the results for quantum dot LSCs in Gallagher et al. (2007). Since the results in Figure 5.13 compare efciency relative to normal efciency, it provides insights in relative losses and therefore acts as correction factor that accounts for the angular performance of LSCs. Correction factors of the same kind are frequently employed for modeling solar thermal collectors where they are described as incidence angle modiers (IAM) or (K ) (Dufe and Beckman, 2006). In this thesis, the same terminology is adopted. An empirical relation for this IAM was obtained by tting a curve through the measured datapoints. The shape of this curve shows similarities with the specular reection component of clear oat glass. Therefore, inspiration for candidate curve-ts was found in functions that describe angular resolved properties of windows (Rubin et al., 1998; Karlsson et al., 2001). In principle, it would have also been possible to model this behavior based on rst-principles by using Fresnel equations and Beer-Lamberts law. The following reasons justify the use of an empirical rather than the theoretical approach: It is unknown how luminescence effects inuence specular behavior of the sample. Spectral refractive index, spectral extinction coefcient and thickness of all substrate layers needs to be characterized accurately in the theoretical approach (Singh and Garg, 2010). Considering the required level of detail, an empirical approach is computationally the most effective. The tangent function in Equation (5.3) was found to be a simple yet effective method for representing the desired behavior. K = 1 tanx 2 (5.3)

By denition, these functions return 1 when is zero and 0 when radiation hits the sample in parallel direction. The curvature of the function is tuned by selecting the most appropriate value for x. The best curve t was obtained by calculating root mean square errors (R M SE ) as a function of x (Equation (5.4)), and subsequently selecting the value for x that minimizes this function. The best matching tangent function was found to have a value of x = 3, 5 and is plotted with the black line in Figure 5.13 1 7 (y (10n) y t (10n))2 8 n=0 exp

RM SE (x) =

(5.4)

When the results for this IAM correlation are now combined with the expression for calculation of optical collection efciency (Equation (5.2)), this results in the semi-empirical relationship for calculating 68

5.4. Simulation model optical output as a function of angle of incidence in Equation (5.5). Assuming that K is a characteristic property of SEG, no angular dependent measurements are required for characterizing the optical Af output of SEG. The shape effects are included via the geometric gain factor G = A e ; the efciency factor opt, accounts for the effects of different window states. I opt = I tot G cos() opt () = I tot G cos() K opt, (5.5)

Thus far, the development of the electrical model has only dealt with the optical collection of output emitted at the edges of the window relative to the incoming photon ux at the front surface. The next step is then to include the photovoltaic conversion process in the model. Photovoltaic conversion efciency ( pv ) includes both coupling losses at the PV interface and the quantum efciency of the PV cells (Currie et al., 2008), and is best calculated by a dedicated photovoltaic model. The electrical output of a SEG window (P el ) is obtained by multiplication of the optical output power (P opt ) with the photovoltaic conversion efciency (Equation (5.6)): P el = P opt pv = 4 I tot A f cos() K opt, pv (5.6)

Where the factor four in Equation (5.6) accounts for the difference between single-edge collection to full-perimeter conversion efciencies of square windows. The product I tot A f cos() represents the amount of incident solar radiation. The variable properties of changing SEGs window state are included in the model via opt, ; a factor that is characteristic for the specic geometry and window state. When (measured) electrical output of SEG is compared to incoming radiation, the expression of Equation (5.6) can be further simplied by using the total power conversion efciency ( pc ) of the window which is dened as: 4 K opt, pv , and results in the expression of Equation (5.7): P el = I tot A f pc (5.7)

5.4 Simulation model

5.4.1 Simulation strategy
The main objective of the simulation process is to quantify the overall performance of SEG. This allows the user to gain insight in the trade-offs between thermal and visual comfort on the one hand and associated energy demand on the other hand. This simulation requirement necessitates that thermal, optical and electrical models are considered simultaneously, and therefore conrms that a combination of stand-alone tools is inappropriate for the purpose. The BPS tools overview of Table 3.1 contains two programs that can cope with integrated simulations of the thermal, optical and electrical domains: ESP-r and EnergyPlus. The possibility to change the window properties in three states during runtime is identied as a pre-requisite for successful simulation of SEG. Chapter 3 draws the conclusion that TRNSYS appears to be the single tool that is able to aptly handle this adaptive behavior. Unfortunately, TRNSYS does not offer the possibility to predict the buildings daylighting performance (Table 5.2).
Table 5.2: Capabilities of tools to simulate SEG.

Adaptation ESP-r EnergyPlus TRNSYS

Thermal

Optical

Electrical

As a consequence, integrated simulation approaches are also inadequate to fulll the simulation objective, and the only remaining simulation strategy is co-simulation. The decision was made to use 69

5. M ODELING S MART E NERGY G LASS TRNSYS for performance prediction in the thermal and electrical domain, and to couple this model with the results of dynamic daylight simulations in DAYSIM. Daylight simulations for all window states are rst conducted in a preprocessing stage. This data is then supplied to TRNSYS that selects the right data during runt-time corresponding to the controlled window state. Data is owing across the three domains at every time-step and therefore data exchange is dynamic. The simulators in DAYSIM and TRNSYS are however invoked consecutively and therefore the workow is sequential (Figure 5.14). Zhai et al. (2002) have identied this simulation approach as virtual dynamic coupling (or bin coupling), originally developed as a way to reduce computational costs of detailed thermal-airow simulations. A schematic representation of this simulation strategy for SEG is given in Figure 5.15, and is discussed in more detail in the following sections.
Workflow
B11 B12

....

B1n B21 B22

....

B2n A1 D1 A2 D2

.... ....

An Dn

DAYSIM

TRNSYS

Figure 5.14: Workow of the simulation process, where A,B and D indicate the thermal, optical and electrical domain respectively.

5.4.2 DAYSIM
DAYSIM is a dynamic daylight simulation tool that uses the general accepted Radiance algorithms in combination with a daylight coefcient method and Perez sky model (Perez et al., 1990) to calculate annual time-series of light distribution in a room. The simulation process starts with specication of (i) scene and (ii) scenario for every window state. This includes modeling of external obstructions, room geometry, surface reectance and daylight openings. According to the conceptual model, the characteristics of SEG can be modeled by using and changing the visible transmittance (Tvis ) of the fenestration system in each window state. Thin glass surfaces like SEG are best represented by using the glass primitive (Larson and Shakespeare, 1998), which computes multiple internal reections without tracing rays; thus saving rendering time without compromising accuracy. In Radiance/DAYSIM, glass is characterized by transmissivity (t n ) at normal incidence. Therefore it is necessary to convert the measured visible transmittance derived by OPTICS to the theoretical value for transmissivity (Larson and Shakespeare, 1998). In DAYSIM, the scenario is specied by selecting an appropriate weather le. These les contain hourly values of solar irradiance and therefore do not acknowledge the ckle supply of daylight (Section 3.2). Shorter time-step data, upt to one minute can be obtained by using the model-based stochastic Skartveit-Olseth approach implemented by Walkenhorst et al. (2000). After the building scene and scenario is modeled, the simulation process continues with full-year daylight simulations for all window states . Raw luminance and illuminance data are computed at specied sensor points and stored in data-les on a time-step basis. Analysis of these results also yields
Although SEG can be switched in three states, the decision was made to only include the bright and dark state in the initial model while excluding the translucent state. This simplication was made for the following three reasons: (i) from the energy point of view, the effects of translucency are inferior to switching between the dark to bright state, (ii) the translucent state is the least mature technology with a currently only a mediocre haze factor, and (iii) the goniophotometer equipment required to accurately characterize the optical scattering properties was not at our disposal. Switching the window in three states was however part of the initial requirements of the model. The simulation model thus enjoys the benet of high exibility and can be easily extended with additional window state(s).

70

5.4. Simulation model the amount of supplementary articial lighting needed to sustain sufcient illuminance levels and the additional internal heat gains associated with this contribution. Another distinct quality of DAYSIM is its occupancy model that is based on Lightswitch-2002 (Reinhart, 2004). Since this model uses behavioral patters, it exhibits a relatively high level of detail and is consequently favored above the simple time-schedules in TRNSYS.

DAYSIM Window state TYPE 56 Electricity generation Internal heat gain & electricity consumption

Ti History Control

time state ill, lum Daylight data

Simulation time

One time-step

Window state TYPE 56 Electricity generation

PVmodel

Internal heat gain & electricity consumption

Ti History Control

time state ill, lum Daylight data

Window state TYPE 56 Electricity generation

PVmodel

Internal heat gain & electricity consumption

Ti

Figure 5.15: Simulation strategy for the SEG model.

5.4.3 TRNSYS
Thermal The inuence that SEG exerts on thermal performance of a zone or building is evaluated by using the TRNSYS multizone building model (TYPE 56). This dynamic model explicitly calculates many possible outputs including, air temperature, surface temperatures, heating and cooling power demand and thermal comfort indicators. In the simulations, only one building model is used that dynamically changes its window properties during run-time with a function called variable window ID. Environmental conditions are ensured to be identical to those subject to the daylight model by selecting the same weather le. Internal heat gains of the building depend on the state of the window since the amount of articial lighting changes with daylight availability. Together with occupants presence, this data is imported from the data les precalculated by DAYSIM. 71

5. M ODELING S MART E NERGY G LASS Electrical The electrical model calculates electrical output of SEG according to Equation (5.6). The amount of radiation incident on the window is calculated by the weather data processor (TYPE 15). This ux of energy is then multiplied in equation-TYPES by the incidence angle modier (K ) and the optical collection efciency ( opt, ) whose value is adaptive and depends on the window state as imposed by the control component. The remaining photovoltaic conversion process is calculated by using TRNSYS TYPE 180; the photovoltaic array with data le. This type is specically developed for characterizing photovoltaic output and is based the equivalent one-diode model (Wenham et al., 2007). The parameters of this model include short-circuit current (I sc ), open-circuit voltage (Voc ), and voltage (Vmpp ), current (I mpp ), and power (P max ) at the maximum power point (MPP), all measured at standard test conditions (STC). In addition, the short circuit current temperature coefcient (I,sc ) and opencircuit voltage temperature coefcient (V,oc ) are required. This data is typically available in the product information provided by the manufacturer of the PV cells. Control The control component plays a central role in SEGs simulation model as indicated in Figure 5.15. The crude output les from DAYSIM are rst rearranged in a spreadsheet tool and then imported in TRNSYS via free format data readers in expert mode (TYPE 9e). Every time-step, this data is accessed by the control component which decides upon the right adaptive actions on the basis of an imposed control strategy. This control logic is implemented via equations containing conditional statements that compare model output (e.g. temperature, window luminance, workplane illuminance) to target values and return window ID for the next time-step as output. This window ID is passed on to the thermal and electrical model and used in the respective calculations for the next time-step. TRNSYS assembles and manipulates the controlled state variables and stores them in output les. Once the simulation is nished, these results are used to calculate the desired performance indicators of SEG for the given operational strategy.

72

CHAPTER
6.1 Introduction
73

S MART E NERGY G LASS MODEL VALIDATION

Performance predictions in support of decision making is identied as a controversial use of simulation (Sokolowski and Banks, 2009). This is because validity of predictions and associated decisions entirely rests on the assumption that the simulation model is valid too. In this chapter, the validity of the Smart Energy Glass (SEG) model is tested on the basis of an empirical validation study. It is not the intention to question or demonstrate the validity of the underlying algorithms of the software tools. Instead, the specic assumptions regarding the SEG model will be evaluated.

Up till now, SEG has only been tested as small samples under the controlled conditions of a laboratory environment. Doing experiments on a larger scale, in windows exposed to atmospheric conditions, is however indispensable for gaining more insight in thermal, optical and electrical performance, and the potential advantages of SEG. Furthermore, this approach can also reveal potential practical problems (e.g. scaling effects, homogeneity problems, performance degradation) related to the integration of SEG in a real faade, that will not show up in laboratory settings. To be able to as well evaluate the performance of SEG under conditions outside those experienced during the measurements, one can take advantage of the qualities of BPS. These models are however based on a number assumptions, simplications and uncertainties. In order to use the model with sufcient condence, one needs to assure that the model results are reliable. One way to check the validity of BPS-models is by comparing simulated results with the outcomes of an experiment; a process called empirical validation (Judkoff and Neymark, 1995). Although it is the rst time SEG is being tested under real-world conditions, it is no novelty that the effects of switchable glazing are investigated under atmospheric conditions. Several research projects reported on the use of experiments for determining the performance of electrochromic glazing. These tests were carried out in different settings, ranging from (i) reduced-scale test boxes with small window samples (Piccolo et al., 2009; Piccolo, 2010), and (ii) highly standardized and instrumented test-cells for thermal and solar building research (PASSYS) (Aleo et al., 2001; Assimakopoulos et al., 2007), to (iii) fullscale windows in ofce buildings with occupants as subjects (Platzer, 2003; Lee et al., 2006; Zinzi, 2006). A selection of methodological aspects in these studies serves as reference for setting up the experiments in this project.
Including environmental boundary conditions, orientation, occupant behavior, design variables, material properties, etc.

6. S MART E NERGY G LASS MODEL VALIDATION

6.2 Methodology
Empirical validation studies in the light of BPS essentially consist of the following three steps (Judkoff, 2008; Strachan, 2008): 1. Monitoring the boundary conditions and the performance of a test object in an adequate experimental setup; 2. Creating a simulation model of the experimental setup and test environment, and undertaking simulations using the measured climatic data; 3. Comparing predicted performance with measured performance and deciding upon validity of the model. If the third step turns out to be successful (i.e. the results are within certain bounds), it proves that the simulation program can correctly model the components characteristics when that component is subject to arbitrary dynamic outdoor conditions lying within the tested range. In addition, the validation process then also gives sufcient condence that the model reliably predicts performance in new situations, provided that the extrapolated conditions are selected with care. Empirical validation however encompasses more than simply comparing predictions with empirical data for the binary decision whether the model is right or wrong. The procedure also provides an opportunity for new insights and/or enhanced understanding of the system, and this new knowledge can be used for amelioration of the initial model and advancements of the product design (Palomo Del Barrio and Guyon, 2004). The following two subsections describe the methodology that was applied in the rst two steps of the empirical validation study. Subsequently, Section 6.3 discusses the comparisons between measurements an simulations. The focus in this validation study is on the daylight model and electrical model. Rigorous studies have already demonstrated that solar energy gains through glazing systems are accurately modeled in TRNSYS (Loutzenhiser et al., 2006; Judkoff, 2008). Replicating these experiments for SEG would be a cost- and time intensive undertaking that requires an extensive set of measurements and a large degree of control on disturbances. Because thermal behavior of SEG resembles that of conventional windows reasonably well, it is assumed that the decision to rely TRNSYS algorithms for thermal performance predictions of SEG is justied.

6.2.1 Experimental setup

All measurements took place at the test faade facility of the Vertigo building on the campus of Eindhoven University of Technology in the Netherlands (Figure 6.1). The orientation of this faade is west, no signicant obstructions that block direct solar radiation are present.

Figure 6.1: Overview of the test-faade with the SEG-prototype installed.

74

6.2. Methodology The performance of SEG is tested on the basis of a working prototype. In this custom-made window (Figure 6.4(a)), the SEG-coating is sandwiched between two layers of 3 mm clear Guardian glass to form the internal pane of the double glazing unit. The external pane is formed by a 6 mm clear glazing with UV-lter. This UV-blocking layer was selected to protect the SEG-coating and to increase its durability. In SEGs envisioned nal product design, the UV-protection layer will be mixed in as part of the glass sandwich. Consequently, the SEG layer can move to the external pane which directly has benecial effects for the solar heat gains and internal window layer temperature in summertime.

1 2

8 7

54 cm

35 cm

3
40 cm 72 cm
108 cm

6 4 5

Figure 6.2: Dimensions of the test-box, the window is oriented to the west.

Figure 6.3: Segments of PV cells for electrical measurements, view from the inside.

Given the size of the available SEG-prototype (35 x 40 cm2 ), the decision was made to install a test-box, to mimic the effects of a zone behind the faade. The dimensions of the breboard box are given in Figure 6.2 and correspond to a typical full-scale equivalent ofce room, scaled with a factor 1 : 5. In order to asses SEGs daylighting performance correctly, it is important that the internal surfaces of the box (walls, oor, ceiling) are painted in representative colors (Bodart and Deneyer, 2006). The oor of the test-box was painted grey, all other interior surfaces were colored with matt white paint, pursuant to the monitoring protocol developed in IEA SHC Task 21 / IEA ECBCS Annex 29 (Velds and Christoffersen, 2001). Light distribution in the test-box was measured via an array of photocell illuminance sensors, all calibrated in advance with a reference cell in integrating sphere measurements. The arrangement of these sensors inside the box is given in Figure 6.4 (c).

(a)

(b)

(c)

Figure 6.4: Photos of the experimental setup, showing (a) fabrication of the prototype, (b) the window installed in the faade, and (c) the test-box with illuminance sensors.

Electrical performance of SEG was evaluated by monitoring the output voltage of the PV cells. For this purpose, the electrical circuit of PV cells was subdivided in eight segments, as illustrated in Figure 6.3. To 75

6. S MART E NERGY G LASS MODEL VALIDATION demonstrate the energy generating potential, electrical wires were uncoupled from the window frame (Figure 6.4 (b)) and each circuit was connected to a xed load with a resistance of 980 . The voltage across each resistance was measured and recorded in time-steps of one minute. The last essential part of the experiments is the monitoring of occurring operational conditions. The state of the window (dark, bright or translucent) was manually controlled via a remote control system. Actual switching was accomplished via a Moeller Xcomfort wireless control switch. This signal was also sent to the data logger and continuously monitored during the entire measurement period. Environmental boundary conditions were monitored every two minutes with a Vaisala WXT510 weather station that measures wind speed and direction, liquid precipitation, barometric pressure, air temperature and relative humidity. Direct and diffuse solar irradiance was additionally measured via a cm10 pyranometer.

6.2.2 Simulations
Building performance simulations were carried out with the integrated thermal-optical-electrical simulation models as detailed in Chapter 5. The model was executed in TRNSYS v16 in combination with daylight results from DAYSIM v2.1. In both tools, building models were created following the dimensions of the test-box given in Figure 6.2, without any external obstructions. The internal surfaces of the test-box are treated like perfect Lambertian surfaces with reectance = 0.3 for the oor and = 0.8 for walls and ceiling. SEG is modeled as windows and the respective material properties for the bright and dark state are given in Table B.1. Simulations in DAYSIM were carried out with a time-step of two minutes, while in TRNSYS a time-step of ten minutes was selected. The appropriate measured boundary conditions were supplied to the models, and representative days were selected for comparisons of the results.

6.3 Comparing results

The ultimate or absolute validation standard would be a comparison of simulation results with a perfectly performed empirical experiment, the inputs for which are perfectly specied to the simulationists (Neymark et al., 2002). Obviously, this ideal situation does not exist in practice. Multiple potential error sources are simultaneously operative which makes interpretation of the results of validation exercises a challenging task. A mismatch between measured and simulated output does not necessarily imply that the assumptions of the simulation model is wrong. The following subsections present the comparison of results and judgement about the models validity. This starts with the daylight model in Section 6.3.1, and then discusses the validation outcomes of the electrical model in Section 6.3.2.

6.3.1 Daylight model

Figure 6.5 shows the comparison between measurements and simulation results for the three sensor positions closest to the window. The results present interior illuminance values on a sunny day when the window was switched in the bright state. The comparisons in Figure 6.5 show generally good agreement between results of measurements and simulations, with a maximum overprediction at the peak of approximately 15 % . The results can be used to test the assumption that SEG is reliably modeled by using DAYSIMs glass primitive and visible transmittance (Tvis ) of the fenestration system. The daylight coefcient approach in DAYSIM is based on the principle of superposition which makes that Tvis correlates to the magnitude of illuminance
An error of 15 % may seem unduly high. It should however be noted that the human eye is a logarithmic sensor, which makes that the accuracy of the simulation model corresponds to the sensitivity of illuminance our eyes can detect (Reinhart, 2010).

76

6.3. Comparing results

20000

Sensor A
16000

Sensor B Sensor C

Illuminance (lux)

12000

8000

4000

0 0 4 8 12 Time (h) 16 20 24

Figure 6.5: Comparison between measurements and simulations on a sunny day, with: the measured illuminance (solid lines) and simulated illuminance (dashed lines) for three different sensor points.

levels. The observed errors in the results compare well to the results in e.g. Reinhart and Walkenhorst (2001) and Reinhart and Breton (2009), and are only slightly higher than the condence bounds as specied by Mardaljevic (2000). The differences in measured illuminance among the sensor points A, B and C are prompted by their respective positions in the test-box. The window is mounted in the west faade, which makes Sensor B the rst to receive direct radiation. Sensor C receives direct radiation later in the afternoon, and at this particular day, sensor A does not receive any direct radiation at all. The results of the simulations tend to predict the same pattern, however, the starting time of the sudden increase for sensors B and C shows a considerable offset. Figure 6.6 introduces the most important error sources in daylight performance predictions.

Figure 6.6: Main causes of error in daylighting performance assessment (Thanachareonkit, 2008).

Discrepancies between measurements and predictions stem from the compound error of multiple error sources and cannot be traced back to model photometry only. The observed time shift in Figure 6.5 depends on the moment when the sensor starts to observe direct radiation. Therefore, the discrepancy 77

6. S MART E NERGY G LASS MODEL VALIDATION is believed to be attributable to model geometry rather than model photometry. Indeed, this seems a plausible explanation, since window sills and mullions, and the shading obstructions of the test setup were not included in the building model. This simplication was corrected in the results of Figure 6.7 where the building geometry was set up in more detail and compared to the measurements and the results of the original model.
20000

Measured 16000 Original model Detailed model

Illuminance (lux)

12000

8000

4000

0 0 4 8 12 Time (h) 16 20 24

Figure 6.7: Comparison between measurements and simulations with the original and detailed model for sensor position A.

When shading obstructions are included in the model, the overall residuals are reduced. Although the time shift is only of minor importance for visual comfort levels, it does represent a signicant contribution of daily solar gains. Therefore it is concluded that modeling building geometry with sufcient accuracy deserves special attention in the simulation process. Currently, no general accepted standard or reference values exists for evaluation of empirical validation studies related to daylight models (Reinhart and Breton, 2009). Errors in the presented results do however compare well to other prominent validation studies. Consequently, the conclusion is made that the simulation approach in DAYSIM is condentially used for assessing SEGs daylighting performance in the bright state, provided that the model geometry is included at an appropriate level of detail. The outcomes of SEG in the dark state showed error bounds comparable to the bright state, which is a direct consequence of the superposition principle in DAYSIMs daylight coefcient method. In addition, DAYSIM can also be used with condence for performance prediction of SEG in in the translucent privacy state. Reinhart and Andersen (2006) found that DAYSIM is fully capable of simulating the short time step dynamics of indoor illuminances due to daylight for a design that features translucent panels. For this purpose, the bidirectional light distribution function of SEG rst needs to be established in transmission (BTDF) and reection (BRDF) measurements.

6.3.2 Electrical model

Figure 6.8 shows the comparison between results from measurements and simulations for a day with full overcast sky. It can be observed that the measured electrical output in all segments (see Figure 6.3) is proportional to the intensity of incoming solar radiation and that the results of simulation tend to follow the same pattern. The measured output features a high degree of variability that resembles the erratic supply of sunlight. In the simulations however, a timestep of 10 minutes was used which makes that the output is averaged and the variability is attenuated. Figure 6.8 further shows that the PV cells still 78

6.3. Comparing results

0,8
PV1 [V] PV3 [V] PV5 [V] PV7 [V] Simulation [V] PV2 [V] PV4 [V] PV6 [V] PV8 [V] Solar radiation [W/m]

80

0,7

70

0,6

60

Electrical output (V)

0,4

40

0,3

30

0,2

20

0,1

10

0 7 9 11 13 Time (h) 15 17 19

Figure 6.8: Comparison between measurements and simulations on a day with overcast sky, with: the measured electrical output (PV1 to PV8) in [V]; measured global solar irradiance on the faade in [W/m2 ]; and simulated electrical output in [V].

produce electricity around sunset whereas the model predicts zero output. This discrepancy is probably caused by the fact that twilight and reected radiation from surrounding objects to the window is not explicitly included in the model. Another difference is introduced by the rectangular dimensions of the SEG-prototype (35 x 40 cm2 ), where the model assumes square window shapes. The results of measurements for a sunny day are given in Figure 6.9. Output on a sunny day apparently differs one order of magnitude compared to cloudy days. The results in Figure 6.9 also show that electrical output on a sunny day is not the same for all edges of the window, but instead shows substantial directional effects. The model as developed in Chapter 5 is based on the waveguiding effect only, and therefore assumes equal electrical output at all four sides of the window. This assumption introduces signicant underprediction of electrical output that is not justiably neglected and therefore deems the initial model unacceptable. To overcome this deciency, the electrical model was extended with a component that accounts for the radiation that falls directly onto the PV cells. The modied expression for electrical output is given in Equation (6.1)

P el

= =

P el 4 I tot A f cos() K opt, pv

+ +

P direct
4 i =1

(I dir,i I re,i ) A pv,i pv

Irradiance (W/m)

0,5

50

(6.1)

Where P el is equal to Equation (5.6), I dir is direct beam irradiance hitting the PV cell [Wm2 ], I re is reected radiation at the interface [Wm2 ], A pv,i is area of the PV cell [m2 ] and i represents the position of the PV cell in the frame. Since this new component only considers direct beam radiation, obstruction elements are used in the model to determine at what time of the day the additional radiation strikes the PV cells. The expression includes a summation for the four edges, because the position that the PV cell occupies in the window 79

6. S MART E NERGY G LASS MODEL VALIDATION

4 top [V] bottom [V] 3 Measured electrical output (V) solar radiation [W/m] left [V] right [V]

800

3,5

700

600

400

1,5

300

200

0,5

100

0 7 9 11 13 Time (h) 15 17 19

Figure 6.9: Measured electrical output voltage on a sunny day in [V ] for the four edges of the window, together with measured solar irradiance on the faade in [Wm2 ] (green line).

frame and thus its orientation and shading determines the amount of received radiation. This effect is modeled in TRNSYS by obstructing the respective views to the sky via TYPE 34 shading components (overhang and wingwall shading). Finally, the reection component is introduced in the model because a part of the radiation that would have hit the bare PV cells is lost in advance due to specular reection at the external SEG window pane.

Figure 6.10 again shows the results for the same sunny day, but now in comparison with the simulation results of the modied electrical model. For pragmatic reasons, it was not possible to assess these outcomes on the basis of absolute comparisons: In the experiments, the PV cells were connected to xed electrical resistances that choose an arbitrary working point on the IV-curve of the PV cell, depending on available photocurrent. Because of the high variability in supply that is introduced by direct radiation, the PV cells often operate in suboptimal working points. The modeled PV cells on the other hand are equipped with MPP-trackers, and therefore they do always operate in maximum power point (MPP). This disparity between experiment and model prohibits fair comparisons on the basis of output voltage only, and also prevents useful conversion from one quantity to another via Ohms law. Nevertheless, the results still do contain valuable information that is relevant for assessing validity of the modied electrical model.

The PV cells facing downwards (top) and facing the north (left) do not receive any direct beam radiation and therefore the simulation results for these PV arrays, as displayed in Figure 6.10, are identical and similar to results of the initial model. It can be observed that the shapes of the simulated curves compare reasonably well to those measured. In addition, the relative magnitude of increments in output between the segments is also represented in the model. Given the relatively large amount of sources of uncertainty (e.g. temperature dependency, reected radiation from the environment, losses in electrical circuits), the model predicts the actual electrical output with fairly good similarity. Therefore, the conclusion is made that the model is valid within the boundaries and purposes of this study. 80

Irradiance (W/m)

2,5

500

6.4. Conclusion

3 top [V] 2,5 bottom [V] top/left [W] 2 bottom [W] Voltage (V) 1,5 right [V] right [W] left [V]

0,225

0,150 Power (W) 0,075 0,000 7 9 11 13 Time (h) 15 17 19

0,5

Figure 6.10: Comparison between measurements and simulations on a sunny day, with: the measured electrical output voltage for the four edges in [V] (solid lines) and corresponding simulated electrical output power in [W] (dashed lines).

6.4 Conclusion
This chapter presented the ndings of the empirical validation study that was carried out to check the validity of the optical and electrical SEG models. Errors predicted by the daylight model were found to be commensurable with the results of other validation studies, and consequently deemed acceptable. The electrical model however, needed to be extended with an additional component that models the radiation directly falling onto the PV cells, in order to yield accurate results. On the basis of this study, it can be concluded that the results of the integrated SEG model are accurate enough to be used for comparisons of alternative design variants and for investigations into the effects of different control strategies. The results of the simulation study that was carried out with this validated model are presented in Chapter 7.

81

CHAPTER
7.1 Methodology
7.1.1 Model setup
83

P ERFORMANCE SIMULATION OF S MART E NERGY G LASS

In the third step of the case study, performance simulations of smart energy glass (SEG) are carried out. The results of this study are primarily used for quantifying this integrated performance and for pointing to future directions of product development. In the second place, the simulation process is also useful for clarication of practical points of attention associated with performance simulation of CABS in general. This chapter rst introduces the methodology for this application study in Section 7.1. Subsequently, the results of three examined cases are presented. Finally, this chapter concludes with a discussion that places the specics of the simulation results in a broader perspective.

The performance of SEG is assessed on the basis of the simulation strategy as developed in Section 5.4.1 and illustrated in Figure 5.15. The modied expression for electrical output (Equation (6.1)) was used to predict electrical output. Three different cases are examined in this chapter: Energy efcient ofce, Advanced renovation and SEG in a continental climate. This section discusses the common characteristics for these simulations, and introduces the performance indicators used in the analysis. Specic assumptions and motivations for each case are presented together with the results in Sections 7.2 to 7.4 respectively.

Smart energy glass is a coating that can be applied to almost any glazing system. The technology is still under development and the performance of the dye continues to be improved. Developing the specications of SEG has to satisfy multiple and competitive requirements, including: aesthetics, efciency, bandwidth, UV-stability and homogeneity. In this evaluation, the properties of SEG were selected to be a fair projection of what is possible in the foreseeable future. The switchable SEG-layer is assumed to form the external pane of the double glazing system . Thermal window performance is further enhanced with a low emissivity (low-E) coating and argon lled cavity to meet current building energy standards. Earlier simulation studies with electrochromic windows indicated that selection of the reference case is of essential importance for assessing feasibility of the entire concept (Lee and Tavil, 2007). Comparisons to invalid reference values can disguise the true potential of the technology and is on the other
This contrasts with the composition of the SEG prototype used in the validation study of Chapter 6, where SEG formed the internal window pane because of UV-stability reasons.

7. P ERFORMANCE SIMULATION OF S MART E NERGY G LASS hand misused to manipulate stakeholders opinions. In this study, the SenterNovem reference ofce building (SenterNovem, 2009) is used for baseline comparisons, because this reference case is specically developed for obtaining rational indications of the energy saving potential of innovative building technologies. The analysis in this study is restricted to a two-person south-facing perimeter ofce zone, situated at an intermediate oor of the reference building. The ofce space is assumed to be surrounded by identical ofce spaces. These adiabatic boundary conditions were selected to ascertain that observed performance differences are effectively attributable to SEG. Storage of thermal energy in internal partitions is taken into account, and typical ofce equipment amounts to a heat load of 10 W/m2 . The dimensions and geometry of the zone are given in Figure 7.1. The window-to-wall ratio of the south faade equals 35 %. The opaque faade elements are well-insulated with an R c -value of 3.5 m2 K/W.

2,7 m

1,4 m

1,25 m
m 5,4

3,6 m

Figure 7.1: Schematic representation of the ofce room used in performance simulations.

Occupancy in the ofce room follows DAYSIMs probability-based ve day workweek-schedule with intermediate and lunch breaks. Articial lighting (15 W/m2 installed power) switches according to this same schedule and is continuously dimmable up to an indoor illuminance of 500 lx on the basis of the LIGHTSWITCH-2002 algorithms (Reinhart, 2004). Lighting is controlled based on a work plane photosensor at a distance of 1.8 m from the envelope. Thermal conditions in the zone are controlled on the basis of indoor air temperature, with setpoints for heating (20 C) and cooling (24 C) between 8 a.m. and 17 p.m., and a heating setpoint of 16 C outside working hours. Identical weather les in EPW-format are supplied to both DAYSIM and TRNSYS. Stochastically generated short time-step (ve minute) solar irradiance data les (Walkenhorst et al., 2000) are created once for every case, and are used for predicting daylighting performance. The number of possible strategies for controlling SEGs adaptive behavior is virtually innite. The aim of this chapter is to explore their potential and provide some rst insights in the cause-and-effect relationships of various options. It was not intended to seek for the best possible strategy nor to account for the practical implementation aspects of these algorithms. Table 7.1 provides an overview of control strategies that were studied.

7.1.2 Performance indicators

The underlying motive behind this application study is to gain insight in the interplay between thermal and visual comfort and associated impacts on energy demand for heating, cooling and lighting. For this purpose, it is important to establish the performance indicators (PIs) that form the basis for comparisons beforehand. Values for annual heating, cooling and lighting energy demand are obtained after 84

7.1. Methodology

Table 7.1: Overview of the six investigated control strategies.

A B C D E F

Reference case SEG always switched in the bright state SEG always switched in the dark state SEG switched to the dark state when indoor air temperature 21 C SEG switched to the dark state when daylight illuminance on work plane (E h ) 700 lx SEG switched to the dark state when window luminance (L v ) 1500 cd/m2

summation of the momentary power that is required to satisfy setpoints on a time-step basis. These values are converted to the same base unit (kWh) because this allows for (i) calculation of overall annual energy demand and (ii) insights in the relative importance of individual contributions. Apart from integrated energy consumption, peak power demand is also considered as important PI. Peak capacities not only determine investment costs, but also inuence operating system efciencies, and dimensions of equipment and distribution systems. Both overall annual energy demand and peak power should be aimed to be as low as possible. This goals however needs to be achieved without making compromises on performance in terms of the comfort requirements as discussed below. Minimum comfort levels with SEG are guaranteed by proper sizing and control of HVAC and articial lighting equipment. Too cold thermal environments are of minor concern in well-insulated ofce buildings, because internal heat gains contribute to quick warm-up times. The same is true for minimum illuminance levels, because lighting systems are dimensioned to deliver adequate visual performance under dark outside conditions. The crux of controlling SEGs performance is to counteract discomfort due to an oversupply of solar gains without the penalty of expending too much energy. In this thesis, thermal comfort is assessed on the basis of overheating risk. This is accomplished by counting the number of hours that indoor air temperature exceeds 25 C. Allowing discomfort during maximum 5 % of working hours is usually seen as realistic and economic target value in the trade-off between energy and comfort. As a result, this amounts to an allowed number of 100 overheating hours. Visual comfort is evaluated by considering the risk of glare, which is dened as the sensation produced by luminance within the visual eld that is sufciently greater than the luminance to which the eyes are adapted to cause annoyance, discomfort or loss in visual performance and visibility (IESNA, 2000). Multiple assessment methods and performance indices are in use to evaluate the occurrence of glare (Bellia et al., 2008). The current DAYSIM distribution does not come with a module for direct assessment of the risk of glare . The tool is however capable of calculating surface luminance values, and these can be transformed into the desired performance indicator. Occurrence of glare is found to be affected by many interrelated factors, including: absolute luminance of the glare source and surrounding surfaces, transient adaptation of the eye, surface reectances, type of work (screen, paper), position in the room, etc. In this study, the risk of discomfort caused by glare is assessed by counting the number of times when the ratio between window luminance (L v ) and paper task luminance (L h ) is higher than 10:1 (IESNA, 2000). Occurrence of glare is further extended with the additional clause that L v 2000 cd/m2 , in order to exclude from analysis cases with high contrast but low brightness. Levels of maximum allowed visual discomfort typically exist in terms of design recommendations rather than xed standards. All these guidelines agree that occurrence of glare should be as low as possible; the trade-offs in relation to other PIs should rationalize for each situation whether this strive is justiable.

An extension with dynamic daylight glare probability (DGP) functions based on glare rating classication and frequency distributions (Wienold, 2009) is planned for the next version of DAYSIM.

85

7. P ERFORMANCE SIMULATION OF S MART E NERGY G LASS

7.2 Case 1 - Energy efcient ofce

Product development of SEG initially focuses on commercial buildings, although applications in e.g. residential buildings and automotive settings are also possible. With their large casual gains and high comfort standards, commercial buildings are seen as the primary segment for early market introduction. In this rst case, the performance of SEG is investigated under the climatic conditions of the Netherlands. Dutch ofces typically feature equivalent heating and cooling demand, therefore energy savings by controlling admission of daylight are expected in both summer and winter. This potential is further illustrated in Figure 7.2 that shows a temporal map of annual distribution of solar irradiance in Amsterdam. This graph shows high variability in supply of sunlight, hence it is expected that adaptivity in window transparency can be valuable throughout the whole year.
W/m2
24 1200

1080

960 18 840

Time of the Day

720

12

600

480

360 6 240

120

0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Day of the Year

Figure 7.2: Temporal map of direct horizontal solar irradiance in the Netherlands (NASA, 2010).

In each of three application studies, the building with SEG is compared to a reference. Here, the reference case (case A) assumes equal window areas as is the case for SEG (Figure 7.1), but is now equipped with solar control glazing. The presence of a metallic reective coating in these windows offers a moderate but xed reectance in the visible, and high reectance in the near infrared wavelength area. Application of such windows reduces cooling load because solar heat gains are blocked while views to the outside are preserved. In addition, shading systems can be omitted in buildings with solar control glazing. A comparison of fenestration product properties between the reference situation and SEG is given in Table B.2; the simulation results are presented in Figure 7.3. The overall annual energy demand for the reference case is only 1236 kWh or 64 kWh/m2 . These values are comparable to target values for low-energy ofce buildings in most West European countries (Jensen et al., 2009). The energy efcient reference case is realized with an acceptable number of overheating hours, the occurrence of glare on the other hand appears to be unacceptable. In practice this would unavoidably lead to application of screens or other types of brightness control. Consequently, the results of this study contradict with advertisements for solar control glazing arguing that these devices are not necessary. Beside higher investment and maintenance costs, this would also inuence views to the outside and daylight utilization and thus move away from the optimal results as presented in Figure 7.3(A). Considering SEG as design alternative can provide an adequate solution for fullling both thermal and visual comfort requirements. Figure 7.3 however also shows the natural consequence that application of SEG has not so much to gain in terms of annual energy demand. If continuously in the bright state 86

7.3. Case 2 - Advanced renovation

Energy 1500
Lighting Cooling Heating

Comfort
Overheating Glare

500

1200 320

408

400 609 513 364 400 420 300

900 151 246 250 246 245

249 600 182 765 300 104 557 101

200

153 595 102 562 102 566 566 100 99

99 58

20

20

Figure 7.3: Comparison between energy and comfort performance of the reference case (A) and SEG (case B to F) for a new ofce building in the Netherlands. With on the left axis: heating, cooling and lighting energy consumption [kWh], and on the right axis: risk of overheating [h] and glare [times].

(case B), energy consumption gets reduced, but glare is still a problem. When always switched to the dark state (case C), the occurrence of glare drops drastically, but at the same time this also introduces an undesirable higher energy demand for articial lighting. Implementation of an appropriate control strategy (e.g. case D to F) is the key that allows for proting from the benets of both the dark and the bright state. As expected, occurrence of glare is counteracted when SEG is switched to the dark state at times when supply of solar gains is high. At the same time, this relieves cooling load since the transmitted part of solar radiation is reduced. However, increasing visual comfort in terms of glare is nearly always adversely compromised by an increase in electricity demand because daylight illumination is rmly reduced as well. Case D demonstrates that it may even be detrimental to switch to the dark state when the indoor air temperature is high. This is because the control strategy sometimes leads to situations with simultaneous demand for cooling and high demand of articial lighting, where the latter then directly adds to the cooling load of the zone. Better results are obtained when window state is controlled based on stimuli from the luminous environment (case E and F). In these cases, daylight is only allowed when desired and blocked when unwanted. It can be observed that electricity generation potential of SEG is not included in the results. This is because the produced amount of electricity is only marginal compared to the scale of hundreds of kWh in Figure 7.3. Currently, the focus is on making adaptivity of the window self-sufcient, which appears to be a realistic ambition. Treating SEG as decentral power supply source to feed back to the grid or to be consumed locally will be considered in future generations of SEG. The model as developed in the previous chapters provides the basis for computational analyses of this potential.

7.3 Case 2 - Advanced renovation

With an annual construction renewal rate of only 1-2 %, it is obvious that energy targets in the built environment cannot be met by considering new buildings only (WBCSD, 2009). Faades have been identied to offer greatest potential for pursuing sustainable renovation projects in the existing building stock (Kuhn, 2010). In this light, the prospects of SEG are especially promising because it is a self-sufcient device that can be installed according to the plug-and-play principle. This allows for introduction of the 87

Overheating (hours per year), Glare (times per year)

Energy demand (kWh)

7. P ERFORMANCE SIMULATION OF S MART E NERGY G LASS positive aspects of adaptivity to the building shell, in a cost-effective way, without the need for major overhauls. Appraisals pertaining to the renovation potential of SEG requires comparisons to be made against a different reference situation. The performance assessment in this second case assumes the same building geometry and orientation as Case 1, but other building characteristics are taken to be representative for the 1970s. For the window, this is manifested via conventional double glazing with the properties given in Table B.3. Insulation levels of opaque construction elements were also revised to have a R c -value of 0.67 m2 K/W, a typical value for ofce buildings constructed around 1975 (Petersdorff et al., 2006). Solar shading and brightness control in the reference case is achieved via manually controlled internal venetian blinds. Operation of blinds in the simulations is controlled in DAYSIM on the basis of the Active users prole. This stochastic algorithm assumes that blind settings are rearranged on a regular basis with the aim of maximizing daylight availability while also trying to exclude glare (Reinhart, 2004).
2000 707

Energy
Lighting Cooling Heating

Comfort
Overheating Glare

750

1600

278 609

600

Energy demand (kWh)

499 1200 429 364

437

420 450

193 800

191

193

191

191 300

989 400 149 100 873 182

914

878 159 100 20 102

885

883

150

98 21

98 20 0

Figure 7.4: Comparison between energy and comfort performance of the reference (A) and SEG (case B to F) in an advanced renovation case. With on the left axis: heating, cooling and lighting energy consumption [kWh], and on the right axis: risk of overheating [h] and glare [times].

The simulation results in Figure 7.4 suggest that cooling energy demand after window replacement with SEG is cut by more than a factor two. In addition, installed cooling power capacity was found to be safely reduced with more than 30 % (from 1.0 kW to 0.67 kW) when SEG is installed, while still achieving sufcient comfort levels. Figure 7.4 further shows that heating energy demand for the basecase compares well to that for SEG, even though the window U -value of SEG is much lower as a result of the presence of a low-E coating. Closer inspection at the energy balance reveals that the transmission losses after renovation are indeed lower, but this difference is almost compensated by the decrease in valuable passive solar gains. The lower visible transmission values of SEG also give rise to a relatively large electricity demand for lighting . The combined effects of high heating and lighting energy demand make savings in overall energy demand in this particular case marginal. The actual renovation potential of SEG however is likely to be more differentiated, because of the following reasons:
Luminous efcacy of the light sources is assumed identical in both cases. A more conservative reference situation would result in higher energy saving potential.

88

Overheating (hours per year), Glare (times per year)

7.4. Case 3 - SEG in a continental climate Any attempt to model occupant control of shading devices adds to the uncertainty in outcomes of the simulation. The reference situation in this case assumes the optimistic setting of active blind control. Blinds are opened and closed as to maximize daylight utilization and thus represent the best case scenario for manual blind control. In practice, a much more passive operation of blinds, and thus degradation in performance is likely to be expected (Daum and Morel, 2009). SEG on the other hand is easily integrated in the building automation system (BAS), which makes that difference are likely to become more clear. All cases assume continuous dimming control, where daylight levels are almost perfectly supplemented by articial lighting. In both new and old buildings, this might be a too idealistic representation of reality. When internal heat gains are considered to be more independent of ingress of daylight, the energy saving potential of SEGs adaptivity is supposed to be higher. More aspects than only operational energy demand and comfort levels play a role in the decision for renovation or preservation. A distinct advantage of SEG is the fact that shading devices are no longer required and therefore the access to outdoor is views preserved and maintenance costs reduced. In addition, renovation with SEG prevents early demolition, and thus increases economized exploitation of the building and extends resource efciency of the buildings life cycle.

7.4 Case 3 - SEG in a continental climate

It is not only control strategy and building design parameters, but also prevailing weather conditions that determines to a large extent the performance of switchable glazings (Jonsson and Roos, 2010). In both aforementioned cases, buildings with SEG were exposed to weather conditions of the Netherlands. The Dutch maritime climate is characterized by its moderately cool summers and comparatively warm winters, and consequently has a narrow annual range of temperatures. The assumption is that SEGs performance becomes more pronounced in regions with more divergent wintersummer climatic features. Testing of this hypothesis by replication in other climates is relatively straightforward in BPS as opposed to real-world experiments and happens by simply selecting a different climate le. In this third case, the continental climate of Chicago, Illinois (OHare International Airport) was selected, because it is known for its large annual range of temperatures.
2500

Energy
Lighting Cooling Heating

Comfort
Overheating Glare

500

553 2000 295 396 201 425 400 300 299

1500

577

684

696

300 697 690 690

1000 201 180

200

1172 500

1055 103

1114 127

1058 127

1063

109 64

1061

110

100

53 25 0

25 0

Figure 7.5: Comparison between energy and comfort performance of the reference case (Case A) and SEG (Case B to F) in a continental climate. With on the left axis: heating, cooling and lighting energy consumption [kWh], and on the right axis: risk of overheating [h] and glare [times].

Overheating (hours per year), Glare (times per year)

Energy demand (kWh)

89

7. P ERFORMANCE SIMULATION OF S MART E NERGY G LASS With the exception of climate le, all other building model parameters are taken identical to the modern ofce space of case 1 in Section 7.2. In this case the zone is subject to more extreme summer and winter conditions, therefore, the capacity of heating and cooling system was adjusted accordingly in order to meet thermal comfort standards. The results of simulations are presented in Figure 7.5. The results in Figure 7.5 show the same tendency that energy efciency of the reference case can hardly be surpassed by SEG. The results again do show that in this particular application, SEG can especially score on prevention of glare. As opposed to the mild Dutch climate, conditions in Chicago are more severe, thus thermal energy considerations become dominant, while lighting energy is less important. In addition, the effects of switching window state on the risk of overheating become perceptible because SEGs impacts on admission of solar gains is more pronounced. Controlling window state on the basis of luminance seems the most promising, as it allows to take advantage of low glare occurrence while at the same time avoiding excessive electricity consumption. More work is required to evaluate if more advanced control strategies can further enhance performance of SEG

7.5 Discussion
The work presented in this chapter uncovered the potential of SEG by investigation of three different cases. In this way, it provides a good point of departure for further product development of SEG. Indications of performance benets are however better substantiated once a more elaborate parametric study is conducted. A sensitivity analysis can be helpful to elucidate which factors are most inuential, and thus have the highest priority to be considered for further development. The results of this study might not be as promising as one would expect on the basis of recent studies with other switchable glazings (Jonsson and Roos, 2010; Mardaljevic and Nabil, 2008; Piccolo, 2010). The deviations may be explained by the following reasons: The bandwidth for switching SEG is relative narrow compared to other switchable windows . In addition, SEG only switches in either one of three states, without the possibility for gradual transitions in between. Most comparative studies are based on thermal or daylight performance only, whereas this study considers the integral performance aspects simultaneously. Consequently, performance tradeoffs are included in the analysis and this removes the unconscious bias towards solutions that are out of balance (i.e. too much in favor of either thermal or visual performance). Switching the optical properties of SEG primarily takes place in the visible wavelength area. Blocking solar gains is therefore followed by a proportional increase in lighting energy use. The net result is that solar gains are exchanged for internal gains, and consequently part of the energy saving potential is counterbalanced. The specications of SEG in this thesis have been characterized as being Version 0.1 and more work is still required towards optimization of the nal product design. SEG introduces the possibility of selecting and tuning window properties for the dark, bright and translucent state. These properties are however subject to some fundamental restrictions as illustrated in Figure 7.6. A high visible transmittance in the bright state (1) is attainable, but this also involves a relatively high level of transparency in the dark state. On the other hand, a larger switching range (2) is also possible, but this then puts a constraint on the maximum level of transparency in the bright state. The simulation model as developed in this project can be used to investigate which optical properties and control strategy yield the optimum balance for heating, cooling and lighting energy demand. At the moment, this is still an open question. Moreover, expectations are that eventually a catalog of SEGs with different properties will coexist, each tuned to specic climatic conditions and building designs. Again, building performance simulations can be a helpful tool to accomplish this goal.
In comparative studies, T vis was smaller than 0.1 in the dark state and larger than 0.55 to 0.8 in the bright state.

90

7.5. Discussion

Bright state Dark state

Visible transmittance

Dye intensity

Figure 7.6: Conceptual representation that shows the restrictions for optimization of SEGs optical properties.

Thus far, this chapter assessed the potential of SEG 2on the basis of annual energy consumption and thermal and visual comfort requirements. The remainder of this chapter complements this analysis by evaluating the performance of SEG on the basis of two alternative performance indicators.
Option space

7.5.1 Dynamic daylight performance

Controlled ingress of daylight in buildings is generally known for its energy saving potential and contribution to sustainability in the built environment. Nowadays, architects are becoming increasingly aware of the great importance and qualities of daylighting in perception of comfort and well-being of occupants. Daylight not only acts as major determinant for the perceived quality of the space, it is also favored because of its role in supporting levels of motivation, mood, alertness and views to the outside (Altomonte, 2008). Together with the emerging use of dynamic daylight simulations, there is a strong demand for new evaluation methods that do appreciate the temporal dynamics of daylight illumination (Reinhart et al., 2006; Mardaljevic et al., 2009). One of the climate based dynamic daylight performance evaluation metrics that was developed for this purpose is the useful daylight illuminance (U D I ) (Nabil and Mardaljevic, 2005). This performance metric subdivides internal daylight illuminance levels in three bins according to their utility. Daylight with illuminance values below 100 lx creates gloomy atmospheres and is considered to be too low to be valuable as addition to articial light. Illuminance values of 2000 lx are seen as the upper limit for comfortable, or tolerable conditions in daylit spaces. As a result, values above this level are undesirable and considered as a measure for the propensity of visual discomfort. The intermediate bin is composed of values between 100 and 2000 lx, and thus ranges from (minimum) desirable to (maximum) tolerable daylight levels. Consequently, values within this bin are regarded to constitute useful illuminance levels.

Table 7.2 shows U D I values for SEG under Dutch climate with clear double glazing and internal blinds as reference case. Evidently, designers should aspire to realize buildings with the largest share of illuminance values in the intermediate bin (U D I 100-2000 ). No consensus has yet been reached about 91

Visible transmittance

The optical properties of SEG in this study only switch in the Bright state wavelength area. The luminescent visible dye technology however makes extension of the switching range to other parts of the spectrum viable. Dark state The ultimate solution would be a window that is capable of switching visible and infrared transmittance independently. Such an ideal window has been identied as the holy grail of faade technology (Selkowitz and Lee, 1998). Research efforts that aspire this aim are underway, but currently the luminescent materials active in the infrared wavelength area still suffer from low stability, low quantum 1 efciency, and a relatively narrow absorption spectrum (Goldschmidt, 2009).

7. P ERFORMANCE SIMULATION OF S MART E NERGY G LASS

Table 7.2: Useful daylight illuminance for the reference renovation case (case A) and SEG with different operational strategies (cases B to F).

A U D I <100 U D I 100-2000 U D I >2000 [%] [%] [%] 14 70 16

B 55 41 4

C 77 23 0

D 69 28 3

E 60 40 0

F 56 44 0

acceptable or target values that satisfy both thermal and visual criteria (Mardaljevic et al., 2009). Table 7.2 reveals that application of SEG can avoid an oversupply of daylight, but on the other hand the U D I <100 value is comparatively high even when continuously switched in the bright state. These results reinforce the notion that it is worthwhile to explore possibilities to broaden the radius of action of SEG (i.e. difference between bright and dark state) to be able to modulate daylight in a variety of conditions that is as wide as possible. U D I can thus be a useful indicator to nd the best operating point on the curves of Figure 7.6. The results in Table 7.2 only consider one sensor point in space. U D I however, is well-suited to provide insights in the spatial distribution of daylight. It would be interesting to also see in visualization plots how SEG with different operational logic inuences spatial distribution of daylight in the zone.

7.5.2 Radiant asymmetry

Acceptable thermal comfort conditions are not only determined by the average thermal environment; people are also sensitive to the effects of local disturbances. Experimental studies have demonstrated that asymmetric radiant elds result in uneven body temperature and consequently lead to a higher probability of complaints about thermal comfort (Fanger et al., 1985). In Figure 7.7, the percentage of people dissatised (PPD) as a function of the radiant temperature asymmetry caused by a warm ceiling, a cool wall, a cool ceiling or by a warm wall, is shown.

Figure 7.7: Percentage of people expressing discomfort due to asymmetric radiation to a warm ceiling, cool wall, cool ceiling or warm wall or as a function of the radiant temperature asymmetry (Fanger et al., 1985).

Especially in the dark state, SEG absorbs a relatively large amount of incident solar radiation and for that reason creates an increase in window temperature. As a result, application of SEG can be a potential source of discomfort. The risk of radiant asymmetry due to SEG as warm wall was assessed on the basis of results from the simulation model. Figure 7.8 shows the simulated surface temperature of the internal SEG window pane, in comparison with ambient air temperature and the mean surface temperature of the other zone surfaces. The temperature difference between wall and window reaches a maximum of 9 C on the hottest day with the Dutch climate le. 92

7.5. Discussion

40

35

Tmax = 9 C

30

Temperaure (C)

25

20

15

10

Surface: window Surface: wall Air: ambient

0 0 1 2 3 Time (days) 4 5 6

Figure 7.8: Surface temperature of wall and window, and ambient air temperature during six days in summer.

Radiant temperature asymmetry, instead of surface temperature, is the determinant of local thermal discomfort and is dened as the difference between the plane radiant temperature Tpr on two opposite sides of a small black element. Consequently, it is not only temperature of the window but also the view factor from the occupant to the window that is important in assessing local discomfort. As Figure 7.9 illustrates, this view factor depends on window size, room geometry and occupants location in the room.

Figure 7.9: Schematic diagram illustrating how window geometry and observer position inuences the view factor for long-wave radiant heat transfer (Huizenga et al., 2006).

According to ASHRAE (2001), radiant temperature asymmetry Tpr for the conguration of Figure 7.9 can be calculated with the expression of Equation (7.1): Tpr = Tpr, left Tpr, right = (Twindow F occwindow + Twall F occwall ) Twall 1

(7.1)

ISO 7730 species an allowable threshold value of 23 C for radiant temperature asymmetry, which corresponds to a PPD of 5 %. Assuming that the view factor to the window (F occwindow ) in spaces with moderate window-to-wall ratios go up to approximately 0.5, this results in a maximum allowable window temperature of 65 C. On the basis of the simulations it can be safely concluded that for the investigated conguration there is no cause for concerns regarding discomfort due to the high window pane temperature. On the contrary, cold window surfaces likely offer a much higher risk of discomfort because the allowable temperature difference is much lower.

93

CHAPTER
8.1 Concluding remarks
95

C ONCLUSIONS

The main objective of this study was to evaluate the role that building performance simulation (BPS) can play in design of buildings with climate adaptive building shells (CABS). To clarify the characteristic properties that distinguish CABS, a total number of one hundred different application examples was examined. These projects feature a large diversity in concepts and also show a growing interest in climate adaptive architecture to promote building performance in terms of people, planet and prot. It was found that it is not only adaptivity, but also multi-ability and evolvability that are responsible for the potential success of CABS. Before becoming mainstream, rst some barriers that are rooted in the complex nature of CABS need to be bypassed. More than traditional building envelopes, design of CABS needs to address (i) maintenance issues, (ii) the human aspects associated with dynamic behavior of the envelope, and (iii) how the new technology merges with the existing practices of the conservative building industry. By denition, CABS also have to deal with trade-offs in performance because requirements imposed by changing user requirements and environmental conditions are competitive or even conicting. Caused by the impact of these factors together, it turns out that successful design and operation of CABS might not be straightforward. Nevertheless, energy policies increasingly advocate the necessity for more sustainability, and ways to suppress greenhouse gas emissions. The building sector is often regarded as the low-hanging fruit on the tree (rge Vorsatz and Novikova, 2008; UNEP, 2009) because the cumulative mitigation potential is large, and changes are readily achievable as well as highly visible to the community. CABS hold the promise to play a major role in this pursuit, and can contribute without having the need to make compromises in terms of comfort, provided they are designed in an effective way. Essential is that adaptivity should be based on informed design decisions rather than being adventitious properties as is the case in most of the 100 CABS in the overview. To achieve this, both the opportunities and risks of making faades adaptive should be made transparent to remove controversy and rationalize uncertainties already in the design stage. This was also recognized by Nobel prize winner and U.S. Secretary of Energy, Steven Chu who has phrased it in the following way (Turner et al., 2009): "Low-hanging fruit will remain on the tree as long as those making decisions about building construction, renovation, and operation are unaware of its value or do not have an easy way to harvest it."

8. C ONCLUSIONS This study demonstrates that BPS can turn out to become the powerful toolbox that is required as catalyst for this job. BPS is able to predict building performance by coupling the theory of building physics with behavioral models and climate data. The system dynamics, ranging from short-term impacts to seasonal cycles can be analyzed, and BPS also makes it possible to explore the integrated effects of various operational control strategies. Because of these properties, BPS allows for gaining increased understanding of the interdependencies between energy and comfort and dynamics of the systems in multiple physical domains. Consequently, BPS can be a valuable tool exploiting the latent potential of CABS. What may be most promising is that current BPS tools are still amenable to continuous developments. Enhanced domain integration and incessant extension of capabilities will certainly improve the utility of model outcomes. Software developments also promote more exibility which enables modelers to analyze systems that are ahead of technology rather than only having the opportunity to use predened functionalities. The ever increasing computational power in addition favors shorter simulation periods which opens the door for consideration of more design and control alternatives. The assertion that BPS can be a valuable instrument in designing buildings with CABS was conrmed in the case study of Smart Energy Glass (SEG). BPS not only displayed its traditional capabilities as a design aid, but also proved to be useful as active tool in product design and development. This study is unique in its attempt to bridge the gap between the fundamental research in luminescent solar concentrators and its practical application in architecture. A consequence of this novelty was the fact that outcomes of an empirical validation study were necessary input to ensure reliability of the simulation model. It should however be noted that such test facilities are often not available, and efforts often not justiable in routine design processes. On the basis of the simulations it is not yet possible to give conclusive answers about SEGs energy saving potential. The concept however is promising because energy savings can be achieved while occurrence of discomfort is reduced. The same results do also suggest that there is still room for improvement. The focus of attention for developing the concept of SEG should be in the laboratory as well as in the actual large-area integration in the building shell. The next section shows that BPS can continue to play a role in these ongoing developments.

8.2 Recommendations for future work

While the CABS overview examined the entire spectrum of CABS, the case study zoomed in to the details of only one concept. A corollary is that less attention was paid to other types of systems, especially those operative in the airow domain. Performance of CABS is not only application-specic but also context-dependent, which makes that extrapolating the potential of SEG in universally applicable projections for CABS is senseless. More work is required to develop understanding of effects of other kinds of adaptive behavior on building performance and to identify the potential role of e.g. airow simulation tools. The strategy for modeling and simulation of CABS in Chapter 4 provides a good starting point for research in this direction. Currently, adaptive features in BPS are primarily available in the research-grade simulation tools only. These tools require a thorough understanding of domain knowledge and often lack convenient user interfaces. It is questionable to what extent these tools will effectively assist in consultancy and design of CABS in practice. More efforts should be directed to valorization of CABS models and tools by making them easily accessible to a wider community, to support early design decisions without losing the powerful modeling capabilities. After having been used in the design stages, BPS models can also be used in the post-construction phase. One of the envisioned applications is to couple BPS with the building automation system for simulation-assisted predictive control. Although practical application examples of simulation-assisted control are still limited, the viability of the concept has been demonstrated successfully (Clarke et al., 2002; Henze et al., 2005; Mahdavi, 2008). Exploring the role that BPS can play in operation of adaptive behavior of CABS is a challenge for future research. 96

8.2. Recommendations for future work The fragmented developments in CABS technology are typically driven by advances in science and technology (i.e. technology push). Unknown is however what kind of adaptive behavior of the faade really yields the most favorable building performance (i.e. demand pull). The true value of making building envelopes adaptive is most likely only partially accessible with current CABS concepts. Smart use of BPS can help to divert from such sub-optimal solutions. A so-called inverse modeling approach can explore the full option space to derive the optimal dynamic properties of a building shell. These results can then be used to outline future R&D paths that should lead towards this goal. Apart from its role in developing CABS concepts in general, BPS can also continue to play a role in research, design and development processes of SEG. No two building projects are alike and therefore every (new) building needs to be designed in harmony with its local context and specic functional requirements. The technology behind SEG makes it possible within certain limits to freely choose the switchable window properties (U-value, SHGC and Tvis for the dark, bright and translucent state) for a given situation. This makes it possible to develop, tune and optimize multiple SEG-coatings, each for a different application area. It would be very interesting to employ BPS to answer the following questions: What are ideal specications of SEG in a range of different applications ?; What are the associated trade-offs and performance benets for each of these cases in terms of energy consumption and thermal and visual comfort?; Will it be sufcient to focus on the development of a single, generalpurpose window, or should Peer+ invest in tailoring the properties of SEG for each particular situation? Unlike faades in the scenery of western movies, CABS do not exist in isolation. The performance evaluations in this study however only considered ctive heating and cooling loads, and did not explicitly model interactions with other building components. The assumption is that good opportunities for SEG lie in the interaction with slow-responding low-exergy building systems like aquifer thermal storage and concrete core conditioning. These kinds of systems typically offer a tempered, steady interior climate, but cannot respond fast to changing conditions, which may lead to uncomfortable conditions or needless waste of energy. Because SEG has the ability to quickly react to changes in performance requirements and environmental boundary conditions, it is potentially possible to mitigate these hazardous effects. To test this hypothesis, further use of the BPS model is inevitable.

Including window-to-wall ratio, orientation, type of building, insulation standard, renovation vs. new building, climate, etc.

97

R EFERENCES

Addington, M. (2009). Contingent Behaviours. Architectural Design, 79(3):1217. Addington, M. and Schodek, D. (2005). Smart materials and new technologies: for the architecture and design professions. Amsterdam: Elsevier. Aldersey-Williams, H. (2004). Zoomorphic: New animal architecture. London: Lawrence King. Aleo, F., Pennisi, A., Scalia, S., and Simone, F. (2001). Optical and energetic performances of an electrochromic window tested in a PASSYS cell. Electrochimica Acta, 46(1314):22432249. Altomonte, S. (2008). Daylight for Energy Savings and Psycho-Physiological Well-Being in Sustainable Built Environments. Journal of Sustainable Development, 1(3):316. Aschehoug, . and Andresen, I., editors (2008). Annex 44 - Integrating Environmentally Responsive Elements in Buildings. IEA - ECBCS. Aschehoug, ., Andresen, I., Kleiven, T., and Wyckmans, A. (2005). Intelligent Building Envelopes Fad or Future? In Proceedings of the 7th Symposium on Building Physics in the Nordic Countries, Reykjavik, Iceland. ASHRAE (2001). ASHRAE Handbook of fundamentals. Atlanta, GA. Assimakopoulos, M., Tsangrassoulis, A., Santamouris, M., and Guarracino, G. (2007). Comparing the energy performance of an electrochromic window under various control strategies. Building and Environment, 42(8):28292834. Augenbroe, G. (2005). A framework for rational buildng performance dialogues. In Kolarevic, B. and Malkawi, A., editors, Performative Architecture Beyond Instrumentality, pages 98109. New York: Spon Press. Augenbroe, G. (2010). The role of simulation in performance-based building. In Hensen, J. and Lamberts, R., editors, Building Performance Simulation for Design and Operation. Bailey, C. (2009). Science and Technology Developments in Structural Fire Engineering. Structural Engineering International, 19(2):155164. Balci, O. (1998). Verication, Validation and Testing. In Banks, J., editor, Handbook of simulation: principles, methodology, advances, applications, and practice, pages 335394. New York: John Wiley & Sons. Barlow, J. (2009). Simplication: ethical implications for modelling and simulation. In Proceedings of the 18th World IMACS / MODSIM Congress, Cairns, Australia. Barnaby, C. and Crawley, D. (2010). Weather data for building performance simulation. In Hensen, J. and Lamberts, R., editors, Building Performance Simulation for Design and Operation. Batchelder, J., Zewail, A., and Cole, T. (1979). Luminescent solar concentrators. 1: Theory of operation and techniques for performance evaluation. Applied Optics, 18(18):30903110. 99

R EFERENCES BBC (2009). Living wall now very much dead. BBC News, 21 August 2009, Web: http://news.bbc. co.uk/2/hi/uk_news/england/london/8215035.stm. Becker, R. (2008). Fundamentals of Performance-based Building Design. Building Simulation, 1(4):356 371. Bellia, L., Cesarano, A., Iuliano, G., and Spada, G. (2008). Daylight glare: a review of discomfort indexes. In Proceedings of Visual quality and energy efciency in indoor lighting: today for tomorrow, Rome, Italy. Bendig, M., Hanika, J., Dammertz, H., Goldschmidt, J., Peters, M., and Weber, M. (2008). Simulation of Fluorescent Concentrators. In Proceedings of the IEEE symposium on Interactive Ray Tracing, pages 9398, Los Angeles, CA. Benson, D. and Branz, H. (1995). Design goals and challenges for a photovoltaic-powered electrochromic window covering. Solar Energy Materials and Solar Cells, 39(2-4):203211. Benyus, J. (1997). Biomimicry: innovation inspired by nature. New York: Perennial. Berdahl, P Akbari, H., Levinson, R., and Miller, W. (2008). Weathering of roong materials - an overview. ., Construction and Building Materials, 22(4):423433. Biancardo, M., Taira, K., Kogo, N., Kikuchi, H., Kumagai, N., Kuratani, N., Inagawa, I., Imoto, S., and Nakata, J. (2007). Characterization of microspherical semi-transparent solar cells and modules. Solar Energy, 81(6):711716. Bloomframe (2010). Bloomframe homepage. Web: http://www.bloomframe.com/ retrieved: 05-032010. Bodart, M. and Deneyer, A. (2006). A guide for the building of daylight scale models. In Proceedings of Plea2006 The 23rd Conference on Passive and Low Energy Architecture, Geneva, Switzerland. Bokel, R. and van der Voorden, M. (2005). Inuence of the facade on the energy demand: an introduction to climate adaptive skins. In Proceedings of IBPSA-NVL 2005. Bougdah, H. and Sharples, S. (2010). Environment, Technology and Sustainability. London & New York: Taylor & Francis. Bourgeois, D., Reinhart, C., and Macdonald, I. (2006). Adding advanced behavioural models in whole building energy simulation: A study on the total energy impact of manual and automated lighting control. Energy and Buildings, 38(7):814823. Brager, G. and de Dear, R. (1998). Thermal adaptation in the built environment: a literature review. Energy and Buildings, 27(1):8396. Brand, S. (1994). How buildings learn: What happens after theyre being built. New York: Viking Press. Bueno Unzeta, B., Norford, L., and Britter, R. (2009). An urban weather generator coupling building simulations with a physically based urban model. In Proceedings of the seventh International Conference on Urban Climate, Yokohama, Japan. Chatten, A., Barnham, K., Buxton, B., Ekins-Daukes, N., and Malik, M. (2003). A new approach to modelling quantum dot concentrators. Solar Energy Materials and Solar Cells, 75(3-4):363371. Chen, Q. (2009). Ventilation performance prediction for buildings: A method overview and recent applications. Building and environment, 44(4):848858. Chiras, D. (2001). Insulation for Passive Solar. Web: http://www.greenhomebuilding.com/QandA/ solarheat/insulation.htm retrieved: 05-03-2010. 100

References Chwif, L., Barretto, M., and Paul, R. (2000). On simulation model complexity. In Proceedings of the 2000 Winter Simulation Conference, pages 449455, Orlando, Florida. Cisar, H. (2003). Vertigo: Weiterbauen als ubersetzungsarbeit. Werk, Bauen und Wohnen, 6:3237. Citherlet, S. (2001). Towards the Holistic Assessment of Building Performance based on an Integrated Simulation approach. PhD thesis, Swiss Federal Institute Of Technology (EPFL), Lausanne, Switzerland. Claridge, D. E. (2010). Building simulation for practical operational optimization. In Hensen, J. and Lamberts, R., editors, Building Performance Simulation for Design and Operation. Clarke, J. (2001). Energy simulation in building design. Oxford: Butterworth - Heinemann, 2nd edition. Clarke, J. (2009). Integrated Building Performance Simulation: Achievements and Requirements. Keynote Presentation Building Simulation 2009, Glasgow, Scotland. Clarke, J., Cockroft, J., Conner, S., Hand, J., Kelly, N., Moore, R., OBrien, T., and Strachan, P. (2002). Simulation-assisted control in building energy management systems. Energy and Buildings, 34(9):933940. Clements-Croome, D. (2009). Editorial. Intelligent Buildings International, 1(1):34. Cole, R. and Brown, Z. (2009). Reconciling human and automated intelligence in the provision of occupant comfort. Intelligent Buildings International, 1(1):3955. Cook, J. (1996). Architecture indigenous to extreme climates. Energy and Buildings, 23(3):277291. Crawley, D., Hand, J., Kummert, M., and Grifth, B. (2008). Contrasting the capabilities of building energy performance simulation programs. Building and Environment, 43(4). Currie, M., Mapel, J., Heidel, T., Goffri, S., and Baldo, M. (2008). High-Efciency Organic Solar Concentrators for Photovoltaics. Science, 321:226228. Datta, S., Hanan, S., and Pitts, G. (2009). Experiments with stochastic processes: Facade subdivision based on wind motion. International Journal of Architectural Comuting, 7(3):389402. Daum, D. and Morel, N. (2009). Assessing the saving potential of blind-controller via multi-objective optimization. Building Simulation, 2(3):175185. Davies, M. (1981). A wall for all seasons. RIBA Journal, 88(2):5557. Debije, M. (2010). Solar Energy Collectors with Tunable Transmission. Advanced Functional Materials, 20(9):14981502. DesignAdvisor (2008). The MIT Design Advisor. Web: http://designadvisor.mit.edu/design/ retrieved: 31-08-2009. Ding, G. (2008). Sustainable construction the role of environmental assessment tools. Journal of Environmental Management, 86(3):451464. Djunaedy, E. (2005). External coupling between building energy simulation and computational uid dynamics. PhD thesis, Eindhoven University of Technology, Eindhoven, The Netherlands. DOE (2010).

buildings/tools_directory/ retrieved: 29-03-2010.

Building energy software tool directory.

Web: http://apps1.eere.energy.gov/

Douglass, D. (2008). Dening a sustainable aesthetic: A new paradigm for architecture. PhD thesis, University of Southern California, Los Angeles, United States. 101

R EFERENCES Dounis, A. and Caraiscos, C. (2009). Advanced control systems engineering for energy and comfort management in a building environment - A review. Renewable and Sustainable Energy Reviews, 13(67):12461261. Drake, S. (2007). The third skin; architecture, technology & environment. Sydney: UNSW Press. Dufe, J. and Beckman, W. (2006). Solar Engineering of Thermal Processes. Hobroken: Wiley. Elkington, J. (1998). Cannibals with forks: the triple bottom line of 21st century business. Gabriola Island, Canada: New Society Press. Engel, R. (1984). Using simulation to optimize solar greenhouse design. In Proceedings of the 17th annual symposium on Simulation, pages 119139, Piscataway, NJ. Erell, E. and Williamson, T. (2006). Simulating air temperature in an urban street canyon in all weather conditions using measured data at a reference meteorological station. International Journal of Climatology, 26(12):16711694. Evans, M. and Kelly, N. (1996). Modelling Active Building Elements with Special Materials. Technical report, Energy Simulation Research Unit - University of Strathclyde, Glasgow. Fanger, O., Ipsen, B., Langkilde, G., Olessen, B., Christensen, N., and Tanabe, S. (1985). Comfort limits for asymmetric thermal radiation. Energy and Buildings, 8(3):225236. Ferguson, S., Lewis, L., Siddiqi, A., and de Weck, O. (2007). Flexible and recongurable systems: nomenclature and review. In Proceedings of the ASME 2007 Design Engineering Conference, Las Vegas, Nevada. Fox, M. and Kemp, M. (2009). Interactive Architecture. New York: Princeton Architectural Press. Fox, M. and Yeh, B. (1999). Intelligent kinetic systems in architecture. In Proceedings of Managing Interactions in Smart Environments (MANSE), London, UK. Gallagher, S., Norton, B., and Eames, P (2007). Quantum dot solar concentrators: Electrical conversion . efciencies and comparative concentrating factors of fabricated devices. Solar Energy, 81(6):813821. Geebelen, B. (2003). Daylighting computation methods - from dot chart to digital simulation. Syllabus Studiedag Daglichttoetreding in gebouwen. Goetzberger, A. and Greube, W. (1977). Solar energy conversion with uorescent collectors. Applied Physics A: Materials Science & Processing, 14(2):123139. Goldschmidt, J. (2009). Novel Solar Cell Concepts. PhD thesis, Universitt Konstanz, Germany. Gowri, K. (2004). Green Building Rating Systems: An Overview. ASHRAE Journal, November:5659. Granqvist, C. (1991). Solar energy materials. Applied Physics A: Materials Science & Processing, 52(2):83 93. Gruber, P (2008). The signs of life in architecture. Bioinspiration & biomimetics, 3(2):19. . Guan, L. (2009). Preparation of future weather data to study the impact of climate change on buildings. Building and Environment, 44(4):793800. Haeusler, M. (2009). Media facades: history, technology, content. Ludwigsburg: Avedition. Haldi, F. and Robinson, D. (2008). On the behaviour and adaptation of ofce occupants. Building and Environment, 43(12):21632177. Haynes, B. (2008). The impact of ofce comfort on productivity. Journal of Facilities Management, 6(1):3751. 102

References Heaberlin, C. (2004). Weathering: the ever-changing nish. Masters thesis, University of Cincinati. Heerwagen, J. (2000). Buildings, Organizational Success, and Occupant Productivity. Building Research and Information, 28(5):353367. Heim, D. (2010). Isothermal storage of solar energy in building construction. Renewable Energy, 35(4):788796. Henriksen, J. (2008). Taming the Complexity Dragon. Journal of Simulation, 2(1):317. Hensen, J. (1999). Simulation of building energy and indoor environmental quality - some weather data issues. In Proceedings International Workshop on Climate data and their applications in engineering, Prague, Czech Republic. Hensen, J., Djunaedy, E., Radoevi , M., and Yahiaoui, A. (2004). Building Performance Simulation for c better design: some issues and solutions. In Proceedings of Plea2004 - The 21st Conference on Passive and Low Energy Architecture, pages 11851190, Eindhoven, The Netherlands. Hensen, J. and Lamberts, R. (2010). Introduction to building performance simulation. In Hensen, J. and Lamberts, R., editors, Building Performance Simulation for Design and Operation. Henze, G., Kalz, D., Felsmann, C., and Knabe, G. (2005). Experimental analysis of model-based predictive optimal control for active and passive thermal storage. HVAC&R Research, 11(2):189213. Henze, G. and Neumann, C. (2010). Modelling and simulation in building automation systems. In Hensen, J. and Lamberts, R., editors, Building Performance Simulation for Design and Operation. Hoes, P Hensen, J., Loomans, M., de Vries, B., and Bourgeois, D. (2009). User behavior in whole building ., simulation. Energy and Buildings, 41(3):295302. Hoffman, A. and Henn, R. (2008). Overcoming the Social and Psychological Barriers to Green Building. Organization & Environment, 21(4):390419. Hopfe, C. (2009). Uncertainty and sensitivity analysis in building performance simulation for decision support and design optimization. PhD thesis, Eindhoven University of Technology, Eindhoven, The Netherlands. Hu, H. (2009). Risk-conscious design of off-grid solar houses. PhD thesis, Georgia Insitute of Technology, Georgia, Atlanta, United States. Huizenga, C., Zhang, H., Mattelaer, P Yu, T Arens, E., and Lyons, P. (2006). Window performance for ., human thermal comfort. Technical report, Final Report National Fenestration Rating Council. Hutcheon, N. (1963). Requirements for exterior walls. Canadian Building Digest, 48. IESNA (2000). Illuminating Engineering Society of North America Lighting Handbook. New York. Ishii, K. (2000). Structural design of retractable roof structures. London: WIT Press. Janak, M. (2009). Integrated building energy and lighting simulation in the framework of EU PV-LIGHT project. In Proceedings of Building Simulation 2009, the Eleventh International IBPSA Conference, pages 16711677, Glasgow, Scotland. Jencks, C. (2005). Iconic Building: The Power of Enigma. London: Frances Lincoln. Jensen, O., Wittchen, K., and Engelund, K. (2009). Towards very low energy buildings Energy saving and CO2 emission reduction by changing European building regulations to very low energy standards. Technical report, Danish Building Research Institute. Jentsch, M., Bahaj, A., and James, P (2008). Climate change future proong of buildings - generation . and assessment of building simulation weather les. Energy and Buildings, 40(12):21482168. 103

R EFERENCES John, G., Clements-Croome, D., and Jeronimidis, G. (2005). Sustainable building solutions: a review of lessons from the natural world. Building and Environment, 40(3):319328. Jonsson, A. and Roos, A. (2010). Evaluation of control strategies for different smart window combinations using computer simulations. Solar Energy, 84(1):19. Judkoff, R. (2008). Testing and validation of building energy simulation tools. Final Task Management Report - IEA SHC Task 34 / IEA ECBCS Annex 43. Judkoff, R. and Neymark, J. (1995). International Energy Agency Building Energy Simulation Test (BESTEST) and Diagnostic Method. Technical report, National Renewable Energy Laboratory and national laboratory of the U.S. Department of Energy. Karlsson, J., Rubin, M., and Roos, A. (2001). Evaluation of predicitve models for the angle-dependent total solar energy transmittance of glazing materials. Solar Energy, 71(1):2331. Kelly, N. (1998). Towards a design environment for building integrated energy systems: The integration of electrical power ow with building simulation. PhD thesis, University of Strathclyde, Glasgow, UK. Klooster, T. (2009). Smart Surfaces and their Application in Architecture and Design. Basel: Birkhuser. Klotz, L., Mack, D., Klapthor, B., Tunstall, C., and Harrison, J. (2010). Unintended anchors: Building rating systems and energy performance goals for U.S. buildings. Energy Policy, 38(7):35573566. Knaack, U., Klein, T., and Bilow, M. (2008). Imagine 01Faades. Rotterdam: 010 Publishers. Knaack, U., Klein, T., Bilow, M., and Auer, T. (2007). FaadesPrinciples of construction. Basel, Boston, Berlin: Birkhaser. Kolarevic, B. and Malkawi, A. (2005). Performative architecture: Beyond instrumentality. Oxford: Taylor & Francis. Kotiadis, K. and Robinson, S. (2008). Conceptual modelling: Knowledge acquisition and model abstraction. In Proceedings of the 2008 Winter Simulation Conference, pages 951958, Miami, Florida. Kuhn, T. (2010). EU FP7 project Cost-Effective Convert facades into multifunctional, energy gaining components. Web: http://www.cost-effective-renewables.eu retrieved: 05-03-2010. Kula, D. and Ternaux, E. (2009). Materiology: the creatives guide to materials and technologies. Basel: Birkhuser. Kuo, T., Lin, J., Takeuchi, Y., and Huang, Y. (2009). Intelligent daylight panel control system based on fyzzy control for green buildings. In Proceedings of WCE 2009 the World Congress on Engineering, London: United Kingdom. Kuznik, F., Virgone, J., and Johannes, K. (2010). Development and validation of a new TRNSYS type for the simulation of external building walls containing PCM. Energy and Buildings, 42(7):10041009. Larson, G. and Shakespeare, R. (1998). Rendering with Radiance: the art and science of lighting visualization. San Francisco: Morgan Kaufman. LBNL (2010). Window & Daylighting software tools. Web: http://windows.lbl.gov/software/ default.htm retrieved: 12-04-2010. Leal, V., Erell, E., Maldonado, E., and Etzion, Y. (2004). Modelling the solvent ventilated window for whole building simulation. Building Services Engineering Research and Technology, 25(3):183195. Lee, E., DiBartolomeo, D., and Selkowitz, S. (2006). Daylighting control performance of a thin-lm ceramic electrochromic window: Field study results. Energy and Buildings, 38(1):3044. 104

References Lee, E. and Tavil, A. (2007). Energy and visual comfort performance of electrochromic windows with overhangs. Building and Environment, 42(6):24392449. Lehar, M. and Glicksman, L. (2003). A simulation tool for predicting the energy implications of advanced facades. In Carmeliet, J., Hens, H., and Vermeir, G., editors, Research in Building Physics Proceedings of the Second International Conference on Building Physics, pages 513518, Leuven, Belgium. Leydecker, S., Kroto, H., Veith, M., Klbel, M., and Peters, S. (2008). Nano Materials in Architecture, Interior architecture and Design. Basel: Birkhuser. Lichtenberg, J. (2005). Slimbouwen. Boxtel: neas. Lomanowski, B. and Wright, J. (2009). Modeling fenestration with shading devices in building energy simulation: a practical approach. In Proceedings of Building Simulation 2009, the Eleventh International IBPSA Conference, pages 976983, Glasgow, Scotland. Loutzenhiser, P Manz, H., Strachan, P Felsmann, C., Frank, T., Maxwell, G., and Oelhafen, P. (2006). An ., ., Empirical Validation of Modeling Solar Gains Through a Glazing Unit Using Building Energy Simulation Programs. HVAC&R Research, 12(4):10971116. Lstiburek, J. (2008). Why green can be wash. ASHRAE Journal, 2008(November):2836. Lu, X., Clements-Croome, D., and Viljanen, M. (2009). Past, present and future mathematical models for buildings. Intelligent Buildings International, 1(1):2338. Mahdavi, A. (2008). Predictive simulation-based lighting and shading systems control in buildings. Building Simulation, 1(1):2535. Mahdavi, A., Mohammadi, A., Kabir, R., and Lambeva, L. (2008). Occupants operation of lighting and shading systems in ofce buildings. Journal of Building Performance Simulation, 1(1):5765. Malkawi, A. (2004). Developments in environmental performance simulation. Automation in construction, 13(4):437445. Mardaljevic, J. (1995). Validation of a lighting simulation program under real sky conditions. Lighting Research and Technology, 27(4):181188. Mardaljevic, J. (2000). Simulation of annual daylighting proles for internal illuminance. Lighting Research and Technology, 32(3):111118. Mardaljevic, J., Heschong, L., and Lee, E. (2009). Daylight metrics and energy savings. Lighting Research and Technology, 41(3):261283. Mardaljevic, J. and Nabil, A. (2008). Electrochromic glazing and facade photovoltaic panels: a strategic assessment of the potential energy benets. Lighting Research and Technology, 40(1):5576. Maver, T. (1980). Appraisal in design. Design studies, 1(3):160165. McDowell, T., Bradley, D., Thornton, J., and Kummert, M. (2004). Simulation synergy: expanding trnsys capabilities and usability. In Proceedings of SimBuild 2004 the First IBPSA-USA National Conference, Boulder, CO. Meagher, M., van der Maas, D., Abegg, C., and Huang, J. (2009). Dynamic Ornament: Climatically Responsive Surfaces in Architecture. In Proceedings of CAADFutures 2009, Montreal, Canada. Meir, I., Garb, Y., Jiao, D., and Cicelsky, A. (2009). Post-Occupancy Evaluation: An Inevitable Step Toward Sustainability. Advances in building energy research, 3:189220. Moloney, J. (2007). A Framework for the Design of Kinetic Faades. In Proceedings of Computer-Aided Architectural Design Futures 2007, pages 461474, Sydney, Australia. 105

R EFERENCES Nabil, A. and Mardaljevic, J. (2005). Useful daylight illuminance: a new paradigm for assessing daylight in buildings. Lighting Research and Technology, 37(1):4159. Nakhi, A. (1995). Adaptive construction modeling within whole building dynamic simulation. PhD thesis, University of Strathclyde, Glasgow, UK. NASA (2010). Surface meteorology and Solar Energy, release 6.0. Web: http://eosweb.larc.nasa. gov/sse/ retrieved: 28-02-2010. Neymark, J., Judkoff, R., Knabe, G., Le, H., Drig, M., Glass, A., and Zweifel, G. (2002). Applying the building energy simulation test (BESTEST) diagnostic method to verication of space conditioning equipment models used in whole-building energy simulation programs. Energy and Buildings, 34(9):917 931. Nicol, J. and Humphreys, M. (2009). New standards for comfort and energy use in buildings. Building Research & Information, 37(1):6873. Ochoa, C. and Capuleto, I. (2009). Advice tool for early design stages of intelligent faades based on energy and visual comfort approach. Energy and Buildings, 41(5):480488. Olewnik, A., Braun, T., Ferguson, F., and Lewis, K. (2004). A framework for exible systems and its implementation in multiattribute decision making. Journal of Mechanical Design, 126:412419. Page, J., Robinson, D., Morel, N., and Scartezzini, J. (2008). A generalised stochastic model for the simulation of occupant presence. Energy and Buildings, 40(2):8398. Palomo Del Barrio, E. and Guyon, G. (2004). Theoretical basis for empirical model validation using parameters space analysis tools. Energy and Buildings, 35(10):985996. Papaefthimiou, S., Syrrakou, E., and Yianoulis, P. (2009). An alternative approach for the energy and environmental rating of advanced glazing: An electrochromic window case study. Energy and Buildings, 41(71):1726. Pedersen, C. (2007). Advanced zone simulation in EnergyPlus: incorporation of variable properties and phase change material (pcm) capability. In Proceedings of Building Simulation 2007, the Tenth International IBPSA Conference, pages 13411345, Beijing, China. Pedersen Zari, M. (2007). Biomimetic approaches to architectural design for increased sustainability. In Proceedings of the SB07 Sustainable Building Conference, Auckland, New Zealand. Peerplus (2009). Peer+ Smart Energy Glass. Web: http://www.peerplus.nl retrieved: 20-10-2009. Perez, R., Ineichen, P Steals, R., Michalsky, J., and Stewart, R. (1990). Modeling daylight availability and ., irradiance components from direct and global irradiance. Solar Energy, 44(5):271289. Petersdorff, C., Boermans, T., and Harnisch, J. (2006). Mitigation of CO2 Emissions from the EU-15 Building Stock. Beyond the EU Directive on the Energy Performance of Buildings. Environmental Science and Pollution Research, 13(15):350358. Piccolo, A. (2010). Thermal performance of an electrochromic smart window tested in an environmental test cell. Energy and Buildings, 42(9):14091417. Piccolo, A., Pennisi, A., and Simone, F. (2009). Daylighting performance of an electrochromic window in a small scale test-cell. Solar Energy, 83(6):832844. Platzer, W., editor (2003). Architectural and Technical Guidelines - Handbook for the Use of Switchable Facades Technology. Freiburg: Fraunhofer Institute for Solar Energy Systems. Plaza, B. (2000). Evaluating the inuence of a large cultural artifact in the attraction of tourism. Urban Affairs Review, 36(2):264274. 106

References Porteous, C. and MacGregor, K. (2005). Solar architecture in cool climates. London: Earthscan. Prasad, D. and Snow, M. (2005). Designing with solar power : a source book for building integrated photovoltaics (BiPV). London: Earthscan. Prez-Lombard, L., Ortiz, J., and Pout, C. (2008). A review on buildings energy consumption information. Energy and Buildings, 40(3):394398. Randl, C. (2008). Revolving architecture: a history of buildings that rotate, swivel, and pivot. New York: Princeton Architectural Press. Reinhart, C. (2004). Lightswitch-2002: a model for manual and automated control of electric lighting and blinds. Solar Energy, 77(1):1528. Reinhart, C. (2010). Simulation-based Daylight Performance Predictions. In Hensen, J. and Lamberts, R., editors, Building Performance Simulation for Design and Operation. Reinhart, C. and Andersen, M. (2006). Development and validation of a Radiance model for a translucent panel. Energy and Buildings, 38(7):890904. Reinhart, C., Bourgeois, D., Dubrous, F., Laouadi, A., Lopez, P., and Stelescu, O. (2007). DAYLIGHT 1-2-3 a state-of-the-art daylighting/energy analysis software for initial design investigations. In Proceedings of Building Simulation 2007, the Tenth International IBPSA Conference, pages 16691676, Beijing, China. Reinhart, C. and Breton, P (2009). Experimental Validation of Autodesk 3ds Max Design 2009 and . Daysim3.0. Leukos, 6(1). Reinhart, C. and Fitz, A. (2006). Findings from a survey on the current use of daylight simulations in building design. Energy and Buildings, 38(7):824835. Reinhart, C. and 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(2):167187. Reinhart, C., Mardaljevic, J., and Rogers, Z. (2006). Dynamic Daylight Performance Metrics for Sustainable Building Design. Leukos, 3(1):731. Reinhart, C. and Voss, K. (2003). Monitoring manual control of electric lighting and blinds. Lighting Research and Technology, 35(3):243258. Reinhart, C. and Walkenhorst, O. (2001). Validation of dynamic RADIANCE-based daylight simulations for a test ofce with external blinds. Energy and Buildings, 33(7):683697. Richardson, I., Thomson, M., and Ineld, D. (2008). A high-resolution domestic building occupancy model for energy demand simulations. Energy and Buildings, 40(8):15601566. Riederer, P Keilholz, W., and Ducreux, V. (2009). Coupling TRNSYS with Simulink A method to auto., matically export and use TRNSYS models within Simulink and vice versa. In Proceedings of Building Simulation 2009, the Eleventh International IBPSA Conference, pages 16281633, Glasgow, Scotland. Ritter, A. (2007). Smart materials in architecture, interior architecture and design. Basel: Birkhuser. Rivard, H., Bdard, C., Fazio, P and Ha, K. (1995). Functional analysis of the preliminary building ., envelope design process. Building and Environment, 30(3):391401. Roaf, S., Crichton, D., and Nicol, F. (2005). Adapting Buildings and Cities for Climate Change; A 21st century survival guide. Oxford: Architectural Press. Robinson, S. (2008). Conceptual modelling for simulation Part I: denition and requirements. Journal of the Operational Research Society, 59(3):278290. 107

R EFERENCES Rogers, Z. (2007). Overview of Daylight Simulation Tools. In Proceedings of the second Velux daylight symposium, Bilbao, Spain. Rubin, M., von Rottkay, K., and Powles, R. (1998). Window Optics. Solar Energy, 62(3):149161. Sailor, D. (2009). A green roof model for building energy simulation programs. Energy and Buildings, 40(8):14661478. Sakamoto, T., Ferre, A., and Kubo, M., editors (2008). From control to design. Barcelona: Actar. Saleh, J., Mark, G., and Jordan, C. (2009). Flexibility: a multi-disciplinary literature review and a research agenda for designing exible engineering systems. Journal of Engineering Design, 20(3):307323. Samuel, A., Imbabi, M., Peacock, A., and Strachan, P. (2003). An engineering approach to modelling of dynamic insulation using esp-r. In Proceedings of CIBSE/ASHRAE Conference, Edinburgh, Scotland. Sargent, R. (2005). Verication and validation of simulation models. In Proceedings of the 2005 Winter Simulation Conference, pages 130143, Orlando, Florida. Schler, A., Kostro, A., Galande, C., Valle del Olmo, M., de Chambrier, E., and Huriet, B. (2007). Principles of Monte-Carlo Ray-Tracing Simulations of Quantum Dot Solar Concentrators. In Proceedings of ISES World Congress 2007, pages 10331037, Beijing, China. Selkowitz, S., Aschehoug, ., and Lee, E. (2003). Advanced interactve facades critical elements for future green buildings. In Proceedings of GreenBuild, the annual USGBC International Conference. Selkowitz, S. and Lee, E. (1998). Advanced Fenestration Systems for Improved Daylight Performance. In Proceedings of the Daylighting 98 Conference, Ontario, Canada. Senagala, M. (2005). Kinetic, responsive and adaptive: a complex-adaptive approach towards smart architecture. In Proceedings of SIGRADI, Lima, Peru. SenterNovem (2009). EP Varianten Utiliteitsgebouwen. Report: DV1.3.189. Siddiqi, A. and de Weck, O. (2008). Modeling Methods and Conceptual Design Principles for Recongurable Systems. Journal of Mechanical Design, 130(10). Singh, M. and Garg, S. (2010). An empirical model for angle-dependent g-values of glazings. Energy and Buildings, 42(3):375379. Sokolowski, J. and Banks, C., editors (2009). Principles of Modeling and Simulaton A Multidisciplinary Approach. New York: John Wiley & Sons. Spitler, J. (2010). Thermal load and energy performance prediction. In Hensen, J. and Lamberts, R., editors, Building Performance Simulation for Design and Operation. Stevens, S. (2001). Intelligent facades: occupant control and satisfaction. International Journal of Solar Energy, 21(2-3):147160. Strachan, P (2008). Simulation support for performance assessment of building components. Building . and Environment, 43(2):228236. ri, M. (2005). Solar resource data and tools for an assessment of photovoltaic systems. In JgerWaldau, A., editor, Status Report 2006, pages 9398, Luxembourg. Ofce for Ofcial Publications of the European Communities. Szokolay, S. (2004). Introduction to architectural science: the basis of sustainable design. Oxford: Architectural Press. Teuffel, P (2004). Entwerfen Adaptiver Strukturen. PhD thesis, Universitt Stuttgart, Stuttgart, Germany. . 108

References Thanachareonkit, A. (2008). Comparing Physical and Virtual Methods for Daylight Performance Modelling Including Complex Fenestration Systems. PhD thesis, Swiss Federal Institute Of Technology (EPFL), Lausanne, Switzerland. Tr ka, M. (2008). Co-simulation for Performance Prediction of Innovative Integrated Mechanical Energy c Systems in Buildings. PhD thesis, Eindhoven University of Technology, Eindhoven, The Netherlands. Tr ka, M. and Hensen, J. (2010). Overview of HVAC system simulation. Automation in construction, c 19(2):9399. Turcotte, D., Ross, M., and Sheriff, F. (2001). Photovoltaic hybrid system sizing and simulation tools: status and needs. In PV Horizon: Workshop on Photovoltaic Hybrid Systems, Montreal, Canada. Turner, J., Vaughan, E., and McCormick, C. (2009). Moving Toward High-Performance Buildings. Innovations: Technology, Governance, Globalization, 4(4):213234. UNEP (2009). Buildings and Climate Change Summary for decision makers. Technical report, United Nations Environment Programme. Urban, B. and Glicksman, L. (2006). The MIT Design Advisor A fast, simple tool for energy efcient building design. In Proceedings of SimBuild 2006 the Second National IBPSA-USA Conference, pages 270276, Cambridge, MA. Urbikain, M. and Sala, J. (2009). Analysis of different models to estimate energy savings related to windows in residential buildings. Energy and Buildings, 41(6):687695. rge Vorsatz, D. and Novikova, A. (2008). Potentials and costs of carbon dioxide mitigation in the worlds buildings. Energy Policy, 36(2):642661. van Maarsseveen, C. and van Duijnen, T., editors (2008). Gevelbouw Prestatieboek 2008. VMRG. van Maarsseveen, C. and van Duijnen, T., editors (2009). Gevelbouw Prestatieboek 2009. VMRG. van Sark, W., Barnham, K., Slooff, L., Chatten, A., Bchtemann, A., Meyer, A., Mc.Cormack, S., Koole, R., Farrell, D., Bose, R., Bendbendige, E., Burgers, A., Budel, T., Quilitz, J., Kennedy, M., Meyer, T., Wadman, S., van Klink, G., van Koten, G., Meijerink, A., and Vanmaekelbergh, D. (2008). Luminescent Solar Concentrators a review of recent results. Optics Express, 16(26):2177321792. van Treeck, C. (2010). Indoor thermal quality performance prediction. In Hensen, J. and Lamberts, R., editors, Building Performance Simulation for Design and Operation. Velds, M. and Christoffersen, J. (2001). Monitoring Procedures for the Assessment of Daylighting Performance of Buildings. Report of IEA SHC Task 21 / IEA ECBCS Annex 29. Vermillion, J. (2002). Phototropic architecture: intelligent responses to sunlight stimuli. Masters thesis, Ball State University, Muncie, Indiana, United States. Von Stamm, B. (2008). The role of prototypes. In Managing Innovation, Design and Creativity, pages 183193. West Sussex: John Wiley & Sons. Walkenhorst, O., Luther, J., Reinhart, C., and Timmer, J. (2000). Dynamic annual daylight simulations based on one-hour and one-minute means of irradiance data. Solar Energy, 72(5):385395. Wang, W. and Beausoleil-Morrison, I. (2009). Integrated simulation through the source-code coupling of component models from a modular simulation environment into a comprehensive building performance simulation tool. Journal of Building Performance Simulation, 2(2):115126. Wargocki, P (2009). Ventilation, thermal comfort, health and productivity. In Mumovic, D. and San. tamouris, M., editors, A Handbook of Sustainable Building Design and Engineering: An Integrated approach to energy, health and operational performance, pages 181196. London: Earthscan. 109

R EFERENCES WBCSD (2009). Energy Efciency in Buildings Transforming the Market. Technical report, World Business Council for Sustainable Development. Wenham, S., Green, M., Watt, M., and Corkish, R. (2007). Applied photovoltaics. London: Earthscan. Wetter, M. (2009a). Modelica-based modelling and simulation to support research and development in building energy and control systems. Journal of Building Performance Simulation, 2(2):143161. Wetter, M. (2009b). Modeling, Simulation and Analysis of Integrated Building Energy and Control Systems. Presented at LBNL Seminar Environmental Energy Technologies Division. Wetter, M. and Haves, P (2008). A modular building controls virtual test bed for the integration of het. erogeneous systems. In Proceedings of SimBuild 2008 the Third National IBPSA-USA Conference, pages 6976, Berkeley, CA. Widn, M. and Wckelgrd, E. (2010). A high-resolution stochastic model of domestic activity patterns and electricity demand. Applied Energy, 87(6):18801892. Wienold, J. (2009). Dynamic Daylight Glare Evaluation. In Proceedings of Building Simulation 2009, the Eleventh International IBPSA Conference, pages 944951, Glasgow, Scotland. Wigginton, M. and Harris, J. (2002). Intelligent skins. Oxford: Butterworth - Heinemann. Williams, K. and Dair, C. (2007). What is stopping sustainable building in england? barriers experienced by stakeholders in delivering sustainable developments. Sustainable Development, 15(3):135147. Wong, P Shimoda, Y., Nonaka, M., Inoue, M., and Mizuno, M. (2008). Semi-transparent PV: Thermal ., performance, power generation, daylight modelling and energy saving potential in a residential application. Renewable enrgy, 33(5):10241036. Woudhuysen, J. and Abley, I. (2004). Why is consruction so backward? West-Sussex: Wiley-Academy. Wyckmans, A. (2005). Intelligent Building Envelopes - Architectural concepts & applications for daylighting quality. PhD thesis, Norwegian University of Science and Technology, Trondheim, Norway. Xu, X. and Van Dessel, S. (2008). Evaluation of a prototype active building envelope window-system. Energy and Buildings, 40(2):168174. Yannas, S., Erell, E., and Molina, J. (2006). Roof cooling techniques: a design handbook. London: Earthscan. Yeadon, P (2008). Smarticles: Nanotechnology materializes. ArchitectureBoston, 11(4):2629. . Yi, Y. and Malkawi, A. (2008). Site-specic prediction for energy simulation by integrating computational uid dynamics. Building Simulation, 1(3):270277. Zerkin, A. (2006). Mainstreaming high performance building in New York City: A comprehensive roadmap for removing the barriers. Technology in Society, 28(1-2):137155. Zhai, Z., Chen, Q., Haves, P and Klems, J. (2002). On approaches to couple energy simulation and ., computational uid dynamics programs. Building and Environment, 37(8-9):857864. Zinzi, M. (2006). Ofce worker preferences of electrochromic windows: a pilot study. Building and Environment, 41(9):12621273. Zuo, W. and Chen, Q. (2010). Fast and informative ow simulations in a building by using fast uid dynamics model on graphics processing unit. Building and environment, 45(3):747757. Zupan i , B. and Sodja, A. (2008). Object oriented modelling of variable envelope properties in buildcc ings. WSEAS Transactions on Systems and Control, 12(3):10461056.

110

APPENDIX
R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

OVERVIEW OF 100 C LIMATE A DAPTIVE B UILDING S HELLS

Climate Adaptive Building Shells

What can we simulate?

Overview of Climate Adaptive Building Shells

100

R.C.G.M. Loonen

Please refer to the separate booklet for the actual overview of 100 climate adaptive building shells.

111

APPENDIX
Bright state Tvis SHGC U -value [-] [-] [Wm2 K1 ] Dark state Reference case 1 Tvis SHGC U -value [-] [-] [Wm2 K1 ] 0.52 0.29 1.31 SEG bright state SEG dark state Reference case 2 Tvis SHGC U -value [-] [-] [Wm2 K1 ] 0.79 0.70 2.70 SEG bright state SEG dark state 113

S MART E NERGY G LASS PRODUCT INFORMATION

The product data of Smart Energy Glass is withheld from public disclosure because it is considered condential business information. The data is available upon request.
Table B.1: Material properties of the SEG prototype in the experimental setup of the validation study in Chapter 6

Table B.2: Material properties of the SEG window and the reference window for Case 1 Energy efcient ofce

Table B.3: Material properties of the SEG window and the reference window for Case 2 Advanced renovation

B. S MART E NERGY G LASS PRODUCT INFORMATION

Table B.4: Material properties of the SEG window and the reference window for Case 3 SEG in a continental climate

Reference case 3 Tvis SHGC U -value [-] [-] [Wm2 K1 ] 0.52 0.29 1.31

SEG bright state

SEG dark state

114

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

Climate Adaptive Building Shells

What can we simulate?

Overview of Climate Adaptive Building Shells

100

R.C.G.M. Loonen

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

Overview of 100

Climate Adaptive Building Shells

part of MSc. Thesis

CABS What can we simulate?

21 June 2010
Student id Master Faculty R.C.G.M. Loonen 0570677 prof. dr. ir. J.L.M. Hensen Building Services Architecture, Building & Planning Eindhoven University of Technology dr. dipl.-ing. M. Trka D. Cstola, MSc Advisors

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

Preface
This booklet presents the overview of 100 climate adaptive buildings shells (CABS) that was developed as part of the MSc project CABS What can we simulate?. The overview is preferably read in conjunction with the main body of the thesis, however, it was drafted in such a way that it also invites to be examined in a stand-alone manner. The concept of CABS is very broad and described by a multitude of different terms. Many building envelopes do exhibit some form of dynamism; in this overview only those concepts were included that fulfill the following definition: A climate adaptive building shell has the ability to repeatedly and reversibly change its functions, features or behavior over time in response to changing performance requirements and variable boundary conditions. By doing this, the building shell effectively seeks to improve overall building performance in terms of primary energy consumption while maintaining acceptable thermal and visual comfort conditions. Beside being used as reference book for the thesis, the material presented in this booklet serves more purposes. By synopsizing case studies, prototypes and research projects it offers a technology scan that shows the state-of-the-art in adaptive faade technology. This information can be used by professionals engaged in development of building envelopes, and also provides a point of inspiration for designers exploring possibilities to enrich the domain of adaptive building shell technology. The overview of 100 CABS is the result of an extensive literature survey, collected from different kinds of information sources. It attempts to provide a comprehensive database that covers the whole spectrum of CABS, ranging from built examples that successfully operate for many years, to the wildest utopian concepts. The field of CABS is relatively young and in a continuous state of flux. Therefore the expectation is that this compilation will soon be superseded by even more appealing concepts and better performing technologies. Before starting the actual overview, the next page first provides a short introduction that (i) helps to understand the logic of classification, and (ii) shows a legend to the symbols and abbreviations used in the overview.

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop open-loop Control of adaptivity

MACRO MICRO

Introduction
Architect / developer| place | year

Name of the concept, listed in alphabetical order

Adaptive features

In two or three sentences, this section gives a short description of the working principles of the concept. The functions and performance aspects of the concept are indicated and substantiated with claims by the author or developer. For more details and additional information, the interested reader is encouraged to investigate the references in the bottom right corner

00

Superscripts in the main text refer to these CABS numbers

The relevant physics of Chapter 4 is presented in this section

This area features some pictures and schematic drawings that demonstrate the working mechanisms and appearance of the concept

Adaptive surface, indicated via thick lines

This area provides references to the original information sources and gives recommendations for further reading Where possible, two or more independent references are provided in order to keep clear of too biased opinions

State of development, indicated via colors

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MICRO

Active Building Envelope

Rensselaer Polytechnic Institute | Troy, New York

In ABE systems, photovoltaic (PV) cells are used to transform solar energy directly into electrical energy. This electrical energy is subsequently used to power thermoelectric (TE) heatpumps. Both TE and PV systems are integrated in one enclosure surface. Depending on the direction of electrical current applied to the TE system, ABE systems can either operate in a heating or cooling mode.

01

[RPI, Steven Van Dessel]

[Khire et al., 2006]

Khire, R., Messac, A. and van Dessel, S. (2006) Design of thermoelectric heat pump unit for active building envelope systems, International Journal of heat and mass transfer 48(19-20): 4028-4040 Xu X. and van Dessel, S. (2008) Evaluation of an Active Building Envelope window-system, Building and environment 43(11): 1785-1791

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MICRO

Active window / micromirror

University of Kassel | Germany

Large-areas of micromirror arrays (150 x 400 m2) enable translucent protection from solar glare and sunshading, controllable both in terms of time and activation area. Each mirror can be moved within a prescribed angular range, by application of small voltages between the mirror layer and a planar ITO electrode. In contrast to macroscopic solutions, such micromirror-based solutions are nearly invisible, weatherproof and maintenance free.

02

[Viereck et al., 2009]

[Viereck et al., 2009]

[Viereck et al., 2009]

Klooster, T. (2010) Smart Surfaces and their application in architecture and design, Berlin: Birkhuser Viereck, V., Li, Q., Jkel, A. and Hillmer, H. (2009) Large-area applications of optical MEMS: micromirror arrays guide daylight, optimize indoor illumination, Photonik International 2009(2): 48-49 Viereck, V., Ackermann, J., Li, Q., Jkel, A., Schmidt, J. and Himmler, H. (2008) Sun glasses for Buildings on Micro Mirror Arrays, in IEEE Proceedings of Networked Sensing Systems 2008. pp: 135-139

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Adaptive Fritting
Hoberman Associates | New York

Integrated glass units with movable fritted patterns that allow for a broad range of possible transparencies. Utilizes a graphic pattern surface treatment, inspired by nature, in order to control heat gain and modulate light, while allowing sufficient transparency for viewing. Winner of the Wyss Prize for Bioinspired Adaptive Architecture 2009.

03

[Hoberman, 2009]

[Hoberman, 2009]

[Hoberman, 2009]

Brownell, B. (2010) Transmaterial 3: A catalog of materials that redefine our physical environment, Princeton Architectural Press Hoberman, C. (2009) Adaptive fritting http://hoberman.com/portfolio/gsd.php?projectname=Adaptive%20Fritting

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

AeroDimm
Stefan Pfaffstaller | Vienna, Austria | 2004

The biomimetic surface is inspired by the color change of cephalopod skin. Elastic membranes change volume between two layers of the faade and in this way control levels of daylight and solar gains. Pneumatic pressure is the driving force behind the adaptive mechanisms, and the ventilation tubes are integrated into the external layer.

04

[Gruber, 2008]

Gruber, P. (2008), The signs of life in architecture, Bioinspiration & Biomimetics 3(2): 1-9

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Aegis Hyposurface
Mark Goulthorpe | dECOi Architects | 1999-2001

The system consists of transformable triangulated metal plates, driven by a bed of 896 pneumatic pistons. Deforms in real-time, based on various environmental stimuli, including the sounds and movement of people, weather and electronic information. The piece marks the transition from autoplastic (determinate) to alloplastic (interactive, indeterminate) space.

05

[http://hyposurface,org]

Brownell, B. (2006) Transmaterial: A catalog of materials that redefine our physical environment, Princeton Architectural Press Ataman, O. and Rogers, J. (2005) Toward New Wall Systems: Lighter, Stronger, Versatile, in Proceedings of the 10th Iberoamerican Congress of Digital Graphics, Santiago de Chile, Chile. pp: 248-253

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Aldar Central Market

Foster and Partners | Abu Dhabi | 2006-2010

The grid system provides panels of controlled permeability that vary smoothly between a covered state and a largely open state. In its covered configuration, the shading roof appears similar to a traditional coffered Islamic roof. When retracted, the roof becomes a slender lattice that complements the designs by the Foster team for fixed shading.

06

[http://hoberman.com]

Aldar Central Market, Abu Dhabi, Foster+Partners Projects. http://www.fosterandpartners.com/Projects/1431/Default.aspx Aldar Central Market, Hoberman Associates Works. Website: http://hoberman.com/portfolio/aldarcentralmarket.php Grid System, Adaptive building initiative. Website: http://www.adaptivebuildings.com/cladding-shading-systems.html

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Allwetterdach
ESCO | Germany

When the slats are open, heat is prevented from building up in glazed rooms and conservatories because warm air can escape via the chimney effect. In winter, when the roof is closed, the infrared radiation (IR) can pass through in only one direction. The warmth thus remains under the roof.

07

[www.allwetterdach.com]

Smout, L. (2008) Slat roof, European Patent Application EP1992756,

http://www.freepatentsonline.com/EP1992756.html

Allwetterdach Flyer. http://allwetterdach.com/downloads/EscoProspektDT2009.pdf

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Aperture
Frdric Eyl and Gunnar Green | University of the Arts Berlin, Germany | 2005

This faade installation consists of a matrix of iris diaphragms. Like the human eye's iris and irises in camera objectives, they react to light, widening and contracting with corresponding increases and decreases in intensity of incoming light. The degree of opening is controlled via Light Dependent Resistors (LDR) that actuate servo-motors.

08

[Eyl and Green, 2005]

Klooster, T. (2010) Smart Surfaces and their application in architecture and design, Berlin: Birkhuser Eyl, F.and Green, G. (2005) Aperture faade installation, Website: http://www.fredericeyl.de/aperture/

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Arabe Intitute du Monde

Jean Nouvel | Paris, France | 1987

Arabe Institute du Monde is perhaps the most widely known example of a building with a climate adaptive building shell. The automatically controlled shutters are a technical interpretation of the traditional Arabic sun screens in integrated pane systems. The mechanical irises open and close in response to different light conditions over the course of the day.

09

[http://wikimedia.org]

[http://mimoa.eu]

[http://edwardlifson.blogspot.com/]

Coelho, M. and Maes, P. (2009) Shutters: A Permeable Surface for Environmental Control and Communication, in Proceedings of the Third International Conference on Tangible and Embedded Interaction, Cambridge, UK Arab World Institute, Jean Nouvel, Projects, Paris. Website: http://www.jeannouvel.com

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Balloon Sun Shading

Marcel Bilow, Tillman Klein | TU Delft, The Netherlands | 2006

The entire system is a combination of deflated (faade) and inflated (balloons) construction elements. Inflating the balloons creates an integrated active mechanism for daylight regulation and solar protection. An additional positive feature is that the shading system also makes profit of the acoustic qualities of the balloons.

10

[Knaack et al, 2008]

Knaack, U. and, Bilow, M. (2009) Innovationen im Fassadenbau, Bauportal 121(5):254-258 Knaack, U., Klein, T. and, Bilow, M (2008) Imagine 02 Deflatables, Rotterdam: 010 Publishers

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Beadwall
Zomeworks | New Mexico, USA | 1980

A form of movable insulation and shading that uses tiny polystyrene beads blown into the space between two window panes. Beadwall windows are a great application for greenhouses with the ability to drain and let the sun in during the day and then fill up to maintain temperature at night.

11

[Bear, 2009]

[www.commonwealthsolar.com]

[Bear, 2009]

Baer, S. (2009) Some passive solar buildings with a focus on projects in New Mexico, Presentation for the Albuquerque chapter of AIA, January 15, 2009 BeadWall, Website: http://www.commonwealthsolar.com/cwbead.htm

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Bengt Sjostrom Starlight theatre

Studio Gang Architects | Rockford, Illinois | 2003

The faceted roof consists of six triangular, stainless-steel-clad panels. Under the folded, origami-like roof, an intimate social setting is created with a porous boundary to the landscape. The roof can provide weather protection, if necessary. But when the roof is fully open, the panels form a six-point star through which the audience views the starlit sky.

12

[http:/rockvalleycolege.edu]

[http:/archdaily.com]

[http:/archdaily.com]

[Fox and Kemp, 2009]

Starlight Theatre, Uni-systems, featured projects, Website: http://www.unisystems.com/Projects/FeaturedProjectDetails.aspx?ProjectID=3&PN=Starlight%20Theater Bengt Sjostrom Starlight Theatre / Studio Gang Architects, ArchDaily 05 August 2009. http://www.archdaily.com/28649/bengt-sjostrom-starlight-theatre-studio-gang-architects/ Fox, M. and Kemp, M. (2009) Interactive Architecture, New York: Princeton Architectural Press

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Bi-directional thermodiodes
Cheju National University| Republic of Korea

Like electric diodes, thermodiodes have a favorable direction of heat flow. This forward direction can reversibly be changed via rotatable joints. The tested thermal conductivity of the thermodiode panels in forward mode is 35 times higher than that in backward mode. Three different system types have been studied: tilted rectangular loops (see figures), bayonet-shaped diodes and loop-and-tank diodes.

13

[Chun, 2009]

Chun, W., Chen, K. and Kim, H. (2002) Performance Study of a Bi-Directional Thermodiode Designed for Energy-Efficient Buildings, Journal of Solar Engineering 124(3): 291-299 Chun, W., Ko, Y., Lee, H., Han, H., Kim, J. and Chen, K. (2009) Effects of working fluids on the performance of a bi-directional thermodiode for solar energy utilization in buildings, Solar Energy 83(3): 409-419

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Bionic Breathing Skin

Lidia Badarnah | TU Delft, The Netherlands | 2007

The envelope transforms into a distributed ventilation system via contractions and expansions of an elastic membrane. The skin performs a process of inhaling and exhaling via lung like chambers inspired by the respiratory surface of sea sponges. The constantly changing rate of deformation of the functional units is controlled via piezo-electric actuators.

14

[Badarnah and Knaack, 2007]

Badarnah, L. and Knaack, U. (2007) Bio-Inspired Ventilating System for Building Envelopes, in Proceedings of the International Conference of the 21st Century: Building Stock Activation, Tokyo, Japan. pp: 431-438 Badarnah, L. and Knaack, U. (2007) Bionic breathing skin for buildings, in Proceedings of the International

Conference SB07: Sustainable Construction, Materials and Practices.

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

BioTower
Dennis Dollens,

BioTower adds value to the building with its leaf-cluster systems for air filtration, sound baffling and heat control. The cladding consists of a series of branch panels with an origami-like folded paper skin modeled from the observation of leaves. The outer biomechanical sensor-node pods embed biological filters, and passive cooling system embodied in digital leaf panels.

15

[Dollens, 2009]

Jodidio, P. (2009) Green Architecture Now!, Kln, Germany: Taschen Dollens, D. (2009) D-BA2: Digital Botanic Architecture, Website: http://www.tumbletruss.com/TrussImages/DBA2-150.pdf

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MICRO

Blight
Vincent Gerkens | Belgium | 2009

Rather than only blocking the suns rays, the design re-envisions blinds as sun-soaking solar panels that store energy during the day and illuminate the interior at night. With the revolving blades, the course of the sun is followed in order to catch a maximum of energy and convert the energy in PV cells. The electricity is stored in a battery, and re-emitted at night via electroluminescent foil.

16

[www.core77.com]

Core 77 Greener Gadgets Design Competition (2009) BLIGHT, Website: http://www.core77.com/greenergadgets/entry.php?projectid=41 Cruden, D. (2009) Rethinking Design environmental/ecological design: materials design interior. http://tur-www1.massey.ac.nz/~interior/blog/wp-content/uploads/2009/03/302_09_ecology-booklet.pdf

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Bloomframe Balcony
Hurks Geveltechniek | The Netherlands | 2007-2010

An innovative window frame that can be transformed into a balcony-ondemand: users can add light, air and space to a space with a single push on a button. The dynamic balcony offers an opportunity to add outdoor space to compact apartments, offices or hotels in inner city areas. Winner of the prestigious Red Dot design award 2008 and Wallpaper design award 2009.

17

[Hurks geveltechniek, 2008]

Brownell, B. (2008) Transmaterial 2: A catalog of materials that redefine our physical environment, Princeton Architectural Press Hurks geveltechniek (2008), Bloomframe brochure, http://website.hurks-gdtechniek.nl/beheer/uploads/docs/bloomframe_folder_english.pdf

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

Breathing wall
The Environmental Building Partnership Ltd | United Kingdom

Breathing wall or dynamic insulation panels form the basis for a distributed ventilation air supply system where the wall functions as (i) a supply source, (ii) heat exchanger and (iii) filter for airborne pollutants. The building fabric is used as cross-flow heat exchanger through which the movement of air from outside to inside is controlled. As a consequence, the U-value is a function of the air-flow. The amount of airflow can be both passive, or actively controlled.

18

[www.environmental-building.com]

[www.environmental-building.com]

[www.environmental-building.com]

[Imbabi, 2006]

Imbabi, M. (2006) Modular breathing panels for energy efficient, healthy building construction, Renewable Energy 31(5): 729-738 Taylor, B., Cawthorne, D. and Imabi, S. (1996) Analytical investigation of the steady-state behaviour of dynamic and diffusive building envelopes, Building and Environment 31(6): 519-525

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Bremtex
VOW architecten | Oisterwijk, The Netherlands | 1996

The two storey tall aluminum shutters transform this circular building into a real eye-catcher. While individually adapting in response the sun, the fins take care of an optimized balance between daylight utilization and solar gains. The whole system is controlled via a coupling to the overall building energy management system.

19

[www.vowarchitecten.nl]

[www.vowarchitecten.nl]

[www.architectenweb.nl]

Bremtex, VOW architecten, projecten. Website: http://www.vowarchitecten.nl/architecten/projecten/project03/project03-buiten.htm Schoepenzonwering, architectenweb. http://www.architectenweb.nl/aweb/producten/product_detail.asp?productID=3231

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Burke Brise soleil

Santiago Calatrava | Milwaukee, Wisconsin | 2001

The brise soleil covering a museum pavilion is made up of 72 steel fins, ranging in length from 8 to 32 meter, which makes that its span is comparable to that of a Boeing 747-400. The wings open TuesdaySunday at 10 a.m. with the Museum, close/reopen at noon, and close again with the Museum at 5 p.m. Whenever wind speeds exceed 37 km/h for more than 3 seconds, the wings close automatically.

20

[www.wrightinwisconsin.org]

[www.calatrava.com]

[www.333cn.com]

Milwaukee Art Museum, Santiago Calatrava, Recent projects. Website: http://www.calatrava.com Ataman, O., Rogers, J. and Ilesanmi, A. (2006) Redefining the wall: Architecture, Materials and Macroelectronics, International Journal of architectural computing 4(4): 125 - 136

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MICRO

Chanel Ginza
Peter Marino Associates | Tokyo, Japan | 2004

LED elements are combined with the steel mesh and sandwiched between layers of switchable PRIVA-LITE glass. This innovative system allows the building to take on different appearances during the course of the day. During the day, the electrooptic glass is switched to the transparent or trannsluscent state, but at night, the faade becomes a projection screen for the 700.000 LEDs.

21

[www.quantumglass.com]

Ritter, A. (2007) Smart Materials in architecture, interior architecture and design. Basel: Birkhuser Block, R. and Constantinou, M. (2009) Media faades Interactive brand communication in the urban space http://www.arclighting.de/weblog_01/wp-content/uploads/2009/05/04_03_media-facades_text_version.pdf

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

City of justice
Foster + Partners | Madrid, Spain | 2006-2011

Hexagonal shading units will occupy the central circular atrium as well as the eight peripheral atria of the judicial complex. When extended, the system will cover the entire triangulated roof grid. When retracted, their profile will 'disappear' into the structural profile of the roof. A custom algorithm combining historic solar gain data with real-time lightlevel sensing will control the shading units.

22

[www.hoberman.com]

City of Justice, Madrid, Foster+Partners Projects. Website: http://www.fosterandpartners.com/Projects/1453/Default.aspx Audencia Provincial, Hoberman Associates Works. Website: http://www.hoberman.com/portfolio/audenciaprovincial.php

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

open-loop

MICRO

Climate Adaptive Skin

Bas Hasselaar | TU Delft, The Netherlands | 2005-2009

All functions for heating, cooling, solar control, energy storage, ventilation and electricity generation are integrated in a single skin. The system is universally applicable on every building because it is able to operate independently of building services. Maintenance and servicing is kept as low as possible because the number of moving parts is limited.

23

[Hasselaar et al. 2007-2008]

Hasselaar, B. and Looman, R. (2007) The climate adaptive skin, the integral solution to the conflict of comfort and energy performance, in Proceedings of the CIB World Congress, Cape Town, South Africa. pp: 111 - 1125 Hasselaar, B., van de Spoel, W., Bokel, R. and Cauberg, H. (2008) Simulation of innovative climate control strategies using passive technologies, in Proceedings of The 1st International Conference On Industrialised, Integrated, Intelligent Construction, Loughborough, UK. pp: 239 - 249

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Curtain Wall House

Shigeru Ban Architects | Tokyo, Japan | 1995

The pliable fabric used in the walls construction makes it possible for them to fully retract, channeling more light and air into the core of the house. In winter, the curtains close to retain heat. In cold weather, a set of glazed doors can completely enclose the house for insulation and privacy. The building shell is described as the literal interpretation of a curtain wall faade.

24

[http://www.3dallusions.com]

[www.architetture21.eu]

[http://skellis.net]

[www.architettura21.eu]

Quinn, B. (2006) Textiles in architecture, Architectural Design 76(6): 22-26 Curtain Wall House, Works Shigeru Ban Architects, Website: http://www.shigerubanarchitects.com/SBA_WORKS/SBA_HOUSES/SBA_HOUSES_15/SBA_Houses_15.html

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Da Vinci Rotating Tower

David Fisher | Dubai | 2010-2011

The floor segments rotate independently and individually; one rotation can take place in 90 minutes. The rotating tower is the first skyscraper to be entirely assembled from factory-produced pre-fabricated parts. The technology that will enable the floors to rotate autonomously is triggered according to tenants' wishes via voice commands.

25

[Fisher, 2010]

Fox, M. and Kemp, M. (2009) Interactive Architecture, New York: Princeton Architectural Press Fisher, D. (2010) Dynamic Architecture, Website: http://www.dynamicarchitecture.net

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

DBU conference pavilion

Herzog + Partner | Osnabrck, Germany | 2002

Rotatable blinds and a translucent insulation material together form this functional roof that is covered by a single ETFE layer. Adaptivity of the roof allows for maximum flexibility of interior and furnishing in the windowless conference rooms below. Because thermal insulation and the blinds are sandwiched between polymer layers, these systems are protected from rain and maintenance is minimized.

26

[www.hightexworld.com]

[www.hightexworld.com]

[www.dbu.de]

Jeska, S. (2007) Transparent plastics: design and technology, Basel: Birkhuser Cremers, J. (2007) Multifunctional Membrane Systems for Intelligent Buildings, Intelligent Glass Solutions 1(1): 29-32

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

Deployable External Insulation

Stephen Gage, Bartlett School of Architeture | London, UK | 2008

Deployable External Insulation offers the possibility of actively modulating building fabric thermal performance. The movement of insulative shutters is passively controlled via phase change wax piston actuators. One piston (the cold piston) opens the shutter when the temperature rises above a minimum. A second piston (the hot piston) closes the shutter when the temperature exceeds the maximum threshold.

27

[www.deployable.org.uk]

Deployable External Insulation Project Website: http://www.deployable.org.uk Borden, I. (2009) Sustainability and architectural design, Pallette, UCLs journal of sustainable cities, Summer 2009,

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

Documentation Centre
Axel Ritter | Proposal for Former Concentration Camp | Hinzert, Germany | 2004

The kinetic external cladding acts as shading device and produces electricity via thin film solar cells. The degree of shading is automatically controlled via a thermal expansion material (linear actuator) with PTC thermistors. The expansion material is wrapped with insulation in order to ensure that light is the only controlling stimulus.

28

[Ritter, 2007]

Ritter, A. (2007) Smart Materials in architecture, interior architecture and design. Basel: Birkhuser

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Dr. Jockisch Building

Architecten HBH | Landshut, Germany | 2000

The mobile photovoltaic sunshading system covers 3/8 of the faade and follows the sun while it floats around the building. Motorized wheels support the structure of 182 PV-panels on its way around the building. A sloped solar shield mounted on the roof provides additional electricity generation.

29

[HBH, 2009]

[Inglin, 2007]

[www.architekten-hbh.de]

[www.architekten-hbh.de]

[Inglin, 2007]

Inglin, C. (2007) Sunny side up, with BIPV, Malaysian International Building Exposition (MALBEX) ; http://www.mbipv.net.my/dload/SunnySideUp(PhoenixSolar-ChristopheInglin)-070919.pdf HBH architekten (2009) HBH Broschre mit ausgewhlten Projekten Stand 04/2009, http://www.architekten-hbh.de/fileadmin/img/Download/Buerobroschuere_04-2009_web.pdf

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Eclipsis shutter shade

Virginia Techs entry to US DOE Solar Decathlon competition | 2009

Eclipsis shutter shade is made from stainless steel and a translucent polycarbonate panel filled with aerogel. The shutter shades can slide along the north and south faades, providing protection from direct sunlight while simultaneously allowing for indirect, natural lighting, views to the exterior and privacy to those inside. The homes smart system gathers information from its rooftop weather station to automatically control the screens.

30

[http://lumenhaus.blogspot.com]

[www.flickr.com/afagen]

[www.solardecathlon.org]

Solar Decathlon, Virgina Tech. Website: http://www.solardecathlon.org/2009/team_virginia_tech.cfm Lumenhaus homepage, a brighter way, everyday. http://www.solar.arch.vt.edu/

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Edge Monkey
Stephen A. Gage and Will Thorne | 2005

These imaginary devices are perhaps the buildings environmental control system of the future. Their function would be to patrol building faades by regulating energy consumption and indoor comfort conditions. Basic duties of these autonomous agents include closing unattended windows, checking thermostats, activating movable insulation panels and adjusting blinds.

31

[Sullivan, 2006]

Sullivan, C. (2006) Robo Buildings: Pursuing the interactive building envelope. Architectural Record April 2006:149-156 Lim, C (2006) Devices: a manual of architectural + spatial machines, Architectural Press

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MICRO

Electrochromic glazing
An electrochromic coating enables the glass to switch its optical properties and hence color, while maintaining transparency. The optical properties can be controlled on-demand or via intelligent supervisory control systems. Because of the adaptability, thermal and visual comfort can be improved while energy consumption is minimized.

32

[www.quantumglass.com]

[www.quantumglass.com]

[www.sage-ec.com]

[www.sage-ec.com]

Lee, E., Selkowitz, S., Clear, R., DiBartolomeo, D., Klems, J., Fernandes, L., Ward, G., Inkarojrit, V. and Yazdanian, M. (2006) Advancement of Electrochromic Windows, California Energy Commission, PIER. Publication number CEC-500-2006-052. Quantum Glass Brochure inspiring solutions by Saint-Gobain. http://www.quantumglass.com/download-pdf.php?type=quantumglass

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MICRO

Electrically heated glass

The window can perform three functions: heating, anti-condensation and snow melting. Active insulating glazing with low-emissivity coating and embedded electrodes also improves the thermal distribution in the room. This concept offers highest potential in regions with extreme climates, like Scandinavia, Siberia and Russia.

33

[www.quantumglass.com]

[www.prelco.ca]

[www.quantumglass.com]

Kurnitski, J., Jokisalo, J., Palone, J., Jokiranta, K., Seppnen, O. (2004), Efficiency of electrically heated windows, Energy and Buildings 36(10):1003-1010 Quantum Glass Brochure inspiring solutions by Saint-Gobain. http://www.quantumglass.com/download-pdf.php?type=quantumglass

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

ETFE roofs
ETFE (ethylene tetrafluoroethylene) is increasingly used in buildings because of its lightweight properties, high daylight transmittance and the potential for energy savings. Compared to conventional glazed roofs, ETFE cushions can provide thermal insulation with reduced initial costs and less structural supports. Application of a reflective frit to an inflatable intermediate cushion can dynamically alter the roofs optical properties.

34

[LeCuyer, 2008]

[www.drmm.co.uk]

[LeCuyer, 2008]

LeCuyer, A. (2008) ETFE Technology and Design, Basel: Birkhuser Poirazis, H., Kragh, M. and Hogg, C. (2009) Energy modelling of ETFE in building applications, in Proceedings of Building Simulation 2009, Glasgow, Scotland. pp: 696-703

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

EWE Arena
ASP Architekten | Oldenburg, Germany | 2005

The integrated photovoltaic shading system creates a temporary double skin faade while floating around the building. The kinetic PV-wall provides shading when and where desired, and at the same time maximizes electrical output. Daylight and view to the outside are unobstructed in the rooms that are not exposed to direct sunlight.

35

[www.panoramio.com]

[www.asp-stuttgart.de]

[www.asp-stuttgart.de]

Weller, B. and Jakubetz, S. (2008) Fassaden als Kraftwerke, Deutsches Ingenieurblatt, 2008(4): 24-28 Lling, C. (2009) Energizing Architecture, Berlin: Jovis Mller, H. (2007), Building Integrated Photovoltaics, in Proceedings of PREA Workshops,

Stellenbosch, Dar es Salaam, Kampala,

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Flare faade
WhiteVoid interactive art and design | Berlin, Germany

Acting like a living skin, flare faade allows a building to express, communicate and interact with its environment. The system consists of a number of tiltable metal flake bodies supplemented by individually controllable pneumatic cylinders. By reflecting ambient or direct sunlight, the individual flakes of the FLARE system act like pixels formed by natural light.

36

[www.flare-facade.com]

Klooster, T. (2010) Smart Surfaces and their application in architecture and design, Berlin: Birkhuser Flare kinetic ambient reflection membrane, Website: http://www.flare-facade.com Tscherteu, G. (ed) (2008) Media Faades Catalogue, German Centre for Architecture, http://www.mediaarchitecture.org/wp-content/uploads/2008/11/media_facades_exhibition_companion.pdf

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Fluidized Glass Faade

Fraunhofer Institute for Solar Energy | Germany | 2005

The concept consists of a four layer glazing element with two fluid filled chambers and one gas filled chamber in the middle. Colored water in the external cavity is used for control of solar gains; water in the inner cavity serves as transparent large area heating or cooling element. Heating, cooling and shading levels are regulated by controlling fluid flow levels and water color.

37

[Platzer et al. 2005]

Platzer, W. (2005) Steps towards a vision a multifunctional fluidized glass faade, Intelligent glass solutions 2: 54-56 Platzer, W., Hube, W., Adelmann, W., Crisinell, M. and Vollmar, T. (2005) Development of a glass faade with internal fluid flow, in Proceedings of the International Conference Glass Processing Days 2005. pp 303-307

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Gemini Haus
ArchBro Kaltenegger | Weiz, Austria | 2001

About 40 m of window-area is always orientated towards the sun for maximum passive solar heating. The solar curtain at the front (output 4.0 kWp) rotates uni-axially around the vertical axis of the building and helps to avoid overheating. The PV modules integrated in the faade receive the sunlight via hologram foils, thus enhancing the electrical output.

38

[www.erwin-kaltenegger.at]

ArchBro Kaltenegger, Projekte, Wohnbau. Website: http://www.erwinkaltenegger.at/projekte/wb_ef.php?was=090915_1547 Mit dem Energieberschuss die Hypothek abzahlen, Frankfurter Allgemeine Zetitung, 16 July 2002, page T6

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Gerontology Centre
Siegert Architekten | Bad Tlz, Germany | 2003

The first building in the world that uses a single layer of transparent film, which spans the complete height of the faade. This was the only way to create the double curvature spiral form that covers a multi-story void. During the night, vents in the faade open up which enables effective stackventilation of the heat that is built up during the day.

39

[www.kockmembranen.de]

[www.hightexworld.com]

[www.grp.hwz.uni-muenchen.de]

Jeska, S. (2007) Transparent plastics: design and technology, Basel: Birkhuser Cremers, J. (2007) Multifunctional Membrane Systems for Intelligent Buildings, Intelligent Glass Solutions 1(1): 29-32

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

Girasol
COLT international Ltd. | United Kingdom

Girasol (Italian for sunflower) is a controllable external shading system consisting of glass louvers with photovoltaic cells integrated into the glass. The innovative, autonomous passive sun-tracking system does nor consume any electricity. Two absorber evaporator tubes cause different expansion or contraction of a working fluid. This difference in pressure in turn actuates the thermohydraulic drive mechanism.

40

[www.coltinfo.co.uk]

Janak, M. and Kainberger, R. (2009) Integrated building energy and lighting simulation in the framework of EU PV-LIGHT project, in Proceedings of Building Simulation 2009, Glasgow, Scotland. pp: 1671-1677 Starlinger, M. (2007) Ensuring that louvered shading systems always follow the sun Applied bionics in the building skin Architektur und Sonnenschutz 2007(3)

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

open-loop

MICRO

GlassX crystal
GlassX AG | Zrich, Switzerland

GlassX crystal integrates four functionalities into a single unit: transparent insulation, overheating protection, energy conversion and thermal storage. The key component is the slim transparent PCM heat storage module which has a storage capacity equivalent to about 20 cm concrete. Solar heat is stored in the PCM by means of a melting process. At night, the stored heat is delivered to the interior during re-crystallization.

41

[www.glassx.ch]

Ritter, A. (2007) Smart Materials in architecture, interior architecture and design. Basel: Birkhuser GlassX Speichern, Wrmen, Kuhlen, Website: http://www.glassx.ch

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MICRO

GreenPIX
Simone Giostra & Partners | Beijing, China | 2008

The worlds largest display area (2200 m2 - 2292 LEDs) combines sustainability with digital media technology. The system is also called Zero Energy Media Wall since no external power is required to fuel the diodes. In addition, the photovoltaic panes provide a means for both natural ventilation and solar shading.

42

[www.greenpix.org]

Sinclair Eakin, J. (2007) A glam in the eye China makes room for an energy-efficient media wall, I.D. The International Design Magazine 54(1): 47-48 Klooster, T. (2010) Smart Surfaces and their application in architecture and design, Berlin: Birkhuser GreenPix, Press release http://www.greenpix.org/press/PDF/GreenPix_press-kit_EN.pdf

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

GROW
SMIT | New York | 2005-2010

GROW is a hybrid energy delivery device that provides power via the sun and wind. The solar leaves act as micro-power stations by employing thin film photovoltaics with piezoelectric generators and screen printed conductive ink encapsulated in ETFE fluoropolymer laminates.

43

[www.S-M-I-T.com]

Klooster, T. (2010) Smart Surfaces and their application in architecture and design, Berlin: Birkhuser SMIT - Sustainably Minded Interactive Technology, Webstite: http://www.S-M-I-T.com Cochran, T. (2008) Sustainable Minded Interactive Technology (SMIT) Design, in Proceedings of Sustainable design: at the crossroads of art, policy and science, Boston, MA

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Habitat 2020
Philips Design Probe | 2009

This concept is exploring the possibility of using sensitive textile skins on buildings to create energy independent structures. Electronics and bio-chemical functionalities are being introduced in the building shell. The membrane is used as transporter for collecting and channeling air, water and light, from the outside feeding into the inside space.

44

[www.design.philips.com]

Off the grid: Sustainable Habitat 2020. Website: http://www.design.philips.com/probes/projects/sustainable_habitat_2020/index.page Kim, H. and Stephens, B. (2009) A conceptual tool for critical interaction design for sustainability, In Proceeding of International Conference on Designing Pleasurable Products and Interfaces, Compiegne, France.

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Heliotrop
Rolf Disch | Freiburg, Germany | 1993

The building has been named after heliotropism which refers to plants that grow in response to the stimulus of the sun. Via a powerful engine, the building is able to vary its orientation in response to the sun. One-half of the building is highly glazed, the other side is well insulated. A central control system decides which side will be exposed to the sun.

45

[www.rolfdisch.de]

Wigginton, M. and Harris, J. (2002) Intelligent Skins, Oxford: Butterword-Heinemann Heliotrope in the spotllights of success, Renewable Energy Journal 2000(10), http://ec.europa.eu/energy/res/publications/doc1/rej_10.pdf Heliotrop, Robert Disch, Projects, Website: http://www.rolfdisch.de/project.asp?id=45&sid=-1402147115

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Homo Lumens MMVII

Lszl Benczr, Pter Sugr, Lszl Kara | Budapest, Hungary | 2007

The environment-responsive kinetic glass faade consists of an array of transparent lamellas and panels of six LEDs embedded in the window sills. Each lamella responds to temperature and wind signals from roof-mounted weather sensors. The whole faade moves slowly when the weather is quiet, and faster when it is windy. Color patterns and shades of each hue are programmed to change subtly with time and temperature.

46

[http://lanchid19.blogspot.com]

Homo Lumens MMVII blogspot, http://lanchid19.blogspot.com/ Dumiak, M. (2008) Homo Lumens moves, inviting many interpretations, Architectural Record may 2008

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MICRO

Hotel habitat
Ruiz-Geli Cloud9 | Barcelona, Spain | 2006-2010

Because the envelope acts as thermometer, it gives a special aesthetic appearance to the building. Each node of the steel mesh consists of one photovoltaic cell, a battery, CPU and a RGB LED which creates an artificial light cloud around the hotel. The light mesh has sensors that will read the daylight sun amplitude and then at night each node will give off color according to how much that node had absorbed.

47

[www.ruiz-geli.com]

Hotel Habitat, Ruiz-Geli, Website: http://www.ruiz-geli.com/download/project.pdf Interactive Architecture (2006) LED light mesh facade James Clar & Cloud9, http://www.interactivearchitecture.org/led-light-mesh-facade-cloud9-james-clar.html

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

iHomeLab
Lischer Partner Architekten| Lucerne, Switzerland | 2008

Every other vertical louver can be forced to curve, and in that way control the level of daylight and solar gains entering the building. During its operation, the lamellae very much act like the gills of a fish. At night, the membrane changes color when and where people approach the building.

48

[http://www.lischer-partner.ch]

[http://www.lischer-partner.ch]

[Boog, 2009]

Boog, P. (2009) Es bewegt sich was! Fassade / Faade 2009(2): 39 - 42 Coltpost (2009) Colt Group Limited. http://www.coltgroup.com/view.aspx?/aboutcolt/colt_post_2009.pdf

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

IC solar faade system

Rensselaer Polytechnic Institute | Troy, New York | 2007-2010

Solar concentrating tracking technology is transferred to a day-lighting system within a double-skin faade. The lens directly concentrates the light (>400:1) onto a high-efficiency multijunction PV cell that recently demonstrated an efficiency of 39.4%. Absorbed power that is not converted to electricity is captured via a coolant flow and is used for domestic hot water generation and space heating.

49

[www.case.rpi.edu]

Dyson, A., Jensen, M. and Stark, P. (2007), Integrated Concentrating (IC) Solar Faade System, DOE Solar Energy Technologies Program Review Meeting RPI, Center for Architecture, Science and Ecology, Website: http://www.case.rpi.edu/projects/ICsolar.html

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Kameleon Concept
Sublean | Sliedrecht, The Netherlands | 2009

The concept consists of a rotatable aluminium box girder with four different surfaces and replaceable coffers. Depending on the prevailing weather, the faade can perform different functions: e.g., energy generation, heat storage, air purification, advertising and rainwater drainage. The system is designed to create maximum value, 24 hours a day.

50

[www.sublean.nl]

Beerda, E (2010) Kameleontisch oppervlak werkt verduurzamend Cobouw 15 January 2010, pp: 3/4 Submission to the competition: Het ei van Columbus, Kameleon concept, Website: http://www.ei-van-columbus.nl/inzendingen/inzending/419

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MICRO

Kinetic light
Christian Mller and Rdiger Kramm | Frankfurt, Germany | 1992

Behind a screen of perforated aluminum, 120 floodlights fade from blue to yellow, illuminating the front of the building. The overall image is directed by a weather station on top of the building: The ambient temperature determines the amount of yellow on the blue wall; The yellow patches move in line with the direction of the wind and rain; LED lines in the upper area of the faade visualize noise in the street in real-time.

51

[www.christian-moeller.com]

Christian Moeller, works, Kinetic Light Scuplture, Website: http://www.christian-moeller.com/display.php?project_id=30 Interactive Architecture (2005) Kinetic Light Sculpture of the Zeilgalerie, Website: http://www.interactivearchitecture.org/kinetic-light-sculpture-of-the-zeilgalerie.html

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Kinetic roof Medina house

MIT | Entry for Medina 3rd annual design competition | 2001

Because the sun changes position, the roof keeps moving like a living structure. Winter day: the kinetic structure opens to the sun and its warm light. Winter night: the roof closes to keep the heat inside. Summer day: the roof denies the sun while facilitating cross ventilation. Summer night: the kinetic roof opens for night ventilation.

52

[Magnoli et al., 2001]

Abdel Kawi, A. (2001) Medina Third Annual Design Competition: Youth ideas competition, Smart Village, Egypt, Medina 19: 50 - 63 Magnoli, C., Bonanni, L. Khalaf, R. and Fox, M. (2001) Designing a DNA for responsive architecture: a new built environment for social sustainability, MIT Media Lab Development by design Workshop

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Le mur neutralisant
Le Corbusier | 1929

Le mur neutralisant is often considered as the revolutionary predecessor of the double skin faade. It consists of two glass panels within which tempered air is flowing at a constant temperature of 18 degrees Celsius. This air layer neutralizes the effects of the cold downdraught in winter or keeps solar heat outside in summer, thereby enhancing thermal comfort conditions throughout the whole year.

53

[Knaack et al., 2007]

[www.quintademosteiro.com]

[www.quintademosteiro.com]

Banham, R. (1984) The architecture of the well-tempered environment, University of Chicago Press Knaack, U., Klein, T., Bilow, M. and Auer, T. (2007) Faades - Principles of construction, Basel, Boston, Berlin: Birkhuser

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Living glass
The Living | New York | 2005

Elastic shape memory alloy (SMA) wires control the level of carbon dioxide in a room. High CO2 concentrations force SMA wires to contract and pull open slits etched in the window to allow fresh air to flow in. Inspired by the gill, the respiratory organ that controls the absorption of oxygen and the secretion of carbon dioxide in most aquatic organisms.

54

[www.thelivingnewyork.com]

Manfra, L. (2006) Envisioning architecture that performs like natural organisms, Metropolis 25(5): 52-54 Living glass: What if architecture responded to you? http://www.thelivingnewyork.com/lg.htm

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

Living walls and roofs

These self-sufficient vertical or horizontal surfaces restore the level of flora and fauna in the urban context. The systems add mass and thermal insulation to the building shell and are also useful for controlling rainwater runoff. The plants can help to reduce overall temperature of building surfaces via evaporative cooling which reduces energy consumption

55

[www.massstudies.com]

[http://rpbw.r.ui-pro.com]

[www.verticalgardenpatrickblanc.com]

Blanc, P. (2008) The vertical garden: from nature to the city , New York: Norton Leenhardt, J. and Lambertini, A. (2007) Vertical gardens : bringing the city to life, London: Thames & Hudson

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Lhemisferic
Santiago Calatrava | Valencia, Spain | 1998

A large shelter over a spherical IMAX theatre that opens up in the same way as a retractable eyelid. The aluminum awnings can fold up individually or collectively over the full breadth (90 meter) of the building. The system is actuated via telescoping hydraulic cylinders which allow fresh breezes to enter the inside.

56

[www.cac.es]

[www.britannica.com]

[www.cac.es]

[Calatrava Sketchbook]

City of sciences and arts in Valencia, Santiago Calatrava, Recent projects, Website: http://www.calatrava.com Ataman, O., Rogers, J. and Ilesanmi, A. (2006) Redefining the wall: Architecture, Materials and Macroelectronics, International Journal of architectural computing 4(4): 125 - 136

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Liquid faade
Ulrich Knaack | TU Delft, The Netherlands | 2006

Water is used for its thermal mass and energy storing capabilities and the system also has a load bearing capacity. Insulation values can be controlled pneumatically by changing the position of the water-filled chamber. In future generations the water can be used for collection of thermal energy to supply heating and cooling to the building systems.

57

[Knaack et al., 2008]

Knaack, U., Klein, T. and, Bilow, M (2008) Imagine 01 Faades, Rotterdam: 010 Publishers

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Metal shutter houses

Shigeru Ban Architects | New York, USA | 2007-2010

The building can literally become a uniform minimal closed cube, or it can open completely. Perforated metal shutters on the outside can modulate incoming light and acts as privacy screen. Additionally, sliding glazed doors on the inside can be opened to further blur the boundary between inside and outside.

58

[www.metalshutterhouses.com]

Metal Shutter Houses, Shigeru Ban West Chelsea, Website: http://www.metalshutterhouses.com/ Eyes wide open - Shigeru Bans "metal shutter houses" under construction in Manhattan, World Architecture News: editorial 28 October 2007

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MICRO

Microblind
Boris Lamontagne and Christophe Py | National Research Council of Canada

An array of overhanging pre-stressed invisible microblinds with a size of 100 micrometer covers the window. Curling of microblind-electrodes is electrostatically activated. Arrays can be deposited on the flat glass by magnetron sputtering like regular low-E coatings, and then patterned by laser. The microblinds have several advantages including switching speed (milliseconds), UV durability, customized appearance and transmission.

59

[Lamontagne and Py, 2006]

Lamontagne, B. and Py, C. (2006) Microblinds and a method of fabrication thereof, US Patent Application 20060196613 Lamontagne, B. (2009) Next generation of smart windows based on micro-blinds (MEMS) and Silicon photonics at IMS-NRC: from bio-sensing to space applications, Lecture in Mininova, Helsinky University of Technology

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

Nano Vent Skin

Augustin Otegui | 2008

This conceptual project tries to make (existing) objects greener with a skin made out of micro windturbines. Kinetic energy of the wind is converted into electricity and at the same time, CO2 is absorbed via organisms inside the turbines. The outer skin of the structure absorbs sunlight through an organic photovoltaic skin.

60

[Otegui, 2008]

Otegui, A (2008) Nano Vent Skin weblog. http://nanoventskin.blogspot.com/ May Young, N. (2008) Living skin could be the future in sustainable design, World Architecture News: editorial June 2008

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

Nocturnal ventilator
Nick Browne | Bartlett School of Architecture | London, UK | 2005

The system stores thermally induced movement of a wax piston over a days temperature change in a spring mechanism. The spring is then released during the night via an ingenious mechanism, and this creates a breeze that lasts for hours. Arrays of ventilators can be retrofitted to existing dwellings, and provide energy-free localized ventilation.

61

[Gage, 2008]

[Lim, 2006]

Gage, S. (2008) The wonder of trivial machines, System Design and behavioral science 26(3):771-778 Lim, C (2006) Devices: a manual of architectural + spatial machines, Architectural Press

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Paper art museum

Shigeru Ban Architects| Shizuoka, Japan| 2002

The faade of the museum consists of translucent glass-fiber-reinforced panels that can be pivoted over their whole length. By opening the stacking shutters and awnings (shitomido), a spatial continuity between the interior and exterior is achieved. This enables outside air to freely flow into the building while at the same time protecting the building from overheating via solar shading.

62

[www.shigerubanarchitects.com]

Jeska, S. (2007) Transparent plastics: design and technology, Basel: Birkhuser Paper art museum, Works Shigeru Ban Architects, Website: http://www.shigerubanarchitects.com/SBA_WORKS/SBA_OTHERS/SBA_OTHERS_9/SBA_Others_9.html

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Pfalzkeller
Santiago Calatrava | St. Gallen, Switzerland | 1998

Calatrava designed a symmetrical system, covering the glazed roof of a largely underground building. Two arched tubes position the articulated slats on each side of the roof, and in this way dynamically control daylight. The system was the first application of a movable roof cover based on slats in combination with a mechanical hoist system.

63

[www.calatrava.com]

Pfalzkeller Emergency Service Center, Santiago Calatrava, Past projects, Website: http://www.calatrava.com Santia Calatrava St Gallen, Kinetic Architecture blog. http://kineticarchitecture.blogspot.com/2008/05/santiago-calatrava-stgallen_27.html

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Phantom house
Diller Scofidio + Renfro | 2007

The roof consists of a deployable shelter that can serve two functions: 1) collecting rainwater and 2) providing sunshading. Water sensors on the roof double the areas able to catch rainwater when necessary and store the water for use in toilets and the garden. The house has a dynamic overhang which allows valuable solar gains in the winter but avoids overheating in summer.

64

[Diller et al., 2007]

Jodidio, P. (2009) Green Architecture Now!, Kln, Germany: Taschen Diller, E., Scofidio, R., Renfro, C., Baccarini, C., Donnelly, R. and Gesauldi, F. (2007) An eco-house for the future, The New York Times Magazine, May 20, 2007

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

open-loop

MICRO

Photochromics
These materials reversibly change their color when exited by the effect of light (electromagnetic energy). A common application is its use in self-adjusting sunglasses, but architectural implementation is also pursued for a long time. Relatively high costs, sensitivity to heat and poor long-term behaviour have prevented large scale use of these systems.

65

[www.photochromicwindows.com]

[www.photochromicwindows.com]

[www.x-celoptical.com]

Ritter, A. (2007) Smart Materials in architecture, interior architecture and design. Basel: Birkhuser Smith, G. (1966) Chameleon in the sun: photochromic glass, IEEE Specctrum 3(12): 39-47 Fernndez, J. (2007) Materials for Aesthetic, Energy-Efficient, and Self-Diagnostic Buildings, Science 315(5820): 1807 - 1810

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MICRO

Pixelskin01
Sachin Anshuman | Orangevoid, United Kingdom | 2006

The system consists of a combination of electrochromic glass and ultra-bright electroluminescent tubes controlled by a distributed network of microcomputers and sensor consoles. The architectural surface is designed to switch between a conventional window and a display/advertising/communication screen. Pyroelectric sensors on each disk are used for controlling transparency.

66

[Anshuman, 2006]

Brownell, B. (2008) Transmaterial 2: A catalog of materials that redefine our physical environment, Princeton Architectural Press Anshuman, S. (2006) Pixelskin01, Website: http://www.orangevoid.org.uk

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Pixelskin02
Sachin Anshuman | Orangevoid, United Kingdom | 2006

A matrix of interconnected pixel tiles that are controlled interactively via embedded controllers. The system utilizes two-state material SMA wires for actuating each pixel in one of 255 potential states. The surface can also be used to generate dynamic low resolution images.

67

[Anshuman, 2006]

Brownell, B. (2008) Transmaterial 2: A catalog of materials that redefine our physical environment, Princeton Architectural Press Anshuman, S. (2006) Pixelskin02, Website: http://www.orangevoid.org.uk

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MICRO

PV honeycomb glass faade

Daisuke Nagatomo and Minnie Jan | Japan | 2006

Display skin and structural design are integrated in Honeycomb system. The energy generated by photovoltaic panels will be used to illuminate highefficiency LED lights at night. The system mimics nature: honeycomb structure has durability, like Radiolaria, and optical efficiency, like flys eyes.

68

[www.core77.com]

Photovoltaic Honeycomb Glass Screen Faade (2006) Light Objects, A design competition about sustainability, Website: http://www.core77.com/lightobjects/img/1474/default.asp bright LED (2007) a design competition organized by designboom, Website: http://www.designboom.com/contest/view.php?contest_pk=19&item_pk=13925&p=1

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Responsive kinetic faade

SOM and Sci-Arc | 2008

A composition of hinged glass panes and spandrel panels laminated with solar cells. The flexible skin is capable of changing its shape, controlled via magnetic switches that are powered via the PV cells. The building skin is considered as a pump: depending on the season, air is supplied to or extracted from the room.

69

[www.core.form-ula.com]

SOM + SCI-Arc on CF:Responsive Kinetic Faade, CORE.FORM-ULA 15 April 2009. Website: http://www.core.form-ula.com/2009/04/15/som-sci-arc-on-cfresponsive-kineticfacade/

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Roof pond Hammond house

Jonathan Hammond | Winters, California | 1975

Large water storing bags are covered by insulating hinged lids which are opened and closed by a hydraulic ram. When the panels are raised on sunny winter days, water bags are exposed to the sun, while the panels act as reflectors to enhance absorption of solar radiation. In summer, the panels are raised at night. Absorbed warmth from the inside is released to the cool night sky, via radiation and convection.

70

[Bear, 2009]

[Bear, 2009]

Baer, S. (2009) Some passive solar buildings with a focus on projects in New Mexico, Presentation for the Albuquerque chapter of AIA, January 15, 2009 Anderson, B. and Michal, C. (1978) Passive Solar Design, Annual Review of Energy 3(1):57-100

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Sagami bay house

Foster + Partners | Tokyo, Japan | 1987- 1992

This modern dwelling is designed according to the high transparency guidelines of traditional Japanese architecture. The architects brought even more light into the main house by conceiving the roof as a fifth faade of glass. The roof creates a theatre of shifting light inside, adding another dimension to the variety of light admitted by the perimeter of glass.

71

[www.fosterandpartners.com]

House in Japan, Tokyo, Japan, Foster+Partners Projects. Website: http://www.fosterandpartners.com/Projects/0411/Default.aspx Giovannini, J. (2008) A Japanese Modernism Fusing tradition and progression in a house on the volcanic coastline of Sagami Bay Architectural Digest March 2008

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

Self-regulating ventilation
Via a combination of pivoting and bending mechanisms, the ventilation grille installed in the window frame becomes self-regulating. These vents can achieve stable ventilation levels in spite of the highly irregular variations in natural driven conditions. The system also allows for users intervention via a manually controlled ventilation flap.

72

[www.renson.be]

Breesch H., Descheermaeker K., Janssens A., Willems L. (2004) Evaluation of natural ventilation systems in a landscaped office, Proceedings of PLEA2004, Eindhoven: The Netherlands; pp:1-6 Axley, J., Emmerich, S. and Walton, G. (2002) Modeling the Performance of a Naturally Ventilated Commercial Building With a Multizone Coupled Thermal/Airflow Simulation Tool. ASHRAE Transactions 108(2)

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Showroom Kiefer technic

Ernst Giselbrecht + Partner | Bad Gleichenberg, Austria | 2007

The shades of this dynamic faade can be adapted individually to changing conditions and needs. The faade changes continuously; each day and each hour shows a new face - the faade is turned into a dynamic sculpture. A video clip featuring a choreography dedicated to this dancing faade already attracted more than 160.000 viewers on YouTube.

73

[www.giselbrecht.at]

Showroom Kiefer technic, Giselbrecht + Partner projekte, Website: http://www.giselbrecht.at/sect01_projekte/gewerbe_industriebauten/kiefer/frames_unten.html e-architect, Kiefer Technic Showroom: Architecture Information, Website: http://www.e-architect.co.uk/austria/kiefer_technic_showroom.htm

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Skytherm solar roof pond

Harald Hay | Atascadero, California | 1973

The system consists of a large mass of water, packed in containers and placed on top of a metallic corrugated ceiling deck. The flat roof also holds an insulation panel that moves on rails to cover and uncover the water to the atmosphere. Depending on season and time of the day/night, the roof pond system can act as source of heating and cooling to the zone below.

74

Heating cycle

Cooling cycle

[Gut and Ackerknecht, 1993]

Gut, P. and Ackerknecht, D. (1993) Climate Responsive Building, St. Gallen: SKAT Givoni, B. (1979) Passive cooling of buildings by using natural energies Energy and Buildings 2(4): 279-285

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MICRO

Smart Energy Glass

Peer+ | Eindhoven, The Netherlands

An innovative coating allows the window to switch its optical properties in three different states: dark, bright and privacy. Part of the sunlight is captured and converted into electricity in PV-cells mounted at the edges of the window. When equipped with batteries, the system becomes autonomous and can be integrated in the building envelope without the need for external wires.

75

[www.peerplus.nl]

Peer+ Smart Energy Glass, Website: http://www.peerplus.nl Kemper, A. (2008) Schone energie dankzij superslimme ruiten Volkskrant 14 oktober 2008

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

SmartScreen
Martina Decker and Peter Yeadon | New York

The shading device is automatically activated by changes in local interior air temperature. The R-Phase SMA (shape memory alloy) smart material simultaneously acts as both a sensor and a motor. Adaptation is triggered automatically and consequently does not consume any fuel or electricity.

76

[Decker and Yeadon, 2008]

Decker, M. and Yeadon, P (2008) SmartScreen: Controlling solar heat gain with shape memory systems, Boston Society of Architects Design Research Grants Morell, D. (2007) The role of memory in architecture and materials, MetropolisMag.com

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

Smart shade
Lance Hosey | ATMO / Atelier Modern | Washington DC | 2005

A passive solar shading device that automatically adjusts to changing light and heat conditions. Smart shade employs the thermodynamics of zinc and steel to control the amount of sunlight passing into a building's interior. Expansion and contraction of these sandwiched materials in response to temperature cause the blinds to curl up in winter (allowing more sunlight in) and curve down in summer (allowing less).

77

[Manfra, 2006]

Manfra, L. (2006) Envisioning architecture that performs like natural organisms, Metropolis 25(5): 52-54 Application design Interaction design, Robotecture lecture, http://robotecture.com/spaceelevatorbase/lectures/applicationdesign.pdf

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

open-loop

MICRO

SmartWrapTM
Kieran Timberlake Associates | Philadelphia, US | 2003

Rolled and printed onto fabrics and plastics to fulfill the following functions: shelter, climate control, lighting, information display and power supply. The construction material features organic photovoltaics, thin film batteries and OLEDs, all inspired by the printing industry. The inner skin supports thermal storage and insulation via pockets filled with phase change material and aerogel.

78

[http://kierantimberlake.com]

Ritter, A. (2007) Smart Materials in architecture, interior architecture and design. Basel: Birkhuser SmartWrapTM: The Building Envelope of the future, http://kierantimberlake.com/research/smartwrap_research_8.html#

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Solar barrel wall

Steve Baer | Corrales, New Mexico | 1971

During the day, thermal energy from the sun is collected and stored in a stack of 200 liter water-filled steel oil drums. At night, the external lids are closed while the internal lid is opened, so the stored energy is discharged to the room. The ends of the barrels are painted black, and the shutter is made highlyreflective; both effects together improve storage of thermal energy.

79

[Bear, 2009]

[Knaack, 2007]

[Bear, 2009]

Knaack, U., Klein, T. Bilow, M. and Auer, T. (2007) Faades Principles of construction, Basel: Birkhuser Baer, S. (2009) Some passive solar buildings with a focus on projects in New Mexico, Presentation for the Albuquerque chapter of AIA, January 15, 2009

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

SolaRoof
Harvey Rayner | Suffolk, UK | 2005

Bubbles in the transparent building cavity can provide insulation or shade as required. A liquid film instead of bubbles acts as solar energy harvester or liquid cooling film. The generation and movement of soap bubbles and the rainbow color reflection of soap bubbles provides a fascinating and charming atmosphere.

80

[www.solarbubblebuild.com]

Han, H.J., Jeon, Y.I., Lim, S.H., Kim, W.W. and Chen, K. (2010) New developments in illumination, heating and cooling technologies for energy-efficient buildings, Energy 35(6): 2647-2653 SolarBubbleBuild SolaRoof technology, Website: http://www.solarbubblebuild.com/

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

Solar Ivy
S-M-I-T | New York | 2005 - 2009

A system of solar powered panels disguised to look like ivy leaves and made out of 100% recyclable polymer solar cells. Solar Ivy's unique visual appeal and flexibility brings a technology traditionally restricted to the roof to almost any architectural surface. It has the ability to provide varying degrees of opacity to modulate heat gain, actuated by the vagaries of the wind.

81

[www.solarivy.com]

Klooster, T. (2010) Smart Surfaces and their application in architecture and design, Berlin: Birkhuser Solar Ivy, Website: http://www.solarivy.com Muis, R. and Griffioen, R. (2009) Gevels begroeid met zonnecellen Architectenweek Magazine 29: 84

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

SolVent window system

The system consists of an innovative reversible frame, incorporating two glazing assemblies: a clear glazing to provide a weatherproof seal, and an absorptive glazing to provide solar control. The two glazing assemblies and the ventilated channel between them rotate together through 180 to enable the transformation from winter mode to summer mode. In winter, the glazing system allows passive solar space heating. In summer, penetration of unwanted radiation is reduced, thus reducing overheating and visual glare. External shading devices are made unnecessary.

82

[Erell et al, 2004]

[Leal et al, 2004]

[Erell et al, 2004]

Erell, E., Etzion, Y., Carlstrom, N., Sandberg, M., Molina, J., Maestre, I., Maldonado, E., Leal, V. and Gutschker, O. (2004) SolVent: development of a reversible solar-screen glazing system, Energy and Buildings 36(5): 467-480 Leal, V., Erell, E., Maldonado, E. and Etzion, Y. (2004) Modelling the SolVent ventilated window for whole building simulation, Building Services Engineering Research and Technology 25(3): 183-195

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Sonomorph
Natasa Sljivancanin | 2007

Sonomorph is a kinetic building system that changes its behavior in response to various acoustic stimuli. It consist of cellular components with aluminum outer panels and glassreinforced plastic inner panels. By opening or closing these panels, the faade can change from a sound reflective to a sound absorptive state.

83

[http://bifokal.no]

Brownell, B. (2008) Transmaterial 2: A catalog of materials that redefine our physical environment, Princeton Architectural Press Bifokal Projects, Sonomorph: interactive absorbant. http://bifokal.no/project/architecture/

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

Stomata Inspired Skin

Michael Murauer | Austria, Vienna | 2004

A functional biomimetic translation of stomata, the openings that control the gas exchange in the leaves of plants. At high relative humidity, the blue Cithosan polymer swells and separates the two layers of fabric, which increases airflow through the faade. The selfregulating smart material provides distributed ventilation and enhances the quality of the internal climate.

84

[Gruber, 2008]

Gruber, P. (2008), The signs of life in architecture, Bioinspiration & Biomimetics 3(2): 1-9

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Sunbender
ZomeWorks Passive Energy Products | New Mexico, USA

The key to the Sunbenders performance is the adjustable reflector that can be mounted to every existing skylight. In the lowered summer position, it shades the skylight and greatly reduces heat gain, while still allowing diffuse light to enter. Sunbender allows collection of desirable heat during cold weather months, the curved shape of the reflector concentrates the reflected sunlight.

85

[http://zomeworks.com]

Zomeworks Passive Energy Products, Sunbender Brochure http://zomeworks.com/files/sunbenderTM/Sunbender%20Brochure%202009.pdf Baer, S. (2009) Some passive solar buildings with a focus on projects in New Mexico, Presentation for the Albuquerque chapter of AIA, January 15, 2009

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Sndreyer
SNDREYER GmbH | Treia, Germany | 2005 - 2010

The round buildings and turnable roofs ensure optimal utilization of solar radiation. The heliotropic horizontal sun-tracking roof produces a 25% surplus of electrical energy compared to a static configuration. A central electro-hydraulic unit drives the system via small wheels on rails.

86

[www.suendreyer.de]

Sndreyer GmbH, Website: http://www.suendreyer.de Sndreyer Drehbares dach, Kinetic Architecture blog. Website:

http://kineticarchitecture.blogspot.com/2008/05/sundreyer-drehbares-dach_06.html

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

Super Cilia Skin

MIT Media Lab | Massachusetts, USA | 2003

Super Cilia Skin (SCS) is basically a surface for displaying haptic and visual information but is also envisioned as dynamic cladding material. Electromagnetic energy generators make it possible to harness energy from the wind blowing over the buildings exterior. SCS can be both a billboard-like ambient display surface and a source for alternative energy that is also visually appealing.

87

[Raffle et al., 2003-2004]

Raffle, H., Joachim, M.W. and Tichenor, J. (2003) Super Cilia Skin: An Interactive Membrane. In Proceedings of the Conference on Human Factors in Computing Systems (CHI'03), Ft. Lauderdale, FL, USA. Raffle, H., Tichenor, J. and Ishii, H. (2004) Super Cilia skin, A Textural Interface, Textile, The journal of cloth and culture 2(3): 238-247

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

Swindow
Tateyama aluminium inc. | Japan | 2007

The basic configuration consists of a horizontally pivoted window that is hinged just above mid-height. The window starts moving when the wind blows, hence, the natural ventilation rate is a function of wind speed and direction. Shock absorbers and a balancing weight ensure calm operation and a steady ventilation flow rate.

88

[Annex 44]

[http://buildingsash.net]

[www.nav-window21.net]

[Annex 44]

Aschehoug, . and Andresen, I. (2008) State-of-the-art review, IEA-ECBCS Annex 44: Integrating environmentally responsive elements in buildings Lee, SW., Kato, S., Takahashi, T., Chang H, and Murakami, S. (2000) Field Test on Natural Ventilation Performance of Swinging Windows in a Building, in Proceedings of the Meeting of the Society of Heating, Air-Conditioning and Sanitary Engineering, pp 261-264

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MICRO

Switchable insulation
ZEA Bayern | 1999-2001

The system is based on the controllable release and re-absorption of hydrogen in metal hydride, which changes thermal conductivity. Thermal conductivity can reversibly be changed by a factor 40, after applying a small amount of electricity. Passive solar heat is transferred to the inside, but only at times when this is desired and advantageous.

89

[www.zae.uni-wuerzburg.de]

[Horn et al., 2000]

Horn, R., Neusinger, R., Meister, M., Hetfleisch, J., Caps, R. and Fricke, J. (2000) Switchable thermal insulation: results of computer simulations for optimization in buildings, High Temperatures High pressures, 32(6) 669-675 ZAE Bayern projects, Switchable Insulation for Utilizing Solar Energy, Website: http://www.zae.uniwuerzburg.de/english/division-2/projects/archive/si.html

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

The B&W house

Universidad Politcnica de Madrid| US DOE Solar Decathlon competition | 2009

The Roof is self-orientating along the day, following the sun. Moving the PV roof gives 10.3% extra power compared to a horizontal fixed roof. The movable faades are clad with semi-transparent PV-cells. These faade panels are also oriented towards the sun in order to maximize the capture of solar energy.

90

[Universidad Politcnica de Madrid]

Solar Decathlon, Team Spain, Universidad Politcnica de Madrid. Website: http://www.solardecathlon.org/2009/team_spain.cfm UPM Solar Decathlon 2009, Website: http://www.solardecathlon.upm.es/

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

The esplanade
DP Architects and Michael Wilford & Partners | Singapore | 2002

Movable triangular sunshades admit diffuse daylight while at the same time excluding direct sunlight. The shapes and sizes of each of the 7139 shutters were determined on the basis of solar optimization studies. Provides the building with climate control mechanism, much as an animal can raise the hair on its back.

91

[http://mikebm.wordpress.com]

[http://app.mica.gov.sg/]

[http://keropokman.blogspot.com]

Aldersey-Williams, H. (2004) Towards biomimetic architecture, Nature / materials 3(5):277-279 The realisation of complex roof geometries, Roof & Faade, Asia, March 2004.

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

open-loop

MICRO

Thermochromics
Thermal energy (heat) to the material alters its molecular structure. Which in turn changes spectral reflectivity of the thermochromic material. The smart material found its way in architecture via applications in switchable windows, concrete and wallpaper. Thermochromic windows become opaque in warm environment. Unfortunately, the system does not allow for manual override.

92

[www.shiyuan.co.uk]

[Nitz and Hartwig, 2005]

[Nitz and Hartwig, 2005]

Arutjunjan, R. (2003) Thermochromic Glazing for Zero Net Energy House, in Proceedings of the International Conference Glass Processing Days 2003. pp 299-301 Nitz, P. and Hartwig, H. (2005) Solar control with thermotropic layers, Solar Energy 79(6): 573-282

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

The Sliding House

dRMM architects | Suffolk, United Kingdom | 2008

A 20 ton mobile roof-and-wall enclosure traverses the site on recessed railway tracks, and in this way gives the building envelope its adaptive properties. Sliding house offers radically variable spaces, and also provides different levels of shelter, insulation and ingress of daylight throughput the day. Winner of: RIBA 2009 award, Grand design award 2009 and WAF09 award.

93

[Russel, 2010]

May Young, N. (2008) Extraordinary design concept veiled within a seemingly ordinary exterior, World Architecture News: editorial March 2008 Russel, R. (2010) Sliding house, Website: http://www.therussellhouse.org/ dRMM projects, Sliding house, Website: http://www.drmm.co.uk/dRMM/projects/by%20name/sliding%20house/sliding%20house.pdf

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Turning the place over

Richard Wilson | Liverpool, United Kingdom | 2007

This piece of public artwork is a slowly rotating circle of the faade itself, offering recurrent glimpses of the interior during its constant cycle during daylight hours. The oval section cuts back into the building and veers out over the pavement as it goes round. The revolving faade rests on a specially designed giant rotator, usually used in the shipping and nuclear industries.

94

[Croft, 2007]

Future of cities: When does public art become plop art, Guardian, 1 October 2008.

http://www.guardian.co.uk/society/2008/oct/01/regeneration.public.art

Croft, C. (2007) Sculptor Richard Wilson and Price & Myers turn a Liverpool office faade inside out, Building Design Magazine Envelope, July 2007

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

open-loop

MICRO

Wall integrated PCM

Encapsulated latent heat storage material is incorporated in the construction elements of the building. Heat or cold is stored and automatically released when indoor or outdoor temperature rises or falls beyond the melting point. Phase change materials also have the ability to dampen temperature fluctuations without consuming energy.

95

[www.basf.com]

Baetens, R., Jelle, B. and Gustavsen, A. (2010) Phase change materials for building applications: A state-of-the-art review, Energy and Buildings 42(9): 1361-1368 Tyagi, V and Buddhi, D. (2007) PCM storage in buildings: A state of art, Renewable and Sustainable Energy Reviews 11(6): 1146-1166

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Wall street ferry terminal

Smith-Miller + Hawkinson Architects | New York | 1995

The fully retractable wall diminishes the threshold to the city and its living waterfront. In fair weather, the building opens to the south, erasing the line between inside and out. The terminal can also provide shelter, a place not only for the traveler, but for the local community as well.

96

[www.smharch.com]

Pier 11 Wall Street Ferry Terminal and Transit Hub, SMH+, Website: http://www.smharch.com/project_template.php?id=29&bgId=&category=planning Turner, J. and Hejduk, J. (2000) Between spaces: Smith-Miller + Hawkinson Architecture, New York: Princeton Architectural Press

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

Weather Sensitive Envelope

Axel Ritter | 1997

A polyreactive mechanomembrane, bionically inspired by the various functions of the human skin. This technology demonstrator has the ability to change color, exploit evaporative cooling and raise hairs in response to changes in temperature and humidity of the surrounding air. The concept works via a complex system of thermobimetals, water absorbing ceramics, color filters, springs and elastics bands all attached to a flexible textile membrane.

97

[Ritter, 2007]

Ritter, A. (2007) Smart Materials in architecture, interior architecture and design. Basel: Birkhuser Currey, M. (2007) Performance enhancers, Metropolis 26(9): 172-175

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

MACRO
open-loop

Wind Shaped Kinetic Pavilion

Michael Jantzen | 2007

Each of its six main segments twists freely around its central support frame in response to fluctuations in the wind. The rotation of the light-weight slices generates enough electricity as it moves to light the pavilion at night. The view from inside literally changes at the whims of the weather.

98

[Jantzen, 2010]

Jantzen (2010) Portfolio Michael Jantzen, Website: http://www.michaeljantzen.com/portfolio/ Dove, C. (2007) Wind shaped pavilion, a mind-blowing experience, Modern design, architecture and art, 2007(6): 33

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Zollverein school
Sanaa Architects | Essen, Germany | 2007

The wall construction only consists of a 30 cm single leaf concrete layer with embedded heating ducts, acting as active thermal insulation. Symbiosis is achieved because excess mine-water at 30C that is pumped from a nearby colliery to prevent collapse of the mine, is fed to the radiant tube heat exchanger to modulate heat transfer across the envelope. This circulation of warm water does not only serve as heating supply in winter, but it also renders conventional thermal insulation superfluous.

99

[Risen, 2008]

[www.thomasmayerarchive.de]

[Risen, 2008]

Knaack, U., Klein, T. Bilow, M. and Auer, T. (2007) Faades Principles of construction, Basel: Birkhuser Risen, J. (2008) Zollverein school building: active thermal insulation, Greenlineblog. Website: http://greenlineblog.com/2008/02/zollverein-school-building-active-thermal-insulation/

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010

closed-loop

MACRO

Zon-Wel SmartFaade
SenterNovem EET project | The Netherlands | 2007

Zon-Wel concept improves daylight entrance with window integrated retroreflective lamellae, saves energy with high insulating vacuum panels and generates energy with integrated photovoltaic modules. The 'smartbox, part of Zon-Wel, is a small, faade integrated, unit that incorporates all climate control functions like cooling, heating and ventilation with heat and moisture recovery in the building envelope.

100

[www.smartfacade.nl]

de Boer, B et al. (2008) Zon-Wel Dynamische zonnegevel, openbaar eindrapport Jelgersma, M. (2007) Halvering energiegebruik - Smartbox is opgenomen in gevel en regelt klimaat van achterliggende ruimte, Bouwen op innovatie, december 2007

R.C.G.M. Loonen | CABS What can we simulate? | MSc Thesis | TU/e | 2010