Intelligent Residential Buildings and the Behaviour of the Occupants

SPRINGER BRIEFS IN APPLIED SCIENCES AND TECHNOLOGY

Pedro F. Pereira Nuno M. M. Ramos João M. P. Q. Delgado

Intelligent Residential Buildings and the Behaviour of the Occupants State of the Art 123

SpringerBriefs in Applied Sciences and Technology

SpringerBriefs present concise summaries of cutting-edge research and practical applications across a wide spectrum of fields. Featuring compact volumes of 50– 125 pages, the series covers a range of content from professional to academic. Typical publications can be: • A timely report of state-of-the art methods • An introduction to or a manual for the application of mathematical or computer techniques • A bridge between new research results, as published in journal articles • A snapshot of a hot or emerging topic • An in-depth case study • A presentation of core concepts that students must understand in order to make independent contributions SpringerBriefs are characterized by fast, global electronic dissemination, standard publishing contracts, standardized manuscript preparation and formatting guidelines, and expedited production schedules. On the one hand, SpringerBriefs in Applied Sciences and Technology are devoted to the publication of fundamentals and applications within the different classical engineering disciplines as well as in interdisciplinary fields that recently emerged between these areas. On the other hand, as the boundary separating fundamental research and applied technology is more and more dissolving, this series is particularly open to trans-disciplinary topics between fundamental science and engineering. Indexed by EI-Compendex, SCOPUS and Springerlink.

More information about this series at http://www.springer.com/series/8884

Pedro F. Pereira Nuno M. M. Ramos João M. P. Q. Delgado •

Intelligent Residential Buildings and the Behaviour of the Occupants State of the Art

123

Pedro F. Pereira CONSTRUCT-LFC, Faculty of Engineering University of Porto Porto, Portugal

João M. P. Q. Delgado Department of Civil Engineering, Faculty of Engineering University of Porto Porto, Portugal

Nuno M. M. Ramos CONSTRUCT-LFC, Faculty of Engineering University of Porto Porto, Portugal

ISSN 2191-530X ISSN 2191-5318 (electronic) SpringerBriefs in Applied Sciences and Technology ISBN 978-3-030-00159-9 ISBN 978-3-030-00160-5 (eBook) https://doi.org/10.1007/978-3-030-00160-5 Library of Congress Control Number: 2018956593 © The Author(s), under exclusive license to Springer Nature Switzerland AG 2019 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

Over the last years, intelligent buildings and the behaviour of the occupants have been the scope of many studies. The number of studies of these areas is growing, as they appear to be the next step to optimize the energy efficiency of the buildings. The concept of intelligent building is associated with the creation of a management system that takes into account the requirements of its occupants in terms of thermal comfort and the activities of their daily life, maintaining a good indoor air quality and minimizing the energy consumption. Thus, there is a need to study and combine these issues to obtain the new generation of buildings. In commercial or office buildings, these systems are already in an intermediate stage of implementation. However, in the residential sector, it still does not have a significant implementation. In mild climate regions, where the interactions of the occupants with the building mechanisms are the primary way to meet their comfort and ventilation requirements, the importance of occupant behaviour studies and its incorporation in the algorithms of the intelligent buildings become even more important. The main benefit of the book is that it contains a state of the art of two areas that have to be treated together in order to emerge a new concept of buildings, that are more efficient, more comfortable and healthier. Porto, Portugal

Pedro F. Pereira Nuno M. M. Ramos João M. P. Q. Delgado

v

Acknowledgements

This work was developed on the frame of the Project POCI-01-0145-FEDER007457—CONSTRUCT—Institute of R&D In Structures and Construction funded by FEDER funds through COMPETE2020—Programa Operacional Competitividade e Internacionalização (POCI)—and by national funds through FCT— Fundação para a Ciência e a Tecnologia.

vii

Contents

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 3

. . . . . . . . . . .

5 5 5 5 12 14 16 23 30 35 42

3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

51 54

2 State of the Art . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Intelligent Buildings Adapted to Their Occupants . . . 2.2.1 Concept of Intelligent Buildings . . . . . . . . . . 2.2.2 Intelligent Residential Buildings . . . . . . . . . . 2.3 Occupant Behaviour in Residential Buildings . . . . . . 2.3.1 Detection of the Actions of the Occupants . . 2.3.2 Assessment of the Drivers of the Occupants . 2.3.3 Assessment of the Impacts of the Occupants . 2.3.4 Modelling the Occupant Behaviour . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

ix

Abbreviations

ADL AIIB ASHRAE ASTM BAS BMS BP BRI CEN CIBSE CIE CMV DGEG DNAS DSF E EMS EPBD EU HMM HVAC IAQ IB IBG IBI ICT INE ISO

Activities of daily living Asian Institute of Intelligent Buildings American Society of Heating, Refrigeration and Air Conditioning Engineers American Society for Testing and Materials Building automation system Building management system Building performance Building-related illness European Committee for Standardization Chartered Institution of Building Services Engineers International Commission of Lighting (France) Centralized mechanical ventilation General Direction of Geology and Energy Bibliographic Framework of the Behaviour of Buildings Occupants— Drivers, Needs, Actions and Systems Double skin façade Energy Energy management systems European Directive on Energy Performance of Buildings European Union Hidden Markov models Heating, ventilation and air conditioning Indoor air quality Intelligent buildings Intelligent Building Group Intelligent Building Institute Information and communication technologies National Institute of Statistics (Portugal) International Organization for Standardization

xi

xii

JIBI KDD MCMC NDIR NIOSH OSHA PIR QEM SBS WHO

Abbreviations

Japanese Intelligent Buildings Knowledge Discovery in Databases Markov chain Monte Carlo Non-dispersive infrared National Institute for Occupational Safety and Health Occupational Safety and Health Administration Passive infrared Quality Environment Modules Sick building syndrome World Health Organization

Chapter 1

Introduction

Energy consumption can be divided by sector of activity, but the building consumption differs from country to country. It is close to 40% in the United States of America (USA) and Europe (2010/31/EU 2010); however, it is about 30% in the countries located in the south of Europe, including Portugal (Garrido-Soriano et al. 2012; INE and DGEG 2011). In Portugal, according to the data collected by INE and DGEG (2011), the largest energy consumption is in the kitchen with a percentage of 39.1%, followed by heating food products with 23.5%, heating the environment with 21.5% and household appliances with 10.9%. By EU imposing, the member states will have to transpose into internal regulations the Energy Performance of Buildings Directive (EPBD) transcribed in Directive 2010/31/EU (2010). This Directive stipules a set of targets to be met by 2020 that entail a 20% reduction in energy consumption and CO2 emissions and a minimum of 20% of renewable energy as source of energy. Among the suggestions presented in Directive 2010/31/ EU (2010) is the increasing use of active control systems and intelligent systems that enhance the energy efficiency of new buildings or the target of major renovations. In particular, the new EPBD encourages the use of “automation, control and monitoring systems that aim to save energy”. According to the literature (Clements-Croome 2004; Wang 2010), “true” intelligence in buildings is only achieved when the buildings respect the demands of their occupants, using the least amount of energy possible. In order to respect the requirements of the occupants, it is first necessary to know their behaviour and the way in which it influences the interior space of a residential fraction. The study of the occupant’s behaviour of residential buildings is therefore considered an area of major importance, and the results will be the basis for possible modes of operation of intelligent buildings (IBs). On the other hand, occupant behaviour is widely recognized as being a preponderant factor for the existence of uncertainty in the energy performance of buildings (Yan et al. 2015). Several studies have analysed the effect of occupant’s behaviour. The study developed by Sonderegger (1978) was one of the first to be carried out, in which a value of 33% was presented for the contribution of the occupants’ behaviour to the total energy consumption. More recently, stud© The Author(s), under exclusive license to Springer Nature Switzerland AG 2019 P. F. Pereira et al., Intelligent Residential Buildings and the Behaviour of the Occupants , SpringerBriefs in Applied Sciences and Technology, https://doi.org/10.1007/978-3-030-00160-5_1

1

2

1 Introduction

ies on the subject point to a contributions in the order of 20% (Kleiminger et al. 2014; Luo et al. 2014; Gao and Whitehouse 2009), while other researchers such as Guerra Santin et al. (2009) found contributions lower than 10% and the researches of Gram-Hanssen (2010) and D’Oca et al. (2014) present contributions higher than 50%. Regarding the influence of the occupants on the building ventilation, Kvisgaard and Collet (1986) found that the actions of the occupants provide 63% of the total air changes per hour (ACH) of the dwellings with natural ventilation. These actions corresponded mainly to the handling of doors and windows. In a similar study, Iwashita and Akasaka (1997) found a direct relationship between occupant influence and housing ACH with contributions that reached 87%, due to different occupant patterns in window and door handling. The work presented by Wallace et al. (2002) demonstrated that the typical window aperture affects the ACH value in 1 h−1 and Pereira et al. (2017) quantified the impact of some occupant actions on a building, finding contributions of up to 4 h−1 . These values were obtained in studies of different localities and different climates, not being able to be compared among them, giving, however, an idea about the impact that the occupants can have in a dwelling. The investigation related to the occupants’ behaviour of residential buildings and their impact on them has been worthy of study by several authors with an increase in the number of studies by years (D’Oca et al. 2018). The main motivations for the study of this theme are related to two distinct situations (Jia et al. 2017; Hong et al. 2017): • The need to optimize building automated systems (BASs) | energy management systems (EMSs) | building management systems (BMSs), adapting them to occupants’ habits in terms of energy consumption, indoor air quality (IAQ), thermal comfort and occupant habits; • The need to reduce the mismatch between the results of the numerical simulation programs and the actual performances of the buildings. According to Hong et al. (2015), due to the growing interest of researchers in this area of study, there was a need to organize it in subareas, which led to the creation of an ontology to represent the behaviour of occupants in buildings. By this way, the authors created a research framework divided into four subareas: drivers, needs, actions and systems, giving rise to the acronym DNAS. In these four areas, motivations are understood as the environmental factors perceived by the occupants in the outer world that provoke a stimulus in their inner world to fulfil a physical, psychological, or physiological need. Thus, needs are the requirements in the three dimensions referred to the inner world of the occupant that need to be met in order to the occupant feel satisfied in the environment in which he is inserted. The actions then emerge as the way the occupant has to fulfil his needs, thereby bridging the inner and outer world through interaction with the systems because of his need. The systems are therefore the equipment or mechanisms belonging to the built environment of the building that can be manipulated by the occupant to influence the interior environment.

References

3

References 2010/31/EU, Directive (2010) Energy Performance of Buildings Directive (EPBD) Clements-Croome D (2004) Intelligent buildings—design, management and operation. Thomas Telford D’Oca S, Fabi V, Corgnati SP, Andersen RK (2014) Effect of thermostat and window opening occupant behaviour models on energy use in homes. Build Simul 7(6):683–694. http://www.sco pus.com/inward/record.url?eid=2-s2.0-84906315309&partnerID=40&md5=3bfa70c0f24b3cfa 9d27df2036e2b215 D’Oca S, Hong T, Langevin J (2018) The human dimensions of energy use in buildings: a review. Renew Sustain Energy Rev 81:731–742. https://www.scopus.com/inward/record.uri?eid=2-s2. 0-85027558816&doi=10.1016%2fj.rser.2017.08.019&partnerID=40&md5=d10b6408f890100 165480c42b1d31238 Gao G, Whitehouse K (2009) The self-programming thermostat: optimizing setback schedules based on home occupancy patterns. Comunicação apresentada em 1st ACM Workshop on Embedded Sensing Systems for Energy-Efficiency in Buildings, BUILDSYS 2009, in Conjunction with ACM SenSys 2009, em Berkeley, CA Garrido-Soriano N, Rosas-Casals M, Ivancic A, Álvarez-Del Castillo MD (2012) Potential energy savings and economic impact of residential buildings under national and regional efficiency scenarios. A Catalan case study. Energy Build 49:119–125. http://www.scopus.com/inward/reco rd.url?eid=2-s2.0-84861832911&partnerID=40&md5=973148e727ada66fff245746dbf5fbf7 Gram-Hanssen K (2010) Residential heat comfort practices: understanding users. Build Res Inf 38(2):175–186. http://www.scopus.com/inward/record.url?eid=2-s2.0-77649289979&partne rID=40&md5=fa2f02c5f3e4483a219d4245718b66c8 Guerra Santin O, Itard L, Visscher H (2009) The effect of occupancy and building characteristics on energy use for space and water heating in Dutch residential stock. Energy Build 41(11):1223–1232. http://www.scopus.com/inward/record.url?eid=2-s2.0-69249216516&partn erID=40&md5=b316031a3378c0ded482e6f861f6c07b Luo M, Cao B, Zhou X, Li M, Zhang J, Ouyange Q, Zhu Y (2014) Can personal control influence human thermal comfort? A field study in residential buildings in China in winter. Energy Build 72:411–418. http://www.scopus.com/inward/record.url?eid=2-s2.0-84893145375&partnerID=4 0&md5=dce1f1fce7dfc407ced0f5e6070f2bc9 Hong T, D’Oca S, Turner WJN, Taylor-Lange SC (2015) An ontology to represent energy-related occupant behaviour in buildings. Part I: Introduction to the DNAs framework. Build Environ 92:764–777. https://www.scopus.com/inward/record.uri?eid=2-s2.0-84929489501&doi=10.101 6%2fj.buildenv.2015.02.019&partnerID=40&md5=392f5ec7c1125675d279d62ea3d18e69 Hong T, Yan D, D’Oca S, Chen CF (2017) Ten questions concerning occupant behaviour in buildings: the big picture. Build Environ 114:518–530. https://www.scopus.com/inward/record. uri?eid=2-s2.0-85009508216&doi=10.1016%2fj.buildenv.2016.12.006&partnerID=40&md5=e 0d0a24fe1c51b759b4fb520f6405262 INE, Instituto Nacional de Estatistica, Direção Geral de Energia e Geologia DGEG (2011) Inquérito ao Consumo de Energia no Sector Doméstico 2010 Iwashita G, Akasaka H (1997) The effects of human behaviour on natural ventilation rate and indoor air environment in summer—a field study in southern Japan. Energy Build 25(3):195–205. http://www.scopus.com/inward/record.url?eid=2-s2.0-0031142616&part nerID=40&md5=b7f3e5e4921343f8be2eae346bb6d1a4 Jia M, Srinivasan RS, Raheem AA (2017) From occupancy to occupant behaviour: an analytical survey of data acquisition technologies, modeling methodologies and simulation coupling mechanisms for building energy efficiency. Renew Sustain Energy Rev 68:525–540. https://www.scopus.com/inward/record.uri?eid=2-s2.0-84994000394&doi=10.101 6%2fj.rser.2016.10.011&partnerID=40&md5=d8bc76cffaa73b5f0878e8321d6edb89 Kleiminger W, Mattern F, Santini S (2014) Predicting household occupancy for smart heating control: a comparative performance analysis of state-of-the-art approaches. Energy Build

4

1 Introduction

85:493–505. http://www.scopus.com/inward/record.url?eid=2-s2.0-84908324413&partnerID=4 0&md5=0f0b58bc0a2619bac90bee0052d41410 Kvisgaard B, Collet PF (1986) Occupants’ influence on air change in dwellings. Comunicação apresentada em 7th AIC Conference, em Stratford-upon-Avon, UK Pereira PF, Ramos NMM, Almeida RMSF, Simões ML, Barreira E (2017) Occupant influence on residential ventilation patterns in mild climate conditions. Energy Procedia 132(Suppl C):837–842. http://www.sciencedirect.com/science/article/pii/S1876610217348166 Sonderegger RC (1978) Dynamic models of house heating based on equivalent thermal parameters. Princeton University, Princeton Wallace LA, Emmerich SJ, Howard-Reed C (2002) Continuous measurements of air change rates in an occupied house for 1 year: the effect of temperature, wind, fans, and windows. J Expo Anal Environ Epidemiol 12(4):296–306. https://www.scopus.com/inward/record.uri?eid=2-s2.0-0036 068753&partnerID=40&md5=63e1c3a9fcb9ffa807a234b56bb935d0 Wang S (2010) Intelligent buildings and building automation. Taylor & Francis Yan D, O’Brien W, Hong T, Feng X, Burak Gunay H, Tahmasebi F, Mahdavi A (2015) Occupant behaviour modeling for building performance simulation: current state and future challenges. Energy Build 107:264–278. https://www.scopus.com/inward/record.uri?eid=2-s2.0-849405132 01&partnerID=40&md5=3d626d60485cc9a7b34f92bca19b1a57

Chapter 2

State of the Art

2.1 Motivation The present book was the result of an extensive bibliographical research in order to collect information on the state of the art of intelligent buildings and the behaviour of their occupants. The following subchapters present the state of the art in the following areas: • • • • •

Intelligent buildings adapted to their occupants; Monitoring of occupants of buildings; Motivations for the actions of occupants of buildings; Impacts of occupant behaviour of buildings; Modelling the behaviour of occupants of buildings.

The form of interconnection of these areas can be observed in the scheme present in Fig. 2.1.

2.2 Intelligent Buildings Adapted to Their Occupants 2.2.1 Concept of Intelligent Buildings The word “intelligence” has its etymological origin in the Latin term “intelligere”. This is a word composed by two terms: inter (“between”) and legere (“to choose or pick out”). Thus, due to the etymological origin of the word, there is intrinsic to the concept of the capacity to choose between several options in function of the objectives of the decision maker. As an object of philosophical study, the concept of intelligence has undergone several changes due to its close connection with the

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2019 P. F. Pereira et al., Intelligent Residential Buildings and the Behaviour of the Occupants , SpringerBriefs in Applied Sciences and Technology, https://doi.org/10.1007/978-3-030-00160-5_2

5

6

2 State of the Art

Fig. 2.1 Interconnection between the target areas of the state-of-the-art study

sociopolitical and ideological contexts, assuming different versions depending on the capacities extolled by each society at different times. It is also possible to obtain different concepts of intelligence at the same time and in the same society by people belonging to different areas (Afonso 2007). There is, however, another kind of intelligence disconnected from the natural sciences and sociology, the artificial intelligence. According to Winston (1993), artificial intelligence is the computational study that makes perception, reasoning and action possible. In this way, it can be mentioned that the great difference of the artificial intelligence with respect to the human intelligence is the focus in the computational means, being in turn different from other studies connected to the science of the computers for if it is intended that the machines endowed with intelligence that perceive, reason and act according to automatic learning. There are other definitions of artificial intelligence; for example, the authors Russel and Norvig (2003) decided to group their definitions into four quadrants. The authors found definitions that concerned the process of thinking and reasoning and other definitions that focused more on the action itself. Two variants were found, definitions that linked artificial intelligence to an attempt to simulate human intelligence and others that linked artificial intelligence to the most rational intelligence, which is not always human. The research area of intelligent buildings (IB), although current, is not new. Several authors consider that the beginning of the discussions related to this subject, at least theoretically, reassembles to the 1980s. On 1 December 1983, the New York Times (Marcus 1983) reports the opening of the first intelligent building in the world, in Hartford, Connecticut. It was an office building called “CityPlace” with an area of about 0.12 km2 . The intelligence of the building meant that its services would be controlled by a computer system and connected by a fibre optic network running

2.2 Intelligent Buildings Adapted to Their Occupants

7

through its core. Several functions such as heating, ventilation, lighting, transportation, security, fire protection and, more important, telecommunications and electronic office services would be integrated, providing relevant savings in the construction and management of the building (So and Chan 1999). On 13 May 1984, the same newspaper wrote an article describing the concept of a new generation of buildings with the capacity to “think” for themselves. This article described a house that controlled the temperature, the lighting and some constructive elements of the building according to the occupation and the occupant’s requirements (Sinopoli 2010). Since the beginning of the researches in this area, the definition and field of application have been altered, mainly due to the development of relevant technologies and the needs of changing the built environment. In resume, the approach of the IB can be described as follows: • Until 1985: intelligent buildings are buildings controlled automatically for a function. • From 1986 to 1991: intelligent buildings are buildings capable of responding to the needs of change. • From 1992 to present: intelligent buildings are buildings with characteristics that effectively meet the changing needs. In the above-mentioned summary chronology, it is referred to the change of scope, and it should be noted that the concept and scope of IB are not yet fully stabilized or parameterized. The discussion has been extended for decades, including new technologies, new interface platforms, and expanded the area of influence of the automation of buildings without being able to standardize the field of intelligence of the building. It is difficult to achieve a stabilization of evolution since, as they are buildings that incorporate technology and this is the most creative and innovative areas of all, this does not seem plausible (Wang 2010). The definition of IB is then considered to be vast and controversial, for example, Wigginton and Harris (2002) found, in 2002, about 30 definitions of building-related intelligence. In evolutionary terms, Wong et al. (2005a) consider that the first definitions of IB focused on the purely technological properties of buildings and did not mention interaction with the occupants. This purely technological view has been criticized by many researchers since the mid-1980s, with variants of this view supporting IB to respond to occupant requirements. At present, there are some large groups of intelligent buildings research, which are described in the following paragraphs. According to the Intelligent Building Institute (IBI), an IB should provide a productive and economically efficient environment by optimizing four basic components and their interconnection; • • • •

Structure; Systems; Services; Management.

8

2 State of the Art

Thus, the interconnection of the four elements of an IB aims to satisfy the needs of the occupants and owners of the buildings consuming the least possible resources. Due to the close linking of an IB with the needs of occupants/owners, it is not credible that the IB classification is attributed because they do not have a list of minimum eligible characteristics. According to IBI, the only feature that intelligent buildings have is a structure designed to accommodate changes at a controlled cost (So and Chan 1999; Wang 2010). According to the Intelligent Building Group (IBG), an IB is one that creates an environment that maximizes the productivity of building occupants without simultaneously compromising the efficient management of resources and costs in their lifetime. From the concept, it is deduced that an IB is one that incorporates the best available concepts, materials, systems and integration technologies to achieve a building that meets or exceeds the performance requirements of building stakeholders, which include owners and occupants, as well as the local and global community (So and Chan 1999; IBG 2014; Wang 2010). According to the Japanese Intelligent Buildings (JIBI), an IB has convenient communication and automation services for occupant use. The fundamental aspects of JIBI are the following (Wang 2010): • Ensuring the satisfaction and comfort of people working indoors; • Rationalize the management of buildings to provide more attractive low-cost administrative services; • Rapid, flexible and economical capacity to change according to the sociological changes in the environment, diversified and complex means of work and business strategies; • Serve as a place to receive and transmit information and support efficient management. According to the European Smart Accelerate project (Nikolaou et al. 2004), an IB is one that provides an economically and technically efficient environment by optimizing its four basic components: structure; systems; services and management and having as a concern the interrelations between them in order to achieve: • The benefit of the owners and their requirements for the interior comfort of the building; • Maximizing occupant efficiency; • Effective management of resources with minimum life cycle costs; • Facilitating building management and optimizing environmental and economic impact due to changes in the interior environment. The definition of IB in China and Singapore (So and Chan 1999) implies the fulfilment of certain characteristics that should be taken into account: • Automatic control systems to monitoring various installations; • Data infrastructure for a good flow of information between floors; • Adequate telecommunications facilities.

2.2 Intelligent Buildings Adapted to Their Occupants

9

According to So and Chan (1999) and Wang (2010), an IB should be labelled as “3A” or “5A” if they have three or five automatic functions: • • • • •

Automatic communication; Automation in the office; Automation in building management; Fire detection system; Automation in the maintenance system.

The authors of the International Building Index (Wong et al. 2005b) indicate that the Japanese model is best suited to achieve a global uniformity of an IB. In this sense, the Asian Institute of Intelligent Buildings (AIIB) decided to create a strategy to classify intelligent buildings stratified into ten levels, known as Quality Environment Modules (QEM): • M1: environmental friendliness—health and energy conservation; • M2: space utilization and flexibility; • M3: cost-effectiveness—operation and maintenance with emphasis on effectiveness; • M4: human comfort; • M5: working efficiency; • M6: safety and security measures—fire, earthquake, disaster and structural damages, etc.; • M7: culture; • M8: image of high technology; • M9: construction process and structure; and • M10: health and sanitation. Each of these ten levels has a number of key aspects hierarchically ordered at the priority level. In this way, the authors redefine the IB concept as one in which the design and construction should be based on the strict selection of the QEM in order to meet the requirements of the occupants. According to the author, this new definition has two aspects: • Technology; • Needs of the occupants. According to Sinopoli (2010), an IB implies the installation and use of systems of advanced technology and integrated in the buildings. These systems should include the following fields: • • • • •

Building automation; Safety; Telecommunications; Interfaces with occupants; Facilitates management systems.

10

2 State of the Art

Sinopoli (2010) also demonstrates that the traditional way of constructing causes each system to be treated individually without any interconnection or integration of the different systems, which is the main difference from an IB. The creation of an IB begins in the project where the infrastructures of each of the several systems can be reduced by their integration. An IB is also pointed out as essential for achieving intelligent electricity grids. Due to the different and sometimes contradictory ways of idealizing IB, Wang (2010) presents an approach to defining an IB, grouping into three categories as follows: • Performance-based definitions; • Service-based definitions; • System-based definitions. Performance-based definitions are those that define an IB by listing the requirements that each building must have. In this approach, it is easy to perceive the importance that the occupants have in defining the requirements. This philosophy gives more importance to the performance of the buildings and the requirements of the occupants rather than the technologies or systems provided. In this approach, it is important that owners and developers of buildings need to understand correctly what kind of buildings they want and also how to satisfy continuously the increasing demands of occupants. Energy and environmental performances of buildings are certainly among the important issues of an IB. It is argued that this type of building should have the ability to adapt itself quickly in response to internal and external conditions and to meet the changing demands of occupants. The principles advocated by IBG and IBI are included in this approach. However, Wong et al. (2005a) argue that the definitions of each of these institutes about IB are different; i.e. the definition in IBG focuses IB on occupant requirements, while the definition in IBI suggests more equipment. Service-based approaches define IB from the point of view of the services offered and the quality with which these services are provided. The services are thus emphasized to the requirements of the occupants, and these are placed at their disposal, even if the occupants at any given time do not feel the need or the advantage in their use. JIBI shares this philosophy. System-based approaches define IB according to the systems that comprise them. The focus for this type of approach lies in the technology of the systems used. The IB definitions in China and Singapore share this philosophy. According to Clements-Croome (2004), an IB should be sustainable, healthy, with current technology, responsive to the needs of the occupants and the business it houses, and should be flexible and adaptable to deal with change. This definition emphasizes the importance of design, construction and management of equipment and facilities. According to Wong et al. (2005a), IB research focused essentially on three interrelated areas that provide insight into different phases of IB life. These areas are essentially focused on the search for new technologies, ways to evaluate the performance of IB and the evaluation of investments.

2.2 Intelligent Buildings Adapted to Their Occupants

11

Yang and Peng (2001) and Wigginton and Harris (2002) argue that the IB concept has to be attached to the capacity of the building to learn and adapt in function of its surroundings and occupancy. Although it is considered important that an IB has the ability to automatically adapt in an intelligent way, it is important that it “loses intelligence” so that it can also adapt to the wishes of the occupants. In this way, it is considered essential that an IB should suit the occupants, but in the most efficient way possible. As is to be expected in an area in permanent and rapid evolution, the technologies of the contemporary IB are very different from the one existing at the beginning of the development of this area. The integration and interaction between IB subsystems are increasingly used as a way of information sharing. This principle already begins to extrapolate the building object so that the interaction extends to the city. Systems integration between IB may be based on safety, energy efficiency or city/block management issues under the 2010/31/EU (2010) Directive. An IB without the inclusion of technology could have been a possibility a few years ago, and at present, it is expected that an IB includes technology. Thus, it is considered that the concept of IB is intrinsic to the association of information and communication technologies (ICTs). Although the inclusion of technology in an IB is an inevitable way, it is not the amount of technology that gives the building a greater or lesser degree of intelligence, but rather the help that this technology can give the occupants in order to make the building more efficient and comfortable for the occupants. In this way, building technology is expected to be a means, not an end. The bibliography previously exposed has focused essentially on the definition of IB according to the active part of the building. Although, for Wang (2010), the main focus of IB is the efficient use of technology in which ICTs stand out, there are however other ways of contributing to the intelligence of buildings and these are essential ways to achieve high performance in buildings. This requires, in addition to technology, the design of an IB with efficient architecture and components. Thus, it is understood that an IB should have an architecture, structure and materials that optimize it and improve the performance of the building and its occupants. The concept of intelligent architecture, according to Wang (2010), is subdivided into three parts: • Intelligent design—a building with intelligent architecture should be in harmony with the surroundings using natural resources efficiently. It must also take into account the cultural, political and economic contexts of the area in which it is situated so as to be able to adapt to local requirements and occupants; • Appropriate use of smart technology—the study of occupants’ habits and preferences will be paramount in choosing the best technology, but above all the one that best fits the habits of the people. If this study is not done, there is a risk of creating a building without intelligence because it does not adapt to its occupants; • Smart use and maintenance of buildings—an IB should be thought of as a function of the life cycle. Regardless of its complexity, it should be easy to use, maintain and modify.

12

2 State of the Art

The definitions presented allow us to conclude that the definition of IB is not only one and consensual, but there are common positions. One of the important points in defining an IB lies in the fact that it can only be considered a smart building if it performs better than an identical one without “intelligence”. There is also another important point that will have to do with the interaction and satisfaction of the requirements of the occupants. Thus, it is assumed that an IB does not contradict the wills of the occupants, but that it achieves them as efficiently as possible.

2.2.2 Intelligent Residential Buildings Due to the links between IB and technology and automation in Portugal, the concept of IB is associated with the concept of domotics when it comes to residential buildings. Etymologically, home automation comes from two words domus (from Latin for home) and robotics (from robotic Czech). Domotics is commonly known as the application of information technology, electronics and robotics/automation to buildings in order to facilitate the interaction of occupants with them. Globally, this term is known as building automation systems (BAS), energy management system (EMS), building management system (BMS) or simply home automation, referring to the automation of buildings that can be residential or commercial/industrial. The control system of an IB consists of technological equipment, terminals that communicate with control devices or servers. Communication between the equipment takes place through communication protocols. The equipment connected to the network can be of three types: sensors, actuators and controllers, which are interconnected through a network. The sensors are electronic or mechanical components responsible for converting a physical variable into an electrical signal capable of being acquired by the system. Actuators are components capable of operating in the environment, controlled by an electrical signal. The control network is implemented so that there is communication between all components of the system, allowing access to the data acquired by the sensors and allowing the control components to send commands to the actuators. Communication is done through physical means such as cables of various types or through wireless transmitters and receivers. There is usually a network administrator system that manages the network and interacts with the user. Existing IB systems are based on standardized communication networks allowing the following (Sinopoli 2010): • Communication between applications; • More efficiency and monetary savings in materials, labour and equipment; • Functionality of systems with equipment from different manufacturers. Something quite important in a home automation system is its interaction with the occupants. At this level, there has been a great evolution, and at present, the provision of software for smartphones is becoming widespread so that the user can interact with the house when it is far from it.

2.2 Intelligent Buildings Adapted to Their Occupants

13

The automation of building components is related to the existence of a BAS (Wigginton and Harris 2002; Wang 2010; Sinopoli 2010). This system, through a preprogrammed algorithm, allows the control of several building components. At present, there is technology that allows to promote automation in the following levels: • • • • • •

Artificial lighting; Natural lighting; Solar radiation; Electricity generation; Ventilation; Air conditioner.

It is common to find this automation in commercial buildings and services. However, in residential buildings, their use is still low. According to the work done by Wyckmans (2005), building facades can adapt to external conditions in order to promote better interior conditions. The author focused his research on devices that influence the natural lighting of buildings. Examples of mobile, automated devices have been described, such as: • • • • •

Exterior doors; Venetian blinds; Blinds on exterior blades; Building-integrated photovoltaics (BIPV); Solar monitoring and reflection systems.

Several other authors have studied algorithms for the operation of door frames, blinds and mobile shading devices to improve the luminosity, thermal comfort and energy efficiency of buildings (Bourgeois et al. 2006; Lee et al. 1998; Van Den Wymelenberg 2012, Gomes et al. 2014). Generally, the study of the authors is focused on the choice of a methodology that defines the regulation of the devices in function of the interior and exterior parameters. At the moment, these devices are already in commercialization phase, and there is often the possibility of occupants overlapping their will by manually changing the state of the devices against the optimized programming defined by the manufacturer. The “Home for Life” residential building, based on the specifications of Active House, has automation at the level of the blinds and ventilation. Initially, the blinds were configured as a function of the interior temperature; however, the occupants manually adjusted the blinds as a function of the external radiation, and the algorithm was modified accordingly. On-demand ventilation is programmed as a function of CO2 , operating above 850 ppm, and relative humidity, operating above 60% indoor relative humidity and at top speed as the indoor relative humidity reaches 80% (Foldbjerg et al. 2011). This housing is based on the ventilation system “on demand”, and its study has already been solved and in commercial product phase, with a wide distribution. The ventilation can be promoted by mechanical extraction systems or by the simple opening of windows, functioning naturally as a function of wind. There are

14

2 State of the Art

devices that automatically open to close the windows. These systems typically operate in a reduced section of the window for security-related intrusion issues. Promoted ventilation usually occurs at night in order to take advantage of the temperature gradient to cool the house (Ochoa and Capeluto 2009). It is now possible to place an electromagnetic system on the windows. This type of technology allows a change in the solar reflection of the glass. The great advantage of this technology is to be able to automatically control the amount of energy and lighting that goes through the window (Clear et al. 2006; Jelle 2013). The system controlling these variables can be programmed according to parameters of the external environment such as temperature and solar radiation or internal illuminance, for example. In the work done by Lee et al. (2012), the author studied the application of this technology having found an energy saving of 48%. The double skin facades (DSF) system is a solution that can contribute to improving the energy efficiency and thermal comfort of buildings. However, this is only possible with the conjugation of shading devices within the air cavity and ventilation thereof. In this work, the ventilation system and the shading devices are usually associated with this constructive solution (Marques da Silva et al. 2015; Zhou and Chen 2010; Gratia and De Herde 2004; Manz 2003).

2.3 Occupant Behaviour in Residential Buildings As detailed in Sect. 2.2, IB can be completely automated and has a purely informative system that suggests occupants the most “intelligent” actions to take or combine a set of automatic actions and alerts. Regardless of the type of functioning of an EI, the literature is consensual in stating that only true intelligence is achieved in buildings when they respect the requirements of their occupants, using the least amount of energy possible. In order for the requirements of the occupants to be respected, it is first necessary to know them. The study of the occupant’s behaviour of residential buildings is therefore considered as an area of great importance, and the results will be the basis of possible modes of operation of an EI. It should be noted that there are parameters that any IB must respect and that they are not a conscious necessity of the occupants but are part of a set of requirements that contribute to the improvement of their health (indoor air quality). On the other hand, as reported in ZeroCarbonHub (2015) and Delzendeh et al. (2017), there is a gap between the performance simulated in the design phase and the post-occupation phase. This incongruence is due to different motifs, and two stages can be identified that are at their origin (Fig. 2.2). There is a gap between what the designers thought and what was actually constructed. This is due to several types of conscious and unconscious changes to the initial project (Sinnott and Dyer 2011, Calì et al. 2016b). Considering the conditions of the actual construction, there is still another gap for the conditions of use, due, among others, to the impact of occupant behaviour (Nguyen and Aiello 2013; Delzendeh et al. 2017).

2.3 Occupant Behaviour in Residential Buildings

15

Fig. 2.2 Gaps from the design phase to the post-occupation, adapted from Delzendeh et al. (2017)

According to the aforementioned, it reinforces the already mentioned in Chap. 1, indicating the main motivations for the study of this theme (Jia et al. 2017; Hong et al. 2017): • The need to optimize BAS/EMS/BMS, adapting them to the occupants’ habits; • The need to reduce the gap between the forecast at the design stage and the actual building performances. The heterogeneity of human behaviour creates extra difficulties for the occupant behaviour area of studies, because systematization and parameterization are difficult to achieve, at least in a global way. Thus, it is common in this area of study to use statistical databases and to divide the sampling by clusters (D’Oca and Hong 2015). However, according to some studies (Nguyen and Aiello 2013), the human behaviour could be characterized and standardized but a dynamic and adaptive functioning of the building systems is essential to achieve substantial energy savings combined with occupant satisfaction. According to Yan and Hong (2014), energy efficiency studies in buildings have focused essentially on improving systems, drawing attention away from the study of occupant behaviour. The same authors defined the in-depth study of the behaviour of the occupants indicating that the simple recognition of the influence of the actions of the occupants in the buildings is insufficient. The area of study of occupant behaviour in buildings has undergone some recent evolution (D’Oca et al. 2018) and has been very useful for the significant reduction of energy costs in several buildings, mainly services. In the residential sector, there are some studies developed although the practical application of the results of the investigation is still little spread, mainly with dynamic systems and adaptable to the occupants. In the following subchapters, we sought to separate the information collected by the bibliography into four areas in order to answer the following questions: • • • •

How to detect occupant actions; What conditions the behaviour of the occupants; What are the impacts of occupant behaviour; How to model the behaviour of occupants.

16

2 State of the Art

2.3.1 Detection of the Actions of the Occupants The identification of the main activities of occupants in a dwelling is generally related to their impacts, but may have different motivations (Nguyen and Aiello 2013). In this way, it is difficult to define the main actions carried out by the occupants in a generalized way. However, according to the state of the art raised by the authors Stazi et al. (2017b), in residential buildings, the main actions aimed at scientific studies were as follows: • Operation of windows; • Regulation of air conditioning systems. In another perspective, the authors Johansson et al. (2010) studied the main activities with steam production in a dwelling. This study ranks the five main actions as sources of water vapour production and are decreasingly the following: nonventilated drying of clothes; human respiration; take a shower; food confection (dinner) and the existence of aquariums. Hendron and Engebrecht (2010) analysed the actions related to water consumption such as bathtub use, shower, dishwasher, washing machine, lighting and general appliances and lighting. However, the identification of the activity may not be sufficient, and it may be necessary to characterize it according to its duration, frequency, day and time of occurrence (Bonte et al. 2014). According to Yan et al. (2015), the measurement of occupant behaviour can be developed in three ways: • Monitoring processes—although they are the most used by the scientific community, there are several difficulties in the studies that use monitoring processes, and it is not easy to apply them in a residential context. The monitoring system has to be almost imperceptible to the occupants, operate well with few maintenance needs and accurately record values. These requirements are difficult to achieve simultaneously so some have to be devalued by others. In addition, there is always a risk that the data collected have corrupted periods and non-measurement periods, which requires a labour-intensive preprocessing of data. There is also the difficulty in choosing the ideal sensors for the study to be developed. There is a great variability of commercial sensor modules in the market (Messerve et al. 2010) and of sensors ready to be mounted in Arduíno (Ali et al. 2016). However, the best choice has to be taken into account—reliability, autonomy and response time. In a transversal way, the existence of at least 1 year of monitoring is considered important. • Conducting surveys and interviews—despite the greater accuracy of the observation studies compared with the surveys, some authors carried out studies using exclusively surveys, analysing significant amounts of data; e.g. Andersen et al. (2009) analysed 1569 surveys and Wilke et al. (2013) 7949. These authors tried to minimize possible deviations using samples of significant size. However, there are some reservations related to the accuracy of the inquiries. According to the authors Hnat et al. (2011) and Nguyen and Aiello (2013), long surveys tend to

2.3 Occupant Behaviour in Residential Buildings

17

produce non-negligible errors, which may be accurate or long lasting, but rarely both at the same time. In this sense, most of the authors defined the existence of monitoring campaigns as the most effective way of identifying the main actions of the occupants, and the surveys can be used in coordination with the monitoring system used (Jia et al. 2017). • Laboratory studies—in the area of thermal comfort, this kind of studies has been carried out in the past (Fanger 1970). Currently, its use is limited. The measurement performed according to the first process is non-intrusive, allowing monitoring of undistorted or conditioned behaviours. The rest may already suffer some deviations from the natural behaviour of the occupants. There have been several studies in the area of occupant behaviour based on monitoring studies and surveys, and the following are the most relevant aspects of this chapter. The E3SoHo project (Messerve et al. 2010), which aimed to reduce the consumption of social housing, summarizes the main sensors used in residential monitoring. In the scope of the project, the sensors available in the market were evaluated and some of its characteristics were analysed and compared: • Temperature—different sensors were evaluated: thermocouples, resistive sensors and thermistors. From this group, the resistive sensors were considered the most adequate because they were the most stable, accurate and with good measurement range, despite having a relatively high response time and the highest cost. • Relative humidity—the resistive and capacitive sensors were analysed in detail. The most suitable sensors are the capacitors because they have a wider measuring range and a higher temperature range, but also have a lower response time and better accuracy. • Carbon dioxide concentration—can be based on the use of infrared light or electrochemical sensors. Due to the difficult maintenance, electrochemical sensors are not usually used. • Lighting—there are two types, the photocell and the photodiode. The big difference between them is that the former is more sensitive, but slower. • Occupation—there are infrared, pressure, ultrasonic, cameras and other sensors combined. The various sensors can be compared essentially to the ability to detect movement, number of occupants, location of persons, detection of physical activity and price. The sensors capable of responding to all previous requirements are infrared cameras, video cameras and PIRs with 360° reading. Of these three types, the 360° PIRs are those that present better price, being therefore the most advisable. • Elements opening—three types were evaluated: the contact sensors (used to know the opening state of the spans), the pressure sensors (used to detect openings of gaps at a global level, but not at an individual level) and the reed switch sensors. • Consumption—can be used per fraction and therefore be located in the counters or the use of point sensors. In the case of point-of-use electricity sensors, they can measure intensity and voltage. The most recommended water and gas sensors are low-cost, low-maintenance ultrasonic flow metres.

18

2 State of the Art

The work carried out by the authors Ahmed et al. (2013) aimed at the recognition and standardization of occupant actions in residential buildings. For this, two cameras were used in each room, these being connected to a computer for data acquisition. Through the local binary pattern (LBP) method, it is possible to recognize facets and activities of building occupants through video images. The study by Bao et al. (2011) aimed to increase the available knowledge about the energy consumption of the occupants of the buildings in order to equal the balance between required energy and generated energy. The study focused on the use that the inhabitants give the equipment installed inside the dwellings. In this study, the “on” and “off” states of each machine were recorded as well as the particularities of the used operation program. According to the authors project Barbato et al. (2009), temperature, lighting and presence sensors were placed in each division. They used low-cost PIR sensors to detect presence. Wirelessly connected sensors were generally used. Due to the fact that each dwelling is relatively small, the ideal would have been to use only one networked system. However, due to economic and commercial issues, heterogeneous wireless networks were used and integrated into the information management program. The “CASAS-Sustain System” project (Chen et al. 2013) analysed two apartments on the campus of the Washington State University campus. Several sensors were used for monitoring. In order to evaluate the presence of the occupants, motion sensors were placed on the ceilings, with the installation of seven motion sensors in a room with a minimum of four. Temperature sensors, illuminance, hot water and cold water metres were also installed such as oven detection sensors. The work of the authors D’Oca et al. (2014b) monitored 15 dwellings in Copenhagen with natural ventilation. The values of indoor temperature, CO2 concentration and illuminance of living rooms and bedrooms were recorded for each dwelling. Regarding the external climate, outdoor temperature, relative humidity, global solar radiation, wind speed and number of hours of sun per day were collected from the closest meteorological stations. The opening or closing of the windows and the values placed in the programming of the heating system were also registered. The study developed by Kleiminger et al. (2014) used an algorithm based on the occupants’ smartphones to build occupancy profiles. The purpose of the algorithm was to infer when each user was at home. Thus, the algorithm calculates the occupancy schedule of each user. To do this, the algorithm uses the Wi-fi visible access point’s registers. Mobile phones have been programmed to cyclically scan the Wi-fi networks visible as such in the vicinity of the user. The input data for the algorithm consist of a list of these records, from which only the dates and the identifier (MAC address) of the mobile devices are used by the algorithm. The data collection for the development of occupancy profiles was developed over 18 months. The study developed by Iwashita and Akasaka (1997) was based on questionnaires to predict the operation of windows and doors and temperature conditioning equipment as well as their presence in the dwellings. The ventilation of the spaces was obtained by means of the tracer gas method and temperature and relative humidity through sensors in eight fractions of a housing complex in Japan.

2.3 Occupant Behaviour in Residential Buildings

19

According to the study by Bonte et al. (2014), the occupants were monitored in order to obtain eight different characteristics. There studied were the type of activity; its duration; the day of week it occurs; if it occurs in a weekday or on weekend; the time of the day; number of times each motion sensor is triggered during an activity; number of motion sensors that are activated in all rooms; number of sensors that were triggered. Bekö et al. (2010) monitored 500 Danish rooms using temperature, relative humidity and carbon dioxide sensors. Exterior weather data from a weather station within 20 km of the monitored dwellings were also used. The occupants were also requested to complete an inquiry related to the windows and doors conditions of the monitored rooms, as well as the night occupants of those rooms. Each house was monitored for a minimum of 48 h. Lu et al. (2010) used passive infrared (PIR) X10 wireless sensors, such as motion sensors and the detection of sleeping occupants, and magnetic contact readers (reed sensors) at the entrance door of the dwellings. The temperature and relative humidity of the monitored spaces were also evaluated. The choice of these sensors to detect the occupation of the occupants instead of others, such as the recording of the electrical consumption of all divisions, the use of portable sensors in the occupants or the use of video cameras, was taken for economic reasons and to avoid being intrusive in the house. The cost of the sensors used was around $5 each. In the eight houses monitored, a sensor was placed in each room, using a total of nine movement sensors plus one door sensor at the entrance, for a total cost of about $50 per house. The authors also say that the sensors used are very simple and therefore with less precision and reliability compared to more evolved ones. Despite these limitations, the authors considered that the sensors used were suitable for residential monitoring. In another work by some of the same authors (Whitehouse et al. 2012), as a future challenge the authors propose to improve the monitoring of occupancy by housing rooms, which will serve as a basis for the objective of individual climatic zoning of housing. According to the study developed by Möllers et al. (2014), wireless communication used in residential building monitoring systems is a potential source of insecurity and privacy breach. According to the authors, many available systems provide little or no security. In the two facilities that served as a case study, safety faults were detected with the possibility of all the information collected by the installed sensors to be accessible by third parties, among which the habits of the occupants and their presence were highlighted. In general, systems without any type of encryption provide a great deal of information to any experienced computer programmer. No prior knowledge about the installation or victim is required to perform this type of attack. The authors suggest that steps have to be taken to make these monitoring systems safer. Although encryption schemes are available and can be easily applied, it is considered that protection against traffic analysis attacks on wireless monitoring systems has yet to be further developed. The authors consider that the generation of fictitious traffic would be an effective and efficient way to increase the security of these systems. In order not to decrease the battery of the sensors distributed throughout the house, the authors suggest that the base station produces fictitious traffic since it is usually connected to electricity.

20

2 State of the Art

Ramos et al. (2015) used temperature and relative humidity measurement equipment, CO2 sensors and a ventilation door to measure the airtightness, the change of which was then evaluated from the point of view of the occupants’ behaviour. Occupancy data and occupants’ actions were obtained by survey responses. Guerra Santin (2011) carried out a work of association of patterns of behaviour and profiles of occupants of residential dwellings with the consumption of energy for heating, through the conducting of surveys. The surveys were conducted at two different times and planned according to the characteristics of the houses. The author states that the response rate was very low, with only 313 fractions from the 6000 selected. The surveys covered questions about the socio-demographic profile, lifestyle and characteristics of the fractions, correlating these characteristics with the energy consumption. The authors’ work presented by Wilke et al. (2013) was based on the questionnaire response completed between 1998 and 1999, corresponding to a total of 7949 families and 15,441 individuals. These questionnaires related to the report of the activities carried out by the occupants of the dwellings during 24 h with a precision of 10 min. In the study presented by Weng and Agarwal (2012), some types of sensors were compared to point out the best way to monitor the occupant behaviour in order to optimize the operation of an IB. PIR, reed switch and a prototype developed by the authors were compared to detect the presence of the occupants in the compartments. It was also used in experimental campaigns a prototype of measurement of consumption of the sockets. Yang et al. (2014) carried out a work of modelling occupant occupancy through parameters measured by indoor environment sensors in order to create an algorithm that controlled the operation of a HVAC system. Fifty prototypes of monitoring modules were created and were placed in 50 rooms of a student residence building. Each module consisted of an Arduino with the following sensors: illuminance; sound; movement; CO2 ; T; RH; infrared; reed switch sensor. Through these sensors, 11 variables were created that were recorded minute by minute. The results showed that CO2 , reed switch and illuminance sensors were the most useful for the purpose of the study. The work of D’Oca et al. (2014a) consisted in the creation of a system of global energy measurement for the residential system associated with an information system on real, historical and comparative consumption with similar families. The metering system consisted essentially of smart sockets that communicated via wireless to the central system. The system was tested on 31 dwellings. In the study developed by Calì et al. (2016a), the window openings of 67 housing fractions were analysed. For this purpose, monitoring modules with the following sensors were used: volatile organic compounds (VOC), CO2 concentration, illuminance, reed switch, T and RH. Authors Yao and Zhao (2017) used reed switch sensors to monitor the open/closed state of windows and eight environmental parameters, including temperature and relative humidity of indoor and outdoor air, indoor CO2 concentration, PM2.5 concentration outdoor and outdoor wind speed and direction.

2.3 Occupant Behaviour in Residential Buildings

21

The work developed by Hnat et al. (2011) gave rise to a guide for the monitoring of residential buildings. The work was based on 10 previous studies using over 1200 sensors. The type of sensors used varied with the following: movement; use of objects; opening of doors and windows; measuring the height of people passing through the doors; power metres; temperature and relative humidity; measurement of air intake in each room. The monitoring system used by Hnat et al. (2011) is constituted by several subsystems due to the different sensors used and different protocols used for their communication. Due to the reduced costs of the sensors and their ease of placement, motion sensors with a unit cost of around $5, bonded to walls with double-sided adhesive tape and X10 communication, were used in almost all studies. At the level of object utilization sensors, X10 sensors were initially used for their low cost. However, the fact that they are too intrusive led to the change by sensors with Z wave technology. The temperature and relative humidity sensors used were manufactured by Onset and La Crosse Weather Direct. The air intake sensors in the rooms were designed by the authors of the study as well as the actuators to facilitate the closing and the opening of the air intakes, remotely. The different subsystems used had different forms of energy storage and power supply, and the authors analysed the various faults that occurred during the monitoring time. The authors’ experience has allowed us to identify some of the main problems felt during the monitoring: • Electrical power: An unsustainable decision would lead most people to use electric current instead of batteries to power the sensors. However, these authors found greater reliability in the batteries. Household wall outlets have shown that they are often sparse and often punctually used by the occupants, causing the sensors to be switched off. In addition to three times greater losses in sensors connected to plugs than those using batteries, more interventions were required on the spot. The direct connection to electricity without using a plug has also been studied; however, this intervention is more intrusive and expensive. • Wireless indoors: Although homes are generally small, Wi-fi and Z wave are often not connected to all sensors because there are points where the network is weaker. The use of the electrical network as a means of communication between sensors can be used for some subsystems, depending on the communication protocols, although there are problems in connecting the sensors to the electrical network of the house, as explained in the previous point. • Occupants of the houses: the systems, being intrusive in a house, are subject to interaction of their occupants causing disturbances in the measurements. Children and animals are the occupants who most negatively influence monitoring. Sensors with bright screens and lights should be avoided. The wires of some sensors are also a source of disturbance to the house environment being often accidentally disturbed. • Visits to monitored sites: Although each sensor requires only one-minute operation, the placement of dozens of sensors can lead to many hours of visiting. Prolonged visits should be avoided for interfering with mealtimes, which makes the occupants displeased with the presence of the investigators. It is recommended

22

2 State of the Art

that the location of the sensors is prepared in the laboratory and that photographs are taken at the site so that sensor placement and maintenance operations are rapid. It is advisable to test the system in the laboratory so as not to carry out experiments at the monitoring site, and it is advisable to check the installation soon after leaving the site to be monitored. • Collaboration of the occupants: in five studies where people were asked to collaborate in the surveys, it was observed that it can obtain great precision results or surveys for long periods, but never both situations simultaneously. It is advisable to use redundant sensors in order to validate one another, especially those that depend on human action, such as the sensors that need to be transported by the occupants in order to identify their constant location in the house. • Aesthetic alteration caused by sensors: Most people accept that their home has sensors for research purposes, but does not accept them as decorative objects, especially for long periods of time. Beyond this situation, no trace can be left after monitoring. The use of an efficient and non-marking gluing system is essential. The lights and the noise of the equipment were also considered disturbances not accepted by the occupants. • Availability of home systems with large number of sensors: high correlations were found between the increase of the sensors with increasing errors and data omissions. The different manufacturers have been developing different communication protocols for home automation sensors, so their use in a home implies the use of several subsystems. If more economic systems are used, they are more thoughtful for non-professional use, so they become less reliable as network complexity increases. In order to avoid the successive occurrence of errors, Hnat et al. (2011) proposed a number of checks to be observed: • • • • • • •

Connection status with each system and subsystem; Status of the connection to the data collection system; Ask for a sensor response every time period; Stipulate a minimum communication frequency for all sensors; Ensure that the mean values measured by the sensors are acceptable; Ensure uniformity of sensor hours; Ensure that there is sufficient storage space.

Although there are more recent studies in the area of occupant behaviour using specific sensors to detect occupant actions (Jia et al. 2017), there are other studies (Stazi et al. 2017a; Guerra-Santin et al. 2016) in which the sensors used to determine the behaviour of the occupants are the same used to measure indoor environmental parameters (more precisely, temperature, relative humidity and concentration of CO2 ). There are obvious reasons for the use of this type of strategy since the use of specific sensors for the detection of occupant actions, besides costly, involves a complex installation, periodic maintenance actions and a diverse and therefore laborious data processing.

2.3 Occupant Behaviour in Residential Buildings

23

The temperature, relative humidity and CO2 concentration sensors, besides being useful in studies of thermal behaviour and air quality, are very important to the knowledge of the behaviour of the occupants (Candanedo et al. 2017). This extraction of time series knowledge is based on the principle of analysis of the variation of the consecutive parameters of a complete time series due to occupants’ activity (Szczurek et al. 2016). The authors Pereira and Ramos (2018) used a statistical tool to know the occupant activities of daily living through temperature, relative humidity and CO2 sensors. On the other hand, the use of temperature sensors, relative humidity has been trivialized by its low cost and the relevance of the information for both technicians and occupants. Although CO2 sensors are not widely used, they are already showing up on some non-professional air quality metres. Table 2.1 shows the measurement strategy of the behaviour of occupants in several studies.

2.3.2 Assessment of the Drivers of the Occupants The study of motivations/drivers was one of the four areas defined in the work of Hong et al. (2015) in the framework (DNAS) of occupant behaviour research. For this purpose, different studies have found different motivations for actions (Fabi et al. 2012; Stazi et al. 2017b). According to Peng et al. (2012), the motivations for human behaviour in buildings can be divided into three categories: • Temporal issues: actions based on a daily routine whose repeatability can be found in the same time periods. • Environmental issues: some actions are motivated by the logical perception of internal and external environmental parameters. • Random issues: Actions that are not regular do not follow logical principles or may follow unknown or complex motivations. On the other hand, Fabi et al. (2012) divided motives for occupant behaviour into five categories: • • • • •

Factors of the physical environment; Contextual factors; Psychological factors; Physiological factors; Social factors.

More recently Yan et al. (2017) considered the existence of four types of factors that influence the behaviour of the occupants: • • • •

Physiological; Individuals; Environmental impacts; Adjustments.

X

Hnat et al. (2011)

Wilke et al. (2013)

Weng and Agarwal (2012)

Guerra Santin (2011)

X

X

Bekö et al. (2010)

Bao et al. (2011)

X

X

X

X

X

X

Messerve et al. (2010)

X

Lu et al. (2010) X

Barbato et al. (2009)

X

RH

X

Exterior climate

X

X

X

X

X

X

X

X

X

X

CO2 Illuminance Occupancy Occupancy Occupancy Presence (PIR) (video (distance (mobile camera) metre) phone)

Sensors used in monitoring

T

Iwashita and X Akasaka (1997)

Authors

Table 2.1 Measurement strategy of the behaviour of occupants

X

X

Air flow

X

X

X

Reed switch

X

X

X

X

X

X

X

Surveys

(continued)

Consumptions Others

24 2 State of the Art

X

X

X

Ramos et al. (2015)

Calì et al. (2016a)

X

X

X

X

X

X

RH

X

Exterior climate

X

X

X

X

X

X

X

X

X

X

X

X

X

CO2 Illuminance Occupancy Occupancy Occupancy Presence (PIR) (video (distance (mobile camera) metre) phone)

Sensors used in monitoring

T

Yang et al. (2014)

Kleiminger et al. (2014)

D’Oca et al. (2014a)

D’Oca et al. (2014b)

Bonte et al. (2014)

Chen et al. (2013)

Ahmed et al. (2013)

Authors

Table 2.1 (continued) Air flow

X

X

X

Reed switch

X

X

X

Surveys

(continued)

VOC

Sound

Consumptions Others

2.3 Occupant Behaviour in Residential Buildings 25

X

X

X

X

Pereira and Ramos (2018)

X

X

Candanedo et al. (2017)

X

X

X

X

X

CO2 Illuminance Occupancy Occupancy Occupancy Presence (PIR) (video (distance (mobile camera) metre) phone)

Yao and Zhao (2017)

X

Exterior climate

X

X

RH

Sensors used in monitoring

T

Szczurek et al. (2016)

Guerra-Santin et al. (2016)

Authors

Table 2.1 (continued) Air flow

Reed switch

PM2.5

Consumptions Others

X

X

Surveys

26 2 State of the Art

2.3 Occupant Behaviour in Residential Buildings

27

According to the authors Wei et al. (2014), the existing bibliography points to several factors that condition the occupants’ behaviour regarding the heating of buildings, and the authors divided these factors into four groups: • Environmental—exterior climate and indoor relative humidity; • Factors related to the building and its systems—type of housing, age of the dwelling, size of the dwelling, type of rooms, house thermal insulation, type of heating system, type of power supply of the heating system and type of control system temperature; • Factors related to occupants—age, gender, culture/race, social class, family size, family income, type of house formerly occupied, owner/tenant, thermal sensation, IAQ perception, noise and health; • Other factors—time of day, day of the week, period of occupation, price of heating and awareness of energy use. According to Delzendeh et al. (2017), personal parameters (physiological and psychological) were considered in many studies (approximately 30% of the articles reviewed) as factors that influence energy consumption. The most recent studies suggest consideration not only of the individual and personal characteristics of the occupants, but also of the particular characteristics of their social context. However, only 10% of the articles reviewed focused on social and personal factors. Therefore, the authors believe that multidisciplinary approaches are necessary to combine sociopersonal parameters through psychological cognitive behavioural methods that can provide new knowledge to the domain. Studies of the motivations of occupants’ actions were mainly concerned with finding triggers that precipitated the occurrence of actions. Thus, the data recorded before the occurrence of the action were essentially evaluated (see Fig. 2.3).

Fig. 2.3 Data used to study the drivers for occupants’ actions

28

2 State of the Art

In this work, it was analysed several studies that focused on determining the drivers for the occurrence of the actions of occupants in residential buildings, which are described in the following paragraphs. According to Wood and Newborough (2007), the use of information screens in a smart home can be a way to motivate occupants to save energy in their homes. The authors identified that setting energy consumption targets as a good method to motivate occupants to reduce consumption. According to the authors Jain et al. (2013), occupants can change their behavioural patterns and consequently of consumption due to receiving eco-feedback from their consumptions. The study by Andersen et al. (2009) involved the carrying out of investigations in 933 dwellings in Denmark in the cooling station and 636 in the heating season. The first phase of the study served for the occupants to characterize the dwellings. The electronic questionnaires were followed where the occupants responded to the state of the house at that time on four levels: window opening status, sun protection, lighting and heating. The study also included the meteorological information of 25 Danish stations with parameters such as outdoor air temperature, wind velocity 10 m above ground and horizontal solar radiation. The study related the opening of windows with the outside temperature. The heating was also correlated with the outside temperature and with the presence of wood stoves. Lighting was correlated with solar radiation, outdoor temperature and light perception. The authors Guerra Santin and Itard (2010) carried out a study on the factors that would influence the behaviour of the occupants. It was found that the type of temperature control of buildings is the factor that has the greatest influence on occupant behaviour. The ventilation system of the building also showed influence on the occupants’ behaviour. The authors concluded that homes with manual heating and ventilation systems tend to have behaviours that lead to a reduction in energy consumption compared to dwellings with programmable thermostats and balanced ventilation. A specific pattern of consumption of the elderly was also found, and this group tends to heat the houses more and ventilate less. For homes with children, they tend to be less ventilated. The study developed by Guerra Santin (2011) attempted to find correlations between the various variables of occupant behaviour with energy consumption. The author was unable to find clear correlations between energy consumption and behavioural patterns and household characteristics. However, clear relationships have been established between the characteristics of the household and the behaviour of the occupants. The study conducted by the authors Fabi et al. (2012) indicated that there is some contradiction between the drivers found for the operation of the windows. In this way, the authors distinguished the motivations that can enter into the models to predict the operation of the windows. There are parameters that are generally drivers, parameters that do not have any influence on occupant actions and parameters whose influence depends on other factors. The inclusion of the latter parameters (wind speed, solar radiation and occupant age) in models must be well founded.

2.3 Occupant Behaviour in Residential Buildings

29

Authors Frontczak et al. (2012) obtained 645 responses to a questionnaire directed to dwellings in Denmark. This questionnaire, in addition to identifying the factors that occupants privilege in terms of comfort, identified the occupants’ common behaviours. The results showed that occupants prefer to manually operate electric lights, shading and window opening. In terms of temperature control, on the contrary, they prefer to be automatic. The work developed by Wilke et al. (2013) found some variables that can significantly influence occupant behaviour. These variables are the information of the country, the employment status of the household members and the income of the household. According to the authors D’Oca et al. (2014b), there was a need to create three types of occupant profiles in their study due to large behavioural differences. These types were defined according to the frequency of actions with the buildings: active profile, average profile and passive profile. It has been shown that some building occupants are very aware of their energy bills (heating and electricity) and tend to act in a way that reduces their billing rather than maintaining a high level of comfort. In contrast, occupants without so much financial constraint or awareness of the implications of their actions on the electric bill tend to interact freely with the heating control system and the opening of the windows in order to improve the comfort conditions in their homes. The results of this research have emphasized that occupant interactions in building systems are closely related to the quest for personal comfort, indoor air quality and overall energy performance of buildings. In the study by Luo et al. (2014), conducted in two Chinese cities, 139 apartments were monitored. This study indicated that occupants with the ability to control heating systems are more likely to feel neutral to the indoor temperature than those who do not have the means to operate the heating system. It was thus found that a lower degree of environmental control does not necessarily equate to a decrease in thermal expectations. On the contrary, the occupants with more control possibilities had less motivation to change their thermal environment. In order to lessen the dissatisfaction of the occupants of the buildings, it is essential that they are provided with means to change the interior conditions. According to the authors Wei et al. (2014), occupants have a large influence on the energy consumption due to their actions in the heating system in residential buildings, especially in the winter season. A survey of the motivations for this behaviour identified 27 different motivations in the literature. The work of Ramos et al. (2015) evaluated the behaviour of the occupants of fractions of a rehabilitated social neighbourhood at the level of window opening. It was concluded that the occupants, in winter, preferred not to open both the windows, thereby worsening the air quality inside, but achieving a reasonable thermal comfort. This study observed that the heating is punctual and very little used, due to economic reasons.

30

2 State of the Art

In the work carried out by Calì et al. (2016a), where 60 apartments from 3 similar buildings were studied, the most common motivation for the occupants to open the windows was the time of day, followed by CO2 concentration. The most common motivation for closing a window was the daily mean temperature, following the time of day. Yao and Zhao (2017) investigated the factors that influence the behaviour of the window opening of occupants in 19 residences in Beijing. The results indicated that outdoor air temperature was the most influential factor, followed by internal CO2 concentration, indoor air temperature, outdoor and indoor relative humidity, environmental concentrations of PM2.5 and external wind direction and wind speed. In the bibliographic research carried out by Stazi et al. (2017b), they found reference to several motivations for the following elements: • Windows—window use is driven by internal and external temperatures and CO2 concentration. The increase in use is mainly related to domestic activities. • Shades and curtains—they are rarely adjusted and almost always to achieve visual comfort. However, it is still difficult to assess which environmental parameter is the main factor for the operation by the occupants. • Air conditioning, thermostats, fans and doors—they are driven by indoor and outdoor temperatures. Only the use of conditioned air is also dependent on the time of day and day of the week. Taking into account the actions of the occupants and their main motivations in residential buildings, the scheme of Fig. 2.4 is constructed, indicating in a nonexhaustive way the interior and exterior parameters that functioned as motives for the occurrence of the actions and the categorization of the motivations according to the proposal of Peng et al. (2012).

2.3.3 Assessment of the Impacts of the Occupants As shown in Fig. 2.2, there is a discrepancy between the behaviour expected for a building and the actual measured when it is in operation and, thus, after the variable “occupant behaviour” has been added. This is due to the erroneous consideration of occupant impacts on the building’s global balances in building simulation programs (Delzendeh et al. 2017). For this reason, the number of studies on the impact of occupants on buildings is growing because of the need to narrow the gap between the prediction of consumption achieved through simulation programs and the actual performance of buildings (Jia et al. 2017; Hong et al. 2017; Delzendeh et al. 2017). According to Delzendeh et al. (2017), the passive and active behaviours of occupants with an impact on the energy consumption of buildings, including: window opening, use of solar shading and blinds, adjustment of the heating system, ventilation and air conditioning (HVAC), water use hot, etc., are not correctly considered in the simulation programs.

2.3 Occupant Behaviour in Residential Buildings

31

Fig. 2.4 Relationship between occupants’ actions and their drivers in residential buildings. 1 Stazi et al. (2017a, b), 2 Hong et al. (2017), 3 Andersen et al. (2009), 4 Alders (2017), 5 Johansson et al. (2010), 6 Hendron and Engebrecht (2010), 7 Fabi et al. (2012), 8 Guerra-Santin et al. (2016), 9 Ramos et al. (2015), 10 Luo et al. (2014), 11 Frontczak et al. (2012), 12 Bekö et al. (2010), 13 Howard-Reed et al. (2002), 14 Calì et al. (2016a), 15 Wei et al. (2014), 16 Yao and Zhao (2017), 17 Delzendeh et al. (2017); 18 Van Den Wymelenberg (2012), 19 D’Oca et al. (2014b), *few studies in residential buildings

The study of the impacts although not in the origin in the acronym DNAS in the work of Hong et al. (2015) is an important part of the initial “A”—actions since it represents the consequence of their existence. Regarding the more practical part of the study of occupant impacts, the data analysed to detect the impacts of the actions are obtained in the insights after their occurrence (see Fig. 2.5).

Fig. 2.5 Data used to study the impacts of occupants’ actions

32

2 State of the Art

The study of occupant impacts on residential buildings can be evaluated through several parameters. According to the state of the art collected, the main impacts studied by the bibliography are indicated: • Energy and water consumption; • Indoor environmental parameters (T and RH); • Air quality (CO2 and ACH). The work analysed on the study of occupant impacts on energy performance, hygrothermics and ventilation of residential buildings are indicated below. The work done by Gao and Whitehouse (2009) studied the behaviour of the occupants as a means to reduce the energy consumption of the dwellings. An air conditioning control system was created based on the occupants’ definition of the balance between energy and comfort and their presence in the house. The system can reduce energy costs by about 15% compared to a system with simple thermostat. The work carried out by Guerra Santin et al. (2009) aimed to quantify the effect that the occupant behaviour has on the heating consumption of residential buildings. The study showed that 42% of the energy consumption variations in Dutch buildings analysed was the responsibility of the constructive characteristics of the buildings, while only 7.1% corresponded to the occupants’ behaviour, 4.2% of the heating consumption. Another work carried out by the same authors found a strong correlation between the opening of the windows and the existence of ventilation grills with the energy consumption. The work done by Lu et al. (2010) whose experimental part was carried out in eight dwellings had an objective to establish a system that quickly detected the presence of the occupants and the moment in which they went to sleep as well as the activation of the air conditioning. Energy savings of about 28% were found on the level of energy spent on HVAC, not only without sacrificing comfort, but still improving it. However, the savings were reduced to 6.8% if no motion and port state sensors were placed in the building to allow better adaptation to the events of the house. The work done by Gram-Hanssen (2010), in Copenhagen houses, analysed in detail 5 families of an initial sample of 30 and obtained the energy expenses of each one of them and the average temperatures of each one of the houses for its 3 floors (basement, ground floor and first floor). It was possible to show how the behaviour of occupants in identical houses can result in three times higher energy consumption for heating without modifying greatly the conditions of comfort. This consumption value was obtained by comparing two families with continuous heating with two families with variable heating. However, three families with variable heating over time were also compared, with consumption references of about 50% and achieving higher temperatures on two floors and lower only on the upper floor. The E3SoHo project (Messerve et al. 2010; Brassier et al. 2014) was developed in 62 social housing houses located in 3 European cities. It aimed to develop an interface system with the occupants informing them about their energy expenditure and suggesting corrective measures. The houses were monitored before and after the installation of the interface systems. The results indicated that after the interface system was implemented, an energy saving of more than 20% was achieved in terms

2.3 Occupant Behaviour in Residential Buildings

33

of heating, hot water and other electricity in one of the target cities. The measures suggested by the system used were based on the characteristics of the buildings and the preferences of the occupants. Questionnaires were prepared for occupants where their needs and preferences were addressed. The main issues addressed in the surveys were related to the occupants’ attitude towards the environment, energy consumption habits and openness to energy savings by segment. The work developed by Whitehouse et al. (2012) has shown that the use of temperature control systems using sensors and operating in a reactive way according to the monitored parameters leads to energy consumption of about 10% more than the preprogrammed system. The authors, in order to improve this performance, have created two other programming systems, the “self-programming thermostat” and the “smart thermostat”. The first one proved to be useful when assessing an optimum operating time according to the history of the occupancy profile of a dwelling. This type of system worked especially well in cases where routines remained normally constant. The smart thermostat system, besides taking into account the history of occupancy profiles, had the ability to respond in real time to occupancy changes. This last system allowed savings of around 28%. The authors are developing a system that achieves even more savings through house zoning. Thus, the energy expended for heating and cooling would just be spent where there is actually use of the dwelling and not on the whole. This promising theme has proved difficult to achieve due to the difficulty of monitoring the presence of occupants in the rooms and the difficulty in compartmentalizing the heating due to the little habit that exists in closing the doors in the residential buildings. The study developed by Kleiminger et al. (2014) developed an algorithm for the control of the heating system based on predictive methods of occupancy profiles that allowed an energy saving between 6 and 17%, reducing the risk of loss of comfort for the occupants. The authors Yang et al. (2014) simulated with DesignBuilder and EnergyPlus programs the energy behaviour of a HVAC system during 3 months of a dormitory building. The authors compared the actual control of the HVAC with a model created by them based on the data obtained through on-site monitoring. They concluded that the operation of the HVAC system controlled by a model according to the actual needs of the occupants would lead to a reduction of 20% of gas and 18% of electricity. The work of D’Oca et al. (2014a) used combinations of occupant information media as real-time graphs and histological data on household consumption of each dwelling and other similar dwellings. The aim was to instil an energy-saving awareness by encouraging the competitiveness of similar housing and providing personalized tips as a means of saving energy. The study concluded that persuasive communication about energy savings can directly contribute to an average reduction in energy consumption between 18 and 57%. In the study developed by Calì et al. (2016b), a monitoring system was used that allowed a comparison between the actual and expected energy consumption of the buildings during 4 years. During the monitored period, those differences reached to a maximum gap of 287%. On average, the difference between the expected and the measured energy was of 117% in 2011, 107% in 2012, 41% in 2013 and 60% in 2014.

34

2 State of the Art

The occupant behaviour was identified as one of the causes of the differences found. Other causes were identified as problems in the workmanship and malfunction of the systems installed in the buildings. The authors Johansson et al. (2010) performed a very comprehensive state of the art on the effect of the occupants on the quality of the indoor environment in terms of humidity. In this study, the following sources of moisture production were addressed: bath/jacuzzi; shower; steam room; food preparation; dishwashing by hand; dishwashing machine; clothing treatment; ironing; wash the floor; human metabolism; metabolism of pets; aquariums; plants. The work also presented average values of the steam production of families according to the type of housing. These values are based on studies of several authors, and the average value for single-family housing ranges from 5.1 to 9.8 kg. The study developed a simulation work with 10,000 iterations. According to the simulations elaborated in this study, it was possible to present the five sources with the greatest contribution to the production of water vapour. The authors Hendron and Engebrecht (2010) in a benchmarking study presented profiles with the impacts of the use of residential buildings in terms of the use of bathtub, shower, dishwasher, washing machine, lighting and general appliances. According to the study by Kvisgaard and Collet (1986), the ACH value of the residences studied with natural ventilation was lower than the value in the residences with mechanical ventilation. They found relations between the occupants’ actions with the ACH of the natural ventilation houses, with occupants accounting for 63% of ACH rates. The study also concludes that the main actions with influence in the ACH of the housing were the operations of the windows and doors. The study presented by Iwashita and Akasaka (1997) found that occupant behaviour accounts for 87% of the air renewal rates. The air renewal rates were measured by tracer gas method in empty dwellings, and the same procedure was adopted when the dwellings were in use. In this study was not found a good correlation between energy consumption and ventilation rates. The authors Howard-Reed et al. (2002), who studied two homes in the east and west of the USA, found that window opening produced the highest increase in air change rate compared to the meteorological effects with closed windows (temperature difference between indoors/outdoors and wind). The authors considered that the operation of windows by the occupants was the main influence of the occupants in the ventilation of the dwellings. The authors also found that opening the windows in a few centimetres generally led to a rapid influx of air, but restricted to a small volume of the house. The study by Wallace et al. (2002) found a great influence of the opening of windows of the occupants in ACH. ACH ranged from a few tenths to the 4 h−1 peak, albeit only for short periods of time. The authors also found that temperature differences between the interior and exterior have a very good correlation with the ACH value. They have shown that the typical window opening effect influences the amount of air changes per hour in 1 h−1 . Window opening patterns were obtained by concluding that occupants keep the windows open half the time in summer and a reduced percentage in winter. This behaviour is one of the main factors in the increase

2.3 Occupant Behaviour in Residential Buildings

35

Fig. 2.6 Relationship between the occupants’ actions and their impacts on the interior environment, in residential buildings. 1 Kvisgaard and Collet (1986), 2 Iwashita and Akasaka (1997), 3 HowardReed et al. (2002), 4 Wallace et al. (2002), 5 Bekö et al. (2010), 6 Fabi et al. (2015), 7 Huang et al. (2014), 8 Mora et al. (2017), 9 Gao and Whitehouse (2009), 10 ZeroCarbonHub (2015), 11 Calì et al. (2016a), 12 Jang and Kang (2016), 13 Rijal et al. (2015), 14 Cheng and Steemers (2011), 15 de Meester et al. (2013), 16 Gill et al. (2010), 17 Steemers and Yun (2009), 18 Van Den Wymelenberg (2012), 19 Johansson et al. (2010), 20 Gram-Hanssen (2010), 21 Guerra Santin et al. (2009), 22 Andersen (2012), 23 Lu et al. (2010), 24 Kleiminger et al. (2014), 25 D’Oca et al. (2014a), 26 Pereira et al. (2017), *no correlation was found, # few residential building studies

of renovations in the summer. The value of ACH in summer was three times higher than in winter. In a study conducted by Pereira et al. (2017), it was found that the occupants’ operation of the air grill extraction vents could vary the total ACH value up to 92% and the windows operation could contribute with a change of around 435%. According to the authors Ramos et al. (2015), it was possible to measure the impact of occupant behaviour on the ventilation and airtightness of buildings. They analysed data from 49 autonomous fractions of 2 social neighbourhoods, one rehabilitated and one not rehabilitated. The data analysed allowed us to conclude that the influence of the user caused a mean ACH50 change from 7.7 to 4.3 h−1 . It was also possible to observe a change in the average ACH throughout the year. ACH values ranged from 0.35 h−1 in December to 1.01 h−1 in August. Taking into account the exposed bibliography about the occupant actions’ impacts, a resume scheme is presented in Fig. 2.6.

2.3.4 Modelling the Occupant Behaviour The occupant behaviour models constitute an important tool for this area of study, contributing decisively to its two main objectives, previously indicated. In general, occupant behaviour prediction models are defined according to the drivers behind

36

2 State of the Art

occupants’ actions, but there may also be models of occupant behaviour in the form of type profiles, in which case the impacts of these occupant actions are used. The utility of occupant behavioural models is comprehensive, and these are used both in the design phase of the building and in the service phase. In order to ensure that the models are reliable to use in the design phase, the models must accurately predict the occupants’ behaviour. However, in this case, the data used to construct the models are based on data collected in other buildings without being the target project. This particularity requires that the model distinguishes the main characteristics of the occupants and finds the specific and non-extrapolated particularities (Wilke et al. 2013; D’Oca et al. 2014b; Yan et al. 2017). On the other hand, there are authors who use the models to define the triggers of certain actions of a BAS/EMS/BMS (Stazi et al. 2017c; D’Oca et al. 2014a). There are still more and more authors opting for probabilistic rather than deterministic models (Wilke et al. 2013; D’Oca et al. 2014b; Fabi et al. 2013), and the authors D’Oca et al. (2014b) indicate that the use of predictive models of occupant behaviour will bring advantages throughout the building cycle: • Project phase: predict the actual energy use of the buildings in the service phase. The improved simulation model will support decision making in the initial phase of the project; • Operational phase: using predictive models and algorithms on the behaviour of occupants built into BAS/EMS/BMS to provide indications on the ideal use through “intelligent” communication; • Construction: evaluating the impact of occupant behaviour on different types of construction technology solutions; • Building management: enabling the construction of control systems, use of appliances and mapping of comfort level; • Construction code and policy: contribute to the development of construction standards, quantifying the variation of energy saving of technologies related to the behaviour of the occupants. Behavioural models are usually focused on a single action, triggered by one or more environmental variables. However, in more recent studies there is an introduction to more complex approaches considering much behaviour, interaction between occupants and different lifestyles. The connection between behavioural models and simulation software is moving towards a co-simulation approach, in order to obtain more realistic results (Stazi et al. 2017b). According to Melfi et al. (2011), there are three dimensions to the modelling of buildings: temporal; spatial and occupational. This study indicates that the temporal dimension is concerned with the definition of the time in which the events will occur. The spatial dimension defines the place where these events take place and the occupational one focuses on the behaviour itself, in terms of the number of occupants, their activities, the type of occupants and their state, for example. Behavioural models are then developed to predict the likelihood of an occupant interacting with the building. Implicit models are used to understand the motives

2.3 Occupant Behaviour in Residential Buildings

37

behind occupant behaviour or to predict the state of a building system or the occurrence of an occupant’s action, based on an indicator variable. Explicit models are used to provide a personalized description or future prediction of the state of a building system or occupants’ actions, based on the actual, monitored behaviour of the occupant himself. In both cases, data mining and statistical models are used to obtain information about repetitive patterns of occupant behaviour and interactions with the building, providing information on occupant profiles (Hong et al. 2015). As referred by Stazi et al. (2017a), when it is intended to relate to the action more than one parameter, the use of logistic regressions is the most adequate. In the study by Yao and Zhao (2017), the predictive models of the opening behaviour of the windows of the occupants were established based on multivariate linear logistic regression. Previous studies (Calì et al. 2016a) also used logistic regressions to analyse binary window-state models to associate them with the motivations that caused their state change. In the same sense, D’Oca et al. (2014b) used logistic regressions in the probabilistic simulation models of the behaviour of the occupants in the operation of windows and thermostats. Rijal et al. (2015) used the logistic functions to predict the probability of operating windows and fans. Herkel et al. (2008) used the “logit function” proposed by Nicol (2001) to relate the opening of windows with the outside temperature. Previously, Hastie and Tibshirani (1986) had used logistic regressions to evaluate binary models in medicine. The logistic regressions present the same principle of the linear regressions, with the purpose of adjusting the mathematical model to a data series. However, in the logistic model can be used binary categorical variables while in linear regressions the variable has to be continuous. There is no need for a normal distribution although atypical values are to be avoided. In order to use the logistic model, it is necessary to test the multiple collinearity, which means that there could not be a strong correlation between the dependent variables. Once there are categorical variables, the use of the Pearson coefficient is discouraged, and the Spearman coefficient can be used for the same effect. On the other hand, the Pearson coefficient is discouraged for series with non-normal distribution whereas the Spearman coefficient can be used in these cases (Lehman et al. 2013). The value of the Spearman coefficient (ρ) is given by Eq. (2.1) according to Hollander et al. (2013): n di2 6 i1 (2.1) ρ 1− n2 − n where n is the number of pairs (xi, yi) and d i is the difference between the stations of xi and yi. The Spearman coefficient varies between −1 and 1. The extreme values correspond to perfect correlations and the value of 0 to the absence of any correlation. The positive sign indicates a positive correlation; i.e. the higher categories of a given variable correspond to the higher categories of the other variable with which the correlation is being tested.

38 Table 2.2 Qualification of the correlation coefficients

2 State of the Art Correlation coefficients

Interpretation

ABS (0.9–1.0)

Very high correlation

ABS (0.7–0.9)

High correlation

ABS (0.5–0.7)

Moderate correlation

ABS (0.3–0.5)

Low correlation

ABS (0.0–0.3)

Reliable correlation

According to Hinkle et al. (2003), the correlation coefficients could be interpreted as follows (Table 2.2). Correlation coefficients are useful for relating actions to motivations or impacts in the form of a prediction according to the learning obtained through a series of training data. The logistic regressions, since they are expressed as probability of occurrence taking values between 0 and 1, are easy to analyse and are more suitable to predict the behaviour of the model near its extremes (0 and 1) since linear regressions are not appropriate for models whose variables are not normally distributed (Gunay et al. 2013). According to Nicol (2001), the probability distribution of a particular event can be calculated by the logit function defined by Eq. (2.2). p(Ai)

exp(α + β1 X 1 + βn X n ) 1 + exp(α + β1 X 1 + βn X n )

(2.2)

where p(Ai) is the probability of the action i (Ai) occurring; α is a constant of the function relates with interception; β is a constant of the function relates to the slope; and X is the variable to which the probability of occurrence of the action is related. According to Yan et al. (2015), there are three main stochastic forms of occupant behaviour modelling: • Bernoulli processes—the probability of an event occurring is independent of the previous state (memoryless). These models are applicable to the whole building, being useful for large-scale energy modelling but not useful for individual descriptions, not being able to predict the time of occurrence of behaviours. • Discrete-time Markov chain—these models depend on the previous state to predict the probability. These models are widely used to predict individual behaviours and the motivations for these actions. Markov extensions use agent-based models. An interesting agent-based model for the characterization of occupant behaviour of buildings is the belief–desire–intention. • Survival analysis—it is widely used to predict the likelihood of lifetime of individuals and is used in other circumstances to predict the time it takes to occur in a particular event. Regarding this theme, the researches of several authors were analysed, not exhaustively, and some of the most relevant ones are indicated below.

2.3 Occupant Behaviour in Residential Buildings

39

In the Mozer project (2005), a concept of intelligent building adaptable to the occupants’ behaviour was created. The house collects information from occupants and creates usage profiles to anticipate the occupants’ needs. The AIM project (Barbato et al. 2009) aimed to develop technologies that allowed the creation of profiles and optimization of energy consumption patterns. In the project, wireless sensor networks were used to monitoring the physical parameters (such as light and temperature), as well as the presence of occupants in the house divisions. The data collected by the sensor system were the basis for plotting occupant behaviour. Several daily profiles were created by individualizing different parameters, such as temperature profile, occupancy profile and lighting profile. Clusters were then created by grouping similar profiles. Based on occupant behavioural profiles and real-time monitoring, a predictive (self-adaptive) algorithm was created to optimize energy consumption and cost, while ensuring the level of the required comfort. When occupants change their habits due to unpredictable events, the system is able to detect erroneous predictions based on the information obtained in real time and modify the behaviour of the system, accordingly. The work carried out by Hawarah et al. (2010) had the objective to create a predictive model for the consumption needs of the occupants of dwellings. This model is intended to avoid consumption peaks. The predictive model is based on the Bayesian network to diagnose and predict the behaviour of the occupants in the dwellings. The model was based on data collected from the use of home appliances from the REMODECE project (2015). User profiles have been created by clustering the houses according to the number of occupants and separating the summer from the winter, as well as the weekends and holidays. Lu et al. (2010) used the hidden Markov model to estimate the probability of occupants to be in three stages: away from home, at home and awake and at home but with all occupants sleeping. Occupancy patterns were considered essential to improve the functioning of the system. These patterns of occupancy were constructed based on the data collected in: this study in 8 dwellings; previous data obtained in the monitoring of 41 dwellings and data from two IB databases. In order to simulate and optimize the operation of the system, these authors used EnergyPlus, and the final result was validated in two buildings. The system created worked generally for the entire housing and could be further optimized if it was segmented by compartment. In the study developed by Dong et al. (2011), predictive models of occupation and meteorology were created. For the occupation models, two types of data analysis models were used: the Gaussian mixture models and the Markov model. Based on the forecasts, a climate control system was created, with savings of about 18% compared to a preprogrammed system, which was able to reduce the time when the ideal temperature was not reached with occupancy. The study by Antretter et al. (2011) was carried out in 17 residential buildings in Germany and aimed to define a predictive model of window opening. For this purpose, the dwellings were monitored for 2 years at the temperature and relative humidity of indoor and outdoor air, outside air velocity and window conditions measured in seconds of opening per hour.

40

2 State of the Art

The study by Bao et al. (2011) assumed that the use of home electrical equipment is initiated by occupants but triggers a series of subsequent actions that can be planned and anticipated. Based on the forecast of energy consumption, an autonomous system was created to manage the energy balance between needs and energy production and the coordination of decentralized power plants. Two types of methods were used for predictive energy consumption methods: one based on the first-order semi-Markov model and another one named “Model of the day type”. The latter assumes that there is a certain regularity of use of equipment and habits per day, and it is possible to create clusters of so-called type days. The ThinkHome project (Reinisch et al. 2011) is carried out in Vienna in five dwellings and attempted to develop the concept of a smart home in order to optimize occupant comfort while improving energy efficiency. One of the ways to reduce energy costs without losing thermal and visual comfort was to monitor the presence of the occupants in the various divisions of the house. In this project, automatons were used to simulate the activities of the occupants in the buildings. However, the behaviour of the occupants was also studied and the occupants were divided according to some characteristics (age, gender, etc.) and according to usage profiles. The external climate conditions were also considered fundamental to reach the project objective. One of the most important points of the project was the creation of house use profiles. To establish usage profiles, the information collected by sensors was used, stored and grouped by days. The data have then been used as inputs in clustering programs whose objective is to obtain a behavioural pattern that will serve as the basis for the programming of the later days. Results were compared between a system that connects the heating only by the presence of occupants and another with the variable of the occupants’ profile. Although the system that only works due to the presence of occupants has presented less energy costs, the system with the inclusion of occupant profiles significantly improves occupant comfort. The work of Wilke et al. (2013) develops models of residential occupants’ activities based on the probability of occurrence of these activities and their likely duration. The data obtained in questionnaires were grouped into subgroups to avoid considerable deviations. The chaining of the occupants’ tasks was also modelled based on the Markov chains. The results of the validation tests showed the effectiveness of this procedure in the prediction of task chaining and in the connection of individual characteristics with the probability of beginning and distribution and duration of activities. The objective of this model is to support the more precise representation of occupant behaviours related to energy consumption in simulation programs. The authors’ work Fabi et al. (2013) presented a methodology to model the behaviour of occupants in the context of real energy use and applied to a case study. The methodology, based on a medium-/long-term monitoring, aims to change to a probabilistic approach in the modelling of human behaviour related to the control of the interior environment. With this approach, it was possible to construct models based on monitored data obtaining distributions of probabilities of energy consumption and internal environmental quality depending on the behaviour of the occupants. The study conducted by the authors Ahmed et al. (2013) uses the SVM lights (Support Vector Machines) tool. This tool is a method commonly used to recognize

2.3 Occupant Behaviour in Residential Buildings

41

patterns or classify events, particularly tailored to work with large databases of text or images, although it also achieves good results with small databases. This tool was used to learn about occupants’ behaviours and to model these behaviours. In the case in question, 11 models were created. At a later stage, this tool classifies the collected images, corresponding to the actions of the occupants, and assigns to these actions their correspondence to the existing models. In this way, through a camera, it is possible to build a database with the actions of the occupants. The “CASAS-Sustain System” project (Chen et al. 2013) created a predictive model based on a previously trained algorithm. All the activities that the occupants perform have some relation with a measurable parameter, such as the time of day, the movement patterns of the occupants and the on/off status of various electrical devices. These activities are directly or indirectly associated with a number of electrical appliances and therefore have a unique standard of energy consumption. It should be noted that there are some appliances that are always connected, such as the heater (in winter), refrigerator, phone charger. Thus, activities were correlated with the use of energy. The project also compared three methods of training the algorithms: Bayesian Networks; support vector machines and artificial neural networks. In this project, accuracy levels of the forecasts were reached between 70 and 90%. A study carried out in Belgium (Aerts et al. 2013) created a behavioural profile of the occupants of residential buildings taking into account the data of 6400 people of 3474 dwellings. The profiles created aim to increase the knowledge of the behaviour of the occupants for the simulations of the buildings. The study allowed the creation of a profile for occupants at home and awake, sleeping and away from home. The work of the authors D’Oca et al. (2014b) detected a reduced accuracy in the dynamic simulation of residential buildings because there was no good representation of the occupants’ behaviour. To improve the simulations, the authors developed a study of occupant behaviour in residential buildings based on a probabilistic approach. Some physical and behavioural parameters of the occupants who constructed the statistical base were monitored. A statistical program was used to determine the physical parameters that most influenced the opening of the windows and the change of the set points of the heating system. A multiple logistic regression formula was then created to gauge the likelihood of each of the three occupant profiles opening the window or changing the set points of the heating system. The authors used the profiles of occupants created for dynamic simulations in three regions: a Mediterranean climate (Athens), a Nordic climate (Stockholm) and a continental climate (Frankfurt). The study developed by Kleiminger et al. (2014) used occupancy profiles to improve the performance of an intelligent heating control system. The authors used 45 dwellings for the construction of the algorithm of prediction of occupation. The algorithm developed achieved an accuracy of 80% in the prediction of the occupation. The research developed by Silva and Ghisi (2014) aimed to analyse and quantify the uncertainties in a numerical simulation of a Brazilian building type. The simulation program used was EnergyPlus. The results showed that there are uncertainties related to the behaviour of the occupants and uncertainties related to the physical parameters of the building. Regarding energy expenditure for heating, uncertainties

42

2 State of the Art

were found regarding the occupants’ behaviour of 36.5 and 38% for heating and cooling, respectively. The authors considered of great importance the quantification of this uncertainty and the determination of the reliability of the simulations. In the work developed, these parameters were only obtained through sensitivity analyses. The authors identified the following variables: number of occupants; occupants’ schedules and equipment power, such as those whose determination must be taken care of. According to Wei et al. (2014), 27 parameters were found to modelling the behaviour of the occupants. In a building performance simulation program, only a few of these factors are considered. However, the advantages of including these 27 parameters in the simulation program are not obvious since the simulation time would be much higher as well as the time needed to define them properly, despite the obvious accuracy advantages. The authors concluded that it would be important to define the main factors influencing the behaviour of buildings so that only these are included in the simulation in order to achieve the best relation of accuracy with the resources spent. One of the proposed future developments is the quantification of the most determinant factors. Yang et al. (2014) used the following machine learning classification algorithms that are the most used in the literature, for modelling the occupation: support vector machine, artificial neural network, naïve Bayesian, naïve Bayes network and decision trees. To implement these algorithms, they used the WEKA v. 3.7 from the University of Waikato, New Zealand. In the study developed by Calì et al. (2016a), the window openings of 67 housing fractions were analysed. According to this study, the use of logistic functions is adequate to determine the probability of occurrence of a binary event as a function of internal and/or external parameters.

References 2010/31/EU, Directive (2010) Energy Performance of Buildings Directive (EPBD) Aerts D, Minnen J, Glorieux I, Wouters I, Descamps F (2013) Discrete occupancy profiles from time-use data for user behaviour modelling in homes. Comunicação apresentada em Proceedings of BS 2013: 13th Conference of the International Building Performance Simulation Association Afonso MJRS (2007) Paradigmas Diferencial e Sistémico de Investigação da Inteligência Humana. Perspectivas sobre o lugar e o sentido do construto. Faculdade de Psicologia e de Ciências da Educação, Universidade de Lisboa Ahmed HS, Faouzi BM, Caelen J (2013) Detection and classification of the behaviour of people in an intelligent building by camera. Int J Smart Sens Intell Syst 6(4):1317–1342 Alders EE (2017) Adaptive heating, ventilation and solar shading for dwellings. Archit Sci Rev 60(3):150–166. https://www.scopus.com/inward/record.uri?eid=2-s2.0-85017288064&doi=1 0.1080%2f00038628.2017.1300132&partnerID=40&md5=072e013f9db16d450f941139fd8c 3297

References

43

Ali AS, Zanzinger Z, Debose D, Stephens B (2016) Open Source Building Science Sensors (OSBSS): a low-cost Arduino-based platform for long-term indoor environmental data collection. Build Environ 100:114–126. http://www.sciencedirect.com/science/article/pii/S036013231 6300476 Andersen RK (2012) The influence of occupants’ behaviour on energy consumption investigated in 290 identical dwellings and in 35 apartments. Comunicação apresentada em 10th International Conference on Healthy Buildings, em Brisbane, Australia Andersen RV, Toftum J, Andersen KK, Olesen BW (2009) Survey of occupant behaviour and control of indoor environment in Danish dwellings. Energy Build 41(1):11–16. http://www.sco pus.com/inward/record.url?eid=2-s2.0-55549145744&partnerID=40&md5=12083e0c10b1413 ca8e2084af4a54e26 Antretter F, Mayer C, Wellisch U (2011) An approach for a statistical model for the user behaviour regarding window ventilation in residential buildings. Comunicação apresentada em 12th Conference of International Building Performance Simulation Association, em Sydney Bao K, Allerding F, Schmeck H (2011) User behaviour prediction for energy management in smart homes. Comunicação apresentada em Proceedings—2011 8th International Conference on Fuzzy Systems and Knowledge Discovery, FSKD 2011 Barbato A, Borsani L, Capone A, Melzi S (2009) Home energy saving through a user profiling system based on wireless sensors. Comunicação apresentada em 1st ACM Workshop on Embedded Sensing Systems for Energy-Efficiency in Buildings, BUILDSYS 2009, in Conjunction with ACM SenSys 2009, em Berkeley, CA Bekö G, Lund T, Nors F, Toftum J, Clausen G (2010) Ventilation rates in the bedrooms of 500 Danish children. Build Environ 45(10):2289–2295. https://www.scopus.com/inward/record.uri? eid=2-s2.0-77953541402&partnerID=40&md5=bc95f8d1bf33da920c2a086b0631ced5 Bonte M, Thellier F, Lartigue B (2014) Impact of occupant’s actions on energy building performance and thermal sensation. Energy Build 76:219–227. http://www.scopus.com/inward/record.url?ei d=2-s2.0-84897426997&partnerID=40&md5=f62d5a6d1a44ad506a758736bc036d69 Bourgeois D, Reinhart C, Macdonald I (2006) Adding advanced behavioural models in whole building energy simulation: a study on the total energy impact of manual and automated lighting control. Energy Build 38(7):814–823. http://www.scopus.com/inward/record.url?eid=2-s2.0-33 646349491&partnerID=40&md5=e8a4456671a31a45ca73ba837e127e74 Brassier P, Adell G, Salmon N (2014) Energy consumption feedback to users: lessons learnt from pilot cases in social housing and tertiary buildings. Comunicação apresentada em BEHAVE2014—Behaviour and Energy Efficiency Conference Calì D, Andersen RK, Müller D, Olesen BW (2016a) Analysis of occupants’ behaviour related to the use of windows in German households. Build Environ 103:54–69. https://www.scopus.com/ inward/record.uri?eid=2-s2.0-84962835183&doi=10.1016%2fj.buildenv.2016.03.024&partnerI D=40&md5=ec72bdf520d8b049a6427e9a7ec813f0 Calì D, Osterhage T, Streblow R, Müller D (2016b) Energy performance gap in refurbished German dwellings: lesson learned from a field test. Energy Build 127:1146–1158. https://www.scopus.c om/inward/record.uri?eid=2-s2.0-84978245108&doi=10.1016%2fj.enbuild.2016.05.020&partn erID=40&md5=2d446a9ec96162a540538730e3fc9265 Candanedo LM, Feldheim V, Deramaix D (2017) A methodology based on Hidden Markov Models for occupancy detection and a case study in a low energy residential building”. Energy Build 148:327–341. https://www.scopus.com/inward/record.uri?eid=2-s2.0-85019956608&doi=10.10 16%2fj.enbuild.2017.05.031&partnerID=40&md5=98c3567d43ae0ed698febb0a46762efb Chen C, Cook DJ, Crandall AS (2013) The user side of sustainability: modeling behaviour and energy usage in the home. Pervasive Mobile Comput 9(1):161–175. http://www.scopus.com/in ward/record.url?eid=2-s2.0-84873124513&partnerID=40&md5=3bbea8f3b8c922f2d941f4ae9 36d56dd Cheng V, Steemers K (2011) Modelling domestic energy consumption at district scale: a tool to support national and local energy policies. Environ Model Softw 26(10):1186–1198. http://www. sciencedirect.com/science/article/pii/S1364815211001034

44

2 State of the Art

Clear RD, Inkarojrit V, Lee ES (2006) Subject responses to electrochromic windows. Energy Build 38(7):758–779. https://www.scopus.com/inward/record.uri?eid=2-s2.0-33646533950&pa rtnerID=40&md5=fb24026dab83f614e1f456db2178312b Clements-Croome D (2004) Intelligent buildings—design, management and operation. Thomas Telford de Meester T, Marique AF, De Herde A, Reiter S (2013) Impacts of occupant behaviours on residential heating consumption for detached houses in a temperate climate in the northern part of Europe. Energy Build 57(Suppl C):313–323. http://www.sciencedirect.com/science/article/pi i/S0378778812005944 Delzendeh E, Wu S, Lee A, Zhou Y (2017) The impact of occupants’ behaviours on building energy analysis: a research review. Renew Sustain Energy Rev 80:1061–1071. https://www.sco pus.com/inward/record.uri?eid=2-s2.0-85019968192&doi=10.1016%2fj.rser.2017.05.264&part nerID=40&md5=3e43566b591891a66d8aa563d5c3e98f D’Oca S, Hong T (2015) Occupancy schedules learning process through a data mining framework. Energy Build 88:395–408. https://www.scopus.com/inward/record.uri?eid=2-s2.0-8492026897 1&doi=10.1016%2fj.enbuild.2014.11.065&partnerID=40&md5=fad95ec2d0e44a338d97da08e 3ff309a D’Oca S, Corgnati SP, Buso T (2014a) Smart meters and energy savings in Italy: determining the effectiveness of persuasive communication in dwellings. Energy Res Soc Sci 3(C):131–142. https://www.scopus.com/inward/record.uri?eid=2-s2.0-84924024473&doi=10.1 016%2fj.erss.2014.07.015&partnerID=40&md5=3a869ef9a15edeac1e56c528470e4ecd D’Oca S, Fabi V, Corgnati SP, Andersen RK (2014b) Effect of thermostat and window opening occupant behaviour models on energy use in homes. Build Simul 7(6):683–694. http://www.sco pus.com/inward/record.url?eid=2-s2.0-84906315309&partnerID=40&md5=3bfa70c0f24b3cfa 9d27df2036e2b215 D’Oca S, Hong T, Langevin J (2018) The human dimensions of energy use in buildings: a review. Renew Sustain Energy Rev 81:731–742. https://www.scopus.com/inward/record.uri?eid=2-s2. 0-85027558816&doi=10.1016%2fj.rser.2017.08.019&partnerID=40&md5=d10b6408f890100 165480c42b1d31238 Dong B, Lam KP, Neuman CP (2011) Integrated building control based on occupant behaviour pattern detection and local weather forecasting. Comunicação apresentada em Proceedings of Building Simulation 2011: 12th Conference of International Building Performance Simulation Association Fabi V, Andersen RV, Corgnati S, Olesen BW (2012) Occupants’ window opening behaviour: a literature review of factors influencing occupant behaviour and models. Build Environ 58:188–198. ://WOS:000309332800018 Fabi V, Andersen RV, Corgnati SP, Olesen BW (2013) A methodology for modelling energy-related human behaviour: application to window opening behaviour in residential buildings. Build Simul 6(4):415–427. https://www.scopus.com/inward/record.uri?eid=2-s2.0-84881604918&doi=10.1 007%2fs12273-013-0119-6&partnerID=40&md5=2a460e6c66973f955ac5b954694ef7bb Fabi V, Sugliano M, Andersen RK, Corgnati SP (2015) Validation of occupants’ behaviour models for indoor quality parameter and energy consumption prediction. Procedia Eng 121(Suppl C):1805–1811. http://www.sciencedirect.com/science/article/pii/S1877705815029884 Fanger PO (1970) Thermal comfort—analysis and applications in environmental engineering Foldbjerg P, Rasmussen C, Asmussen T (2011). Thermal comfort in two European active houses: analysis of the effects of solar shading and ventilative cooling Frontczak M, Andersen RV, Wargocki P (2012) Questionnaire survey on factors influencing comfort with indoor environmental quality in Danish housing. Build Environ 50:56–64 Gao G, Whitehouse K (2009) The self-programming thermostat: optimizing setback schedules based on home occupancy patterns. Comunicação apresentada em 1st ACM Workshop on Embedded Sensing Systems for Energy-Efficiency in Buildings, BUILDSYS 2009, in Conjunction with ACM SenSys 2009, em Berkeley, CA

References

45

Gill ZM, Tierney MJ, Pegg IM, Allan N (2010) Low-energy dwellings: the contribution of behaviours to actual performance. Build Res Inf 38(5):491–508. http://www.scopus.com/inwar d/record.url?eid=2-s2.0-77956217823&partnerID=40&md5=98d327685ffb37dbe948e8ae84bf 21e7 Gomes MG, Santos AJ, Rodrigues AM (2014) Solar and visible optical properties of glazing systems with venetian blinds: numerical, experimental and blind control study. Build Environ 71:47–59. https://www.scopus.com/inward/record.uri?eid=2-s2.0-84885739342&partnerID=40 &md5=43c2dc457194bfe15794f42bcb78f99e Gram-Hanssen K (2010) Residential heat comfort practices: understanding users. Build Res Inf 38(2):175–186. http://www.scopus.com/inward/record.url?eid=2-s2.0-77649289979&partne rID=40&md5=fa2f02c5f3e4483a219d4245718b66c8 Gratia E, De Herde A (2004) Optimal operation of a south double-skin facade. Energy Build 36(1):41–60. https://www.scopus.com/inward/record.uri?eid=2-s2.0-0242335863&partnerID=4 0&md5=8f955d520008cf57a604788a9071e943 Guerra Santin O (2011) Behavioural patterns and user profiles related to energy consumption for heating. Energy Build 43(10):2662–2672. http://www.scopus.com/inward/record.url?eid=2-s2. 0-80054806317&partnerID=40&md5=ed716e58768c7d5d8fdae7a578f00fd9 Guerra Santin O, Itard L (2010) Occupants’ behaviour: determinants and effects on residential heating consumption. Build Res Inf 38(3):318–338. http://www.scopus.com/inward/record.url? eid=2-s2.0-77951591082&partnerID=40&md5=713f31486d565cb77a0a03f3feb00bd8 Guerra Santin O, Itard L, Visscher H (2009) The effect of occupancy and building characteristics on energy use for space and water heating in Dutch residential stock. Energy Build 41(11):1223–1232. http://www.scopus.com/inward/record.url?eid=2-s2.0-69249216516&partn erID=40&md5=b316031a3378c0ded482e6f861f6c07b Guerra-Santin O, Romero Herrera N, Cuerda E, Keyson D (2016) Mixed methods approach to determine occupants’ behaviour—analysis of two case studies. Energy Build 130:546–566. https://www.scopus.com/inward/record.uri?eid=2-s2.0-84985011810&doi=10.10 16%2fj.enbuild.2016.08.084&partnerID=40&md5=97151bbbe1da227d86fdcfbc8813cf3c Gunay HB, O’Brien W, Beausoleil-Morrison I (2013) A critical review of observation studies, modeling, and simulation of adaptive occupant behaviours in offices. Build Environ 70(Suppl C):31–47. http://www.sciencedirect.com/science/article/pii/S0360132313002187 Hastie T, Tibshirani R (1986) Generalized additive models. Stat Sci 1(3):297–318 Hawarah L, Ploix S, Jacomino M (2010) User behaviour prediction in energy consumption in housing using Bayesian networks. In: 10th International Conference on Artificial Intelligence and Soft Computing, ICAISC 2010. Zakopane. http://www.scopus.com/inward/record.url?eid= 2-s2.0-77955406185&partnerID=40&md5=1b3aa40acc49bf327e92ff12732e232b Hendron R, Engebrecht C (2010) Building America research benchmark definition Herkel S, Knapp U, Pfafferott J (2008) Towards a model of user behaviour regarding the manual control of windows in office buildings. Build Environ 43(4):588–600. http://www.sciencedirect. com/science/article/pii/S036013230600326X Hinkle DE, Wiersma W, Jurs SG (2003) Applied statistics for the behavioural sciences. In Mifflin H (ed), 5th edn. Malawai Med J, Boston Hnat TW, Srinivasan V, Lu J, Sookoor TI, Dawson R, Stankovic J, Whitehouse K (2011) The hitchhiker’s guide to successful residential sensing deployments. Comunicação apresentada em 9th ACM Conference on Embedded Networked Sensor Systems, SenSys 2011, em Seattle, WA Hollander M, Wolfe DA, Chicken E (2013) Nonparametric statistical methods Hong T, D’Oca S, Turner WJN, Taylor-Lange SC (2015) An ontology to represent energy-related occupant behaviour in buildings. Part I: Introduction to the DNAs framework. Build Environ 92:764–777. https://www.scopus.com/inward/record.uri?eid=2-s2.0-84929489501&doi=10.101 6%2fj.buildenv.2015.02.019&partnerID=40&md5=392f5ec7c1125675d279d62ea3d18e69

46

2 State of the Art

Hong T, Yan D, D’Oca S, Chen CF (2017) Ten questions concerning occupant behaviour in buildings: the big picture. Build Environ 114:518–530. https://www.scopus.com/inward/record. uri?eid=2-s2.0-85009508216&doi=10.1016%2fj.buildenv.2016.12.006&partnerID=40&md5=e 0d0a24fe1c51b759b4fb520f6405262 Howard-Reed C, Wallace LA, Ott WR (2002) The effect of opening windows on air change rates in two homes. J Air Waste Manag Assoc 52(2):147–159. https://www.scopus.com/inward/recor d.uri?eid=2-s2.0-0036164398&partnerID=40&md5=e957783389be5c94c1667f62a2db8840 Huang KL, Feng GH, Li HX, Yu S (2014) Opening window issue of residential buildings in winter in north China: a case study in Shenyang. Energy Build 84:567–574. ://WOS:000345182000056 IBG (2014) Intelligent Building Group. Accessed 15 May. http://www.ibgroup.org.uk/ Iwashita G, Akasaka H (1997) The effects of human behaviour on natural ventilation rate and indoor air environment in summer—a field study in southern Japan. Energy Build 25(3):195–205. http://www.scopus.com/inward/record.url?eid=2-s2.0-0031142616&part nerID=40&md5=b7f3e5e4921343f8be2eae346bb6d1a4 Jain RK, Gulbinas R, Taylor JE, Culligan PJ (2013) Can social influence drive energy savings? Detecting the impact of social influence on the energy consumption behaviour of networked users exposed to normative eco-feedback. Energy Build 66:119–127. http://www.scopus.com/inward/record.url?eid=2-s2.0-84882327709&partnerID=4 0&md5=efeeb6f210f9b50956f431a9d50eb45d Jang H, Kang J (2016) A stochastic model of integrating occupant behaviour into energy simulation with respect to actual energy consumption in high-rise apartment buildings. Energy Build 121(Suppl C):205–216. http://www.sciencedirect.com/science/article/pii/S037877881630192X Jelle BP (2013) Solar radiation glazing factors for window panes, glass structures and electrochromic windows in buildings—measurement and calculation. Sol Energy Mater Sol Cells 116:291–323. http://www.scopus.com/inward/record.url?eid=2-s2.0-84879319338&partnerID= 40&md5=0571fe9da86e8da50fae0e0a078552f4 Jia M, Srinivasan RS, Raheem AA (2017) From occupancy to occupant behaviour: an analytical survey of data acquisition technologies, modeling methodologies and simulation coupling mechanisms for building energy efficiency. Renew Sustain Energy Rev 68:525–540. https://www.scopus.com/inward/record.uri?eid=2-s2.0-84994000394&doi=10.101 6%2fj.rser.2016.10.011&partnerID=40&md5=d8bc76cffaa73b5f0878e8321d6edb89 Johansson P, Pallin S, Shahriari M (2010) Risk assessment model applied on building physics: statistical data acquisition and stochastic modeling of indoor moisture supply in Swedish multifamily dwellings Kleiminger W, Mattern F, Santini S (2014) Predicting household occupancy for smart heating control: a comparative performance analysis of state-of-the-art approaches. Energy Build 85:493–505. http://www.scopus.com/inward/record.url?eid=2-s2.0-84908324413&partnerID=4 0&md5=0f0b58bc0a2619bac90bee0052d41410 Kvisgaard B, Collet PF (1986) Occupants’ influence on air change in dwellings. Comunicação apresentada em 7th AIC Conference, em Stratford-upon-Avon, UK Lee ES, DiBartolomeo DL, Selkowitz SE (1998) Thermal and daylighting performance of an automated venetian blind and lighting system in a full-scale private office. Energy Build 29(1):47–63. https://www.scopus.com/inward/record.uri?eid=2-s2.0-0032305060&partnerID=4 0&md5=bce0eacef27acb8d267e9e3327869365 Lee ES, Claybaugh ES, Lafrance M (2012) End user impacts of automated electrochromic windows in a pilot retrofit application. Energy Build 47:267–284. http://www.scopus.com/inward/record. url?eid=2-s2.0-84857047798&partnerID=40&md5=555eb139cb44e044e518ee9b3a49d0a8 Lehman A, O’Rourke N, Hatcher L, Stepanski E (2013) JMP for basic univariate and multivariate statistics: methods for researchers and social scientists, 2nd edn. SAS

References

47

Lu J, Sookoor T, Srinivasan V, Gao G, Holben B, Stankovic J, Field E, Whitehouse K (2010) The smart thermostat: using occupancy sensors to save energy in homes. Comunicação apresentada em 8th ACM International Conference on Embedded Networked Sensor Systems, SenSys 2010, em Zurich Luo M, Cao B, Zhou X, Li M, Zhang J, Ouyang Q, Zhu Y (2014) Can personal control influence human thermal comfort? A field study in residential buildings in China in winter. Energy Build 72:411–418. http://www.scopus.com/inward/record.url?eid=2-s2.0-84893145375&partnerID=4 0&md5=dce1f1fce7dfc407ced0f5e6070f2bc9 Manz H (2003) Numerical simulation of heat transfer by natural convection in cavities of facade elements. Energy Build 35(3):305–311. https://www.scopus.com/inward/record.uri?eid=2-s2.00037333127&partnerID=40&md5=41c757f8774b8cb85555fe5d4f581e8b Marcus SJ (1983). The ‘intelligent’ buildings. Accessed 15 Apr. http://www.nytimes.com/1983/1 2/01/business/the-intelligent-buildings.html Marques da Silva F, Gomes MG, Rodrigues AM (2015) Measuring and estimating airflow in naturally ventilated double skin facades. Build Environ 87:292–301. https://www.scopus.com/ inward/record.uri?eid=2-s2.0-84923323321&partnerID=40&md5=639d3d31d999b28c1fa1dd2 2a039727a Melfi R, Rosenblum B, Nordman B, Christensen K (2011) Measuring building occupancy using existing network infrastructure. Comunicação apresentada em 2011 International Green Computing Conference and Workshops, IGCC 2011 Messerve T, Duszczyk K, Martins N, Dymarski P, Samset H, Salmon N, Decorme R, Ramiro M (2010) E3SoHo—ICT for energy efficiency in social housings Möllers F, Seitz S, Hellmann A, Sorge C (2014) Short paper: extrapolation and prediction of user behaviour from wireless home automation communication. Comunicação apresentada em WiSec 2014—Proceedings of the 7th ACM Conference on Security and Privacy in Wireless and Mobile Networks Mora D, Carpino C, de Simone M (2017) Energy consumption of residential buildings and occupancy profiles. A case study in Mediterranean climatic conditions. Energy Effi 1–25. https://www.scopus.com/inward/record.uri?eid=2-s2.0-85027015736&doi=10.1007%2fs 12053-017-9553-0&partnerID=40&md5=91abc93c5be4b7ac057bb78a2f8b6036 Mozer MC (2005) Lessons from an adaptive house. Smart environments: technologies, protocols, and applications, pp 273–294 Nguyen TA, Aiello M (2013) Energy intelligent buildings based on user activity: a survey. Energy Build 56:244–257. http://www.scopus.com/inward/record.url?eid=2-s2.0-84870676065&partne rID=40&md5=00792548dfd45be40bcfaef278342d51 Nicol JF (2001) Characterising occupant behaviour in buildings: towards a stochastic model of occupant use of windows, lights, blinds, heaters and fans. In: Proceedings of the 7th International IBPSA Conference (7), pp 1073–1078 Nikolaou T, Kolokotsa D, Stavrakakis G, Ghiaus C, Inard C, Zinzi M, Fasano G, Liao Z, Stherland G, Karatasou S, Santamouris M (2004) Intelligent building assessmengt methodology—smart accelerate project Ochoa CE, Capeluto IG (2009) Advice tool for early design stages of intelligent facades based on energy and visual comfort approach. Energy Build 41(5):480–488. https://www.scopus.com/in ward/record.uri?eid=2-s2.0-62649125162&partnerID=40&md5=390e11f354f3212e1ab57c9d0 3090f93 Peng C, Yan D, Wu R, Wang C, Zhou X, Jiang Y (2012) Quantitative description and simulation of human behaviour in residential buildings. Build Simul 5(2):85–94. https://doi.org/10.1007/s1 2273-011-0049-0 Pereira PF, Ramos NMM (2018) Detection of occupant actions in buildings through change point analysis of in-situ measurements. Energy Build 173:365–377. http://www.sciencedirect.com/sci ence/article/pii/S0378778818306066

48

2 State of the Art

Pereira PF, Ramos NMM, Almeida RMSF, Simões ML, Barreira E (2017) Occupant influence on residential ventilation patterns in mild climate conditions. Energy Procedia 132(Suppl C):837–842. http://www.sciencedirect.com/science/article/pii/S1876610217348166 Ramos NMM, Curado A, Almeida RMSF (2015) Analysis of user behaviour profiles and impact on the indoor environment in social housing of mild climate countries. Comunicação apresentada em 6th International Building Physics Conference, IBPC 2015, em Turim Reinisch C, Kofler MJ, Iglesias F, Kastner W (2011) ThinkHome energy efficiency in future smart homes. EURASIP J Embed Syst—Special issue on networked embedded systems for energy management and buildings 2011 REMODECE (2015) Residential monitoring to decrease energy use and carbon emissions in Europe. http://remodece.isr.uc.pt/ Rijal HB, Humphreys M, Nicol F (2015) Adaptive thermal comfort in Japanese houses during the summer season: behavioural adaptation and the effect of humidity. Buildings 5(3):1037–1054. https://www.scopus.com/inward/record.uri?eid=2-s2.0-84958774002&do i=10.3390%2fbuildings5031037&partnerID=40&md5=ea8a637300e924c58d8e743d79075548 Russel SJ, Norvig P (2003) Artificial intelligence: a modern approach Silva AS, Ghisi E (2014) Uncertainty analysis of user behaviour and physical parameters in residential building performance simulation. Energy Build 76:381–391. http://www.scopus.co m/inward/record.url?eid=2-s2.0-84897451117&partnerID=40&md5=02fe7b15dd9a1513f828b aaaf6881091 Sinnott D, Dyer M (2011) Airtightness of dwellings in Ireland: design, workmanship and control, em Salford Sinopoli J (2010) Smart building systems for architects, owners, and builders. Elsevier Inc. So AT, Chan WL (1999) Intelligent building systems. Kluwer Academic Publishers Stazi F, Naspi F, D’Orazio M (2017a) Modelling window status in school classrooms. Results from a case study in Italy. Build Environ 111:24–32. https://www.scopus.com/inward/record.uri?eid= 2-s2.0-84993929609&doi=10.1016%2fj.buildenv.2016.10.013&partnerID=40&md5=08f9785a f4151a7dad41e57e7e42eada Stazi F, Naspi F, D’Orazio M (2017b) A literature review on driving factors and contextual events influencing occupants’ behaviours in buildings. Build Environ 118:40–66. https://www.scopus. com/inward/record.uri?eid=2-s2.0-85015988086&doi=10.1016%2fj.buildenv.2017.03.021&pa rtnerID=40&md5=3bcee46a430f7e4cc70b09db5bf1b941 Stazi F, Naspi F, Ulpiani G, Di Perna C (2017c) Indoor air quality and thermal comfort optimization in classrooms developing an automatic system for windows opening and closing. Energy Build 139:732–746. https://www.scopus.com/inward/record.uri?eid=2-s2.0-85012281765&doi=10.10 16%2fj.enbuild.2017.01.017&partnerID=40&md5=9980fd3edf1e2a98e83e8ec33a40f76a Steemers K, Yun GY (2009) Household energy consumption: a study of the role of occupants. Build Res Inf 37(5–6):625–637. https://www.scopus.com/inward/record.uri?eid=2-s2.0-703496 45107&doi=10.1080%2f09613210903186661&partnerID=40&md5=698ce0e33dd3e30e0cf36 e647c612469 ˙ G, Dolega A (2016) Detection of Szczurek A, Maciejewska M, Wyłoma´nska A, Zimroz R, Zak occupancy profile based on carbon dioxide concentration pattern matching. Meas: J Int Meas Confederation 93:265–271. https://www.scopus.com/inward/record.uri?eid=2-s2.0-849909245 35&doi=10.1016%2fj.measurement.2016.07.036&partnerID=40&md5=3fe8d7507d0d399f34d a180b70c6b2fc Van Den Wymelenberg K (2012) Patterns of occupant interaction with window blinds: a literature review. Energy Build 51:165–176. http://www.scopus.com/inward/record.url?eid=2-s2.0-84861 832771&partnerID=40&md5=3a343e6e07ee8f12fb4c488f35c70806 Wallace LA, Emmerich SJ, Howard-Reed C (2002) Continuous measurements of air change rates in an occupied house for 1 year: the effect of temperature, wind, fans, and windows. J Expo Anal Environ Epidemiol 12(4):296–306. https://www.scopus.com/inward/record.uri?eid=2-s2.0-0036 068753&partnerID=40&md5=63e1c3a9fcb9ffa807a234b56bb935d0 Wang S (2010) Intelligent buildings and building automation. Taylor & Francis

References

49

Wei S, Jones R, de Wilde P (2014) Driving factors for occupant-controlled space heating in residential buildings. Energy Build 70:36–44. http://www.sciencedirect.com/science/article/pii/S03 7877881300683X Weng T, Agarwal Y (2012) From buildings to smart buildings-sensing and actuation to improve energy efficiency. IEEE Des Test Comput 29(4):36–44. https://www.scopus.com/inward/record. uri?eid=2-s2.0-84867821726&doi=10.1109%2fMDT.2012.2211855&partnerID=40&md5=33a 70c34a78418a31e07805576002c0e Whitehouse K, Ranjan J, Lu J, Sookoor T, Saadat M, Burke CM, Staengl G, Canfora A, Haj-Hariri H (2012) Towards occupancy-driven heating and cooling. IEEE Des Test Comput 29(4):17–25. http://www.scopus.com/inward/record.url?eid=2-s2.0-84867814988&partnerI D=40&md5=44c8ea621f834bfc087251ad745084da Wigginton M, Harris J (2002) Intelligent skins. Editado por Elsevier Wilke U, Haldi F, Scartezzini JL, Robinson D (2013) A bottom-up stochastic model to predict building occupants’ time-dependent activities. Build Environ 60:254–264. http://www.sciencedi rect.com/science/article/pii/S0360132312002867 Winston PH (1993) Artificial intelligence Wong JKW, Li H, Wang SW (2005a) Intelligent building research: a review. Autom Constr 14(1):143–159. http://www.scopus.com/inward/record.url?eid=2-s2.0-10444258169&partne rID=40&md5=b4a5e230237eabe800377af7074f1a37 Wong KC, So TPA, Leung YTA (2005b) Intelligent building index. Editado por Asian Institute of Intelligent Buildings, 3rd edn Wood G, Newborough M (2007) Energy-use information transfer for intelligent homes: enabling energy conservation with central and local displays. Energy Build 39(4):495–503. http://www.s copus.com/inward/record.url?eid=2-s2.0-33846939481&partnerID=40&md5=79b4ff51bbd76f e263023075d0e00c27 Wyckmans A (2005) Intelligent Building envelopes—architectural concept & applications for daylighting quality. Department of Architectural Design, History and Technology Norwegian University of Science and Technology Yan D, Hong T (2014) EBC Annex 66—definition and simulation of occupant behaviour in buildings Yan D, O’Brien W, Hong T, Feng X, Burak Gunay H, Tahmasebi F, Mahdavi A (2015) Occupant behaviour modeling for building performance simulation: current state and future challenges. Energy Build 107:264–278. https://www.scopus.com/inward/record.uri?eid=2-s2.0-849405132 01&partnerID=40&md5=3d626d60485cc9a7b34f92bca19b1a57 Yan D, Hong T, Dong B, Mahdavi A, D’Oca S, Gaetani I, Feng X (2017) IEA EBC Annex 66: definition and simulation of occupant behaviour in buildings. Energy Build 156:258–270. https://www.scopus.com/inward/record.uri?eid=2-s2.0-85030723429&doi=10.10 16%2fj.enbuild.2017.09.084&partnerID=40&md5=00cc7c4b1482297f6f9642fbf4ef979c Yang J, Peng H (2001) Decision support to the application of intelligent building technologies. Renew Energy 22(1):67–77. Accessed 1–28 Feb 1999. http://www.scopus.com/inward/record.u rl?eid=2-s2.0-0035216491&partnerID=40&md5=da74689df329970ef4ce3b81e30b9e67 Yang Z, Becerik-Gerber B, Li N, Orosz M (2014) A systematic approach to occupancy modeling in ambient sensor-rich buildings. Simulation 90(8):960–977. https://www.scopus.com/inward/r ecord.uri?eid=2-s2.0-84905979634&doi=10.1177%2f0037549713489918&partnerID=40&md 5=79ea79f902a925130a4de8eef1541b58 Yao M, Zhao B (2017) Window opening behaviour of occupants in residential buildings in Beijing. Build Environ 124:441–449. https://www.scopus.com/inward/record.uri?eid=2-s2.0-850283203 08&doi=10.1016%2fj.buildenv.2017.08.035&partnerID=40&md5=0c884bdfa4a7affffa6b395d 1b7912d2 ZeroCarbonHub (2015) Post-occupancy evaluation. Em Rowner Research Project Phase Two Zhou J, Chen Y (2010) A review on applying ventilated double-skin facade to buildings in hot-summer and cold-winter zone in China. Renew Sustain Energy Rev 14(4):1321–1328. https://www.scopus.com/inward/record.uri?eid=2-s2.0-76349088257&partne rID=40&md5=23b892436132b601214d31d06b42f758

Chapter 3

Conclusions

This literature review evidenced a sustained growth, over the last few years, concerning intelligent building studies and occupant behaviour in buildings. Considering the analysed works, it was possible to detect the two main objectives of this area of study: • Reduce the gap between the results of the numerical simulation programs and the real performances of the buildings; • Optimize building automated systems (BASs), energy management systems (EMSs) or building management systems (BMSs), adapting them to occupant habits taking into account ventilation and hygrothermal parameters. The two stated objectives relate to different facets of building life. The former will help to reduce the gap between the simulated numerical program and that actually monitored in the building. The second objective concerns the occupancy phase of the building and will serve to optimize the operation of an intelligent building. In this way, the literature considers that an intelligent building cannot be absent from the behaviour of its occupants and must comply with its main requirements and habits. On the contrary, we may also say that the genesis of the study of occupant behaviour is related to the allocation of intelligence to buildings. The second objective of this study, described above, is directly related to intelligent buildings in their operational phase. However, it is also mentioned in the literature that an intelligent building does not start only in the use phase, and all phases, design, construction and maintenance must have implemented this philosophy. In this way, the first objective for the study of the behaviour of the occupants contributes to an intelligent and more real design of the buildings, allowing a more efficient operation in the occupancy phase. It was also possible to verify the following facts of the state of the art: • Implementation of the IB concept at the level of commercial and service buildings is already at an intermediate stage of implementation. The main reason for its acceptance and scope is related to the type of use of these buildings. These buildings have an occupation in a well-defined period of time, and there are no significant changes in the occupancy profile. Complementarily, the occupation is very hetero© The Author(s), under exclusive license to Springer Nature Switzerland AG 2019 P. F. Pereira et al., Intelligent Residential Buildings and the Behaviour of the Occupants , SpringerBriefs in Applied Sciences and Technology, https://doi.org/10.1007/978-3-030-00160-5_3

51

52

•

•

•

•

•

•

3 Conclusions

geneous, so the strategy of centralizing systems management (BAS/EMS/BMS) turns out to be the most obvious; In residential buildings, their implementation is not yet frequent, due to the difficulty in responding to the specific requirements of the occupants of each dwelling. The IB concept is always related to the implementation of technology and automation in buildings that sometimes perform actions that are not accepted by the occupants; It is highlighted as one of the most important concepts for the implementation of an IB the need to fulfil the will of the occupants in a more efficient way against a “nonintelligent” building. This brings out the concept of intelligence associated with saving resources without neglecting comfort, air quality and other requirements of the occupants. There is also an agreement in the bibliography regarding the need to adopt an “intelligent” architecture that enhances the use of passive systems as the basic procedure for achieving an IB; As mentioned, an IB presupposes some control of the housing through its automation. It is considered that there is currently insufficient research work for the implementation of IB in housing developments in Portugal, mainly due to the lack of methodologies to apply in housing so that the BAS/EMS/BMS learn to recognize the actions of the occupants and interpret them. Although there are already a number of reasonable systems to install in homes, these systems are considered to be still far from the objectives of the IB. The systems in commercialization are mainly concerned with the remote control of the equipment and the consumption of energy of the dwelling; In the area of studies of occupant behaviour, a framework proposed by Hong et al. (2015), consisting of four subareas, the study of actions; motivations, needs and systems (DNAS—drivers; needs; actions; systems). However, the study of the behaviour of the occupants is not restricted to these four vectors; there are other data of extreme relevance that do not directly concern these areas such as the study of occupant impacts, discussed in this work; There were different studies in residential buildings where difficulties were mentioned that began with monitoring and the most correct system to adopt. Although there are many sensors to collect different kind of information, it is noticeable that they function in a separately way. For this, there is still the need to resort to a number of non-integrated systems that make the monitoring process complicated and susceptible to errors. It is therefore considered important to select well the monitoring system to adopt since it is the basis of any algorithm behind the BAS/EMS/BMS; Three types of studies have been found in function to the data collection strategies: using sensors that allow the direct detection of occupants’ actions (Bao et al. 2011; Calì et al. 2016; Jia et al. 2017; Yao and Zhao 2017); the use of sensors that allow the indirect detection of occupants’ actions through mathematical and statistical models (Zouba et al. 2009; Zhang et al. 2012; Ahmed et al. 2013; Candanedo et al. 2017; Pereira and Ramos 2018) and using sensors to measure indoor environmental parameters, using surveys to determine the occupants’ actions (Bekö et al. 2010; Wilke et al. 2013; Ramos et al. 2015; Guerra-Santin et al. 2016);

3 Conclusions

53

• The use of surveys, if necessary, should be used in short periods and as a complement to a monitoring system whose duration should be at least 1 year; • Different authors point out the need to perceive the occupants’ behaviour from the point of view of their preferences and needs as one of the main ways to reduce the consumption of buildings; • The occupant actions drivers indicate that different motivations were found for the same actions in residential buildings. For example, in the study by Kvisgaard and Collet (1986), no correlations were found between indoor temperature, outdoor temperature or number of occupants with window opening, and in the study by Bekö et al. (2010), no correlation was found between the average outside temperature and the opening of windows/doors. The review of the state of the art by Wei et al. (2014) and by Delzendeh et al. (2017) listed several motivations found in the bibliography, showing articles where the correlations between the action of the regulation of the heating and its motivations were not found although the majority of articles have evidenced it. This situation is explained by D’Oca et al. (2017) that emphasizes the need to opt for multidisciplinary approaches to truly understand the specificities of the human factor in buildings, removing the influence of climatic, cultural and socio-demographic factors characteristic and valid only for some regions as proposed by Mills and Schleich (2012); • The main impacts studied by the analysed bibliography were energy consumption; internal environmental parameters (temperature and humidity); indoor air quality (carbon dioxide, CO2 and hourly air changes, ACH) and water consumption. • Regarding the modelling of occupants’ behaviour, this is usually created according to the motivations of the occupants; • Although the energy habits of several countries for the residential sector (Eurostat 2000; INE and DGEG 2011) are well known, the specificities and heterogeneities of the behaviour of occupants of residential buildings does not allow this data to be considered valid for all dwellings (Wei et al. 2014; Delzendeh et al. 2017). In addition, there may be behavioural heterogeneities in the home itself due to different types of motivations that occupants have about the same type of action (Fabi et al. 2012; Stazi et al. 2017). In this way, it is considered appropriate to have a comprehensive monitoring strategy throughout the dwelling so that the actions of the occupants and their motivations and impacts can be determined individually and by each room. • Wrong consideration of the impacts of occupants on the overall balance of the building is one of the main points for a high gap between the numerical simulation programs and those actually monitored in the buildings in service. Passive and active behaviour of occupants with an impact on the energy consumption of buildings, including window opening, use of solar shading and shutters, adjustment of HVAC system, use of hot water, etc., are not correctly considered in the simulation programs.

54

3 Conclusions

References Ahmed HS, Faouzi BM, Caelen J (2013) Detection and classification of the behaviour of people in an intelligent building by camera. Int J Smart Sens Intell Syst 6(4):1317–1342 Bao K, Allerding F, Schmeck H (2011) User behaviour prediction for energy management in smart homes. Comunicação apresentada em Proceedings—2011 8th International Conference on Fuzzy Systems and Knowledge Discovery, FSKD 2011 Bekö G, Lund T, Nors F, Toftum J, Clausen G (2010) Ventilation rates in the bedrooms of 500 Danish children. Build Environ 45(10):2289–2295. https://www.scopus.com/inward/record.uri? eid=2-s2.0-77953541402&partnerID=40&md5=bc95f8d1bf33da920c2a086b0631ced5 Calì D, Andersen RK, Müller D, Olesen BW (2016) Analysis of occupants’ behaviour related to the use of windows in German households. Build Environ 103:54–69. https://www.scopus.com/ inward/record.uri?eid=2-s2.0-84962835183&doi=10.1016%2fj.buildenv.2016.03.024&partnerI D=40&md5=ec72bdf520d8b049a6427e9a7ec813f0 Candanedo LM, Feldheim V, Deramaix D (2017) A methodology based on Hidden Markov Models for occupancy detection and a case study in a low energy residential building. Energy Build 148:327–341. https://www.scopus.com/inward/record.uri?eid=2-s2.0-85019956608&doi=10.10 16%2fj.enbuild.2017.05.031&partnerID=40&md5=98c3567d43ae0ed698febb0a46762efb Delzendeh E, Wu S, Lee A, Zhou Y (2017) The impact of occupants’ behaviours on building energy analysis: a research review. Renew Sustain Energy Rev 80:1061–1071. https://www.sco pus.com/inward/record.uri?eid=2-s2.0-85019968192&doi=10.1016%2fj.rser.2017.05.264&part nerID=40&md5=3e43566b591891a66d8aa563d5c3e98f D’Oca S, Chen CF, Hong T, Belafi Z (2017) Synthesizing building physics with social psychology: an interdisciplinary framework for context and occupant behaviour in office buildings. Energy Res Soc Sci 34:240–251. https://www.scopus.com/inward/record.uri?eid=2-s2.0-85028714403&do i=10.1016%2fj.erss.2017.08.002&partnerID=40&md5=8d328df7fd94a78d74c697b8938cabe4 Eurostat (2000) Harmonized European Time User Survey—HETUS. editado por Eurostat. https:// www.h5.scb.se/tus/tus/default.htm Fabi V, Andersen RV, Corgnati S, Olesen BW (2012) Occupants’ window opening behaviour: a literature review of factors influencing occupant behaviour and models. Build Environ 58:188–198. ://WOS:000309332800018 Guerra-Santin O, Romero Herrera N, Cuerdae E, Keyson D (2016) Mixed methods approach to determine occupants’ behaviour—Analysis of two case studies. Energy Build 130:546–566. https://www.scopus.com/inward/record.uri?eid=2-s2.0-84985011810&doi=10.1016%2fj.enbui ld.2016.08.084&partnerID=40&md5=97151bbbe1da227d86fdcfbc8813cf3c Hong T, D’Oca S, Turner WJN, Taylor-Lange SC (2015) An ontology to represent energy-related occupant behaviour in buildings. Part I: Introduction to the DNAs framework. Build Environ 92:764–777. https://www.scopus.com/inward/record.uri?eid=2-s2.0-84929489501&doi=10.101 6%2fj.buildenv.2015.02.019&partnerID=40&md5=392f5ec7c1125675d279d62ea3d18e69 INE, Instituto Nacional de Estatistica, Direção Geral de Energia e Geologia DGEG (2011) Inquérito ao Consumo de Energia no Sector Doméstico 2010 Jia M, Srinivasan RS, Raheem AA (2017) From occupancy to occupant behaviour: an analytical survey of data acquisition technologies, modeling methodologies and simulation coupling mechanisms for building energy efficiency. Renew Sustain Energy Rev 68:525–540. https:// www.scopus.com/inward/record.uri?eid=2-s2.0-84994000394&doi=10.1016%2fj.rser.2016.10. 011&partnerID=40&md5=d8bc76cffaa73b5f0878e8321d6edb89 Kvisgaard B, Collet PF (1986) Occupants’ influence on air change in dwellings. Comunicação apresentada em 7th AIC Conference, em Stratford-upon-Avon, UK Mills B, Schleich J (2012) Residential energy-efficient technology adoption, energy conservation, knowledge, and attitudes: an analysis of European countries. Energy Policy 49:616–628. https:// www.scopus.com/inward/record.uri?eid=2-s2.0-84865587173&doi=10.1016%2fj.enpol.2012.0 7.008&partnerID=40&md5=02a062ebd806a554bcf5a36441ddf5ce

References

55

Pereira PF, Ramos NMM (2018) Detection of occupant actions in buildings through change point analysis of in-situ measurements. Energy Build 173:365–377. http://www.sciencedirect.com/sc ience/article/pii/S0378778818306066 Ramos NMM, Curado A, Almeida RMSF (2015) Analysis of user behaviour profiles and impact on the indoor environment in social housing of mild climate countries. Comunicação apresentada em 6th International Building Physics Conference, IBPC 2015, em Turim Stazi F, Naspi F, D’Orazio M (2017) A literature review on driving factors and contextual events influencing occupants’ behaviours in buildings. Build Environ 118:40–66. https://www.scopus. com/inward/record.uri?eid=2-s2.0-85015988086&doi=10.1016%2fj.buildenv.2017.03.021&pa rtnerID=40&md5=3bcee46a430f7e4cc70b09db5bf1b941 Wei S, Jones R, de Wilde P (2014) Driving factors for occupant-controlled space heating in residential buildings. Energy Build 70:36–44. http://www.sciencedirect.com/science/article/pii/ S037877881300683X Wilke U, Haldi F, Scartezzini JL, Robinson D (2013) A bottom-up stochastic model to predict building occupants’ time-dependent activities. Build Environ 60:254–264. http://www.scienced irect.com/science/article/pii/S0360132312002867 Yao M, Zhao B (2017) Window opening behaviour of occupants in residential buildings in Beijing. Build Environ 124:441–449. https://www.scopus.com/inward/record.uri?eid=2-s2.0-85028320 308&doi=10.1016%2fj.buildenv.2017.08.035&partnerID=40&md5=0c884bdfa4a7affffa6b395 d1b7912d2 Zhang S, McClean SI, Scotney BW (2012) Probabilistic learning from incomplete data for recognition of activities of daily living in smart homes. IEEE Trans Inf Technol Biomed 16(3):454–462. https://www.scopus.com/inward/record.uri?eid=2-s2.0-84860685669&doi=10.1109%2fTITB.2 012.2188534&partnerID=40&md5=e6e4656edbcc6474dc9ff7d3e9c5cb68 Zouba N, Bremond F, Thonnat M (2009) Multisensor fusion for monitoring elderly activities at home. Comunicação apresentada em 6th IEEE International Conference on Advanced Video and Signal Based Surveillance, AVSS 2009