Approaches for computational performance optimization ofinnovative adaptive façade conceptsCitation for published version (APA):Loonen, R. C. G. M. (2018). Approaches for computational performance optimization of innovative adaptivefaçade concepts. Technische Universiteit Eindhoven.

Document status and date:Published: 06/09/2018

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

Please check the document version of this publication:

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

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

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

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

Take down policyIf you believe that this document breaches copyright please contact us at:openaccess@tue.nlproviding details and we will investigate your claim.

Download date: 02. Aug. 2020

/ Department of the Built Environment

bouwstenen 253Roel Loonen

Approaches for computational performance optimization of innovative adaptive façade concepts

APPROACHES FOR COMPUTATIONAL

PERFORMANCE OPTIMIZATION OF

INNOVATIVE ADAPTIVE FAÇADE

CONCEPTS

PROEFSCHRIFT

ter verkrijging van de graad van doctor aan de Technische Universiteit Eindhoven, op gezag van de rector magnificus prof.dr.ir. F.P.T. Baaijens, voor een commissie aangewezen door het College voor Promoties, in het openbaar te verdedigen op

donderdag 6 september 2018 om 13.30 uur

door

Roel Cornelis Gerardus Maria Loonen

geboren te Grubbenvorst

Dit proefschrift is goedgekeurd door de promotoren en de samenstelling van de promotiecommissie is als volgt: voorzitter: prof.ir. E.S.M. Nelissen

1e promotor: prof.dr.ir. J.L.M. Hensen

co-promotor: dr.dipl.-ing. M. Trčka (United Technologies Research Center)

leden: prof.dr. J.A. Clarke (Strathclyde University)

prof.dr. M. Perino (Politecnico di Torino)

prof.dr.ing. U. Knaack (Technische Universiteit Delft)

prof.dr. A.P.H.J. Schenning

prof.dr.ing. A.L.P. Rosemann

Het onderzoek dat in dit proefschrift wordt beschreven is uitgevoerd in

overeenstemming met de TU/e Gedragscode Wetenschapsbeoefening.

APPROACHES FOR COMPUTATIONAL

PERFORMANCE OPTIMIZATION OF

INNOVATIVE ADAPTIVE FAÇADE CONCEPTS

The research reported in this dissertation was supported by RVO through the

Long-Term Energy Research Strategy (EOS-LT) project, FACET: Façade als

Adaptief Comfortverhogend Energiebesparend Toekomstconcept

© Roel Loonen, 2018

A catalogue record is available from the Eindhoven University of Technology

library

ISBN: 978-90-386-4554-4 NUR: 955

Published as issue 253 in the Bouwstenen series of the Department of the Built Environment, Eindhoven University of Technology

Cover design: GRFK grafisch ontwerp

Printed by: TU/e print service - Dereumaux Eindhoven, the Netherlands

v

Contents

Summary .........................................................................................................................vii

Acronyms ......................................................................................................................... xi

1 Introduction .................................................................................................................. 1

1.1 Trends in the built environment ................................................................... 1

1.2 The promise of adaptive façades .................................................................. 5

1.3 Adaptive façades – research and development needs ........................... 6

1.4 Adaptive façades – modeling, simulation and optimization .................. 8

1.5 Aims and objectives ........................................................................................ 10

1.6 Scope of the work ............................................................................................ 11

1.7 Research methods .......................................................................................... 12

1.8 Thesis outline .................................................................................................. 14

2 State-of-the-art: Adaptive facades ....................................................................... 17

2.1 Introduction ..................................................................................................... 17

2.2 Adaptive façades: background and characteristics ................................ 18

2.3 Decision support for product development of new adaptive façade concepts ............................................................................................................ 21

2.4 Façade adaptation at longer time-scales – an optimization example

............................................................................................................................ 30

3 Principles and requirements for simulation-based optimization of adaptive façades ............................................................................................................................. 41

3.1 Introduction ..................................................................................................... 41

3.2 Modelling and simulation techniques for performance prediction of adaptive façades ............................................................................................. 43

3.3 Control aspects – modeling and optimizing the operation of adaptive

façades ............................................................................................................... 57

3.4 Performance aspects and performance indicators ............................... 62

vi

4 Development of a toolchain for performance optimization of adaptive

façades ............................................................................................................................ 65

4.1 Introduction .................................................................................................... 65

4.2 Performance simulation of adaptive façades ......................................... 66

4.3 Co-simulation approach ............................................................................... 70

4.4 Coupling between ESP-r and BCVTB ......................................................... 71

4.5 MBC implementation using Matlab, BCVTB, ESP-r and Radiance ..... 74

4.6 Overview of the toolchain ............................................................................ 87

5 Quantifying the performance potential of adaptive façades – a case study

........................................................................................................................................... 91

5.1 Introduction ..................................................................................................... 91

5.2 Description of the case study ..................................................................... 92

5.3 Results – performance potential compared to reference case ......... 98

5.4 Discussion and conclusion ......................................................................... 106

6 Investigating trade-offs between performance and façade system

complexity .................................................................................................................... 107

6.1 Introduction ................................................................................................... 107

6.2 Description of the case study .................................................................... 109

6.3 Simulation strategy and settings ............................................................... 113

6.4 Results .............................................................................................................. 113

6.5 Discussion and conclusion .......................................................................... 118

7 Concluding remarks ................................................................................................. 121

References .................................................................................................................... 131

Curriculum Vitae ......................................................................................................... 149

List of Publications ..................................................................................................... 155

Acknowledgements ..................................................................................................... 151

vii

Summary

Adaptive façades give buildings the flexibility to act in response to varying weather conditions and occupant preferences. They are increasingly

recognized as a promising option for achieving high levels of indoor environmental quality (IEQ), while offering potential for low-energy building

operation. Despite the opportunities that arise from the fact that adaptive façades can consolidate the complementary benefits of passive and active

building design strategies, the number of adaptive façade applications in the current building stock is still limited, and the present focus is mostly on

individual projects rather than on solutions that promote widespread adoption. In research and product development contexts, however, a rapid

increase in activities that explore the potential of innovative adaptable building envelope components can be observed.

This doctoral dissertation sets out to investigate how computational

modeling, simulation and optimization techniques can be used to support, stimulate and accelerate the transition towards next-generation adaptive

building envelopes. More specifically, the objective of this work is to develop a computational methodology that can be used to facilitate design analysis

and performance-based design space exploration in the product

development process of innovative adaptable building envelope components and concepts. The computational tools are developed, tested and evaluated

with the aim of assessing the viability of adaptive façades as a design strategy in general, and to guide research and development (R&D) processes towards

the most promising solutions.

The development of the computational methodology begins by analyzing the characteristics of the problem at hand. A literature review that covers the

interface between adaptive façades and simulation support for R&D of innovative building envelope components identifies the needs and

possibilities for building performance simulation (BPS)-based support in this domain. Preliminary simulation studies and an analysis of software

capabilities furthermore identify that current simulation workflows and software lack the necessary capabilities to model the effects of time-varying

viii

building shell properties. Together, these findings serve as input for the

functional requirements of the newly-developed simulation-based optimization approach.

The focus in this thesis is on developing a toolchain that can help in gaining

a better understanding of the effect of adaptive façades on the dynamic interactions between energy saving potential and IEQ in terms of thermal and

visual performance. To achieve these goals, a co-simulation strategy using a middleware software was deployed to couple high-resolution tools for

building energy simulation and dynamic daylight simulation. The code of existing simulation programs was modified to enable performance prediction

of façades that change their properties during simulation run-time. In buildings with adaptive façades, there is a tight coupling between design and

operational aspects. To address this tight coupling in the performance prediction framework, a receding time horizon approach for optimization-

based supervisory control of adaptive façades was implemented in the simulation framework. This simulation-assisted control approach makes use

of genetic algorithms to efficiently search the large option space of possible façade adaptation scenarios.

The capabilities of the whole toolchain are illustrated in a typical commercial office zone in Dutch climate conditions. The results show that adaptive

façades are able to improve the performance of all relevant aspects compared to the best-performing design with non-adaptable façade. These gains are

achieved by influencing the varying trade-offs between multiple competing performance requirements over time. The toolchain is also able to identify

the most promising region in the large option space of possible façade adaptation sequences.

The usability of the newly developed toolchain is further demonstrated by

showing how it can help in addressing the trade-offs between façade complexity and performance. The focus in this second case study is on a

fenestration system that has the ability to selectively control the transmission/reflection of (i) visible light, (ii) near-infrared solar radiation,

(iii) and longwave infrared radiation. The results show that the method is able to identify high-potential directions for further R&D in building envelope

materials and components. Moreover, it enables a better comprehension of the causal relationships between adaptability of building shell parameters

and performance.

ix

By showing the development, testing and evaluation of a computational

approach for analysis and performance-based design space exploration of adaptable building envelopes, this thesis demonstrates how modeling and

simulation techniques can be a valuable tool for supporting both design and product development. Continued application of such virtual testbeds can

help to exploit the lingering potential of adaptive façades.

x

xi

Acronyms

ACH Air changes per hour

BCVTB Building controls virtual test bed

BES Building energy simulation

BPS Building performance simulation

CABS Climate adaptive building shell

CTF Conduction transfer function

DGP Daylight glare probability

EC European Commission

GA Genetic algorithm

HVAC Heating, ventilation and air conditioning

IAQ Indoor air quality

IEA International Energy Agency

IEQ Indoor environmental quality

IPC Inter-process communication

MOO Multi-objective optimization

MBC Model-based control

MPC Model predictive control

PCM Phase change material

PI Performance indicator

RBE Responsive building element

UDI Useful daylight illuminance

WWR Window-to-wall ratio

xii

1

1

Introduction

1.1 Trends in the built environment

The future of buildings and cities will be shaped by a variety of factors. Among the most influential drivers that will affect the way we design and operate

buildings are the need for decarbonization and supply of energy from clean and renewable sources. As the world’s population, average standard of living

(GDP) and urbanization rate continue to rise, the necessity to develop more environmentally-conscious buildings will only become more compelling.

Under a business-as-usual scenario, market forecasts project a twofold increase in global building-related energy consumption (heating and cooling)

and associated emissions by the year 2050 compared to 2010 (Figure 1.1) [IPCC 2014].

The built environment is partly responsible for the current situation1, but also

offers ample opportunities for new solutions to address the societal challenges of climate change and sustainable development.

1 In 2010, the energy consumption in residential and commercial buildings accounted for 35% of the total end-use

energy and 30% of energy-related CO2 emissions [IEA 2013b; Ürge-Vorsatz et al. 2015].

2

Figure 1.1. Trends and drivers for global heating and cooling thermal energy consumption in commercial buildings under a frozen efficiency scenario [IPCC 2014].

The European Commission (EC) has identified the building sector as a key

enabler in its long-term decarbonization strategy, by targeting a reduction in

CO2 emissions of at least 80% by the year 2050 [EC 2011]. This strategy has been translated into immediate action plans and policy instruments that

address various aspects of design, construction, operation and refurbishment of buildings and equipment with less use of fossil fuels.

Nowadays, the technical means are available to develop positive-energy

buildings, which, on an annual basis, consume less energy than they generate with on-site renewable energy systems [Kolokotsa et al. 2011; Cole and

Fedoruk 2015]. However, the deployment of such buildings is so far mostly limited to demonstration projects, while wider market penetration is

impeded by technological as well as non-technological barriers [Ryghaug and Sørensen 2009; Hakkinen and Belloni 2011; Prieto et al. 2017].

A recent analysis by the International Energy Agency (IEA) shows that the

buildings sector is off-track with long-term sustainability objectives, and that concerted efforts are needed to adjust this direction [IEA 2016].

Technological change, driven by research and development (R&D) of innovative building materials and components, with attention for high market

3

penetration potential, forms a key element throughout all phases of this

transition.

One of the core needs for technology development centers around building envelope systems [IEA 2014]. Building envelopes form the boundary between

the conditioned interior spaces of a building and the outdoor environment. As such, they play a dominant role in a building’s energy balance2, and due to

the large building envelope area in the total building stock, they represent a huge cumulative mitigation potential.

Over the past four decades, the main driver for progress in building envelope

technologies has been the need for heating and cooling energy demand reductions [Sadineni et al. 2011; Konstantinou and Knaack 2013]. Significant

results with profound impacts have been achieved, as is for example illustrated by the continuous reductions in window U-values3 (Figure 1.2).

Figure 1.2. Development of U-values and window technologies between 1910 and 2020. The overview is based on literature, interviews with window manufacturers and other

professionals related to the industry in Sweden (adapted from: Kiss and Neij [2011]).

2 Estimates of the total energy consumption in buildings that is directly affected by the design and construction of

building envelopes range from 20 to 60% [IEA 2013a]. 3 Other significant facade-related advances include better airtightness, solar control coatings, natural ventilation

systems and low-conductivity opaque insulation materials.

4

Two notable observations can be derived from the trend in Figure 1.2:

⎯ Progress in building energy efficiency has gone hand in hand with

technological innovations. Regular technology breakthroughs, supported by suitable policy interventions, have always been the

main driving force for moving to the next level of insulated windows [Kiss and Neij 2011; Jakob and Madlener 2004; Papadopoulos 2016].

⎯ Window improvement through reduced U-values, like most energy efficiency measures, is subject to the law of diminishing returns [IEA

2013a]. More resources (e.g. money, time, material) are needed for continued improvements, but the incremental gains are not

proportional to the investments. To break with this trend and attain the next generation of building envelopes, it is worthwhile to explore

alternative strategies and pursue the integration of multiple functions in the building envelope [Knaack et al. 2008].

The upgrade of building envelope components, with an emphasis on energy

conservation and cost saving arguments, undeniably forms an important step towards meeting the goals for a more sustainable built environment.

However, this strategy is also subject to debate [Ucci and Yu 2014; Perino and

Serra 2015]. The strong focus on reducing energy losses tends to approach sustainable building design from only a single direction. To create buildings

that deliver high performance in multiple dimensions, there are several other, occupant-related performance aspects that deserve at least an equal amount

of attention. It is contended that these opportunities are sometimes overlooked with energy efficiency in the spotlight [Bluyssen 2010; Altomonte

et al. 2015; Marszal et al. 2011]. This viewpoint is further supported by a number of recent studies demonstrating that design solutions that are

optimized for low building energy demand are not always conducive to creating a comfortable, healthy, productive and/or positive indoor

environment for occupants in a robust way [Deuble and de Dear 2012; Sundell et al. 2011; Roulet et al. 2006; Baird and Field 2013; McLeod et al. 2013].

Given this context, the premise is that a reconsideration of widely accepted

building design principles is needed to come up with new prospects and integrated solutions that can spearhead the development of future high-

performance, environmentally-conscious buildings. This thesis investigates adaptive façades as a possible design strategy for reconciling the seemingly

contradictory requirements of low-energy building operation with high indoor environmental quality (IEQ).

5

1.2 The promise of adaptive façades

Adaptive façades, sometimes referred to as climate adaptive building shells (CABS), can adjust their properties or configuration over time, in response to

variable weather conditions and comfort preferences [Loonen et al. 2013]. This gives buildings with adaptive façades the unique ability to manage

energy flows for the benefit of occupants, by taking advantage of the dynamic conditions it is exposed to [Heiselberg 2009] (Figure 1.3).

Figure 1.3. Illustration of the dynamic energy flows and interactions in buildings with adaptive façades (from: IEA EBC Annex 44, adapted by Fernández Solla [2014]).

In this way, adaptive façades represent a distinct alternative to the passive design approach, which achieves energy savings, primarily through air-tight, highly-insulated building envelopes, thereby reinforcing a disconnection

between a building and its environment. In contrast, adaptive façades have

potential to influence the multivariate building envelope trade-offs in a dynamic way, and can actively take advantage of the natural energy sinks and

sources in its environment [Davies 1981]. For this reason, many technology roadmaps identify adaptive façades as a promising direction for achieving

forthcoming targets for design and operation of nearly zero-energy buildings [IEA 2013a; EC 2013; US DOE 2014].

Well-known examples of adaptive façades include switchable glazing

[Baetens et al. 2010], dynamic insulation [Kimber et al. 2014; Claridge 1977] or façade systems with phase-change materials [Favoino et al. 2014].

Nevertheless, the number of adaptive façade applications in the current

6

building stock is still limited, with a focus on individual projects rather than

larger-scale solutions [Loonen et al. 2013]. Increased deployment and market uptake of advanced innovative building envelope technologies is needed to

move from niche projects to mainstream construction and to play a role in meeting the 21st century sustainability targets [IEA 2013a; US DOE 2014].

1.3 Adaptive façades – research and

development needs

In recent years, the number of proposals for new adaptive façade systems is increasing at an exponential rate [Loonen 2018; Loonen 2015]. Two main

development directions can be discerned: (i) kinetic components with actuation of movable parts via mechanical systems; and (ii) responsive

materials, with the ability to change the physical behavior of building elements in response to dynamic conditions. Especially the developments in

material science laboratories have been identified as promising for the construction sector, because of the possibilities for scalable solutions with

potential for wide applicability [Bastiaansen et al. 2013; Loonen, Singaravel, et al. 2014]. After discovery of a new principle or material, academic research

groups (e.g. in physics and chemistry labs) typically have the expertise, interest and resources to develop such concepts up to a relatively low

technology readiness level (TRL4) (Figure 1.4).

4 Technology readiness levels were defined by NASA in the 1980s as a system for measuring and comparing the

maturity of innovative technologies. This rating system has been adopted by many organizations and governmental

agencies, and is nowadays frequently used as a policy-making instrument for research and innovation [Mankins

2009].

7

Figure 1.4. Overview of activities at different technology readiness levels (TRL). A detailed description of different R&D challenges at TRL 4-7 is given.

The subsequent phases of testing, refining, technology transfer and

commercialization into marketable products tend not to be straightforward

[Shove 1999; Arora et al. 2014]. Several reasons can be identified for this challenging situation:

⎯ Basic research at low TRLs is mainly done with public funding,

whereas private investors are mostly interested in working towards commercial viability. This leaves a void in the middle.

⎯ Many resources are generally needed throughout the whole R&D cycle, but the certainty of success is relatively low. Only few

technology concepts will eventually develop into successful commercial façade products.

⎯ There is a lack of tools that can provide insights into building-

integration issues at an early R&D phase (TRL 1-5). This results in a

mismatch between information need and availability and complicates decision-making to move future iterations of the product towards the

direction of high-potential solutions.

⎯ The process requires an interdisciplinary approach. The right combination of skills and expertise (materials development and

building application) may not always be available.

8

Figure 1.5. Availability of resources for new product development at various TRLs. The gap in the middle is sometimes referred to as “The Valley of Death” (adapted from: Coyle

[2011]).

The “Valley of Death” is sometimes used as an analogy to describe the aforementioned discontinuity in innovation processes (Figure 1.5) [Markham

et al. 2010; Hensen et al. 2015]. Developing methods and tools that can bridge this valley is identified as an essential stepping stone for practical deployment

of innovative sustainable technologies, and is therefore high on the agenda of policy programmes, such as Horizon 2020, the EU Framework Programme for

Research and Innovation [EC, 2013].

The work presented in this thesis is part of this ongoing development in

relation to emerging building envelope components. More specifically, computational building performance simulation and optimization is put

forward as a tool for supporting informed decision-making in research and development processes of innovative, next-generation adaptive façade

concepts.

1.4 Adaptive façades – modeling, simulation and

optimization

Over the last decades, building performance simulation (BPS) has evolved into

a well-established design support tool in the construction industry [Clarke and Hensen 2015]. BPS takes into account the dynamic interactions between

9

building shape and structure, systems, user behavior and climatic conditions,

and is therefore regarded as a valuable tool in many building design processes [Hand 1998; Bleil de Souza and Tucker 2014]. Because of these attributes, BPS

can also be used as a virtual test-bed for supporting informed decision-making in the R&D phase of emerging adaptive façade concepts. However,

due to (i) the limited flexibility for modeling adaptable building envelope components in state-of-the-art simulation software, and (ii) difficulties with

implementing advanced façade control strategies during the simulation, such possibilities have thus far only been explored to a limited extent [De Klijn-

Chevalerias et al. 2017].

Another opportunity for advanced innovation support with the use of BPS, is to formulate the requirements for building envelope design as an

optimization problem. This leads to an inverse approach in which intelligent algorithms are used to drive a parameter search, until a design solution is found that meets the specified objectives in the best way possible. Through

such structured design space explorations, the trade-offs between various performance aspects can be analyzed in a systematic way [Deb and Srinivasan

2006; Radford and Gero 1987; Turrin et al. 2011]. With application to conventional, static façades, numerous studies have successfully

demonstrated the value of combining BPS models with optimization techniques, such as genetic algorithms [Attia et al. 2013; Evins 2013].

Applications of optimization for adaptive building envelopes are not yet extensively studied in scientific literature [Loonen et al. 2017; Favoino,

Overend, et al. 2015].

Figure 1.6. Building performance simulation has been linked to adaptive façades, optimization and product development in an isolated way. The aim of this research is to

combine these concepts in one framework.

10

It is hypothesized that the combination of computational modeling,

simulation and optimization can form an essential resource in supporting, stimulating and accelerating the development process of high-performance

adaptive façade concepts. However, this is a complex task, and currently available simulation tools and workflows lack capabilities to support this

process (Figure 1.6).

1.5 Aims and objectives

The main aim of this research is to develop, test and evaluate a computational approach that can be used to support design analysis and performance-based

design space exploration in the development process of innovative adaptable building envelope components and concepts.

Four objectives are directly related to this research aim:

⎯ To develop an effective modeling and simulation strategy for

integrated performance prediction of buildings with time-varying

building shell properties.

⎯ To develop and test a computational approach, based on simulation and optimization techniques, that is capable of exploring the

performance potential/limits of adaptive façades.

⎯ To illustrate, on a case study basis, how this approach can be used to

identify high-potential, i.e. high-performance, low-complexity, directions for future adaptive façade concepts (Figure 1.7).

⎯ To better understand the causal relationships between adaptability of building shell parameters and performance (energy and comfort)

in a number of demonstration examples.

11

Figure 1.7. High-potential adaptive façades are identified as having relatively low complexity, but high performance.

1.6 Scope of the work

The field of adaptive façades is a multi-disciplinary design and research area

that covers a relatively broad scope and receives attention from a diverse set of stakeholders. The work presented in this thesis targets a specific subarea

within the larger field of adaptive façades. Before describing the methods and research approach in more detail, a further clarification of the specific scope

of this work is given first.

Performance aspects: Building performance can be analyzed from many

different perspectives, including e.g. structural, aesthetic, air quality, lighting,

financial, durability, energy (operational and embodied) and acoustic aspects. In this thesis, the assessment of building performance concentrates on

energy saving potential for space conditioning (HVAC and lighting) and indoor environmental quality in terms of thermal and visual comfort.

Type of adaptability: The adaptation frequency in adaptive façades can range

from the time-scale of seconds (e.g. responding to variations in wind speed) to decades (e.g. responding to global warming or changes in building use as a

result of functional changes) [Fernández Solla 2014]. In this thesis, the main focus is on façade adaptation in the middle of this range (minutes-hours-

days).

Envelope construction types: Adaptability of building shell features focuses

on thermophysical and optical properties of the exterior façade. For example,

Com

plex

ity

high

low

12

airflow in multiple skin façades, adaptable roof systems and adaptability in

natural ventilation apertures are not within the scope of this study.

Level of abstraction and role in the building design process: The modeling

and simulation work described in this thesis targets to support the product development phase of innovative building envelope components. The

computational tools are developed with the aim of assessing the viability of adaptive façades as a design strategy in general, and to guide future research

and development processes towards the most promising solutions. It is developed for scoping or pre-feasibility studies that take a higher-level

perspective, i.e., the aim is to cover various different adaptive façade concepts, without focusing on the practical implementation aspects of

specific commercial building envelope components or prototypes. In

addition, the work is not necessarily limited to currently existing/available materials, but intends to facilitate the exploration of next-generation

adaptive façade concepts. The goal of the tools is to offer guidance for several stakeholders, including façade manufacturers, consulting engineers and

material scientists, to focus their attention on the most promising development directions with the highest impact.

Software development: The computational tools are developed in the form

of a “software prototype”, i.e. they are developed in a rapid way for research

purposes [Trčka 2008]. In this phase, little attention is being paid to facilitate

robust use for practitioners and other third-party users. Consequently, the software prototype does not yet include extensive documentation, tutorials,

interface development, etc.

1.7 Research methods

The work described in this thesis focuses on the research activities regarding development, testing and deployment of a toolchain5 for computational

performance simulation and optimization of buildings with adaptive façades. The research methodology follows the steps that are taken in the workflow

of software prototyping [Boehm 1988]. In this thesis, this is approached by iteratively addressing the various phases of the modeling and simulation life

5 A suite of software tools, scripts and algorithms, linked together in such a way to achieve capabilities that cannot

be achieved by the individual software tools.

13

cycle, as described by Balci [1994]. The four steps in this process are

described below.

1 Analysis of requirements

Development of the computational approach for performance-based analysis of innovative adaptive façade concepts begins by decomposing the

characteristics of the problem at hand. A literature review that covers the interface between adaptive façades and simulation support for R&D of

innovative building envelope components is conducted first. Together with

preliminary simulation studies and an analysis of capabilities of state-of-the-art BPS programs, including an inter-model comparison, this leads to the

specification of functional requirements for the toolchain.

2 Prototype development

The second step consists of model development and implementation and subsequent initial testing of the toolchain, leading to the development of a

working software prototype. After an initial phase of developing conceptual

models [Robinson 2007], three methods are used to accomplish this step:

⎯ Code modifications in existing BPS software to allow performance prediction with controllable time-varying building envelope

properties;

⎯ Integration of simulation programs in a receding time horizon control

framework with optimization algorithms to address the tight coupling between design and control aspects in dynamic systems such as adaptive façades;

⎯ Development of a co-simulation strategy to enable run-time communication between simulation solvers in the thermal and visual

domains.

3 Review of functionality

After a first version of the toolchain is developed, its performance is evaluated

with respect to the formulated requirements, to validate the functional performance of the toolchain and to examine if further refinements are

necessary. Iterative loops in combination with step 2 are carried out. This research step includes:

⎯ Illustration of the main features and capabilities of the whole

toolchain in a typical application example.

14

⎯ Critical examination of the quality (i.e. sensibility) of optimization

outcomes by comparing the outcomes with principles from building physics and results published in literature.

4 Demonstrate usability

Finally, the usability of the computational approach is further demonstrated in an application-specific case study. This demonstration aims to show the

potential for giving more advanced decision-support in response to requirements posed by real-world R&D processes at different modeling

resolutions and levels-of-detail. Moreover, it helps in building a better understanding of the causal relationships between adaptability of building

shell parameters and performance.

1.8 Thesis outline

Chapter 2 presents an analysis of the unique characteristics of adaptive façades and introduces the opportunities for computational simulation to

support research and development of innovative building envelope components. Together with the lessons learned from preliminary simulation

studies, this information is used to derive the functional needs for effective performance prediction and optimization of adaptive façades.

Chapter 3 forms the link between the research context (Chapter 2) and the

development of the new computational approach (Chapter 4). It presents higher-level requirements for performance prediction of façades with time-

varying properties, and discusses the limitations and opportunities for accomplishing this in existing BPS software. This chapter further introduces

the specific challenges that arise from optimization of dynamic properties in

adaptive façades, and discusses the performance aspects and corresponding performance indicators that are used throughout this thesis.

Chapter 4 provides considerations and detailed information about the

toolchain implementation. It describes the software programs that are used, how they were modified, and in which way they are connected in one co-

simulation framework, that couples thermal and daylighting aspects, and ensures high-performance supervisory control strategies for time-varying

façade properties.

Chapter 5 illustrates various capabilities of the toolchain in a typical application example. The focus in this chapter is on exploring the

15

performance potential of future-oriented adaptive façade concepts that can

simultaneously adjust optical properties and thermal resistance. This study intends to help identifying high-potential R&D directions for emerging

adaptive façade materials.

Chapter 6 demonstrates the usability of the newly developed toolchain, by showing how it can help in addressing the trade-offs between façade

complexity and performance. The subject in this case study is a fenestration system that has the ability to selectively control the transmission/reflection

of (i) visible light, (ii) near-infrared sunlight, (iii) and longwave infrared radiation.

Chapter 7 summarizes the main outcomes, discusses limitations of this work

and provides directions for future research.

16

17

2

State-of-the-art: Adaptive facades

2.1 Introduction

There are numerous research activities all over the world that seek to advance the design and development of buildings with adaptive façades.

These developments cover many different aspects that have conventionally been scattered across various engineering disciplines. In recent years, a

number of dedicated scientific conferences on adaptive building envelopes have been held, and efforts are coordinated via international collaborations

such as IEA ECB Annex 44: Responsive Building Elements [Heiselberg 2009],

and EU COST Action TU1403: Adaptive Façades Network [Luible 2015]. The work presented in this thesis connects to the search for new solutions in

these ongoing developments. A broad overview of challenges and achievements in the larger field of adaptive façades’ research can be found in

Loonen et al. [2013]. The purpose of this chapter is to provide an abridged review of background concepts and new research findings, with relevance for

narrowing down the scope of this thesis. While doing so, the details of two

18

preliminary simulation studies are briefly introduced (Sections 2.2 and 2.3).

These case studies show the type of performance information that can currently be generated in terms of simulation support for adaptive façades,

but also serve to illustrate a number of shortcomings of these existing approaches. Together, this then serves as the basis for functional

requirements of the new computational simulation approach in Chapter 3.

2.2 Adaptive façades: background and

characteristics

The envelope of a building performs many different functions [Hutcheon

1963]. On the one hand, it offers safety, security, privacy, and protection against fire, wind and rain. On the other hand, it functions as the connecting

element between building occupants and the outside world, by regulating the exchange of energy and admitting access to e.g. views, daylight and fresh air.

The design of a building envelope is of crucial importance for the quality of

the indoor environment it encloses. A growing awareness of the positive influences of good indoor conditions on health, well-being and occupant

productivity [Bluyssen 2010; Jin and Overend 2014] prompts increasing interest for finding new ways of achieving this with less energy consumption

[Brager et al. 2015]. Building envelopes are exposed to an environment that is highly dynamic – weather conditions change continuously with seasonal and

daily patterns, and also the influence of occupants (presence, activity, comfort needs) varies with time. Regular building envelopes are designed as

static elements and have limited ability to act in response to these changes. Adaptive façades, on the other hand, try to exploit these dynamics.

Many different adjectives are in use to refer to the virtues of buildings with time-varying façade characteristics. The most common variations include:

adaptable, adaptive, dynamic, flexible, kinetic, intelligent, interactive,

responsive, smart and switchable. Throughout this thesis, we use adaptive façades, and define it as:

An adaptive façade has the ability to repeatedly and reversibly change some of its functions, features or behavior over time in response to changing

performance requirements and variable boundary conditions, and does this

with the aim of improving overall building performance in terms of energy use and IEQ.

19

More specifically, this research draws upon three concepts from systems

engineering literature to describe the positive features of adaptive façades: adaptability, multi-ability and evolvability. The next subsections describe

these concepts in more detail.

2.2.1 Adaptability

The definition of adaptability that is used in this thesis is the one proposed by Ferguson et al. [2007]: the ability of a system to deliver intended functionality during operation, considering multiple criteria under variable

conditions through design variables that change their physical values over

time.

Building shells with this attribute can seize the opportunity to deliberately act in response to changes in IEQ requirements and ambient conditions, such

as solar radiation, wind speed and direction, temperature, rainfall, etc. Doing this offers a potential for energy savings compared to conventional, static

buildings because the valuable energy resources in the local environment can be actively exploited, but only at times when these effects are deemed

favorable. adaptive façades can thus act as weather mediator, negotiating between IEQ needs and what is available in the ambient environment

[Wigginton and Harris 2002]. With embedded adaptability, façades no longer have to be a compromise solution for the whole year, performing acceptable

under a wide range of conditions, but never optimal regarding a specific situation [Ochoa et al. 2012; Goia 2016; C. S. Lee 2017]. Moreover, it gives

opportunities to adjust to the individual user [Bakker et al. 2014; Sadeghi et al. 2016], rather than a best average for all, as usually prescribed in comfort

standards.

In addition to the immediate effect, adaptability of building envelopes also

helps in achieving gains through smart utilization of the thermal storage capacity of building constructions. By controlling the energy flux to/from

structural thermal mass, adaptive façades can help in mitigating overheating risk, and also limit redundancy in installed heating and cooling capacity [Hoes

and Hensen 2016; De Witte et al. 2017].

20

2.2.2 Multi-ability

The concept of multi-ability originates from the existence of non-simultaneous performance requirements, or the need to fulfil new roles over

time. The ‘balcony that can be folded’ [Hofman and Dujardin 2008] and the Velux Cabrio system [Petersen et al. 1990] are illustrative examples of a

responsive building envelope that features multi-ability: depending on ambient conditions and users’ preference, it changes function from window

to a balcony-on-demand [Weaver et al. 2008].

Multi-ability differs from adaptability in the sense that multiple requirements

are fulfilled successively, not at the same time [Ferguson et al. 2007]. Unlike conventional systems, designed to satisfy a single set of conditions, it allows

for addressing change via a plurality of individually-optimized states. In this way, multi-ability promotes more efficient use of resources, which adds to

the list of benefits already mentioned in the previous paragraph. A second application of multi-ability is the potential for spatial versatility. At the same

moment in time, the properties of the building envelope can be different for various positions of the building shell. In this way, different orientations of

the building can independently react to the ambient conditions or to distinct comfort preferences requested by individual users in separate zones.

2.2.3 Evolvability

Whereas adaptability and multi-ability mostly handle short-term changes, evolvability is a property of flexible systems that deals with variations over a

longer time-horizon [Silver and de Weck 2007]. Perhaps even more than uncertainty in day-to-day operation, future building requirements and

boundary conditions are highly unpredictable, or cannot even be known in the design stage [Holmes and Hacker 2007; De Wilde et al. 2011; Grant and

Ries 2013]. Building shells that can evolve over time are a means of extracting value from the uncertainty of these unforeseen events [Struck et al. 2014].

Evolvability is nevertheless considered more a positive side-effect, rather than a main design objective; the ability to keep options open preserves

opportunities to react to changes as they unfold in the future or perhaps might not occur. Concerning the building and construction industry,

evolvability, or sometimes called survivability or resilience [Stevenson et al. 2016], can be used to deal with changing conditions coming from the outside

(e.g. climate change, changing urban environment, deterioration of building

21

components) or from the inside (e.g. organizational function changes of the

building, new space layout). In all these cases, having a façade that can react to changes increases the chances that the building can continue operation as

intended, without getting impaired by potential negative impacts of unforeseen future conditions [Fawcett et al. 2012].

2.3 Decision support for product development

of new adaptive façade concepts

2.3.1 Introduction

Transitioning adaptive façades from niche application to mainstream design

option will require continued development of innovative adaptive façade materials, components and systems. However, it turns out that just having a

good idea is not always enough; the success or failure of innovations in the construction industry is influenced by several other factors. The ability of

innovation teams to show that their new product will reduce cost and

enhance quality or performance, has been recognized as one of the key enablers for success in this context [Toole 2001; Loosemore and Richard 2015;

Klein 2014]. A lack of effective communication about performance aspects, by contrast, was identified as one of the significant barriers hindering technical

innovation in construction [Gambatese and Hallowell 2011]. In Loonen et al. [2014], fifteen recently published projects describing R&D steps of various

innovative adaptable building envelope components were reviewed and classified according to four phases of development: laboratory scale,

reduced-scale experiment, full-scale mock-up, and pilot study. Figure 2.1 presents a summary of these projects.

[Xu et al. 2007; Xu and Van Dessel 2008; Chun et al. 2008; Imbabi 2006; Elsarrag et al. 2012; Granqvist et al. 2010;

Piccolo and Simone 2009; Assimakopoulos et al. 2007; E. S. Lee et al. 2012; Lienhard et al. 2011; Viereck et al. 2010;

He and Hoyano 2011; Ismail 2001; Grynning et al. 2013; Goia et al. 2013; Goia et al. 2014; Davidsson et al. 2010; Runsheng et al. 2003; Givoni 2011; E. S. Lee et al. 2006; Erell 2004; Leal and Maldonado 2008; Carbonari et al. 2012;

Debije 2010; Nitz and Hartwig 2005; Seeboth et al. 2010; E. S. Lee et al. 2013]

22

Figure 2.1. Classification of adaptive façades product development publications in relation to the characteristic R&D phases.

23

A number of inefficient elements in the R&D process were identified from

these fifteen studies:

⎯ Mismatch between information need and availability: Product

development of building envelope components often takes a linear solution approach (e.g. the stage-gate approach), even though the

design problems tend to be ill-defined [Eppinger et al. 1997; Kiss and Neij 2011; De Klijn-Chevalerias et al. 2017]. Decisions in the early

stages have the highest impact on the end result, but in the absence of detailed whole-building performance information, tend to be

based on intuition rather than analysis. Because the available performance information tends to be limited in scope and level of

detail, it is difficult to set goals, and evaluate whether they are

achieved. Moreover, early-stage identification of most promising directions for further development is a challenging task.

⎯ Disconnect between material science and building scale: Multiple

stakeholders with diverse interests are involved in the product

development process. Each stakeholder has a different perception of the value of the future façade product [Den Heijer 2013]. As a result,

there is a need for objective performance information to assist in decision-making trade-offs. The innovation process moreover

typically spans across multiple engineering disciplines. The expertise required for contributing to progress in technology development

tends to be in a different domain than the expertise required to assess what impacts this has on the built environment. In the existing

workflow, material scientists have limited guidance as to which properties are optimal [E. S. Lee et al. 2013], and it may not be

straightforward to appraise how modifications on the material scale affect performance aspects on the building level (e.g. indoor

environmental quality). There is a demand for integrated, multi-scale, multi-disciplinary tools to provide such complex insights.

⎯ Lack of information on building integration issues: The longitudinal

performance of building envelope components is strongly coupled to building-specific design attributes (e.g. glazing percentage,

orientation) and dynamic disturbances (e.g. climatic conditions, occupants’ behavior). Component-level performance metrics like U-

value and solar heat gain coefficient, which are determined under a single set of standard test conditions, can capture this type of

complexity only partially [Dave and Andersen 2011; De Wilde et al.

24

2002]. For example, Saelens [2002] has shown that active façades

feature a significant dependence between environmental conditions and U-value or g-value. Pushing component-level properties of

adaptive façades towards either high or low extremes might not always be the best solution considering the multitude of competing

performance criteria over the whole building life cycle [Ochoa et al. 2012]. Moreover, component-level performance metrics have very

limited use in assessing the dynamic performance of buildings with adaptive façades.

⎯ Limited scope of experiments: The task of obtaining reliable

performance information on the basis of experiments is not always

straightforward. This is the case for measuring the different building

energy flow paths [Judkoff et al. 2008], and also applies to objective quantification of thermal and visual comfort perception. Moreover,

conducting series of experiments with different product variants is a time-consuming and labor-intensive activity. Because of planning

and budget constraints, the number of product iterations in experiments often needs to be kept as low as possible [Strachan

2008].

⎯ No what-if analysis: In the conventional product development

process, only a limited number of scenarios, related to such factors

as orientation, building typology and climate can be examined. It is difficult to make projections of performance outside the range of

tested conditions on the basis of this bounded and incomplete knowledge. Technological product development can also be

hampered by the state of innovation itself. The envisioned directions for further development may be clear, but the technology is still

immature, or evaluated on the basis of semi-functional prototypes [E. S. Lee et al. 2013]. Test output from experiments may consequently

give a distorted view of reality, and thereby introduce the risk that the actual performance of the end product is misinterpreted

[Thomke 1998]. Physical tests thus provide only limited insights into possibilities of products with specifications that push the edge of

what is possible. To explore future directions, it can sometimes be worthwhile to assess the performance of visionary, hypothetical

product variants with properties and/or dimensions that cannot yet be manufactured [O’Connor and Veryzer 2001; C. S. Lee 2017]. The

25

virtual world of computer simulations is well-suited for supporting

this type of analyses.

Considering these limitations in existing product design and development processes, it is likely that some emerging adaptive façade technologies do

currently not reach their full potential. Innovation processes in the

construction industry tend to be based on technology-push (what is

possible?), instead of market-pull (what is needed?) [Nam and Tatum 1992]. This situation potentially leads to sub-optimal solutions that have negative

impact on the competitive position of product developers, but is also a missed opportunity for the innovations to contribute to sustainability goals.

This thesis explores the use of BPS to assist decision-making in product development processes of adaptive façades. The application example

presented in Section 2.3.2 will illustrate the added value of the current use of BPS in a typical R&D project. The opportunities can be contrasted with the

inefficiencies presented in this section. At the same time, the case study also highlights some of the existing limitations in the use of simulations with

respect to the specific characteristics of adaptive façades.

2.3.2 Case study and simulation method

Windows with switchable reflection/transmission properties in the near-

infrared (non-visible) part of the solar spectrum are considered as a viable future option for significant reductions in the energy use for heating, cooling

and lighting in buildings [DeForest et al. 2013; Ye et al. 2013; Khandelwal et al. 2017]. Recently, such a switchable IR reflector with high transparency in the

visible wavelength range has been fabricated with the use of cholesteric liquid crystals [Khandelwal et al. 2015]. Throughout this development process,

simulations acted as a virtual test-bed, providing feedback to the developers by evaluating the whole-building performance impacts of switchable IR

reflectors with different properties.

A medium-size office building (Table 2.1), with properties as defined by the U.S. Department of Energy reference buildings, was simulated using TRNSYS

simulation software [TRNSYS 2017]. In these dynamic whole-building performance predictions, the effect of the switchable IR reflector on

potential primary energy savings for heating, cooling and artificial lighting was analyzed. The window was switched to the reflecting state when the

indoor operative temperature was above 22 °C during daytime. In this way,

26

solar heat gains are allowed to the building when it is cold, but reflected when

there is a risk of indoor overheating.

Table 2.1. Details and assumptions of the case study building. See [Deru et al. 2011] for a full description.

Building description

Window-to-wall ratio 48%

U-value window 1.3 W/m2K

Opaque façade elements Heavyweight, thermal insulation

according to ASHRAE 90.1 [2013]

Room depth 6.4 m

Conversion to primary energy

Heating system efficiency 0.9

COP heat pump 3

Primary energy conversion factor for

electricity

2.5

Building usage scenario

Climate data TMY2 weather files

Occupied hours 09:00 – 17:00 on weekdays

Occupant heat loads 6 W/m2

Equipment loads 13 W/m2

Lighting power density 9 W/m2

Heating and cooling setpoint (Top) 20°C and 26°C

The window properties that were used in the simulations were obtained from spectral measurements on small-scale samples, and converted to the right format, using the calculation methods in the software Optics [LBNL 2017a] (Table 2.2).

27

Table 2.2. Window properties of the double glazing units in the simulations

Type of Window Tsol Tvis g-value

Reference 0.60 0.76 0.70

Switchable IR reflector – transparent state 0.59 0.72 0.68

Switchable IR reflector – reflecting state 0.48 0.69 0.56

Three different climates were selected to understand the relation between environmental conditions and energy savings in buildings with switchable IR

reflectors: (1) Abu Dhabi, United Arab Emirates (2) Amsterdam, the Netherlands and (3) Madrid, Spain. The results are compared to a reference

configuration employing normal double glazing (Ref, Figure 2.2) for a south facing office. We have also compared the results of the switchable IR reflector

(R-IR) with a static permanent IR reflector (StIR) in the ‘off’ state [Khandelwal et al. 2014]. This evaluation shows the impact that a switchable system may

have in comparison to static IR reflectors that are already available in the market.

2.3.3 Results

The simulations show that the energy-saving potential of switchable infrared reflectors in office buildings strongly depends on the local environmental

conditions (Figure 2.2). In Abu Dhabi, which has a warm and sunny climate throughout the year, the application of a switchable IR reflector (R-IR, 178.1

kWh/m2.yr) leads to cooling energy savings of > 15% compared to a normal double glazing window (Ref, 211.2 kWh/m2.yr). However, in such locations,

the demand for cooling is high throughout the year, which makes non-visible solar gains unwanted at all times. Given the current window switching

control strategy, the window would be in the transparent state for only 23 hours of the year. Hence, similar energy savings are achieved with static IR

reflection (Figure 2.2, StIR versus R-IR). In the Amsterdam environment, the window would be switched from the reflective state to the transmission state

for a considerable amount of time (1684 hours). The simulation results indicate that in heating dominated climates such as Amsterdam, there is no

need for either switchable or static IR reflectors. The decrease in utilization of passive solar heat gains would lead to an increase in demand of energy

required for heating.

28

Figure 2.2. Comparison of energy use intensity for a normal double glazed window (Ref), static IR reflector (stIR) and the switchable (responsive) infrared reflector (R-IR) for

three different climates.

In climates such as Madrid which have more seasonal influences than

Amsterdam or Abu Dhabi, for example, trade-offs between heating and cooling are better balanced. The ability of the windows to switch between a

transparent and reflecting state thus becomes more interesting, as well-controlled solar gains have benefits during both heating and cooling periods.

The total energy that can be saved by replacing the normal double glazing window (Ref, 126.1 kWh/m2.yr) with the fabricated switchable IR reflector (R-

IR, 110.6 kWh/m2.yr) is 12.3 %. For the same situation, a static IR reflector (121.9 kWh/m2.yr) would only result in 3% energy savings. The climate in

Madrid is predominantly sunny, yet has a relatively large number of cold days, so continuous reflection of NIR sunlight is not always a good strategy. The

window would need to be switched to the transparent state for an estimated 870 hours per year. It is advised that switchable IR reflectors are used in this

case to utilize passive solar gains in a dynamic way.

2.3.4 Discussion and lessons learned

Although relatively simple in scope and application, the example of windows

with controllable IR reflection properties clearly illustrates the value of using BPS to support the innovation process of an innovative adaptive façade

concept in an early phase of the R&D process. With respect to the inefficiencies introduced in Section 2.3.1, it shows that simulations can be

used to:

29

⎯ Benchmark the performance of new components, relative to current-

practice design solutions and other competing innovations.

⎯ Quantify the performance of building envelope products on the

whole-building level considering the influence of dynamic boundary conditions and indoor comfort criteria.

⎯ Provide insight into the performance under different climatic

conditions. The parametric analyses can easily be extended to include variations in e.g. façades orientations, window-to-wall ratios,

building usage scenarios, etc.

Despite all the opportunities, the simulation process also faced a number of

challenges, most of which were caused by software limitations. The following limitations and needs for further development are identified:

⎯ The computational algorithms in most building performance

simulation programs focus on a single physical aspect (e.g. thermal, ventilation, daylight) [Crawley et al. 2008], with no possibility to

explicitly model the interactions between domains. Proper analysis of advanced building envelope systems requires an integrated

performance assessment method that covers all relevant physical domains and interactions at an appropriate level of detail [Citherlet

et al. 2001]. In this present case, it was not possible to account for all interactions between daylight utilization, visual comfort, electrical

light usage and passive solar energy gains. It is therefore difficult to

investigate the trade-offs between these complementary domains.

⎯ The possibilities to model time-varying façade properties in available simulation tools are very limited. This limits the scope and level-of-

detail at which different adaptive façade solutions can be assessed. The implications of such simulation-specific issues will be further

explored in Chapter 3.

⎯ Only very simplified control strategies for façade adaption could be

modeled. The control logic had to be defined before the start of a simulation. Alternative modeling and simulation approaches are

needed to support performance prediction with more advanced façade control strategies.

⎯ The current simulation study is mostly a post-hoc analysis that

rationalizes the performance of windows with properties that were

30

already fabricated in the lab. To use the power of simulation to its full

extent, the process could also be inversed, by examining the potential of materials with properties that are not (yet) fabricated in the lab. In

that case, the innovation process would shift from technology-push

toward market-pull. However, simulation-based testing of the performance of many window variants with different properties is a labor-intensive task. There is a need for a structured approach to

streamline this design space exploration process.

2.4 Façade adaptation at longer time-scales –

an optimization example

2.4.1 Introduction

The bulk of research on adaptive façades focuses on components that can

change their properties on a high-frequency basis, i.e., in the order of seconds, minutes or hours [Wigginton and Harris 2002]. Examples include

the deployment of switchable glazing technology [Baetens et al. 2010], dynamic thermal insulation [Kimber et al. 2014], advanced solar shading

systems [Schumacher et al. 2010; Fiorito et al. 2016], and materials with variable solar absorptance/emittance properties [Park and Krarti 2016].

These short-term adaptation mechanisms enable the façade dynamics to synchronize with changing boundary conditions, and as a result, they are

expected to represent the highest potential in terms of performance improvements [Selkowitz and Bazjanac 1981; Davies 1981; Moloney 2011]. It is

also the reason why the new computational approach in this thesis focuses on short-term adaptive façades.

An alternative to short-term adaptive façades are façades that can adapt their

properties in response to changing conditions over the seasons [Loonen et al. 2011]. Previous studies have identified that building designs with seasonal

adaptation strategies can enhance energy and comfort performance compared to the best static situation (e.g. [Claridge 1977; Lorenz 1998; Dubois

1999]). In this context, improvements in building performance have already been demonstrated in cases with seasonal variation in: solar shading design,

window properties, thermal insulation, thermal mass, natural ventilation

strategies and temperature setpoints [Kasinalis et al. 2014]. It is noted that adaptive façade technologies with short response time can also be operated

31

in response to long-term variations. However, compared to short-term

adaptability, seasonal façade adaptation has a number of possible advantages:

⎯ The technical feasibility, robustness and durability are expected to be higher, since long-term adaptive façades are more likely to be built as

low-cost add-on solutions, with less challenging technologies.

⎯ There is no need for complex control strategies and user interaction

mechanisms as the façade needs to change only few times during the year.

⎯ The risk for disturbance or reduced occupant satisfaction due to

moving façade components is very low.

Additionally, it turns out that the analysis of the performance potential of

seasonal façade adaptation is not affected by most of the simulation process-related limitations for short-term adaptive façades (as presented in Section

2.3.4), and can be done using already existing simulation approaches. Long-term adaptive façades furthermore allow for easy integration with existing

building envelope optimization workflows. The next example illustrates the benefits of such an approach.

2.4.2 Case study and simulation method

The purpose of this case study is to analyze the possible improvement in energy performance and indoor environmental quality due to monthly

adaptation of six building envelope parameters. The simulations are based on coupled building energy and daylight simulations, conducted under multi-

objective optimization (MOO) scenarios with genetic algorithms, following the method that is introduced in Kasinalis et al. [2014].

The study considers a single-person south-facing perimeter office zone (3.6

m × 5.4 m × 2.7 m), situated at an intermediate floor, surrounded by identical office zones and a corridor at the back. The building, which is evaluated under

Dutch climate conditions, is occupied on weekdays from 8 to 17 h. Heating is assumed to be supplied with an ideal system with unlimited capacity, but no

active cooling system is employed. Ventilation with outside air is provided at a rate of 2 ACH during occupied hours. The opaque part of the external façade

is modeled as a single layer with a thickness of 0.35 m. Window-to-wall ratio (WWR) as well as thermophysical and optical material properties are

32

determined by optimization. Table 2.3 gives the dynamic range of the design

parameters that are considered in this study.

Table 2.3. Overview and range of design parameters.

Parameter Range Unit

Density (ρ) 50 - 3000 [kg/m3]

Specific heat (cp) 0.8 – 2.0 [kJ/kgK]

Thermal conductivity (λ) 0.1 – 2.5 [W/mK]

External surface absorptance (α) 0.1 – 0.9 [-]

Window to wall ratio (WWR) 0.1 – 0.8 [-]

Glazing ID (see Table 2.4) 1-7 [-]

Allowing the different optical and thermal glazing properties to vary independently from each other can easily lead to fenestration typologies that

are unrealistic from a physical point of view. To avoid this, we used existing window systems with meaningful properties. Table 2.4 shows the properties

of the glazing types that correspond to the glazing IDs from Table 2.3.

Table 2.4. Overview of glazing properties.

Glazing ID U-value [W/mK] g-value [-]

1 5.7 0.86

2 2.8 0.76

3 1.4 0.59

4 1.4 0.62

5 1.3 0.59

6 1.3 0.62

7 0.9 0.59

An exterior shading system (venetian blinds) is applied to control the

admittance of solar gains. These blinds are “ideally” controlled on the basis of

33

an active users profile [Reinhart 2004]. This stochastic algorithm assumes

that the blinds are rearranged on a regular basis with the aim of maximizing daylight availability while preventing glare and direct sunlight on the work

plane. Artificial lighting with a lighting power density of 10 W/m2 is switched on, only when daylight availability is not sufficient to meet the illuminance

target of 500 lx on the work plane.

Dynamic simulations were conducted using the multi-zone building model in TRNSYS [TRNSYS 2017]. These energy simulations were coupled with the

outcomes of daylight simulations in DAYSIM [DAYSIM 2015]. The long-term adaptation was investigated using twelve separate analyses, in which each

month of the year was addressed by a different set of simulations and optimization. The annual performance was then obtained by the combination

of monthly results as described in detail in Kasinalis et al. [2014]. By combining these simulations in a multi-objective optimization framework,

the following two optimization objectives were minimized:

⎯ Primary energy consumption [kWh/m2] (for heating and artificial

lighting).

⎯ Number of hours per year when |Tdiff| > 3 K, where Tdiff is the

temperature difference between the indoor operative temperature and the comfort/neutral temperature (corresponding to Category II

in EN 15251 [2007]).

2.4.3 Results

Optimal performance of seasonal adaptation in adaptive façades

Figure 2.3 shows the annual results for different façade configurations in two dimensions: energy consumption and thermal comfort. Each dot in this

scatter plot corresponds to a possible combination of façade properties, i.e. a façade design. The black dots represent the Pareto front with annual

optimization results for the façade with non-adaptive properties (note that

the cloud with dominated non-adaptive design solutions is omitted in this figure). Optimization results for the adaptable façade are obtained by making

all possible combinations from Pareto solutions of the individually optimized monthly simulations. For the case study, this cloud of results is shown with

pink dots in Figure 2.3, with its corresponding Pareto set in red.

34

Figure 2.3. Comparison of the annual Pareto fronts for the static (black) and monthly adaptive façades (red).The pink dots represent the cloud of available solutions for the

case of the adaptive façade. See text for descriptions of points A, B, B’ and C.

It is worth noting that all solutions that were aggregated from the sets of monthly optimal results are non-dominated with respect to the Pareto-

optimal design solutions with a static façade. Due to the large number of possible combinations originating from the points in the twelve Pareto fronts,

the density of the created cloud is significant. Therefore, it is more difficult to distinguish individual solutions (dots) in comparison with the cloud

representing the performance of the same zone with a non-adaptive façade. This finding indicates that many adaptive façades solutions lead to

approximately similar performance, and that additional preferences could be introduced to select the best alternative. When evaluated under only the

scope of energy consumption reduction, the application of a monthly adaptable façade shows a moderate improvement. For example, if the design

team accepts discomfort in the range of 240 hours per year (with respect to EN 15251 standard, category II), then the reduction in energy consumption

amounts to 18% compared to the best performing static building shell design (compare points B and B’ in Figure 2.3). This saving potential is not remarkably

high, but it should be considered that these savings are compared to an optimized static façade, and most buildings do not have optimized static

façades. Moreover, savings around 18% would provide, for example, four

credits in the LEED green building certification system, which is a good result

35

considering that the intervention is focused only on seasonal adaptability of

the building envelope.

The potential of seasonal adaptive façades stands out better when not only the effects on energy conservation, but simultaneously also the impact on

indoor environmental quality is considered. The office zone with adaptive features achieves a high-quality indoor environment (up to zero discomfort

hours, point C) while achieving energy savings around 15% in comparison with the best performing static design, point B’. This performance gain points

out that further developments of innovative adaptable building shell concepts seem worthwhile, particularly considering the positive correlation between

indoor environmental quality and increased employee productivity in office spaces.

Optimum properties of seasonal adaptive façades

In addition to the assessment of the performance potential of a building shell with monthly adaptable properties, it is also useful to gain better

understanding of the optimum design characteristics of such an innovative façade.

Figure 2.4. Optimum window-to-wall ratio of monthly adaptive façades for the case study office building in the Netherlands. Points A, B and C correspond to the three

designs in Figure 2.3. The dotted line represents B’, an optimal non-adaptive design solution.

Figure 2.4 shows the time evolution of optimal WWR values for the three

representative points of Figure 2.3. The three lines represent point B (with 240 discomfort hours), and the two Pareto extremes with maximum (point A)

36

and minimum discomfort level (point C). The fixed value of the best

performing static façade design, point B’, is indicated with the dashed line.

The results in Figure 2.4 indicate a clear preference for low WWR values during winter months. This result implies, in accordance with previous

findings [Peippo et al. 1999; Ochoa et al. 2012], that under winter conditions in temperate climates, passive solar gains and daylight utilization are

insufficient to compensate for conduction losses through windows. Due to the lower thermal resistance of fenestration compared to opaque wall parts,

the given optimization formulation consistently favors low window to wall ratios in winter. In mid-season months, however, the reduced heating

demand in combination with the low risk for overheating promotes higher levels of daylight utilization. Therefore, relatively high values of WWR are the

optimum during these months. In the summer period, the substantially higher risk of thermal discomfort leads to smaller window areas than in the

mid-season, but larger in comparison to winter months. The differences in performance between points A, B and C in Figure 2.3 are clearly reflected in

the optimum WWR for each case. During the winter months, all three points have the same values due to the absence of overheating risk, and hence the

mild conflict between performance objectives. During the mid-season and summer, two groups can be identified. On one hand, the designs with priority

in energy consumption (points A and B) have larger windows, reducing the use of artificial lighting but increasing the number of overheating hours. On

the other hand, the design with priority for comfort (point C) has reduced windows sizes, spending more energy for artificial lighting but avoiding

comfort problems caused by excessive solar heat gains.

A comparison between the properties of point B and the static optimum shell

(B’ represented by the dashed line) in Figure 2.4 shows that the improvements in performance are based on relatively large adaptations in WWR. It is

remarkable that, for most months in the year, the adaptable WWR values are larger than the static value. Larger WWR values indicate that in this case

study, adaptive façades also increase the view to the outside, which is an important performance aspect in buildings [Aries et al. 2010]. As such, this

finding shows a clear example of the promising role of adaptive façades, together with the type of concessions [Radford and Gero 1980] that is needed

to design static building envelopes which are supposed to perform well under the wide range of different boundary conditions.

37

Figure 2.5. Optimum properties of monthly adaptive façades for the case study office building in the Netherlands.

Figure 2.5 presents the optimum monthly values of the other building shell

properties in the case study. The overall trend for most properties is in

accordance with the expected behavior. External surface absorptance has higher values in winter (improving passive solar gains) and has lower values

in summer (reflecting solar radiation to reduce overheating). Thermal conductivity has lower values in the winter and mid-season (reducing

conduction losses) and has higher values in summer to improve heat losses and avoid overheating. The combined effects of density and specific heat

suggest that optimal building shells are lightweight in winter, and heavyweight in summer. This finding is consistent with the results in Hoes et

al. [2011]. Finally, the evolution of glazing ID over time points out that it is important to have high-performance fenestration properties, i.e. low U-value

(0.9 W/m2K) and relatively low g-value (0.59), throughout the entire year.

Monthly optimized adaptive façade properties in Figure 2.5 also show some erratic behavior that cannot be directly explained by physical principles. Part

of this behavior may be caused by limitations of the genetic algorithm, which cannot guarantee that the outcome of the optimization process is the global

optimum. Although measures have been taken to ensure convergence of the optimization, an in-depth analysis of the impact of different settings of the

algorithm was not the focus of this research. These results, however, do

38

demonstrate the need for further research on robustness of optimization

outcomes with respect to adaptive façades. Another reason for the erratic behavior of some properties pertains to sensitivities in the input-output

relationship between design parameters and objectives. Two different effects play a role here:

⎯ There is an inherent difference in sensitivity of the performance

indicators to each adaptive façade property in the design option space [Struck et al. 2009; Tian 2013].

⎯ The effect of certain design variables (e.g. solar absorptance) can become negligible when another variable (e.g. thermal insulation) has

a first-order effect in driving the optimization search [Brownlee and Wright 2012].

The effect of different input-output relationships on the performance of

adaptive façades is further exemplified in Figure 2.6 which presents the cloud of design options investigated during the optimization run for the month

April.

Figure 2.6. Bubble plots of all the design solutions evaluated in the optimization process for the month of April, with colors based on the values for (A) thermal conductivity and

(B) specific heat capacity.

In Figure 2.6, each design alternative is colored according to the value of the conductivity (Figure 2.6A) and specific heat capacity (Figure 2.6B). For

thermal conductivity, it can be noticed that design solutions converge to a specific tint of the color map as the design solutions approach the Pareto

front. Such kind of ordered patterns in the optimization cloud and along the

front contain useful cues about the significance of design variables to the optimal trade-off in multi-objective problems [Chichakly and Eppstein 2013].

39

Based on the results in Figure 2.6A, conductivity seems to have a dominant

influence on performance as all Pareto-optimal solutions converge to approximately the same value. This dominant effect reduces the erratic

behavior of this property in the optimization results, as seen in the winter months of Figure 2.5B. This is in contrast with the seemingly random spread

of values in the Pareto front for specific heat capacity, and in this case the optimum solution in Figure 2.5D should be adopted with caution.

2.4.4 Discussion and lessons learned

The previous paragraphs presented an example of the analysis of adaptive

façades with coupled daylighting and thermal simulations. Although the focus was on thermal comfort and energy consumption, it did also take daylighting

and electricity use for lighting into account. Moreover, the study highlighted the merits of automated design space explorations through the coupling

between BPS tools, optimization algorithms and post-optimization analysis of the results. The study shows a significant performance potential for

seasonal adaptability of building façades because they take care of energy

savings while upgrading the indoor environment. Whereas the optimization of the static façade has to make compromises to achieve satisfactory

performance over the whole year (e.g. resulting in small window areas), long-term adaptive façades can strategically take advantage of the variable

conditions by adapting over time (e.g. large window areas during most part of the year).

It is hypothesized that adaptive façades with shorter-term (e.g. hourly)

adaptation options can achieve even better performance, because the adaptive façade actions/responses can be synchronized with changes in

internal and external boundary conditions. The here presented assessment method is, however, not well-suited to explore this potential. In these

simulations, the changing of façade configurations did not take place during simulation run-time, but was approximated by coupling the results from a

series of disconnected, monthly simulations. With a higher façade adaptation frequency, the simulation periods would get shorter, meaning that

shortcomings in the initialization process (i.e. the end state of one simulation (surface and construction node temperatures) are different from the starting

conditions of the subsequent simulation) effects start playing a more prominent role. Using a series of decoupled simulations, the correctness of

thermal history effects at the transition between two successive simulations

40

cannot be guaranteed [Loonen, Hoes, et al. 2014]. With short-term adaptation

cycles (e.g. hours), in particular, this can lead to significant prediction errors, as the repeated start-ups would almost defeat the purpose of dynamic

simulations.

41

3

Principles and requirements for

simulation-based optimization of

adaptive façades

3.1 Introduction

The relationship between façades and building performance is multi-faceted.

In comparison with conventional buildings, the various IEQ elements (e.g. thermal, visual, air quality, etc.) tend to be more interconnected in buildings

with adaptive façades. Moreover, their interactions and priorities are changing dynamically, and can be influenced over time. Modelling and

simulation studies that consider adaptive building envelopes have to accurately represent a sequence of time-varying building envelope system

states (or properties), instead of a static representation of the building enclosure. Moreover, for effective performance prediction of adaptive

building envelope systems, it is essential to simultaneously consider multiple

levels, in terms of (i) spatial scales, (ii) time resolutions, and (iii) physical domains.

42

Compared to simulation-based design optimization of conventional, static

façades, two major additional requirements for the computational performance prediction framework are identified:

⎯ Modeling time-varying façade properties: Façade specifications (i.e.

material properties or position of components) need to be

changeable during simulation run-time to properly account for transient heat transfer and energy storage effects in building

constructions [Loonen, Hoes, et al. 2014]. The motivation for doing this, as well as the challenges and opportunities for accomplishing

this in existing state-of-the-art BPS software, are discussed in Section 3.2.

⎯ Modeling the dynamic operation of façade adaptation: The dynamic

interactions in adaptive façades give rise to a strong mutual dependence between design and control aspects. The performance

of adaptive systems fully depends on the scheduling strategy (i.e. control logic) for façade adaptation during operation. Moloney [2011]

describes it as: “The design outcome in a project with kinetic façades

is a process, rather than a static object or artifact”. Thus, to identify

the characteristics of optimal adaptive façade systems, it requires not only design considerations (i.e. façade system design parameters), but

also insights into high-performance automated operation strategies of the dynamic façade to be considered, already during the product

development or design phase [Hoes et al. 2012]. Moreover, effective design and operation of dynamic façade systems depends also on the

integration with operations of the other building services. For example, limited lighting energy savings could be achieved if the

operation of dynamic solar shading is not integrated with a lighting dimming system. Similarly, the integration with HVAC, and renewable

energy systems needs to be carefully considered. This coupling between engineering domains leads to a bi-level optimization

formulation [Fathy et al. 2001; Evins and Orehounig 2014], which is further discussed in Section 3.3.

Details from the translation of the above two requirements into a software

implementation are described in Chapter 4. An important characteristic in the development of the simulation-based optimization framework is that the

optimization of building performance primarily takes place via changes in building shell configurations. This is a relatively unique situation, which

requires that the performance aspects and corresponding performance

43

indicators that are used in the software implementation need to match this

situation. The way in which the framework addresses these issues is therefore also described first, at the end of the current Chapter, in Section

3.4.

3.2 Modelling and simulation techniques for

performance prediction of adaptive façades

3.2.1 Modelling challenges

A large number of software tools is available for predicting the energy and comfort performance of buildings6. Each program has unique features in

terms of modelling resolution, solution algorithms, intended target audience, modeling options, ease of use vs. flexibility, etc. Generally, these tools are

used to support informed decision-making in the later phases of building and HVAC system design/sizing [Hensen and Lamberts 2011]. The simulation

tools with most powerful modeling capabilities, and which have undergone

most rigorous validation studies (e.g. ESP-r, EnergyPlus, TRNSYS), are legacy software programs [Crawley et al. 2008]. Although these tools have active

development communities, and receive regular updates and extension of modeling capabilities, their underlying concepts and basic software

architecture do not change. Most tools stem from a time when adaptability of building components was not a primary consideration [Ayres and Stamper

1995; Oh and Haberl 2015]. Consequently, the building shape and material properties are usually not changeable during simulation run-time in these

tools, which restricts the options for modelling adaptive façades. There are three main reasons for the present difficulties:

1. User interface: Input for constructions and material properties to

BPS programs is normally given in the form of scalar values. These parameters are either directly entered by the user, or imported from

pre-configured databases. The same static representation is implemented for the size, geometry and orientation of the various

surfaces that together form the building envelope. In the most common approach, this information is then processed once, prior to

6 The database of building energy analysis software maintained by the International Building Performance

Simulation Association USA currently consists of 174 different tools [IBPSA-USA 2018]

44

the actual simulation run, and is not updated further during the

simulation. In many simulation programs, users have very limited flexibility to extend the functionality for modelling adaptive façades

through the non-modifiable user interface and the restricted access to the source code of (proprietary) simulation tools. Typical

exceptions of non-constant elements in most simulation programs are the deployment of solar shading systems and operable windows

for natural ventilation, both of which can be functions or time series.

2. Solution routines for transient heat conduction through building elements. Many of the widely used BPS tools such (e.g. Trnsys and

EnergyPlus) adopt response factor techniques (e.g. Thermal Response Factors (TRF) or Conduction Transfer Functions (CTF)) to

solve the differential equations governing the heat transfer phenomena through opaque building elements [Spitler 2011]. These

methods are optimized for computational efficiency, but by virtue of their design, they can only work with time-invariant thermophysical

properties (i.e. density, specific heat capacity, and thermal conductivity) [Clarke 2001]. This is because the coefficients that are

used in the equations are constant and determined only once for each building envelope element at the beginning of the simulation. As such,

response factor methods do not permit variations in thermophysical material properties during simulation run-time [Delcroix et al. 2012;

Pedersen 2007]. The default solution routine in EnergyPlus uses the CTF methods, but the software was recently extended with a new

finite difference scheme for conduction, to allow for modelling temperature- or time-dependent material properties [Pedersen

2007; Tabares-Velasco and Griffith 2012]. Practical use of these new algorithms is still limited [Roberz et al. 2017], and its potential largely

unexploited. The code in the simulation program ESP-r, on the other hand, is based on the control volume method. This numerical

methods adopt an iterative procedure, thereby allowing for updates

of the matrix coefficients that describe heat transfer, as time steps of the simulation proceed. This makes the simulation of variable

thermo-physical properties possible [Clarke 2001].

3. Control strategies. Control strategies in BPS models provide the link

between sensed variables and actuator actions by means of a certain control logic. This feature is mostly used for the control of HVAC

systems, but other opportunities also exist. The (non-)availability of

45

actuator options is what in the end determines the types of adaptive

façade technologies that can be modelled in a simulation tool. The general architecture for the control of building systems (including

adaptive building envelope systems) in BPS tools can be divided into 3 parts: (i) sensors level (climatic boundary conditions, building

internal boundary conditions and occupant preferences); (ii) control logic level and (iii) actuators level, that is, any building component

that can be controlled (including HVAC, artificial lighting and adaptive building envelope systems).

Most BES tools use simplified expressions such as time-based

schedules or hardcoded if-then-else statements as strategies for building systems control. Moreover, they allow a limited range of

sensor and actuator options [Hoes et al. 2012]. Advanced control of dynamic properties is, however, identified as one of the major

elements needed for performance assessment of adaptive façades. The lack of options is currently a significant barrier for performance

prediction of advanced operation strategies with adaptive façade components as time-varying actuators.

3.2.2 Opportunities using existing simulation software

A short review of the possibilities for modeling adaptive façades in state-of-the-art BPS tools was conducted to identify the current possibilities and

existing development needs. This analysis is based on the information in user manuals, software tutorials, release notes and contextual help facilities of the

BPS tools, as well as communication with their development teams. Furthermore, scientific articles, dissertations and the information exchange

in mailings lists online forums were used to gather input. Two types of modeling features are distinguished: (i) application-specific and (ii) general-

purpose. Here, application-specific indicates that the simulation model was implemented in the software with a specific adaptive façade concept in mind.

The adaptive mechanism and how it is triggered are, therefore, already embedded in the specific model, and users can activate it easily by means of

the GUI, but are limited to the presets available. The general-purpose features, on the other hand, are not restricted to a specific technology, but

offer flexibility for user-defined combinations of adaptive thermo-physical property variations and/or triggering mechanisms. This higher abstraction

level affords more freedom for exploring innovative adaptive building

46

envelope systems, although it requires the BPS user to define and code the

control mechanism that triggers adaptation in the building element.

1 - Application-specific capabilities

An analysis of the capabilities in six commonly-used software tools resulted in six different application-specific adaptive features (Table 3.1); each of them

is discussed below.

Table 3.1. Overview of application-specific features for modeling adaptive façades in six commonly used BPS tools.

EnergyPlus7 ESP-r eQUEST IDA ICE IES VE TRNSYS

Electrochromic glazing x x x o x o

Thermochromic glazing x x o o o

Insulating solar shading x o x x o

Phase change materials x x o x

Green walls and roofs x o x

Movable insulation x o o o

x - The modeling capability is readily available from the graphical user interface and can

directly be used by advanced users.

o - The modeling capability is present in the software, but can only be activated by expert

users/developers via source-code modifications or custom-made scripts.

Different types of switchable windows, including electrochromic glazing, are commercially available, and many research papers have been written about their application in buildings and architecture [Baetens et al. 2010]. As a result

of their presence in the market, options for modelling switchable glazing technologies are embedded in several simulation tools. All the software tools

analyzed offer the possibility to control the properties of the fenestration system during simulation run-time. The differences between the various

implementations are the number of possible window states (e.g. on/off versus gradual transitions) and the simulation state variables that can be used

for the control of adaptation (e.g. room temperature, ambient temperature and incident radiation).

Thermochromic windows are slightly more complicated to simulate than other switchable window types because of their ‘intrinsic’ control character; adaptation of the fenestration properties is directly triggered by window

surface temperature instead of a control signal that is based on more general simulation variables [Favoino, Cascone, et al. 2015]. A thermochromic window

7 Some of the features in native EnergyPlus are also available in GUIs that use EnergyPlus as simulation engine e.g.

DesignBuilder, OpenStudio, Simergy, Sefaira.

47

capability was implemented in EnergyPlus (since v3.1, 2009) and ESP-r [Evans

and Kelly 1996]. The input of these models consists of sets of window properties at various temperatures. During the simulation, the

thermochromic layer temperature of the previous time step is automatically fed into a window control algorithm, which then selects the window

properties that best match with the given temperature. In IDA ICE, IES VE and Trnsys, it is also possible to model thermochromic windows, but a

significantly higher level of work and expertise is required from the user side.

The GUIs of EnergyPlus, IDA ICE and EnergyPlus offer the possibility to give dynamic shading devices additional thermal resistance properties. This

makes it possible to simulate the performance of insulating solar shading

systems [Hashemi and Gage 2012]. In such an implementation, dynamic thermal insulation and solar shading are coupled; their separate effects cannot be analyzed.

Prediction models for wall-integrated phase change materials (PCM) are present in EnergyPlus [Tabares-Velasco et al. 2012], ESP-r [Heim and Clarke 2004], IDA ICE [Plüss et al. 2014] and TRNSYS [Kuznik et al. 2010]. These

models influence heat transfer in constructions via either the ‘effective heat capacity’ or the ’additional heat source’/‘enthalpy’ method. The need to

implement PCM features led the developers of EnergyPlus to abandon the Conduction Transfer Functions approach and introduce a numerical finite

difference conduction algorithm [Pedersen 2007]. This new algorithm also has a provision for including a temperature coefficient that makes thermal

conductivity variable during the simulation [Tabares-Velasco and Griffith 2012]. No applications of this latter model were found in literature.

EnergyPlus, ESP-r, and TRNSYS support the simulation of green walls and

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 [Sailor 2008; Djedjig et al. 2015]. In the EnergyPlus model, it is possible to include material properties that

change over time with fluctuations in plant growth and moisture content [Sailor and Bass 2014].

Finally, EnergyPlus and IDA ICE [Bionda et al. 2014] have the option to predict

the performance of building envelopes with moveable insulation. A controllable layer can be applied to the interior or exterior side of a construction (not windows) to temporarily increase its thermal resistance.

48

These materials are massless, which means that no thermal energy can be

stored in a moveable insulation layer.

2 - General-purpose modeling options

The application-specific adaptive features presented in the previous paragraph only allow for a restricted level of flexibility. The tendency of BPS

tools to lag behind the market availability of adaptive technologies limits the number of application-specific modelling capabilities available in a specific

BPS tool, compared to what is technologically available. As such, there are many adaptive building envelope systems (either at prototype or at product

stage), whose performance cannot be evaluated yet with the existing application-specific simulation models. Examples include tunable switchable

windows with selective light transmission in different parts of the solar spectrum [Llordés et al. 2013], water-carrying transparent façades with solar

control [Ritter 2014], solar-tracking photovoltaic façades [Nagy et al. 2016] and luminescent solar concentrating façade elements with variable

transmission and scattering properties [Sol et al. 2018]. Therefore, from a product development and innovation point-of-view, it is more desirable to

allow for bottom-up or general-purpose approaches to be able to simulate emerging or not-yet-existing adaptive building envelope materials and

technologies.

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 building shell behavior in a generic way. A brief

overview of adaptive features in both tools is included here to provide further background for the decisions that are made in the newly-developed

simulation approach.

TRNSYS

The approach that TRNSYS takes towards managing complexity in the built environment is characterized by breaking down the problems into a series of

smaller components. One of these components is a multi-zone building model — in TRNSYS called TYPE 56 — that can be connected to a large number

of other components, including weather data, HVAC systems, occupancy schedules, controllers, output functions, thermal energy storage, renewable

(solar) energy systems, etc. This particular configuration allows the user to set up and manipulate the connections between the building and various

other subsystems/components in the simulation environment.

49

TRNSYS TYPE 56 offers the possibility to change the thermal and optical

window properties during run-time with a function called variable window ID. Additionally, it is also possible to control the ratio of window/frame area,

which influences the degree of transparent façade elements. Recently, TRNSYS was extended with a bidirectional scattering distribution function

(BSDF) model that can be changed at every time step of the simulation [Hiller and Schöttl 2014]. All the other adaptive mechanisms in TRNSYS are not found

in the (non-modifiable) building model itself, but in the connections with

other components. Using equations in TRNSYS enables the application of Boolean logic and algebraic manipulations to almost all state variables in the

simulation. This flow of information can then be used to drive a control algorithm that is able to dynamically ‘switch on’, ‘switch off’ or modulate, for

example, 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 modules (TYPES 94, 180 and 194). In addition, it is also possible to adjust the

connections with weather files and radiation processors. In this way, the effects of changing orientations (e.g. rotating buildings) can be mimicked.

Even more control flexibility can be achieved by connecting TRNSYS models to the W-editor (Type 79) [Keilholz et al. 2009]. Type 79 makes use of W, a

simple programming language that can influence the connection between the inputs and outputs of TRNSYS components at every iteration of the

simulation.

The standard TRNSYS distribution already comes with an extensive library of components. Yet, one of the distinct benefits of TRNSYS’ modular structure

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 adaptive façade behavior in a new TRNSYS TYPE which can then

be linked to the building model. Due to constraints in TRNSYS’ CTF method, coupling of these new TYPES with the building envelope model works in a

rather indirect way via the so-called slab-on-grade approach. In TRNSYS it is not possible to substitute building shell constructions/properties during

simulation run-time. Instead, developers can impose the desired behavior by overwriting the inside surface layer temperatures of adjacent zones and the

respective heat transfer coefficients. With respect to adaptive façades, Kuznik et al. [2010] and Claros-Marfil et al. [2014] recently demonstrated this

approach for a new PCM wallboard TYPE, and Djedjig et al. [2015] developed a model for green walls.

50

ESP-r

ESP-r is a multi-domain research-oriented BPS tool with an active development community and a source code that is accessible and modifiable.

The tool is used for prediction of thermal, visual, indoor air, electrical and acoustic performance of buildings. Users can extend or modify existing

features in the source code, according to their needs. Over the course of the years, several functionalities that can be used to model adaptive behavior in

the building shell have already been implemented. Nevertheless, the use of these capabilities has remained limited, possibly because the features lack

extensive documentation or are concealed somewhere in the distributed menu-structure of ESP-r. This section summarizes five of such features:

⎯ One of the control laws in ESP-r is called thermophysical property

substitution mode. It is the only strategy that is not used for controlling the operation of HVAC systems. Instead of this, this

control strategy can replace the thermophysical properties (λ, cp, ρ) 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 [MacQueen 1997]. Unfortunately, this feature does

not allow for full flexibility 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 the 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-files) based on time, temperature, solar radiation level or illuminance level.

Restrictions are that no more than two window states are supported without the possibility for gradual transitions. Recently, the

capabilities of ESP-r have been further extended with the

implementation of two new facilities for modeling transparent façade systems. Both the complex fenestration constructions (CFC)

[Lomanowski and Wright 2012] and the advanced optics [Kuhn et al. 2011] module have powerful options for façade systems with dynamic

fenestration properties.

51

⎯ 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. The

special material subroutines can actively modify the matrix coefficients of these specific nodes at every time-step. By doing this,

it directly changes basic thermophysical or optical properties and/or the associated energy flows at the equation-level, 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-defined special materials; however this may require time-intensive programming work.

⎯ ESP-r offers the unique possibility to use roaming files. This facility is used to change the location of a building as a function of time, and

was originally intended to be used for cruise ships. Because this roaming file not only includes coordinates but also orientation of the

zone, it is very well suited for simulation of rotating buildings.

⎯ Nakhi [1995] introduced variable thermophysical properties in ESP-r

with the aim to model heat transfer in building slabs in a more accurate way. The model takes into account that the properties of

most construction materials are not constant, but change as a function of temperature and/or moisture content. This dependency

is implemented via transient thermophysical material properties (λ, cp, ρ) that are linear or polynomial functions of layer temperature or

moisture content. The same functionality can be used to model certain types of adaptive façades.

3.2.3 Simulation strategies and workarounds

Aside from the above-mentioned possibilities, researchers and engineers have developed numerous customized simulation strategies for predicting

the performance of RBEs in their preferred whole-building performance simulation program [Loonen, Hoes, et al. 2014]. Such approaches often call

upon the use of workaround simulation strategies [Brahme et al. 2009], that are devised in view of various legitimate reasons such as the complete

absence of existing models for certain adaptive façade technologies, a lack of user expertise/experience or limited project resources (time and money) to

52

move towards more complex models and the absence of advanced control

options for actuating the adaptive façade system. In many of these cases, the reuse of validated, high-resolution models is an important argument in favor

of using existing software instead of the development of custom-made simulation code from scratch [Wetter 2011a], such as the approach taken by

Liu et al. [2014]. A main drawback of using workarounds is that they tend to rely on approximations or simplifications that might infringe the physics of

model representations and, consequently, also put the credibility of simulation outcomes at risk.

Arguably, the simplest approach for representing adaptive façades is by

subdividing the simulation period into several simulation runs with shorter periods (e.g. seasons), each with distinct building properties [Joe et al. 2013;

Hoes et al. 2011; Loonen et al. 2011] (Figure 3.1A). The downside of this approach is that it cannot accurately model short-term adaptive façade

dynamics, as already described in Section 2.4.4., and thus results in ‘continuity’ errors.

Figure 3.1. Schematic representation of workaround strategies for modeling the performance of adaptive façades. Case A represents the discrete approach that combines a number of short-term simulations. Case B represents the approach that assembles the

results of simulations with static façades during post-processing.

An alternative approach uses separate models for the whole simulation period, each with static properties that represent different states of the

adaptive building envelope system. At a post-processing stage, the results of these independent simulation models are combined in a single

representation of the performance of the building, according to a certain control strategy for the adaptive façade (Figure 3.1B). Examples of this

approach are presented by DeForest et al. [2013], who used simulations in the EnergyPlus-based program COMFEN, to predict the performance of smart

windows that switch optical properties in the infrared wavelength range, and

53

by Du Montier et al. [2013], who used IES VE to predict the performance of

movable insulation panels. This modelling approach can have the advantage of (i) mimicking more advanced building operation controls than what is

offered by the simulation software and/or (ii) simulating adaptive building envelope technologies and materials for which a model does not exist yet.

Even though such a modelling method is well able to capture switching of instantaneous solar gains, for example, due to changing window-to-wall ratio

[Goia and Cascone 2014] or glazing properties [DeForest et al. 2013; C. S. Lee 2017], it fails to account for the effect of delayed thermal response due to the

capacitance of building components (i.e. slabs, walls and internal partitions). Therefore, in cases where thermal mass is involved in adaptive building

envelope operations, the use of these approximate models would probably lead to significant errors in the results, because they do not correctly handle

transient thermal energy storage effects [Erickson 2013]. These inaccuracies may eventually compromise decision-making based on simulation outcomes,

but little information about this issue is reported in the literature.

3.2.4 Comparison between modeling approaches

This section presents the results of an inter-model comparison that was set

up to assess the impact of different adaptive façade modeling approaches on the accuracy of simulation results. The simple, discontinuous approach,

which combines the results from separate simulation runs with fixed properties (Figure 3.1B) is compared with the “exact” solution, obtained with

run-time adaptation using ESP-r. The simulation results that are analyzed concern the opaque exterior construction (Rc value: 5 m2K/W) of a single-

floor detached residential building in the Netherlands. Thermotropic coatings, which can change the surface properties of a material depending

on temperature [Karlessi et al. 2009; Park and Krarti 2016], are investigated as the adaptive façade concept. At low temperatures, these thermotropic

layers absorb most of the incoming solar radiation, whereas at high temperatures, the coatings help reduce cooling load via increased reflection

and/or enhanced longwave radiation exchange. Different thermotropic technologies are currently under development, but most of them are still in

the earlier phases of the innovation process [Agrawal and Loverme 2011; Bergeron et al. 2008; Zheng et al. 2015].

54

Two types of thermotropic coatings were investigated in this study. This was

done by changing the properties of all opaque interior or exterior surfaces according to the values in Table 3.2.

Table 3.2. Material properties of the thermotropic coatings.

Low High Default Absorptivity (α) (λ: 0.28 – 2.8 μm) 0.3 0.7 0.65 Emissivity (ε) (λ > 3 μm) 0.3 0.9 0.84

The coatings are modelled to switch instantaneously, but only one of the properties at a time. This means that when the case with variable absorptivity

is investigated, the value for emissivity is left in the default state.

Thermotropic coatings are an example of self-adjusting, intrinsic adaptive

façades [Loonen et al. 2013]. In this study, the threshold surface temperature for state switching was assumed to be 20°C. Two situations are investigated.

The first one considers application of the thermotropic coating on the exterior surface. The second situation investigates application on the

innermost surface.

Results – outdoor application

Figure 3.2A shows the surface temperature of the exterior roof layer for three days in summer (4 - 6 July). In the situation with fixed high absorptance

(dashed line), higher temperatures are reached than is the case for fixed low absorptance (solid black line). Surface temperature of the thermotropic

coating closely follows one of the two states with static properties around the switching point of 20°C.

(A) (B)Figure 3.2. Exterior surface temperature. (A) Thermotropic α coating and fixed low and high absorptivity (4 – 6 July); (B) Thermotropic ε coating and fixed low and high emissivity, (30

August – 1 September).

0

10

20

30

40

50

Tem

pera

ture

[oC]

High abs. Low abs. Thermotropic α

day 1 day 2 day 30

10

20

30

40

50

Tem

pera

ture

[oC]

Low ems. High ems. Thermotropic ε

day 1 day 2 day 3

55

Similar behavior is observed in the results with variable emissivity (Figure

3.2B, period: 30 August – 1 September). In this situation, a temperature difference between the high and low case is not only present during the day,

but also at night when the radiant heat transfer coefficient from the roof to the sky and surroundings changes as a result of the difference in emissivity

value.

To evaluate the effect of using the two different modelling strategies, a comparison of heating energy consumption and thermal comfort, predicted

by the two methods is presented in Table 3.3. The differences in heating energy consumption are very small (less than three percent). The difference

in discomfort hours is also negligible. Use of the simplified, discontinuous, modelling approach in this case could therefore be justified, because the

predicted difference will likely not lead to a different recommendation or design decision.

Table 3.3. Comparison of results for the two modelling approaches (discontinuous and run-time).

Discontinuous “Exact”

Error

Abs. Rel. Heating Energy (kWh)Thermotropic α 2492 2525 33 1.3%

Thermotropic ε 2321 2393 72 2.9%

Thermal Comfort (wPPDh)Thermotropic α 65 67 2 3.0%

Thermotropic ε 74 76 2 2.6%

This result is not unexpected because the coating is applied outside of the thermal insulation layer. Therefore, temperature changes immediately follow

switching actions, because almost no thermal energy is stored in the construction.

Results – indoor application

Thermotropic coatings are not only applied to the exterior surfaces of

buildings, but can also be useful for indoor applications, especially to control the release of energy to/from constructions with thermal mass. Figure3.3A

shows the indoor surface temperature for the case with an internal thermotropic emissivity coating for the period 7-10 March. In contrast to the

exterior application, the thermotropic coating is in direct contact with

56

materials that have high thermal storage capacity. Because the switching of

surface properties in this case significantly influences the thermal history of the construction, there is hardly any overlap between the exact solution and

any of the two static surface temperature curves. In terms of surface temperature, the thermotropic coating is therefore not well-represented by

either one of the static cases. A discontinuous modelling approach would therefore not lead to reliable results, especially when high-resolution thermal

comfort evaluations are desired.

(A) (B)Figure 3.3. Interior surface temperature. Thermotropic ε coating and fixed low and high

emissivity during two periods: (A) 7-10 March, (B) 27-29 April.

Figure 3.3B shows that the thermal mass phenomenon can even lead to more unexpected effects (27-29 April). In several periods in this interval, the

temperature of the thermotropic layer rises higher than what would have happened in the static case. The exact reason for such effects is hard to

identify, but comes from multimode heat transfer effects, including non-linearities of longwave heat transfer and convection regimes. The occurring

heat transfer phenomena also depends on whether, during state transitions, the construction is already charged with energy, relative to its surroundings

and heating system operation.

The effect in Figure 3.3B is not reproducible with discontinuous simulations. Only when adaptation of construction properties takes place during

simulation run-time, it is possible to analyze and quantify the impacts of this effect.

18

20

22

24

Tem

pera

ture

[oC]

Low ems. High ems. Thermotropic ε

day 1 day 2 day 3 day 418

20

22

24

Tem

pera

ture

[oC]

Low ems. High ems. Thermotropic ε

day 1 day 2 day 3

57

Summary

The above presented simulation study has demonstrated that:

⎯ Thermal mass has a big influence on the proper selection of

performance prediction strategies for adaptive façades.

⎯ In cases where the operation of adaptive façades is decoupled from

thermal storage (e.g. exterior coatings with varying surface properties), a decoupled simulation approach is adequate.

⎯ When the operation of adaptive façades does affect the amount of

energy stored in the thermal mass, these dynamic effects have to be

taken into account during simulation run-time.

⎯ The simplified approach is not always able to capture all heat transfer phenomena during RBE state transitions.

⎯ Choosing a non-appropriate simulation strategy can lead to significant prediction errors that, in turn, can result in sub-optimal

design decisions.

3.3 Control aspects – modeling and optimizing

the operation of adaptive façades

3.3.1 Control considerations

In buildings with adaptive façades, there is a strong coupling between design

and (façade) control aspects. It is not difficult to understand that a good control strategy is of principal importance for materializing the performance

potential of adaptive façades during operation. To identify directions for high-potential adaptive façade design concepts, these control considerations

therefore deserve significant attention.

Two types of supervisory strategies for automation of dynamic systems in buildings are commonly distinguished: rule-based control (RBC) and model-

based control (MBC) [Oldewurtel et al. 2012; Henze and Neumann 2011]. RBC

is powerful for simple control problems. It connects measurements (sensors) to actions (actuators) via if-then-else control statements. These heuristic

58

control strategies, often represented in the form of decision trees [CIBSE

2009], can, however, get excessively complex in situations with multiple interacting sub-systems [Spindler and Norford 2009; Liu et al. 2015]. As a

consequence, RBC lacks the flexibility that is necessary to ensure high building performance in complex multi-input, multi-output problems.

One of the difficulties of RBC comes from the fact that sensed environmental

variables, setpoints, thresholds, and the order of control actions have to be predefined. MBC approaches, on the other hand, are less rigid, because the

best control strategy can be found after comparing different control

scenarios on the basis of model predictions. MBC approaches describe what the controller should achieve (e.g. minimizing energy use while maintaining

comfort) without prescribing how the controller should do this. This flexibility increases the option space of the control system to reconcile varying performance trade-offs over time in a way that best responds to the

dynamically changing conditions as they occur during building operation. Whereas RBC is a rigid approach that decouples optimal design and

operational aspects of a system, MBC integrates the two and allows for concurrent optimization of both features [Evins and Orehounig 2014]. This is

the main reason why the use of MBC is explored in this research.

3.3.2 Model-based control of adaptive façades

Figure 3.4 introduces the principles of the optimization-based approach for

offline control of dynamic building shell properties that is developed in this

thesis. In this simulation-based, in silico study, the building in the upper part of Figure 3.4 is represented by a whole-building performance simulation

model with actuators in the form of time-varying building envelope properties. This model can be considered as a detailed virtual representation

of the real-world building with adaptive façades.

59

Figure 3.4. Relationships between different components in the model-based adaptive façade control strategy

The task of the building shell controller, as represented by the bottom part in

Figure 3.4, is to drive the search for control sequences (i.e. time varying façade system states) that would optimize building performance. Each of

these control sequences represents a possible scenario of dynamic façade behavior, which is evaluated in the second building model that is integrated

in the building shell controller. This so-called embedded building model is repeatedly called to evaluate the performance of various sequences of time-varying façade properties over a predicted optimization horizon. It takes into

account scenarios for future ambient conditions and occupant behavior. The search for the best control sequence is driven by an optimization algorithm

that ranks the performance of each solution (i.e. sequence). Façade objectives/goals are transformed in an objective/penalty function, based on

the difference between desired (i.e. ideal) and actual responses. The best control sequence is found by minimizing this penalty function within the

optimization window. When an optimization loop is completed, the control sequence with the best performance is selected, and the corresponding set

of dynamic façade parameter values for the current control horizon is transferred from the controller to the building.

MBC is based on iterative, finite-horizon optimization runs with function

evaluations in the embedded building model. Three different time horizons/periods are considered in this control approach, as Figure 3.5

illustrates.

Sensors Actuators

Input OutputBuilding model

Objective function

Future disturbances

Best control sequence

Optimization algorithm

60

Figure 3.5. Illustration of different time horizons in the building and controller models. In this situation with three optimization horizons, each control horizon consists of six adaptation periods. Note that only the shaded façade states are actually applied to the

building.

⎯ Optimization horizon: Using a sequence of multiple, relatively short

optimization horizons (e.g. hours or days) in the building shell controller, the simulation process successively steps forward in time.

This approach is used to evaluate the longer-term (e.g. annual) performance of adaptive façades. Selecting an adequate length of the

optimization horizon should find a balance between conflicting requirements. The look-ahead period needs to be sufficiently long to

exploit dynamic energy storage effects and include future occupancy changes, but short enough to keep prediction uncertainties and the

size of the search space manageable [Zavala et al. 2009; Gunay et al. 2014].

⎯ Control horizon: In MBC approaches, feedback is introduced by only

implementing the first step(s) of the optimized control strategy in the building. Then, calculations are repeated, starting from the new

current state, and shifting the optimization horizon forward in time.

It is for this reason that MBC is sometimes also referred to as receding

time horizon control. The term control horizon is used to describe how many façade states from one optimization horizon (open-loop) are

applied to the building. The length of a control horizon can be as short as one adaptation period until the length of the whole optimization

horizon.

Optimization horizon

Control horizon

Adaptationperiod

Time

61

⎯ Adaptation period: The adaptation period represents the shortest

period of time during which façade properties are constant. Each

adaptation period is therefore characterized by one façade state. One adaptation period can consist of multiple simulation time steps. It

generally makes sense to synchronize the adaptation period with the rate-of-change in exterior boundary conditions and occupancy

conditions inside the building. In most practical building performance simulation studies that use typical meteorological years

(TMY) or other standard weather files, this means that the adaptation period is at least one hour.

3.3.3 Model resolution

To ensure good performance in MBC approaches, it is important that there is a high level of agreement between controller model predictions and the

situation as it would actually take place in the building [May-Ostendorp and Henze 2013]. Achieving this good fit in real-time building-integrated model

predictive control (MPC) tends to be a challenging task, because constraints on computation time prompt developers to use models that either have

relatively low modeling resolution (e.g. reduced-order models) or are not based on first principles (e.g. grey box models) [Prívara et al. 2012]. In

addition, difficulties with parameter estimation and setting of initial conditions may compromise the performance of full-scale MPC [Chen and

Athienitis 2003]. In this research, which is situated in the pre-design and product development phase, the optimization framework is used to calculate

adaptive façade control strategies “offline”, i.e. not coupled to the operation of a real-world system [Coffey 2013]. Computation-time reduction measures

are therefore not a primary concern, and as such, it appears adequate to make use of controller models at a high modeling resolution [Clarke et al. 2002;

Coffey et al. 2010; Hoes et al. 2012]. The integration of high-resolution models in the MBC formulation offers a number of advantages, because the method:

⎯ preserves the interrelated physical effects of all energy flow paths in the simulation process. As a result, energy and comfort performance

can be analyzed at a high level of detail (see Section 3.4 for more details).

⎯ does not rely on availability of model identification data, which is

favorable considering the exploratory nature of the optimization

problem.

62

⎯ is scalable, and can be applied to different climates and building types.

⎯ offers flexibility to deal with loosely specified and innovative building

envelope components (as long as they can be modeled in the selected

simulation program).

⎯ uses dynamic models that take transient effects into account in an explicit way.

⎯ allows for defining a performance bound8 based on “ideal” control actions. The predicted performance of adaptive façades is not

affected by interference from model-plant mismatch.

3.4 Performance aspects and performance

indicators

3.4.1 Thermal comfort and energy performance

Energy performance of buildings is closely linked to the levels of indoor

environmental quality they maintain. Better indoor comfort can generally be obtained by increasing energy use intensity, and vice versa. In this

dissertation, the focus is on analyzing the impact of adaptable façade design on the trade-offs between indoor environmental quality and the energy

consumption for heating, cooling and artificial lighting. In terms of space conditioning for thermal comfort, the goal is to keep the building in free-

running mode (i.e. no energy consumption for active heating or cooling) as much as possible. This goal is achieved by adjusting façade properties in such

a way that operative temperature is steered toward the deadband zone for HVAC control.

The HVAC system in this study is assumed to be an ideal system. Therefore,

temperature limits can always be satisfied in situations when façade

adaptation alone is not sufficient. As a consequence, thermal comfort is not considered as a primary performance indicator, but its importance is

8 The performance bound is a virtual construct that describes the performance of an ideal controller with an infinite

prediction horizon and perfect knowledge about the system dynamics and all future disturbances. It is often used

as a benchmark in control systems design, because no other controller will perform better [Oldewurtel et al. 2012;

Hoes 2014].

63

incorporated by imposing it as a constraint in the optimization process. The

performance of various façade alternatives is directly reflected in the energy consumption that is necessary to achieve satisfactory thermal comfort

conditions.

3.4.2 Visual comfort

Façade-related visual performance is influenced by a complex interplay among numerous different factors [Reinhart and Wienold 2011]. In terms of

daylight, an ideal façade would continuously provide, at least, (i) sufficient levels of well-distributed daylight illuminance, (ii) the absence of discomfort

glare for all occupants, and (iii) full view to the outside. In this research, three performance indicators are considered for evaluating to what extent a

certain adaptable façade configuration is in agreement with this ultimate solution:

⎯ Useful daylight illuminance9 (UDI). Daylight utilization is assessed by

computing the percentage of occupied hours that daylight

illuminance on the work plane falls within certain bounds. Four different categories according to the classification of [Mardaljevic et

al. 2012] are considered: fell short (<100 lux), supplementary (100-300 lux), autonomous (300-3000 lux) and exceeded (>3000 lux).

⎯ Daylight glare probability (DGP). Discomfort glare is assessed using

DGP, as it was shown to be the most robust glare metric in a comparative study under a wide range of ambient conditions

[Jakubiec and Reinhart 2011]. Research that explores the impact of view direction and user occupant influence on glare discomfort is still

in the early phases [Sarey Khanie et al. 2017; Bakker et al. 2014]. Throughout this thesis, a conservative glare scenario is considered in

which the occupant’s view direction is always facing the façade, to investigate the potential of a dynamic façade to mitigate glare risk

under all conditions. Three different categories for discomfort glare are distinguished, according to the classification in [Reinhart and

9 In the original definition, UDI is calculated for whole-year periods. Because this research also uses shorter

simulation periods, the results might not always be directly comparable with results from other studies.

64

Wienold 2011], where the risk is (i) intolerable (DGP > 0.45), (ii)

disturbing (0.35 < DGP < 0.45), or (iii) not disturbing (DGP < 0.35).

⎯ View to outside. Although the effects on visual performance and user

satisfaction are difficult to quantify, it is widely believed that improved view type and quality can affect these aspects in a positive

way [Aries et al. 2010; Hellinga and Hordijk 2014]. In this research, view performance is assessed by computing the potential for visual

interaction with the outside environment. This is done by taking the sum of averaged façade element transparencies over a period of time,

and quantifying this relative to the fully-glazed configuration (i.e. 100% view). This PI is therefore not based on simulation output, but

directly calculated for each façade configuration sequence on the

basis of visible transmittance.

65

4

Development of a toolchain for

performance optimization of

adaptive façades

4.1 Introduction

Taking the functional requirements and performance indicators that were introduced in the previous Chapters as a starting point, this Chapter

describes the details of how these specifications were interpreted in the form of a software implementation. This Chapter is split into five parts that each

cover a different aspect. Section 4.2 introduces characteristics of the simulation strategy for daylight and energy performance prediction of

façades with time-varying properties. Section 4.3 gives a higher-level description of the co-simulation approach that facilitates (i) the mutual

coupling between these energy and daylight simulation models, and (ii) the coupling between models representing the adaptive façades’ ‘real’ building

model, and those embedded in the building shell controller. The code modifications necessary for the coupling between the dynamic thermal

simulation program (ESP-r) and the middleware software that facilitates co-

66

simulation (BCVTB) are described in Section 4.4. Section 4.5 gives an overview

of the integration of four different software tools in the MBC approach. Finally, Section 4.6 presents how to various parts come together, by giving a

graphical overview of the interaction among all elements in the toolchain.

4.2 Performance simulation of adaptive façades

4.2.1 Dynamic daylight simulations

Daylight simulations are carried out using the Radiance three-phase method.

This method is a validated, new addition to the Radiance suite of daylight simulation tools [Ward and Shakespeare 1998], which was specifically

developed to enable high-resolution annual daylight performance prediction at reduced computational cost [Saxena et al. 2010; McNeil and Lee 2013]. It is

called the three-phase method because the calculation of light transport between sky patches and illuminance sensor points is separated into three

parts: exterior transport, fenestration transmission and interior transport.

Each phase of light transport is simulated independently and stored in a matrix form. The resulting climate-based indoor illumination (sensor points

or rendered images) is obtained using matrix multiplication [McNeil and Lee 2013]. The main reason for using the three-phase method in this research is

the possibility for analyzing the dynamic performance of multiple façade configurations in a quick, yet accurate way. By splitting up the daylight

prediction in three parts, different façade/fenestration options can be tested, without recomputation of the other two parts (i.e. light transport

inside and outside the building). The reuse of the already calculated ray-tracing results is responsible for a significant increase in computational

efficiency compared to traditional daylight simulation (e.g. Daysim or

rvu/rpict modules in Radiance), and therefore especially useful for modeling scenes with variable fenestration optics [Saxena et al. 2010].

To fit the specific purpose of this work, i.e. optimization of adaptability in

adaptive façades, a slightly modified workflow for the “standard” three-phase

method was developed, which puts emphasis on the changeability of individual elements in the façade (Figure 4.1).

67

Figure 4.1. Schematic overview that shows the interactions between the various components of the daylight performance prediction strategy.

The simulation process in this workflow is subdivided into a pre-processing

phase and a run-time phase as follows:

Pre-processing

i Sky vectors are generated in gendaylit for each daylight hour in the

simulation period. The Radiance function gendaylit works with the

Perez all weather model [Perez et al. 1993] to create exterior sky scenes using a standard weather file with hourly values as input for

direct and diffuse solar irradiance.

ii Daylight matrices (DMX) represent the luminous flux from sky

divisions to façade elements. They are calculated using rtcontrib to

establish the relationship between each façade surfcace and the sky patches from step 1, using a daylight coefficient method coupled to a

Reinhart sky division of 2305 patches plus ground surface.

iii Optical façade properties are characterized by bidirectional scattering distribution functions (BSDF). These transmission matrices

are calculated in WINDOW7 software [LBNL 2017b]. One BSDF is created for each of the possible fenestration properties, which are

then assigned to the transparent parts of the façade. Using this

68

approach, both specular windows and optically complex light-

redirecting glazing systems can be evaluated.

iv For each façade element, rtcontrib is used to create indoor view matrices (VMX), which model interior light transport by relating outgoing façade directions to observation points and planes in the

zone. To be able to calculate both UDI and DGP, (i) a grid of sensor points for daylight illuminance evaluation, and (ii) a fish-eye

projection image for assessment of glare discomfort are defined.

Run-time

i The contribution that each individual façade surface/element makes

to overall daylighting performance is calculated on a time step basis, by combining sky vectors, DMXs, BSDFs and VMXs in the

dctimestep module. These calculations are based on fast matrix multiplications; no ray-tracing is performed during this run-time simulation phase.

ii The total daylight illuminance coming from all different transparent

façade surfaces/elements is reconstructed using rlam and rcalc for a grid of sensor points. The values for daylight illuminance obtained

in this step are later also used to calculate the required energy demand and internal gains for artificial lighting.

iii A superposition approach using pcomb is employed to generate a single high dynamic range (HDR) image for the entire scene. This

image is supplied to the evalglare program [Wienold 2015] , which is used to calculate glare discomfort performance.

In the daylight performance prediction approach described above, the majority of computational load for the simulations is shifted to the pre-

processing phase, which takes place before the actual optimization process. Within the run-time optimization loop, only fast matrix multiplications are

needed; no ray-tracing calculations. A validation study by McNeil and Lee [2013] shows that the increased simulation speed does not come at the

expense of inaccuracy of predicted results. Currently, the bottleneck in computation time comes from the relatively resource-intensive pixel analysis

of DGP calculations in evalglare. Future work should investigate the potential of using other glare indices (e.g. simplified DGP) [Wienold 2009] or calculation methods (e.g. radiosity-based) [Kleindienst and Andersen 2012]

for this part of the analysis.

69

4.2.2 Building energy simulation with adaptable façade properties

Building energy simulations (BES) are carried out using ESP-r [Clarke 2001]. ESP-r is an integrated, research-oriented BPS tool with a long validation

history and an active development community. The tool can be used for prediction of thermal, visual, indoor air, electrical and acoustic performance

of buildings and systems. Unlike most other high-resolution BES tools, ESP-r does already offer some possibilities to model building components with

time-varying thermophysical properties (Section 3.2.2). In this research, these capabilities were further extended. The limited flexibility in the default

ESP-r “BCL99 control” facility for thermophysical property substitution, as implemented by MacQueen [1997], was avoided by re-programming parts of

the source code to replace these properties during simulation run-time by reading from an external signal. Section 4.3 will describe where this external

signal comes from.

Performance prediction with variable thermophysical properties requires

that the storage and transfer properties of energy conservation equations are estimated at each time step [Nakhi 1995]. The main difference with normal

ESP-r calculations for non-adaptive façades is that not just in the beginning of a simulation run, but during every time step, the building-side matrix

equation is re-established after reading new construction properties for the future time row [MacQueen 1997]. ESP-r has two main subroutines for setting

coefficients of the implicit difference equations. Normally, MZCOE1 sets up those coefficients (or partial coefficients) of the equations that are constant. For regular building envelopes, the thermophysical properties (thermal

conductivity (W/m.K), density (kg/m3) and specific heat capacity (J/kg.K)) are independent of time and need therefore only be computed once per

simulation run. Subroutine MZCOE3, on the other hand, initiates the computation of the energy transport that varies with time (e.g. solar gains

impinging on internal and external surfaces, external longwave radiation exchanges, internal casual gains, and ventilation exchanges), and its

coefficients are therefore computed during every time step. In the present implementation, the thermophysical building envelope properties are made

variable by making extra calls to MZCOE1 on a time-step basis, after reading updated construction properties for recomputing the matrix coefficients.

This step happens right before the call to MZCOE3.

70

4.3 Co-simulation approach

To accomplish implementation of the performance prediction framework as proposed in the previous Chapters, a UNIX-based toolchain was developed

that integrates the solvers for daylight (Radiance) and energy performance (ESP-r) prediction together with optimization routines and scripts for

receding time horizon control. This so-called co-simulation approach is a valuable method to deal with this type of coupled simulation studies, and has the advantage that it can reuse existing code, models and toolboxes in

distributed software programs [Trcka et al. 2009; Wetter 2011b].

In the present toolchain implementation, the co-simulation process is supervised by master algorithms implemented in Matlab [Matlab 2015]. The

master algorithms are responsible for starting and finishing the simulation, preparing simulation input files, extracting outputs, and a number of other

tasks as will be described in the next Sections. Matlab was chosen for these tasks because it offers a high-level programming language in a flexible

environment.

The actual run-time communication (i.e. data exchange) between the various

programs in the toolchain takes place through the Building Controls Virtual Test Bed (BCVTB) as a middleware tool [Wetter 2011b]. BCVTB uses a quasi-

dynamic coupling scheme (Figure 4.2). The inter-process communication (IPC) between software programs happens at fixed time steps, without

iteration. Many simulation programs, including Matlab and Radiance, are already linked to the BCVTB. ESP-r has been coupled to the BCVTB as part of

this research10. This software development was needed to enable controlled adaptation of building shell properties during simulation run-time (Figure

4.3). More details about the changes that were made to ESP-r are presented next.

10 The connection with ESP-r is now part of the standard BCVTB distribution. It can be downloaded from the BCVTB

website and includes example models and user instructions.

71

Figure 4.2. Quasi-dynamic coupling scheme used in BCVTB. The dashed lines indicate data exchange between the tools per time

step; t = time step, yt =data of tool 1 at time step t, ut = data of tool 2 at time step t.

Figure 4.3. Coupling of tools in the co-simulation framework. Master algorithms are implemented in Matlab. The communication through BCVTB takes place between ESP-r (‘real’ building) and Matlab

(building shell controller).

4.4 Coupling between ESP-r and BCVTB

The coupling between ESP-r and BCVTB is based on the FORTRAN90 libraries

that are included in the standard BCVTB distribution. The functions in these libraries make use of Berkeley software distribution (BSD) sockets; BCVTB’s

mechanism of connecting different software programs. They fulfill three main functions:

i establish a connection with BCVTB (establishclientsocket);

ii exchange data (exchangedoubleswithsocket);

iii close the connection (ipcclose).

To couple these functions with ESP-r, new subroutines were added, and other subroutines were adapted to integrate the additional functionality in

the source code of ESP-r11. In addition, a new header file (BCVTB.h) was added to all relevant subroutines to declare the parameters and variables that are

used for communication with BCVTB. The most important variables are those that are used during the data exchange with BCVTB:

11 A version of ESP-r with basic BCVTB functionality is available for download from the ESP-r source code repository

(branch ESP-r_BCVTB). It also includes example models and user instructions.

72

⎯ bcvtb_y: an array that contains all sensor variables sent to BCVTB;

⎯ bcvtb_u: an array that contains all actuator variables (control signals)

received from BCVTB.

Figure 4.4. Overview of ESP-r code modifications to enable coupling with BCVTB, showing modified subroutines (left) and new subroutines (right). ESP-r subroutine

names are underlined.

73

Figure 4.4 presents an overview of all the additions and changes that were

made to the ESP-r code. The calls to the new subroutines are made from

MZNUMA, the backbone subroutine of ESP-r that includes its simulation clock. Before the start of the first simulation time step, the connection with BCVTB

is established by calling bcvtbestablish. The second step in the preparation phase is to exchange initial values of sensor and actuator values

by calling bcvtbexchange. During the actual simulation loop, every time

increment in MZNUMA is interrupted with a data communication call via

bcvtbexchange. Finally, after the last time step in the simulation, the

connection with BCVTB is terminated, by calling bcvtbclose.

The coupling between ESP-r and BCVTB is designed to be as generic as possible. It can therefore be used for various purposes such as solar shading

operation, window opening strategies, or for controlling the temperature and/or flow rate in thermally activated building systems [Hoes et al. 2012;

Hoes 2014]. The specific application is determined by the user-defined ESP-r simulation variable(s) that gets overwritten with actuator signal(s) received

through IPC with BCVTB (i.e. bcvtb_u).

With respect to performance prediction of adaptive façades, the coupling between ESP-r and BCVTB in this research takes place at two different levels,

(i) between Radiance and ESP-r, and (ii) between Matlab (building shell control) and ESP-r. More details are described in the next sub-sections.

4.4.1 Coupling of visual and thermal domains to obtain integrated

results.

The basics of the Radiance-to-ESP-r integration between the thermal and

daylighting domains (Figure 4.1) are very similar to the principles for simulation of daylight-responsive building systems as described in e.g. Janak

[1997]. However, the implementation in the present work is different. Instead of using a direct link to Radiance functions in the ESP-r code, the coupling is

established through BCVTB (as indicated by the blue and red arrows in Figure

4.5). At a time-step basis, daylight illuminance values are computed, and fed into the lighting controller of the thermal simulation model, which then

assigns/updates the right level of internal gains corresponding to the current façade configuration and climatic conditions.

74

4.4.2 Data exchange and synchronized communication between

building model and building shell control.

The time-varying building shell properties are assigned to ESP-r through an external signal (Section 4.2.2). This external signal is based on the outcome of

the model-based control strategy, which is implemented in Matlab. In this process, BCVTB is responsible for synchronizing the data exchange between

both programs. Details of the coupling between building model and building shell control, and the role of BCVTB in this process, are described in the next

Section.

Figure 4.5. Schematic overview of coupling between the various software tools. The two large boxes indicate the master algorithms in Matlab. Black arrows indicate data

exchange through BCVTB.

4.5 MBC implementation using Matlab, BCVTB,

ESP-r and Radiance

4.5.1 Implementation of the receding time horizon control

approach

To achieve the MBC performance evaluation as described in Section 3.3, a procedure is proposed in which an instance of the same high-resolution BPS

model that is used to predict adaptive façade performance, is also embedded in the controller that determines its operation. With respect to their

Build

ing

Build

ing

shel

l con

trol

75

connection to BCVTB, there are two main differences between these models,

as is also indicated with the differences between left and right in Figure 4.3:

⎯ Contrary to the ‘real’ building model where communication takes place as simulation time proceeds, these external simulation calls

from Matlab to the embedded building model are run from start to finish, for the length of one optimization horizon (Figure 3.5). The

calls are implemented via Matlab system commands to (i) Unix shell

scripts, for ESP-r simulation (bps) and results extraction (res), and (ii) the run-time part of Radiance functions from Section 4.2.1.

⎯ For the model that represents the ‘real’ building, the coupling is

directly integrated in ESP-r’s algorithms (Section 4.4). In the building shell control, BCVTB does not directly communicate with the

simulation tools, but with a master algorithm, implemented in Matlab (Figure 4.5). Apart from this communication, the master algorithm

takes care of time step bookkeeping, preparing simulation input files with correct start/stop time, making calls to Radiance and ESP-r (i.e.

new instantiations of the embedded building model), handling post-processing, and managing the optimization procedure

The optimization procedure in the receding time horizon formulation

consists of seven steps. A graphical representation is given in Figure 4.6.

76

Figure 4.6. Schematic overview of the optimization procedure. Each cell represents one façade configuration, while different colors (greyscale) show how it changes over time. The numbers correspond to the different steps in the optimization procedure described

in Section 4.5.1. Only three iterations of the optimization algorithm are shown. The arrow (step 5) indicates the iterative loop that ends when a stopping condition is

satisfied.

i At the beginning of time step tn, simulation of the ‘real’ building is

temporarily paused, to start a new search for optimized control

sequences in the next control horizon (from tn to tn+m). The task of the controller is to decide if and how the façade configuration should be

adjusted in the upcoming period.

ii All state variables representing the thermal state of the building (i.e.

node and surface temperatures) after time step tn-1 are stored in a file.

iii New versions of the BPS model are instantiated at time tn. In these thermal simulations, the initial states of the models are overwritten

with stored state variable values inherited from the previous time step of the ‘real’ building simulation model. This step is needed to

ensure consistency of transient thermal effects, because the default initialization would not always lead to accurate results, as is

described in Section 4.5.2.

iv Multiple scenarios with different façade adaptation strategies are evaluated in the MBC embedded BPS models (both Radiance and ESP-

r). These simulations are executed over a future optimization horizon

77

of m time steps, assuming perfect knowledge about future boundary conditions and building use. An overview of source code

modifications to accomplish this is presented in Section 4.5.3.

v The search for optimal control sequences is driven by a selected optimization algorithm which iteratively suggests new sequences of

façade parameters as it attempts to minimize an objective function (Section 4.5.4). The optimization loop is repeated until the given

stopping criterion (e.g. based on time or relative performance improvement) for the current optimization horizon is satisfied. In this

study, genetic algorithms are used. Additional considerations about the search algorithm and its modifications are discussed in Section

4.5.5.

vi The values of time-varying façade properties that together constitute

the optimal control sequence for the current control horizon (open-loop) are sent as actuator values from Matlab through BCVTB to the ‘real’ building simulation model. In addition, the artificial lighting

needs that are required to complement daylighting in the selected scenario are sent to the thermal model.

vii Performance simulation of the ‘real’ building simulator is resumed

with optimized dynamic façade states from step 6 and continues until an updated adaptive façade sequence is requested at the end of the

control horizon (Figure 3.5). This cycle repeats until the total simulation period is expires.

4.5.2 Explicit state initialization

Despite multiple favorable aspects of the use of high-resolution BPS models

as function evaluators in simulation-based or experimental MBC studies [Clarke et al. 2002], the number of applications in literature is very limited

[Prívara et al. 2012]. One of the main drawbacks of using detailed BPS tools for MBC are the time-consuming pre-conditioning periods. In reduced-order

models or data-driven models, one can explicitly impose the correct initial conditions of embedded building models to avoid continuity erros. With

detailed BPS models, warm-up periods, which lead to considerable overhead in computation time, are usually needed to guarantee correct starting

conditions [Wetter and Haugstetter 2006; Coffey et al. 2010]. Depending on operational schedules and thermal time constants of constructions, this

78

initialization period may consume up to 80% or more of the actual simulation

time (e.g. four days of pre-conditioning for an optimization horizon of one day) [Nghiem and Pappas 2011; Corbin et al. 2013; Favoino et al. 2017].

In earlier MBC studies with BPS as embedded building model, it was not, or only partly [Coffey 2013] possible to remove/reduce this initialization period,

because the programs that were used (e.g. TRNSYS and EnergyPlus) did not allow access to the state variables (the underlying reason for this can be

found in the solution routines for energy balance equations, as described in Section 3.2.1). An advantage of ESP-r is that it does give the user access to the

building and system’s state variables. This flexibility was used to circumvent the need for repeatedly running initialization periods.

Clarke [2001] uses the following matrix notation to describe the set of conservation equations that are solved simultaneously during each time step

of the simulation: 𝑨𝜽𝒏 𝟏 = 𝑩𝜽𝒏 + 𝑪

where 𝑨 and 𝑩 are sparse matrices containing the nodal equation coefficients

for energy conservation that describe the coupling between nodal regions of

various construction elements in the model. The column matrix 𝑪 contains

the influence of ambient boundary conditions, and the column matrices 𝜽𝒏 𝟏

and 𝜽𝒏 contain the nodal temperature terms and heat injection/extraction at the future and present time-rows, respectively. To enable explicit state

initialization of the embedded building model, state variables in the matrix 𝜽𝒏 of the last time step of the ‘real’ building model, before the start of a new MBC loop, are stored in a file. These values are then used to overwrite the default

initial values of the embedded building model. To completely avoid the computational overhead caused by repeated initializations, a simulation start-up period of zero days was used in the embedded building model.

4.5.3 Code modifications in ESP-r

Previous Sections already introduced the ESP-r code modifications that were needed to include time-varying façade properties (4.2.2), and to connect ESP-

r with BCVTB (Section 4.4). This Section discusses the modifications that were additionally needed to implement the receding time horizon control

formulation, in both the ‘real’ building model and the embedded building model.

79

‘Real’ building model

State variables of the thermal model are written to an external file at the right

time step during the simulation, i.e. at the end of the control horizon, just before a new call to the embedded building model is made. A new variable,

MBCupdate with an incremental counter is added to determine the right time step.

Embedded building model

Because the ESP-r algorithm of the embedded building model is not

connected to BCVTB, another way of introducing the dynamic façade properties and lighting gains to ESP-r was implemented here. Matlab writes

each façade sequence as suggested by the optimization algorithm to an external file. This file is opened by ESP-r, and in every time-step of the

simulation, the corresponding values are read and updated by the program.

The ESP-r code is designed to simulate full days, always starting at 00:00 h

and running for x times 24 hours. This can be inconvenient in a receding time horizon control formulation, because it restricts the flexibility of horizon

sizes. However, to be able to start a simulation in the middle of the day would require a major overhaul of ESP-r, with impacts in many parts of the code.

The workaround solution that was implemented to simulate incomplete days (either starting before 00:00h, ending before 24:00h, or both) is presented in

Figure 4.7.

Figure 4.7. Optimization horizon starting after 00:00h. t1 = start of the optimization horizon; t2 = end of the optimization horizon. Explicit state initialization happens at t1.

Simulation results before t1 and after t2 are discarded.

An extra variable MBCtsread was added as a counter for the number of simulation time steps until t1; the time at which the state variable are

overwritten. ESP-r receives control signals with “dummy” variables for the

80

periods before t1 and after t2, and all results in these periods are not included

in the performance evaluation.

4.5.4 Objective function

Function evaluations in the embedded building model are ranked and compared with respect to an objective function that takes the multiple

performance criteria of interest (Section 3.4) into account. More specifically,

performance is expressed using penalty functions (P) and weighting factors

(w) for both energy and daylighting aspects.

As described in Section 3.4.2, visual comfort evaluation is further subdivided

into three components: UDI, DGP and view to outside. These factors are also embedded in the objective function so that the desired goals can be achieved

by optimization. The approach adopted in this research is similar to the concept of “daylight credits” [Gagne and Andersen 2012; Kleindienst and

Andersen 2012] and “preference functions” [Mahdavi 2008]. However, to allow for minimization of both performance criteria (energy and visual

comfort) in the mathematical sense, the credits are transformed into penalty functions. Figure 4.8 introduces the penalty functions for daylight

illuminance, glare and view.

Figure 4.8. Penalty functions for daylight performance aspects.

For each daylight hour in the optimization horizon, penalty values are

computed for the various façade configurations, and then aggregated into one value using weighting factors that are determined by the decision-maker:

𝑃 = 𝑤 ∙ 𝑃 + 𝑤 ∙ 𝑃 + 𝑤 ∙ 𝑃 4.1

81

By combining daylight performance with primary energy consumption for

heating, cooling and lighting, the total objective function can be described as:

𝑃 = 𝑓 𝑤 ∙ 𝑃 + 𝑤 ∙ 𝑃 4.2

Further enhancements to this objective function formulation, and ways of assessing tradeoffs between energy and daylight performance are included

in the next Section.

4.5.5 Genetic algorithm + modifications

A genetic algorithm (GA) is used to drive the parameter search in the

optimization of façade properties. GAs are metaheuristic, population-based optimization algorithms. Numerous studies have shown that these algorithms

are effective and robust for reconciling the competitive requirements in optimization of building envelope design based on “black-box” simulation

models, such as ESP-r, EnergyPlus and Trnsys [Evins 2013; Attia et al. 2013]. In this study, the task of the algorithm is to propose new façade adaptation

sequences, to be evaluated in the controller model. This approach is markedly different from conventional design optimization studies where the algorithm

acts on time-invariant parameters of a simulation model.

Because the properties of each adaptable façade element can be described as

a discrete set of predefined options, the problem is encoded in an integer optimization formulation, with one identifier per façade element. GA

functions in Matlab were used in this implementation. To reduce optimization time, the default GA algorithm was extended with code that

avoids already simulated solutions from being evaluated again.

The dynamic behavior in adaptive façades raises a few specific challenges for the optimization procedure. A number of additional measures were taken in

response to these challenges, to ensure a search process that is both efficient and effective. These are discussed in this Section (Figure 4.9).

82

Figure 4.9. Overview of process steps in the genetic algorithm. Implications of adaptive façade dynamics on four aspects (indicated with numbers) are further explored in this

Section.

1. Seeding of initial population

In addition to the default random way, a number of members in the initial

population are seeded with manually generated solutions. Seeding with high-potential solutions based on problem-specific knowledge is an established

technique for improving optimization efficiency with evolutionary algorithms [Hamdy et al. 2011], and has been mentioned as an important supporting

mechanism for reaching convergence in design optimization of cellular façades [Wright et al. 2014]. In the present approach, all uniform façade

configurations with non-adaptive properties are added to the initial population. In addition, the best solution from the same period in the

previous day (if available), and the tail part of the optimal solution in the previous optimization horizon (indicated in white in Figure 3.5), are added to

the initial population. Initial trials revealed that, without seeding, it is very difficult to obtain good optimization results in the large search space of

possible solutions. In the context of this project, seeding hopes to enhance optimization efficiency in the following ways:

⎯ Ensuring good spread and coverage of the entire design option space.

⎯ Making sure that end points of the design option space belong to the searched solutions. Previous research shows that lower and upper

bounds of the search space are the places where good solutions tend to be [Radford 1981]. This also minimizes the risk that “conventional”

solutions are missed by the optimization algorithm.

83

⎯ Speeding up convergence because genes of good candidate solutions

are likely introduced early in the search process.

2. Objective function and normalization

The dynamic optimization process seeks to improve two different performance aspects at the same time; energy consumption and visual

comfort. This translates into a two-dimensional performance space in which Pareto optimal solutions can be identified [Radford and Gero 1980]. The

optimization of adaptive façades is based on successive decision moments in the MBC control formulation. Because the dynamic optimization procedure

needs to move forward automatically (i.e. no human intervention), manual inspection of Pareto fronts in each control horizon is not possible. As an

alternative, performance preferences that rank the two non-commensurable

objectives have to be articulated a priori.

Which of the façade adaptation sequences gets identified as most preferred solution by the genetic algorithm depends on two elements:

i the relative magnitude of each performance indicator compared to

the other indicators (multi-objective optimization);

ii the range of potential gains that can be achieved within each criterion

(i.e. the swing from worst to best) [Choo et al. 1999].

In conventional façade design optimization, the weights in Equation 4.2 can be tuned in such a way that a balanced trade-off point is found that is in line

with the goals of the design team [Hopfe 2009; Mela et al. 2012]. In adaptive façade optimization, this tuning of weights is less trivial because the

importance of trade-offs is intrinsically subject to change. Due to dynamic

effects, the values for 𝑷𝒆𝒏𝒆𝒓𝒈𝒚 and 𝑷𝒗𝒊𝒔𝒖𝒂𝒍 can differ several orders of

magnitude from horizon to horizon. This can for example happen when the building is in free-running mode (i.e. no energy consumption) for an entire

optimization horizon; in this horizon, the energy penalty (Penergy) would be zero, whereas in the next horizon, a minimum energy consumption of several

kWh could be found. This large variability in the results makes it difficult to find balanced trade-off solutions with respect to the other performance

criterion, in this case visual comfort. As a consequence, it is not straightforward to determine fixed criteria weights that will lead to

satisfactory performance over the whole simulation period.

84

A solution to this difficulty is found by selection of optimal solutions with

reference to an anchor point. The position of this anchor point is recalculated in every optimization horizon by determining the performance of all possible

non-adaptive façade options for the current optimization horizon. Note that these non-adaptive solutions are part of the initial population as an outcome

of the seeding procedure described in step 1 above. Based on predefined weights that represent the relative importance of energy and visual

performance, these values are scaled in an iterative process (i.e. vector normalization) until the preferred non-dominated solution is located at the

position [1,1] in an xy-coordinate system. A graphical summary of this procedure is presented in Figure 4.10.

Figure 4.10. Overview of the multi-criteria decision-making approach. The left figure shows the performance for different adaptive façade system sequences (dots) as well as

reference solutions with fixed façade configuration (diamonds). The axis values are imaginary, but are chosen to indicate that the values for different indicators can be

widely different. The right figure shows the normalized performance indicators with an anchor solution at [1,1]. The circle arcs indicate the shortest Euclidean distance to the ideal solution [0,0], in this case indicating that many adaptive façade solutions can

improve performance on both objectives, compared to the best performing static façade solution.

The normalization weights are re-calculated for every optimization horizon and applied to the fitness function values of all adaptive façade solutions by

dividing the raw penalty value with the values of the solution at the anchor point.

Finally, the best performing sequence is selected using a TOPSIS decision-

making approach. This method finds the compromise solution as the point with the shortest Euclidean distance to the ideal solution, and the farthest

distance from the negative-ideal solution [Opricovic and Tzeng 2004; Triantaphyllou 2000], with:

Objective 1

Objective 2

Objective 1, normalized

Objective 2, no

rmalized

1.0

1.0

0.10.2

0.3

100 200 300Anchor point, best non-adaptive solutionNon-adaptive design solutionCandidate solution, adaptive facadeOptimum solution

85

⎯ Ideal solution: the Utopia solution (point [0,0]) [Zavala 2012] having

zero energy consumption and zero comfort penalty.

⎯ Negative-ideal solution: the anchor point [1,1] representing the best

possible non-adaptable solution. In this way, the distance from worst to best becomes an explicit part of the decision-making process in

the MBC formulation.

The mathematical representation for the TOPSIS case where normalized

penalty is calculated as the shortest Euclidean distance to the ideal solution is presented in Equation 4.3.

P = w ∙ PP | + w ∙ PP | [-] 4.3

3. Stopping criterion

Stopping criteria are an important consideration in population-based

optimization studies, as they determine the interaction between optimality of the results and computational efficiency. In most building-related

optimization studies, the optimization search is stopped after a maximum number of generations in the GA is exceeded [Attia et al. 2013]. In is not

uncommon, however, that the optimum solution lies at the edge of the search space, and is evaluated already as part of the seeded initial population. In such

cases, it is unnecessary to run the GA for all generations. A check was implemented to track the relative improvement in normalized objective

function values. If the relative improvement in equation 4.2 is less than 5% in five consecutive generations, we assume that convergence has been reached,

and terminate the optimization.

4. Recombination

Recombination, or crossover, forms an essential part of the working principle

of GAs. In its default implementation with uniform crossover, however, there is a risk that unfavorable crossover points hinder the convergence because

they can break the gene sequence of good candidate solutions. In order to prevent such effects, in the current approach, crossover points are only

allowed per entire façade sequence, i.e. between adaptation periods (Figure 3.5), not within. Mutation, on the other hand, can take place on any gene to

support sufficient diversity throughout the whole sequence.

86

5. Soft constraints

The optimization formulation as presented in this Chapter, gives rise to an

enormous number of possible façade adaptation strategies; the size of the search space grows exponentially as the optimization horizon gets longer.

However, as it turns out, many of these options lead to equivalent performance. For example, at times when the temperature difference

between inside and outside is small, the thermal resistance of the façade plays only a minor role. Similarly, at night, the influence of façade transparency can

be ignored. In terms of performance as it is defined in the objective function (Equation 4.2), it is often impossible for the optimization algorithm to

differentiate between all the different alternatives. The consequence is that there is a risk that random components or confounding variables enter the

decision making process in lieu of meaningful variations. When one is only interested in the performance potential of adaptive façades, these effects

have no direct consequences. However, when the aim is to also derive the properties of high-performance adaptive façade concepts, this can become

problematic, because it is not directly possible to tell which suggested adaptation actions actually drive the parameter search, and which come from

meaningless variations that go without physical explanation. To avoid such issues as much as possible, the objective function is extended with two

additional hierarchic preference terms in the form of a soft constraint [Kerrigan et al. 2000]. These preference terms are based on the following

three considerations:

⎯ Undue complexity of building envelope components should be

avoided as much as possible. As long as the performance differences are negligibly small, the façade sequence with lowest complexity (i.e.

most limited in frequency and extent of adaptation) is preferred.

⎯ Excessive changes in the façade zone may lead to increased chances

for disturbance and discomfort [Stevens 2001; Bakker et al. 2014]. From the perspective of user acceptance, the façade sequence that is

least obtrusive (i.e. most limited in frequency and extent of adaptation) is preferred, in cases with otherwise equal performance.

⎯ When the optimization algorithm cannot discriminate between

multiple options, it keeps unintentional diversity in the gene pool, rather than narrowing down on a specific subset of high-

performance solutions. Due to a lack of consistency, such erratic changes complicate the way to convergence. Additional preference

87

terms are useful for facilitating effective ranking of all alternative

solutions. There is a risk however, that the optimization gets stuck in a local minimum. Crossover and mutation settings need extra

attention to avoid premature convergence.

Taking Equation 4.3 as a starting point for calculating 𝑃 , these considerations translate into the following hierarchic preference terms: 𝑃 = 𝑃 + 0.01 ∗ 𝑃 , + 0.01 ∗ 𝑃 , [-] 4.4

Where: 𝑃 , limits the variation within an adaptation period: if

performance of more than one design alternative is equivalent on all other

performance aspects, the one with the most uniform appearance is selected;

and: 𝑃 , limits the variation from adaptation period to adaptation

period: if the performance of more than one design alternative is equivalent

on all other performance aspects, the one that requires the lowest number of changing elements is selected.

4.6 Overview of the toolchain

This Chapter is concluded by giving an overall view of the various elements

in the toolchain, and their interactions (Figure 4.11). Different combinations of control horizons and optimization horizons are possible in the current

implementation. This overview shows data exchange for the simplified example of an optimization horizon of 3 time steps with a control horizon of

2 time step (see also Figure 3.5).

The “sensor” variables y are sent to BCVTB by the ‘real’ building model, the

“actuator” values u are sent by the controller. State variables are only transferred at the start of a new optimization horizon (t=1 and t=3; indicated

with dark grey dots). Depending on whether it is in the middle of a control horizon (e.g. at t=2) or at the start of a new one, Matlab sends either data

stored from a previous optimization horizon (e.g. at t’=2) or invokes a new call to the optimization algorithm to receive the data after the optimization (ESP-

r and Radiance) run is finished. In the first situation, the data exchange happens very fast, whereas in the second case, possibly thousands of

scenarios are evaluated in the embedded building model, which may take minutes to hours.

88

Figure 4.11. Example of the data exchange per time step. t= time step, yt = sensor value at time step t, ut = control signal at time step t. Optimization horizon length is 3 time steps

and control horizon length is 2 time steps.

Figure 4.12 demonstrates how the toolchain steps forward in time. Three consecutive adaptation periods are shown, indicating the optimization

process finds different optimal values when more information about the future becomes available.

89

Figure 4.12. Example of façade optimization sequences over time. Three consecutive adaptation periods are shown. The length of the adaptation period in this example is

46h, the control horizon is 12h and the adaptation period is 6h.

90

91

5

Quantifying the performance

potential of adaptive façades –

a case study

5.1 Introduction

So far, this dissertation has mostly focused on the development of a methodology for computational performance optimization of adaptive

façades. This has resulted in a toolchain that couples different simulation tools for building energy and daylight performance prediction with a control

approach that can ensure high-performance adaptive façade operation. In the present and the following Chapter, two selected case studies are

presented to illustrate potential applications of the toolchain, and to show how it can be used to analyze the performance of innovative adaptive façade

concepts.

92

5.1.1 Background and objectives of the case study

The case study that is presented in this chapter is framed within the context of EOS-lt project FACET [Loonen et al. 2015]. The goal of the FACET project

was to identify the ideal adaptable building envelope properties for different building types in the Netherlands. To achieve this objective, the project

proposed an inverse approach. Instead of starting from a certain adaptive façade material or technology and analyzing its performance, the project

began by postulating the desired end result (i.e. optimal IEQ conditions with minimal energy expenditure) and a subsequent search for adaptable façade

system states that best approach the ideal end result. The hypothesis is that, based on this theoretically-derived optimum which can also include building

materials with hypothetical properties, promising suggestions for R&D paths can be outlined.

5.2 Description of the case study

5.2.1 Building model and simulation scenario

This study considers a typical single-zone perimeter office space with

dimensions (LxWxH) of 3.6 m x 5.4 m x 3 m. The office faces south, is located on an intermediate floor, and is surrounded by similar spaces and a corridor

at the back. Construction and material properties of the external façade are determined via optimization as described in Section 5.2.2; the properties of

the other construction elements are given in Table 5.1.

Table 5.1. Constructions and material properties of the case study.

Thickness(mm)

Conductivity(W/mK)

Density (kg/m3)

Specific heat (J/kgK)

Floor Floor tiles 6 0.60 500 750

Concrete blocks 140 1.06 1950 1000

Ceiling Suspended ceiling (gypsum) 13 0.42 1200 837

Internal walls Lightweight concrete blocks 150 0.44 1500 650

Plasterboard 12 0.18 800 837

93

Occupancy and usage

On working days, the zone is occupied by two persons. The office workers

are present from 8:00 to 17:00 and account for internal heat gains of 65 W per person (metabolism) and 120 W per person for appliances. For glare

evaluation, we assume that the persons are looking in the direction of the window (worst-case scenario).

Heating, cooling, lighting and ventilation

Operative temperature inside the room is controlled on the basis of adaptive

temperature setpoints according to EN 15251, class B. The setpoints vary over time according to the running mean outdoor temperature in the previous five

days [Nicol and Humphreys 2010]. Outside occupied hours, temperature setpoints for heating and cooling are adjusted to 17°C and 30°C respectively.

Dimmable artificial lighting with a maximum lighting power density of 10 W/m2 is used. Daylight is controlled automatically to maintain a minimum

work plane illuminance of 500 lux. Illuminance sensors for controlling artificial lights are placed in the center of the room at work plane level (0.9

m).

During occupied hours, outdoor air with a flow rate of 2 ACH is used for ventilation, with a heat recovery efficiency of 90%. The infiltration rate is

assumed to be constant, and set at 0.24 ACH.

Weather conditions

Simulations are carried out with the weather data reference year for energy

simulations for the Netherlands [NEN5060 2008]. The analysis focuses on two typical weeks, one in winter, and one in spring. Figure 5.1A shows incident

solar radiation on the south façade and Figure 5.1B shows ambient temperature for these weeks.

94

A

B Figure 5.1. Weather conditions (the Netherlands) during selected weeks in winter, spring

and summer.

Primary energy conversion

To make fair comparisons with respect to the various energy flows, overall energy performance is assessed on the basis of primary energy consumption.

In the case study, we assume that the seasonal heating efficiency = 0.9, cooling COP = 3, and the primary energy conversion factor for electricity =

2.56 [NEN7120 2011].

5.2.2 Adaptive façades: adaptable properties and adaptation

ranges

To analyze the effect of adaptability of adaptive façades, the external façade is subdivided into a number of smaller elements, or cells. The physical

properties of each of these elements can be controlled individually. Instead of having fixed walls and fenestration areas, such a decomposition enables

flexible optimization of shape, number, position, transparency and thermal properties of “window” elements. Previous research by Wright et al. [2014]

shows that a subdivision of the façade into smaller elements can be a valuable approach, because it allows for façade optimization in response to the

temporal and spatial characteristics of daylight. In contrast to previous work, however, this research not only deals with design optimization of ‘cellular’

façades, we also account for the fact that in adaptive façades, these properties can change over time. Discretization of the façade is done to gain

insight into performance trends and dynamic spatial effects of adaptive façades in an abstract but generic way. Actual adaptive façade concepts

95

deriving from this approach may or may not have this type of grid-like

appearance.

In this demonstration example, the façade is subdivided into a grid of 4*4 elements. Material properties in terms of solar energy transmittance (Tsol)

and U-value of each of these elements (Table 5.2) are adaptable per hour. The option space for thermal transmittance consists of six steps, ranging from

thermal resistances equivalent of a conventional glazing system up to the passive house insulation standard. The option space for solar transmittance

consists of four steps, ranging from values equivalent of a clear glazing to a dark tinted glazing or roller shade system. Overall, this leads to a design

option space of 6*4=24 possible combinations per element, and 24^(4*4)=1.2e22 possible façade configurations per adaptation period.

Table 5.2. Option space of material properties for the adaptive façade.

Design option 1 2 3 4 5 6

U-value [W/m2K] 0.2 0.5 1 2 3 4

Tsol [-] 0.05 0.25 0.50 0.75

5.2.3 Objective function

Multiple performance criteria are transformed into in a single objective optimization function, following the formulation in Section 4.5.4 (Equation

4.4) to compute the dimensionless total penalty function 𝑃 .

𝑃 = 𝑤 ∙ | + 𝑤 ∙ | + 0.01 ∗ 𝑃 , + 0.01 ∗ 𝑃 , [-] 5.1

In this case study, the weighting factors for energy (w ) and visual (w ) performance are chosen to have an equal value of 0.5. These equal weights

ensure that the optimization formulation can find a balanced trade-off point between these competing objectives.

As described in Section 3.4.2, visual comfort is further subdivided into three

components: UDI, DGP and view to outside. These factors are also embedded in the objective function so that the desired goals can be achieved by

optimization. The threshold values for daylight illuminance, glare and view

96

are indicated in Figure 5.2. For each daylight hour in the optimization horizon,

penalty values are computed, and then aggregated into one value using weighting factors. The penalty functions consist of ramps instead of discrete

step functions, as this aids in detecting slight differences between the façade configurations, and thus expedites the design space exploration.

Figure 5.2. Penalty function for daylight performance with threshold values

To achieve the desired performance, it is important to choose appropriate weighting factors for the combined visual comfort performance function

(Equation 4.1). Several factors need to be taken into account. For example, glare risk occurs during only a relatively limited period of the year. There are

many instances when glare considerations are not relevant for determining façade operation, but when the risk occurs, it is important that the

consequences are effectively blocked. On the other hand, view to outside is only of interest after the basic performance criteria have been met. By giving

view to outside a relatively low weight, the parameter search will be nudged towards solutions that maximize openness of the façade while not

compromising on daylight illuminance and glare discomfort. After performing several test runs, it was found that the weights in Equation 5.2

lead to the desired results for calculating 𝑃 as in input for Equation 5.1. Note that only the relative difference between the weights is important,

because the aggregated result will always be normalized.

𝑃 = 1 ∙ 𝑃 + 10 ∙ 𝑃 + 0.2 ∙ 𝑃 [-] 5.2

It should also be noted that the weightings and setting of threshold values involve subjective and contextual considerations. The present case study

aims to demonstrate capabilities of the performance simulation framework. Future work should focus on identifying the impacts of different weights on

performance. As indicated before by Mahdavi [2008] performance

preferences do not have to be constant over time.

97

5.2.4 Optimization horizon and control update

In the present case study, an optimization horizon of 24 hours was chosen. Ideally, we would want to work with a quasi-infinite horizon, because it

theoretically leads to the best results [Zavala et al. 2009]. A disadvantage of increasing the length of this forward-looking horizon is that it leads to an

exponential increase of the search space per control sequence, and as such, the increased risk of finding only a local optimum with non-exhaustive

optimization algorithms. On the other hand, a too short horizon also has drawbacks. When the horizon is too short, e.g. 1 hour, there are potentially

many cases with for example no energy consumption and no discomfort, in which the objective function is not able to rank performance of the different

alternatives. In such cases, however, this does not mean that each façade adaptation option is equally good. Because of transient energy storage

effects, the current control decision affects the system states that can be reached in the future, possibly leading to performance degradation (e.g.

higher energy demands or comfort violations) that could not be foreseen in

the short horizon. For buildings with very low thermal mass, such as highly-glazed spaces, the consequences might be insignificant [C. S. Lee 2017], but

for typical office buildings, these temporal effects are worth considering. In addition, a longer forward-looking horizon is needed to plan façade

sequences with smooth transitions (Section 4.5.5), because some fluctuating façade schedules could only be prevented with consideration of further

points in the future [Coffey et al. 2010].

The case study uses an adaptation horizon, and also update frequency of 12 hours. In this way, the positive effects of receding, partly overlapping, time

horizons are accomplished. Without receding time horizon, thermal energy storage potential of construction elements would be just exhausted/emptied

at the end of each horizon, and indoor temperature could be at the edge of becoming uncomfortable, leaving unfavorable starting conditions for the

next horizon. The selected adaptation period is one hour and the simulation time step for thermal simulations is 10 minutes.

5.2.5 Reference façade

To establish a baseline for further comparison, an office zone with typical façade properties is defined. The building envelope of this reference situation

has a window-to-wall ratio of 40%. The opaque parts have an Rc-value of 5

98

m2K/W, while for the windows, low-emissivity double glazing (g-value: 0.50,

U-value: 1.78 W/m2K), with external solar shading screens is applied. The shading system is activated when incident solar radiation on the façade is

higher than a fixed threshold value of 300 W/m2.

5.3 Results – performance potential compared

to reference case

5.3.1 Spring week

Performance

Figure 5.3 shows a comparison of the energy performance for different design

options in a week in spring (18 – 22 May). Case 1 to 4 represent the energy performance for a façade with non-adaptive properties at the endpoints of

the values given in Table 5.2. The performance of these design options was selected as a baseline, because these points can give an indication of the

potential in best/worst case scenarios for single objectives, but often fail to perform well on the other aspects. Column 5 represents the office with

reference building envelope as presented in Section 5.2.5, and column 6 shows the optimized result for the adaptive façade.

99

Figure 5.3. Optimization results for a spring week. A: Energy performance, B: glare discomfort, C: daylight illuminance.

The results indicate that thermal insulation has a dominant effect on the

energy consumption in spring. With a highly insulated envelope (e.g. case 2 or case 3), the building needs a considerable amount of mechanical cooling

to maintain comfortable conditions because the heat that is trapped inside cannot be easily released to the environment. In addition, the results show

that a reduction in the contribution for lighting, by making a transparent façade, can have a significant impact on the building’s energy balance. With

adaptive façades, the overall energy consumption can be reduced by 60% compared to the reference situation (Figure 5.3A). This is, however, not the

lowest possible energy demand, which would have been achieved with a fully glazed, low-insulation design variant (Case 1). Looking at the visual

performance indicators DGP (Figure 5.3B) and UDI (Figure 5.3C), it is clear why the adaptive façade solution is chosen as optimal scenario. Even though

UDI levels are favorable, a fully glazed façade leads, as expected, to a high risk of intolerable glare discomfort. In the reference solution, the solar shading

system is closed throughout most of the day, resulting in restricted view to outside and very low levels of UDI in the autonomous category. An optimized

adaptive façade, on the other hand, is able to find a good trade-off point considering all performance aspects.

0

20

40

60

80

100

120

U=4,T=0.75

U=0.2,T=0.75

U=0.2,T=0.05

U=4,T=0.05

Ref CABS

Pri

mar

y en

ergy

co

nsu

mp

tion

[k

Wh

]

lighting

cooling

heating

(A)

0

20

40

60

80

100

U=4,T=0.75

U=0.2,T=0.75

U=0.2,T=0.05

U=4,T=0.05

Ref CABS

DG

P c

lass

[%

]

Intolerable

Perceptible

Imperceptible

(B)

0

20

40

60

80

100

U=4,T=0.75

U=0.2,T=0.75

U=0.2,T=0.05

U=4,T=0.05

Ref CABS

UD

I cl

ass

[%]

exceeded

autonomous

supplementary

fell-short

(C)AdaptiveRefU=4

T=0.05U=0.2T=0.05

U=0.2T=0.55

U=4T=0.75

AdaptiveRefU=4T=0.05

U=0.2T=0.05

U=0.2T=0.55

U=4T=0.75

AdaptiveRefU=4T=0.05

U=0.2T=0.05

U=0.2T=0.55

U=4T=0.75

100

Figure 5.4. Scatter plot showing progression of the design space exploration. Green dots indicate static façade configurations. Grey dots indicate adaptive façade configurations; the darker the color, the higher the iteration number in the GA. The red dot represents

the optimized solution that was selected by the GA.

Figure 5.4 shows a scatter plot with performance considering the two optimization objectives: visual performance (horizontal) and energy

performance (vertical). Each dot represents the performance of a certain façade adaptation strategy for the optimization horizon of 18 May from 0h00

to 24h00. In green, the non-adaptive façade solutions are shown as reference

points. Points closer to the origin (Utopia point) have better performance because they have the lowest penalty values. The greyscale colors give an

indication of the progression towards better performing solutions over time (the darker the color, the higher the generation of the genetic algorithm), and

the red point is the façade solution that is finally chosen by the optimization objective.

Design considerations

To better understand the performance of the optimized adaptive façade, it is important to analyze how the façade properties vary over time. Figure 5.5

shows the optimized façade configurations for the 55 occupied hours during the week.

101

Figure 5.5. Optimized façade configurations for the spring week. Each block represents a façade configuration as seen from the inside. Black corresponds to Tsol = 0.05, dark grey

to Tsol = 0.25, light grey to Tsol = 0.50 and white to Tsol = 0.75

The following principles can be observed:

⎯ In the mornings, the façade is highly transparent to enable maximum

daylight utilization and a view to outside.

⎯ During periods of high incident radiation (midday), the optimization

algorithm suggests a façade configuration with low solar transmittance. On the overcast day (Thursday) the façade remains

mostly transparent.

⎯ Later in the afternoon, passive solar energy gains are unwanted

because the building is already warmed up. Configurations with intermediate transparency are found as a balanced trade-off solution

considering all performance aspects involved. The non-uniform façade patterns are the result of (unwanted) randomness that is

sometimes present in the optimization outcomes. In these cases, the soft constraints were not effective in ensuring uniform façade

appearance.

⎯ Sun-tracking behavior of opaque façade elements can be observed to some degree. In the mornings, a higher density of dark cells is

situated on the left (East) side, whereas in the afternoon, the majority of dark cells shifts to the right (West) side. This effects is most

noticeable on the Wednesday, and indicates how the optimization framework is able to identify façade regions with high luminance

contrast to promote the effect of glare-free daylight utilization. The

fact that this effect is not evident on all sunny days highlights that the

16 17 189 10 11 12 13 14 15

MonTueWedThuFri

8

102

existence of multiple local optima and the existence of the large

design option space in combination the flatness of the fitness landscape (i.e. the fact that there are many solutions with nearly

identical performance) make it difficult to reach clear patterns in the final solution.

⎯ On days with similar weather conditions (e.g. Monday and Tuesday)

and IEQ requirements, the algorithm finds a different optimal trade-off solution during many hours of the day. This observation provides

another illustration of the complexity of the optimization problem.

The façade-averaged solar transmittance and U-value for each hour of the

simulation period are represented in Figure 5.6. By analyzing this data, information can be extracted to develop a better understanding of the

adaptive façade properties that will lead to high performance. Three notable findings can be observed:

⎯ The ability to chance thermal insulation properties is effectively

utilized. On evenings and nights with relatively cool outside temperatures, this feature is used for passive cooling of the interior

space.

⎯ Whereas solar transmittance values frequently use all values of the

design option space, U-value relies almost exclusively on the end points. The use of developing façade materials or technologies with

many intermediate thermal insulation states seems limited.

⎯ There are many instances where the optimization algorithm suggests

the combination of high thermal resistance with high solar transmittance. It seems worthwhile to focus R&D efforts on

developing transparent insulation systems that also have the ability to change visible light transmittance.

103

Figure 5.6. Optimization results: surface-averaged dynamic façade properties for the spring week.

5.3.2 Winter week

Performance

Figure 5.7 shows a comparison of the energy performance over the whole winter week (12 – 16 Jan). Compared to spring, the contribution of useful solar

gains is small. More artificial light is needed, and the differences between extremes in transparency are relatively small. Depending on the insulation

level, either heating or cooling energy is needed. In the case with adaptive façade, heating energy can completely be eliminated, while only a small

fraction of the cooling energy is still needed. Compared to the reference case, an overall energy reduction of 43% can be achieved, which can mostly be

attributed to smart use of passive solar heating. Due to the low angle of the sun, and the possibility for low visible light transmittance with the adaptive

façade, there is no risk for glare discomfort during this week, provided that façade transparency is properly controlled on sunny days.

00.511.522.533.544.5

T sol[-] and U-v

alue [W/m2 K]

Mon Tue Wed Thu FriTsol U-value

104

Figure 5.7. Optimization results for a winter week.

Design considerations

Compared to the week in spring, the average transparency of the façade is

much higher in winter (Figure 5.8). There is a high correlation between façade transparency and incident solar radiation during this period. At times when

irradiance levels are low, and there is little or no risk for glare discomfort, the design exploration algorithm suggests highly transparent façade states that

maximize the view to outside. Sun tracking behavior is less visible during this winter week. This is likely due to the position of the sun, which is lower in the

sky during this period. As a consequence, the façade configurations that are found by the genetic algorithm resemble the characteristics of a conventional

solar shading system to a large extent.

Figure 5.8. Optimized façade configurations for the winter week. Each block represents a façade configurations as seen from the inside. Black corresponds to Tsol = 0.05, dark grey

to Tsol = 0.25, light grey to Tsol = 0.05 and white to Tsol = 0.75

0

20

40

60

80

100

U=4,T=0.75

U=0.2,T=0.75

U=0.2,T=0.05

U=4,T=0.05

Ref CABS

Pri

mar

y en

ergy

con

sum

pti

on

[kW

h]

lighting

cooling

heating

AdaptiveRefU=4T=0.05

U=0.2T=0.05

U=0.2T=0.55

U=4T=0.75

15 16 17 189 10 11 12 13 14

MonTueWedThuFri

8

105

The combination of insulation and transmission properties of the optimized

façade states is presented in Figure 5.9. During most of the time, a transparent insulation state is selected as the optimal solution, but during

sunny moments, the solar transmittance goes down to reduce glare risk. A low thermal resistance value is identified as beneficial on two of the

afternoons. In this way, heat trapping can be prevented – and cooling load avoided – by releasing heat to the cooler ambient environment.

Figure 5.9. Optimization results: surface-averaged dynamic façade properties for the winter week.

00.511.522.533.544.5

T sol[-] and U-

value [W/m2 K]

Mon Tue Wed Thu FriTsol U-value

106

5.4 Discussion and conclusion

This Chapter has demonstrated the capabilities of the toolchain, by presenting a case study that explores the performance potential of adaptive

façades for an office building in the Netherlands. Results that have been analyzed in detail for one week in spring and one week in winter indicate that

the simulation-based optimization methodology is able to highlight the capabilities of adaptive façades in terms of finding balanced trade-off

solutions considering multiple competing objectives. In this way, significant performance benefits compared to state-of-the-art façade alternatives can

be obtained.

A rather large option space of dynamic façade properties was used to be able to provide recommendation for further R&D and product development of

adaptive façade technologies and materials. It was found that dynamic thermal insulation and sun-tracking solar shading solutions are mostly

responsible for the improved building performance with adaptive façades. More R&D to develop new design solutions in these directions is therefore

recommended.

107

6

Investigating trade-offs between

performance and façade system

complexity

6.1 Introduction

The goal of this chapter is to demonstrate application-specific aspects of the toolchain on the basis of a case study. The intention is to extend the usability

of the simulation methods, and to illustrate how it can help in developing a better understanding of the tradeoffs between façade performance and

complexity.

6.1.1 Background and objectives of the case study

Careful selection of the properties and dimensions of windows is essential for

optimizing the indoor environmental quality and energy performance of

108

buildings. Nowadays, special layers, such as low-emissivity coatings and solar

control films, are routinely applied for enhanced control of solar heat gains and higher thermal resistance. Although these technologies have

substantially improved the performance of fenestration systems, they also have some drawbacks. One of the limitations of such glazing systems is that

the properties are fixed over time, and can therefore not respond to, or take advantage of, the variability in climatic conditions. A solar control film, for

example, will typically help reduce cooling energy in summer, but also impairs view to outside and blocks useful solar gains in periods when no

cooling is needed. As a result, compromises in the design phase or often

needed to achieve satisfactory performance under all conditions.

Breakthroughs in materials science now open up a range of new possibilities that will allow building designers to influence window-related trade-offs in a

more dynamic way. Recently, materials with switchable reflection/transmission properties in three parts of the electromagnetic

spectrum have been reported in literature. The ability to selectively control the transmission of (i) visible light, (ii) near-infrared sunlight, (iii) and

longwave infrared radiation (emissivity) generates many promising opportunities because through this adaptability, multiple optimized states

can be reached. There is, however, still a lack of guidance on how such innovative technologies should be integrated in buildings.

The objective of this study is to evaluate and compare the performance of such windows on a whole-building level under a range of different conditions.

This scoping study is performed by means of a co-simulation approach, as described in Chapter 4 that uses Radiance, ESP-r and BCVTB to

simultaneously predict energy performance, dynamic daylight illuminance, and glare discomfort. In this study, independent switching of glazing

properties is addressed in three parts of the spectrum, as well as the various interactions, and it is investigated which strategy works best depending on

certain circumstances. The parameters that are taken into account include different façade orientations, climatic conditions and window control

strategies. Finally, the findings of this study are synthesized in the form of general recommendations that can be used by both building designers and

the research community as an outline for future materials science development directions.

109

6.2 Description of the case study

The methodology for the demonstration study that is presented in this chapter is described in five steps: (i) baseline model, (ii) window properties,

(iii) performance indicators, (iv) simulation scenarios, and (v) window control strategies.

6.2.1 Baseline model

The baseline simulation model is a single-zone office space (3.6*5.4*2.7 m)

with a window-to-wall ratio of 60%. The perimeter façade is modelled as an external wall while all other surfaces face similar office spaces. All opaque

construction elements in this reference model have medium thermal mass. The external wall has a thermal resistance (R-value) equal to 3.5 m2K/W

The perimeter office zone is occupied by two people who are present on weekdays from 8-18 h. Each occupant represents an internal heat gain of 120

W. Internal gains from electronic equipment (laptop, printer, etc.) corresponds to 50 W/person, operating only during occupied hours.

Artificial lights are controlled by a sensor inside the zone, 1.5 m from the façade at work plane level. The lights with a nominal lighting power density

of 12 W/m2 are continuously dimmed to meet the setpoint of 500 lux.

The heating, ventilation and air conditioning (HVAC) system is modeled as an ideal convective system, because the main interest of this work is on daylight

control and management of solar gains. The HVAC system has a heating setpoint 21ºC starting one hour before the beginning of the occupied hours

until the zone is empty again (7-18 h.) and 14ºC during the rest of unoccupied hours. The cooling setpoint is set to 25ºC for occupied and 30ºC for

unoccupied hours. Ventilation with heat recovery (90% efficiency) is set to 3 air changes per hour (ACH) during occupied hours and 0.2 ACH during

unoccupied hours, while infiltration was assumed to have a constant value of 0.3 ACH throughout the year.

6.2.2 Window properties

Windows with three different types of switchable glazing properties are analyzed in this study. All switchable windows share one common reference

110

state (sref), having the properties of a regular low-emissivity double glazed

window.

Figure 6.1. Relationship between visible transmittance and solar transmittance. Note that window properties below the maximum light-to-solar heat gain (LSG) line and

above the minimum LSG line are physically impossible.

Figure 6.1 shows how the properties of the windows with low visible

transmittance (s1) and high NIR reflectance (s2) relate to this reference situation (sref, low-e). The switchable window state with high NIR reflection

(s2) has a very high light-to-solar heat gain (LSG) ratio. In this case, it means

that all solar energy in the wavelength range λ > 700 nm is reflected, without

affecting the transmittance of visible light. For s1 the case with low visible transmittance, it was assumed that the absorption of sunlight is equally

spread over the entire solar spectrum. This case therefore results in the situation that switching of window properties happens between two points

with equal LSG value. The end point of VLT = 0.05 is selected because it is generally considered that values as low as this lead to a moderate risk for

glare discomfort [E. S. Lee et al. 2013; Piccolo and Simone 2009].

The third type of adaptability (s3), the changing of properties in the longwave radiation range is achieved by varying the emissivity of the inside surface of

the outside glazing pane from 0.05 to 0.84. The algorithms in LBNL Window 7 [LBNL 2017b] indicate that doing this would lead to an increase in window

U-value from 1.6 W/m2K to 2.7 W/m2K.

A summary of all relevant properties of the different switchable glazing

systems is presented in Table 6.1.

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1

Tso

l

Tvis

NIR

sref

Max. LSG

Min. LSG

s2s1

111

Table 6.1. Overview of the switchable glazing properties in the [i] visible (VLT), [ii] near-infrared (NIR) and [iii] longwave part of the solar spectrum.

Tvis[-]

Tsol[-]

𝜺[-]

U-value[W/m2K]

VLT s1 0.05 0.22 0.05 1.6

sref 0.74 0.54 0.05 1.6

NIR s2 0.74 0.35 0.05 1.6sref 0.74 0.54 0.05 1.6

Longwave s3 0.74 0.54 0.84 2.7

sref 0.74 0.54 0.05 1.6

6.2.3 Performance indicators

Two aspects are considered in the switchable window performance assessment: energy and comfort.

To evaluate energy performance, the total primary energy use intensity (EUI) in kWh/m2 for heating, cooling and lighting is computed. Throughout this

Chapter, we assume that the seasonal heating efficiency = 0.9, cooling COP = 3, and the primary energy conversion factor for electricity = 2.56, according

to (NEN7120 2011).

Façade-related visual comfort is influenced by a number of factors with complex interactions [Reinhart and Wienold 2011]. The target is to use the

switchable windows to contribute to well-distributed daylight illuminance in the absence of discomfort glare. This is evaluated with two complementary

indicators:

⎯ Useful daylight illuminance (UDI). Daylight utilization is assessed by

computing the percentage of occupied hours that daylight

illuminance on the work plane falls within certain bounds. Four different categories according to the classification of [Mardaljevic et

al. 2012] are considered: fell short (<100 lux), supplementary (100-300 lux), autonomous (300-3000 lux) and exceeded (>3000 lux).

⎯ Daylight glare probability (DGP). Discomfort glare is assessed using

DGP, as it was shown to be the most robust glare metric in a comparative study under a wide range of ambient conditions

[Jakubiec and Reinhart 2011]. In this study, a conservative glare scenario is considered in which the occupants’ view direction is

112

always facing the façade. Three different categories for discomfort

glare are distinguished, according to the classification in [Reinhart and Wienold 2011], where the risk is (i) intolerable (DGP > 0.45), (ii)

disturbing (0.35 < DGP < 0.45), or (iii) not disturbing (DGP < 0.35).

6.2.4 Simulation scenarios

Various integrated daylight and thermal simulations are conducted to assess the performance of the switchable window types. Here, one baseline

simulation scenario is defined that considers variations in terms of climate, façade orientation and window-to-wall ratio. Table 6.2 gives an overview of

the different scenarios that are investigated in this study; items in bold represent the baseline scenario.

Table 6.2. Overview of variants in the parametric study (bold represents baseline scenario).

Climate Stockholm, Amsterdam, Berlin, Madrid

Orientation East, South, West

WWR 30%, 60%, 90%

The four locations were selected because together they cover a wide range

of climate types, representative for the European continent. Table 6.3 presents an overview of heating and cooling degree days for each of the four

cities as an indication of the range of climate conditions that is covered in this study.

Table 6.3. Overview of heating degree days (HDD) and cooling degree days (CDD) for the four climates.

HDD CDD

Stockholm 4286 49Amsterdam 3038 65

Berlin 3317 147Madrid 2023 612

113

6.2.5 Window control strategies

The definition of an appropriate window control strategy is of essential importance for analyzing the performance of switchable glazing technologies

[Jonsson and Roos 2010]. This demonstration example uses rule-based control strategies, to be able to study the performance of the different

window systems in a realistic application environment. Table 6.4 summarizes the control strategies that are considered in this study.

Table 6.4. Window control strategies

Switch from sref other state, if

VLT Incident solar radiation > 250 W/m2

NIR Incident solar radiation > 250 W/m2

Longwave Ambient temperature < indoor temperature

AND building is not in heating mode

6.3 Simulation strategy and settings

This study uses a coupled simulation approach to analyze the interactions

between daylight performance and energy efficiency. Annual dynamic

daylight simulations for the different window states are carried out using the Radiance three-phase method, while Building energy simulations (BES) are

performed using ESP-r. The control strategies are directly implemented in the BCVTB using actors from its Ptolemy II library. This coupling strategy also

takes care of assigning the correct internal heat gains for lighting, depending on daylight illuminance corresponding to the values that were computed in

the pre-processing phase.

6.4 Results

6.4.1 Daylight performance

Figure 6.2 compares the simulation results in terms of UDI for the reference

window (sref) and the case with switchable VLT (s1). The difference is most

114

notable for the number of occupied hours that the illuminance threshold of

3000 lux gets exceeded.

Figure 6.2. Useful daylight illuminance results for the reference window (top, sref) and switchable VLT window (bottom, s1). Results are shown for two façade orientations and

three window-to-wall ratios.

The extent of possible discomfort due to the too high daylight illuminance

levels gets significantly reduced in the case of controlled light transmittance,

most notably for the South façade.

During the hours with relatively low daylight illuminance, the difference between the two cases is relatively low. This effect can be explained by the

fact that during most of these hours, the switchable glazing system is switched in the transparent mode, and hence leads to similar lighting

conditions as the regular façade.

The results for glare discomfort are presented in Table 6.5, showing again the comparison between switchable and non-switchable visible light

transmittance. The difference in results for DGP is even more pronounced than the results for UDI. The amount of discomfort hours shows a substantial

reduction when the window transparency is reduced to 5% during times with high incident solar radiation.

0

20

40

60

80

100

South West South West South West

0

20

40

60

80

100

South West South West South West

% o

f occ

up

ied

tim

e%

of o

ccu

pie

d ti

me

WWR = 30 WWR = 60 WWR = 90

WWR = 30 WWR = 60 WWR = 90

Fell short Supplementary Autonomous Exceeded

115

Table 6.5. Daylight glare probability classification for the regular window (top) and switchable VLT window (bottom). Results are shown for tow façade orientations and

three window-to wall ratios.

Orientation WWR DGP Disturbing [h] Intolerable [h]Reference case (sref)

South 30 98 11260 114 198

90 182 325

West 30 179 25660 172 378

90 212 421

Switchable VLT (s1)

South

30 31 4

60 52 990 56 23

West

30 22 12

60 44 2190 51 56

6.4.2 Energy performance

Figure 6.3 shows a comparison of the primary energy demand for different

WWR values (South façade, Amsterdam). The results show that for windows with switchable non-visible properties (NIR and longwave), the results are

highly sensitive to the window area. Especially larger windows lead to a large increase in energy consumption. On the other hand, switching in the visible

part of the spectrum leads to much less difference in the results. This finding indicates the possibility of designing buildings with large glazed façades,

without having a significant penalty in terms of increased energy consumption.

116

Figure 6.3. Energy performance of the three switchable window types for a façade with South orientation in Amsterdam. Results are shown for window-to-wall ratios of 30, 60

and 90%.

In Figure 6.4, the results for the four different climates are presented. These results show that lighting energy consumption has a quite large influence on

the overall energy consumption. The windows that switch in the non-visible part of the spectrum need considerably less energy for lighting. However,

these cases also show a higher energy demand for cooling. The lowest cooling energy demand is observed in the situation with visible switching. Compared

to longwave switching, the dynamic NIR reflectivity also reduces cooling energy demand, but not as effective as the case with VLT control. These

results correspond with the values for solar transmittance, as presented in Table 6.1. However, these static material properties are not sufficient to

characterize the performance of switchable glazing systems, because (i) they do not account for the dynamic effects that occur at different control

strategies, (ii) do not take the total building energy balance in consideration, and (iii) do not distinguish whether most of the incident radiation is absorbed

(VLT case) or whether it is reflected (NIR case). The complex balance between heating, cooling and lighting energy use in Figure 6.4 clearly indicates the

need to develop and select switchable glazing technologies in response to the needs of local climatic conditions. For example, a switchable VLT window

offers the lowest total energy demand for Madrid, but such a window would

not be recommended for energy-saving purposes in Stockholm.

0

50

100

150

200

250

30 60 90 30 60 90 30 60 90Pri

mar

y en

ergy

con

sum

pti

on [k

Wh

/m2]

VLT NIR Longwave

Heating Cooling Lighting

117

Figure 6.4. Energy performance of the three switchable window types for a façade with South orientation and 60% window-to-wall ratio. Results are shown for the climates of

Stockholm, Amsterdam, Berlin and Madrid.

Finally, Figure 6.5 shows the energy performance results as a function of façade orientation. Although these results are useful to obtain a better global

understanding of the performance of switchable glazing types, they do not lead to remarkable new insights. For all three window types, the lowest

energy consumption is achieved for the South façade. In all three cases, the East façade has the highest heating energy consumption. In the situation

where the admittance of solar gains is reduced (VLT and NIR), the heating load on the East façade is extra high because the contribution of passive solar

gains to room heating gets reduced when the room is heating up in the morning hours. The consequences of this orientation effect are also visible

for the West façade. By the time that high intensity solar radiation reaches the West façade, the office room is already warmed up. Because the longwave

case has no ability to control the amount of solar gains, we can see there a much larger increase in cooling energy demand, compared to the other two

cases (VLT and NIR).

0

50

100

150

200

250

300

STO AMS BER MAD STO AMS BER MAD STO AMS BER MAD

Pri

mar

y en

ergy

con

sum

pti

on [k

Wh

/m2]

VLT NIR Longwave

Heating Cooling Lighting

118

Figure 6.5. Energy performance of the three switchable window types for a façade with 60% window-to-wall ratio in Amsterdam. Results are shown for three orientations,

East, South and West

6.5 Discussion and conclusion

This chapter has demonstrated the use of a high-resolution modeling approach to evaluate how three advanced switchable window types, active in

three different parts of the electromagnetic spectrum, are able to meet the goal of well-controlled daylight and solar gains in buildings.

The results are in line with the outcomes of various simulation studies that

present an in-depth analysis of the individual switchable window types, and with a strong focus on either the thermal or daylight aspect [DeForest et al.

2015; Fernandes et al. 2013; Loonen, Singaravel, et al. 2014; Tavares et al. 2014]. This is an important finding, because it gives additional information to

confirm that the co-simulation implementation that is presented in this thesis is able to produce meaningful results. Another unique characteristic of

the results presented here, is that it allows for a rapid side-by-side comparison of the relative performance of several technologies, because

their evaluation was subject to the same simulation scenarios and boundary conditions.

Pri

mar

y en

ergy

con

sum

pti

on [k

Wh

/m2]

VLT NIR Longwave

Heating Cooling Lighting

0

50

100

150

200

250

East South West East South West East South West

119

This application study has led to the following main conclusions:

⎯ The concept of one-size-fits-all does not apply to switchable window

technology. To achieve good performance, product developers and building designers should adapt the properties of switchable glazing

technologies to the needs of individual situations.

⎯ Dynamic modulation of sunlight transmittance in the visible

wavelength range not only influences solar heat gains, but also has a positive effect on visual comfort conditions and lighting energy use.

Among the three tested glazing types, glazing systems that are active in this part of the spectrum have the highest energy-saving potential.

It is essential to use coupled daylighting and thermal simulations to gain insights into such effects.

⎯ Solar shading systems, such as blinds and screens or window coatings

with controllable VLT transmission are needed to ensure acceptable

visual comfort conditions, especially in cases with large window areas. It is not always fair to compare the energy performance of

windows with and without shading system, because they do not lead to equivalent indoor conditions [Ochoa et al. 2012; Goia 2016].

⎯ The performance of switchable windows is very climate-dependent.

Windows with switchable NIR reflection are most effective in climates which have both a considerable heating and cooling energy

demand, whereas the ability to switch VLT is most

effective/promising in sunny climates.

⎯ Compared to the other two technologies, the performance potential of adaptive low-emissivity coatings is relatively low. Its effect appears

to be almost insensitive to the façade orientation.

This study also led to a couple of limitations that are worth to be mentioned.

For example, the potential co-benefits of having switchable properties in multiple parts of the solar spectrum was not investigated. It is expected,

however, that independent control of transmitted visible and non-visible solar energy can lead to extra advantages for low-energy building operation

with high indoor environmental quality. This study only assessed two states per switchable window type, at the end points of the dynamic range. It would

be interesting to analyze if adding the option for switching to intermediate states would lead to different conclusions. Although it is known that the

120

chosen window control strategy can have a significant influence on the

performance of switchable glazing, only a limited number of options were examined here.

121

7

Concluding remarks

7.1 Conclusions

Driven by the need to reduce the negative environmental impact of building energy use, and motivated by the business opportunities that arise from

providing spaces with superior indoor environmental quality, there is currently a rapid transition towards more sustainability in the built

environment. In light of these developments, adaptive façades have been identified as a promising design solution, because their application can

enable high energy performance without compromising indoor comfort conditions. To further support the development and adoption of high-

performance adaptive façade systems, there is a need for tools and methodologies that can be used to facilitate informed decision-making in the

R&D process of innovative adaptive façade materials and technologies. Through literature review and initial simulation studies, it was identified in

Chapter 2 and Chapter 3 that currently available building performance

122

simulation software and workflows have a number of important

shortcomings that limit their usability in the context of buildings with adaptable façade properties. The main aim of the research presented in this

thesis was therefore to develop, test and evaluate computational approaches that can be used to support design analysis and performance-based design

space exploration in the development process of innovative adaptable building envelope components and concepts.

This dissertation has addressed four objectives that are directly related to

this research aim:

⎯ To develop an effective modeling and simulation strategy for

integrated performance prediction of buildings with time-varying building shell properties.

⎯ To develop and test a computational approach, based on simulation

and optimization techniques, that is capable of exploring the performance potential/limits of adaptive façades.

⎯ To illustrate, on a case study basis, how this approach can be used to identify high-potential, i.e. high-performance, low-complexity,

directions for future adaptive façade concepts.

⎯ To better understand the causal relationships between adaptability

of building shell parameters and performance (energy and comfort) in a number of demonstration examples.

As a first step in this process, the requirements that are characteristic for

performance prediction and optimization of buildings with adaptive façades were identified. It was found that special attention needs to be paid to

adequately incorporating time-varying façade properties and dynamic thermal energy storage effects into the simulation models. In addition, when

predicting the performance of high-performance façades, there is a need for computational approaches that can simultaneously take into account the

interactions and trade-offs between thermal and daylight domains. A third important requirement in the present context is the necessity to consider

advanced control strategies because of the tight coupling between design and operational aspects, and the fact that different levels of control

complexity would lead to different optimal solutions.

In response to these requirements, a simulation toolchain was developed

using the building energy simulation program ESP-r and the three-phase

123

simulation approach in Radiance for dynamic daylight predictions. A co-

simulation implementation using the building controls virtual test bed (BCVTB) was set up to enable on-the-fly coupling between the models in

these two software environments. The possibility to incorporate time-varying façade construction properties was achieved by extending the

thermophysical property substitution feature that was implemented by MacQueen [1997]. In this way, a flexible yet accurate method of simulating

dynamic thermophysical façade properties was established.

An important element of this study is the idea to formulate the design exploration process of adaptable building envelopes as an optimization

problem. The fact that adaptive façades change over time makes this study quite different from conventional design optimization studies, because it is

not just one façade configuration, but a whole sequence of time varying system states (i.e. façade properties) that needs to be considered in the

optimization. In this thesis, this challenging task was addressed through the development of an offline receding-horizon model-based control strategy.

This approach contains two instances of the BPS models; one model is a

virtual representation of the real building, and the other model is used inside the simulated control system to evaluate the performance of various façade

adaptation strategies. This whole process was coordinated using Matlab. A genetic algorithm was used to efficiently navigate through the large search

space of façade adaptation strategies, while an approach for explicit state initialization was developed to avoid the significant computational overhead

that would occur in the repeated warm-up phases of the simulations.

Two case studies were presented to illustrate the usability of the tool chain

in different application areas.

The first case study was set up to quantify the performance potential of future-oriented adaptive façade systems that can simultaneously change

both thermophysical and optical material properties. The façade was subdivided into a grid of elements with individually controllable façade

properties to provide insights into performance trends and dynamic spatial effects of adaptive façades in an abstract but generic way. Through an

analysis of optimized solutions, it was found that the toolchain leads to meaningful results that match well with principles known from building

physics and other results published in literature. Based on this positive confirmation, the results were then further analyzed to suggest R&D

pathways in materials science that could be pursued in the coming years to

124

bring the identified high-performance adaptive façade systems within reach

of façade designers.

The second case study focused on spectrally selective switchable glazing systems with adaptive properties in different parts of the electromagnetic

spectrum. Multiple window-to-wall ratios, façade orientations and climatic conditions were investigated to evaluate the performance of the fenestration

system in different scenarios. The results from this demonstration example showed how different glazing types can be developed to meet the specific

requirements of different building types and boundary conditions. As such, it could be illustrated that it is not always necessary to have the widest adaption

range possible, but that tailored solutions can lead to similar performance at much lower complexity. It was found that the toolchain that is developed in

this thesis is a useful instrument for providing decision support in settings where such complex problems are being addressed.

7.2 Limitations

During the development of the toolchain, many decisions about the structure

of the models and inter-process communication between the models had to be made. Although these decisions were always based on careful analysis of

the advantages and drawbacks of state-of-the-art options, it was not always possible to verify that they would lead to the most desirable result. In other

words, even though the case studies demonstrated several advantageous characteristics of the simulation approach, it is not possible to claim with

confidence that the approach to simulation and optimization taken here is the most effective or efficient.

For example, the toolchain contains a considerable number of simulation parameters (e.g. simulation time step) and optimization settings (e.g. length

of optimization horizon) that can be tuned by the user. It has been observed that setting these user-defined variables appropriately is very important,

because implementing incorrect settings can lead to sub-optimal results, thereby jeopardizing the effectiveness of the toolchain. Despite this

importance, due to limited resources, it was not possible to carry out a systematic study that would allow advice to be provided to the end user on

how to select robust simulation settings. It is also worth noting that many of these settings are very case-specific (e.g. the length of the optimization

horizon should be chosen in accordance with the thermal time constant of

125

the building), which makes it difficult to arrive at generally applicable

guidelines.

The optimization method that is currently implemented in the toolchain is based on genetic algorithms (GA). GAs are the preferred method of choice for

providing optimal solutions in many multi-criteria building performance problems, and, also in this study, it was found that they can work well. This

kind of meta-heuristic algorithm, however, also introduces an important limitation in the context of the present work. When using GAs, it is not

possible to guarantee that the best design solutions, as identified by the algorithm, are actually global optimum solutions. This is not a serious issue

when one is interested in quantifying the performance potential (e.g. energy-saving or comfort improvement) of a certain adaptive façade technology.

However, in many cases, it is also interesting to gain an in-depth understanding of which design parameters and façade adaptation strategies

are responsible for the prospective high performance. In such circumstances, it is important that a certain level of consistency can be identified in the

optimized results. When this consistency is missing, it is very difficult to distinguish random artefacts of the optimization algorithm (e.g. due to

finding local instead of global optima) from meaningful differences that are actually driving the parameter search. Because of this difficulty, it can

become highly problematic to discern patterns, and to identify optimal designs for adaptive façades. Several cases with inconsistencies in optimal

façade properties were found in this study. These inconsistencies can be partly attributed to the intrinsic complexity of the optimization problem and

to the fact that the fitness landscape is rather flat, but they are also partly due to the local optima that are found by the GA. Various measures such as soft

constraints and careful selection of the bounds of the design option space were implemented to eliminate the negative consequences of this effect as

much as possible. Nevertheless, it was not possible within the limitations of this thesis to test whether different optimization algorithms of the same type

(e.g. particle swarm optimization), or algorithms of a different category (e.g. deterministic algorithms or gradient-based algorithms) could possibly lead to

more consistency in the results. Consequently, there is a need for comparison

studies that test the relative effectiveness of optimization algorithms in the context of adaptive façades.

It has been demonstrated in this thesis that formulating the design space

exploration of adaptive façade systems in the form of an optimization problem can be a very powerful approach. Nevertheless, there are also points

of attention, arising from the fact that solving the optimization problem at

126

every n time steps of the simulation drastically increases the computational time that is needed to arrive at optimal or near-optimal solutions. Possible

ways to reduce computation time are through parallel computing, or by improving the search efficiency of optimization algorithms. Here it is argued

that there are also application-specific elements in which knowledge of the optimization problem at hand can be exploited to make the search process

more efficient. For instance, adaptation periods could have a variable length, to allow for frequent transitions of façade properties at times when there are

more dynamics in the environment (e.g. when direct sunlight is incident on the façade), but with a lower degree of adaptability under stable boundary

conditions (e.g. during nighttime). In addition, the idea of using pattern recognition could be explored. Using this approach, the façade control

intelligence could detect certain trends or characteristic periods in IEQ requirements and environmental boundary conditions. Based on this

information, valuable computation time can be saved by using an archive or look-up table function to retrieve high-performance façade sequences that

have proven to work well for similar conditions in the past. Comparable approaches have already been demonstrated for control of HVAC systems

[Coffey 2013], and it is expected that this knowledge can also be transferred to the domain of adaptive façades. In the current implementation, a high-

resolution ESP-r model is used to assess the performance of different proposed façade sequences inside the virtual controller. It would be

worthwhile to investigate if simplified or data-driven models (e.g. grey-box models) could be used instead of the high-resolution first-principles based

model to reduce computation time without significant loss of prediction quality.

In Chapter 1, it was identified that high-potential adaptive façade systems are the ones that combine high performance with relatively low complexity in

terms of e.g. the number of adaptive vs. fixed design variables, their modulation range (i.e. high and low value of adaptive design variables) and

corresponding adaptation frequency. Using the case of switchable glazing as an example, Chapter 6 subsequently demonstrated how the developed

toolchain can help designers and product developers to quantify the prevailing trade-offs between performance and complexity of façade

systems, thereby enabling them to make informed decisions that can lead to desired outcomes. By means of this demonstration, only a fraction of the

possibilities of how the toolchain could be used for taming complexity were addressed. Many other possibilities were left unexplored, but the toolchain

127

has been developed in such a way that it can be used to help answering the

following questions:

⎯ What is the relationship between spatial resolution (e.g. the number of adaptable façade elements) and façade performance?

⎯ Do different façade orientations (e.g. north, no direct sun vs. sun-exposed façades) and preferences for performance criteria (e.g.

different scenarios for glare perception, multiple occupants) lead to different optimal façade solutions?

⎯ Is there a critical minimum adaptation frequency to achieve certain

performance targets with different design variables? Does this

frequency differ for buildings with low and high thermal mass?

⎯ What is the loss in performance when the façade is changing its configuration on a less frequent basis?

⎯ Should this frequency be constant throughout the year, or can we identify periods where less frequent transitions would lead to the

same performance?

7.3 Future research

While developing, implementing and testing the functionality of the toolchain for performance optimization of adaptive façades, this study has opened up

various promising directions for future research. On one hand, these suggestions for further work focus on broadening the scope and application

domain of the toolchain. On the other hand, they refer to the integration of knowledge or methodologies from adjacent research disciplines to further

enhance the potential added value of the approach.

The performance of adaptive façades has so far been analyzed for a relatively

small number of performance indicators, mostly related to energy efficiency and thermal and visual comfort. It is expected that the application of adaptive

façades in buildings can lead to a number of other benefits, such as cost-reductions due to smaller installed capacity of HVAC systems, increased

robustness/resilience with respect to unexpected events (e.g. heatwaves or persistent changes in building usage), or synergistic interactions with other

building services. The toolchain that has been developed in this thesis has all

128

the capabilities to become a useful asset in future studies that quantify such

effects for various stakeholders. In doing so, it would be worthwhile to make the toolchain part of a larger virtual test environment that also contains the

necessary elements for extensive uncertainty propagation (e.g. weather variability and occupant behavior) and sensitivity analysis studies. It might be

wise to further develop dedicated methodologies such as dynamic sensitivity analysis [Loonen and Hensen 2013], or to adjust existing approaches, to take

full advantage of the intrinsic value of these advanced computational support tools in the context of adaptive façades.

The application potential of the work presented in this thesis could be

broadened by adapting it in such a way that additional façade technologies can also be included. It would, for example, be interesting to extend its use

towards onsite harvesting of renewable energy (e.g. façade-integrated PV or solar thermal), or to include controlled use of façade openings to optimize

the design of natural ventilation and night-time ventilative cooling. There is also potential for incorporating interior slabs with adaptable thermal storage

[Hoes and Hensen 2016].

By embedding the simulation programs in a receding time horizon control

framework, this research has established a thorough coupling between the design and control aspects that determine the performance of adaptive

façades. Advanced, optimization-based control sequences were considered to evaluate how the operation of the façade plays a defining role in identifying

optimal adaptive façade properties. All the analyses have targeted the product development and/or façade design phase, prior to the building’s

operation. The actual control of adaptive façade systems when the building is in use was beyond the scope of this study. Nevertheless, it seems possible

to use the tools and methods of this study to support the operational phase of buildings with adaptive façades. For example, the receding horizon

approach could be integrated into a real-time simulation-assisted control framework, similar to the work described by e.g. Clarke et al. [2002] and

Xiong and Tzempelikos [2016]. As indicated by e.g. Favoino et al. [2017], the ability to use a digital twin [Boschert and Rosen 2016] of an adaptive façade

system can be very beneficial for supporting complex decision-making during operation. Another option would be to use the simulation framework

to extract control rules that can be implemented in a conventional building management system [May-Ostendorp et al. 2013].

In a wider sense, it would be profitable to apply the toolchain developed in

this work to further investigate and improve the mutual interaction between

129

adaptive façades and building occupants. The developed toolchain can be

used to investigate the impact of user influence (e.g. control override actions) on adaptive façades and in turn support the development of interfaces that

enable a better human-façade interaction experience. Moreover, instead of using the toolchain to quantify how adaptive façades can minimize the

occurrence of discomfort compared to other façade solutions, there is also scope for increased use of simulation-based research to further explore how

adaptive façades can lead to increased occupant satisfaction, wellbeing and productivity [Loonen et al. 2015].

Researchers in the field of flexible systems engineering have developed a

range of design support tools that can be used to appraise the sustained value of reconfigurable or adaptive systems over their entire lifecycle [Cardin 2013].

Furthermore, these tools can be used to systematically evaluate the relative influence of adaptive vs. fixed design variables, and to prioritize which

variables should be changeable in operation. Notable examples include epoch-era analysis with Pareto trace [Fitzgerald and Ross 2012], selection-

integrated optimization [Khire and Messac 2008] and time-expanded decision networks [Silver and de Weck 2007]. Thus far, these methodologies

have mostly been applied for e.g. automotive and aerospace applications, and to improve the design of manufacturing systems. Because of the many

similarities with adaptive façades, it would be interesting to transfer such methodologies to the domain of building engineering. The neighboring field

of multi-criteria decision-making could also provide interesting connection points to further enrich design support for adaptive façades. It would, for

example, be interesting to process the outcomes of the toolchain using post-optimization analysis [Brownlee and Wright 2012], visualization [Chichakly

and Eppstein 2013], and data-mining approaches [Bandaru et al. 2017] as a step forward in reaching a better understanding of the causal relationships

between adaptive façade design and performance.

The toolchain that is presented in this thesis is developed using the building

energy simulation tool ESP-r. Several decisions about the design of the toolchain were made in response to specific characteristics in terms of

structure and algorithms of the program. ESP-r is a legacy simulation program. It is a robust tool that is evolving on top of many thousands of man-

hours of software development and testing. The core of ESP-r was developed in the 1970s and has remained largely unchanged since then. As a result, there

are some archaic features in the program (e.g. need for recompilation, and the relatively monolithic code structure) that limit the flexibility and options

for versatile use of the toolchain. Modern scripting and modeling languages

130

such as Python and Modelica offer a number of advantages that would enable

scalability, modular extension of capabilities, options for interfacing with external software (e.g. through functional mock-up units) and user-friendly

access to advanced simulation settings [Nouidui et al. 2012; Wetter 2009]. It is expected that the lessons that have been learned during the development

of the present toolchain can provide a valuable starting point for the implementation of a similar workflow in another software environment, or in

a hybrid fashion, combining the strengths of both approaches.

131

References

Agrawal, A., and F.W. Loverme. 2011. “Variable Emissivity Coatings and Their Applications - US Patent, Patent Number 20110111147.”

Altomonte, S., P. Rutherford, and R. Wilson. 2015. “Human Factors in the Design of Sustainable Built Environments.” Intelligent Buildings International 7 (4): 224–41.

Aries, M.B.C., J.A. Veitch, and G.R. Newsham. 2010. “Windows, View, and Office Characteristics Predict Physical and Psychological Discomfort.” Journal of Environmental Psychology 30 (4): 533–41.

Arora, S.K., R.W. Foley, J. Youtie, P. Shapira, and A. Wiek. 2014. “Drivers of Technology Adoption - the Case of Nanomaterials in Building Construction.” Technological Forecasting and Social Change 87: 232–44.

ASHRAE. 2013. ASHRAE Standard 90.1-2013 - Energy Standard for Buildings Except Low-Rise Residential Buildings.

Assimakopoulos, M., A. Tsangrassoulis, M. Santamouris, and G. Guarracino. 2007. “Comparing the Energy Performance of an Electrochromic Window under Various Control Strategies.” Building and Environment 42 (8): 2829–34.

Attia, S., M. Hamdy, W. O’Brien, and S. Carlucci. 2013. “Assessing Gaps and Needs for Integrating Building Performance Optimization Tools in Net Zero Energy Buildings Design.” Energy and Buildings 60 (May): 110–24.

Ayres, J.M., and E. Stamper. 1995. “Historical Development of Building Energy Calculations.” ASHRAE Transactions 101 (1): 841–49.

Baetens, R., B.P. Jelle, and A. Gustavsen. 2010. “Properties, Requirements and Possibilities of Smart Windows for Dynamic Daylight and Solar Energy Control in Buildings: A State-of-the-Art Review.” Solar Energy Materials and Solar Cells 94 (2): 87–105.

Baird, G., and C. Field. 2013. “Thermal Comfort Conditions in Sustainable Buildings – Results of a Worldwide Survey of Users’ Perceptions.” Renewable Energy 49 (January): 44–47.

Bakker, L.G., E.C.M. Hoes-van Oeffelen, R.C.G.M. Loonen, and J.L.M. Hensen. 2014. “User Satisfaction and Interaction with Automated Dynamic Facades: A Pilot Study.” Building and Environment 78 (August): 44–52.

Balci, O. 1994. “Validation, Verification, and Testing Techniques throughout the Life Cycle of a Simulation Study.” Annals of Operations Research 53: 215–20.

Bandaru, S., A.H.C. Ng, and K. Deb. 2017. “Data Mining Methods for Knowledge Discovery in Multi-Objective Optimization: Part A - Survey.” Expert Systems with Applications 70: 139–59.

Bastiaansen, C.W.M., A.P.H.J. Schenning, M.G. Debije, and D.J. Broer. 2013. “Nano-Textured Polymers for Future Architectural Needs.” Journal of Facade Design & Engineering 1: 97–104.

Bergeron, B.V., K.C. White, J.L. Boehme, A.H. Gelb, and P.B. Joshi. 2008. “Variable Absorptance and Emittance Devices for Thermal Control.”

132

Journal of Physical Chemistry C 112 (3): 832–38.

Bionda, D, U.P. Menti, and H Manz. 2014. “Kann Eine Gebaüdehülle Mit Steuerbarem U-Wert Zur Reduktion Des Kühlenenergiebedarfs Beitragen?” In Proceedings of BauSIM2014, the Fifth German-Austrian IBPSA Conference, 649–55.

Bleil de Souza, C., and S. Tucker. 2014. “Thermal Simulation Software Outputs: A Framework to Produce Meaningful Information for Design Decision-Making.” Journal of Building Performance Simulation 8 (2): 57-78.

Bluyssen, P.M. 2010. “Towards New Methods and Ways to Create Healthy and Comfortable Buildings.” Building and Environment 45 (4): 808–18.

Boehm, B.W. 1988. “A Spiral Model of Software Development and Enhancement.” Computer 21 (5): 61–72.

Boschert, S., and R. Rosen. 2016. “Digital Twin - The Simulation Aspect.” In Mechatronic Futures: Challenges and Solutions for Mechatronic Systems and Their Designers, edited by P Hehenberger and D. Bradley, 59–74. Springer International Publishing.

Brager, G.S., H. Zhang, and E. Arens. 2015. “Evolving Opportunities for Providing Thermal Comfort.” Building Research & Information 43 (3): 274–87.

Brahme, R., Z. O’Neill, W. Sisson, and K. Otto. 2009. “Using Existing Whole Building Energy Tools for Designing Net-Zero Energy Buildings – Challenges and Workarounds.” In Proceedings of Building Simulation 2009, 9–16. Glasgow, Scotland.

Brownlee, A., and J.A. Wright. 2012. “Solution Analysis in Multi-Objective Optimization.” In Proceedings of Building Simulation and Optimization, 317–24. Loughborough, UK.

Carbonari, A., R. Fioretti, B. Naticchia, and P. Principi. 2012. “Experimental Estimation of the Solar Properties of a Switchable Liquid Shading System for Glazed Facades.” Energy and Buildings 45 (February): 299–310.

Cardin, M.A. 2013. “Enabling Flexibility in Engineering Systems: A Taxonomy of Procedures and a Design Framework.” Journal of Mechanical Design 136 (1): 11005.

Chen, T.Y., and A.K. Athienitis. 2003. “Investigation of Practical Issues in Building Thermal Parameter Estimation.” Building and Environment 38 (8): 1027–38.

Chichakly, K.J., and M.J. Eppstein. 2013. “Discovering Design Principles from Dominated Solutions.” IEEE Access 1 (iii): 275–89.

Choo, E.U., B. Schoner, and W.C. Wedley. 1999. “Interpretation of Criteria Weights in Multicriteria Decision Making.” Computers & Industrial Engineering 37 (3): 527–41.

Chun, W., S. J. Oh, H. Han, J. T. Kim, and K. Chen. 2008. “Overview of Several Novel Thermodiode Designs and Their Application in Buildings.” International Journal of Low-Carbon Technologies 3 (2): 83–100.

CIBSE. 2009. CIBSE Guide H - Building Control Systems.

Citherlet, S., J.A. Clarke, and J. Hand. 2001. “Integration in Building Physics Simulation.” Energy and Buildings 33 (5): 451–461.

133

Claridge, D. 1977. “Window Management and Energy Savings.” Energy and Buildings 1 (1): 57–63.

Clarke, J.A. 2001. Energy Simulation in Building Design. Oxford: Butterworth-Heinemann.

Clarke, J.A., J. Cockroft, S. Conner, J.W. Hand, N.J. Kelly, R. Moore, T. O’Brien, and P. Strachan. 2002. “Simulation-Assisted Control in Building Energy Management Systems.” Energy and Buildings 34 (9): 933–940.

Clarke, J.A., and J.L.M. Hensen. 2015. “Integrated Building Performance Simulation: Progress, Prospects and Requirements.” Building and Environment 91 (September): 294–306.

Claros-Marfil, L.J., A. Dentel, F.F. Padial, and B. Lauret. 2014. “Active and Passive PCM Walls Simulation – a New TRNSYS PCM-Type.” In Proceedings of the 1st International Congress on Research in Construction and Architectural Technologies.

Coffey, B. 2013. “Approximating Model Predictive Control with Existing Building Simulation Tools and Offline Optimization.” Journal of Building Performance Simulation 6 (3): 220–35.

Coffey, B., F. Haghighat, E. Morofsky, and E. Kutrowski. 2010. “A Software Framework for Model Predictive Control with GenOpt.” Energy and Buildings 42 (7): 1084–92.

Cole, R.J., and L. Fedoruk. 2015. “Shifting from Net-Zero to Net-Positive Energy Buildings.” Building Research & Information 43 (June 2015): 111–20.

Corbin, C.D., G.P. Henze, and P. May-Ostendorp. 2013. “A Model Predictive Control Optimization Environment for Real-Time Commercial Building Application.” Journal of Building Performance Simulation 6 (3): 159–74.

Coyle, P.E. 2011. The Missing Middle. National Defense Industrial Association Science & Engineering Technology Division.

Crawley, D., J. Hand, M. Kummert, and B. Griffith. 2008. “Contrasting the Capabilities of Building Energy Performance Simulation Programs.” Building and Environment 43 (4): 661–73.

Dave, S., and M. Andersen. 2011. “A Comprehensive Method to Determine Performance Metrics for Complex Fenestration Systems.” In 27th Conference on Passive and Low Energy Architecture (PLEA), 2–7.

Davidsson, H., B. Perers, and B. Karlsson. 2010. “Performance of a Multifunctional PV/T Hybrid Solar Window.” Solar Energy 84 (3): 365–72.

Davies, M. 1981. “A Wall for All Seasons.” RIBA Journal 88 (2): 55–57.

DAYSIM. 2015. “Daysim v3.1 - Advanced Daylight Simulation Software.” http://daysim.ning.com/.

De Klijn-Chevalerias, M.L., R.C.G.M. Loonen, A. Zarzycka, D. De Witte, M.V. Sarakinioti, and J.L.M. Hensen. 2017. “Assisting the Development of Innovative Responsive Façade Elements Using Building Performance Simulation.” In Proceedings of SimAUD2017, Symposium on Simulation for Architecture and Urban Design, 243–50. Toronto, Canada.

De Wilde, P., G. Augenbroe, and M. Van der Voorden. 2002. “Design Analysis Integration: Supporting the Selection of Energy Saving Building Components.” Building and Environment 37 (8–9): 807–16.

134

De Wilde, P., W. Tian, and G. Augenbroe. 2011. “Longitudinal Prediction of the Operational Energy Use of Buildings.” Building and Environment 46 (8): 1670–80.

De Witte, D., M.L. De Klijn-Chevalerias, R.C.G.M. Loonen, J.L.M. Hensen, U. Knaack, and G. Zimmermann. 2017. “Convective Concrete: Additive Manufacturing to Facilitate Activation of Thermal Mass.” Journal of Facade Design and Engineering 5 (1).

Deb, K., and A. Srinivasan. 2006. “Innovization: Innovating Design Principles through Optimization.” In Proceedings of the 8th Annual Conference on Genetic and Evolutionary Computation (GECCO’06), 1629–36.

Debije, M.G. 2010. “Solar Energy Collectors with Tunable Transmission.” Advanced Functional Materials 20 (9): 1498–1502.

DeForest, N., A. Shehabi, J. O’Donnell, G. Garcia, J. Greenblatt, E.S. Lee, S. Selkowitz, and D.J. Milliron. 2015. “United States Energy and CO2 Savings Potential from Deployment of near-Infrared Electrochromic Window Glazings.” Building and Environment 89: 107–17.

DeForest, N., A. Shehabi, G. Garcia, J. Greenblatt, E. Masanet, E.S. Lee, S. Selkowitz, and D.J. Milliron. 2013. “Regional Performance Targets for Transparent near-Infrared Switching Electrochromic Window Glazings.” Building and Environment 61 (March): 160–68.

Delcroix, B., M. Kummert, A. Daoud, and M. Hiller. 2012. “Conduction Transfer Functions in TRNSYS Multizone Building Model: Current Implementation, Limitations and Possible Improvements.” In Proceedings of SimBuild 20102, the Fifth National Conference of IBPSA-USA, 219–26. Madison, WI.

Den Heijer, A. 2013. “Assessing Facade Value - How Clients Make Business Cases in Changing Real Estate Markets.” Journal of Facade Design and Engineering 1: 3–16.

Deru, M., K. Field, D. Studer, K. Benne, B. Griffith, P. Torcellini, B. Liu, et al. 2011. U.S. Department of Energy Commercial Reference Building Models of the National Building Stock - NREL/TP-5500-46861.

Deuble, M.P., and R.J. de Dear. 2012. “Green Occupants for Green Buildings: The Missing Link?” Building and Environment 56 (October): 21–27.

Djedjig, R., E. Bozonnet, and R. Belarbi. 2015. “Analysis of Thermal Effects of Vegetated Envelopes: Integration of a Validated Model in a Building Energy Simulation Program.” Energy and Buildings 86: 93–103.

Du Montier, C., A. Potvin, and C.M.H. Demers. 2013. “Adaptive Façades for Architecture : Energy and Lighting Potential of Movable Insulation Panels.” In Proceedings of PLEA2013 - Conference, Sustainable Architecture for a Renewable Future, 1–6. Munich, Germany.

Dubois, M.C. 1999. “The Design of Seasonal Awnings for Low Cooling and Heating Loads in Offices.” In Proceedings of the 5th Symposium on Building Physics in the Nordic Countries, 505–12.

EC. 2011. A Roadmap for Moving to a Competitive Low Carbon Economy.

EC. 2013. Energy-Efficient Buildings - Multi-Annual Roadmap for the Contractual PPP under Horizon 2020.

Elsarrag, E., Y. Al-Horr, and M.S.E. Imbabi. 2012. “Improving Building Fabric

135

Energy Efficiency in Hot-Humid Climates Using Dynamic Insulation.” Building Simulation 5 (2): 127–34.

EN 15251. 2007. “CEN Standard EN 15251: Indoor Environmental Input Parameters for Design and Assessment of Energy Performance of Buildings Addressing Indoor Air Quality, Thermal Environment, Lighting and Acoustics.” Comité Européen de Normalisation (CEN).

Eppinger, S.D., M.V. Nukala, and D.E. Whitney. 1997. “Generalised Models of Design Interation Using Signal Flow Graphs.” Research in Engineering Design 9: 112–23.

Erell, E. 2004. “‘SOLVENT’: Development of a Reversible Solar-Screen Glazing System.” Energy and Buildings 36 (5): 467–80.

Erickson, J. 2013. “Envelope as Climate Negotiator: Evaluating Adaptive Building Envelope’s Capacity to Moderate Indoor Climate and Energy.” PhD Thesis, Arizona State University.

Evans, M., and N.J. Kelly. 1996. Modelling Active Building Elements with Special Materials. ESRU Occasional Paper, University of Strathclyde, Glasgow.

Evins, R. 2013. “A Review of Computational Optimisation Methods Applied to Sustainable Building Design.” Renewable and Sustainable Energy Reviews 22 (June): 230–45.

Evins, R., and K. Orehounig. 2014. “A Bi-Level Design and Operation Optimisation Process Applied to an Energy Centre.” In Proceedings of Building Simulation and Optimization BSO2014. London, UK.

Fathy, H.K., J.A. Reye, P.Y. Papalambros, and A.G. Ulsog. 2001. “On the Coupling between the Plant and Controller Optimization Problems.” In Proceedings of the American Control Conference, 1864–69.

Favoino, F., Y. Cascone, L. Bianco, F. Goia, M. Zinzi, M. Overend, V. Serra, and M. Perino. 2015. “Simulating Switchable Glazing with EnergyPlus: An Empirical Validation and Calibration of a Thermotropic Glazing Model.” In Proceedings of Building Simulation 2015. Hyderabad, India.

Favoino, F., F. Goia, M. Perino, and V. Serra. 2014. “Experimental Assessment of the Energy Performance of an Advanced Responsive Multifunctional Façade Module.” Energy and Buildings 68 (January): 647–59.

Favoino, F., Q. Jin, and M. Overend. 2017. “Design and Control Optimisation of Adaptive Insulation Systems for Office Buildings. Part 1: Adaptive Technologies and Simulation Framework.” Energy 127: 301–9.

Favoino, F., M. Overend, and Q. Jin. 2015. “The Optimal Thermo-Optical Properties and Energy Saving Potential of Adaptive Glazing Technologies.” Applied Energy 156 (October 2015,): 1–15.

Fawcett, W., M. Hughes, H. Krieg, S. Albrecht, and A. Vennström. 2012. “Flexible Strategies for Long-Term Sustainability under Uncertainty.” Building Research & Information 40 (5): 545–57.

Ferguson, S., A. Siddiqi, K. Lewis, and O. De Weck. 2007. “Flexible and Reconfigurable Systems: Nomenclature and Review.” In Proceedings of ASME 2007 - International Design Engineering Technical Conferences & Computers and Information in Engineering Conference. Las Vegas.

136

Fernandes, L.L., E.S. Lee, and G. Ward. 2013. “Lighting Energy Savings Potential of Split-Pane Electrochromic Windows Controlled for Daylighting with Visual Comfort.” Energy and Buildings 61 (2010): 8–20.

Fernández Solla, I. 2014. “Time Scale and Adaptive Envelopes – Minutes or Decades?” In Façade2014 – Conference on Building Envelopes, 18–37. Luzern, Switzerland.

Fiorito, F., M. Sauchelli, D. Arroyo, M. Pesenti, M. Imperadori, G. Masera, and G. Ranzi. 2016. “Shape Morphing Solar Shadings: A Review.” Renewable and Sustainable Energy Reviews 55: 863–84.

Fitzgerald, M.E., and A.M. Ross. 2012. “Mitigating Contextual Uncertainties with Valuable Changeability Analysis in the Multi-Epoch Domain.” 2012 IEEE International Systems Conference SysCon 2012, March, 1–7.

Gagne, J., and M. Andersen. 2012. “A Generative Facade Design Method Based on Daylighting Performance Goals.” Journal of Building Performance Simulation 5 (3): 141–54.

Gambatese, J.A., and M. Hallowell. 2011. “Factors That Influence the Development and Diffusion of Technical Innovations in the Construction Industry.” Construction Management and Economics 29 (5): 507–17.

Givoni, B. 2011. “Indoor Temperature Reduction by Passive Cooling Systems.” Solar Energy 85 (8): 1692–1726.

Goia, F. 2016. “Search for the Optimal Window-to-Wall Ratio in Office Buildings in Different European Climates and the Implications on Total Energy Saving Potential.” Solar Energy 132: 467–92.

Goia, F., and Y. Cascone. 2014. “The Impact of an Ideal Dynamic Building Envelope on the Energy Performance of Low Energy Office Buildings.” Energy Procedia 58 (1876): 185–92.

Goia, F., M. Perino, and V. Serra. 2013. “Improving Thermal Comfort Conditions by Means of PCM Glazing Systems.” Energy and Buildings 60 (May): 442–52.

Goia, F., M. Perino, and V. Serra. 2014. “Experimental Analysis of the Energy Performance of a Full-Scale PCM Glazing Prototype.” Solar Energy 100 (February): 217–33.

Granqvist, C.G., S. Green, G.A. Niklasson, N.R. Mlyuka, S. Von Kræmer, and P. Georén. 2010. “Advances in Chromogenic Materials and Devices.” Thin Solid Films 518 (11): 3046–53.

Grant, A., and R. Ries. 2013. “Impact of Building Service Life Models on Life Cycle Assessment.” Building Research & Information 41 (2): 168–86.

Grynning, S., F. Goia, E. Rognvik, and B. Time. 2013. “Possibilities for Characterization of a PCM Window System Using Large Scale Measurements.” International Journal of Sustainable Built Environment 2 (1): 56–64.

Gunay, H.B., J. Bursill, B. Huchuk, W. O’Brien, and I. Beausoleil-Morrison. 2014. “Shortest-Prediction-Horizon Model-Based Predictive Control for Individual Offices.” Building and Environment 82 (December): 408–19.

Hakkinen, T., and K. Belloni. 2011. “Barriers and Drivers for Sustainable Building.” Building Research & Information 39 (3): 239–55.

137

Hamdy, M., A. Hasan, and K. Siren. 2011. “Applying a Multi-Objective Optimization Approach for Design of Low-Emission Cost-Effective Dwellings.” Building and Environment 46 (1): 109–23.

Hand, J.W. 1998. “Removing Barriers to the Use of Simulation in the Building Design Professions.” PhD Thesis, University of Strathclyde, Glasgow, UK.

Hashemi, A., and S. Gage. 2012. “Technical Issues That Affect the Use of Retrofit Panel Thermal Shutters in Commercial Buildings.” Building Services Engineering Research and Technology 35 (1): 6–22.

He, J., and A. Hoyano. 2011. “Experimental Study of Practical Applications of a Passive Evaporative Cooling Wall with High Water Soaking-up Ability.” Building and Environment 46 (1): 98–108.

Heim, D., and J. Clarke. 2004. “Numerical Modelling and Thermal Simulation of PCM–gypsum Composites with ESP-r.” Energy and Buildings 36 (8): 795–805.

Heiselberg, P., ed. 2009. IEA ECBCS Annex 44 Integrating Environmentally Responsive Elements in Buildings - Expert Guide Part 1: Responsive Building Concepts.

Hellinga, H., and T. Hordijk. 2014. “The D&V Analysis Method: A Method for the Analysis of Daylight Access and View Quality.” Building and Environment 79 (May): 101–14.

Hensen, J.L.M., and R Lamberts. 2011. Building Performance Simulation for Design and Operation. London: Spon Press.

Hensen, J.L.M., R.C.G.M. Loonen, M. Archontiki, and M. Kanellis. 2015. “Using Building Simulation for Moving Innovations across the ‘Valley of Death.’” REHVA Journal 52 (3): 58–62.

Henze, G.P., and C. Neumann. 2011. “Building Simulation in Building Automation Systems.” In Building Performance Simulation for Design and Operation, edited by J.L.M. Hensen and R. Lamberts, 402–40. Routledge.

Hiller, M., and P. Schöttl. 2014. “Modellierung Komplexer Verglasungssysteme in Trnsys.” In Proceedings of BauSIM2014, the Fifth German-Austrian IBPSA Conference, 387–94.

Hoes, P. 2014. “Computational Performance Prediction of the Potential of Hybrid Adaptable Thermal Storage Concepts for Lightweight Low-Energy Houses.” Doctoral dissertation, Eindhoven University of Technology.

Hoes, P., and J.L.M. Hensen. 2016. “The Potential of Lightweight Low-Energy Houses with Hybrid Adaptable Thermal Storage: Comparing the Performance of Promising Concepts.” Energy and Buildings 110: 79–93.

Hoes, P., R.C.G.M. Loonen, M. Trčka, and J.L.M. Hensen. 2012. “Performance Prediction of Advanced Building Controls in the Design Phase Using ESP-R, BCVTB and Matlab.” In Proceedings of Building Simulation and Optimization. Loughborough, UK.

Hoes, P., M. Trčka, J.L.M. Hensen, and B. Hoekstra Bonnema. 2011. “Investigating the Potential of a Novel Low-Energy House Concept with Hybrid Adaptable Thermal Storage.” Energy Conversion and Management 52 (6): 2442–47.

Hofman, M., and B. Dujardin. 2008. “European Patent EP1644591: Balcony

138

That Can Be Folded.”

Holmes, M., and J. Hacker. 2007. “Climate Change, Thermal Comfort and Energy: Meeting the Design Challenges of the 21st Century.” Energy and Buildings 39 (7): 802–14.

Hopfe, C.J. 2009. “Uncertainty and Sensitivity Analysis in Building Performance Simulation for Decision Support and Design Optimization.” PhD Thesis. Eindhoven University of Technology.

Hutcheon, N. 1963. “Requirements for Exterior Walls.” Canadian Building Digest 48.

IBPSA-USA. 2018. “Building Energy Software Tools Directory.” https://www.buildingenergysoftwaretools.com/software-listing.

IEA. 2013a. Technology Roadmap - Energy Efficient Building Envelopes. International Energy Agency.

IEA. 2013b. Transition to Sustainable Buildings - Strategies and Opportunities to 2050. International Energy Agency.

IEA. 2014. Tracking Clean Energy Progress 2014. International Energy Agency.

IEA. 2016. Energy Technology Perspectives 2016 - Towards Sustainable Urban Energy Systems. International Energy Agency.

Imbabi, M. 2006. “Modular Breathing Panels for Energy Efficient, Healthy Building Construction.” Renewable Energy 31 (5): 729–38.

IPCC. 2014. Buildings. In: Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Edited by O. Lucon and A. Ürge-Vorsatz.

Ismail, K. 2001. “Thermally Effective Windows with Moving Phase Change Material Curtains.” Applied Thermal Engineering 21 (18): 1909–23.

Jakob, M., and R. Madlener. 2004. “Riding down the Experience Curve for Energy-Efficient Building Envelopes: The Swiss Case for 1970–2020.” International Journal of Energy Technology and Policy 2 (1/2): 153–78.

Jakubiec, J., and C. Reinhart. 2011. “The ‘Adaptive Zone’ - A Concept for Assessing Discomfort Glare throughout Daylit Spaces.” Lighting Research and Technology 44 (2): 149–70.

Janak, M. 1997. “Coupling Building Energy and Lighting Simulation.” In Proceedings of Building Simulation 1997. Prague, Czech Republic.

Jin, Q., and M. Overend. 2014. “A Prototype Whole-Life Value Optimization Tool for Façade Design.” Journal of Building Performance Simulation 7 (3): 217–32.

Joe, J., W. Choi, H. Kwon, and J. Huh. 2013. “Load Characteristics and Operation Strategies of Building Integrated with Multi-Story Double Skin Facade.” Energy and Buildings 60 (May): 185–98.

Jonsson, A., and A. Roos. 2010. “Evaluation of Control Strategies for Different Smart Window Combinations Using Computer Simulations.” Solar Energy 84 (1): 1–9.

Judkoff, R., D. Wortman, B. O’Doherty, and J. Burch. 2008. A Methodology for Validating Building Energy Analysis Simulations. NREL Technical Report

139

550-42059.

Karlessi, T., M. Santamouris, K. Apostolakis, A. Synnefa, and I. Livada. 2009. “Development and Testing of Thermochromic Coatings for Buildings and Urban Structures.” Solar Energy 83 (4): 538–51.

Kasinalis, C., R.C.G.M. Loonen, D. Cóstola, and J.L.M. Hensen. 2014. “Framework for Assessing the Performance Potential of Seasonally Adaptable Facades Using Multi-Objective Optimization.” Energy and Buildings 79 (August): 106–13.

Keilholz, W., P. Sette, and E. Soussi. 2009. The W Programming Language User & Reference Manual. CSTB France.

Kerrigan, E.C., A. Bemporad, D. Mignone, M. Morari, and J. Maciejowski. 2000. “Multi-Objective Prioritisation and Reconfiguration for the Control of Constrained Hybrid Systems.” In Proceedings of the American Control Conference, 1694–98. Chicago, Ill.

Khandelwal, H., R.C.G.M. Loonen, J.L.M. Hensen, M.G. Debije, and A.P.H.J. Schenning. 2015. “Electrically Switchable Polymer Stabilised Broadband Infrared Reflectors and Their Potential as Smart Windows for Energy Saving in Buildings.” Scientific Reports 5 (April): 11773.

Khandelwal, H., R.C.G.M. Loonen, J.L.M. Hensen, A.P.H.J. Schenning, and M.G. Debije. 2014. “Application of Broadband Infrared Reflector Based on Cholesteric Liquid Crystal Polymer Bilayer Film to Windows and Its Impact on Reducing the Energy Consumption in Buildings.” Journal of Materials Chemistry A 2 (35): 14622–27.

Khandelwal, H., A.P.H.J. Schenning, and M.G. Debije. 2017. “Infrared Regulating Smart Window Based on Organic Materials.” Advanced Energy Materials 7 (14): 1602209.

Khire, R.A., and A. Messac. 2008. “Selection-Integrated Optimization (SIO) Methodology for Optimal Design of Adaptive Systems.” Journal of Mechanical Design 130 (10): 101401.

Kimber, M., W.W. Clark, and L. Schaefer. 2014. “Conceptual Analysis and Design of a Partitioned Multifunctional Smart Insulation.” Applied Energy 114 (February): 310–19.

Kiss, B., and L. Neij. 2011. “The Importance of Learning When Supporting Emergent Technologies for Energy efficiency—A Case Study on Policy Intervention for Learning for the Development of Energy Efficient Windows in Sweden.” Energy Policy 39 (10): 6514–24.

Klein, T. 2014. “Integral Facade Construction. Towards a New Product Architecture for Curtain Walls.” PhD Thesis. Delft University of Technology.

Kleindienst, S., and M. Andersen. 2012. “Comprehensive Annual Daylight Design through a Goal-Based Approach.” Building Research & Information 40 (2): 154–73.

Knaack, U., T. Klein, and M. Bilow. 2008. Imagine 01 - Façades. Rotterdam: 010 Publishers.

Kolokotsa, D., D. Rovas, E. Kosmatopoulos, and K. Kalaitzakis. 2011. “A Roadmap towards Intelligent Net Zero- and Positive-Energy Buildings.”

140

Solar Energy 85 (12): 3067–84.

Konstantinou, T., and U. Knaack. 2013. “An Approach to Integrate Energy Efficiency Upgrade into Refurbishment Design Process, Applied in Two Case-Study Buildings in Northern European Climate.” Energy and Buildings 59: 301–9.

Kuhn, T.E., S. Herkel, F. Frontini, P. Strachan, and G. Kokogiannakis. 2011. “Solar Control: A General Method for Modelling of Solar Gains through Complex Facades in Building Simulation Programs.” Energy and Buildings 43 (1): 19–27.

Kuznik, F., J. Virgone, and K. Johannes. 2010. “Development and Validation of a New TRNSYS Type for the Simulation of External Building Walls Containing PCM.” Energy and Buildings 42 (7): 1004–9.

LBNL. 2017a. “Optics6.” https://windows.lbl.gov/software/Optics/optics.html.

LBNL. 2017b. “WINDOW7.” http://windows.lbl.gov/software/window/window.html.

Leal, V., and E. Maldonado. 2008. “The Role of the PASLINK Test Cell in the Modelling and Integrated Simulation of an Innovative Window.” Building and Environment 43 (2): 217–27.

Lee, C.S. 2017. “Simulation-Based Performance Assessment of Climate Adaptive Greenhouse Shells.” PhD Thesis, TU Eindhoven.

Lee, E.S., E.S. Claybaugh, and M. LaFrance. 2012. “End User Impacts of Automated Electrochromic Windows in a Pilot Retrofit Application.” Energy and Buildings 47 (2010): 267–84.

Lee, E.S., S Kim, S. Kim, Y Park, and J Park. 2006. “Experimental Research on Performance of the Solar Heat Gain (SHG) Insulation Panel.” Building and Environment 41 (3): 336–42.

Lee, E.S., X. Pang, S. Hoffmann, H. Goudey, and A. Thanachareonkit. 2013. “An Empirical Study of a Full-Scale Polymer Thermochromic Window and Its Implications on Material Science Development Objectives.” Solar Energy Materials and Solar Cells 116 (September): 14–26.

Lienhard, J., S. Schleicher, S. Poppinga, T. Masselter, M. Milwich, T. Speck, and J. Knippers. 2011. “Flectofin: A Hingeless Flapping Mechanism Inspired by Nature.” Bioinspiration & Biomimetics 6 (4): 45001.

Liu, M., K.B. Wittchen, and P.K. Heiselberg. 2015. “Control Strategies for Intelligent Glazed Façade and Their Influence on Energy and Comfort Performance of Office Buildings in Denmark.” Applied Energy 145: 43–51.

Liu, M., K.B. Wittchen, and P.V. Heiselberg. 2014. “Development of a Simplified Method for Intelligent Glazed Façade Design under Different Control Strategies and Verified by Building Simulation Tool BSim.” Building and Environment 74 (April): 31–38.

Llordés, A., G. Garcia, J. Gazquez, and D.J. Milliron. 2013. “Tunable near-Infrared and Visible-Light Transmittance in Nanocrystal-in-Glass Composites.” Nature 500: 323–27.

Lomanowski, B.A., and J.L. Wright. 2012. “The Complex Fenestration Construction: A Practical Approach for Modelling Windows with Shading

141

Devices in ESP-r.” Journal of Building Performance Simulation 5 (3): 185–98.

Loonen, R.C.G.M. 2015. “Bio-Inspired Adaptive Building Skins.” In Biotechnologies and Biomimetics for Civil Engineering, edited by F. Pacheco-Torgal, J.A. Labrincha, M.V. Diamanti, C.-P. Yu, and H.K. Lee, 115–24. Springer International Publishing.

Loonen R.C.G.M. 2018. “Climate Adaptive Building Shells. http://www.pinterest.com/CABSoverview/.”

Loonen, R.C.G.M., B. de Boer, W. Kornaat, L.G. Bakker, and J.L.M. Hensen. 2015. “Zoektocht Naar de Ideale Adaptieve Gevel.” TVVL Magazine 44 (7–8): 36–38.

Loonen, R.C.G.M., F Favoino, J.L.M. Hensen, and M. Overend. 2017. “Review of Current Status, Requirements and Opportunities for Building Performance Simulation of Adaptive Facades.” Journal of Building Performance Simulation 10 (2): 205–23.

Loonen, R.C.G.M., and J.L.M. Hensen. 2013. “Dynamic Sensitivity Analysis for Performance-Based Building Desing and Operation.” In Proceedings of Building Simulation 2013.

Loonen, R.C.G.M., P. Hoes, and J.L.M. Hensen. 2014. “Performance Prediction of Buildings with Responsive Building Envelopes - Some Challenges and Solutions.” In Proceedings of Building Simulation and Optimization, IBPSA BSO14, 1–8. London, UK.

Loonen R.C.G.M., M.G.L.C. Loomans, and J.L.M. Hensen. 2015. “Towards predicting the satisfaction with indoor environmental quality in building performance simulation.” In Proceedings of Healthy Buildings Europe 2015, 1-8. Eindhoven, The Netherlands.

Loonen, R.C.G.M., S. Singaravel, M. Trčka, D. Cóstola, and J.L.M. Hensen. 2014. “Simulation-Based Support for Product Development of Innovative Building Envelope Components.” Automation in Construction 45 (September): 86–95.

Loonen, R.C.G.M., M. Trčka, D. Cóstola, and J.L.M. Hensen. 2013. “Climate Adaptive Building Shells: State-of-the-Art and Future Challenges.” Renewable and Sustainable Energy Reviews 25 (September): 483–93.

Loonen, R.C.G.M., M. Trčka, and J.L.M. Hensen. 2011. “Exploring the Potential of Climate Adaptive Building Shells.” In Proceedings of Building Simulation 2011, 2148–55. Sydney, Australia.

Loosemore, M., and J. Richard. 2015. “Valuing Innovation in Construction and Infrastructure: Getting Clients Past a Lowest Price Mentality.” Engineering, Construction and Architectural Management 22 (1): 38–53.

Lorenz, W. 1998. “Design Guidelines for a Glazing with a Seasonally Dependent Solar Transmittance.” Solar Energy 63 (2): 79–96.

Luible, A. 2015. “COST Action 1403.” In Proceedings of Energy Forum 2015. Bern, Switzerland.

MacQueen, J. 1997. “The Modeling and Simulation of Energy Management Control Systems.” PhD Thesis, University of Strathclyde, Glasgow.

Mahdavi, A. 2008. “Predictive Simulation-Based Lighting and Shading

142

Systems Control in Buildings.” Building Simulation 1 (1): 25–35.

Mankins, J.C. 2009. “Technology Readiness Assessments : A Retrospective.” Acta Astronautica 65 (9–10): 1216–23.

Mardaljevic, J., M. Andersen, N. Roy, and J. Christoffersen. 2012. “Daylighting Metrics: Is There a Relationship between Useful Daylight Illuminance and Daylight Glare Probability.” In Proceedings of Building Simulation and Optimization, 189–96. Loughborough, UK.

Markham, S.K., S.J. Ward, L. Aiman-Smith, and A.I. Kingon. 2010. “The Valley of Death as Context for Role Theory in Product Innovation.” Journal of Product Innovation Management 27 (3): 402–17.

Marszal, A.J., P. Heiselberg, J.S. Bourrelle, E. Musall, K. Voss, I. Sartori, and A. Napolitano. 2011. “Zero Energy Building – A Review of Definitions and Calculation Methodologies.” Energy and Buildings 43 (4): 971–79.

Matlab. 2015. “Matlab.” http://nl.mathworks.com/products/matlab/.

May-Ostendorp, P., and G. Henze. 2013. “Experimental Investigation of Model Predictive Control-Based Rules for a Radiantly Cooled Office.” HVAC&R Research, no. July 2013: 37–41.

May-Ostendorp, P., Gregor P. Henze, B. Rajagopalan, and C. Corbin. 2013. “Extraction of Supervisory Building Control Rules from Model Predictive Control of Windows in a Mixed Mode Building.” Journal of Building Performance Simulation 6 (3): 199–219.

McDowell, T.P., D.E. Bradley, J.W. Thornton, and M. Kummert. 2004. “Simulation Synergy: Expanding TRNSYS Capabilities and Usability.” In Proceedings of SIMBUILD 2004, IBPSA-USA Conference on Building Sustainability and Performance Through Simulation, 1–6.

McLeod, R.S., C.J. Hopfe, and A. Kwan. 2013. “An Investigation into Future Performance and Overheating Risks in Passivhaus Dwellings.” Building and Environment 70 (December): 189–209.

McNeil, A., and E.S. Lee. 2013. “A Validation of the Radiance Three-Phase Simulation Method for Modelling Annual Daylight Performance of Optically Complex Fenestration Systems.” Journal of Building Performance Simulation 6 (1): 24–37.

Mela, K., T. Tiainen, and M. Heinisuo. 2012. “Comparative Study of Multiple Criteria Decision Making Methods for Building Design.” Advanced Engineering Informatics 26 (4): 716–26.

Moloney, J. 2011. Designing Kinetics for Architectural Facades: State Change. Routledge.

Nagy, Z., B. Svetozarevic, P. Jayathissa, M. Begle, J. Hofer, G. Lydon, and A. Willmann. 2016. “The Adaptive Solar Facade: From Concept to Prototypes.” Frontiers of Architectural Research 5 (2): 143–56.

Nakhi, A.E. 1995. “Adaptive Construction Modelling within Whole Building Dynamic Simulation.” PhD Thesis, University of Strathclyde, Glasgow.

Nam, C.H., and C.B. Tatum. 1992. “Strategies for Technology Push: Lessons from Construction Innovations.” Journal of Construction Engineering and Management 118 (3): 507–24.

NEN5060. 2008. “Hygrothermal Performance of Buildings - Climatic

143

Reference Data.” Netherlands Standardization Institute (NEN).

NEN7120. 2011. “Energy Performance of Buildings - Determination Method.” NEN.

Nghiem, T.X., and G.J. Pappas. 2011. “Receding-Horizon Supervisory Control of Green Buildings.” In Proceedings of the American Control Conference, 4416–21.

Nicol, F., and M.A. Humphreys. 2010. “Derivation of the Adaptive Equations for Thermal Comfort in Free-Running Buildings in European Standard EN15251.” Building and Environment 45 (1): 11–17.

Nitz, P., and H. Hartwig. 2005. “Solar Control with Thermotropic Layers.” In Solar Energy, 79:573–82.

Nouidui, T.S., A. McNeil, and E.S. Lee. 2012. Development of a Simulation-Based Controls Framework for Implementation of Control Algorithms for Complex Fenestration Systems.

O’Connor, G.C., and R.W. Veryzer. 2001. “The Nature of Market Visioning for Technology-Based Radical Innovation.” Journal of Product Innovation Management 18 (4): 231–46.

Ochoa, C.E., M.B.C. Aries, E.J. van Loenen, and J.L.M. Hensen. 2012. “Considerations on Design Optimization Criteria for Windows Providing Low Energy Consumption and High Visual Comfort.” Applied Energy 95 (July): 238–45.

Oh, S., and J.S. Haberl. 2015. “Origins of Analysis Methods Used to Design High-Performance Commercial Buildings: Whole-Building Energy Simulation.” Science and Technology for the Built Environment 4731 (October): 1–20.

Oldewurtel, F., A. Parisio, C.N. Jones, D. Gyalistras, M. Gwerder, V. Stauch, B. Lehmann, and M. Morari. 2012. “Use of Model Predictive Control and Weather Forecasts for Energy Efficient Building Climate Control.” Energy and Buildings 45 (February): 15–27.

Opricovic, S., and G.H. Tzeng. 2004. “Compromise Solution by MCDM Methods: A Comparative Analysis of VIKOR and TOPSIS.” European Journal of Operational Research 156: 445–55.

Papadopoulos, A.M. 2016. “Fourty Years of Regulations on the Thermal Performance of the Building Envelope in Europe: Achievements, Perspectives and Challenges.” Energy and Buildings 127: 942–52.

Park, B., and M. Krarti. 2016. “Energy Performance Analysis of Variable Reflectivity Envelope Systems for Commercial Buildings.” Energy and Buildings 124: 88–98.

Pedersen, CO. 2007. “Advanced Zone Simulation in EnergyPlus: Incorporation of Variable Properties and Phase Change Material (PCM) Capability.” In Proceedings of Building Simulation 2007, 1341–45. Beijing, China.

Peippo, K., P.D. Lund, and E. Vartiainen. 1999. “Multivariate Optimization of Design Trade-Offs for Solar Low Energy Buildings.” Energy and Buildings 29 (2): 189–205.

Perez, R., R. Seals, and J. Michalsky. 1993. “All-Weather Model for Sky Luminance Distribution—preliminary Configuration and Validation.” Solar

144

Energy 50: 235–45.

Perino, M., and V. Serra. 2015. “Switching from Static to Adaptable and Dynamic Building Envelopes: A Paradigm Shift for the Energy Efficiency in Buildings.” Journal of Facade Design and Engineering 3 (2): 143–63.

Petersen, F.O., C. Lindgren, B. Moller, and S.F. Vigenberg. 1990. “US Patent - US5335461A: Balcony Window with Side Railing for Installation in an Inclined Roof.”

Piccolo, A., and F. Simone. 2009. “Effect of Switchable Glazing on Discomfort Glare from Windows.” Building and Environment 44 (6): 1171–80.

Plüss, I., P. Kräuchi, D. Bionda, M. Schröcker, S. Felsenstein, and G. Zweifel. 2014. “Modellbildung Eines Phasenwechsel-Fassadenelements in IDA ICE.” In Proceedings of BauSIM2014, the Fifth German-Austrian IBPSA Conference, 374–78.

Prieto, A., U. Knaack, T. Auer, and T. Klein. 2017. “Solar Façades - Main Barriers for Widespread Façade Integration of Solar Technologies.” Journal of Facade Design & Engineering 5 (1): 51–62.

Prívara, S., Z. Váňa, E. Žáčeková, and J. Cigler. 2012. “Building Modeling: Selection of the Most Appropriate Model for Predictive Control.” Energy and Buildings 55 (September): 341–50.

Radford, A. 1981. “Optimization, Simulation and Multiple Criteria in Window Design.” Computer-Aided Design 13 (6): 345–50.

Radford, A., and J.S. Gero. 1987. Design by Optimization in Architecture, Building, and Construction. New York: John Wiley & Sons, Inc.

Radford, A., and J.S. Gero. 1980. “Tradeoff Diagrams for the Integrated Design of the Physical Environment in Buildings.” Building and Environment 15 (1): 3–15.

Reinhart, C.F. 2004. “Lightswitch-2002: A Model for Manual and Automated Control of Electric Lighting and Blinds.” Solar Energy 77 (1): 15–28.

Reinhart, C.F., and J. Wienold. 2011. “The Daylighting Dashboard – A Simulation-Based Design Analysis for Daylit Spaces.” Building and Environment 46 (2): 386–96.

Ritter, V. 2014. “Challenges in the Development of the Adaptive Solar Thermal Facade System Fluidglass.” In Proceedings of Façade2014 – Conference on Building Envelopes. Luzern, Switzerland.

Roberz, F., R.C.G.M. Loonen, P. Hoes, and J.L.M. Hensen. 2017. “Ultra-Lightweight Concrete: Energy and Comfort Performance Evaluation in Relation to Buildings with Low and High Thermal Mass.” Energy & Buildings 138: 432–42.

Robinson, S. 2007. “Conceptual Modelling for Simulation Part I: Definition and Requirements.” Journal of the Operational Research Society 59 (3): 278–90.

Roulet, C.A., F. Flourentzou, F. Foradini, P. Bluyssen, C. Cox, and C. Aizlewood. 2006. “Multicriteria Analysis of Health, Comfort and Energy Efficiency in Buildings.” Building Research & Information 34 (5): 475–82.

Runsheng, T., Y. Etzion, and E. Erell. 2003. “Experimental Studies on a Novel Roof Pond Configuration for the Cooling of Buildings” Building and

145

Environment 28: 1513–22.

Ryghaug, M., and K.H. Sørensen. 2009. “How Energy Efficiency Fails in the Building Industry.” Energy Policy 37 (3): 984–91.

Sadeghi, S.A., P. Karava, I. Konstantzos, and A. Tzempelikos. 2016. “Occupant Interactions with Shading and Lighting Systems Using Different Control Interfaces: A Pilot Field Study.” Building and Environment 97: 177–95.

Sadineni, S.B., S. Madala, and R.F. Boehm. 2011. “Passive Building Energy Savings: A Review of Building Envelope Components.” Renewable and Sustainable Energy Reviews 15 (8): 3617–31.

Saelens, D. 2002. “Energy Performance Assessment of Single Storey Multiple-Skin Facades.” PhD Thesis. KU Leuven.

Sailor, D.J. 2008. “A Green Roof Model for Building Energy Simulation Programs.” Energy and Buildings 40 (8): 1466–78.

Sailor, D.J., and B. Bass. 2014. “Development and Features of the Green Roof Energy Calculator (GREC).” Journal of Living Architecture 1 (3): 36–58.

Sarey Khanie, M., J. Stoll, W. Einhäuser, J. Wienold, and M. Andersen. 2017. “Gaze and Discomfort Glare, Part 1: Development of a Gaze-Driven Photometry.” Lighting Research and Technology 49 (7): 845–65.

Saxena, M., G. Ward, T. Perry, L. Heschong, and R. Higa. 2010. “Dynamic Radiance - Predicting Annual Daylighting with Variable Fenestration Optics Using BSDFs.” In Proceedings of SimBuild 2010, 402–9.

Schumacher, M., O. Schaeffer, and M. Vogt. 2010. Move: Architecture in Motion - Dynamic Components and Elements. Basel: Birkhäuser.

Seeboth, A., R. Ruhmann, and O. Mühling. 2010. “Thermotropic and Thermochromic Polymer Based Materials for Adaptive Solar Control.” Materials 3 (12): 5143–68.

Selkowitz, S., and V. Bazjanac. 1981. “Thermal Performance of Managed Window Systems.” ASHRAE Transactions 85 (1): 392–408.

Shove, E. 1999. “Gaps, Barriers and Conceptual Chasms : Theories of Technology Transfer and Energy in Buildings.” Energy Policy 26 (15): 1105–12.

Silver, M.R., and O.L. de Weck. 2007. “Time-Expanded Decision Networks : A Framework for Designing Evolvable Complex Systems.” Systems Engineering 10 (2): 167–86.

Sol, J.A.H.P., G.H. Timmermans, A.J. van Breugel, A.P.H.J. Schenning, and M.G. Debije. 2018. “Multistate Luminescent Solar Concentrator ‘Smart’ Windows.” Advanced Energy Materials 8 (12): 1702922.

Spindler, H.C., and L.K. Norford. 2009. “Naturally Ventilated and Mixed-Mode buildings—Part II: Optimal Control.” Building and Environment 44 (4): 750–61.

Spitler, J.D. 2011. “Thermal Load and Energy Performance Prediction.” In Building Performance Simulation for Design and Operation, edited by J.L.M. Hensen and R. Lamberts, 84–142. Spon Press.

Stevens, S. 2001. “Intelligent Facades: Occupant Control and Satisfaction.” International Journal of Sustainable Energy 21 (2): 147–60.

146

Stevenson, F., M. Baborska-Narozny, and P. Chatterton. 2016. “Resilience, Redundancy and Low-Carbon Living: Co-Producing Individual and Community Learning.” Building Research and Information 44 (7): 789–803.

Strachan, P. 2008. “Simulation Support for Performance Assessment of Building Components.” Building and Environment 43 (2): 228–36.

Struck, C., P.J.C.J. de Wilde, C.J. Hopfe, and J.L.M. Hensen. 2009. “An Investigation of the Option Space in Conceptual Building Design for Advanced Building Simulation.” Advanced Engineering Informatics 23 (4): 386–95.

Struck, C., D. Markov, P. Stankov, G. Ilic, M. Serafimov, D. Bionda, and A. Seerig. 2014. “Towards Compensating HVAC System Degradation Phenomena with Adaptable Building Elements.” In Proceedings of SET2014, the 13th International Conference on Sustainable Energy Technologies, 1–8.

Sundell, J., H. Levin, W. Nazaroff, W.S. Cain, W.J. Fisk, D.T. Grimsrud, F. Gyntelberg, et al. 2011. “Ventilation Rates and Health: Multidisciplinary Review of the Scientific Literature.” Indoor Air 21 (3): 191–204.

Tabares-Velasco, P.C., C. Christensen, and M. Bianchi. 2012. “Verification and Validation of EnergyPlus Phase Change Material Model for Opaque Wall Assemblies.” Building and Environment 54 (August): 186–96.

Tabares-Velasco, P.C., and B. Griffith. 2012. “Diagnostic Test Cases for Verifying Surface Heat Transfer Algorithms and Boundary Conditions in Building Energy Simulation Programs.” Journal of Building Performance Simulation 5 (5): 329–46.

Tavares, P.F., A.R. Gaspar, A.G. Martins, and F. Frontini. 2014. “Evaluation of Electrochromic Windows Impact in the Energy Performance of Buildings in Mediterranean Climates.” Energy Policy 67: 68–81.

Thomke, S.H. 1998. “Managing Experimentation in the Design of New Products.” Management Science 44 (6): 743–62.

Tian, W. 2013. “A Review of Sensitivity Analysis Methods in Building Energy Analysis.” Renewable and Sustainable Energy Reviews 20 (April): 411–19.

Toole, T. 2001. “Technological Trajectories of Construction Innovation.” Journal of Architectural Engineering 7 (4): 107–14.

Trčka, M. 2008. “Co-Simulation for Performance Prediction of Innovative Integrated Mechanical Energy Systems in Buildings.” PhD Thesis, Eindhoven University of Technology.

Trčka, M., J.L.M. Hensen, and M. Wetter. 2009. “Co-Simulation of Innovative Integrated HVAC Systems in Buildings.” Journal of Building Performance Simulation 2 (3): 209–30.

Triantaphyllou, E. 2000. Multi-Criteria Decision Making Methods: A Comparative Study. Kluwer Academic Publishers.

TRNSYS. 2017. “TRNSYS v17 - Transient System Simulation Tool.” http://www.trnsys.com/.

Turrin, M., P. von Buelow, and R. Stouffs. 2011. “Design Explorations of Performance Driven Geometry in Architectural Design Using Parametric Modeling and Genetic Algorithms.” Advanced Engineering Informatics 25

147

(4): 656–75.

Ucci, M., and C. W. F. Yu. 2014. “Low-Carbon Buildings, Health and Wellbeing: Current Perspectives and Critical Challenges.” Indoor and Built Environment 23 (3): 335–39.

Ürge-Vorsatz, D., L.F. Cabeza, S. Serrano, C. Barreneche, and K. Petrichenko. 2015. “Heating and Cooling Energy Trends and Drivers in Buildings.” Renewable and Sustainable Energy Reviews 41 (January): 85–98.

US DOE. 2014. Windows and Building Envelope Research and Development: Roadmap for Emerging Technologies.

Viereck, V., Q. Li, A. Ja, and H. Hillmer. 2010. “Design of Self-Assembling Micromirror Arrays for Light Guiding Applications,” Microsystem technologies 16 (5): 895-899.

Ward, G., and R. Shakespeare. 1998. Rendering with Radiance - The Art and Science of Lighting Visualization. Morgan Kaufmann Publishers.

Weaver, J., K.L. Wood, and D. Jensen. 2008. “Transformation Facilitators: A Quantitative Analysis of Reconfigurable Products and Their Characteristics.” In Proceedings of ASME 2008 - International Design Engineering Technical Conferences & Computers and Information in Engineering Conference.

Wetter, M. 2009. “Modelica-Based Modelling and Simulation to Support Research and Development in Building Energy and Control Systems.” Journal of Building Performance Simulation 2 (2): 143–61.

Wetter, M. 2011a. “A View on Future Building System Modeling and Simulation.” In Building Performance Simulation for Design and Operation, edited by J.L.M. Hensen and R. Lamberts, 481–504. Spon Press.

Wetter, M. 2011b. “Co-Simulation of Building Energy and Control Systems with the Building Controls Virtual Test Bed.” Journal of Building Performance Simulation 4 (3): 185–203.

Wetter, M., and C. Haugstetter. 2006. “Modelica versus TRNSYS–A Comparison between an Equation-Based and a Procedural Modeling Language for Building Energy Simulation.” In Proceedings of the 2nd SimBuild Conference.

Wienold, J. 2009. “Dynamic Daylight Glare Evaluation.” In Proceedings of Building Simulation 2009, 944–51.

Wienold, J. 2015. “Evalglare v1.11.”

Wigginton, M., and J. Harris. 2002. Intelligent Skins. Oxford: Butterworth-Heinemann.

Wright, J.A., A.E.I. Brownlee, M.M. Mourshed, and M. Wang. 2014. “Multi-Objective Optimization of Cellular Fenestration by an Evolutionary Algorithm.” Journal of Building Performance Simulation 7 (1): 33–51.

Xiong, J., and A. Tzempelikos. 2016. “Model-Based Shading and Lighting Controls Considering Visual Comfort and Energy Use.” Solar Energy 134: 416–28.

Xu, X., and S. Van Dessel. 2008. “Evaluation of a Prototype Active Building Envelope Window-System.” Energy and Buildings 40 (2): 168–74.

148

Xu, X., S. Dessel, and A. Messac. 2007. “Study of the Performance of Thermoelectric Modules for Use in Active Building Envelopes.” Building and Environment 42 (3): 1489–1502.

Ye, H., Xi. Meng, L. Long, and B. Xu. 2013. “The Route to a Perfect Window.” Renewable Energy 55 (July): 448–55.

Zavala, V. 2012. “Real-Time Resolution of Conflicting Objectives in Building Energy Management: An Utopia-Tracking Approach.” In Proceedings of SimBuild 20102, the Fifth National Conference of IBPSA-USA, 1–8. Madison, WI.

Zavala, V., E. Constantinescu, T. Krause, and M. Anitescu. 2009. “On-Line Economic Optimization of Energy Systems Using Weather Forecast Information.” Journal of Process Control 19 (10): 1725–36.

Zheng, S., Y. Xu, Q. Shen, and H. Yang. 2015. “Preparation of Thermochromic Coatings and Their Energy Saving Analysis.” Solar Energy 112: 263–71.

149

Curriculum Vitae

Roel Loonen (1986) holds a BSc and MSc (cum laude) in Building Services from Eindhoven University of Technology (TU/e). His PhD research – carried out at the Unit Building Physics and Services (TU/e) – developed approaches for computational performance optimization of innovative adaptive façade concepts. Roel spent the summer of 2012 in Philadelphia, US, for a research stay with Pennsylvania State University at DOE’s Energy Efficient Buildings Hub. Roel’s research and teaching focus on the development and application of modeling and simulation strategies to provide decision support for designing buildings that combine high indoor quality with low or no impact on the environment. Roel collaborates with SMEs in various R&D projects, aspiring to accelerate the process of bringing innovative building envelope technologies and renewable energy systems to the market. In addition, he is an active participant in EU COST Action 'Adaptive Facades Network' and IEA SHC Task 56 'Solar Building Envelopes'. Since 2012, he is board member of IBPSA—NVL and has acted as a reviewer for 21 international scientific journals. Roel has received prizes for his Master thesis (REHVA International Student Competition 2011), innovation competitions (Team Energy Challenge 2015) and the Best PhD Supervisor Award of the Department of the Built Environment (2017). From the summer of 2018, Roel continues as Assistant Professor in the Building Performance group at TU/e.

150

151

Acknowledgements

If we want to think like a scientist more often in life, then there are three key

objectives to keep in mind — be humbler about what we know, more confident

about what’s possible, and less afraid of things that don’t matter.” ~ Tim Urban

The list of metaphors that is being used to give outsiders an idea of what it is like to do a PhD is nearly endless, and includes words such as: a symphony, a

knot, a pilgrim's journey, a tree, a school of fish, a battlefield, learning how to drive, building a ship, etc.

The one that resonates most with me lies in the distinction between puzzles and mysteries, as it was put forward by Gregory F. Treverton12:

"There’s a reason millions of people try to solve crossword puzzles each day. Amid the well-ordered combat between a puzzler’s mind and the

blank boxes waiting to be filled, there is satisfaction along with frustration. Even when you can’t find the right answer, you know it exists. Puzzles can

be solved; they have answers. But a mystery offers no such comfort. It poses a question that has no definitive answer because the answer is

contingent; it depends on a future interaction of many factors, known and unknown. A mystery cannot be answered; it can only be framed, by

identifying the critical factors and applying some sense of how they have interacted in the past and might interact in the future. A mystery is an

attempt to define ambiguities."

Looking at it from this perspective, the PhD process can undeniably be

categorized as mystery solving. As we all know, mysteries can sometimes be wicked, while it also ought to be realized that the methods that are used for

puzzles generally don’t work for solving mysteries, or worse, could even

impair the situation.

As such, contingency becomes the main ingredient for the successful completion of a PhD thesis. Contingency can happen in a variety of different

ways, at work or elsewhere, by reading the work of others, and through a range of possible interactions, both planned or in serendipitous situations. In

any case, contingency is not to be confused with randomness. I have

12 https://www.smithsonianmag.com/history/risks-and-riddles-154744750/

152

experienced that the interaction with people is of particular importance

during a PhD, especially when these interactions go beyond the level of what can reasonably be expected. Numerous of such wonderful interactions have

helped me to accomplish the end result that is reported in this thesis. I am very grateful for this.

First and foremost, I would like to express my gratitude to prof.dr.ir. Jan Hensen. Thank you Jan for your constant support, guidance and trust. Your

enthusiasm about how the proper application of building performance simulation can help to make the built environment a better place to live in is

contagious. The way you manage to combine sincere ambition with congeniality, and the way you give others the chance to develop while always

noticing the important things in every project, is appreciated by many. On top of that, I think that the way you always care for the human behind the

researcher is more than commendable.

I am also grateful to dr.dipl.-ing Marija Trčka. Marija, you have been a great

source of inspiration throughout the first phases of my doctorate research. Under your tutelage, the main ideas for the project were developed, many of

which have found their way into this thesis. I also want to thank you for helping to arrange my research stay in Philadelphia.

I would like to thank the other members of the committee for their comments and valuable suggestions to improve this dissertation: prof.dr. Joe Clarke,

prof.dr. Marco Perino, prof.dr.ing. Ulrich Knaack, prof.dr. Albert Schenning, prof.dr.ing. Alex Rosemann and prof.ir. Elphi Nelissen.

The development phase of this project has greatly benefited from close collaboration with dr.ir. Pieter-Jan Hoes. Pieter-Jan, thank you very much for

the shared developments and all other collegial and scientific advice throughout the years.

During my Master studies, and in the early phases of my PhD, I have learned a lot from dr. Daniel Cóstola. Daniel, thanks for sharing your knowledge and

experiences, related to research as well as many other things.

dr. Bruno Lee and dr. Chul-sung Lee also deserve a special acknowledgement

in this section. Bruno and Chul-sung, thanks a lot for helping to create the friendly environment during the years that we have shared the same office

space.

Further, I would like to thank Ignacio for sharing your knowledge, wittiness

and your couch; Isabella for challenging me with a continuous stream of creative ideas; Rebeca and Petr for laughs, reflection and always making me

look forward to the next rendezvous; Rajesh and Vojta for the many merry

153

moments while you were competently overtaking me; and Hamid for

generous advice, support and simply for being a very good friend.

Having had the fortunate possibility to be co-supervisor of various PhD

candidates and PDEng trainees might not have expedited the completion of this thesis, but has undoubtedly enriched my research and personal

development in countless different ways. I would like to thank Finn, Mohammad, Marie, Hemshikha, Adam, Samuel, Antía and Dmitry for your

enthusiasm and the pleasant collaboration.

Apart from providing an excellent research environment, the Building Physics

and Services Unit at TU/e stimulates an atmosphere that blurs – if not vanishes – the line between colleagues and friends. I consider it as a great

bonus to have the privilege to be part of this cohesive group. In addition to those mentioned above, I would like to thank for all the support, useful

distractions and cultural learning: Adam, Adelya, Alessia, Ana Paula, Asit, Benedetto, Carlos, Christina, Christof, Christos, Erkki, Evangelos, Fotis,

Gerardo, Giovanni, Johann, John, Katarina, Luyi, Marcel, Maryam, Massimo, Mike, Mohamed, Mohammad, Olga, Parisa, Qimiao, Raúl, Rizki, Rongling,

Rubina, Sanja, Sanket, Tomaž, Wiebe, Yasin, and Zahra.

The work described in this thesis has, directly and indirectly, benefitted from

many inspiring discussions with the master thesis students I have worked with. Sundar, Alexander, Babis, Michalis, Charlotte, Daan, Nikos, Bas,

Franziska, João, Riccardo, Wouter, Daniela, Prahaladh, Stefan, Remco, Teun, Adheesh, Canan, David, Helena, Niels, Puji, Suryadi and Vignesh, thank you

all.

Without the cooperation of many people outside the BPS Unit at TU/e, who

have supported me in this research and thesis, it would have never been possible to achieve this end result.

This research was financially supported by RVO through the EOS-lt project: FACET. I would like to thank Bart de Boer from TNO for the skillful project

management, and all project members for all the valuable input and insightful suggestions for my research.

I am very fortunate to have met dr. Fabio Favoino, as an aspiring researcher who is on a similar mission to materialize the benefits of adaptive facades

through increased use of building performance simulation. Fabio, it was a great pleasure to work together and to be able to discuss research ideas and

everyday challenges, without the hassle of needing to explain all the context. I am really happy to see you back in academia, and look forward to continuing

the collaboration. I am also grateful to dr. Mauro Overend for the support and constructive feedback.

154

I am thankful to dr. Hitesh Khandelwal, dr. Michael Debije and prof.dr. Albert

Schenning for introducing me to the wonders of stimuli-responsive functional materials and devices. You have provided me with a great number

of ideas and I appreciate that you were always there to answer my questions.

I would further like to thank the project members of COST Action TU1403

‘Adaptive Facades Network’, IEA SHC Task 56 ‘Solar Building Envelopes’ and the 4TU.Lighthouse initiatives for feedback, support and providing an

international perspective.

Duncan Harkness, thanks a lot for the much-needed support during the last

phases of the project.

Apart from the contacts in academia, this work has also significantly

benefited from information exchange with industrial partners. I would like to thank Teun Wagenaar, Jasper van den Muijsenberg and dr. Casper van Oosten

from Merck Window Technologies and Sam Kin, Stan de Ridder and Matthijs Damen from Kindow / Wellsun for giving me the much-appreciated reality-

checks and for letting me have a first-hand insight into what it takes to really bring an innovative façade concept from initial idea to the market.

My dear friends from Lottum, thanks for being there during the day of the defense, for being curious about my project, at times, but most notably for

always reminding me that the most important things in life happen outside the university.

Finally, a special word of thanks to my parents and to Ingrid and Remon. I don’t express it often enough, but I really appreciate the outstanding home

base and your unquestioned support.

155

List of Publications

Journal articles

1. Capperucci, R., Loonen, R.C.G.M., Hensen, J.L.M. & Rosemann, A.L.P. (2018). Angle-dependent optical properties of advanced fenestration systems - finding a right balance between model complexity and prediction error. Building Simulation. Article in Press.

2. Koenders, S.J.M., Loonen, R.C.G.M. & Hensen, J.L.M. (2018). Investigating the potential of a closed-loop dynamic insulation system for opaque building elements. Energy and Buildings, 173, 409-427.

3. Sarakinioti, M.V., Konstantinou, T., Turrin, M., Tenpierik, M., Loonen, R.C.G.M., de Klijn - Chevalerias, M.L. & Knaack. U. (2018). Developing an integrated 3D-printed façade for thermal regulation in complex geometries. Journal of Façade Design and Engineering, 6(2): 29-40.

4. Roberz, F., Loonen, R.C.G.M., Hoes P. & Hensen, J.L.M. (2017). Ultra-lightweight concrete: energy and comfort performance evaluation in relation to buildings with low and high thermal mass. Energy and Buildings, 138, 432-442.

5. De Witte, D., de Klijn – Chevalerias, M.L., Loonen, R.C.G.M., Hensen, J.L.M., Knaack, U. & Zimmermann, G. (2017). Convective Concrete: Additive Manufacturing to facilitate activation of thermal mass. Journal of Façade Design and Engineering 5(1): 107-117.

6. Loonen, R.C.G.M., Favoino, F., Hensen, J.L.M. & Overend, M. (2017). Review of current status, requirements and opportunities for building performance simulation of adaptive facades. Journal of Building Performance Simulation, 2017(2): 205-223. (most read article in Journal of Building Performance Simulation since 2011)

7. Rezaeian, M., Montazeri, H. & Loonen, R.C.G.M. (2017). Science foresight using life-cycle analysis, text mining and clustering: a case study on natural ventilation. Technological Forecasting and Social Change, 118, 270-280.

8. De Witte, D., Knaack, U., de Klijn - Chevalerias, M.L., Loonen, R.C.G.M., Hensen, J.L.M. & Zimmermann, G. (2017). Convective Concrete: additive manufacturing to facilitate activation of thermal mass. Spool 4(2): 13-16.

9. Sarakinioti, V., Turrin, M., Teeling, M., De Ruiter, P., Van Erk, M., Tenpierik, M., Konstantinou, T., Knaack, U., Pronk, A.D.C., Teuffel, P.M., van Lier, A., Vorstermans, R., Dolkemade, E., de Klijn - Chevalerias, M.L., Loonen, R.C.G.M., Hensen, J.L.M. & Vlasblom, D. (2017). Spong3d: 3D printed facade system enabling movable fluid heat storage. Spool. 4(2):57-60.

10. Khandelwal, H., Loonen, R.C.G.M., Hensen, J.L.M., Debije, M.G., & Schenning, A.P.H.J. (2015). Electrically switchable polymer stabilized broadband infrared reflectors and their potential as smart windows for energy saving in buildings. Nature Scientific Reports 5, 11773.

11. Bakker, L.G., Hoes - Van Oeffelen, E.C.M., Loonen, R.C.G.M. & Hensen, J.L.M. (2014). User satisfaction and interaction with automated dynamic facades: a pilot study. Building and Environment, 78, 44-52.

12. Kasinalis, C., Loonen, R.C.G.M., Cóstola, D. & Hensen, J.L.M. (2014). Framework for assessing the performance potential of seasonally adaptable facades using multi-objective optimization. Energy and Buildings, 79, 106-113.

156

13. Khandelwal, H., Loonen, R.C.G.M., Hensen, J.L.M., Schenning, A.P.H.J. & Debije, M.G. (2014). Application of broadband infrared reflector based on cholesteric liquid crystal polymer network to windows and its impact on reducing the energy consumption in buildings. Journal of Materials Chemistry A, 2, 14622-14627.

14. Loonen, R.C.G.M., Singaravel, S., Trčka, M., Cóstola, D. & Hensen, J.L.M. (2014). Simulation-based support for product development of innovative building envelope components. Automation in Construction, 45, 86-95.

15. Loonen, R.C.G.M., Trčka, M., Cóstola, D. & Hensen, J.L.M. (2013). Climate adaptive building shells: state-of-the-art and future challenges. Renewable & Sustainable Energy Reviews, 25, 483-493.

16. Franchimon, F., Loonen, R.C.G.M. & Bronswijk, J.E.M.H. van (2008). Towards an economically acceptable prevention of Legionnaire's disease. Gerontechnology, 7(2): 107-107.

Book chapters

1. Aelenei, D., Aelenei, L., Loonen, R.C.G.M., Perino, M. & Serra, V. (2018). Adaptive Facades. In Desideri, U. & Asdrubali, F. (Eds.), Handbook of Energy Efficiency in Buildings. Elsevier.

2. Loonen, R.C.G.M. (2015). Bio-inspired adaptive building skins. In F. Pacheco Torgal, J.A. Labrincha, M.V. Diamanti, C.P. Yu & H.K. Lee (Eds.), Biotechnologies and Biomimetics for Civil Engineering (pp. 115-134). Springer International Publishing.

Conference Contributions

1. Bognar, A., Loonen, R.C.G.M., Valckenborg, R. & Hensen, J.L.M. (2018). Modeling Reflected Irradiance in Urban Environments – A Case Study for Simulation-Based Measurement Quality Control for an Outdoor PV Test Site. Conference Paper: Proceedings of EUPVSEC 2018 – European PV Solar Energy Conference and Exhibition. Brussels, Belgium.

2. Tzikas, C., Valckenborg, R., van den Donker, M., Bognar, A., Duque Lozano, D., Loonen, R.C.G.M., Hensen, J.L.M. & Folkerts, W. (2018). Outdoor characterization of colored and textured prototype PV façade elements. Conference Paper: Proceedings of EUPVSEC 2018 – European PV Solar Energy Conference and Exhibition. Brussels, Belgium.

3. Bianco, L., Loonen, R.C.G.M., Koenders, S.J.M., Avesani, S., Vullo, P., Bejat, T., Goia, F., Cascone, Y., Serra, V. & Favoino, F. (2018). Towards new metrics for the characterization of the dynamic performance of adaptive façade systems. Conference Paper: Proceedings of FACADE2018 – final conference of COST TU1403 “Adaptive Facades Network”. Lucerne. Switzerland.

4. De Michele, G., Loonen, R.C.G.M., Saini, H., Favoino, F., Avesani, S., Papaiz, L. & Gasparella, A. (2018). Opportunities and challenges for performance prediction of dynamic Complex Fenestration Systems (CFS). Conference Paper: Proceedings of FACADE2018 – final conference of COST TU1403 “Adaptive Facades Network”. Lucerne. Switzerland.

5. Juaristi, M., Loonen, R.C.G.M., Monge-Barrio, A. & Gómez-Acebo, T. (2018). Dynamic Analysis of Climatic Conditions for deriving suitable Adaptive Façade responses. Conference Paper: Proceedings of FACADE2018 – final conference of COST TU1403 “Adaptive Facades Network”. Lucerne. Switzerland.

157

6. Luna-Navarro, A., Loonen, R.C.G.M., Attia, S., Juaristi, M., Monge-Barrio, A., Rabensheifer, R., Donato, M., & Overend, M. (2018). Occupant-Adaptive Façade Interaction: Relationships and Conflicts. Conference Paper: Proceedings of FACADE2018 – final conference of COST TU1403 “Adaptive Facades Network”. Lucerne. Switzerland.

7. Meskinkiya, M., Loonen, R.C.G.M., Paolini, R. & Hensen, J.L.M. (2018). Calibrating Perez Model Coefficients Using Subset Simulation. Conference Paper: Proceedings of the 9th international SOLARIS Conference. Chengdu, China.

8. Saini, H., Loonen, R.C.G.M. & Hensen, J.L.M. (2018). Simulation-based performance prediction of an energy-harvesting façade system with selective daylight transmission. Conference Paper: ICAE2018 – International Congress on Architectural Envelopes, San Sebastian, Spain.

9. de Klijn, M.L., Loonen, R.C.G.M., Zarzycka, A., de Witte, D., Sarakinioti, M.V. & Hensen, J.L.M. (2017). Assisting the development of innovative responsive façade elements using building performance simulation. Conference Paper: Proceedings of SimAUD 2017, Toronto, Canada. pp:243-250.

10. Favoino, F., Giovannini, L. & Loonen, R.C.G.M. (2017). Smart Glazing in Intelligent Buildings: What Can We Simulate? Conference Paper: Proceedings of Glass Performance Days 2017, Tampere Finland, pp:212-219

11. de Witte, D., de Klijn, M.L., Loonen, R.C.G.M., Hensen, J.L.M., Knaack, U. & Zimmermann, G. (2017). Convective concrete: additive manufacturing to facilitate activation of thermal mass. Conference Paper: Proceedings of POWERSKIN, Munich, Germany.

12. Luna-Navarro, A., Loonen, R.C.G.M., Attia, S., Juaristi, M., Bilir S., Favoino, F., Monteiro Silva, S., Mateus, R., Almeida, M., Petrovski, A. & Overend, M. (2017). A roadmap for capturing user-adaptive façade interaction. Conference Paper: Proceedings of NEXT-Facades 2017: mid-term conference of COST Action TU 1403 Adaptive Facades Network, Munich, Germany. pp: 84-85

13. Bognár, Á., Loonen, R.C.G.M., Valckenborg, R.M.E. & Hensen, J.L.M. (2017). Identifying local PV shading in urban areas based on AC power and regional irradiance measurement. Conference Poster: Sunday conference.

14. Surugin, D., Loonen, R.C.G.M., Valckenborg, R.M.E. & Hensen, J.L.M. (2017). Performance prediction of BIPV: approach for development of a model selection guideline. Conference Poster: Sunday conference.

15. Ghasempourabadi, M., Sinapis. K., Loonen, R.C.G.M., Valckenborg, R.M.E., Hensen, J.L.M. & Folkerts, W. (2016). Towards simulation-assisted performance monitoring of BIPV systems considering shading effects. Conference Paper: Proceedings of IEEE PVSC 43, 5-10 June 2016, Portland, Oregon, US, (pp. 6). IEEE.

16. Loonen, R.C.G.M. & Hensen, J.L.M. (2015). Smart windows with dynamic spectral selectivity : a scoping study. Conference Paper: Proceedings Building Simulation '15, 7-9 December 2015, Hyderabad, India, (pp. 8). IBPSA.

17. Loonen, R.C.G.M., Loomans, M.G.L.C. & Hensen, J.L.M. (2015). Towards predicting the satisfaction with indoor environmental quality in building performance simulation. Conference Paper: Proceedings of Healthy Buildings Europe 2015: 18-20 May 2015, Eindhoven, The Netherlands, (pp. 1-8). ISIAQ International.

18. Loonen, R.C.G.M., Rico-Martinez, J.M., Favoino, F., Brzezicki, M., Menezo, C., La Ferla, G., & Aelenei, L. (2015). Design for facade adaptability - Towards a unified and systematic characterization. In Proceedings of the 10th Energy Forum - Advanced Building Skins. Bern, Switzerland. pp: 1274-1284

19. Attia, S., Favoino, F., Loonen, R.C.G.M., Petrovski, A. & Monge-Barrio, A. (2015). Adaptive façades system assessment: An initial review. In Proceedings of the 10th Energy Forum - Advanced Building Skins. Bern, Switzerland. pp: 1265-1273

158

20. Loonen, R.C.G.M. (2014). Shaping the next generation of adaptive facade concepts with the use of simulations. In A. Luible & U Zihlmann (Eds.), Conference Paper: Facade2014 - Conference on Building Envelopes, (pp. 84-95). Lucerne, Switzerland. Invited paper.

21. Hensen, J.L.M., Archontiki, M., Cóstola, D. & Loonen, R.C.G.M. (2014). Simulation support for research and development of advanced building skin concepts. Conference Paper: Proceedings of the 9th Energy Forum on Advanced Building Skins, 28-29 October 2014, Bressanone, Italy, (pp. 1091-1107).

22. Khandelwal, H., Roberz, F., Loonen, R.C.G.M., Hensen, J.L.M., Bastiaansen, C.W.M., Broer, D.J., Debije, M.G. & Schenning, A.P.H.J. (2014). Infrared reflector based on liquid crystal polymers and its impact on thermal comfort conditions in buildings. In I. C. Khoo (Ed.), Conference Paper: Liquid Crystals XVIII, August 17, 2014, San Diego, USA, (Proceedings of SPIE, 9182, pp. 91820S-1-91820S-8). Bellingham: SPIE.

23. Loonen, R.C.G.M., Hoes, P. & Hensen, J.L.M. (2014). Performance prediction of buildings with responsive building elements. Proceeding of IBPSA-England conference Building Simulation and Optimization (BSO14) 23-34 June 2014, London, UK.

24. Loonen, R.C.G.M. & Hensen, J.L.M. (2014). Considerations regarding optimization of dynamic facades for improved energy performance and visual comfort. Proceedings of COLEB2014, Computational optimization of low-energy buildings, 6-7 March, Zurich, Switzerland.

25. Lee, C., Cóstola, D., Loonen, R.C.G.M. & Hensen, J.L.M. (2013). Energy saving potential of long-term climate adaptive greenhouse shells. Conference Paper: Proceedings of BS2013: 13th Conference of International Building Performance Simulation Association, 26-28 August 2013, Chambery, France, (pp. 954-961).

26. Loonen, R.C.G.M. & Hensen, J.L.M. (2013). Dynamic sensitivity analysis for performance-based building design and operation. Conference Paper: Proceedings of BS2013: 13th Conference of International Building Performance Simulation Association, 26-28 August 2013, Chambery, France, (pp. 299-305).

27. Boer, B.J. de, Bakker, L., Oeffelen, E.C.M. van, Loonen, R.C.G.M., Cóstola, D. & Hensen, J.L.M. (2012). Future climate adaptive building shells - Optimizing energy and comfort by inverse modeling. Oral : Proceedings of the 8th Energy Forum on Solar Building Skins, 6-7 December 2012, Bressanone, Italy, (pp. 15-19). Bressanone.

28. Hoes, P., Loonen, R.C.G.M., Trčka, M. & Hensen, J.L.M. (2012). Performance prediction of advanced building controls in the design phase using ESP-r, BCVTB and Matlab. Oral : Paper presented at the First IBPSA-England conference Building Simulation and Optimization(BSO12) 10-11 September 2012, Loughborough, UK, (pp. 229-236). Loughborough, England.

29. Boer, B.J. de, Ruijg, G.J., Bakker, L.G., Kornaat, W., Zonneveldt, L., Kurvers, S.R., Alders, E.E., Raue, A., Hensen, J.L.M., Loonen, R.C.G.M. & Trčka, M. (2011). Energy saving potential of climate adaptive building shells - Inverse modelling of optimal thermal and visual behaviour. Adaptive Architecture, an international conference, London.

30. Boer, B.J. de, Ruijg, G.J., Loonen, R.C.G.M., Trčka, M., Hensen, J.L.M. & Kornaat, W. (2011). Climate adaptive building shells for the future – optimization with an inverse modelling approach. Proceedings ECEEE Summer Study 2011, Belambra Presqu'île de Giens, France, June 2011, (pp. 1413-1422). Presqu'île de Giens.

31. Loonen, R.C.G.M., Trčka, M. & Hensen, J.L.M. (2011). Exploring the potential of climate adaptive building shells. Conference Paper : Proceedings of Building Simulation 2011: 12th Conference of International Building Performance Simulation Association, Sydney, Australia, November 2011, (pp. 2148-2155). Sydney, Australia.

159

32. Trčka, M., Loonen, R.C.G.M., Hensen, J.L.M. & Houben, J.V.F. (2011). Computational building performance simulation for integrated design and product optimization. Conference Paper : Proceedings of Building Simulation 2011: 12th Conference of International Building Performance Simulation Association, Sydney, Australia, November 2011, (pp. 302-309). Sydney, Australia.

33. Loonen, R.C.G.M., Cóstola, D., Trčka, M. & Hensen, J.L.M. (2010). Prestatiesimulatie van adaptieve gevels - Smart Energy Glass als case-study. Proceedings IBPSA-NVL 2010 Event, Eindhoven, oktober 2010, (pp. 1-8 [Paper ID: 15]). Eindhoven: IBPSA-NVL.

34. Loonen, R.C.G.M., Trčka, M., Cóstola, D. & Hensen, J.L.M. (2010). Performance simulation of climate adaptive building shells - Smart Energy Glass as a case study. In V. Lemort, P. André & S. Bertagnolio (Eds.), Proceedings 8th International conference on System Simulation in Buildings, (pp. 1-19). Liège.

Professional journals and magazines

1. Roberz, F., Loonen, R.C.G.M., Hoes P. & Hensen, J.L.M. (2016). Ultra-lichtgewicht beton: bouwfysische alleskunner of simpelweg een ander type beton?. Bouwfysica, 27(3), 10-14. Best paper award.

2. Hensen, J.L.M., Loonen, R.C.G.M., Archontiki, M. & Kanellis, M. (2015). Using building simulation for moving innovations across the "Valley of Death". REHVA Journal, 52(3), 58-62.

3. Loonen, R.C.G.M., de Boer, B.J., Kornaat, W., Bakker, L.G. & Hensen, J.L.M. (2015). Zoektocht naar de ideale adaptieve gevel. TVVL Magazine, 44(7-8), 36-38.

4. Hensen, J.L.M., Loonen, R.C.G.M., Archontiki, M. & Kanellis, M. (2014). Modelirovaniye zdaniy dlya prodvizheniya innovatsiy cherez « dolinu smerti » [Modeling of buildings to promote innovation through the "valley of death"]. Zdaniâ vysokih tehnologij, 2014 (Autumn), 72-79.

5. Loonen, R.C.G.M. & Hensen, J.L.M. (2014). Bioadaptivnaya obolochka zdaniya [Bio-inspired adaptive facades]. Zdaniâ vysokih tehnologij, 2014 (Summer), 50-57.

6. Berk, A.B.M., Rutten, P.G.S., Loomans, M.G.L.C., Aarts, M.P.J. & Loonen, R.C.G.M. (2013). Gelijktijdig berekenen en beoordelen van gevelprestaties. TVVL Magazine, 42(1), 16-20.

7. Loonen, R.C.G.M., Trčka, M. & Hensen, J.L.M. (2012). Adaptieve gevels: een verkenning van het potentieel. TVVL Magazine, 41(5), 34-36.

Bouwstenen is een publicatiereeksvan de Faculteit Bouwkunde,Technische Universiteit Eindhoven.Zij presenteert resultaten vanonderzoek en andere activiteiten ophet vakgebied der Bouwkunde,uitgevoerd in het kader van dezeFaculteit.

Bouwstenen en andere proefschriften van de TU/e zijn online beschikbaar via: https://research.tue.nl/

KernredactieMTOZ

Reeds verschenen in de serieBouwstenen

nr 1Elan: A Computer Model for Building Energy Design: Theory and ValidationMartin H. de WitH.H. DriessenR.M.M. van der Velden

nr 2Kwaliteit, Keuzevrijheid en Kosten: Evaluatie van Experiment Klarendal, ArnhemJ. SmeetsC. le NobelM. Broos J. FrenkenA. v.d. Sanden

nr 3Crooswijk: Van ‘Bijzonder’ naar ‘Gewoon’Vincent SmitKees Noort

nr 4Staal in de WoningbouwEdwin J.F. Delsing

nr 5Mathematical Theory of Stressed Skin Action in Profiled Sheeting with Various Edge ConditionsAndre W.A.M.J. van den Bogaard

nr 6Hoe Berekenbaar en Betrouwbaar is de Coëfficiënt k in x-ksigma en x-ks? K.B. LubA.J. Bosch

nr 7Het Typologisch Gereedschap: Een Verkennende Studie Omtrent Typologie en Omtrent de Aanpak van Typologisch OnderzoekJ.H. Luiten nr 8Informatievoorziening en BeheerprocessenA. NautaJos Smeets (red.)Helga Fassbinder (projectleider)Adrie ProveniersJ. v.d. Moosdijk

nr 9Strukturering en Verwerking van Tijdgegevens voor de Uitvoering van Bouwwerkenir. W.F. SchaeferP.A. Erkelens

nr 10Stedebouw en de Vorming van een Speciale WetenschapK. Doevendans

nr 11Informatica en Ondersteuning van Ruimtelijke BesluitvormingG.G. van der Meulen

nr 12Staal in de Woningbouw, Korrosie-Bescherming van de Begane GrondvloerEdwin J.F. Delsing

nr 13Een Thermisch Model voor de Berekening van Staalplaatbetonvloeren onder BrandomstandighedenA.F. Hamerlinck

nr 14De Wijkgedachte in Nederland: Gemeenschapsstreven in een Stedebouwkundige ContextK. DoevendansR. Stolzenburg

nr 15Diaphragm Effect of Trapezoidally Profiled Steel Sheets: Experimental Research into the Influence of Force ApplicationAndre W.A.M.J. van den Bogaard

nr 16Versterken met Spuit-Ferrocement: Het Mechanische Gedrag van met Spuit-Ferrocement Versterkte Gewapend BetonbalkenK.B. LubirM.C.G. van Wanroy

nr 17De Tractaten van Jean Nicolas Louis DurandG. van Zeyl

nr 18Wonen onder een Plat Dak: Drie Opstellen over Enkele Vooronderstellingen van de StedebouwK. Doevendans

nr 19Supporting Decision Making Processes: A Graphical and Interactive Analysis of Multivariate DataW. Adams

nr 20Self-Help Building Productivity: A Method for Improving House Building by Low-Income Groups Applied to Kenya 1990-2000P. A. Erkelens

nr 21De Verdeling van Woningen: Een Kwestie van OnderhandelenVincent Smit

nr 22Flexibiliteit en Kosten in het Ontwerpproces: Een Besluitvormingondersteunend ModelM. Prins

nr 23Spontane Nederzettingen Begeleid: Voorwaarden en Criteria in Sri LankaPo Hin Thung

nr 24Fundamentals of the Design of Bamboo StructuresOscar Arce-Villalobos

nr 25Concepten van de BouwkundeM.F.Th. Bax (red.)H.M.G.J. Trum (red.)

nr 26Meaning of the SiteXiaodong Li

nr 27Het Woonmilieu op Begrip Gebracht: Een Speurtocht naar de Betekenis van het Begrip 'Woonmilieu'Jaap Ketelaar

nr 28Urban Environment in Developing Countrieseditors: Peter A. Erkelens George G. van der Meulen (red.)

nr 29Stategische Plannen voor de Stad: Onderzoek en Planning in Drie Stedenprof.dr. H. Fassbinder (red.)H. Rikhof (red.)

nr 30Stedebouwkunde en StadsbestuurPiet Beekman

nr 31De Architectuur van Djenné: Een Onderzoek naar de Historische StadP.C.M. Maas

nr 32Conjoint Experiments and Retail PlanningHarmen Oppewal

nr 33Strukturformen Indonesischer Bautechnik: Entwicklung Methodischer Grundlagen für eine ‘Konstruktive Pattern Language’ in IndonesienHeinz Frick arch. SIA

nr 34Styles of Architectural Designing: Empirical Research on Working Styles and Personality DispositionsAnton P.M. van Bakel

nr 35Conjoint Choice Models for Urban Tourism Planning and MarketingBenedict Dellaert

nr 36Stedelijke Planvorming als Co-ProduktieHelga Fassbinder (red.)

nr 37 Design Research in the Netherlandseditors: R.M. Oxman M.F.Th. Bax H.H. Achten

nr 38 Communication in the Building IndustryBauke de Vries

nr 39 Optimaal Dimensioneren van Gelaste PlaatliggersJ.B.W. StarkF. van PeltL.F.M. van GorpB.W.E.M. van Hove

nr 40 Huisvesting en Overwinning van ArmoedeP.H. Thung P. Beekman (red.)

nr 41 Urban Habitat: The Environment of TomorrowGeorge G. van der Meulen Peter A. Erkelens

nr 42A Typology of JointsJohn C.M. Olie

nr 43Modeling Constraints-Based Choices for Leisure Mobility PlanningMarcus P. Stemerding

nr 44Activity-Based Travel Demand ModelingDick Ettema

nr 45Wind-Induced Pressure Fluctuations on Building FacadesChris Geurts

nr 46Generic RepresentationsHenri Achten

nr 47Johann Santini Aichel: Architectuur en AmbiguiteitDirk De Meyer

nr 48Concrete Behaviour in Multiaxial CompressionErik van Geel

nr 49Modelling Site SelectionFrank Witlox

nr 50Ecolemma ModelFerdinand Beetstra

nr 51Conjoint Approaches to Developing Activity-Based ModelsDonggen Wang

nr 52On the Effectiveness of VentilationAd Roos

nr 53Conjoint Modeling Approaches for Residential Group preferencesEric Molin

nr 54Modelling Architectural Design Information by FeaturesJos van Leeuwen

nr 55A Spatial Decision Support System for the Planning of Retail and Service FacilitiesTheo Arentze

nr 56Integrated Lighting System AssistantEllie de Groot

nr 57Ontwerpend Leren, Leren OntwerpenJ.T. Boekholt

nr 58Temporal Aspects of Theme Park Choice BehaviorAstrid Kemperman

nr 59Ontwerp van een Geïndustrialiseerde FunderingswijzeFaas Moonen

nr 60Merlin: A Decision Support System for Outdoor Leisure PlanningManon van Middelkoop

nr 61The Aura of ModernityJos Bosman

nr 62Urban Form and Activity-Travel PatternsDaniëlle Snellen

nr 63Design Research in the Netherlands 2000Henri Achten

nr 64Computer Aided Dimensional Control in Building ConstructionRui Wu

nr 65Beyond Sustainable Buildingeditors: Peter A. Erkelens Sander de Jonge August A.M. van Vlietco-editor: Ruth J.G. Verhagen

nr 66Das Globalrecyclingfähige HausHans Löfflad

nr 67Cool Schools for Hot SuburbsRené J. Dierkx

nr 68A Bamboo Building Design Decision Support ToolFitri Mardjono

nr 69Driving Rain on Building EnvelopesFabien van Mook

nr 70Heating Monumental ChurchesHenk Schellen

nr 71Van Woningverhuurder naar Aanbieder van WoongenotPatrick Dogge

nr 72Moisture Transfer Properties of Coated GypsumEmile Goossens

nr 73Plybamboo Wall-Panels for HousingGuillermo E. González-Beltrán

nr 74The Future Site-ProceedingsGer MaasFrans van Gassel

nr 75Radon transport in Autoclaved Aerated ConcreteMichel van der Pal

nr 76The Reliability and Validity of Interactive Virtual Reality Computer ExperimentsAmy Tan

nr 77Measuring Housing Preferences Using Virtual Reality and Belief NetworksMaciej A. Orzechowski

nr 78Computational Representations of Words and Associations in Architectural DesignNicole Segers

nr 79Measuring and Predicting Adaptation in Multidimensional Activity-Travel PatternsChang-Hyeon Joh

nr 80Strategic BriefingFayez Al Hassan

nr 81Well Being in HospitalsSimona Di Cicco

nr 82Solares Bauen:Implementierungs- und Umsetzungs-Aspekte in der Hochschulausbildung in ÖsterreichGerhard Schuster

nr 83Supporting Strategic Design of Workplace Environments with Case-Based ReasoningShauna Mallory-Hill

nr 84ACCEL: A Tool for Supporting Concept Generation in the Early Design Phase Maxim Ivashkov

nr 85Brick-Mortar Interaction in Masonry under CompressionAd Vermeltfoort

nr 86 Zelfredzaam WonenGuus van Vliet

nr 87Een Ensemble met Grootstedelijke AllureJos BosmanHans Schippers

nr 88On the Computation of Well-Structured Graphic Representations in Architectural Design Henri Achten

nr 89De Evolutie van een West-Afrikaanse Vernaculaire ArchitectuurWolf Schijns

nr 90ROMBO TactiekChristoph Maria Ravesloot

nr 91External Coupling between Building Energy Simulation and Computational Fluid DynamicsEry Djunaedy

nr 92Design Research in the Netherlands 2005editors: Henri Achten Kees Dorst Pieter Jan Stappers Bauke de Vries

nr 93Ein Modell zur Baulichen TransformationJalil H. Saber Zaimian

nr 94Human Lighting Demands: Healthy Lighting in an Office EnvironmentMyriam Aries

nr 95A Spatial Decision Support System for the Provision and Monitoring of Urban GreenspaceClaudia Pelizaro

nr 96Leren CreërenAdri Proveniers

nr 97SimlandscapeRob de Waard

nr 98Design Team CommunicationAd den Otter

nr 99Humaan-Ecologisch Georiënteerde WoningbouwJuri Czabanowski

nr 100HambaseMartin de Wit

nr 101Sound Transmission through Pipe Systems and into Building StructuresSusanne Bron-van der Jagt

nr 102Het Bouwkundig ContrapuntJan Francis Boelen

nr 103A Framework for a Multi-Agent Planning Support SystemDick Saarloos

nr 104Bracing Steel Frames with Calcium Silicate Element WallsBright Mweene Ng’andu

nr 105Naar een Nieuwe HoutskeletbouwF.N.G. De Medts

nr 106 and 107Niet gepubliceerd

nr 108GeborgenheidT.E.L. van Pinxteren

nr 109Modelling Strategic Behaviour in Anticipation of CongestionQi Han

nr 110Reflecties op het WoondomeinFred Sanders

nr 111On Assessment of Wind Comfort by Sand ErosionGábor Dezsö

nr 112Bench Heating in Monumental Churches Dionne Limpens-Neilen

nr 113RE. ArchitectureAna Pereira Roders

nr 114Toward Applicable Green ArchitectureUsama El Fiky

nr 115Knowledge Representation under Inherent Uncertainty in a Multi-Agent System for Land Use PlanningLiying Ma

nr 116Integrated Heat Air and Moisture Modeling and SimulationJos van Schijndel

nr 117Concrete Behaviour in Multiaxial CompressionJ.P.W. Bongers

nr 118The Image of the Urban LandscapeAna Moya Pellitero

nr 119The Self-Organizing City in VietnamStephanie Geertman

nr 120A Multi-Agent Planning Support System for Assessing Externalities of Urban Form ScenariosRachel Katoshevski-Cavari

nr 121Den Schulbau Neu Denken, Fühlen und WollenUrs Christian Maurer-Dietrich

nr 122Peter Eisenman Theories and PracticesBernhard Kormoss

nr 123User Simulation of Space UtilisationVincent Tabak

nr 125In Search of a Complex System ModelOswald Devisch

nr 126Lighting at Work:Environmental Study of Direct Effects of Lighting Level and Spectrum onPsycho-Physiological VariablesGrazyna Górnicka

nr 127Flanking Sound Transmission through Lightweight Framed Double Leaf WallsStefan Schoenwald

nr 128Bounded Rationality and Spatio-Temporal Pedestrian Shopping BehaviorWei Zhu

nr 129Travel Information:Impact on Activity Travel PatternZhongwei Sun

nr 130Co-Simulation for Performance Prediction of Innovative Integrated Mechanical Energy Systems in BuildingsMarija Trcka

nr 131Niet gepubliceerd

˙

�

nr 132Architectural Cue Model in Evacuation Simulation for Underground Space DesignChengyu Sun

nr 133Uncertainty and Sensitivity Analysis in Building Performance Simulation for Decision Support and Design OptimizationChristina Hopfe

nr 134Facilitating Distributed Collaboration in the AEC/FM Sector Using Semantic Web TechnologiesJacob Beetz

nr 135Circumferentially Adhesive Bonded Glass Panes for Bracing Steel Frame in FaçadesEdwin Huveners

nr 136Influence of Temperature on Concrete Beams Strengthened in Flexure with CFRPErnst-Lucas Klamer

nr 137Sturen op KlantwaardeJos Smeets

nr 139Lateral Behavior of Steel Frames with Discretely Connected Precast Concrete Infill PanelsPaul Teewen

nr 140Integral Design Method in the Context of Sustainable Building DesignPerica Savanovic

nr 141Household Activity-Travel Behavior: Implementation of Within-Household InteractionsRenni Anggraini

nr 142Design Research in the Netherlands 2010Henri Achten

nr 143Modelling Life Trajectories and Transport Mode Choice Using Bayesian Belief NetworksMarloes Verhoeven

nr 144Assessing Construction Project Performance in GhanaWilliam Gyadu-Asiedu

nr 145Empowering Seniors through Domotic HomesMasi Mohammadi

nr 146An Integral Design Concept forEcological Self-Compacting ConcreteMartin Hunger

nr 147Governing Multi-Actor Decision Processes in Dutch Industrial Area RedevelopmentErik Blokhuis

nr 148A Multifunctional Design Approach for Sustainable ConcreteGötz Hüsken

nr 149Quality Monitoring in Infrastructural Design-Build ProjectsRuben Favié

nr 150Assessment Matrix for Conservation of Valuable Timber StructuresMichael Abels

nr 151Co-simulation of Building Energy Simulation and Computational Fluid Dynamics for Whole-Building Heat, Air and Moisture EngineeringMohammad Mirsadeghi

nr 152External Coupling of Building Energy Simulation and Building Element Heat, Air and Moisture SimulationDaniel Cóstola

´

nr 153Adaptive Decision Making In Multi-Stakeholder Retail Planning Ingrid Janssen

nr 154Landscape GeneratorKymo Slager

nr 155Constraint Specification in ArchitectureRemco Niemeijer

nr 156A Need-Based Approach to Dynamic Activity GenerationLinda Nijland

nr 157Modeling Office Firm Dynamics in an Agent-Based Micro Simulation FrameworkGustavo Garcia Manzato

nr 158Lightweight Floor System for Vibration ComfortSander Zegers

nr 159Aanpasbaarheid van de DraagstructuurRoel Gijsbers

nr 160'Village in the City' in Guangzhou, ChinaYanliu Lin

nr 161Climate Risk Assessment in MuseumsMarco Martens

nr 162Social Activity-Travel PatternsPauline van den Berg

nr 163Sound Concentration Caused by Curved SurfacesMartijn Vercammen

nr 164Design of Environmentally Friendly Calcium Sulfate-Based Building Materials: Towards an Improved Indoor Air QualityQingliang Yu

nr 165Beyond Uniform Thermal Comfort on the Effects of Non-Uniformity and Individual PhysiologyLisje Schellen

nr 166Sustainable Residential DistrictsGaby Abdalla

nr 167Towards a Performance Assessment Methodology using Computational Simulation for Air Distribution System Designs in Operating RoomsMônica do Amaral Melhado

nr 168Strategic Decision Modeling in Brownfield RedevelopmentBrano Glumac

nr 169Pamela: A Parking Analysis Model for Predicting Effects in Local AreasPeter van der Waerden

nr 170A Vision Driven Wayfinding Simulation-System Based on the Architectural Features Perceived in the Office EnvironmentQunli Chen

nr 171Measuring Mental Representations Underlying Activity-Travel ChoicesOliver Horeni

nr 172Modelling the Effects of Social Networks on Activity and Travel BehaviourNicole Ronald

nr 173Uncertainty Propagation and Sensitivity Analysis Techniques in Building Performance Simulation to Support Conceptual Building and System DesignChristian Struck

nr 174Numerical Modeling of Micro-Scale Wind-Induced Pollutant Dispersion in the Built EnvironmentPierre Gousseau

nr 175Modeling Recreation Choices over the Family LifecycleAnna Beatriz Grigolon

nr 176Experimental and Numerical Analysis of Mixing Ventilation at Laminar, Transitional and Turbulent Slot Reynolds NumbersTwan van Hooff

nr 177Collaborative Design Support:Workshops to Stimulate Interaction and Knowledge Exchange Between PractitionersEmile M.C.J. Quanjel

nr 178Future-Proof Platforms for Aging-in-PlaceMichiel Brink

nr 179Motivate: A Context-Aware Mobile Application forPhysical Activity PromotionYuzhong Lin

nr 180Experience the City:Analysis of Space-Time Behaviour and Spatial Learning Anastasia Moiseeva

nr 181Unbonded Post-Tensioned Shear Walls of Calcium Silicate Element MasonryLex van der Meer

nr 182Construction and Demolition Waste Recycling into Innovative Building Materials for Sustainable Construction in TanzaniaMwita M. Sabai

nr 183Durability of Concretewith Emphasis on Chloride MigrationPrzemys�aw Spiesz

nr 184Computational Modeling of Urban Wind Flow and Natural Ventilation Potential of Buildings Rubina Ramponi

nr 185A Distributed Dynamic Simulation Mechanism for Buildings Automation and Control SystemsAzzedine Yahiaoui

nr 186Modeling Cognitive Learning of UrbanNetworks in Daily Activity-Travel BehaviorSehnaz Cenani Durmazoglu

nr 187Functionality and Adaptability of Design Solutions for Public Apartment Buildingsin GhanaStephen Agyefi-Mensah

nr 188A Construction Waste Generation Model for Developing CountriesLilliana Abarca-Guerrero

nr 189Synchronizing Networks:The Modeling of Supernetworks for Activity-Travel BehaviorFeixiong Liao

nr 190Time and Money Allocation Decisions in Out-of-Home Leisure Activity Choices Gamze Zeynep Dane

nr 191How to Measure Added Value of CRE and Building Design Rianne Appel-Meulenbroek

nr 192Secondary Materials in Cement-Based Products:Treatment, Modeling and Environmental InteractionMiruna Florea

nr 193Concepts for the Robustness Improvement of Self-Compacting Concrete: Effects of Admixtures and Mixture Components on the Rheology and Early Hydration at Varying TemperaturesWolfram Schmidt

�¸

nr 194Modelling and Simulation of Virtual Natural Lighting Solutions in BuildingsRizki A. Mangkuto

nr 195Nano-Silica Production at Low Temperatures from the Dissolution of Olivine - Synthesis, Tailoring and ModellingAlberto Lazaro Garcia

nr 196Building Energy Simulation Based Assessment of Industrial Halls for Design SupportBruno Lee

nr 197Computational Performance Prediction of the Potential of Hybrid Adaptable Thermal Storage Concepts for Lightweight Low-Energy Houses Pieter-Jan Hoes

nr 198Application of Nano-Silica in Concrete George Quercia Bianchi

nr 199Dynamics of Social Networks and Activity Travel BehaviourFariya Sharmeen

nr 200Building Structural Design Generation and Optimisation including Spatial ModificationJuan Manuel Davila Delgado

nr 201Hydration and Thermal Decomposition of Cement/Calcium-Sulphate Based MaterialsAriën de Korte

nr 202Republiek van Beelden:De Politieke Werkingen van het Ontwerp in Regionale PlanvormingBart de Zwart

nr 203Effects of Energy Price Increases on Individual Activity-Travel Repertoires and Energy ConsumptionDujuan Yang

nr 204Geometry and Ventilation:Evaluation of the Leeward Sawtooth Roof Potential in the Natural Ventilation of BuildingsJorge Isaac Perén Montero

nr 205Computational Modelling of Evaporative Cooling as a Climate Change Adaptation Measure at the Spatial Scale of Buildings and StreetsHamid Montazeri

nr 206Local Buckling of Aluminium Beams in Fire ConditionsRonald van der Meulen

nr 207Historic Urban Landscapes:Framing the Integration of Urban and Heritage Planning in Multilevel GovernanceLoes Veldpaus

nr 208Sustainable Transformation of the Cities:Urban Design Pragmatics to Achieve a Sustainable CityErnesto Antonio Zumelzu Scheel

nr 209Development of Sustainable Protective Ultra-High Performance Fibre Reinforced Concrete (UHPFRC):Design, Assessment and ModelingRui Yu

nr 210Uncertainty in Modeling Activity-Travel Demand in Complex Uban SystemsSoora Rasouli

nr 211Simulation-based Performance Assessment of Climate Adaptive Greenhouse ShellsChul-sung Lee

nr 212Green Cities:Modelling the Spatial Transformation of the Urban Environment using Renewable Energy TechnologiesSaleh Mohammadi

nr 213A Bounded Rationality Model of Short and Long-Term Dynamics of Activity-Travel BehaviorIfigeneia Psarra

nr 214Effects of Pricing Strategies on Dynamic Repertoires of Activity-Travel BehaviourElaheh Khademi

nr 215Handstorm Principles for Creative and Collaborative WorkingFrans van Gassel

nr 216Light Conditions in Nursing Homes:Visual Comfort and Visual Functioning of Residents Marianne M. Sinoo

nr 217Woonsporen:De Sociale en Ruimtelijke Biografie van een Stedelijk Bouwblok in de Amsterdamse Transvaalbuurt Hüseyin Hüsnü Yegenoglu

nr 218Studies on User Control in Ambient Intelligent SystemsBerent Willem Meerbeek

nr 219Daily Livings in a Smart Home:Users’ Living Preference Modeling of Smart HomesErfaneh Allameh

nr 220Smart Home Design:Spatial Preference Modeling of Smart HomesMohammadali Heidari Jozam

nr 221Wonen:Discoursen, Praktijken, PerspectievenJos Smeets

nr 222Personal Control over Indoor Climate in Offices:Impact on Comfort, Health and ProductivityAtze Christiaan Boerstra

nr 223Personalized Route Finding in Multimodal Transportation NetworksJianwe Zhang

nr 224The Design of an Adaptive Healing Room for Stroke PatientsElke Daemen

nr 225Experimental and Numerical Analysis of Climate Change Induced Risks to Historic Buildings and CollectionsZara Huijbregts

nr 226Wind Flow Modeling in Urban Areas Through Experimental and Numerical TechniquesAlessio Ricci

nr 227Clever Climate Control for Culture:Energy Efficient Indoor Climate Control Strategies for Museums Respecting Collection Preservation and Thermal Comfort of VisitorsRick Kramer

nr 228Fatigue Life Estimation of Metal Structures Based on Damage ModelingSarmediran Silitonga

nr 229A multi-agents and occupancy based strategy for energy management and process control on the room-levelTimilehin Moses Labeodan

nr 230Environmental assessment of Building Integrated Photovoltaics:Numerical and Experimental Carrying Capacity Based ApproachMichiel Ritzen

nr 231Performance of Admixture and Secondary Minerals in Alkali Activated Concrete:Sustaining a Concrete Future Arno Keulen

nr 232World Heritage Cities and Sustainable Urban Development: Bridging Global and Local Levels in Monitor-ing the Sustainable Urban Development of World Heritage Cities Paloma C. Guzman Molina

nr 233 Stage Acoustics and Sound Exposure in Performance and Rehearsal Spaces for Orchestras: Methods for Physical MeasurementsRemy Wenmaekers

nr 234Municipal Solid Waste Incineration (MSWI) Bottom Ash:From Waste to Value Characterization, Treatments and ApplicationPei Tang

nr 235Large Eddy Simulations Applied to Wind Loading and Pollutant DispersionMattia Ricci

nr 236Alkali Activated Slag-Fly Ash Binders: Design, Modeling and ApplicationXu Gao

nr 237Sodium Carbonate Activated Slag: Reaction Analysis, Microstructural Modification & Engineering ApplicationBo Yuan

nr 238Shopping Behavior in MallsWidiyani

nr 239Smart Grid-Building Energy Interactions:Demand Side Power Flexibility in Office BuildingsKennedy Otieno Aduda

nr 240Modeling Taxis Dynamic Behavior in Uncertain Urban EnvironmentsZheng Zhong

nr 241Gap-Theoretical Analyses of Residential Satisfaction and Intention to Move Wen Jiang

nr 242Travel Satisfaction and Subjective Well-Being: A Behavioral Modeling Perspective Yanan Gao

nr 243Building Energy Modelling to Support the Commissioning of Holistic Data Centre OperationVojtech Zavrel

nr 244Regret-Based Travel Behavior Modeling:An Extended FrameworkSunghoon Jang

nr 245Towards Robust Low-Energy Houses: A Computational Approach for Performance Robustness Assessment using Scenario AnalysisRajesh Reddy Kotireddy

nr 246Development of sustainable and functionalized inorganic binder-biofiber compositesGuillaume Doudart de la Grée

nr 247A Multiscale Analysis of the Urban Heat Island Effect: From City Averaged Temperatures to the Energy Demand of Individual BuildingsYasin Toparlar

nr 248Design Method for Adaptive Daylight Systems for buildings covered by large (span) roofsFlorian Heinzelmann

nr 249Hardening, high-temperature resistance and acid resistance of one-part geopolymersPatrick Sturm

nr 250Effects of the built environment on dynamic repertoires of activity-travel behaviourAida Pontes de Aquino

nr 251Modeling for auralization of urban environments: Incorporation of directivity in sound propagation and analysis of a frame-work for auralizing a car pass-byFotis Georgiou

nr 252Wind Loads on Heliostats and Photovoltaic TrackersAndreas Pfahl

Propositions

associated with the dissertation

“Approaches for computational performance optimization of innovative adaptive façade concepts”

R.C.G.M. Loonen, September 2018, Eindhoven University of Technology

1. Neither truly passive designs nor building strategies that rely only on active systems to provide occupant comfort can solve all of the challenges necessary to achieve a sustainable built environment. Building elements that can respond to their environment by controlled adaptation to interior and exterior conditions unite the benefits of both approaches and are therefore a more promising way forward. <This dissertation>

2. Innovative adaptive façade systems can push the envelope of building performance by outperforming the best possible building envelope with fixed properties. These adaptive façade systems provide new windows of opportunity for high-performance building design and operation. <This dissertation>

3. Fifteen years ago, Selkowitz et al. remarked that “Delivering dynamic, responsive control of solar gain and glare, but permitting daylight use, is still the holy grail of façade technology”; this is still the case. <Selkowitz et al., 2003* and this dissertation (2018)>

4. ‘In silico’ research can provide insights that are impossible to obtain with physical experiments and is therefore a valuable method in product development of innovative building envelope components. <This dissertation>

5. Researchers who use and develop legacy simulation software are in the fortunate position to be standing on the shoulders of giants. This offers protection, while making it easy to rise fast and reach great heights. These heirs, however, should be aware that giants have not been storied for their accomplishments in software documentation and that some of the methods they are using were state-of-the-art in a different era.

6. The building industry is regularly denounced as being one of the most conservative sectors in science and technology. This tendency to preserve the status quo is a mixed blessing. It is indeed true that society is sometimes deprived of valuable innovations due to process inefficiencies and other seemingly irrational barriers. On the other hand, few aspects in life are regarded as being as vital to our well-being, protection and self-identity as the buildings we live in. The fact that buildings do not become a playground for passing fads and speculative technologies is therefore worth cherishing.

7. When the intricate knot between research and academic education gets disentangled, one ends up with two dysfunctional pieces of yarn. <Inspired by Rik Torfs, former Rector Magnificus KU Leuven>

8. Academic research requires both inspiration and transpiration. Those who plan to cherry-pick the low-hanging fruit will soon discover that it has already been harvested.

9. Stimulated by the virtues of the adage ‘if you can’t measure it, you can’t improve it’, metric-driven decisions increasingly permeate our everyday life. This can backfire when measurement of performance turns into metric fixation. We should be very careful that what is actually measured is a reasonable proxy of what is intended to be measured. Measurement is therefore not a substitute for judgement. <Inspired by Muller, 2018†>

10. When Hofstadter's law ("It always takes longer than you expect, even when you take into account Hofstadter's Law") coincides with Parkinson's law ("Work expands so as to fill the time available for its completion"), the likely result is a positive feedback loop with potentially negative consequences.

* Selkowitz, S., Aschehoug, Ø. and Lee, E.S. 2003. Advanced Interactive Facades – Critical Elements for Future Green Buildings?, in: Proceedings of GreenBuild, the Annual USGBC International Conference and Expo.

† Muller, J.Z. 2018. The Tyranny of the Metrics. Princeton University Press

/ Department of the Built Environment