Professional Documents
Culture Documents
Human–Building
Ecosystems
Edited by Yimin Zhu, Ph.D.
Khee Poh Lam, Ph.D.
Yong Tao, Ph.D.
SUSTAINABLE
HUMAN–BUILDING
ECOSYSTEMS
SELECTED PAPERS FROM THE FIRST INTERNATIONAL SYMPOSIUM
ON SUSTAINABLE HUMAN–BUILDING ECOSYSTEMS
SPONSORED BY
EDITED BY
Yimin Zhu, Ph.D.
Khee Poh Lam, Ph.D.
Yong Tao, Ph.D.
Any statements expressed in these materials are those of the individual authors and do not necessarily
represent the views of ASCE, which takes no responsibility for any statement made herein. No reference
made in this publication to any specific method, product, process, or service constitutes or implies an
endorsement, recommendation, or warranty thereof by ASCE. The materials are for general information
only and do not represent a standard of ASCE, nor are they intended as a reference in purchase
specifications, contracts, regulations, statutes, or any other legal document. ASCE makes no
representation or warranty of any kind, whether express or implied, concerning the accuracy,
completeness, suitability, or utility of any information, apparatus, product, or process discussed in this
publication, and assumes no liability therefor. The information contained in these materials should not be
used without first securing competent advice with respect to its suitability for any general or specific
application. Anyone utilizing such information assumes all liability arising from such use, including but
not limited to infringement of any patent or patents.
ASCE and American Society of Civil Engineers—Registered in U.S. Patent and Trademark Office.
Photocopies and permissions. Permission to photocopy or reproduce material from ASCE publications
can be requested by sending an e-mail to permissions@asce.org or by locating a title in ASCE's Civil
Engineering Database (http://cedb.asce.org) or ASCE Library (http://ascelibrary.org) and using the
“Permissions” link.
Preface
For decades, a significant amount of research effort has been directed towards
developing engineering, economic, and policy approaches to better design and
maintain energy efficient buildings. Until recently, many have realized that occupant
behavior is another missing piece of the equation. As the research community is
quickly growing, we need a venue where researchers can share cutting edge findings
in the integration of human behavioral, social and economic sciences with building
design, engineering and metrology for better understanding building energy
performance, environmental impacts and occupant comfort.
We received over 60 abstracts and 25 high quality papers have been accepted and
included in the proceedings. The abstracts and final papers were peer-reviewed. The
proceedings cover a wide range of topics, generally classified into three subject areas,
namely, occupant behavior modeling and analysis, thermal comfort prediction and
analysis, and innovative design, planning and policies for building energy efficiency.
The proceedings also include a keynote presentation paper on human ecology and
building science.
© ASCE
Sustainable Human–Building Ecosystems iv
Contents
Featured Paper
© ASCE
Sustainable Human–Building Ecosystems v
Integrated Project Delivery and Total Building Automation for the Nearly
Net-Zero-Energy Q1 ThyssenKrupp Headquarters............................................ 168
Thomas Spiegelhalter
© ASCE
Sustainable Human–Building Ecosystems vi
© ASCE
Sustainable Human–Building Ecosystems 1
Ardeshir Mahdavi
INTRODUCTION
© ASCE
Sustainable Human–Building Ecosystems 2
© ASCE
Sustainable Human–Building Ecosystems 3
BOUNDARY CONDITIONS
Motivation. We will address the immediate implications of the human factor for the
built environment (i.e., human requirements and behavior) in sections 4 and 5 of this
paper. But it is important that we consider a number of broader questions pertaining
to the environmental implications of social development and human activity
(Mahdavi, 2012): i) How important is the antecedent consideration of population
growth, lifestyle development, as well as agricultural and industrial production for
the effectiveness of sustainable building efforts? ii) How do the relative resource
needs and environmental loads associated with building activity compare to other
domains of human activity and production such as industry and transportation? iii)
To which extent can contextual factors such as urban planning decisions and
mobility solutions affect and constrain the energy and environmental performance of
individual buildings? iv) How should we account for the impact of user behavior
(including the rebound effect) on the energy and environmental performance of
buildings?
Population and life style. United nation's data (medium growth scenario) projecta for
2050 alone for India and China a combined population number of three billions.
Likewise, the population of Africa – slightly over 200 millions around 1950 – is
projected to approach two billion by 2050. The topic of population growth containment
appears thorny and difficult politically. But even if population growth is seen as an
inevitable and unalterable process, at least the implications for resources, environment,
as well as ecological and social systems should be frankly discussed, rather than
evaded. In human ecological parlance, the ecological valency of an ecosystem can
sustainably support only a finite number of people of a given ecological potency.
Transgressing that limit invariably results in an ecological degradation of the
environment.
The ecological strain resulting from population increase is aggravated by a
parallel process involving the improvement of living standards – at least for some
populations – around the world. For instance, the global primary energy consumption
of China is projected to increase – from the value in 2008 – about 230% to reach
roughly 200 exajoules by the year 2035 (EIA, 2012). Moreover, whereas in the last
twenty years per capita energy use has been stagnating (albeit at a very high level) in
countries such as United States and Germany, both per capita energy use and Gross
National Income (GNI) have been increasing in China, India, and Brazil (Databank,
2012). In fact, the Gross National Product (GNP) of China and India is projected to
increase within a period of 40 years (2009 – 2049) roughly by a factor of 6 and 9
respectively. Lest these assertions are misunderstood: Rise in people's standard of
living is necessary, crucial, and desirable socially and ethically. However, even
though not a necessity, per capita improvement in living standard is typically
mirrored in per capita increase in resource depletion and environmental impact. The
environmental ramifications of such developments can be easily exemplified.
Human ecologically speaking, rise in indicators such as GNP and GNI can be
interpreted as a population's increased ecological potency. A precarious implication
© ASCE
Sustainable Human–Building Ecosystems 4
of this rise is the population's higher capacity to rapidly exploit natural resources and
intervene in the working of natural systems, resulting in all too familiar negative
consequences. Thus, the combined ecological and environmental effects of
concurrent population growth and the rise of the so-called standard of living can
easily undo incremental efficiency improvements in the building sector.
Buildings, Industry, mobility. The relative share of buildings in energy use and
environmental impact, as compared to other areas human activity and production such
as industry, agriculture, and transportation is undoubtedly significant. In European
Union, the relative energy demand for the domains transportation, industry, and
buildings (residential and commercial sector) was in the year 2000, 31%, 28%, and
41% respectively (Janssen, 2004). Comparing buildings' share in resource depletion
with other areas serves here two purposes: At a global level, the investment costs and
efforts for efficiency improvement potential in each domain can be compared to those
in the other domains. Thus, public funds and financial incentives such as tax incentives
could target those areas where maximum energy efficiency improvement and
environmental impact reduction can be achieved. At a more detailed level, where
behavioral decisions of individuals may matter, building-related energy use and
environmental impact issues could be assessed considering individual energy use
profiles and life styles.
This latter point can be further illustrated if we consider the relative energy
allocation to various activities of middle-class individuals in a European country
(Mahdavi, 2010). It seems activities such as driving cars, travelling with airplanes, and
using electronic gadgets are not all insignificant as compared to energy requirements for
heating of buildings. As mentioned earlier, the so-called developing countries
increasingly adapt both production and consumption practices of the so-called
developed countries. A significant case in point in this regard pertains to the automobile
industry. It is expected that the global demand for automobiles will increase from
currently around 70 millions to around 110 in the year 2020. Thereby, the highest rate
of increase will occur in growing markets such as China. An even more irrational case
of energy and resource usage pertains to the world-wide production and deployment of
weapons. Expenditures for weapons dramatically increased in the years 2001 to 2011
(over 80% in USA and Russia: close to 190% in China, around 60% in Saudi Arabia
and India).
The decisive role of the human factor in the sustainability discourse is not
limited to population explosion and increasing affluence in some populations around
the world. A highly important – but insufficiently understood – variable concerns the
behavior of individuals and populations. Broadly speaking, people's behavioral
tendencies may in certain cases favor solutions and products that are disadvantageous
from the energy and environment point of view. For instance, in various analyses of
human mobility (see, for example, Knoflacher, 1996), people, confronted with multiple
options to embark on a trip, have been shown to display a tendency to favor the
reduction of their personal physical exertion. In other words, people may prefer
mobility options such as driving a car (even if they may require – for instance while
searching for a parking space – more time), rather than walking and biking. On the
other hand, the choice of a certain mobility medium also depends on the availability of
options. If a regional and urban setting provides an extensive infrastructure tailored for
© ASCE
Sustainable Human–Building Ecosystems 5
individual transport, it should not come as a surprise if cars would emerge as the
dominant mobility medium. In contrast, increased provision of well-designed and
convenient pedestrian routes and bike paths has shown to result in a correspondingly
higher rate of trips made on foot or on bike. Behavioral patterns with implications for
sustainable development can be influenced, to a certain degree, not only by
consciousness raising measures and information campaigns, but also – and perhaps
more effectively – via proper design and planning measures and strategies as well as
proper economic incentives.
Human behavior and the built environment. The ramifications of people's presence
and actions in buildings are not sufficiently documented. Occupants operate buildings'
control devices such as windows, shades, luminaires, radiators, and fans to bring about
desirable indoor environmental conditions. These control actions can significantly
influence buildings' performance (Mahdavi, 2011a, 2011b). A better understanding of
the patterns of human presence and control-oriented behavior can conceivably facilitate
and guide both technical responses such as occupancy-sensitive environmental control
systems and information campaigns toward improving buildings' energy efficiency and
environmental performance (see chapter 5 for a detailed treatment).
A further example for the relevance of human behavior in the sustainable
buildings discourse concerns the so-called rebound effect (Sorrell, 2007). This effect
refers to the paradoxical circumstance that, under certain conditions, energy
efficiency measures may end up increasing the energy use (see section 5.4).
Instances of rebound effect in the building domain underline the importance of the
human factor in energy efficiency potential assessments and projections. For the
expected efficiency effects of technical measures to materialize, behavioral and
economical boundary conditions must be taken into consideration. This also explains
why it may be a good idea to adjust energy prices in tandem with energy efficiency
improvements.
Summary reflections. The above discussion of the boundary conditions suggests that
there may be an extensive set of essential higher-level measures and actions involving
human populations and individuals that could be undertaken parallel with, if not prior to
measures focused on individual buildings: i) Population increase needs to be addressed
in earnest. Whether the focus is on a region, a country, a continent, or the entire planet,
the respective ecological valency or the available carrying capacity must be considered.
ii) In policy, in education, and all manners of social discourse, the distinction between
the standard (or quality) of living and purely economic measures such as GNP must be
taken seriously. When a society's entire economic and political system is exclusively
focused on narrowly defined monetary gains, it should not come as a surprise if genuine
sustainability concerns can be forgone in favor of superficial and ancillary agenda. iii)
The existing mobility paradigm needs to be reexamined. Individual energy and
emission intensive modes of motorized transportation need to be radically reduced. It is
essential to move away from the practice of designing and organizing cities around cars,
instead of around pedestrians. iv) Individuals and communities, especially in affluent
societies, need to made aware of the environmental consequences of their life style
choices in view of mobility, residency, diet, and recreational activities. v) Overarching
© ASCE
Sustainable Human–Building Ecosystems 6
© ASCE
Sustainable Human–Building Ecosystems 7
Thermal comfort models and standards. Insights gained based on the study of the
human body's thermoregulation have informed the attempts to identify relationships
between physiological factors and the experience of thermal comfort. For example,
indices such as the PMV (Predicted Mean Vote) and PPD (Predicted Percentage of
Dissatisfied) postulate that people would be likely to be in the state of thermal
comfort if their mean skin temperature and sweat secretion are within a certain range
(Fanger, 1972). PMV is a numerical index that expresses the statistical mean of
thermal comfort evaluations by a large group of people. The correlating PPD index is
a predictive measure of the percentage of thermally dissatisfied people in a specific
thermal environment. Based on studies conducted in climate chambers, optimal
thermal comfort conditions were suggested to correlate with personal factors
(metabolic rate, clothing) and environmental conditions (air and radiant
temperatures, humidity, air flow speed). The results of such studies have been
frequently structured in terms of equations and associated rules, tables, and graphic
means that can be used by designers and engineers to infer preferable indoor climate
conditions for people with a certain level of clothing and activity. Various standards
such as ISO 7730 (ISO 2005) and ASHRAE 55 (ASHRAE, 2004) specify building
categories based on PMV ranges. The PMV-based approach is meant to statistically
apply to large groups of people. Moreover it primarily pertains to the human body as
a whole. Further studies and associated formulas have addressed the local thermal
discomfort due to draught risk, radiative asymmetry, cold or warm floor surfaces,
and vertical temperature gradients (Fanger et al., 1988, 1985; Olesen et al., 1979;
Olesen, 2008, 2002).
The above approach to description and prediction of thermal comfort has both
merits and limitations. Amongst the virtues of the approach is the systematic way in
which in a number of personal (level of activity, thermal resistance of the clothing)
and environmental factors (air temperature and humidity, air flow speed, radiant
temperature) serve as the input information to predict thermal comfort level of
people in a room. However, the approach has also been faulted with regard to
intrinsic limitations and predictive performance. On the one hand, precise definition
of the model's input variables (e.g., the exact determination of people's activity or the
thermal resistance of their clothing) is not trivial and may constrain the applicability
of the model to practical situations. On the other hand, laboratory-based models may
not properly capture conditions in the field, where inhabitants are typically adapted
to their living and working environments. Moreover, the steady-state assumptions
underlying classical thermal comfort models do not apply to real – particularly free-
running – buildings. Field studies have documented, particularly in free-running
buildings, considerable deviations of PMV-based thermal comfort predictions from
actual comfort votes by the occupants (see, for example, de Dear and Brager, 2002).
In this context, human-ecologically relevant psychological factors such as people's
expectations as well as their behavioral adaptations have been suggested to play a
key explanatory role. People, when thermally uncomfortable, tend to react in such a
way as to reduce thermal constraint and thus restore thermal comfort (Humphreys
and Nicol, 1998; Auliciems, 1983). Such reflections have led to the conception of an
adaptive approach to thermal comfort definition and prediction (Nicol et al., 2012).
© ASCE
Sustainable Human–Building Ecosystems 8
© ASCE
Sustainable Human–Building Ecosystems 9
The relevance of people's behavior. To appreciate the critical role of the people factor
for building performance, consider a few simple questions (Mahdavi, 2011b) that
building designers are typically expected to answer: i) How much energy will be
needed to heat, cool, ventilate, and illuminate buildings? ii) What kinds of indoor
conditions concerning thermal comfort and air quality are to be maintained in
buildings? iii) What level of daylight can be expected in indoor environment under
dynamically changing outdoor illuminance levels? iv) Will the acoustical environment
in indoor spaces provide the necessary conditions for communication and task
performance? v) Can occupants be safely evacuated from buildings in case of an
emergency such as outbreak of fire?
Obviously, none of these questions can be reliably answered without
considering the role of the people living and working in buildings. People affect the
performance of buildings, due to their presence and their actions. Energy and thermal
performance of buildings is not only influenced by the people's presence as a source
of sensible and latent heat, but also due to their actions, including use of water,
operation of appliances, and manipulation of building control devices for heating,
cooling, ventilation, and lighting. User-based operation of luminaires and shading
devices in a room affect the resultant solar gains, light levels, and visual comfort
conditions. Presence of people in a room and the associated sound absorption
influences the acoustical performance of the room. Safety performance of a building
© ASCE
Sustainable Human–Building Ecosystems 10
Two kinds of effects. Broadly speaking, a useful distinction can be made between
"passive" and "active" effects of users and occupants on buildings' performance.
Passive effects of people on indoor conditions in buildings denote those effects caused
by the mere presence of people in the building. For instance, hygro-thermal conditions
and indoor air quality in architectural spaces are influenced by such passive people
effects: Depending on their activity, people release not only various quantities of
sensible and latent heat, but also water vapor, carbon dioxide, and other execrations and
odorous substances. Likewise, in the building and room acoustics domain, presence of
people in a space has an effect on the sound field via introduction of additional sound
absorption. To capture the passive effects of people's presence in buildings in the design
process, we typically rely on existing data such as occupancy load schedules derived
from measurement results of people's metabolic rates. This is as such a straight-forward
process, barring two possible complexities. Firstly, different levels of resolution are
conceivable regarding temporal and spatial distribution of such passive effects.
Secondly, the passive people effects such as heat emission may depend on the context,
e.g., thermal conditions in occupants' rooms. This interdependence would require the
concurrent consideration of the human agent and its immediate environment.
In most buildings, occupants operate control devices such as windows, shades,
luminaires, radiators, and fans to bring about desirable indoor environmental conditions.
We refer to these control actions as people's active effects. They have a significant
impact on buildings' hygro-thermal and visual performance. To predict and evaluate
buildings' performance we need good knowledge of such control-oriented user
behavior. General information about building type (residential, commercial) and
environmental systems (free-running, air-conditioned) as well as organizational and
administrative information (e.g., working hours) can only provide rough directions
regarding such active effects. More representative people presence and action models
require, however, extensive observational data based on empirical studies of occupancy
and control-oriented user behavior in a large number of buildings. Thereby, possible
relationships between control actions and environmental conditions inside and outside
buildings could provide the underlying basis for derivation of user behavior models.
© ASCE
Sustainable Human–Building Ecosystems 11
© ASCE
Sustainable Human–Building Ecosystems 12
probabilistic nature of most control-oriented control actions. Nor do they capture the
dynamism of actual processes and events in buildings, as stochastic models can – at
least in principle (Fritsch et al., 1990; Nicol, 2001).
The latter models have been used to generate time series of both occupancy
intervals and user control actions that "look" similar to actual (real) processes and
event sequences. Thus, if grounded in quantitatively sufficient and qualitatively
adequate empirical data, stochastic occupancy and control action models could
capture the general occupancy-triggered processes in a building. Such models can be
implemented in simulation applications in terms of autonomous agents with built-in
methods to generate behavioral patterns that appear realistic (Bourgeois, 2005;
Chang and Mahdavi, 2002; Liao et al., 2011). However, the promise of stochastic
occupancy needs to be qualified against both reliability and applicability concerns
(Mahdavi and Tahmasebi 2015).
An argument can be made for the utility of simple (code-base or descriptive)
occupancy-related simulation input assumptions in the design development phase,
where calculations can be used to obtain numeric values for a number of aggregate
performance indicators such as buildings' annual heating and cooling loads. Such
aggregate indicators can support at least two purposes: i) benchmark a specific
building design proposal against applicable codes, standards, and guidelines, or ii)
comparatively assess the likely performance of multiple design alternatives. Thereby,
concise inferences are expected concerning the quality of the proposed building
"hardware" vis-à-vis design variables pertaining to the building's envelope, massing,
orientation, shape, construction, etc. Naturally, this is done under "standardized"
conditions pertaining to external climate, which is typically represented in terms of a
standard weather file, and internal occupancy-related processes, which are typically
represented in terms of fixed, more or less detailed assumptions regarding internal
gains, ventilation rates, etc. Theoretically speaking, the use of a probabilistic
presence and user action models could represent a problem not only for code-based
compliance checking, but also for the performance analyses of design alternatives,
when the aim is to compare multiple alternative designs irrespective of variance in
contextual boundary conditions (weather) and occupancy.
A different circumstance arises, however, if we consider more elaborate
building design analysis scenarios, which require us to consider the implications of
uncertainties associated with occupancy processes in buildings. Differences in
occupancy patterns over time and location can be quite significant. Such differences
can be important especially in view of the variance of thermal loads or conditions in
various zones of a building. Information regarding temporal and zonal load
variations is critically important, for example when calculations provide essential
data for design and sizing of indoor climate control systems. Thus, rigid models of
user presence and behavior that ignore associated stochastic fluctuations and the
resulting uncertainties would be rather problematic, if the detailed configuration of a
building's mechanical equipment is the main concern: While dealing with the
requirement of providing sufficient heating and cooling capacity to different zones of
a building, the variability of required thermal loads cannot be captures via spatially
and temporally averaged occupancy assumptions.
It seems as though different approaches to representation of occupancy-
related processes in building design support may be appropriate given different
scenarios. If consideration of the implications of variance in model input
assumptions is evidently critical to a specific performance inquiry, then probabilistic
models of occupancy presence and control actions would be appropriate. On the
other hand, when the objective of a performance analysis inquiry is to benchmark
design proposals against applicable codes and standards or to parametrically compare
design alternatives, uncritical inclusion of random variations of boundary conditions
© ASCE
Sustainable Human–Building Ecosystems 13
A note on the rebound effect. As noted earlier (see section 3.4), studies in different
fields have shown that energy efficiency improvement measures may miss their targets,
sometimes considerably. Paradoxically, energy efficiency measures may even end up
increasing the energy use. This circumstance is often referred to as the "rebound effect".
It can be defined as the non-materialized fraction of the projected energy saving due to
an energy efficiency improvement measure. Related observations have been also
reported in the building domain, whereby projected energy performance of new
buildings and the energy saving potential of thermal retrofit measures on existing
buildings were found to be overly optimistic. The rebound effect is complex and can
have multiple roots. Thereby, one of the contributing factors may be attributed to
behaviorally relevant circumstances. For example, thermal retrofit measures on a
building can principally reduce the heating energy required to maintain certain thermal
conditions in that building (Housez et al. 2014). However, the reduction potential may
not be exploited in actuality, if occupants modify their behavior in a more energy-
intensive direction. For example, they may change the temperature settings for heating,
or they may ventilate spaces more frequently, or they may turn on the heating in more
rooms. Such behavioral phenomena may explain the results of a number of recent
studies, which documented a lower than expected energy efficiency improvement effect
following thermal retrofit measures pertaining to existing buildings. The rebound effect
may also involve the redirection of energy efficiency gains in one area to increased
consumption in another. For example, monetary benefits from building-related energy
conservation measures in terms of lower heating costs may be redirected toward
increased energy use in higher fuel usage for car driving.
Summary remarks. The importance of people's passive and active effects on building
performance (e.g., indoor environmental conditions, energy use) can be significant.
Accordingly, many recent and ongoing research efforts attempt to understand and
predict passive and active occupancy effects on building performance. Thereby,
physiological and psychological descriptions of occupancy as well as empirically-based
observational data provide the knowledge base. Specifically, long-term high-resolution
empirical data on people's presence and control-oriented actions in buildings can
support the generation of general patterns of user control behavior as a function of
indoor and outdoor environmental parameters such as temperature, air flow, air quality,
illuminance, and irradiance. These patterns can be expressed either in terms of
typologically differentiated aggregate occupancy and control action models or realized
in terms of emergent behavior of a society of computational agents with embedded
stochastic features. Future developments in this area are expected to facilitate detailed
computational models of environmental processes in buildings via comprehensive
multiple-coupled representations that dynamically capture the states of occupancy,
building, and context.
CONCLUSION
In this paper, we reviewed the role of people in the built environment from a
number of different vantage points. Thereby, we explored the potential of human
© ASCE
Sustainable Human–Building Ecosystems 14
REFERENCES
ASHRAE Standard 55 (2004). “Thermal environmental conditions for human
occupancy”. American Society of Heating, Refrigerating and Air-
Conditioning Engineers.
Auliciems, A.(1983) “Psychological criteria for global thermal zones of building
design.” International Journal of Biometeorology. 27 pp.69 – 86.
Bourgeois, D. (2005) “Detailed occupancy prediction, occupancy-sensing control
and advanced behavioral modeling within whole-building energy
simulation.” PhD Thesis, Université Laval, Quebec, Canada.
Boyce, P.(1980) “Observations of the manual switching of lighting.” Lighting
Research & Technology, 12:4 pp.195-205.
Chang, S. and Mahdavi, A.(2002) “A hybrid system for daylight responsive lighting
control.” Journal of the Illuminating Engineering Society, 31:1 pp.147- 157.
Databank World databank.(2012)
URL: http://databank.worldbank.org/ddp/home.do?Step= 1&id=4, (accessed
February.
de Dear, R.J. and Brager, G.S.(2002) “Thermal comfort in naturally ventilated
© ASCE
Sustainable Human–Building Ecosystems 15
© ASCE
Sustainable Human–Building Ecosystems 16
© ASCE
Sustainable Human–Building Ecosystems 17
22.
Olesen, B.W. (2002) “Radiant floor heating in theory and practice.” ASHRAE
Journal 44:7, pp.19 – 24.
Olesen, B.W., Schöler, M., Fanger, P.O. (1979) “Discomfort caused by vertical air
temperature.” in Fanger, P.O. and Valbjorn, O. (eds), Indoor Climate, Danish
Building Research Institute, Copenhagen, pp. 561 – 579.
Rea, M. S. (1984) “Window blind occlusion: a pilot study” Building and
Environment, 19:2, pp.133-137.
Reinhart, C. (2004) LIGHTSWITCH-2002: “A Model for Manual Control of Electric
Lighting and Blinds”, Solar Energy, 77, pp.15-28.
Schönpflug, W. (1981) „Acht Gründe für die Lästigkeit von Schallen und die
Lautheitsregel.“ From Akustik zwischen Physik und Psychologie. In
SCHICK, A. (Ed.), Akustik zwischen Physik und Psychologie. Stuttgart
(Klett-Cotta).
Sorrel, S. (2007) “The rebound effect: an assessment of the evidence for economy-
wide energy savings from improved energy efficiency.” UK Energy Research
Centre, Sussex Energy Group for Technology and Policy Assessment.
Uexküll, J. (1920) “Kompositionslehre der Natur.” (Edited by Thure von Uexküll).
Frankfurt am Main.
© ASCE
Sustainable Human–Building Ecosystems 18
1
Environment, Behavior, and the Built Environment Laboratory (EBBEL),
Department of Apparel, Textiles, and Interior Design, Kansas State University, 225
Justin Hall, Manhattan, KS 66506. E-mail: juliakday@ksu.edu
Abstract
© ASCE
Sustainable Human–Building Ecosystems 19
this one occupant ran the space heater every day during the colder months (October
through February)? This would equate to $213 ($1.42 x 30 days x 5 months).
Perhaps an additional $213 for an electric bill is still not significant for a large office
building. But, what if 10 people operated personal space heaters for five months of
the year ($2,130)? What about 50 people ($10,650)? And, what if ten of those 50
people forgot to turn the space heater off every day when they went home ($13,622)?
The numbers can quickly add up to a more significant total for a company.
In addition, there are other less obvious factors that may quickly increase
energy costs. A temperature sensor, used to control the heating and cooling in a
particular zone of the building, may be placed in the area where the occupant is
operating the space heater. This may throw off the temperature readings in the overall
space, therefore making other occupants cold. Not only can this lead to thermal
discomfort to other occupants, but it may contribute to even more energy use from
the building’s HVAC system as facility managers try to please occupants by turning
up the heating setpoints even higher to compensate. Thermal comfort in buildings is a
complex issue because people have different thermal preferences, and furthermore,
individual behaviors can impact others’ comfort and energy use outcomes.
At first glance, the example above seems simple: someone brought in a space
heater because they were cold. However, in a high-performance building with finely
tuned HVAC setpoints and aggressive energy reduction goals, these behaviors can be
detrimental. The logic behind the example illustrated above can also be applied to
many different types of high-performance strategies in a building such as natural
ventilation (a window left open in the winter leading to extraneous heating), or
daylighting (window blinds left open on the south side of the building on a summer
weekend, which may lead to excess heat gain); both examples lead to superfluous
building energy use that could have been mitigated through informed occupant
behaviors.
Ultimately, in a high-performance building, occupant behaviors and interactions
with the building can negatively or positively affect energy outcomes for a variety of
reasons. The examples above are overly simplified to illustrate a point, however,
these types of occupant interactions within buildings can become more difficult to
predict depending on the types of building strategies / controls, climate, access to
control, social dynamics, cultural factors, comfort, and personal preferences (Brager,
& de Dear, 2000; Cole & Brown, 2009; Goodwin, 2013; Guerin et al., 2011; Hanqing
et al., 2006; Humphreys, 2005; Luo et al. 2014). The goal of this paper is to unpack
some of these complicated issues and to identify important characteristics of occupant
behaviors, as they relate to building energy use, through the lens of a large mixed-
methods study. Specifically, this paper will focus on two of the research questions
from the study that were geared toward occupant behaviors: (1) Why do occupants
interact with the blinds, electric lighting or thermal controls? (2) For what reasons
do occupants choose not to interact with high performance building features?
METHODOLGY
© ASCE
Sustainable Human–Building Ecosystems 20
Gunderson, 2014). In short, the first phase of the study included interviews with
experts in the field, literature review, and examination of building databases, the
second phase of the study included a large-scale survey, and the third phase of the
study was comprised of semi-structured interviews and detailed examination of the
building energy use data, photographs, architectural drawings, and other building
specific data.
This paper will primarily focus on the rich qualitative data analyses within the
third phase. The semi-structured and open-ended interview responses illustrated why
and how occupants were behaving or acting in a certain way within their given
buildings. While each building was unique, collecting data across varying types of
climates and high performance buildings helped to strengthen the overall
generalizability of the qualitative results. Furthermore, these results, coupled with the
quantitative data from the second phase, helped to reinforce the overall research
design and validity (Creswell, 2009).
Participants. The participants for the qualitative phase of the study included
individuals that had completed the survey in the first phase. The interview
participants were selected through convenience sampling, which involves the
selection of easily accessible and willing participants (Plano Clark & Creswell,
2008). One of the last questions on the survey asked if respondents would be willing
to participate in a follow-up interview, and if so, to provide contact information. If a
person added contact information, then he or she was contacted for a follow up
interview. In addition to those who had completed the survey, architects, building
owners and facility managers were also contacted for building-specific interview
questions and requests.
Interviews. The most prominent method of data collection was through semi-
structured and open-ended interviews with occupants, owners and building managers
via telephone and email. Semi-structured interviews fall somewhere between fully
structured interviews, which are inflexible in nature, and unstructured interviews,
which are not typically preconceived before an interview. As such, questions were
developed before the interviews as a guide, but they remained flexible based on the
occupants’ knowledge of the building and the types of strategies used in the building.
This approach allowed for more flexibility during the interview, and it also allowed
for some responses to emerge that may not have with a more defined set of questions.
The majority of telephone interviews were taped and later transcribed. Extensive
notes were taken for the un-taped interviews. Interview questions varied widely and
were based on individual answers from the survey, or the questions were specific to
the building the occupant represented.
Documents, photographs and architectural drawings. Other means of data
collection included the acquisition of photographs, maps, architectural documents and
plans, and occupant training materials if available.
Establishing validity, reliability and credibility. Rigorous data analysis
techniques were used for all phases of this research, but it is important to point out
that the criteria for measuring validity and reliability differs in quantitative and
qualitative research. In quantitative research, validity means that participants’
assessments are meaningful indicators of the particular construct being measured
© ASCE
Sustainable Human–Building Ecosystems 21
(Creswell & Plano Clark, 2011); this definition differs in qualitative research. Four
respected methods were used to validate the qualitative findings to establish
credibility and rigor during the qualitative phase of the study: (1) triangulation –
comparing different types of information (transcripts, documents, survey responses)
(2) rich, thick description of the results; (3) identification of disconfirming evidence
to confirm the accuracy of the findings (Creswell & Plano Clark, 2011; Creswell,
2009); and (4) member checking – a method of providing summaries to participants
to see if the researcher’s interpretation of their conversation was “right” (Marshall, &
Rossman, 2011).
Data Analysis. The recorded interviews were sent to a transcriptionist. After the
transcribed interviews were received, the researcher reviewed the transcripts for
accuracy. After the transcripts were cleaned, they were divided into separate
interview questions before they were imported into the software program for coding.
Some of the questions asked during the interviews varied based on the occupants’
responses to the survey or what type of building strategies had been implemented in
their particular building. Both the open-ended responses from the survey and the
transcribed interviews were analyzed using open and thematic coding in NVivo 10.
Photographs, architectural drawings, and other collected documents were also coded
and grouped by climate type and building code. These additional materials were used
anecdotally during data analysis to help explain findings from either the survey or
interview responses.
RESULTS
© ASCE
Sustainable Human–Building Ecosystems 22
(2) RQ13: For what reasons do occupants choose not to interact with high
performance building features? This question could not be answered with any of the
survey results in the quantitative phase. However, there were many comments made
during the follow-up interviews that helped to explain why people chose not to
manipulate high performance building features. After the interviews were coded, four
themes emerged which clarified why people did not interact with the building or
controls. Example quotes are provided for each theme.
• Social concerns (occupants did not want to affect others) and/or the culture in the
office was not conducive to changing thermal or visual conditions
o “normally in my own house I would certainly go ahead and do it [change
the blinds]. But here, I’ll change my own position so it doesn’t affect 20
other people usually - you know, shift positions even though… If the sun’s
only in my eyes, you know, no big deal.”
• “Not my dime” in reference to why they do not care about saving energy
o “… it feels like if you had some individual control [of the thermostat] you
actually would end up with energy saved, because I’ve had my window
open wasting heat a number of times... I had a father who taught me not to
waste energy in the seventies, but the number of times that I have wasted
energy here… it’s because its not my dime, right? It’s because I’m
uncomfortable.”
• Occupants did not understand how to effectively control the features
o “I have a problem remembering which way to tilt the blinds so they
maximize daylight and reduce glare and heat gain so having a reference
for blind positions at the controls or access to a building operation
manual would be useful.”
• Lack of actual control or perceived control (or asked not to touch it by someone)
o “And if you want to close or open the blinds or window …well the
windows are annoying because it's automated…so you can’t control it.”
o “It is just stupid. We all have to stay completely still for 15 whole minutes
before the lights will turn off so that we can actually see outside. Once
they go off, no one moves. It is very annoying.”
DISCUSSION
© ASCE
Sustainable Human–Building Ecosystems 23
obviously incredibly inconvenient for occupants who needed night vision capabilities
to perform their job at the weather station. This is a reminder that designers must
understand the use and the intent of the building when implementing high
performance strategies into a building. Not only did occupants truly need control of
the lights to efficiently perform their job, but they were also highly unsatisfied and
frustrated with their environment. This example demonstrates the importance of an
integrated design approach (Brown & Cole, 2006), which takes all necessary use
factors into account during the programming and design phases.
Other reasons occupants chose not to interact with high performance building
features, which emerged from the open-ended survey questions and interview
responses, included the following:
• Social concerns (occupants did not want to affect others) and / or culture
in the office not conducive to changing thermal or visual conditions
• “Not my dime” in reference to why they don’t care about saving energy
• Occupants did not understand how to effectively control the features
• Lack of control or perceived control
Each of these emergent themes relates back to the literature: social influence
(Jain et al., 2013), lack of understanding (Day et al., 2012; Hadi & Halfhide, 2011),
and lack of control (Luo et al., 2014). Additionally, many of the reported motivators
for behaviors related to thermal comfort (or discomfort).
High-performance buildings integrate strategies, such as daylighting and natural
ventilation, which may lead to variable visual or thermal conditions. As such, it is
important that occupants have control of these strategies so visual and thermal
comfort can be maintained. It should also be noted that if occupants are expected to
interact with these building systems to maintain their visual and/or thermal comfort,
then it is of the utmost importance they understand how to control and alter their
conditions. To understand building controls, and to effectively alter conditions to
meet their needs, an occupant may require additional education within the context of
their particular building.
Furthermore, in high-performance buildings, where natural ventilation is often
used as a primary design strategy, the temperature may shift even more than those in
conventional, mechanically controlled buildings. Conventional buildings are
mechanically regulated and aim to provide thermal comfort to only 80% of
occupants, meaning that 20% will most likely be uncomfortable at some point during
the day (ASHRAE Standard 55, 2008). It is argued in the literature that occupants
may have to redefine this “acceptable” range for thermal comfort in high-
performance buildings (Cole & Brown, 2009; Kwok & Rajkovich, 2010) because (a)
it is more difficult to maintain constant temperatures in high-performance buildings,
and (b) space heating, ventilation, and air-conditioning utilize such a large amount of
energy in buildings that expanding our thermal comfort standards would reduce
greenhouse gas emissions and help to conserve energy (Kwok & Rajkovich, 2010).
Some studies have found that occupants may be more willing to tolerate wider
temperature ranges in naturally ventilated buildings when they are given the option of
control (i.e. opening windows themselves) (Cole & Brown, 2009; Humphreys, 2005;
Nicol & Humphreys, 2010).
© ASCE
Sustainable Human–Building Ecosystems 24
CONCLUSION
REFERENCES
Brager, G., & de Dear, R. (2000). A standard for natural ventilation. ASHRAE
Journal, 21-28.
Brown, G.Z, and Cole, J. (2006) Rethinking the design process, retrieved October 4,
2010 from http://www.betterbricks.com/graphics/assets/documents/IED
Brown, Z. B., Dowlatabadi, H., & Cole, R. J. (2009). Feedback and adaptive
behaviour in green buildings. Intelligent Buildings International, 1(4), 296–315.
Chen, H.M., Lin, C.W., Hsieh, S.H., Chao, H.F., Chen, C.S., Shiu, R.S., … Deng,
Y.C. (2012). Persuasive feedback model for inducing energy conservation
behaviors of building users based on interaction with a virtual object. Energy and
Buildings, 45, 106–115. doi:10.1016/j.enbuild.2011.10.029
Cole, R. J., & Brown, Z. (2009). Reconciling human and automated intelligence in
the provision of occupant comfort. Intelligent Buildings International, 1(1), 39–
55. doi:10.3763/inbi.2009.0007
Creswell, J. (2009). Research design: Qualitative, quantitative, and mixed methods
approaches. Thousand Oaks, CA: SAGE Publications, Inc.
Creswell, J., & Plano Clark, V. (2011). Designing and conducting mixed methods
research. (2nd ed.). Thousand Oaks, CA: SAGE Publications, Inc.
Day, J., Theodorson, J., & Van Den Wymelenberg, K. G. (2012). Understanding
controls, behaviors and satisfaction in the daylit perimeter office: A daylight
design case study. Journal of Interior Design, 31(1), 17–34.
© ASCE
Sustainable Human–Building Ecosystems 25
Day, J. & Gunderson, D.E. (2014). Understanding high performance buildings: The
link between occupant knowledge of passive design systems, corresponding
behaviors, occupant comfort and environmental satisfaction. Building and
Environment, 84, 114–124. doi:10.1016/j.buildenv.2014.11.003
Galasiu, A. D., & Veitch, J. A. (2006). Occupant preferences and satisfaction with the
luminous environment and control systems in daylit offices: A literature review.
Energy and Buildings, 38(7), 728–742. doi:10.1016/j.enbuild.2006.03.001
Goodwin, S. (2013, July 23). Behavior-based energy efficiency approaches around
the northwest (a table). [Excel spreadsheet]. Retrieved from
https://conduitnw.org/pages/file.aspx?rid=1653.
Guerin, D. A., Kim, H. Y., Brigham, J. K., Choi, S. M., & Scott, A. (2011). Thermal
comfort, indoor air quality and acoustics: A conceptual framework for predicting
occupant satisfaction in sustainable office buildings. International Journal of
Sustainable Design, 1(4), 348–360.
Hadi, M., & Halfhide, C. (2011). Green buildings: Understanding the role of end user
behaviour. Going Green: The Psychology of Sustainability in the Workplace, 31.
Hanqing, W., Chunhua, H., Zhiqiang, L., Guangfa, T., Yingyun, L., & Zhiyong, W.
(2006). Dynamic evaluation of thermal comfort environment of air-conditioned
buildings. Building and Environment, 41(11), 1522–1529.
doi:10.1016/j.buildenv.2005.06.002
Humphreys, M. A. (2005). Quantifying occupant comfort: are combined indices of
the indoor environment practicable? Building Research & Information, 33(4),
317–325.
Jain, R. K., Gulbinas, R., Taylor, J. E., & Culligan, P. J. (2013). Can social influence
drive energy savings? Detecting the impact of social influence on the energy
consumption behavior of networked users exposed to normative eco-feedback.
Energy and Buildings, 66, 119–127. doi:10.1016/j.enbuild.2013.06.029
Kwok, A. G., & Rajkovich, N. B. (2010). Addressing climate change in comfort
standards. Building and Environment, 45(1), 18–22.
doi:10.1016/j.buildenv.2009.02.005
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 and Buildings, 72, 411–418.
doi:10.1016/j.enbuild.2013.12.057
Marshall, C., & Rossman, G. (2011). Designing qualitative research. (5th ed.).
Thousand Oaks, CA: SAGE Publications, Inc.
Nicol, F., & Humphreys, M. (2010). Derivation of the adaptive equations for thermal
comfort in free-running buildings in European standard EN15251. Building and
Environment, 45(1), 11–17.
Otter Tail Power Company. (2009, March). Electric space heaters. Retrieved Feb. 9,
2014, from http://www.conservingelectricity.com/ElectricSpaceHeaters.asp
Plano Clark, V., & Creswell, J. (2008). The mixed methods reader. Thousand Oaks,
CA: SAGE Publications, Inc.
Sweeney, J. C., Kresling, J., Webb, D., Soutar, G. N., & Mazzarol, T. (2013). Energy
saving behaviours: Development of a practice-based model. Energy Policy, 61,
371–381. doi:10.1016/j.enpol.2013.06.121
© ASCE
Sustainable Human–Building Ecosystems 26
© ASCE
Sustainable Human–Building Ecosystems 27
Abstract
INTRODUCTION
© ASCE
Sustainable Human–Building Ecosystems 28
Albuyeh, 2009 - Kang, Park, Oh, & Park, 2014 - Rahimian, Iulo, & Cardoso Llach,
In Press).
Energy feedback technologies are based on the hypothesis that most people
lack awareness and understanding about how their everyday behavior affects the
environment and therefore limits the consumer’s capacity on deciding to take
conservation actions (Lutzenhiser, 1993 - Froehlich, et al., 2010). In microgrids, it is
expected that displaying energy feedback information at the demand side will
increase users’ knowledge on consumption and consequently leads to managing
energy usage by adopting long lasting conservation behaviors. Several studies
suggest that delivering high quality energy-related information on users’ activities
and consumption rates has the potential to motivate users change behavior towards
more energy saving ones (Yu, Fung, Haghighat, Yoshino, & Morofsky, 2011).
However, feedback delivery alone may not suffice to change people’s consumption
behavior in long term (Hargreaves, Nye, & Burgess, 2010). Thus the issue of how to
make energy-related behavior change durable and integrated to one’s lifestyle
remains a topic for further investigation.
BACKGROUND
1As Froehlich et al assert, environmental HCI is the study and design of eco-feedback technologies
which provides feedback on individual or group behaviors with the goal of reducing environmental
impact (Froehlich, Findlater, & Landay, 2010)
© ASCE
Sustainable Human–Building Ecosystems 29
of feedback design and the production of the artifact itself with an emphasis on
understandability, usability and aesthetics (Froehlich, Findlater, & Landay, 2010).
SYSTEM OVERVIEW
The section below discusses how the proposed conceptual prototype energy
exchange system operates and introduces an interface structured upon an incentive
program that motivates users’ participation.
© ASCE
Sustainable Human–Building Ecosystems 30
TH
HE INCENT
TIVE-SRU
UCTURED INTERFA
ACE
© ASCE
Sustainable Human–Building Ecosystems 31
© ASCE
Sustainable Human–Building Ecosystems 32
HOW IT WORKS
© ASCE
Sustainable Human–Building Ecosystems 33
staying in the limits of their debit-energy share and receiving monetary rewards and
some crossing the lines of efficiency and paying back money to the system. By this
conceptual prototype and the introduced framework it is expected that energy and
money transactions constantly occur in the scale of the community and become the
main driver of the energy exchange system. In this scenario the system will be host
to users’ participation in this community endeavor.
While explained the conditions which debit and credit-energy tokens are
being used, a situation might occur that a household’s consumption goes beyond the
limits of the second share of energy (credit-energy token) in a month, leaving the
user in need of extra energy-tokens. In this case a shared community-token account
can be borrowed against, providing household’s access to additional energy tokens.
By using energy-tokens from this shared account, the user owes the community both
energy-tokens and actual money due to the end of the following month. In order to
prevent users from continuously depending on the credit and community energy
accounts for purchasing extra energy, the energy price follows an ascending pattern.
ACKNOWLEDGEMENTS
This research was supported by the Alma Heinz and August Louis Pohland
Scholarship in the Stuckeman School of architecture and landscape architecture.
Professor Ute Poerschke’s support during this research is gratefully acknowledged.
Some information covered in this paper, including a more thorough literature review,
will be published in the forthcoming proceedings for the “Architecture and
© ASCE
Sustainable Human–Building Ecosystems 34
REFERENCES
Amin, M., & Wollenberg, B. F. (2005). Toward a Smart Grid. IEEE Power and
Energy, 34 - 41.
Aune, M. (2007). Energy Comes Home. Energy Policy, 5457 - 5465.
Costanzo, M., Archer, D., Aronson, E., & Pettigrew, T. (1986). Energy Conservation
Behavior, The Difficult Path From Information to Action. American
Psychologist , 521 - 528.
Farhangi, H. (2010). The Path of the Smart Grid. IEEE Power & Energy, 18 - 28.
Fischer, C. (2008). Feedback on household electricity consumption: a tool for saving
energy? Energy Efficiency, 79 - 104.
Froehlich, J., Findlater, L., & Landay, J. (2010). The Design of Eco-Feedback
Technology. In Proceedings of the SIGCHI Conference on Human Factors in
Computing Systems, (pp. 1999-2008). Atlanta.
Goodman, E. (2009). Three Environmental Discourses in Human-Computer
Interaction. In CHI'09 Extended Abstracts on Human Factors in Computing
Systems, (pp. 2535-2544).
Hargreaves, T., Nye, M., & Burgess, J. (2010). Making energy visible: A qualitative
field study of how householders interact with feedback from smart energy
monitors. Energy Policy, 6111-6119.
Ipakchi, A., & Albuyeh, F. (2009). Grid of The Future, Are we ready to transion to a
smart grid? IEEE Power & Energy, 52 - 62.
Kang, S. J., Park, J., Oh, K.-Y., & Park, H. (2014). Scheduling-based real time
energy flow control strategy for buildingenergy management system. Energy
and Building, 239 - 248.
Lutzenhiser, L. (1993). Social and Behavioral Aspects of Energy Use. Annual
Review of Energy and the Environment, 247-289.
Ouyang, J., & Hokao, K. (2009). Energy-saving potential by improving occupants’
behavior in urban residential sector in Hangzhou City, China. Energy and
Buildings, 711 - 720.
Paglia, T. K. (2011). Energy Improvement Districts and Local Energy Production
(Master of Regional Planning Dissertation, Cornell University).
Rahimian, M. (2015). A Participatory Approach for Constructing Energy Resiliency.
University Park, PA, USA: Unpublished Master Dissertation. Pennsylvania
State University.
Rahimian, M., Iulo, L. D., & Cardoso Llach, D. (In Press). The Case for a
Collaborative Energy Sharing Network. Submitted for publication at the
proceedings for the "Architecture and Resilience on a Human Scale"
conference at University of Sheffield, School of Architecture, 10-12
September 2015.
© ASCE
Sustainable Human–Building Ecosystems 35
© ASCE
Sustainable Human–Building Ecosystems 36
One Size Does Not Fit All: Eco-Feedback Programs Require Tailored Feedback
Abstract
INTRODUCTION
© ASCE
Sustainable Human–Building Ecosystems 37
concepts such as social norms (Schultz et al., 2007), goal setting (Erickson et al.,
2013), and gamification (Gamberini et al., 2012) at larger scale at a higher frequency.
Eco-feedback systems that are designed based on social norm effects have
proven to be effective in establishing energy efficient behavior among peers
(Gulbinas, Jain, et al., 2014; Jain, Gulbinas, Taylor, & Culligan, 2013). However,
designing such systems requires a targeted and fair normative comparison
environment to engage all the users regardless of their level of efficiency,
demographics, or equipment. In this paper, we developed two novel energy efficiency
metrics to quantify occupants’ energy-use behavior and conducted multiple
hypothesis tests on occupants’ behavior under various eco-feedback programs. The
results are discussed in detail and future venues of research are suggested to further
increase the efficacy of eco-feedback programs.
BACKGROUND
© ASCE
Sustainable Human–Building Ecosystems 38
METHODOLOGY
© ASCE
Sustainable Human–Building Ecosystems 39
ind
dividuals’ lowwest level of
o energy-use in their ennergy consuumption histtory or uses
the average offf-working ho ours energy value of all employees (whichever is less) and
meaasures the off-working
o hours energ gy savings bby penalizing the excesssive energy
connsumption with
w respect to t their best performance
p e, and rewarrds based on the amount
of energy
e savedd with respeect to work-h hours averagge energy coonsumption. In order to
commpare the reesults and consider
c thee individualss’ energy effficiency pootential, the
norrmalized data is used to reward occu upants’ enerrgy saving effforts. Fig. 1 depicts an
oveerview of thee off-workinng hour energ gy saving reeward algoritthm. The summation off
eacch individuall’s efficiencyy reward (i.ee. the green area) and innefficiency ppenalty (i.e.
the red area) demonstrate
d their energgy efficienccy status com mpared to ttheir peers.
Furrthermore, the normalizzed data red duces the eeffect of equuipment varriability on
occcupants’ EE ranking and provides a fairf compari son measuree.
Fig 1. An overview
o off off-working hour enerrgy saving ccalculation p
process
© ASCE
Sustainable Human–Building Ecosystems 40
Daily Energy-use Intensity. Daily energy consumption is the most common metric
used in eco-feedback systems as a means of normative comparison. This metric is
necessary but not sufficient to establish sustainable energy efficient behavior. Thus,
we calculate the daily average energy-use as a part of our comprehensive metric.
Hypothesis 1, 2, 3: Occupants who are ranked in the top, medium, and bottom
quintile of “Energy Savings” (Group A, B, C), reduced their energy consumption
after the eco-feedback program was launched.
Hypothesis 4, 5, 6: Occupants who are ranked in the top, medium, and bottom
quintile of “Off-Working Hours Energy Savings” (Group D, E, F), reduced their
energy consumption after the eco-feedback program was launched.
Hypothesis 7, 8, 9: Occupants who are ranked in the top, medium, bottom quintile of
“Shared Resources Energy Savings” (Group G, H, I), reduced their energy
consumption after the eco-feedback program was launched.
© ASCE
Sustainable Human–Building Ecosystems 41
© ASCE
Sustainable Human–Building Ecosystems 42
CONCLUSION
New behavioral analysis and ranking systems are needed in order to better
understand occupants’ reactions to eco-feedback programs and to improve the
normative comparison methodologies implemented in current eco-feedback research
in commercial buildings. In this paper, we develop new metrics to analyze occupant
behavior and performed hypothesis tests based on data collected in an eco-feedback
study conducted in a commercial building in Denver, CO. Inconsistent responses
under the unified eco-feedback program indicate that a single eco-feedback system
does not fit all occupants. These findings suggest that there is potential to improve the
efficacy of EE programs by providing occupants with targeted feedback.
AKNOWLEDGMENTS
REFERENCES
© ASCE
Sustainable Human–Building Ecosystems 43
Bradley, P., Leach, M., & Fudge, S. (2014). The Role of Social Norms in
Incentivising Energy Reduction in Organizations.
Delmas, M. A., & Lessem, N. (2014). Saving power to conserve your reputation? The
effectiveness of private versus public information. Journal of Environmental
Economics and Management, 67(3), 353-370. doi:
10.1016/j.jeem.2013.12.009
Erickson, T., Li, M., Kim, Y., Deshpande, A., Sahu, S., Chao, T., . . . Naphade, M.
(2013). The dubuque electricity portal: evaluation of a city-scale residential
electricity consumption feedback system. Paper presented at the Proceedings
of the SIGCHI Conference on Human Factors in Computing Systems.
Gamberini, L., Spagnolli, A., Corradi, N., Jacucci, G., Tusa, G., Mikkola, T., . . .
Hoggan, E. (2012). Tailoring feedback to users’ actions in a persuasive game
for household electricity conservation. Persuasive Technology. Design for
Health and Safety, 100-111.
Gulbinas, R., Jain, R. K., & Taylor, J. E. (2014). "BizWatts": A modular socio-
technical energy management system for empowering commercial building
occupants to conserve energy. Applied Energy.
Gulbinas, R., Khosrowpour, A., & Taylor, J. E. (2014). Segmentation and
Classification of Commercial Building Occupants by Energy-Use Efficiency
and Predictability. IEEE Transaction on Smart Grids.
Gulbinas, R., & Taylor, J. E. (2014). Effects of real-time eco-feedback and
organizational network dynamics on energy efficient behavior in commercial
buildings. Energy and Buildings, 84(0), 493-500. doi:
http://dx.doi.org/10.1016/j.enbuild.2014.08.017
Jain, R. K., Gulbinas, R., Taylor, J. E., & Culligan, P. J. (2013). Can social influence
drive energy savings? Detecting the impact of social influence on the energy
consumption behavior of networked users exposed to normative eco-feedback.
Energy and Buildings, 66, 119-127. doi: 10.1016/j.enbuild.2013.06.029
Noar, S. M., Benac, C. N., & Harris, M. S. (2007). Does tailoring matter? Meta-
analytic review of tailored print health behavior change interventions.
Psychological bulletin, 133(4), 673.
Schmitt, J. (2014). Normative Social Influence and the Moderating Role of Group
Identification: A Field Experiment on Household Electricity Consumption.
Schultz, P. W., Nolan, J. M., Cialdini, R. B., Goldstein, N. J., & Griskevicius, V.
(2007). The constructive, destructive, and reconstructive power of social
norms. Psychological science, 18(5), 429-434.
Smeets, T., Kremers, J., De Vries, H., & Brug, J. (2007). Effects of tailored feedback
on multiple health behaviors. Annals of Behavioral Medicine, 33(2), 117-123.
U.S.EIA. (2014). Annual Energy Outlook 2014 Early Release Overview.
U.S.EPA. (2012). The Inside Story: A Guide to Indoor Air Quality. Retrieved from:
http://www.epa.gov/iaq/pubs/insidest.html#Refguide
© ASCE
Sustainable Human–Building Ecosystems 44
Abstract
© ASCE
Sustainable Human–Building Ecosystems 45
INTRODUCTION
Residential and commercial buildings currently account for 40% of the total
annual energy consumption in the U.S., making them the largest consumer of the
country’s total energy use (U.S. Department of Energy 2011). In particular, the
commercial sector consumes half of all buildings’ energy use (U.S. Department of
Energy 2011), and its energy demand continues to grow faster than other energy-use
sectors (e.g., industry and transportation) (U.S. Energy Information Administration
2014). In particular, more than 80 percent of buildings’ total energy is typically
consumed during the operational phase (United Nations Environment Programme
2007). Therefore, maximizing energy saving during the operational phase when the
building is actually in use is critical.
A growing body of research posits that changing occupants' behaviors is one
of the most cost-effective approaches to achieving energy savings and offers a great
potential for contributing to energy savings (Rieur and Alahmad 2014). Built
environment’s energy use is highly connected to the energy-use behavior of its
occupants, and commercial buildings are estimated to have a potential 6-38% energy
reduction through occupant behavior changes (Azar and Menassa 2013; Kavulya and
Becerik-Gerber 2012; Meier 2006; Staats et al. 2000). Occupant-driven usage in
commercial buildings can also lead to increased waste—excess energy consumption
can account for up to 150% more energy use than necessary (Clevenger et al. 2014).
Unfortunately, such research into occupants’ behavior change significantly relies on
the data availability of occupant-specific energy consumption, but an overview of the
current literature indicates the difficulties in tracking occupant-specific energy
consumption due to the diversity of occupants and complexity of tasks in commercial
buildings (Gulbinas et al. 2015).
To address this issue, our previous study (Chen and Ahn 2014) presented a
concept of non-intrusive occupant load monitoring (NIOLM) that estimates occupant-
specific energy consumption by linking building energy-load variations with
occupancy-sensing data. However, individualities and uncertainties of occupants’
energy use behavior pose a challenge in the pursuit of this approach’s development.
For example, the time lags between occupant’s entry/departure events and the energy-
load variations caused by him/her may vary by occupants and also by each entry and
departure event of an occupant. As part of our effort to address such issues, this study
aims to analyze the repeatability of occupants’ energy-use behaviors, particularly
those related to time lags between occupancy events and energy-use starting/ending
events. This research will provide a basis for developing an algorithm to establish the
linkage between datasets collected by two different systems: occupancy sensing and
building energy management systems (BEMS).
© ASCE
Sustainable Human–Building Ecosystems 46
© ASCE
Sustainable Human–Building Ecosystems 47
Figure 1. NIOLM
N fram
mework
RE
ESEARCH OBJECTIV
O VE AND HY
YPOTHESE
ES
© ASCE
Sustainable Human–Building Ecosystems 48
Hyypothesis 1. There is not a significaant variabilityy in the delaay interval bbetween the
occcupant’s entrry event and d the start off his/her enerrgy-consum
ming behaviors since the
occcupant practtices a consistent delay pattern reppeated over time. H0: T There is no
stattistically siggnificant diifference am mong differrent delay intervals bbetween an
occcupant’s entrry events an nd his/her en nergy consum mption. H1: There is a statistically
signnificant diffference among differentt delay interrvals betweeen an occuppant’s entry
eveents and his/h her energy consumption
c n.
Fiigure 2. Sam
mple of delaay interval
Hyypothesis 2. There is not a significaant variabilityy in the delaay interval bbetween the
endd of energy--consuming behaviors and a the occcupant’s depparture evennt since the
occcupant practtices a consistent delay pattern reppeated over time. H0: T There is no
stattistically sig
gnificant diifference ammong differrent delay intervals bbetween an
occcupant’s energy-consum ming behaviors and his/hher departuree events. H1: There is a
stattistically sig
gnificant diffference am
mong delay intervals bbetween an occupant’s
eneergy-consum ming behavioors and his/heer departure events.
EX
XPERIMEN
NT DESIGN
N AND MET
THODOLO GY
© ASCE
Sustainable Human–Building Ecosystems 49
survey results indicated that participants always leave their smartphone’s Wi-Fi on,
which provided an opportunity for us to capture their Wi-Fi packets. During the
experiment, the energy consumption of each participant was tracked at his/her work
desk by installing a plug-in level meter. The meter—“watts-up?.net”—collected the
real-time electrical energy-load data with 1-second interval resolution. In addition, a
local access point was set up to track the entry/departure events of the participants. A
database was also created to store each participant information from the
entry/departure events and associated energy behavior.
Since we needed to test whether an occupant has a consistent delay interval
pattern repeated over the experiment, we divided collected data for delay interval of
entry/departure events for each participant into two groups: 1st 2-week and 2nd 2-
week. Comparing the two groups would help to find whether individual occupants’
delay interval patterns repeat over time. An appropriate statistical comparison
between the data of the two groups could then address the hypotheses. For this
reason, a two-sample t-test was chosen as the scenario constructed for testing our
hypotheses.
Figure 3 presents the scatter plot for all entry and departure delay intervals of
Participant #1. This figure, as a sample of raw data, shows that the variability across
delay intervals for entry/departure events is low, which indicates that delay intervals
repeat over time. It is noteworthy that across all the raw data, the variability for delay
intervals of all five participants is low; a high variability would likely indicate that a
participant does not have a specific delay interval repeated over time. Furthermore,
figure 4 presents the boxplots for entry and departure delay intervals for two groups
of data for Participant #2. These boxplots, as a sample, show there is no significant
difference between the means of the two groups of data. The boxplots also shows that
the variance of the two groups of data is approximately equal.
Table 1 also presents the results of conducting a t-test on entry/departure
delay intervals for each participant. Based on the results, we fail to reject our null
hypotheses for all five participants, which means that there is no significant
difference among delay intervals for individual participants. The results of the t-test
would therefore indicate that individual participants have their own delay interval
patterns repeated over time.
The specific delay interval pattern for each occupant could be a critical detail
for determining the start/end of energy-consuming behavior. Since the developed
NIOLM framework is founded on occupancy-detection events and there are a large
number of occupants in commercial buildings with different entry and departure
times, it is possible that some entry or departure events happen at the same time. In
such cases, knowing each occupant’s specific delay interval as it is repeated over time
would be helpful in correlating various load variations with various entry/departure
events to find load variations created by each individual occupant.
© ASCE
Sustainable Human–Building Ecosystems 50
Fig
gure 3. Scattter plots forr delay interrvals of Parrticipant #1
Figure
F 4. Box
x plots for delay
d intervvals of Participant #2
Tabble 1. Two-S
Sample t-Teest for Delay y Intervals
Enttry Event Dep parture Eveent
Parrticipant
t-Statistic t-Crittical 1 t-Statistiic Critical 1
t-C
#1 1.085290 2.0599538 1.0096700 2..063898
#2 0.015134 2.0555552 0.959831 2..051831
#3 1.188885 2.0599538 0.8450555 2..079613
#4 2.072843 2.0933024 1.1168122 2..093024
#5 0.132360 2.0799613 1.7303244 2..063898
1
α = 0.05
Based on
o the storedd energy datta in the dattabase, we aalso observeed that each
parrticipant typpically incrreases/decreaases the looad by a consistent amount at
entrry/departuree events. Tab
ble 2 shows the amountts of energy load increaase/decrease
and
d the corresponding freequency for each particcipant. Suchh findings inndicate that
ind
dividual occu upants have their own specific eneergy-use related behavior repeated
© ASCE
Sustainable Human–Building Ecosystems 51
CO
ONCLUSION
This papper introducced and testeed the conceppt of monitooring the delay intervals
betw
ween occupants’ entry/d departure events and theeir energy coonsumption. The results
ind
dicated that individual
i occcupants hav ve unique deelay intervall patterns reppeated over
tim
me. Such speccific delay innterval patteerns can servve as the fouundation for developing
the NIOLM fraamework, whichw will serve as an eefficient toool for gatherring critical
© ASCE
Sustainable Human–Building Ecosystems 52
ACKNOWLEDGMENT
REFERENCES
© ASCE
Sustainable Human–Building Ecosystems 53
Staats, H., van Leeuwen, E., and Wit, A. (2000). “A longitudinal study of
informational interventions to save energy in an office building.” Journal of
Applied Behavior Analysis, 33(1), 101–104.
United Nations Environment Programme. (2007). “Buildings Can Play Key Role In
Combating Climate.”
U.S. Department of Energy. (2011). “Building Energy Data Book.”
U.S. Energy Information Administration. (2014). “Annual Energy Review.”
Zoha, A., Gluhak, A., Imran, M. A., and Rajasegarar, S. (2012). “Non-Intrusive Load
Monitoring Approaches for Disaggregated Energy Sensing: A Survey.”
Sensors, 12(12), 16838–16866.
© ASCE
Sustainable Human–Building Ecosystems 54
Abstract
INTRODUCTION
© ASCE
Sustainable Human–Building Ecosystems 55
© ASCE
Sustainable Human–Building Ecosystems 56
provide the necessary related data on human behavior, these studies could be limited
in terms of their size and generalizability (e.g., small number of participants, one or
limited number of case studies, and etc.). More importantly, there is a possibility of
experimental noise that could affect the results and that cannot be controlled for (e.g.
changes in weather, daylighting availability, interior designs, location and orientation
of building(s) etc.). For instance, (Maniccia et al. 1999) mentions, in their study, the
difference in interior designs among different offices could have impacted the
shading and dimmer use among occupants of different buildings.
In order to reduce the experimental noise, collect accurate occupant related
information, and control different variables that could directly or indirectly affect
occupant behavior, we have utilized immersive virtual environments (IVEs) to create
virtual office spaces with realistic interior designs (e.g., tables, chairs) and lighting
features (e.g., shades, light switches, realistic luminance and illuminance) that a
physical office would have. In recent years, with the advents in fields of virtual and
augmented reality, there has been a significant amount of improvement in visual
quality of such environments, as well as their immersivity and interactivity options,
allowing users to feel fully immersed in the virtual environment. Previous studies
have shown that participants’ sense of presence is similar in IVEs and physical
environments (Adi and Roberts 2014). In our previous study, we benchmarked 150
participants’ performance and sense of presence between a physical office space and
a virtual office space (Heydarian et al. 2015); we concluded that participants perform
similarly between the two environments and have a strong sense of presence within
IVEs. Another important advantage of IVEs is that the experimenter is less salient to
the participant (as participants cannot see the experimenter), facilitating behavior that
is more natural. Therefore, in this study, we utilized IVEs to facilitate our
explorations.
METHODOLOGY
© ASCE
Sustainable Human–Building Ecosystems 57
hadd three windo ows (includiing manually y controlled shades) andd 12 light fixxtures (three
fluoorescent ligh ht bulbs on each
e fixture). The light sswitches in tthe room weere designed
in a way that the t participaant could adjjust the artifficial lightinng to have oone, two, or
threee light bulb bs on each fixture.
f Figure 1 shows the renderedd model of tthe test bed
envvironment.
The parrticipants were
w asked tot read a shhort passagee and answ wer a set off
commprehension n questions based on what w they hhave read. The particiipants were
randomly assig gned either to a room with the deefault settingg of only nnatural light
avaailable (all sh hades open and
a no artifiicial lights oon) or only aartificial lighhts available
(alll artificial liights on and d no shadess open). Thee authors chhose these ttwo default
lighhting settings as they aree the polar opposites
o andd thus allow the authors to examine
the extent to wh hich a defau
ult lighting setting
s influeence a particcipant’s lightting choice.
Theey were instrructed that th heir task is to
t read the pprovided passage and thaat they have
the option of either
e keepiing the defaault light settting or adjusting it to their most
preferred settin ng. The partiicipants had the option tto open/closse each set oof shades to
incrrease/decreaase the availaability of naatural light aand turn the light switchhes on/off to
conntrol the arrtificial ligh ht levels in n the room. Followingg the experriment, the
parrticipants com mpleted a qu uestionnaire assessing thheir environm mental valuees.
Figure
F 1 – Virtual
V Officce Space
Mo odel and Ap pparatus. To o be able to create a reaalistic and innteractive virtual model
andd have the participants
p to be fully immersed inn them, a 33D model off the office
spaace was first designed in n Revit© 2015. The Reviit model onlyy consisted of the basic
geoometry of th he room, inccluding wallls, generic w windows, thee ceiling, annd the light
fixttures. The Revit
R model was then im mported to 3dds Max© to give it a moore realistic
look by adding g furniture, materials,
m teexture, shaddows, reflecttions, and arrtificial and
natuural lighting
g. Once all the
t necessary y componennts to make the room loook realistic
©
werre added in 3ds Max , 32 3 different renderings ((render-to-teexture) were created for
eveery lighting combination n that the paarticipants ccould have iin the IVE. In order to
enssure the lighting settingss in 3ds Max x© were reallistic and reppresentative of real-life
valuues, 32 diffeerent light maps
m were geenerated usinng Ladybugg and Honeyybee plugins
(op
pen source plugins for Grasshopper3
G 3D – Rhino plugin (Rouudsari et al. 2013)) and
© ASCE
Sustainable Human–Building Ecosystems 58
the lux values around the room were calculated. It is importtant to note in order to
keeep everything constant among
a the participants,
p the modeledd office wass located in
Loss Angles an nd the simu ulations weere performeed at 2:00 PM, June 22nd; both
connditions’ avaailable lux values
v were according tto the standaard’s for intterior office
spaace with lux values approximately ranging
r fromm 500 to 10000 lux throoughout the
dayy.
32 models were then exported as a FBX files from 3ds M Max© and were imported
© ©
in Unity
U gamee engine as different
d scen
nes. The Unnity game eengine was uused for two
purrposes: (1) to
o program in nteractive opptions for eacch scene andd (2) to connnect the IVE
equuipment to thhe 3D models. The interractive optioons were proogrammed inn a way that
wou uld allow thhe participaants to havee a more re alistic interaaction with the virtual
models (e.g., being able to turn light bulbs
b on or ooff by standiing in front of the light
swiitch or opening and closing
c the shades by standing inn front of a window,
animmations thaat would show the sh hades openning or cloosing based on users’
inteeraction). Th
he equipmen nt that were used
u to provvide a fully iimmersive eenvironment
inclluded an Occulus DK2 Head-Moun nted Displayy (HMD), ann Xbox-3600 controller,
andd a positional tracker thaat would track the particiipants’ head and neck movements.
Exp perimental Procedure. 64 participants, betweeen ages 18 tto 32, were recruited (a
powwer analysis confirmed thet sample size
s was suffficient to dettect significaant effect at
a power
p of 0.990). The parrticipants were
w all undeergraduate oor graduate students in
diffferent fields of study (e.g. engineeriing, architeccture, life-scciences, art, and etc.) at
the University of Southern n California. Prior to thee experimennt, the participants were
given a brief explanation
e of
o the experriment withoout revealinng any inforrmation that
cou
uld affect their judgmen nt in their deecisions durring the experiment. Following the
exp
planation, th hey were askked to read and sign ann IRB (Instittutional Revview Board)
app
proved conseent form.
© ASCE
Sustainable Human–Building Ecosystems 59
and/or the light switch. The participants were randomly divided into two groups with
different default light settings. They were asked to place the HMD on their head and
walk inside the virtual office. Once they entered the room, at first, they were
instructed to walk around the space and get a feeling of the environment (Figure 2). If
they were in a default light setting with “only natural light available”, they only had
the shades open and all the artificial lights were off, and if they started with “only
artificial lights” available, they had all the shadings closed and maximum artificial
light available.
If they chose to change the default setting, the setting they set as their
“preferred” setting was recorded. Once comfortable with the lighting setting (either
the default setting or the preferred setting), they were then asked to read over the
provided passage on the table. The amount of time that took the participants to read
the passage was recorded. Upon completion of the reading task, they were asked to
remove the HMD and complete a comprehension test based on what they had read.
Following that, they were asked to complete an environmental value questionnaire.
We first examined how the lighting default settings (“all natural lights” and
“all artificial lights) differentially influenced participant’s decision to keep or change
the lighting setting. There was a statistically significant difference in participants’
propensity to keep the default light setting between the two lighting conditions, such
that those with the “all natural lights” default setting were significantly more likely to
keep the default setting than those with the “all artificial lights” default setting, χ2 (1)
= 9.94, p = .002. 74% of the participants with the “all natural lights” default setting
kept the initial lighting setting whereas only 29% of the participants in “all artificial
lights” default kept the initial light setting.
Participants were marginally significantly more likely to keep the initial
lighting setting than at chance (50% chance of keeping and 50% chance of changing)
when they had “all natural lights” as the lighting default, 74% vs. 50%, χ2 (1) = 2.89,
p = .09. However, participants were not significantly more likely to keep the initial
lighting setting than at chance when they had “all artificial lights” as the default, 26%
vs. 50%, χ2 (1) = 2.33, p = .17.
Participants with the “all-natural light” default setting were significantly more
likely to have all of the blinds open as their final choice than those with the “all
artificial light” default setting (94% vs. 32% χ2 (1) = 25.69, p < .0001). Figure 3
shows the distribution of the final lighting choices for each of the default light
settings.
© ASCE
Sustainable Human–Building Ecosystems 60
Overall, the samplle of particiipants in booth the “alll natural” ddefault light
conndition and “all artificial” defau ult light coondition weere similarlyy high on
envvironmental--friendliness according to o the adminiistered questtionnaire, M = 6.13 and
M= = 6.86 (out of a maximum m score of 10), respectivvely. Particippants who wwere high on
envvironmental friendliness were signifi ficantly moree likely to chhange the innitial setting
oveerall, β = -.35, p = .032. However, th here was no significant eeffect of envvironmental
frieendliness on the final cho
oice.
Particip
pants in the “all
“ natural light” defauult setting annswered com mprehension
testt questions correctly marginally
m more
m significcantly than those in thee “artificial
lighht” default condition, β = 12.32, p = .061. Theree were no siignificant diffferences in
read ding speed between
b the two settingss.
ONCLUSION, LIMITA
CO ATIONS AN
ND FUTURE
E WORK
These results
r sugg
gest that deefault settings can havve different effects on
inflluencing parrticipants to keep the in nitial lightingg setting. Lighting defaault settings
with all naturall lights weree more likely y to be kept by participaants than at chance and
more likely to be b kept than n artificial lig
ght default ssettings. As a result, parrticipants in
with the all-nattural light deefault setting g were signiificantly morre likely to hhave all the
shaades open in n their final choice than those in the all artificiial light defaault setting.
Theese participaants also perrformed bettter on the coomprehensioon questionss than those
in the artificial light defaultd settiing. Additiionally, thee authors found that
envvironmentally friendly individuals were w more llikely to chhange the innitial setting
oveerall. These results have implicattions for arrchitects, enngineers, annd building
man nagers. As natural
n lightt use has beeen shown too be an effeective meanns to reduce
eneergy, office rooms shou uld be design ned with wiindows and managers oof buildings
sho
ould have thee blinds opeened as a default to encoourage naturral light use rather than
artiificial light use.
u
This reesearch brid dges the gap p between the human behavior rresearch on
occcupant behav vior and eng gineering/arcchitectural w work on buildding design and energy
© ASCE
Sustainable Human–Building Ecosystems 61
ACKNOWLEDGMENT
REFERENCES
Adi, M., and Roberts, D. (2014). "Using Virtual Environments to Test the Effects of
Lifelike Architecture on People." Technologies of Inclusive Well-Being, A. L.
Brooks, S. Brahnam, and L. C. Jain, eds., 261-285.
Dubois, M.-C., and Blomsterberg, Å. (2011). "Energy saving potential and strategies for
electric lighting in future North European, low energy office buildings: A
literature review." Energy and Buildings, 43(10), 2572-2582.
EPA (2013). "National Awarness of Energy Star ",
<http://www.energystar.gov/sites/default/uploads/about/old/files/2013%2020CEE
%2020Report_2508%2020compliant.pdf>.
EPA (2014). "Energy Star, Building and Plants."
<http://www.energystar.gov/buildings?s=mega>.
© ASCE
Sustainable Human–Building Ecosystems 62
Fabi, V., Andersen, R. V., Corgnati, S. P., and Olesen, B. W. "A methodology for
modelling energy-related human behaviour: Application to window opening
behaviour in residential buildings." Proc., Building Simulation, 415-427.
Galasiu, A. D., and Veitch, J. A. (2006). "Occupant preferences and satisfaction with the
luminous environment and control systems in daylit offices: a literature review."
Energy and Buildings, 38(7), 728-742.
Heydarian, A., Carneiro, J. P., Gerber, D., Becerik-Gerber, B., Hayes, T., and Wood, W.
(2015). "Immersive virtual environments versus physical built environments: A
benchmarking study for building design and user-built environment
explorations." Automation in Construction, 54(0), 116-126.
Hong, T., D'Oca, S., Turner, W. J. N., and Taylor-Lange, S. C. (2015). "An ontology to
represent energy-related occupant behavior in buildings. Part I: Introduction to
the DNAs framework." Building and Environment(0).
Inoue, T., Kawase, T., Ibamoto, T., Takakusa, S., and Matsuo, Y. (1988). "The
development of an optimal control system for window shading devices based on
investigations in office buildings." ASHRAE transactions, 94, 1034-1049.
Kawase, T. (2000). "Office Worker Preferences of Exterior Shading Devices: a Pilot
Study." Eurosun 2000, Copenhagen, Denmark.
Maniccia, D., Rutledge, B., Rea, M. S., and Morrow, W. (1999). "Occupant use of
manual lighting controls in private offices." Journal of the Illuminating
Engineering Society, 28(2), 42-56.
O'Brien, W., Kapsis, K., and Athienitis, A. K. (2013). "Manually-operated window shade
patterns in office buildings: A critical review." Building and Environment, 60(0),
319-338.
Roudsari, M. S., Pak, M., and Smith, A. "ladybug: A Parametric Environmental Plugin
For Grasshopper To Help Designers Create An Environmentally-Conscious
Design." Proc., 13th conference of international building performance
association, chambery, france, 26-28.
Sieminski, A. (2013). "International energy outlook 2013." US Energy Information
Administration (EIA) Report Number: DOE/EIA-0484.
Van Den Wymelenberg, K. (2012). "Patterns of occupant interaction with window blinds:
A literature review." Energy and Buildings, 51, 165-176.
© ASCE
Sustainable Human–Building Ecosystems 63
Abstract
In the recent years, surveys and studies have established the importance of
occupant's behaviour on energy consumption in buildings. Therefore, inclusion of
inhabitants' behaviours is compulsory for the assessment of building energy
management system's (BEMS) strategies, which highly depends on human
behaviour. The purpose of modelling the inhabitants’ behaviour is to see how their
choices and control of household appliances can impact the energy consumption. In
this paper, a co-simulation approach is presented where the inhabitants' behaviours
are co-simulated with the SIMBAD-MOZART thermal model of a reference house
and BEMS. The realization of all the different kinds of inhabitant behaviours into
energy co-simulations will help to improve the smart grid technology and hence
provide inhabitants with better services to save energy and cost while maintaining
their comfort levels.
INTRODUCTION
The advancements in the electric grid technology have led to the concept of a
smart grid that uses the information technology to communicate with the suppliers
and customers about their energy supply and demand needs. The smart grid helps in
improving energy efficiency and sustainability of its production and distribution. The
information that can be provided to the inhabitants consists of availability of energy,
tariff details and energy consumption by different household appliances etc. After
receiving all the different information from the smart grid, the inhabitants must be
intelligent enough to interpret all this information so that they can save energy while
maintaining their comfort. This requires a high cognitive workload to make decisions
about energy management, and the results depend on how intelligently the
information is handled and acted upon.
© ASCE
Sustainable Human–Building Ecosystems 64
inhabitants. Thhe inhabitannts can alsoo communiccate with thhe BEMS and a can exppress
theiir comfort needs,
n occuppancy planss etc. and caan also ask for
f advice.
In ordeer to assess and evaluaate the diffeerent strateggies that aree developedd by
the BEMS, it i is imporrtant to innclude the inhabitantss’ reactive and dynaamic
inteeractions with
w their ennvironment in buildingg energy siimulations. It will helpp to
anaalyze the coontrol of diifferent behhaviours oveer the envirronment annd the resullting
imppact on enerrgy consum mption patterrns. Similarrly, the rolee of BEMS in the preseence
of these reactiive behavioours will be more chaallenging annd will leaad to improoved
funnctionality and
a energyy efficient decision
d m
making. Thee BEMS ussed in the co-
sim
mulation callled G-HomeeTech (Ha et al., 2012) has been developed
d a G-SCOP and
at
commmercializeed by Vesta System (VeestaEnergy,, 2011).
INH
HABITANTS' BEHA
AVIOUR MODEL
M
© ASCE
Sustainable Human–Building Ecosystems 65
howwever, due to
t the externnal environmental consstraints only
y one of theem is conveerted
to tthe agent's intention. Finally,
F bassed on the intention thhe agent peerforms cerrtain
actiions on the environmennt.
CO
O-SIMULA
ATION ENV
VIRONME
ENT
© ASCE
Sustainable Human–Building Ecosystems 66
Figgure 3. Co--simulation
n environm
ment
O-SIMULA
CO ATION SCE
ENARIO AND
A RESU
ULTS
A scen nario of a 2 person fammily, husbannd and wife,, has been implemente
i ed in
Braahms. The husband is an "Eco agent",a wheereas the wife
w is a "N Non-eco" aggent.
Aftter spendingg their day at
a work, thee agents com me back hoome in the evening.
e Ass the
houuse is relativvely warmeer than outsiide, the ageents will perrceive it to be comfortaable
for a while. However,
H fter a short period theyy will start perceiving that the acctual
aft
© ASCE
Sustainable Human–Building Ecosystems 67
tem
mperature iss very low w. The com mfort/discommfort of ann agent is based on the
hommeostasis which
w furtheer depends on the percceived PMV V (Predictivve Mean Vote)
V
valuues in this scenario.
s Ass soon as an
n agent startts to feel waarm (PMV 1 to 2), it taakes
somme action too be comforrtable again.. However, if it does no ot take any action or iff the
actiion does noot result thhe agent beeing comforrtable againn it will sttart feeling hot
(PMMV 2 to 3) or too hot (above 3). Similarly, as a soon as the t agents start
s perceivving
neggative PMV V value, theyy increase thhe temperatuure setpointt to be warm mth.
Since their
t percepption of com mfort does not
n solely depend
d on thhe temperatture,
butt also on othher factors, i.e. what acctivity they are involved in, what clothes
c theyy are
weaaring etc. The
T time att which theyy feel com mfortable varries. As soon as an aggent
starrts to feel warm
w it willl take an acction to be comfortablle again. Thhe EcoHusbband
ageent would prefer
p to decrease
d thee temperatuure by rem moving extraa clothing and
turnning off thee heater wheereas the NoonEcoWife agent woulld like to oppen the winddow
to quickly
q beccome comfoortable, witthout caringg that the heater
h that is
i still workking
andd that it is wearing
w tooo many clotthes. The innformation about
a the control overr the
apppliance/objeect is sent too the SIMB BAD thermaal model, wherew the neew temperaature
for the room iss calculatedd and sent back to Brahhms. Based upon the neew temperaature
the PMV valuees for all thee agents aree again calcuulated.
© ASCE
Sustainable Human–Building Ecosystems 68
unccomfortablee with the seetpoint adjusted by the BEMS and d it did not communicat
c te to
the BEMS. It rather itsellf increasedd the setpoiint to a higgher value and a put on the
sweeater. This is
i shown byy the "Adjust Heater Seetpoint" toool tip and "pput on sweaater"
tool tips on thheis workfraame in EcoH Husband's sspace at aroound 17:40. The blue line
going from thiis workfram me to the woorkframe in Livingroom mHeater's sppace shows that
the EchoHusbaand agent directly
d conttrolled the heater
h withoout any intervention byy the
eneergy manageer. These acctions helpeed the agent to become comfortablle shown byy the
yelllow coloured upward arrow shoowing the juump from one thermaal conditionn to
anoother in figgure 5c at around
a 18:000. The temmperature further
f wennt up to 266°C,
shoown in figurre 5b at aroound 19:30.. Now againn it starts feeling
fe warmm and turnss off
the heater. At this
t point wwhen the temmperature sttarts decreaasing, the BE EMS interruupts
the agents' deccisions and does not lett the temperature fall below
b 23°C
C by controllling
the heating system.
s Thhe state of the heatter is shown in fig gure 5a unnder
"LivingroomH Heater State"" where thee signal first goes to zeero and thenn to one duue to
BEMS interrupption. This is shown by b the "Set Temperaturre Intelligen ntly" tool tip in
EneergyManageer's worksppace at arouund 19:30. ThusT the EcoHusband d agent remmains
commfortable with
w the decision taken by b the BEM MS shown byy the green curve in figgure
5c between 200:00 and 233:00. The teemperature when contrrolled by thhe BEMS, also
helpps NonEcooWife agentt to remainn in the sliightly cool to comforttable condiition
rathher than being cool orr cold (figure 5c, 5d). This is sho own by the light blue and
whiite workfram mes in NonnEcoWife's space in figgure 4 and byb the light blue and grreen
currve in figuree 5d betweenn 20:00 andd 23:00.
(b) T
Temperaturre controlledd by agent and
a
BEMS
(aa) Agent co
ontrols heateer with BEM
MS
© ASCE
Sustainable Human–Building Ecosystems 69
CO
OST AND COMFORT
C T ANALYS
SIS
In this section, an analysis off the cost-coomfort tradeeoff for the situations with
w
andd without th he BEMS iss given. To quantify thhe comfort of o agents, thhe PMV vaalues
obttained after the simulattion runs arre summed up for diffferent PMV levels (figuures
6a and 6b). Siince EcoHuusband agennt is not onlly concerneed by the coomfort but also
the energy savvings and inn this efforrt it remainss less comffortable thann NonEcoW Wife
ageent (figure 6a).
6 Mostly, it remainss in slightlyy cool or slightly warm m due to havving
morre interactions with thhe heater to control the t temperaature. NonE EcoWife aggent,
howwever, remaains more comfortable
c e than EcoH Husband ageent, as it is not concerrned
aboout energy savings
s andd wants to achieve
a commfort at any cost. Figurre 6b showss the
therrmal comfoort durationss of agents with the innclusion of a BEMS inn the system m. In
thiss case, the divergence
d o agents' coomfort levells is reduced and they converge
of c too the
commfortable zoone. Also, the
t agents remain
r commfortable forr a longer tiime durationn as
commpared to before
b i.e. without
w BEM MS. In this case EcoH Husband ageent's comforrt is
bettter than NoonEcoWife agent. The improvemeent in the co omfort is duue to the beetter
deccisions takenn by the BE EMS based on the knoowledge thaat the BEMS S has aboutt the
inteernal and external ennvironmenttal conditioons, weathher forecastts, inhabitaant's
commfort and seelf learning algorithms..
Figure 7 shows the t power consumptioon of the electric heeater while the
envvironment is he BEMS. The
i controlleed by differrent agentss with and without th
highest power consumed is due to th he behaviouur of NonEccoWife ageent since it tries
t
to aachieve com
mfort by oppening and closing thee window. This
T assessm
ment of BE EMS
wheen co-simu ulated with building syystem and iinhabitants shows thatt the BEMS is
cappable of noot only savving the inhhabitants frrom cognitiive workloaad but alsoo of
prooviding them
m with betteer comfort and
a energy savings.
s
© ASCE
Sustainable Human–Building Ecosystems 70
CO
ONCLUSIO
ON
The beehaviour model in thee co-simulaator generattes the proffiles which are
random and dy ynamic. Ass soon as thhe environm mental variaables chang ge, they chaange
ageents’ beliefss and the system reacts in a differeent way thann before. Thhe introducction
of iinhabitants’ reasoning processes
p toowards their actions on n the physiccal environm
ment
will give energy simulatiion tools more
m realism
m. The reacttions to theese grid signals
couuld further be
b diverse and
a complex x dependingg on differen nt types of inhabitants e.g.
bassed on theeir family compositioon, role inn the familly, econom mic conditions,
knoowledge andd concerns about enerrgy problem m. The realiization of alla the diffeerent
kinds of inhabbitant behavviours into energy co--simulationss with the smart s grid will
helpp to improvve the smarrt grid techhnology andd hence pro ovide the innhabitants with
w
bettter services to save eneergy and cosst while maaintaining thheir comfortt levels.
RE
EFERENCE
ES
© ASCE
Sustainable Human–Building Ecosystems 71
Kashif, A., Dugdale, J., and Ploix, S. (2013b) Simulating occupants’ behaviour for
energy waste reduction in dwellings: A multi agent methodology. Advances
in Complex Systems 16, 37.
Noel, J. (2008) Cas d’exemple codyba a partir de la typologie cstb des batiments,
http://www.jnlog.com/pdf/typologie_cstb.pdf >(July. 7, 2015)
Sierhuis, M., Clancey, W., and Van Hoof, R. (2007) Brahms - a multiagent modeling
environment for simulating work practice in organizations. International
Journal of Simulation and Process Modelling 3(3), 134-152.
VestaEnergy, (2011) Solutions logicielles d’energy management dynamique,
http://www.asprom.com/jei/vesta.pdf >(July. 7, 2015)
© ASCE
Sustainable Human–Building Ecosystems 72
Abstract
INTRODUCTION
© ASCE
Sustainable Human–Building Ecosystems 73
Similar work for finding occupancy has been already tackled and various
methods have been investigated. The methods vary from basic single feature
classifiers that distinguish among two classes Presence and Absence to multi-
sensor, multi-feature models. A primary approach, which is prevalent in many
commercial buildings is to use passive infrared (PIR) sensors for occupancy.
However, motion detectors fail to detect presence when occupants remain
relatively still, which is quite common during activities like working on a
computer, or regular desk work. Furthermore, drifts of warm or cold air on objects
can be interpreted as motion leading to false positive detections. This makes the
use of only PIRs for occupancy counting purpose less attractive. Conjunction of
PIRs with other sensors can be useful as discussed in (Agarwal et al. 2010) who
makes use of motion sensors and magnetic reed switches for occupancy detection
to increase efficiency in HVAC systems of smart buildings, which is quite simple
and non-intrusive. Apart from motion, acoustic sensors (Padmanabh et al., 2009)
may be utilized. However, audio from the environment can easily fool such
sensors, and with no support from other sensors it can report many false positive
detections. In the same way, other sensors like video cameras (Erickson et al.,
2011; Milenkovic and Amft, 2013b), which utilize the huge advances in the field
of computer vision and the ever increasing computational capabilities, RFID tags
(Philipose et al., 2004) installed on id cards, sonar sensors (Milenkovic and Amft,
2013a) plugged on monitors to identify presence of a person on the computer,
have been used and have proved to be much better at solving the problem of
occupancy count, yet cannot be employed in most office buildings for reasons like
privacy and cost concerns. Pressure sensors and PIRs has been discussed in
(Nguyen and Aiello, 2012) to determine presence/absence in single desk offices.
They further tag activities based on this knowledge.
A new approach for occupancy recognition is going on by understanding
the relationships existing between carbon dioxide concentration and indoor air
quality IAQ in terms of occupant number. Physical CO2 model built on sensor
networks (Aglan, 2003) have been used extensively in smart office projects to
improve occupancy comfort and building energy use. However in this paper, CO2
physical model is studied to find out the valuable of using it in occupancy
estimation. However, for various applications like activity recognition, or context
analysis within a larger office space, information regarding the presence or
absence of people isn’t sufficient, and an estimation of the number of people
occupying the space is essential. (Lam et al., 2009) investigates this problem in
open offices, estimating occupancy and human activities using a multitude of
ambient information, and compare the performance of HMMs, SVMs and
Artificial Neural Networks. However, none of these methods generate human
understandable rules which may be very helpful to building managers.
In general, an occupancy count algorithm that fully exploits information
available from low cost, nonintrusive, environmental sensors and provides
meaningful information is an important yet little explored problem in office
buildings.
© ASCE
Sustainable Human–Building Ecosystems 74
Learning process. From the large set of features displayed in, (Abhay Arora and
Bandyopadhyay, 2015), some of them may not be worthwhile to consider, to
achieve our target of occupancy classification. These features are ones, which
© ASCE
Sustainable Human–Building Ecosystems 75
In this paper new features are added in addition to the previous one: Audio
microphone detection and occupancy from physical model. The correlation
with occupancy estimation with these features is discussed below.
© ASCE
Sustainable Human–Building Ecosystems 76
to the classifier. Since the path to the leaf may consist of many internal nodes,
each of which may check different feature values, such paths exploit the
correlation among the various features. The decision tree approach offers the
advantage of generating rules that the path towards the leaf node is quite
informative and it clearly points out direct causes for the selection of a particular
class. Unlike methods that use decision boundaries (SVMs, regression
techniques), decision tree analyses are independent of the scale of the input data,
so no conditioning of the data is necessary.
Using this raw training data, previously mentioned features were
extracted. A vector of features and target 〈 , , , … , ; 〉had been generated
for each time quantum, where fi stands for the ith feature and y, for the level of
occupancy.
Occupancy from Acoustic Sensor. Acoustic features are a very important part of
occupancy classification when other non-intrusive sensors offer low class
separation. A single omnidirectional microphone can be used as an important tool,
when it comes to classify occupancy. Omnidirectional microphones are ones,
which can pick up sound from virtually any direction. They are considerably
cheaper than having multiple unidirectional microphones, and prove to be much
advantageous in places where it is required to track/ listen to multiple sources like
in meetings, discussions, (Abhay Arora and Bandyopadhyay, 2015). In this paper,
the recording signal from an office is generally background environmental noise
with a few human voices, some door opening, and tapping events. From the
recording signal RMS amplitude feature is defined, which is the root mean square
(or average) of the amplitude of a sound. However, it is related to the volume of
∑ ( )
the sound: = , where n is the number of samples taken and Si the
th
i sample. High and low RMS value will give indicator to the level of occupants
inside the office, this relationship is easy to visualize in (figure 2, left side), which
represents both, the RMS amplitude in dB for 4 days, and the actual occupancy
profile with respect to time (quantum time is 30 minutes).
© ASCE
Sustainable Human–Building Ecosystems 77
The first step is to find the best parameter values for invariant parameters
S, , , and QD using an iterative nonlinear optimization approach,
taking into account the positions of the door and the window, as shown in table 1.
An objective function is determined to minimize the difference between actual
and measured number of occupants in the room. Optimization covers a long
period of time but it can be imagined that less representative observations could
be sufficient.
The next step is to use these adjusted parameters for calculating the
number of occupants over a time quantum lasting 30 minutes. Occupancy
estimation is obtained from equation (2). Finally, the last step is to use this
estimation of occupants as one feature in the classification model.
Deciding the number of occupancy levels. In this section, how to choose the
number of levels (L) of occupancy for classification is discussed. This number is
not fixed and can be changed according to the required average error (average
distance between actual occupancy numbers and the mid points of estimated
levels). To determine the number of levels and related non overlapping ranges of
occupancy, training data are partitioned into L clusters with 2 ≤ ≤ , where
N is the maximum possible number of occupants. At L = 2, the problem amounts
© ASCE
Sustainable Human–Building Ecosystems 78
too classify presence aand absencce of peopple. Table 2 shows the differeent
diiscretization
ns considereed (N = 4).
R
RESULTING
G OCCUP
PANCY EST
TIMATOR
RS
The C4.5
C decisioon tree algoorithm has been
b used to
o perform reecognition by
b
ussing aggreggated featurees and the labels extraacted from video cameeras. Traininng
daata cover 11 days from 4 May 2015 to 144 May 20155 while tessting data are a
coollected oveer for 4 dayys from 17 May 2015 to 20 May 2015. Oveer the traininng
peeriod, 12000 00 data poinnts have beeen collectedd.
Figurre 1, left sidde, shows thhe result obbtained fromm the learnt decision trree
coonsidering the
t basic sett of featuress (table 3), aas input to the
t detectio on model. The
T
pllot shows both
b actual occupancy
o profile and the estimaated profile as a graph of
nuumber of occupants
o w
with respectt to time (qquantum tim me is 30 minutes).
m T
The
avverage errorr yields to 0.32
0 occupannt.
Figurre 1, right siide, shows the result oobtained froom the decission tree aft
fter
coonsidering the two additional
a f
features of audio miccrophone detection
d annd
occcupancy fromCO2
fr phhysical moodel, in adddition to th he previous basic set of
feeatures. CO2 2 average value
v and CO2
C moved from the initial set
derivatiive are rem s
off features annd replaced by the estim mation of occupancy frromCO2 ph hysical moddel,
eqquation (2). Consideriing these features,
f leaads to impprovement in i occupanncy
esstimation with an averaage error of 0.24occupaant.
Both acoustic preessure (figu ure 2, left sidde) and occcupancy from
m CO2
phhysical mod del (figure 2,
2 right side)) are observved to be on ne of the moost importannt
feeatures for occupancy
o c
classificatio n, accordingg to the finaal Decision tree
n which rankks the featurres ascendinngly due to informationn gain for
cllassification
eaach feature, (figure 3,riight side). Acoustic
A preessure improoves the esttimation in
© ASCE
Sustainable Human–Building Ecosystems 79
occcupancy att high levelss while occuupancy fromm CO2 physsical model decrease thhe
w
whole averag ge error in thhe classificaation.
Finallly, (figure 3, left side) show ws the ressults of average
a errror
coorrespondinng to each leevel. Accorddingly, 5 levels of occuupancy is thhe best option
foor the occuppancy classification.
CONCLUSIIONS
R
REFERENC CES
A
Abhay Aroraa, Manar Amayri,
A V.. R. B. S. P. and Baandyopadhyyay, S. 20115.
Estim
mating occuupancy in an office setting. 1BS-20155 Secretariat,
Hyderabad, Indiaa.
© ASCE
Sustainable Human–Building Ecosystems 80
Agarwa l, Y., Balaji, B., Gupta, R., Lyles, J., Wei, M., and Weng, T. 2010.
Occupancy driven energy management for smart building automation. In
Proceedings of the 2nd ACM Workshop on Embedded Sensing Systems for
Energy-Efficiency in Building, pages 1–6. ACM.
Aglan, H.2003. Predictive model for co2 generation and decay in building
envelopes. JOURNAL OF APPLIED PHYSICS, 93(2).
ASHRAE, Atlanta, G. 1985. Fundamentals American Society of Heating,
Refrigerating and Air-Conditioning Engineers. Fundamentals American
Society of Heating, Refrigerating and Air-Conditioning Engineers.
Erickson, V. L., Carreira-Perpin˜a´n, M. A´., and Cerpa, A. E. 2011. Observe:
Occupancy-based system for efficient reduction of hvac energy. In
Information Processing in Sensor Networks (IPSN), 2011 10th
International Conference on, pages 258–269. IEEE.
Kashif, A., Dugdale, J., and Ploix, S. 2013. Simulating occupants’ behaviour for
energy waste reduction in dwellings: A multi agent methodology.
Advances in Complex Systems, 16:37.
Lam, K. P., H¨oynck, M., Dong, B., Andrews, B., shang Chiou, Y., Benitez, D.,
and Choi, J. 2009. Occupancy detection through an extensive
environmental sensor network in an open-plan office building. In Proc. of
Building Simulation 09, an IBPSA Conference.
Milenkovic, M. and Amft, O. 2013a. An opportunistic activity-sensing approach
to save energy in office buildings. In Proceedings of the fourth
international conference on Future energy systems, pages 247–258. ACM.
Milenkovic, M. and Amft, O. 2013b. Recognizing energy-related activities using
sensors commonly installed in office buildings. Procedia Computer
Science, 19:669 677.
Nguyen, T. A. and Aiello, M. 2012. Beyond indoor presence monitoring with
simple sensors. In PECCS, pages 5–14.
Padmanabh, K., Malikarjuna V, A., Sen, S., Katru, S. P., Kumar, A., Vuppala, S.
K.,Paul, S., et al. 2009. isense: a wireless sensor network based conference
room management system. In Proceedings of the First ACM Workshop on
Embedded Sensing Systems for Energy-Efficiency in Buildings, pages 37–
42. ACM.
Page, J., Robinson, D., and Scartezzini, J. 2007. Stochastic simulation of occupant
presence and behaviour in buildings. Proc. Tenth Int. IBPSA Conf:
Building Simulation, pages 757–764.
Philipose, M., Fishkin, K. P., Perkowitz, M., Patterson, D. J., Fox, D., Kautz, H.,
and Hahnel, D. 2004. Inferring activities from interactions with objects.
Pervasive
Computing, IEEE, 3(4):50–57. Quinlan, J. R. 1986. Induction of decision trees.
Machine learning, 1(1):81–106.
Robinson, D. and Haldi, F. 2009. Interactions with window openings by office
occupants. Energy and Buildings, 44:2378–2395.
Roulet, C., Fritsch, R., Scartezzini, J., and Cretton, P. 1991. Stochastic model of
inhabitant behavior with regard to ventilation. Technical report.
© ASCE
Sustainable Human–Building Ecosystems 81
Abstract
INTRODUCTION
© ASCE
Sustainable Human–Building Ecosystems 82
© ASCE
Sustainable Human–Building Ecosystems 83
, , ,
. . … .
. = … … ⋱ …
, , ,
. . … .
© ASCE
Sustainable Human–Building Ecosystems 84
,
where . denotes the movement from zone to zone between time and
+ 1. . is calculated based on the user specified human movement model and
the state vector . The update equation of is = + . , in which
is a n dimensional column vector with all one entries.
2) The Sensors: It is assumed that there are only two kinds of sensors, namely
in-zone sensors and boundary sensors. Although there are a plentiful types of
different sensors which are capable of providing the information about occupants,
most of these sensors may fall into these two types.
The in-zone sensors can provide the information of the number of occupants
in a zone. At time , the map from the distribution vector = { , … , } to the
observation value is formulated as a probability model. The observation value of
in-zone sensors is denoted as ={ ,…, } , where represents the
observation of the in-zone sensor in zone at time . In our simulation model, it is
assumed that the probability distribution of only depends on . To put the
formulation more clearly, the video camera is used as an example to show our model
of the in-zone sensors. In Wang et al. (2010), the estimation value of the number of
occupants in a zone given by a video camera is modeled as a sum of the true number
of occupants and a Gaussian noise. Generally speaking, the observation of an in-zone
sensor is a random variable, and its distribution depends on the number of occupants
in the zone. So is a function of , and the random vector , namely
= ( , ). The function is specified by the user, using the interface provided
by our platform.
The other type of sensors is called the boundary sensor. When an occupant
goes across the boundary, the boundary sensor provides related information. The
boundary beam sensor in Song et al. (2008) and the camera video in Chen et al.
(2008) are both widely used boundary sensors. From time to + 1, a boundary
sensor observation matrix
, , ,
. . … .
. = … … ⋱ …
, , ,
. . … .
© ASCE
Sustainable Human–Building Ecosystems 85
© ASCE
Sustainable Human–Building Ecosystems 86
Then { } and { } are inputt into the modules m Booundary Sensor Detecction
Sim
mulation andd In-zone Sensor
Se Detecction Simullation respectively as shown
s in Fiig 2.
Theen the Bou undary Sennsor Observvation Dataa { . | = 1, … , 1} and the
In-zzone Sensor Observattion Data { . | = 1, … , 1} are gen nerated by this
module.
Based on the asssumption inn part III.A A that the observation ns of diffeerent
sennsors are independent, the simulaation proceddure of the sensors coould be diviided
intoo smaller sim
mulation prroblems. In each sub-pproblem, onlly a part of the sensorss are
mulated. Thiis property of data sepparability makes
sim m it eaasy to simu ulate the sennsor
deteection in paarallel. In our
o platform m, ”Matlab Parallel
P Com mputing Tooolbox” is used
u
to implement
i the parallell simulationn of the Booundary Sennsor Detecttion Simulaation
andd In-zone Seensor Detection Simulaation module.
The and are then innput to thee module Estimation
E t generate the
to
Esttimated Occcupant Disttribution daata. In the eend, this daata and the Real R Occuppant
Disstribution daata are inpuut into the Evaluation
E module to generate thhe Performaance
dataa of the seensing system. In the Estimationn module, the estimattion algoritthm,
whiich is relateed to the occupant
o movement model
m and sensor
s detecction model, is
given by the users. In thhe Evaluatiion modulee, the valuees of severaal performaance
mettrics are prrovided, inccluding the mean erroor rate, fivee-number suummary off the
estiimation erroor, operationn cost of thee sensing syystem per yeear, etc.
EX
XPERIMEN
NTS
A. The Efficiency Imprrovement of the Simu ulation Plattform. A 1000 zones and
10000 occupantts scale probblem is testted to show w the improv
vement in efficiency
e ussing
parrallel simulaation. The result is shown in Fig. 3, and it sh
hows that wiith the increease
of tthe number of CPUs, thet computiing time deccreases inveersely propo ortionally. This
T
couuld be explaained by thhe Amdahl’s law. In Fig. F 3 we can c see thatt using paraallel
sim
mulation impproves the efficiency
e siignificantly.
F
Figure 3. Im
mprovemen
nt in Efficieency using Parallel
P Sim
mulation
© ASCE
Sustainable Human–Building Ecosystems 87
B. A Case Stu udy. A case study on thhe 1st floor of FIT buillding, Tsingghua Univerrsity
shoows that ouur simulatioon platform m could be used to simulate diffferent kindss of
models and alggorithms. TheT 1st floor of the FIT T building contains
c 81 zones, andd the
movements off 200 occupaants from 8:00 8 am to 88:00 pm in a day are siimulated in our
case study.
Three kinds
k of sennsors are coonsidered inn the case study,
s whichh are the viideo
cammera sensor, the infraared beam sensor andd the passiv ve infraredd sensor. Inn the
sim
mulation, thee video cam mera is used to count thhe number of o occupantss in a zone, and
its rresult is moodeled as a sum of the true numbeer of occuppants and a Gaussian noise
N(0 0,0.5 ). Theen the probability of th he camera too get the truue number ofo the occuppant
is 668%. The infrared
i beaam sensor is used to monitor
m the boundary event.
e Whenn an
occcupant enterrs or leavess a zone, thhe infrared beam senso or could deetect this evvent.
How wever, it suuffers fromm a 20% probability
p o miss dettection erroor. The passive
of
infrrared sensoor detects thet in-zonee occupant movementt. It is asssumed that the
proobability of the sensor tot detect onne occupant is 0.75, an nd the probaability to deetect
morre that one of the n occcupants in th he zone is 1 0.25 .
Using the detectioon results of o these thrree kinds ofo sensors, three kindss of
estiimation algoorithms are used in this case studyy. The first algorithm, which is caalled
Naïïve Estimattion, uses thhe results ofo the infrareed beam sennsors only. When a sennsor
deteects an arrival or a depparture, thee estimationn number pllus one or minus
m one. The
second algoritthm is Infrared Trigggered Videeo Counting. It uses the t informaation
fromm both thee passive innfrared sen nsors and tthe video cameras.
c W
When a passive
infrrared sensorr detects thee in-zone movement
m inn a zone, theen the videoo camera inn the
zonne will be trigger
t to count
c the number
n of occupants
o in
i the zonee. This kindd of
cooordination policy
p is maainly for thee purpose off energy sav ving. The laast algorithm
m is
callled Naïve Estimation
E with Revission, and it uses all thee results of the three kiinds
of ssensors.
Figgure 4. Simu
ulation Ressult of the Case
C Studyy
© ASCE
Sustainable Human–Building Ecosystems 88
The result of the case study is shown in Fig. 4, we could see that the more
number of sensors are used, the better the estimation result is, which accords with the
intuition. And this case study shows that our simulation platform could integrate
different models and different algorithms.
CONCLUSION
ACKNOWLEDGEMENT
REFERENCES
© ASCE
Sustainable Human–Building Ecosystems 89
IEEE Conf. DEST, Menlo Park, CA, USA , Jul. 24–26, 2013.
V. Garg and N. K. Bansal, “Smart occupancy sensors to reduce energy consumption”,
Energy and Buildings, vol. 32, no. 1, pp. 81–87, 2000.
G. Gunathilak, A. Prasannakumar, N. Nazarian, H. Naeimi, “A Generalized Event
Driven Framework for Building Occupancy”, presented at Symp. on
Simulation for Architecture and Urban Design, San Diego, California, USA,
2013.
J. Hutchins, A. Ihler, P. Smyth, “Modeling count data from multiple sensors: a
building occupancy model”, presented at 2nd IEEE Int. Workshop on
CAMPSAP, St. Thomas, VI, Dec. 12–14, 2007.
C. Liao and P. Barooah, “An integrated approach to occupancy modeling and
estimation in commercial buildings, presented at ACC, Baltimore, MD, Jun.
30–Jul. 2, 2010.
S. Meyn, A. Surana , Y. Q. Lin, and S. M. Oggianu,“A Sensor-Utility-Network
Method for Estimation of Occupancy in Buildings”, presented at 48th IEEE
CDC., Shanghai, P.R. China, Dec. 16–18, 2009.
J. Page, D. Robinson, N. Morel and J. L. Scartezzini, “A generalised stochastic
model for the simulation of occupant presence”, Energy and Buildings, vol.
40, no. 2, pp. 83–98, 2008.
J. Song, Y. F. Dong, X. W. Yang, J. H. Gu and P. P. Fan, “Infrared Passenger Flow
Collection System Based on RBF Nueral Net”, presented at Proceedings of
the Seventh International Conference on Machine Learning and Cybernetics,
Kunming, China, Jul. 12–15, 2008.
D. Wang, C. Federspiel and F. Rubinstein, “Modeling occupancy in single person
offices”, Energy and Buildings, vol. 37, no. 2, pp. 121–126, Feb. 2005.
H. T. Wang, Q. S. Jia , C. Song, R. X. Yuan and X. H. Guan, “Estimation of
Occupancy Level in Indoor Environment Based on Heterogeneous
Information Fusion”, presented at 49th IEEE CDC., Atlanta, GA, USA, Dec.
15–17, 2010.
H. T. Wang, Y. L. Lei and X. H. Guan, “Estimation of Occupant Distribution by
Detecting the Entrance and Leaving Events of Zones in Building”, presented
at IEEE Conf. MFI, Hamburg, Germany, Sep. 13–15, 2012.
© ASCE
Sustainable Human–Building Ecosystems 90
Abstract
INTRODUCTION
© ASCE
Sustainable Human–Building Ecosystems 91
© ASCE
Sustainable Human–Building Ecosystems 92
© ASCE
Sustainable Human–Building Ecosystems 93
necessary to be met when the zone is unoccupied. However, when a zone becomes
unoccupied, instant control adjustments may cause discomfort, as an occupant may
reoccupy the zone at any time and reconditioning from a deep setback to a setpoint
after a relatively short period of vacancy may consume more energy than just
maintaining the setpoint. Therefore, a terminal control could be divided into four
periods: (1) the setpoint period: a terminal works to maintain the setpoint, loads are
considered effective; (2) the float period: a terminal is off and the temperature floats
from setpoint to setback, loads are considered ineffective; (3) the setback period: a
terminal works to maintain the setback, loads are considered effective; (4) the
reconditioning period: a terminal works to restore temperature from setback to
setpoint, loads are considered effective (Figure 2). In sum, the transitions between
occupied/unoccupied statuses do not necessarily follow the transitions between
effective and ineffective loads. A portion of the loads during unoccupied periods
should be considered effective for improving energy efficiency.
As shown in Figure 2, the time lag and limit for temperature to float
determine the transitions of effective loads and ineffective loads. They are called
setpoint/setback schedules (e.g., waiting time to trigger setback) and distances
(difference between setpoint and setback) in this paper. Conceptually, different
combinations of setpoint/setback schedules and distances may lead to significantly
different levels of energy efficiencies. The synergetic effects of setpoint/setback
schedules and distances are investigated for energy efficiency. Energy efficiency in
this paper incorporates both energy reduction, which is the absolute amount of energy
savings, and the conditioning miss, which is the length of time during which a space
is occupied but the temperature is outside the comfort range. Conditioning miss is
considered as an equally important component as the energy reduction because it
compromises the basic function of an HVAC system to maintain a comfortable
thermal environment. 23oC is used as the static setpoint when the zone is occupied.
The interval for setpoint/setback distance is 2 K and the interval for setpoint/setback
schedule is 5 minutes. Both setpoint and setback have a deadband of ±1 K and the
conditioning miss is the length of time a zone is occupied but the temperature is
outside the range of [22oC, 24oC] in order to be in compliance with the PMV
(Ghahramani et al. 2015). It is important to note that, the ideal heating/cooling loads
are studied in this paper, thus the loads from HVAC components (primary/secondary
systems and terminals) are neglected. For example, an assumption used in this paper
is that there is no energy loss in the transportation of air. Ventilation is considered to
provide airflow per the ASHRAE requirement (ASHRAE 2010).
© ASCE
Sustainable Human–Building Ecosystems 94
Variant Occup pancy Tran nsitions and d Energy Effficiency. T The energy im mplications
of discrete com mbinations ofo setpoint/ssetback scheedules and ddistances aree simulated
usin ng whole buuilding energ gy simulation. The weigghted sums oof energy redduction and
connditioning miss
m represeent the energy efficienncy of diffe ferent combinations of
setppoint/setbackk schedules and distancees. Since thee presence oof occupants depicts the
onee-time occurrrences of occcupant preseence/absencee changes, thhe variant trransitions of
occcupant preseence mean the differences of occcupied/unocccupied trannsitions for
diffferent days. In order to analyze the effects of vaariant occuppancy transittions on the
eneergy efficienncy of setpoint/setback controls,
c twoo sets of sim
mulations are conducted:
1) is using actual occup pancy with variant traansitions (eaach day haas different
occcupancy), wh hile 2) is usiing repeatedd occupancy with constannt transitionns (each day
hass the same occupancy).
o There are N possible sscenarios foor the second set if the
perriod has N daays, and each h scenario iss called one sample repreesenting onee possibility
of occupancy
o happening
h duuring that peeriod withouut variant traansitions. Thherefore, the
firsst objective of
o this paperr is to test whether
w 1) annd 2) have ssignificant ddifference in
eneergy efficienncy of differeent setpoint//setback scheedules and ddistances. A process for
commparison is designed
d and shown in Figure3. Zoone level preesence is dettermined by
agggregating thee occupancy of rooms in n that zone: ((1) if at leasst one room in a zone is
occcupied, the zone is occup pied; (2) if alll rooms are vacant, the zone is unocccupied.
First thee energy effficiency resuults of differrent setpointt/setback conntrols using
actual occupan ncy are obtaained, and th he estimationn interval fofor each com mbination is
calcculated withhin ±2.5% off the simulatted energy effficiency givven the preddefined 95%
connfidence leveel. Then thee repeated occupancy is used to repplace actual occupancy,
andd the simulaation is ran for
f N times, each time uusing one’s occupancy for N days.
Forr each comb bination of setpoint/setb
s ack schedulle and distannce, the percentages of
eneergy efficien ncy levels of all sam mples (all ppossible sceenarios usinng repeated
occcupancy with hout variant transitions) within the iintervals from m the actuall occupancy
are calculated (it is calleed coverage percentagee in this paaper). The ssmaller the
covverage perceentage is, the more sign nificant the iinfluence thhe variant traansitions of
occcupant preseence on thee absolute energy efficciency of ddifferent schhedule and
disttance combinations.
Fig
gure 3. Comp
parison of en
nergy impliccations betweeen actual annd repeated occupancy
Variant Occup pancy Tran nsitions andd Relative PPerformancee. The seconnd objective
is to
t test whethher actual occcupancy with variant ttransitions aand repeatedd occupancy
witthout varian
nt transitionss have sign
nificant diffeerence in reelative perfoormance of
diffferent setp
point/setbackk schedules and diistances coombinations. Relative
perrformance is defined as the performance of one combinatioon for energyy efficiency
© ASCE
Sustainable Human–Building Ecosystems 95
compared to the best and worst combination in the same scenario. The difference
between absolute energy efficiency and relative performance is: absolute energy
efficiency shows the actual energy efficiency that can be improved, while the relative
performance evaluates differences of selected combinations from the remaining. Even
if the absolute energy efficiency levels from repeated occupancy and actual
occupancy are different, the combinations of setpoint/setback schedule and distance
that outperform other combinations could still be the same (e.g. [20%,40%,60%,80%]
and [17%,29%,41%,54%] are significantly different, but they are similar after
normalization). Therefore, the simulated energy implications of discrete
setpoint/setback combinations are normalized by ∗ = to indicate the
relative energy performance of different setpoint/setback controls. The same process
of comparison analysis (Figure 3) is then applied to test whether there is significant
difference between actual and repeated occupancy in terms of the normalized energy
implications. Similarly, the coverage percentage is calculated, and the smaller the
percentage is, the more significant the influence the variant transitions of occupant
presence on the relative performance of different schedule and distance combinations.
The test bed building used in this paper is the Ralph & Goldy Lewis Hall
(RGL), a typical educational building on the University of Southern California
campus near downtown Los Angeles, California. It is a three-story building with a
footprint of 3,735 m2, and contains 89 mechanically ventilated rooms. The 16 zones
on the third floor consist of 28 rooms and were monitored by 28 wireless sensor units,
each of which includes a number of ambient sensors such as temperature sensor and
CO2 sensor for modeling real-time occupancy (Yang et al. 2013). The ideal loads
(system energy loss is neglected) for the third floor during May 2013 to April 2014
were simulated to implement the proposed methodology. One year was divided into
four periods based on the outside temperature statistics (mean and standard deviation).
23oC (with 1 K as deadband) was used as the static setpoint to maintain a comfortable
thermal environment when the zone was occupied. When there was no occupant, six
different setback values (2 K as the interval) and six setback waiting time (with 5 min
as the interval) were combinated. In this paper, energy reduction and conditioning
miss were considered equally important, thus the energy efficiency was expressed as
50%*(energy reduction)+50%*(conditioning miss). The period of N days for testing
the energy significance of variant occupancy transitions was three-month long and
therefore there were four periods to analyze the consistency of findings.
For each period, the influences of variant occupancy transitions on the energy
efficiency of different setpoint/setback schedules and distances were first investigated.
Actual occupancy and repeated occupancy were used and statistics of coverage
percentages for all 36 combinations were shown in Table 1. The differences of energy
efficiency caused by variant occupancy transitions were around 60%, indicating that
the variant transitions of occupant presence had significant influence on the absolute
energy efficiency of different setpoint/setback schedules and distances. The
normalized energy implications of different setpoint/setback controls from actual
occupancy and repeated occupancy were then compared to investigate the influences
© ASCE
Sustainable Human–Building Ecosystems 96
Tab
ble 1. Coverrage percen ntages of eneergy efficien
ncy for all ccombination
ns
Eneergy Efficien
ncy Relative P Performancee
Mean
M SD Max. Min. M Meaan SD Maax. Min.
Maay-Jul (1) 45% 19% 75% 21 1% May-Juul (1) 94% % 3% 88% % 98%
Aug-Oct (2) 38% 15% 69% 20 0% Aug-Oct (2) 90% % 4% 85% % 95%
Nov-Jan (3) 42% 18% 71% 17 7% Nov-Jaan (3) 91% % 3% 87% % 97%
Feb
b-Apr (4) 40% 16% 73% 19 9% Feb-Appr (4) 92% % 3% 88% % 96%
The covverage perceentages of reelative perforrmances for four periodds were then
stattistically com
mpared usin ng paired T-tests. The results (Tabble 2) show wed that the
covverage percentages of relative peerformances were signnificantly diifferent for
diffferent periodds. Besides, the
t four periiods had diff fferent outsidde temperatuure statistics,,
andd it was foun nd that the difference
d off average temmperature beetween the ttwo periods
wass approximaately linearly associated d with the ddifference oof coverage percentage
betwween the saame periods.. The joint effects
e of ouutside temperature and the variant
occcupancy tran nsitions on energy
e efficciency were more signiificant than on relative
perrformance off setpoint/settback schedu ule and distaance combinations.
Tab
ble 2. Statisstical analyssis of covera
age percentaage differen
nces between
n periods
Relative Performance
P e
Compariison Sig.
Period 1-2
1 0.05
Period 1-3
1 0.02
Period 1-4
1 0.01
Period 2-3
2 0.00
Period 2-4
2 0.00
Period 3-4
3 0.03
CO
ONCLUSIONS
© ASCE
Sustainable Human–Building Ecosystems 97
ACKNOWLEDGMENTS
REFERENCES
© ASCE
Sustainable Human–Building Ecosystems 98
© ASCE
Sustainable Human–Building Ecosystems 99
Abstract
INTRODUCTION
© ASCE
Sustainable Human–Building Ecosystems 100
HVAC systems are responsible primarily for providing satisfactory thermal comfort
conditions. Thermal comfort is one of the most influential factors on overall
satisfaction with indoor environments (Frontczak and Wargocki 2011). Standard
models for thermal comfort conditions (e.g., ASHRAE Standard 55 (Thermal
Environmental Conditions for Human Occupancy) (Standard 2004)) use few selected
environmental and occupant related parameters (e.g., indoor air temperature, airflow
rate, activity, etc.) to predict occupants’ thermal comfort. In addition, standard models
are not adaptive to time dependent variations while individuals’ thermal comfort
ranges may change over time due to many dynamic environmental and human related
variables (Jendritzky and de Dear 2009). Moreover, it has been shown that humans
perceive comfort in a range of environmental thermal conditions (Nicol and
Humphreys 2002). The fact that adjusting the HVAC temperature set point by 1°C
have considerable impacts on the overall energy consumption (Hoyt et al. 2014; Yang
and Becerik-Gerber 2015) suggests that learning this range can potentially lead to
advanced control of building HVAC system for reducing energy consumption.
Human thermal comfort could be obtained using two types of approaches: (1)
survey based approaches; and (2) physiological measurement based approaches.
Survey based approaches use a participatory learning process through a
questionnaire, while physiological measurements based approaches utilize certain
physiological measurements (e.g., heart rate, skin temperature, etc.). Application of
these two categories of approaches requires real-time and continuous monitoring of
building occupants, making it a challenging task. In order to address this challenge, a
correlation analysis could be performed between instant comfort levels driven from
previously mentioned approaches with some other variables, such as environment
related variables (e.g., indoor air temperatures, clothing levels). Consequently, the
selected correlated variables could be used to estimate occupants’ thermal comfort
levels.
In this paper, we first describe our adaptive stochastic modeling approach for
modeling personalized thermal comfort of building occupants (Ghahramani et al.
2015). We then demonstrate the results of time dependent variations of thermal
comfort studying the data from 33 test subjects and statistically analyzing the results.
First, we provide a review of recent studies on personal thermal comfort learning and
modeling. Next, we describe our stochastic approach for modeling personalized
thermal comfort, as well as our time dependent variation analysis procedure. We then
explain the experimental design and present the results for time dependent variations
using the data from the subjects. Finally, we summarize the findings and conclude the
paper.
© ASCE
Sustainable Human–Building Ecosystems 101
few parameters that are easier to monitor (e.g., ambient factors) and occupants’
comfort. For example, in (Liu et al. 2007), a deep artificial neural network (ANN)
learning technique was used for classifying environmental conditions into
comfortable, uncomfortably warm, and uncomfortably cool. The ANN algorithm had
4 input variables (i.e., air temperature, radiant temperature, air flow, air humidity).
The algorithm was trained with comfort votes from subjects under controlled
experiments. However, similar to other studies in literature, the time dependent
variations of thermal comfort were not considered in this study. Researchers also
focused on adaptive thermal comfort modeling. For example, an adaptive technique,
which uses the PMV model as a prior model, was introduced in (Yao et al. 2009).
The model calculated an adaptation coefficient, which decreases or increases the
estimated PMV values. The adaptation coefficient took into account local climate,
culture, and social backgrounds.
Yet, the majority of the efforts in the literature lack the components for
detecting time dependent variations (preference variations over time due to hidden
variables) in thermal comfort. In a previous effort (Ghahramani et al. 2015), we
developed a modeling technique that enabled us detecting time dependent variations
in a systematic manner with no prior assumptions about occupant preferences. This
modeling technique used the data from occupants and the environment without any
pre-defined bias on the personal comfort. The modeling technique has an online
learning structure: when any time a new data point is communicated by an occupant
(i.e., comfort vote), we record the associated environmental condition and the model
gets updated. This learning process improved the accuracy of learning personal
comfort. The opportunity that this modeling technique provides is to mathematically
demonstrate the compliance of comfort predictions with the requirements stated in
standards (ASHRAE Standard 55). This work motivated us to study how personal
comfort varies over time and/or how the range of comfortable temperatures for each
person varies over time.
In this section, we briefly describe our comfort modeling approach that uses
internal parameters to capture the variations in an individual’s thermal comfort
preferences and explain how we quantify the variations, which is the focus of this
paper. The input data to the modeling approach consist of thermal comfort votes and
associated air temperatures. Thermal votes are divided into three categorical
variables: uncomfortably warm; comfortable; and uncomfortably cool. Figure 1
shows a sample dataset for an individual.
We first transform comfort votes/temperature data into a parametric
mathematical model. Upper Limit (UL) and Lower Limit (LL) for temperatures that
comfort can potentially be perceived by a subject and three probability distributions
for uncomfortable and comfortable conditions (LD – Lower Distribution, MD –
Middle Distribution, UD – Upper Distribution) are the parametric models that are
generated. UL is defined as the highest temperature that a subject has communicated
a comfort vote. LL is defined as the lowest temperature that a subject communicates a
comfort vote. LD, MD, and UD are the probability distribution functions defined for
the uncomfortably cool, comfortable, and uncomfortably warm data points,
© ASCE
Sustainable Human–Building Ecosystems 102
resppectively. We
W then inteegrate the efffects of diffferent therm mal comfortt conditions
into
o a single jooint distributtion (JD) by y defining a Bayesian nnetwork to ccombine the
pro
obability disttributions ovver the rang ge of temperratures that evidence suuggests that
commfort can potentially
p be
b perceiveed [LL to UL]. A B Bayesian nettwork is a
pro
obabilistic graphical mo odel (a direected acyclicc graph) thhat representts a set of
ran
ndom variab bles and theeir condition nal dependeencies. The network inn our work
represents thee probabilisstic relation nships betw ween the influential probability
disttributions (L
LD, MD, an nd UD) and d the overalll comfort. A ASHRAE S Standard 55
(Staandard 2004 4) uses the PMV-PPD model to ddefine the reequirements for indoor
therrmal condittions. This standard requires that percentage of dissatisffied people
(PP
PD) to be leess than 20% %, which im mplies that aat least 80% % of the occupants in a
buiilding to be satisfied. Based
B on thee triangle innequality, if we set the probability
threeshold of ouur proposed approach to o 80%, the e xpected percentage of tthe satisfied
occcupants wo ould be grreater than 80%. Theerefore, thee ASHRAE E standard
req
quirements would
w be meet. Thus wee set probabbility threshoold (PT) as the rule of
classsification to
t be 0.8. This
T fact co
ould be usedd to alter thhe conventional multi-
objective (e.g.,, comfort an nd energy) optimizationn problems for buildinngs’ HVAC
systems by transforming comfort ob bjectives too constrain functions oof a single
opttimization problem
p as demonstrateed in ((Ghaahramani ett al. 2014; Yang and
Beccerik-Gerberr 2014)). Fig gure 1 illustrrates the diffferent compoonents of thee model.
Fiigure 1. Seg
gmentation of
o data and PT (Probab bility Thresshold) as a cconstraint
for classifying comfo
ort vs. discoomfort condditions
© ASCE
Sustainable Human–Building Ecosystems 103
data for each subject. In the results section, we demonstrate the range for each
individual at every data point interval. In addition, we calculated the point that had
the highest probability of comfort and averaged over the duration of the experiment
for each subject, as well as the number of data points communicated. We also
calculated the variations in thermal preferences as absolute differences between the
previous and current most comfortable temperature values over an interval. The
intervals studied in this paper are data points or time (i.e., day). Consequently, we
calculated the data point-based variation analysis through calculating the difference in
red crosses in Figure 2 and averaged over all data points. For calculating the daily-
based variations, we multiplied the data point based calculation by the average data
points per day for each test subject.
EXPERIMENT PROCEDURE
RESULTS
© ASCE
Sustainable Human–Building Ecosystems 104
Thee average daata points co ommunicated per day w was 4.23. Thhe major obsservation is
thatt thermal preeferences off individuals change conssiderably evven on small time scales
(i.ee., day to day).
d Consequently, perrsonal therm mal comfortt modeling techniques
sho ould considerr inherent tim
me variation
ns of preferennces as a maajor factor.
Figure
F 2. Six
x sample sub
bjects’ thermal prefereence variations over daata points
© ASCE
Sustainable Human–Building Ecosystems 105
In this paper, we demonstrated how personal thermal comfort varies over time
through studying thermal preferences of 33 subjects. We briefly described the
adaptive stochastic modeling technique that was used to quantify personal thermal
comfort. Our stochastic models are probability distributions in a dynamic Bayesian
network that utilizes a sliding window based algorithm for detecting significant
statistical differences in joint probability distributions. By applying the requirements
for standard ASHRAE 55 to the approach, we calculated comfortable temperature
ranges for each individual as they vary over time. We then calculated the absolute
© ASCE
Sustainable Human–Building Ecosystems 106
difference of comfortable temperature ranges to the previous data point and day. Our
results suggest that personal preferences have considerable variations over time and
thus are not negligible. The average variation was 0.0606 °C with a high standard
deviation of 0.1591 °C. This finding not only shows that personal comfort should be
tracked over time (time is not defined explicitly), but also suggests that comfort
variations vary from person to person.
ACKNOWLEDGEMENTS
REFERENCES
© ASCE
Sustainable Human–Building Ecosystems 107
© ASCE
Sustainable Human–Building Ecosystems 108
Jie Zhao1; Khee Poh Lam1; Vivian Loftness1; and B. Erik Ydstie2
1
Center for Building Performance and Diagnostics, School of Architecture, Carnegie
Mellon University, Pittsburgh, PA 15213. E-mail: jayzhao@cmu.edu
2
Department of Chemical Engineering, College of Engineering, Carnegie Mellon
University, Pittsburgh, PA 15213. E-mail: ydstie@cmu.edu
Abstract
Occupant thermal comfort is a key performance metric for green buildings.
Although numerous studies have used thermal comfort as a criterion to control
mechanical systems and optimize their energy consumptions, few studies examined
individual thermal comfort preferences and their implications. This study aims to
investigate the individual thermal comfort preferences in an office space. An online
dashboard is developed to collect subjective thermal comfort data from office
workers. The data then is compared and correlated with the measured key
environmental data, such as air temperature and relative humidity. The data analysis
results suggested that different occupants have various preferences about thermal
comfort yet each individual has fair consistent thermal preference over time. In future
study, the dashboard could be linked to the real-time control system to provide
individual thermal control for office workers.
INTRODUCTION
Providing thermal comfort to occupants is one of the most critical objectives
of green buildings. Predicted mean vote (PMV) which can be calculated by six major
factors – air temperature, relative humidity, mean radiant temperature, air velocity,
metabolic rate, and clothing insulation, are typically used as a metric to quantify
occupant thermal comfort (ASHRAE, 2010). A number of studies have used PMV as
a control criterion in the heating, ventilation, and air conditioning (HVAC) system
(Daum, Haldi, & Morel, 2011; Murakami, Terano, Mizutani, Harada, & Kuno, 2007;
Yang & Su, 1997; Yu, Loftness, & Yu, 2013).
PMV could be used to define the average thermal comfort level in a group
using calculation methods, but could also be used as a subjective metric to evaluate
the thermal sensation of each individual. Several studies have indicated that
occupants do not have uniformed thermal sensation in the same thermally controlled
space (Klein et al., 2012; Park, 2013). In order to meet the thermal comfort
requirements of different occupants, several studies have developed individual
thermal comfort voting systems to feedback the subjective thermal preference
information into HVAC control systems (Daum et al., 2011; Klein et al., 2012; Lee &
Malkawi, 2014; Murakami et al., 2007). Choi (2010) developed a wearable bio-
sensing adaptive control system that can collect thermal measurement data from
© ASCE
Sustainable Human–Building Ecosystems 109
ind
dividual bio-ssensors to auutomaticallyy adjust HVA AC system ssetpoints (C Choi, 2010).
Thee experimen nt and simulaation resultss of these preevious studiies show thaat providing
occcupants witth individuaal thermal controls nnot only caan improvee occupant
satiisfaction (Chhoi, 2010; Murakami
M ett al., 2007) aand comfortt (Choi, 2010; Daum et
al., 2011; Kleiin et al., 20012; Lee & Malkawi, 22014), but aalso impact the HVAC
eneergy consum mption (Cho oi, 2010; Daaum et al., 2011; Kleinn et al., 20012; Lee &
Maalkawi, 2014; Murakami et al., 2007)).
udy aims to investigate the individuual thermal comfort preeferences in
This stu
an office spacce. An onlin ne dashboarrd is develooped to colllect subjectiive thermal
commfort data frrom office workers.
w Thee data then iss compared and correlatted with the
measured key environmenttal data, succh as air tem mperature annd relative hhumidity, in
ord
der to identify
fy the thermaal comfort prreference diffferences of the office w
workers.
EX
XPERIMEN
NT SETUP
© ASCE
Sustainable Human–Building Ecosystems 110
factor with the question of “What are you wearing?” The thermal comfort band
ranges from -2 to 2 based on the PMV definition. The clothing insulation band ranges
from 0.4 to 1.9, representing from “shorts & T-shirts” to “jacket and overcoat”. The
occupants are required to provide the two pieces of information concurrently when
they are not feeling thermally comfortable. All the data are stored in a web-based
MySQL database with timestamps.
RESULTS
70
60
Comfort vote counts
50
40
30
20
10
0
A B C D E F G H I J K L M N O
Participated occupant
© ASCE
Sustainable Human–Building Ecosystems 111
26 Feeling cold
regardless of
22
20
-2 -1 0 1 2
Comfort vote (PMV)
Feeling cold
60
Relative humidity (%)
regardless of
humidit y
50
40
30
20
-2 -1 0 1 2
Comfort vote (PMV)
2.0
Clothing insulation (Clo)
1.0
0.5
-2 -1 0 1 2
Comfort vote (PMV)
© ASCE
Sustainable Human–Building Ecosystems 112
respectively. The histograms (N = 36) of the variables are shown on the top and right
sides of the plot areas. In general, Subject B feels cold for the majority of votes
regardless of actual air temperature (20 - 25°C) and relative humidity (20 - 60%)
measurements. Figure 8 shows that almost all the “feeling cold” votes (except for one
vote) are correlated with clothing insulation factor around 0.5, which suggests that
Subject B has relatively thin clothes in the office compared to the typical business
dress. Arguably, Subject B could have improved his/her thermal comfort level easily
by adding a jacket or another layer of clothes. This result is very different from the
data of Subject A, who has already had thick clothes but still feels cold.
26 Feeling cold
regardless of
Air temperature (°C)
temperature
24
22
20
-2 -1 0 1 2
Comfort vote (PMV)
60
Feeling cold
Relative humidity (%)
regardless of
50 humidity
40
30
20
-2 -1 0 1 2
Comfort vote (PMV)
© ASCE
Sustainable Human–Building Ecosystems 113
2.0
0.5
-2 -1 0 1 2
Comfort vote (PMV)
26
Air temperature (°C)
24
22
Feeling warm
20
in general
-2 -1 0 1 2
Comfort vote (PMV)
© ASCE
Sustainable Human–Building Ecosystems 114
60
Relative humidity (% )
50
40
30
20 Feeling warm
in general
10
-2 -1 0 1 2
Comfort vote (PMV)
Figure 10. Marginal plot of relative humidity vs. comfort vote of Subject C.
Marginal Plot of Clothing insulation (Clo) vs Comfort vote (PMV)
2.0
Clothing insulation (Clo)
1.5
1.0
Wearing less clothes
but still feeling warm
0.5
-2 -1 0 1 2
Comfort vote (PMV)
Figure 11. Marginal plot of clothing insulation vs. comfort vote of Subject C.
By comparing the three subjects’ data and taking into consideration of the
overall voting statistics, the following conclusions can be drawn. A total of 11 out of
15 volunteers express their opinions on their thermal comforts when they do not feel
comfortable, which indicates thermal comfort is something important for the test
group to consider at workplaces. However, people have very different sensitivities to
thermal comfort. Three subjects voted over 30 times during the three months
experiment periods, but most people voted less than 30 times. In addition, among the
three subjects, the data shows very different comfort preferences and clothing choices.
However, the thermal comfort preference and clothing choice of each individual are
mostly consistent for the majority of time during the entire experiment period.
Therefore, it is necessary to develop a control system that can accommodate
individual occupant thermal comfort preference, and based on the individual
preference, the control system could make consistent adjustment to meet individual
needs.
DISCUSSIONS
This study found very interesting patterns about individual thermal comfort
preference despite the small sample size (15 people) and the relative short duration
(three months). One could imagine that if future studies could incorporate more
© ASCE
Sustainable Human–Building Ecosystems 115
individuals under different seasons and locations that could also represent culturally
behavioral differences, the findings would be very valuable for understanding the
occupant thermal comfort and behavior, and to develop more appropriate occupant-
oriented control systems.
Future work could be focused on developing a control system that linking the
online dashboard into the actual HVAC zone level control, and testing its comfort and
energy performance.
CONCLUSIONS
ACKNOWLEDGEMENTS
REFERENCES
© ASCE
Sustainable Human–Building Ecosystems 116
© ASCE
Sustainable Human–Building Ecosystems 117
Abstract
INTRODUCTION
© ASCE
Sustainable Human–Building Ecosystems 118
Currenttly, most bu uilding therm mal environnmental conntrols and ssystems are
adoopting prediccted mean votev / predictted percentaage of dissattisfied (PMV
V and PPD)
models by utillizing heat balance
b equaations to esttimate therm
mal comfortt conditions
thatt affect a buiilding’s tenaants. Fanger’’s equations are used forr PMV calcuulations of a
larg
ge number of human samples with a particuular thermal condition.. This is a
commbination off dry ball aiir temperatu ure, mean raadiant tempeerature (MR RT), relative
hummidity, air sppeed, metabo olic rate, and
d clothing innsulation (KE
E, 2003) (Figgure 1).
Figure 1. Psychometri
P ic chart-theermal comfoort range off the PMV M Method.
Howeveer, a lot of building
b occcupants repoort their therrmal stress, discomfort,
andd dissatisfacttion during their time in a built ennvironment, in spite of tthe thermal
connditions mon nitored and regulated by y the currennt Fanger’s model. IFM MA reported
thatt too cold and
a too warrm condition ns were thee most criticcal issues afffecting the
occcupants’ ind door environnmental quaality, includding lightingg, spatial, pprivacy, air
quaality, etc. While
W many efforts havee been madee to overcom me the currrent control
appproaches thaat rely on conventional model-baased enviroonmental coontrols, the
therrmal comforrt issue has not
n been reso olved. This iis a critical llimitation inn the current
buiilding enviroonmental co ontrol strateg
gies, and it is essential that a hum man-building
inteegrative fram
mework be developed to t enhance human phyysiological bbenefits and
envvironmental sustainabilitty via optimiization of ennergy use.
© ASCE
Sustainable Human–Building Ecosystems 119
The outcome of this research, in the form of a computational model that uses
real time facial skin temperature data as an input variable, will be applicable to any
existing thermal control system in a building. This especially applies to any
individual control system that is equipped with a personal environmental module /
terminal reheating box, and is situated in an office building or a healthcare facility.
The occupants’ low mobility in these facilities can be helpful for collecting the facial
skin temperature data remotely without being intrusive.
METHODS
During the experiment, the temperature was controlled within a range of 20ºC
to 30ºC. The sequence began from cooling to heating, or vice versa, to prevent any
biased thermal sensation reporting from any test participant. Most human subjects
were either undergraduate or graduate students at the University of Southern
California. We selected 1.5ºC as a temperature change step based on the capacity of
the HVAC systems in the environmental chamber. Each test consisted of seven
change steps and, at the end of each step, the subject was asked to report his/her
© ASCE
Sustainable Human–Building Ecosystems 120
Figure
F 2. Exxperiment prrocedure
A 7-point scale is a very pop pular methood for surveeying user ssatisfaction;
ASHRAE-55 also a adopted this scale to o estimate th ermal comfoort and sensaation. Since
it is
i neither to oo complicaated nor too o simple, a 7-point scaale has beenn popularly
adoopted in the research do omain of ind door environnmental quaality and envvironmental
satiisfaction stu
udies. In ord
der to obtainn stable expeerimental coonditions, suubjects were
req
quested not tot have food d for at leastt 30 minutess before the experiment in order to
maintain consisstent metabo olic rates of the
t individuaal subjects.
Figure 3. An experimenta
e al setting in
n this
study Figure 4. Selected faacial
temperaturre sensing ppoints
RE
ESULTS
Individu
ually collectted skin tem
mperatures aat six facial points of each subject
werre categorizeed based on the thermal sensation annd conditionning mode, ii.e., cooling
andd heating. Figure 6 illlustrates thee generated temperaturees (includinng dry bulb
tem
mperature annd operative temperaturee), and the aaverage skinn temperaturre collected
from
m one samp pled subject during
d xperiment. IIt looks veryy stable in Fiigure 6, but
the ex
© ASCE
Sustainable Human–Building Ecosystems 121
Tab
ble 1. Therm
mal sensatio
on and comffort survey
1. What
W is your overall
o level of
o thermal comfort?
Very Unsatisfied Slightly Neural Slightly Satisfied Very
unssatisfied unsatisfied satisfied satisfied
□ □ □ □ □ □ □
2. What
W is your overall
o therm
mal sensation??
Very cool Slightly Neural Slightly warm Very
Cool cool warm warm
□ □ □ □ □ □ □
Fiigure 5. Tim
me series plo
ot of Tdry (d
dry bulb airr temperatu
ure) and To ((operative
tem
mperature) and T skin (average)
© ASCE
Sustainable Human–Building Ecosystems 122
Based on
o this findin ng, the averaage facial skkin temperatture was seleected as the
representative variable
v for a thermal inndex of a facce. The data from all of tthe subjects
werre assembled d and categoorized into seven
s differeent thermal ssensations, aas shown in
Fig
gure 9. The absolute
a valuues of the tem
mperatures w were totally different, deepending on
the sensing sp pots on a face,
f and allso on the test subjectts. Their phhysiological
therrmoregulatioon principlees were very y similar to each other in order to maintain a
heaat balance beetween the human
h body and the ambbient thermaal condition.. Therefore,
5 minutes
m wass selected ass a time wiindow framee for estimaating a graddient of the
aveerage facial skin temperrature per th hermal sensaation while the temperaature in the
chaamber was co ontinuously being chang ged.
Figure 7. Pattterns of skiin temperatures at six sselected faciial skin poin nts, as air
tempeerature (fro om one samp pled subjectt).
Figure 9 summariizes the diistribution oof the graddient of avverage skin
mperature perr thermal sen
tem nsation, baseed on the collected data from all of tthe subjects..
ANNOVA (Anaalysis of variance) tesst results sshowed signnificant diffferences in
graadients betw
ween any two o different thermal
t sennsations. To increase the statistical
signnificance, th
he cold sennsation data (i.e., -3, -22. -1) weree grouped aas one cool
sen
nsation, and the
t warm sen nsations (i.ee., +3, +2. +11) were also combined aas one warm
sen
nsation, while keeping the neutrall sensation as a singlee baseline nnorm. This
trannsformation of the data generated
g mo ore highly siignificant loow p-values < 0.00.
© ASCE
Sustainable Human–Building Ecosystems 123
Figure
F 8. Ch
hange rate ranges
r of the average sk kin temperaature of all subjects'
facial skin pointts.
DIS
SCUSSION
N
The findings in thiss research leed to the connclusion thaat a gradientt of average
skinn C is a critiical indicatorr for revealin
ng each subjject’s thermaal sensation. In addition,,
eveen though the absolute average
a skin temperaturee was the samme betweenn subjects, it
cou
uld be a goo od indicatorr to illustratee an “overaall” sensationn per person since the
ran
nge of skin temperatures
t s for each sensation w was not widee enough. O On the other
han
nd, an ambieent temperatture was also o recognizedd as a signfiicant parameeter since it
wass good enough for illustrrating the “o overall” therm
mal sensatioons of the subbjects.
© ASCE
Sustainable Human–Building Ecosystems 124
parrameters. By y adopting these param meters and consideringg the humann factor, a
deccision tree (u
using J48) was
w establish hed per the H HVAC systeem model (ii.e., cooling
and
d heating). Figures 9 and d 10 illustratee how the deecision trees was appliedd to AC and
heaater controlss. Depending g on the esttimated therrmal sensatiion, the conndenser and
heaater were “onn” or “off” to
o maintain a subject’s thhermal comffort conditionn, a thermal
neu
utral sensatioon. For thesse decision trees,
t a 10-ffold cross vvaldiation was adopted.
Thiis estimationn approach achieved
a 78% , 86%, annd 86% accuuracy in esttimating the
therrmal sensatiions: cool, neutral,
n and warm
w condittions, respecctively, as a function of
facial skin temmperature an nd the ambieent thermal condition, aas well as ggender as a
humman factor.
CO
ONCLUSION
This stu
udy focused on determin ning the poteential use off facial skin ttemperature
to estimate
e an individual’s
i thermal sennsation (espeecially in a wworkstation ssetting) and
to identify
i any
y significant responses of o facial skiin temperatuure to ambieent thermal
connditions. Forr purposes ofo testing thee relationshipp between ffacial skin teemperatures
andd thermal co onditions, an nd to identiffy potential effects, a seeries of hum man subject
expperiments weere conducteed in an env vironmental chamber loocated at thee University
of Southern
S Caalifornia. Ev
ven though absolute
a leveels of facial skin temperature were
totaally differen
nt, depending on the six x selected ssensing spotts and also the various
sub
bjects, the grradient of thee average faccial skin tem
mperature waas practicallyy consistent
for all test subbjects. In add dition, the average
a skinn temperaturre and the aambient dry
bullb temperatu ure did not reeadily reveaal the individdual thermall sensations, although it
wass possible too provide an n overall therrmal sensati on within ceertain rangess. Based on
these findingss, this stud dy adopted ambient ttemperature,, average facial skin
tem
mperature, an nd its gradieent to develoop a thermall sensation pprediction m model in the
formm of a decision tree (J4 48), with an n estimated pprediction aaccuracy of 881% ± 5%,
deppending on th he thermal sensation.
s
© ASCE
Sustainable Human–Building Ecosystems 125
REFERENCES
ASHRAE. (2013). ASHRAE Standard 55. American Society for Heating, Ventilating
and Air- …. American Society of Heating, Refrigerating, and Air-
Conditioning Engineers.
Huizenga, C., Abbaszadeh, S., Zagreus, L., & Arens, E. A. (2006). Air quality and
thermal comfort in office buildings: Results of a large indoor environmental
quality survey. Proceeding of Healthy Buildings 2006, 3.
KE, C. (2003). Fanger’s Thermal Comfort and Draught Models (IRC-RR-162).
Myhren, J. A., & Holmberg, S. (2008). Flow patterns and thermal comfort in a room
with panel, floor and wall heating. Energy and Buildings, 40(4), 524–536.
Thermal comfort : analysis and applications in environmental engineering / by P.O.
Fanger. - Version details - Trove. (1972).
Wyon, D. P. (1996). Individual microclimate control: required range, probable
benefits and current feasibility. In International Conference of Indoor Air
Quality and Climate (pp. 1067–1072).
© ASCE
Sustainable Human–Building Ecosystems 126
Abstract
BACKGROUND
Numerous case studies have been performed over the years with respect to
Indoor Environmental Quality (IEQ). These case studies have articulated the
standards that determine the qualitative aspects of IEQ. Moreover, they have had a
marked impact on directing various design guidelines such as LEED and the current
WELL Building Standard. This study has conducted an extensive review of case
study literature on office building environments in the U.S. and abroad. It provides a
© ASCE
Sustainable Human–Building Ecosystems 127
threshold for the reconsideration of how the indoor environmental conditions of air
quality, thermal quality, and lighting and acoustics have been evaluated and how user
satisfaction with respect to environmental quality has been differentiated. The goal of
this project has been to focus the research of the past decade. It has collated the
results of these case studies to develop an understanding of how the research is
typically conducted and to determine if the reported results contain information that
supports the development appropriate metrics for measuring commercial office
environment performance by quantifying IEQ influences on behavior.
METHOD
THERMAL COMFORT
© ASCE
Sustainable Human–Building Ecosystems 128
cost benefits have been evaluated. Thirty-eight (38) papers on the issue of thermal
comfort have been studied for the period from 2001 to 2014.
10
Chatzidiakou et al. (2014), -11%
5
Cui et al, (2013), 5.4% in Memory Typing
0
Increase and Decrease in Performance
VENTILATIVE COMFORT
© ASCE
Sustainable Human–Building Ecosystems 129
starting in the late 1990’s to the present, they range in information about productivity,
health, comfort, and creativity related to ventilation rates and air quality. The research
reported in the articles relates to the ASHRAE standards that determine ventilation air
rates. Additional articles have looked beyond the standards to see if more needs to be
done in order to provide better environmental quality. There are also standards such
as LEED and WELL that focus on indoor environmental qualities and have
ventilation standards. These different standards were compared to one another to help
to understand which one is the most effective when it comes to occupants overall
productivity and health. While each standard presents a different set of criteria for
ventilation quality, they all recognize that it is important and necessary for the
building users to have good quality air in order to be comfortable in their
environment. Unlike LEED and ASHRAE, the WELL building standard does go into
more detail about the individuals health and wellbeing, such as understanding
nourishment, physical, and mental aspects of the person, while the other standards
focus more on the buildings requirements, which is important to ventilation
requirements because it is more than just providing air but effects the overall indoor
environmental quality. The articles help show that the individual’s overall comfort is
related to a lot of factors that the building has to have / accomplish in order to meet
the basic needs of the worker. Through the research, it is clear that there should be at
least a ventilation rate of 21 to 40 cfm for optimal occupant health, along with access
to both natural and mechanical ventilation, with this the performance rate of the
workers will increase up to 8% (Fisk, 2000).
A significant number of articles address the importance of ventilation for
healthier building occupants, but not as many actually understand what that ideal
ventilation rate are. ASHRAE 62.1 – 2010 states that a building needs a ventilation
rate of 20 cfm/ per person or .06 cfm/ per square foot for in a mechanically ventilated
office space. For a naturally ventilated space, there must be a minimum size of
openings based on floor area to be ventilated. Most articles follow the ASHRAE
standards as a minimum requirement for ventilating the designated space, but they
compare having a mechanically ventilated versus a mix-mode or completely natural
ventilated space. The article “Mixed Mode Ventilation HVAC meets Mother
Nature,” focuses on occupant preference for operable windows in their office space to
control their own natural ventilation. Brager, Ring, and Powell (2000) study points
out that occupants with access to an operable window are comfortable over a larger
range of temperatures than people in completely ventilated spaces.
There are many factors that affect the comfort of an occupant in a building as
related to both ventilation and air quality in a space. But, few articles discuss the
ventilation rate solely in its relationship to productivity and satisfaction. Each article
(Figure 2) used a specific ventilation method (i.e., mechanical or naturally
ventilated). Some research reported the actual measured rates used. As shown, the
spaces with natural ventilation increased in satisfaction the most, with one article
almost reaching 100% satisfaction with their environment. The least satisfaction was
reached when the ventilation rate was only 17.7cfm, as discussed earlier, this is below
the standard ASHRAE which states a good rate is at least 21cfm. An article presented
by Helsinki University states that improved ventilation rates up to 17L/s per person
improves office task performance and a rate above 10L/s to 25L/s add a significant
© ASCE
Sustainable Human–Building Ecosystems 130
deccrease to sick
k building sy
yndrome. Ventilation noot only provvides new, cllean air, but
also
o reduces thhe amount off pollutants and airbornee viruses in a space. Thherefore, the
abo
ove rates aree an importaant factor in
n the overalll health and productivityy within an
offi
fice.
Figgure 2. Occuupant satisfacction as a facctor of Natuural and mechhanical
venttilation ratess
LU
UMINOUS COMFORT
C T
© ASCE
Sustainable Human–Building Ecosystems 131
part. It possesses both quantitative and qualitative aspects that must be considered in
design, and the role it plays in the workplace especially cannot be overstated. This
research attempts to analyze the benefits of the luminous environment on health,
productivity, and satisfaction. It begins by recognizing two different types of light
daylight and electric light. For each of these types of light, articles were found,
analyzed, and compared with existing standards (LEED and WELL) to find areas
where the standards may be lacking or find potential opportunities for further
research. Overall, 96 articles were taken into account, spanning from 1965 to 2015.
The standards set forth by LEED and WELL go beyond building codes and deal with
environmental responsibility and occupant health respectively. In general, WELL
tends to go beyond LEED in assuring that the built environment is conducive to the
well-being of occupants, with credits that deal with the circadian rhythm, glare
control, and lease depth.
Light has a profound effect on health, largely through its influence on the
circadian rhythm. Certain luminance values and color temperatures tend to increase
alertness, well-being, and vitality. According to a study by K. C. H. J. Smolders, de
Kort, & van den Berg (2013) Brighter light tends to mean higher vitality in
occupants, especially during morning hours (Figure 3). Leichtfried et al. (2015)
found that exposure to bright light around 5000lx from 740 to 810am can improve the
mood of occupants, however, it can also increase distraction thus lowering mental
performance. A study by Sahin & Figueiro (2013) suggests that a 48 minute, post-
lunch exposure to red light (40lx, 630nm) can increase alertness in workers. Figures 3
and 4 illustrate the various illuminance and color values throughout the day tested
within the articles.
6000 8000
Lighting Level (lux)
Color Temperature
5000 6000
4000 4000
(Kelvins)
3000 2000
2000 0
1000
0
© ASCE
Sustainable Human–Building Ecosystems 132
TH
HE IMPACT
T OF LIGH
HTING ON PRODUCT
P TIVITY
Lighting g can effecct productiv vity by makking visual tasks easieer or more
diffficult, by cau
using distracction, and byy manipulatinng the comffort levels off occupants.
Borrisuit, Linhaart, Scartezziini, & Munch h (2014) fouund that subjjects felt sleeepier earlier
in the
t afternoon under elecctric light raather than ddaylight. On a health noote, subjects
alsoo reported loower well-beeing under electric
e lightt. WELL staandards sugggest the use
of controllable
c g within offiice spaces annd the researrch of Juslenn, Wouters,
task lighting
& Tenner
T (20077) suggest th
hat such straategies can inncrease prodductivity by up to 4.5%.
It is
i thought that the sattisfaction of control oover their im mmediate environment
conntributes to this
t productiivity increasee. However,, some of thee research suuggests that
commfort and productivity might actuaally be at oodds. While the most ccomfortable
lighhting conditiions are bettween 401 anda 500lx, thhe optimum m illuminancee for visual
taskks was foun nd to be over 900lx. (Gou, Lau, & Ye, 20014) Mentallly fatigued
inddividuals exp posed to lighht of around 1000lx havee decreased sleepiness aand reaction
tim
me on simple tasks, whilee increasing self-control.. However, oon more com mplex tasks,
the high illumiinance had an adverse effect. (Karrin C. H. J.. Smolders & de Kort,
20114) This ressearch suggeests that a balance
b betwween comfoort and prodductivity be
useed, or periodds of intense illumination n to increasee productivitty with interrvals of low
illu
umination to serve as restt periods.
TH
HE IMPACT
T OF LIGH
HTING ON SATISFAC
S CTION
The liteerature indicaates the link between ligghting and saatisfaction ussed specific
ligh
hting strateg gies or technnologies. A post occupaancy evaluattion (POE) oof the New
Yorrk Times Building
B reveealed occup pants satisfaaction with tthe automatted shading
devvices (Figuree 5) and the automated dimming
d ligghts (Figure 6). Clear, Innkarojrit, &
Leee (2006) stud died occupan nt satisfactio
on with electtrochromic w windows andd found that
Subbjects preferrred variable transmittancce and used venetian bliinds less, butt used more
elecctric lighting
g than fixed transmittancce.
Fig
gure 5. Satisfaction with
h Automated Window Shhades in NY Times Buildding
© ASCE
Sustainable Human–Building Ecosystems 133
COUSTIC COMFORT
AC C
Tab
ble 2. Reverrberation tim
me requiremeents
T60 (sec), at 5000 Hz, 1000
Ro
oom type Applicattion Hz, aand 2000 Hzz
Meeting
g or banquet room < 0.88
Offfice buildin
ng Executiv
ve or privatee office < 0.66
Conferen
nce room < 0.66
Teleconfference room
m < 0.66
Open-plan office without sound < 0.88
g
masking
Open-plan office with sound maasking 0.8
© ASCE
Sustainable Human–Building Ecosystems 134
In 1972, the EPA presented the Noise Control Act, which defined noise
sources and how to control, maintain, and police noise levels in public areas. While
this doesn’t directly correlate to indoor office standards, it set up a way of thinking
about noise and noise control.
LEED v4 for Interior Design and Construction identifies points to acoustic
performance. Its intent is to provide workspaces that promote occupants’ well-being,
productivity, and communications through effective acoustic design. All occupied
spaces must meet the following design requirements, as applicable, for HVAC
background noise, sound isolation, reverberation time, and sound reinforcement and
masking.
Achieve maximum background noise levels from heating, ventilating, and air
conditioning (HVAC) systems per 2011 ASHRAE Handbook, HVAC Applications,
Chapter 48, Table 1; AHRI Standard 885-2008, Table 15; or a local equivalent.
The WELL Building Standard does not include acoustics. It addresses,
however, to include several standards dealing with acoustics, within the comfort tier.
Most of the research however, deals with partition height and room layouts, leaving
very little quantitative data. However, by using WELL as a base case, more articles
can be found by searching for the specific metrics that WELL utilizes. These can then
be used to evaluate WELL and LEED.
CONCLUSIONS
There has been significant work over the years on the development of case
studies for various aspects of indoor environmental quality (i.e. air, visual, thermal
and acoustic quality.). For the most part, such studies have been stove-piped in the
sense that they focus deeply on only one component. Excellent monitoring tools have
been developed and results have been reported in the literature.
OBSERVATIONS
This work is somewhat unique it its emphasis has been on the combined
effects of the different components of IEQ. It is also unique in its attempt to create
metrics focused on business performance. Given that energy tends are such a small
percentage of operating costs, and one that is effectively transparent to most owners
in commercial office environments, such metrics can become useful in re-focusing
questions about energy performance to improved value and business outcomes.
• Most of the research on occupant satisfaction performance and productivity
has focused on assessing one factor of comfort. Temperature, for example, is
seen as the primary indicator of occupant thermal comfort. But, occupants are
exposed to multiple factors simultaneously, such as air movement, humidity
© ASCE
Sustainable Human–Building Ecosystems 135
w of the con
Review ntemporary literature suuggests thatt there is a large gap
betw
ween IEQ comfort categ gories from which cohessive conclussion regardinng occupant
welll-being may
y be extrapollated from im
mperial physsiological stuudies.
NE
EXT STEPS
S
This liteerature searcch supports the observattion that a mmore cohesivve approach
reg
garding behaavioral influeences and thheir role in aarticulating m
more accuratte occupant
sch
hedules in simmulation toools is neededd. These queestions have been deriveed from this
view and are guiding a seecond phase of this IEQ project.
rev
• Can a means
m of asssessing hierrarchical vallue to the m multiple influences that
influencce comfort be b developeed by introdducing occuppants to thee impact of
these vaariables? Workshops,
W poolicy assessmment and caase study willl be guided
by curreent social sccience metho ods for deterrmining occcupant underrstanding of
the influ
uence upon comfort
c percception.
• Can a cross referenttial matrix ofo thermal syystems compponents be developed to
address up-fit alternnatives in rettrofit and rem
modeling proojects?
• A stand dardized forrmat can bee developedd for assesssing occupaant comfort
sufficienncy? If so, can
c it result in a change in the way ooccupant schhedules and
systemss parameters are formulaated in simullation tools?
© ASCE
Sustainable Human–Building Ecosystems 136
Fig
gure 8. Meaasured occup pant energy demand proofile Simulaated vs. actuual – issues
Pollicy and behaavioral undeerstanding off plug load im
mpacts.
AC
CKNOWLEDGEMENT
TS
This research was made possiible throughh funding bby the NSF I/UCRC –
Cennter for Susttainably Inteegrated Build
dings and Siites and our industry parrtners Wells
Farrgo Bank, Baank of Amerrica and Ingeersoll Rand C Corporation.
This reesearch was compiled by b Researchh Assistancce from the School of
Arcchitecture, College
C of Arts
A and Architecture;
A UNC Charrlotte. Ourr thanks to
Niccholas Bradfford, Kelsey Lane, Sheid da Hosseinzaadeh, and NNazanin Moddaresahmadi
for their consid
derable effortt.
BIB
BLIOGRAP
PHIC COM
MPENDIUM
M OF LITER
RATURE R
REVIEW
© ASCE
Sustainable Human–Building Ecosystems 137
2007; Olli Seppänen, William J Fisk, David Faulkner, 2003; Olli A Seppänen,
William J Fisk, 2006; Pawel Wargocki, David P. Wyon ,2013; Pawel Wargocki,
David P. Wyon,2007; R. Kosonen, F. Tan,2004; SINGH, A., SYAL, M.,
KORKMAZ, S. and GRADY, S., 2011; Tanabe, Haned & Nishipana, 2007; Tanabe
S., Kobayashi, K., Kiyota, O., Nishihara, N.,and Haneda, M. 2009; T. Kershaw, D.
Lash, 2013; Valeria De Giuli, Roberto Zecchin, Livio Corain, Luigi Salmaso, 2014;
Weilin Cui, Guoguang Cao, Jung Ho Park, Qin Ouyang, Yingxin Zhu, 2013; W. J.
Fisk, D. Black, G. Brunner,2011; Zs. Bakó-Biró, D.J. Clements-Croome, N.
Kochhara, H.B. Awbia, M.J. Williams, 2012.
© ASCE
Sustainable Human–Building Ecosystems 138
Energy. Environ. Annual Review of Energy and the Environment 25.1 (2000) 537-66.
Web. Fritsch, R., A. Kohler, M. Nygard-Ferguson, and J.L. Scartezzini. 1990. “A
stochastic model of user behaviour regarding ventilation.” Building and Environment
25 (2) 173-181. Heiselberg, P.K. (2002) Principles of Hybrid Ventilation, Hybrid
Ventilation Centre, Aalborg University, Aalborg. Heinzerling, D., Webster, Tom,
Schiavon, Stefano, Anwar, George, and Dickerhoff, Darryl (2013). "A Prototype
Toolkit for Evaluating Indoor Environmental Quality in Commercial Buildings."
HVAC System 29. Henze, G.P., Felsmann, C., Kalz, D.E. and Herkel, S. (2008)
Primary energy and comfort performance of ventilation assisted thermo-active
building systems in continental cli- mates. Energy and Buildings, 40(2), 99–111.
Honnekeri, A., Pigman Margaret C, Zhang, Hui, Arens, Edward, Fountain, Marc,
Zahi, Yongchao, and Dutton, Spencer (2014). "Use of Adaptive and Thermal
Comfort in a Naturally Ventilated Office." Indoor Environmental Quality 9. Hui
Zhang, D. K., Edward Arens, Elena Buchberger, Fred Bauman, and a. C. Huizenga
(2008). "COMFORT, PERCEIVED AIR QUALITY, AND WORK
PERFORMANCE IN ALOW-POWER TASK-AMBIENT CONDITIONING
SYSTEM.” "Increased Ventilation." LEED User. N.p., n.d. Web.
<http//www.leeduser.com/credit/nc-2009/IEQc2. Indoor Air Quality Scientific
Findings Resource Bank.” Indoor Air Quality Scientific Findings Resource Bank
Health and Economic Impacts of Building Ventilation Ventilation Rates and Office
Work Performance. N.p., n.d. Web. 07 May 2015.
<http//www.iaqscience.lbl.gov/vent-office.html>. Jonsson, A. (2013). "THE VALUE
OF VENTILATION FROM THE WEBER-FECHNER LAW." Productivity and
Economics 8. Kats, G., L. Alevantis, A. Berman, E. Mills, and J. Perlman. 2003. The
Costs and Financial Benefits of Green Buildings. Report to California’s Sustainable
Building Task Force, pp.54-55. Ncube M, Riffat S. Developing an indoor
environment quality tool for assessment of mechanically ventilated office buildings in
the UK – A preliminary study. Building and Environment 2012;5326–33. Olli
Seppänen, W. J. F., QH Lei (2006). "VENTILATION AND PERFORMANCE IN
OFFICE WORK." Indoor Air Journal 18 15. Pasut, W. a. D. C., Michele (2012).
"Evaluation of Various CFD Modelling Strategies in Predicting Airflow and
Temperature in a Naturally Ventilated Double Skin Façade." Applied Thermal
Engineering 37 27. Pawel Wargocki, D. P. W., and P Ole Fanger (2000).
"Productivity is Affected by the Air Quality in Offices." International Centre for
Indoor Environment and Energy 6. Schiavon, S., Melijove, Arsen, and Chandra,
Sekhar (2010). "Energy Savings Strategies with Personalized Ventilation in Tropics."
Energy and Buildings 42(5) 16. Schiavon, s., Bauman, Fred, Tully Brad, and
Rimmer, Julian (2011). "Air Change Effectivness in Laboratory Test of Combined
Chilled Ceiling and Displacement Ventilation." Indoor Air 7. Seppänen OA, Fisk WJ,
Mendell MJ. (1999) Association of ventilation rates and CO2- concentrations with
health and other responses in commercial and institutional buildings. Indoor Air J, 9,
252-274. Seppänen, O. and W. Fisk. 2001. “Association of ventilation system type
with sick building symptoms in office workers”, Proceedings Indoor Air, pp. 98-112.
Seppanen, O. (2008). "Scientific basis for design of ventilation for health,
productivity and good energy efficiency." Indoor Air Congress 8. Spindler, H. and
Norford, L. (2009a) Naturally ventilated and mixed-mode buildings – Part I Thermal
© ASCE
Sustainable Human–Building Ecosystems 139
modeling. Building and Environment, 44(4), 736–749. Vivian Loftness, F., Volker
Hartkopf, Ph.D.2, Beran Gurtekin, Ph.D. (2002). "Linking Energy to Health and
Productivity in the Built Environment." Advanced Building Systems Integration
Consortium 12. Wargocki P, Seppanen O, Anderson J, Boerstra A, Clements-
Croome DJ, et al. Indoor climate and productivity in offices guide book 6. Brussels
Federation of European Heating and Air-Conditioning Associations (REHVA); 2006.
WJ Fisk, P. P., D Faulkner, D Sullivan, D Dibartolomeo, C Federspiel, G Liu, and M
Lahiff (2002). "Productivity and Ventilation Rate Analyses of Timeseries data for a
Group of Call-Canter Workers." 6. Li Y, Leung GM, Tang JW, Yang X, Chao CYH,
Lin JZ, et al. Role of ventilation in airborne transmission of infectious agents in the
built environment—a multidisciplinary system- atic review. Indoor Air. 2007;17(1)2–
18.
Luminous Comfort. Aries, 2005; Badia, Myers, Boecker, Culpepper, & Harsh,
1991; Baron, Rea, & Daniels, 1992; S. Begemann, G. Van den Beld, & A. Tenner,
1997; S. H. A. Begemann, G. J. van den Beld, & A. D. Tenner, 1997; Berman,
Jewett, Fein, Saika, & Ashford, 1990; Borisuit, Linhart, Scartezzini, & Munch, 2014;
P. Boyce, Hunter, & Howlett, 2003; P. R. Boyce, 2010; P. R. Boyce, Beckstead,
Eklund, Strobel, & Rea, 1997; P. R. Boyce & Cuttle, 1990; Peter R Boyce et al.,
2006; Cajochen, 2007; Cajochen et al., 2005; Cajochen, Zeitzer, Czeisler, & Dijk,
2000; Chellappa et al., 2011; Clear, Inkarojrit, & Lee, 2006; Daurat et al., 1993; De
Carli, De Giuli, & Zecchin, 2008; Deguchi & Sato, 1992; Dijk & Archer, 2009;
Dikel, Burns, Veitch, Mancini, & Newsham, 2013; "<eScholarship UC item
3km3d2sn.pdf>," ; Escuyer & Fontoynont, 2001; Farley & Veitch, 2001; Ferlazzo et
al., 2014; M. G. Figureueiro, Bierman, Plitnick, & Rea, 2009; Mariana G Figureueiro,
Bullough, Bierman, Fay, & Rea, 2007; Fotios & Cheal, 2010; Golden et al., 2005;
Gou, Lau, & Ye, 2014; Heerwagen & Heerwagen, 1986; Hoffmann et al., 2008;
Hubalek, Brink, & Schierz, 2010; Hwang & Jeong Tai, 2010; Ju, Chen, & Lin, 2012;
Juslen, Wouters, & Tenner, 2007; K. Kaida et al., 2006; K. Kaida, Takahashi, &
Otsuka, 2007; Kosuke Kaida, Takeda, & Tsuzuki, 2012; Kantermann et al., 2012;
Kasper, Rogers, Madden, Joseph-Vanderpool, & Rosenthal, 1990; Katzev, 1992;
Lehrl et al., 2007; Leichtfried et al., 2010; Leichtfried et al., 2015; Lewy, Wehr,
Goodwin, Newsome, & Markey, 1980; Linhart & Scartezzini, 2011; Liu, Chiang, &
Lin, 2010; Logadóttir, Christoffersen, & Fotios, 2011; Martinez-Nicolas, Ortiz-
Tudela, Madrid, & Rol, 2011; McLntyre, Norman, Burrows, & Armstrong, 1989;
Menzies & Wherrett, 2005; Mills, Tomkins, & Schlangen, 2007; Moore, Carter, &
Slater, 2003; Munch, Linhart, Borisuit, Jaeggi, & Scartezzini, 2012; Myers & Badia,
1993; Newsham et al., 2009; "Office Worker Response to Fluorescent Lamps of
Different CCT and Lumen Output," ; Partonen & Lonnqvist, 2000; Revell, Arendt,
Fogg, & Skene, 2006; Roche, Dewey, & Littlefair, 2000; Rodriquez & Pattini, 2012;
Rosenthal, Rotter, Jacobsen, & Skwerer, 1987; Rüger, Gordijn, Beersma, de Vries, &
Daan, 2006; Sahin & Figureueiro, 2013; Sahin, Wood, Plitnick, & Figureueiro, 2014;
Scheer, van Doornen, & Buijs, 1999; Sivaji, Shopian, Nor, Chuan, & Bahri, 2013; K.
C. Smolders, de Kort, & Cluitmans, 2012; Karin C. H. J. Smolders & de Kort, 2014;
K. C. H. J. Smolders, de Kort, & van den Berg, 2013; Stephenson, Schroder,
Bertschy, & Bourgin, 2012; Tabuchi, Matsushima, & Nakamura, 1995; Trinder,
© ASCE
Sustainable Human–Building Ecosystems 140
Armstrong, O'Brien, Luke, & Martin, 1996; Tuaycharoen & Tregenza, 2007; van
Bommel, 2006; Vandewalle et al., 2006; Vandewalle, Gais, et al., 2007; Vandewalle,
Maquet, & Dijk, 2009; Vandewalle, Schmidt, et al., 2007; JA Veitch, Geerts, Charles,
Newsham, & Marquardt, 2005; JA Veitch & Newsham, 2000; J. Veitch, Newsham,
Boyce, & Jones, 2008; Jennifer A. Veitch, 2001; Jennifer A Veitch, 2001; J. A.
Veitch & Gifford, 1996; Jennifer A Veitch, Hine, & Gifford, 1993; J. A. Veitch &
McColl, 2001; J. A. Veitch, Stokkermans, & Newsham, 2011; Viola, James,
Schlangen, & Dijk, 2008; Wang & Boubekri, 2010; Webb, 2006; Wei et al., 2014;
Wells, 1965; Yildirim, Akalin-Baskaya, & Celebi, 2007; Zeitzer, Ruby, Fisicaro, &
Heller, 2011.
© ASCE
Sustainable Human–Building Ecosystems 141
Abstract
INTRODUCTION
Thermal comfort has become a new talking point when people consider energy
efficiency in the residential/commercial buildings. The most complicated and
dynamic part in energy modelling is thermal comfort due to dynamic nature of human
actions and unpredictable nature of human behavior. But, it seems that modelling
human thermal comfort may be possible by different techniques as demonstrated by
Cao et al. (Cao, 2011) by comparing Thermal sensation vote (TSV) to Predicted
mean vote (PMV) and using ASHRAE index of 7 levels to categorize thermal
comfort. Thermal comfort can be studied as function of air temperature (Cui, 2013)
and dynamic modelling of thermal comfort using IMEM model is demonstrated by
Katavoutas et al. (Katavoutas, 2015). Various other modelling approaches have been
studied and described in literature survey below. Here, the intention is to find
© ASCE
Sustainable Human–Building Ecosystems 142
LITERATURE REVIEW
© ASCE
Sustainable Human–Building Ecosystems 143
(Cui, 2013) and motivation for high performance increases with high thermal comfort
(Cui, 2013). Dynamic modelling of thermal comfort has been studied by Katavoutas
et al. They used two node IMEM model for dynamic modelling of thermal transients.
Temporal pattern of body temperatures, skin wittedness and water loss were the
parameters investigated (Katavoutas, 2015). The authors concluded that thermal
adaptation and response to the same neutral thermal environment is different for
individuals based on their activities outdoors (Katavoutas, 2015). The skin
temperature was lower when individual sits under shade (Scenario A) compared to
temperature when individual is under direct solar radiation and sitting (Scenario B)
and also skin temperature was lower than (Scenario C) where individual was walking
(Katavoutas, 2015).
Kolarik et al. simulated thermo active building system (TABS) and compared the
results with performance of all-air VAV ventilation system. The TABS systems are
essentially the hydronic heating/cooling systems also known as radiant
heating/cooling systems. The building modules were considered with heavy and light
construction with two different orientations (east-west and north-south) (Kolarik,
2011). The authors found that TABS saved 50% more primary energy than VAV
system for hot-humid climate, thermal sensation expectations did not change more
than 1% for 60%-70% 0f working hours for TABS system (Kolarik, 2011). Change
of building orientation did not have significant impact on primary energy use while
90% of occupants were thermally satisfied for using TABS system in moderate and
hot-dry climate (Kolarik, 2011). Li et al. performed an experiment in a naturally
ventilated room in Chongqing, China. 300 students were part of the experiment and
the authors found that sensory nerve conduction velocity (SCV) was fine measure of
human response assessment for thermal comfort (Li, 2010). SCV can produce some
adaptive changes in human body to deal with indoor temperature changes due to
adaptive feedback regulation (Li, 2010). Luo et al. conducted field studies in Beijing
and Shanghai, China to find out influence of personal control on thermal comfort.
139 apartments were used for experiments. Occupants with personal control had
neutral temperature lower than occupants with no personal control by 2.6 ºC (Luo,
2014). A review of field studies has been presented by Mishra et al. They conclude
that conditioned areas have restricted thermal comfort zones and that similar thermal
environments are perceived differently by occupants based on their activities
outdoors, ease and effectiveness of the systems (Mishra, 2013). Cultural habits like
sitting on cool floors and opening windows which help thermal comfort in naturally
ventilated buildings are limited to cultures and regions associated (Mishra, 2013).
Moustris et al. developed and applied Artificial Neural Network (ANN). The model
were based on bioclimatic data calculated from Discomfort Index and Cooling Power
Index and microclimatic data like air temperature, humidity etc. The data was
considered for 2001-2005 (Moustris, 2010). The ANN models can successfully
predict thresholds for human discomfort (Moustris, 2010). Conditions of heatstroke in
humans and days with high thermal discomfort can be predicted through ANN
models (Moustris, 2010). Authors suggest to focus on developing operational
forecasting system for urban areas to predict severe human discomfort (Moustris,
2010). Ng et al. studied urban thermal comfort in Hong Kong China. The authors
© ASCE
Sustainable Human–Building Ecosystems 144
found that wind speed of 0.9-1.3 m/s and physiological equivalent temperature (PET)
of 28 ºC is enough for Hong Kong public to experience thermal comfort (Ng, 2012).
Authors suggest proper urban planning, building coverage and proper layout (Ng,
2012). Authors also suggest limiting solar radiations in the streets by providing
canopies, covered streets, walkaways, recesses, etc. (Ng, 2012).
© ASCE
Sustainable Human–Building Ecosystems 145
The present study were performed on five participants and all of them were
students. All participants were male and the experiment was conducted in summer.
Total 25 experiments were performed in five sets. Participants wore short sleeve t-
shirt and light jean pants which has clothing insulation value of 0.34 Clo together. All
the experiments were performed in Zero Energy Lab (ZOE) at University of North
Texas on roughly consecutive days around 2:30 PM over the June 2015. ZOE is a
unique residential test lab thermally controlled by radiant floor system and water-to-
water heat pump (WWHP) systems with forced air. Room conditions during
experiments were checked dynamically through building management system
TACVista. Ambient conditions were maintained at 25.5 °C and the relative humidity
was variable between 73%-77% which effectively changed the dew point temperature
between 21 °C – 22 °C. Fig.1 shows the Flir ThermaCam B2 infra-red camera used
for measurements of palm and wrist temperatures during experiments. It is possible to
set emissivity and reflective temperature of human skin in the ThermaCam. Human skin
emissivity is 0.98 and reflective temperature is 33.33 °C due to the closest temperature to
the average participant’s hand temperature. Reflective temperature setting is provided in
the thermal camera to compensate for the temperature reflected from the objects
surrounding the subject of the experiment.
© ASCE
Sustainable Human–Building Ecosystems 146
Figu
ure 1. Flir Therm
maCAM B2 Fig
gure 2. Skin temp
mperature measuurement Figgure 3. Position of the subjects
RE
ESULTS AN
ND DISCUS
SSION
Results aree presented for correlatiion betweenn human therrmal comforrt sensation
and
d palm and wrist
w skin tem
mperatures for
f 5 particippants in Figuure 4 to Figurre 8.
© ASCE
Sustainable Human–Building Ecosystems 147
© ASCE
Sustainable Human–Building Ecosystems 148
The x-axis denotes comfort level while y-axis shows temperature of the body
part in °C. Because the experiment was done in summer season, ASHRAE index of 0
to +3 for comfort level, is plotted in the graphs by neglecting the -3 index as it is
relevant during cold climate.
The experiment set out to evaluate the possible correlation between human
thermal comfort feeling and selected body parts temperature and also if wrist is a
better spot than palm for capturing the skin temperature. The diagrams illustrate that
the wrist temperature is always less than palm temperature. It also can be seen that
they are convergent together when participants tend to feel warm. Also in the same
feeling levels, the temperature difference between the coldest test and warmest test is
not big on the both spots. Different experiment sessions give similar results and they
could represent precise correlation between comfort feeling level and the two spots
surface temperature on hand. In addition, wider temperature difference on the wrist
which can also be explained by grater slope trend line, makes the wrist relatively
better spot for measuring the whole hand’s part skin surface temperature. Error
analyzing for the temperature variation at 0 level comfort sensation at the biggest
case was 0.7 °C for the palm and 0.8 °C for the wrist. Regarding to the necessity of
repeating this experiment, for future research it is planned to perform the same
experiment in winter season to check the consistency in body physiological behavior
with these results. Although the ambiguity of age, gender and race influence on these
results still exist in the supposed hypothesis, but for next series of experiments,
number of participants will increase and different age groups and genders would be
considered for analysis.
CONCLUSION
Wrist and palm were the potential candidates for evaluating the possible
correlation between human comfort perception and body physiological behavior, due
to their feasibility in making simple apparatus to measure that and also their
sensitivity and accountability to ambient condition changes. It is assumed that in each
thermal comfort level which reported by the participants, each person has his own
individual skin temperature. By over 25 experiment series, the results are in
acceptable concordance with the elementary hypothesis and they show that alongside
rise and falls in skin temperature on different comfort sensation, each person has his
own unique skin temperature on each comfort level scale. Diversely, from this
illustration it can be branched out that by learning each person’s unique wrist skin
index temperature at his neutral or best comfort level, the room condition can be
provide cooling or heating load based on mean time wrist skin temperature deviation
from the comfort index temperature. So there would not be excessive energy injection
to the indoor environment anymore. Also for more than one person in this system, the
average of their comfort index temperature can be login to the system as the set point
temperature for average occupants’ skin temperature and then the central unit HVAC
system provide required load based on the real time difference between these two
values.
© ASCE
Sustainable Human–Building Ecosystems 149
REFERENCES
Cao, B. (2011). “Field study of human thermal comfort and thermal adaptability
during the summer and winter in Beijing” Energy and Buildings, 43,1051-1056.
Kolarik, J. (2011). “Simulation of energy use , human thermal comfort and office
work performance in buildings with moderately drifting operative temperatures”
Energy and Buildings, 43, 2988-2997.
Luo, M. (2014). “Can personal control influence human thermal comfort? A field
study in residential buildings in China in winter” Energy and Buildings, 72, 411-418.
Ng, E. (2012). “Urban human thermal comfort in hot and humid Hong Kong” Energy
and Buildings, 55, 51-65.
© ASCE
Sustainable Human–Building Ecosystems 150
Abstract
This paper presents a novel modeling methodology that integrates the near
building environmental conditions (or microclimate), whole-building design, and
occupant behavior. Accurate predictions of the future building operating conditions
lead to designs that serve the building’s purpose – to support occupants’ tasks. This
study bridges the gap between human factors and architecture to include physical,
cognitive, and organizational systems into building information modeling using
future typical meteorological year climate data, canyon air temperature microclimate
model, and a whole-building energy simulation to investigate the impact of future
microclimate conditions on a “typical” single-occupant office. Additionally, to
capture the effects of building occupant decision-making and adaptive behaviors, an
agent-based model is proposed. Model inputs are task-based, which aim to produce a
more robust model to investigate a variety of human-building control interactions to
ensure high building performance and occupant comfort and satisfaction.
INTRODUCTION
Building operating conditions are the product of the dynamic and complex
interactions between the climate, building, and occupants. Modeling these
interactions can significantly improve building performance and occupant comfort.
However, failing to properly represent these interactions leads to buildings that do not
operate as intended and do not satisfy occupants (Andrews, Yi, Krogmann, Senick, &
Wener, 2014). Hence, it is important to accurately account for the potential factors
influencing building operating conditions so that buildings can be better designed to
serve the purpose and needs of building occupants. Efforts to improve accuracy of
commercial building performance predictions has led to increased attention on
modeling building occupant behavior and microclimate – two main drivers of
building energy use (Bonte, Thellier, & Lartigue, 2014; Bouyer, Inard, & Musy,
2011).
Building modeling overview. The commercial building design phase frequently uses
building simulations to examine the interactions between climate, building, and
occupants. It is common practice to rely on stand-alone building configurations in
© ASCE
Sustainable Human–Building Ecosystems 151
building simulations. However, it is well known that the urban microclimate strongly
affects building operation (Erell & Williamson, 2007; Santamouris et al., 2001). The
urban microclimate is the local atmospheric zone modified by energy exchanges with
local built environment context. Buildings in urban context experience higher
ambient temperatures due to different wind flow patterns caused by building
placement and form, use of materials that effectively store short-wave radiation, and
increased solar radiation because of high-albedo materials on neighboring surfaces.
Microclimate models simulate these interactions to adapt climate weather data from a
stand-alone meteorological station to realistic site-specific air temperatures.
The building interior includes building systems, furniture, equipment, and
occupants. These are typically represented in building simulation as “oversimplified,
predetermined inputs” that are unrepresentative of actual building systems and
occupancy (Lee & Malkawi, 2014). Inputs are reduced to averages that are then
applied to the whole building, an approach that does not represent the dynamic
environment and behaviors observed in a real workspace. Equipment use, light use,
occupant presence, and occupancy count, to name a few, are reduced to fixed
schedules and represented as fixed metabolic heat generators based on historical data
(Mahdavi, 2001), or as a discrete “stimulus-behavior” relationship (Reinhart, 2004).
Further, furniture that affects airflow and radiant heat exchanges are treated as a fixed
mass that is evenly distributed throughout the spaces.
Future conditions are challenging to predict. It is even more challenging to
predict operating conditions for a building that do not yet exist – for example, a
building designed in a new development where no other structures yet exist. On the
other hand, advancements toward more accurate predictions has led to research
identifying probable operating circumstances (Burian, McPherson, Brown, Streit, &
Turin, 2003). For instance, urban planning describes the context in which individual
buildings are constructed, such as building form, building height-to-width ratios, and
street orientation (Elmualim, Valle, & Kwawu, 2012). Climate-based design
strategies can relate to envelope design, façade materials, and climate mitigation
strategies, such as reflective glazed facades and high albedo materials, to radiate,
conduct, and convect the climate to produce desired result of building performance
(EPA, 2013; Schuetter, DeBaillie, & Ahl, 2014). Further, tasks occupants perform in
a building inform various aspects of building interiors: the furniture, equipment,
interior environment conditions required to carry out day-to-day business operations,
and building systems needed to achieve those operating conditions (Leach, Lobato,
Hirsch, Pless, & Torcellini, 2010). Additionally, tasks can provide expected occupant
activity and clothing levels.
To account for uncertainties in occupant behavior, agent-based modeling
(ABM) has gained increasing attention in building simulation (Lee & Malkawi,
2014). ABM is a computer simulation of agents that interact with their environment
through defined rules, and have the ability to learn and adapt in response to changes
in the environment and other agents. Existing ABMs have utilized either an occupant
satisfaction driver of behavior (Andrews et al., 2014) or an environmental affordance
driver (Lee & Malkawi, 2014). While both have the capability to address variability
in human behavior, these models lack generalizable results because there is no
© ASCE
Sustainable Human–Building Ecosystems 152
relaationship beetween thesee behaviors and the truue driver off occupant bbehavior in
commmercial buiildings – thee tasks occup pants perform m.
The obj
bjective of th
his paper is to
t present ann integrated modeling m methodology
thatt takes a task-based approach
a to define parrameters forr building design and
occcupant behaavior in a commercial
c building wwhile accounnting for m microclimate
inflluences on occupant comfort
c connditions andd energy uuse in builddings. This
appproach bridges the gap between
b humman factors aand architectture to includde physical,
coggnitive, and organizationnal systems into
i buildingg informatioon modeling. By basing
model inputs on o tasks rath her than occcupants, the goal is to pproduce a m more robust
model that ev valuates futuure building operatingg conditionss in terms of energy
connsumption anda occupannt comfort. This paperr, first, pressents a dataa collection
metthodology, based in humanh facto
ors methodss, to collect and organnize future
buiilding operaating condittions. Colleccted data w will then bee used to inform the
miccroclimate, building,
b andd occupant components
c iin building ssimulation.
DA
ATA COLLE
ECTION METHODOL
M LOGY
Dettailed informmation is reequired abo out microcli mate, buildding, and occcupants to
acccurately reprresent futuree operating conditions.
c F
Figure 1 outtlines a hum
man factors-
bassed methodology to captu ure data for a Midwest ccommercial ooffice buildiing.
Figuree 1. Research
h methodology used to develop a m
model of thee office
domain.
d
First, an
a ethnographic study y of the office dom main providdes a full
undderstanding of the occup pants and their tasks by documentinng people’s behavior in
the context off their own space. Seco ond, the chaaracterizatioon of buildinng systems
idenntifies and represents
r bu
uilding systeems in a way
ay that highliights both fuunction and
cap
pability. Thirrd, analysis of human behavior
b reppresents the link betweeen occupant
beh
havior, build ding perform
mance, and occupant
o peerformance. The results from each
metthod will be b analyzed d separatelyy, and integgrated into abstractionn hierarchy
dessigned to reflect the coonstraints (i.e. social, phhysical) of the work environment
usinng an Ecolo ogical Interface Design n (EID) appproach, as ooutlined by Burns and
Hajjdukiewicz (2004).
(
Ethhnographic study of the office dom main. Severaal techniques will be useed to gather
task
k-based info
ormation requ
uired to defiine a typical office buildding:
• Review w of comfort standard ds will identify the bbaseline connditions for
occupan nt comfort in an office environm ment. Standaards includee ASHRAE
(2013) Standard 555: thermal en nvironmentaal conditionss for humann occupancy
© ASCE
Sustainable Human–Building Ecosystems 153
Model of the office domain. To construct the model of the office domain, four steps
were followed. First, the system boundary and scale were defined to include anything
users might want to control or have information about, and exclude anything
redesigned in future work, such as system controls and sensors. Second, an
abstraction hierarchy was created using a part-whole decomposition to develop the
© ASCE
Sustainable Human–Building Ecosystems 154
oveerall work do
omain structture. Lastly, the abstracttion hierarchhy model wiill be tested
for completeness using scennarios in whhich the occuupants perforrm various ooffice tasks.
Tassk definitioon. Preliminary results from the coontextual tassk analysis reveal four
inittial common n tasks for examinatio on: 1) electtronic docum ment preparration (e.g.
wriiting or typin
ng a documeent on a commputer), 2) noon-electronicc document preparation
(e.g
g. reading or writing ussing pen and d paper), 3) phone call,, and 4) sm mall meeting
(e.g
g. 2-4 occupants gathered in the samme vicinity foor discussionn). These tassks were the
most frequently performed d, critical to
o the occuppant’s job deescription, aand are the
lon
ngest in durration when being perfformed. Furrther, these four tasks all require
diffferent enviro
onmental coonditions. Using
U these ffour tasks, bbuilding moodel aspects
andd occupant behavior are determined.
BU
UILDING SIIMULATIO
ON METHO
ODOLOGY
Y
Figure 2. Simulation
n process.
Building. The Departmentt of Energy’s new consttruction com mmercial meddium office
ogy (Deru et al., 2011) seerves as the larger buildiing structuree in which a
buiilding typolo
© ASCE
Sustainable Human–Building Ecosystems 155
sing
gle-occupannt office is deefined. A sin
ngle-occupannt office is ccentrally loccated east to
wesst and alongg the north wall
w on the second
s levell because it w
was identified as being
the most affectted by the suurrounding microclimate
m e, based on results from
m a previous
stud
dy (Kalvelag ge, Passe, & Dorneich, 2015).
2 The m model of thee office dommain defines
the single-occu
upant office ini terms of size,
s shape, layout, furniture, and eqquipment in
ord
der to perforrm the four tasks, as shown in Figgure 3. Addditionally, thhe building
systems for thee single-occuupant office include
i overrhead suppliied heating aand cooling,
a th
hermostat wiith a limited
d range of co ontrol, a mannual window w roll shade, and a door
lock
kset.
Figuree 3. Single-o
occupant offfice schemattic.
Occupant. Usiing NetLogo o, an agent-bbased modell (ABM) off the office ddomain will
be developed. Occupants will be reepresented bby autonom mous agentss that have
inteelligence (i.ee., internal lo
ogic), as welll as the abillity to make complex deecisions and
enggage in co omplex inteeractions wiith other aagents and objects w within their
envvironment. The office domain ABM A will simulate occcupant behhavior and
inteeractions thaat occur betw ween buildin ng occupantss and their ooffice environnment as a
fun
nction of tasks performeed (e.g., the agent is perrforming a hhigh activityy level task
andd adjusts its clothing lev vel according gly). The AABM will alsso account ffor dynamic
occcupant adapttations, enab bling agents tot “learn” ovver time andd respond prooactively to
their environmeent (e.g., thee agent loweers the windoow shade beefore leavingg at the end
of the
t day to prevent early y morning su unlight fromm overheatingg the office tomorrow).
Furrther, occupaant actions do not neceessarily havee a guaranteeed effect onn the office
envvironment, an nd the ABM M will allow this
t uncertaiinty to be caaptured.
ABM accommodat
a es the repreesentation off heterogeneeous occupaants via the
devvelopment of multiple different
d ageent types, orr personas. E Each occupaant persona
willl be characterized by a set of uniqu ue, innate peersonal attribbutes and dyynamic state
varriables derivved from th he ethnograp phic study aand the hum man behavioor analysis.
Atttributes incllude objectives, preferrences, expectations, ssatisfaction thresholds,
assigned tasks, prior knowlledge, attitud des, motivatiions, and exxpertise. Statte variables,
which represen nt an agent’’s state at anya point inn time throuughout the simulation,
desscribe an ageent’s currentt task behaviiors (e.g., thee agent is onn the phone)), as well as
outtcomes that depend
d uponn personal attributes
a andd the state oof its environnment (e.g.,
the agent is com mfortable, giiven that the office is at a certain tem mperature).
© ASCE
Sustainable Human–Building Ecosystems 156
The attributes and current state of each individual agent will impact its ability
to perform required tasks, and may affect other occupant environmental
characteristics within the building, thereby influencing other agents’ states and
behaviors. This dynamic and adaptive process of interactions between the building
and its occupants will lead to system-wide performance over time, which will be
measured in terms of energy consumption and occupant productivity and comfort.
The current methodology to evaluate energy efficient building operations (with preset
schedules, set points, and occupation patterns) does not integrate occupancy behavior
nor specific climate features of the near building environment. The prediction, thus,
lacks depth and precision, buildings do not perform as predicted, and occupants
operate buildings differently than anticipated by the designing architecture team
(Andrews et al., 2014). With dominant societal goals to develop high performance
and energy efficient buildings, it has become necessary to develop more refined
models for building performance prediction. This paper presented a new
methodology that incorporates future building operating conditions, as well as
predictions of occupant behavior through an agent-based model, into the modeling of
building-microclimate interactions. The purpose is to improve energy and comfort
design predictions for buildings in urban context and in a changing climate. The
research goal of this project is to provide the inadequate modeling capacity that
enables more refined predictions based on occupant interactions with the building and
occupant task comfort levels. The resulting methodology can be leveraged to further
define model inputs that are more robust to long-term changes and more
generalizable across different building types.
© ASCE
Sustainable Human–Building Ecosystems 157
REFERENCES
Andrews, C., Yi, D., Krogmann, U., Senick, J., & Wener, R. (2014). Designing
buildings for real occupants: An agent-based approach. IEEE Transactions on
Systems, Man, and Cybernetics - Part A: Systems and Humans, tba. doi:
10.1109/TSMCA.2011.2116116
ASHRAE. (2013). ASHRAE Standard 55: Thermal environmental conditions for
human occupancy. In R. a. A.-C. E. American Society of Heating, Inc. (Ed.):
ASHRAE.
Bonte, M., Thellier, F., & Lartigue, B. (2014). Impact of occupant's actions on energy
building performance and thermal sensation. Energy and Buildings, 76(0),
219-227. doi: http://dx.doi.org/10.1016/j.enbuild.2014.02.068
Bouyer, J., Inard, C., & Musy, M. (2011). Microclimatic coupling as a solution to
improve building energy simulation in an urban context. Energy and
Buildings, 43(7), 1549-1559.
Burian, S., McPherson, T., Brown, M., Streit, G., & Turin, H. J. (2003). Urban
environmental modeling and assessment using detailed urban databases. Earth
Science in the City: A Reader, 56, 303-333. doi: 10.1029/056SP 12
Burns, C., & Hajdukiewicz, J. (2004). Ecological interface design: CRC Press.
Deru, M., Field, K., Studer, D., Benne, K., Griffith, B., Torcellini, P., . . . Crawley, D.
(2011). U.S. Department of Energy commercial reference building models of
the national building stock. Golden, Colorado: National Renewable Energy
Laboratory.
DOE. (2013). EnergyPlus (Version 8.2.7) [Computer software]: U.S. Department of
Energy (DOE). Retrieved from www.energyplus.gov
Elmualim, A., Valle, R., & Kwawu, W. (2012). Discerning policy and drivers for
sustainable facilities management practice. International Journal of
Sustainable Built Environment, 1(1), 16-25. doi:
http://dx.doi.org/10.1016/j.ijsbe.2012.03.001
EN. (2007). EN 15251: Indoor environmental input parameters for design and
assessment of energy performance of buildings addressing indoor air quality,
thermal environment, lighting and acoustics. In E. C. f. S. (EN) (Ed.).
Brussels: EN.
EPA. (2013). Reducing urban heat islands: Compendium of strategies. Retrieved
from http://www.epa.gov/heatisland/resources/compendium.htm.
Erell, E., & Williamson, T. (2006). Simulating air temperature in an urban street
canyon in all weather conditions using measured data at a reference
meteorological station. International Journal of Climatology, 26, 1671-1694.
doi: 10.1002/joc.1328
Erell, E., & Williamson, T. (2007). Intra-urban difference in canopy layer air
temperature at a mid-latitude city. International Journal of Climatology, 27,
1243-1255. doi: 10.1002/joc.1469
Feige, A., Wallbaum, H., Janser, M., & Windlinger, L. (2013). Impact of sustainable
office buildings on occupant's comfort and productivity. Journal of Corporate
Real Estate, 15(1), 7-34. doi: http://dx.doi.org/10.1108/JCRE-01-2013-0004
Kalvelage, K., Passe, U., & Dorneich, M. (2015). Investigating the impact of
reflective facades on the microclimate. Building and Environment.
© ASCE
Sustainable Human–Building Ecosystems 158
Leach, M., Lobato, C., Hirsch, A., Pless, S., & Torcellini, P. (2010). Technical
support document strategies for 50% energy savings in large office buildings
(pp. 141). Golden, CO.: National Renewable Energy Laboratory.
Lee, Y. S., & Malkawi, A. M. (2014). Simulating multiple occupant behaviors in
buildings: An agent-based modeling approach. Energy and Buildings, 69(0),
407-416. doi: http://dx.doi.org/10.1016/j.enbuild.2013.11.020
Mahdavi, A. (2001). Simulation-based control of building systems operation.
Building and Environment, 36(6), 789-796. doi:
http://dx.doi.org/10.1016/S0360-1323(00)00065-2
Patton, S. L. (2013). Development of a future typical meteorlogical year with
application to building energy use. (Master of Science Thesis), Iowa State
University, Ames, Iowa. (Paper 13635)
Reinhart, C. F. (2004). Lightswitch-2002: a model for manual and automated control
of electric lighting and blinds. Solar Energy, 77(1), 15-28. doi:
http://dx.doi.org/10.1016/j.solener.2004.04.003
Santamouris, M., Papanikolaou, N., Livada, I., Koronakis, I., Georgakis, C., Argiriou,
A., & Assimakopoulos, D. N. (2001). On the impact of urban climate on the
energy consumption of buildings. Solar Energy, 70(3), 201-216. doi:
http://dx.doi.org/10.1016/S0038-092X(00)00095-5
Schuetter, S., DeBaillie, L., & Ahl, D. (2014). Future Climate Impacts On Building
Design. ASHRAE Journal, 56(9), 36-44.
© ASCE
Sustainable Human–Building Ecosystems 159
Abstract
INTRODUCTION
Human Energy-related Behavior. During the past decades, many studies have
focused on energy efficiency in built environments and the demand for more
sustainable buildings. Even though the deterministic features of the building energy
simulators appear to be quite useful (Haldi and Robinson, 2008), the randomness of
occupant-related factors and the complexity of human nature make it rather hard to
accurately predict occupant behavior. Clevenger and Haymaker (2006) argued that
accuracy and precision of existing energy simulation models are still uncertain since
predicting the actual performance of buildings is highly dependent upon users’
© ASCE
Sustainable Human–Building Ecosystems 160
behavior. Also, “initial results show that predicted energy consumption changes by
more than 150% using all high or all low values for what experts believe reasonably
represents occupant behavior”. Thus, understanding the way that occupants interact
with buildings is a contributing factor to determining energy saving potentials.
IVE Validation Methods. There have been many studies to date focusing on the
validity of IVE experiments and its measures in order to determine its value as a
measuring tool. The major validity types that should be considered in IVE
experiments are given as below.
Ecological validity. Loomis et al. (1999) state that “IVE technology is a highly
promising tool for the study of basic psychological processes. Its primary advantages
are affording more ecological validity without compromising experimental control
and allowing the decoupling of variables that naturally co-vary”. When it comes to
measuring ecological validity, presence or sense of “being there” becomes an
absolutely inevitable concept to consider. The term presence is generally used to
address the human feelings and perceptions of a virtual environment (Barfield et al.,
1995). The reason behind finding participants’ sense of being there is that the more a
person feels present in a virtual environment, the more his/her response to the stimuli
would match those in the physical environment (Villani et al., 2012).
© ASCE
Sustainable Human–Building Ecosystems 161
Along with testing the level of presence in IVE, the face validity of the
experiment has to be confirmed as well, i.e., the experiment material should look like
they are measuring what they intend to measure and be designed as respondent-
friendly as possible (Oppenheim, 2000).
Content validation in IVE is considered as another requirement to determine
the value of IVE as a research tool, which is referred to as the appropriateness of the
simulator in its functionality (Lyons et al., 2013).
RESEARCH OBJECTIVES
© ASCE
Sustainable Human–Building Ecosystems 162
Tests in physical setting. The experiment room was located in a typical office
building on the LSU campus; the test environment was equipped with office
furniture, computer and specialized tools for the test and the room was illuminated
with ceiling lights as well as a south-oriented window. The tests were carried out at a
specific time of day with certain default amount of outdoor illumination. Ten students
participated in the experiment. At first they were asked for their background
information and after that the experiment procedures started by the reading task in a
dimly lit condition, which continued by completing a visual perception questionnaire.
It was intended to make a thorough examination of individuals’ visual perception, so
a previously tested questionnaire from the study of Lee et al. (2014)was applied to
carefully evaluate subjects’ visual perception and also to observe their subsequent
lighting use behavior. After the lighting condition of the room was adjusted based on
subjects’ preferred choice, they were asked to perform another reading task and fill
out the visual perception questionnaire one more time. Their feedbacks were recorded
for further evaluations and comparison with an IVE test.
Tests in virtual setting. The second test setting was a 3-D virtual environment that
represented almost all of the visual qualities of the in-situ test environment. The
virtual scene was also able to reflect the equivalent natural outdoor light and the
ceiling lights. The experiment room was simulated with a powerful game design tool,
Unreal Engine 4.6, so that it can give participants the opportunity to manipulate some
physical features of their virtual test environment such as turning the light on/off and
opening/closing the window blind. Participants had no previous experience in IVEs,
so they received training of maneuvering in the virtual scene and the use of the head
mount device and the controller before the experiment starts. They were taught to
explore the IVE demo scenes, adjust the equipment and make changes to the
illumination of the scene in a comparable way to those in-situ. Once they became
completely familiar with the environment, the test started and similar to the in-situ
exercise, subjects performed a reading task in a dimly lit virtual environment. Then,
they answered the post experiment visual perception questionnaire and they were
given options to improve their visual comfort. Subsequently, after changing the
lighting condition of the room, they performed another reading task and filled out the
visual perception questionnaire once again. After the reading tasks were over, the ITC
Sense of Presence Inventory (ITC-SOPI) was utilized to assess some validity aspects
of the experiments (Lessiter et al., 2001) and at the very end of the test, participants
were asked to give their opinions about their behavior related to lighting consumption
through Theory of Planned Behavior (TPB) questionnaires (Ajzen, 2006, Ajzen,
1991) in order that criterion validity of the experiment can be supported.
© ASCE
Sustainable Human–Building Ecosystems 163
There are
a four lightting choices in each of tthe physical and virtual test setting,
nammely P1, P2, P3 and P4, which
w respeectively reprresent the exxperiment environment
witth natural and
a artificiall lighting toogether, onlly natural llighting, onlly artificial
ligh
hting and finnally no lig ghting source at all. Th ey are illusttrated in thee following
table (see Tablee 1).
able 1: Experiment Environmentss
Ta
Room Condition
C In-situ photo
p IVE illustration
P1 Blind
d open/
Lig
ght off
P2 Blind
d open/
Ligght on
Blind
d closed/
P3 Lig
ght on
P4 Blind
d closed/
Lig
ght off
EX
XPERIMEN
NTAL RESU
ULTS
© ASCE
Sustainable Human–Building Ecosystems 164
© ASCE
Sustainable Human–Building Ecosystems 165
With the above observations, the authors propose a set of hypotheses for
further testing, which has been summarized and presented in the following table (see
Table 2).
Table 2: Summary of analysis methods and hypothesis testing
Type of Hypothesis Quality Method for Data Gathering
Validity Assessed
Ecological IVE and the in-situ context Presence Post-experiment questionnaire
can produce a comparable (ITC-SOPI)
level of spatial presence and Face Post-experiment questionnaire
engagement (ITC-SOPI)
Content Post-experiment questionnaire
(Qualitative IVE reflection
inquiry)
Internal IVE and the in-situ context Cause and Effect Observation and experimental
can produce a comparable Relation stimulus manipulation control
cause and effect relation
Construct IVE and the in-situ context Visual Comfort 1) Post-experiment
can produce comparable Visual questionnaire (Visual
measurement and path Stimulation perception)
coefficients Mood 2) Behavioral inquiry
Criterion IVE and the in-situ context Attitude and Post-experiment questionnaire
can result in comparable Beliefs (Theory of planned behavior)
behavioral predictions. Norms
Control Beliefs
Actual
Behavioral
Control
© ASCE
Sustainable Human–Building Ecosystems 166
© ASCE
Sustainable Human–Building Ecosystems 167
VILLANI, D., REPETTO, C., CIPRESSO, P. & RIVA, G. 2012. May I experience
more presence in doing the same thing in virtual reality than in reality? An
answer from a simulated job interview. Interacting with Computers, 24, 265-
272.
© ASCE
Sustainable Human–Building Ecosystems 168
Integrated Project Delivery and Total Building Automation for the Nearly Net-Zero-
Energy Q1 ThyssenKrupp Headquarters
Thomas Spiegelhalter1
1
Co-Director Structures and Environmental Technologies Lab (SET-Lab), College of Architecture + The
Arts, Florida International University, PCA, 11200 SW 8th St., Miami, FL 33199.
E-mail: tspiege@fiu.edu
Abstract
These days, the design and construction industry is challenged with the question “how are
architects, engineers, contractors, building owners using information technology to further
automate their practices, building operation and benchmarking?” The following case study with
interviews will critically evaluate the ‘Integrated Project Delivery’ process of the ‘Q1
ThyssenKrupp Quarter’ in Essen, Germany built from 2009 to 2010, and decommissioned in
2012. The Q1-Energy Concept of the ThyssenKrupp Headquarters’ primary energy consumption
is measured at 139.6 kWh/m²/yr and its ecological footprint has by 27% less CO2 emissions than
similar buildings in this category. One of the major innovative systems are the Solar Facade
controlled system and the Q1-Total Green Building Automation (DESIGO). Siemens specifically
developed a Building Automation System (BAS) called Green Building Monitor™ to supervise
the building consumption data of Q1 continuously for expert analysis. DESIGO includes a
specially developed energy-efficient operating control concept for 1,133 rooms.
The schematic and automated workflow for the integrated project delivery of the Q1
ThyssenKrupp Quarter includes the early design stage, 3-D-simulations, detailing, file-to-
production process, assembly, commissioning, benchmarking, and the total green building
automation. This case study evaluates how the German Thyssen-Krupp headquarters’ real-time
SIEMENS total green building automation system (BAS) performs with intelligent control
feedback loops and learning algorithms for constantly optimized building performance, security,
and user comfort operation. The BAS also includes a wireless environmental management
system to ensure trend analysis and optimizations toward yearly net-zero-energy certifications
based on the mandatory yearly European Building Energy Performance Certification system.
Interviews about the 2-D to 3-D/4-D/5-D processes. The Q1 lead architect J. Steffens of JSWD
Architekten in Cologne, Germany, described that the computationnel shift in his office into 3-D
also paralleled and engaged a conceptual and pragmatic shift which was required to navigate and
© ASCE
Sustainable Human–Building Ecosystems 169
manage through increased organizational complexity and required flexibility of the 300,000 m²
gross floor area with 10 different occupancy types and zonings”. According to Steffens, the first
shift in the design process from 2-D AutoCAD / SketchUp to 3-D-Max and Rhino/ Grasshopper
happened when the client demanded a more collaborative and transparent working environment.
The second shift occurred when the complexity of the project demanded realtime 3-D/4-D
simulations with VRfx-compatible formats in OpenGL Performer software to share, reiterate,
synchronize, and visualize quick changes and updates in the parameters with the client and
special contractor direct input. The VRfx open format of the OpenGL Performer was developed
by the German Fraunhofer Gesellschaft (FhG) in collaboration with the Competence Center
Virtual Environments (CCVE) of the Fraunhofer IAO and the Institute for Human Factors and
Technology Management (IAT) of the University of Stuttgart. The spatial 3-D visualization for
direct interaction and immersion into the virtual architecture of Q1 was facilitated in particular
by Caetano da Silva and Prof. Guenther Wenzel from Fraunhofer IAO in Stuttgart for JWSD
Architekten and Chaix & Morel et Associés.
The early ‘integrated design project to fabrication delivery process’ was parallel
supported by the development of the real-time OpenGL Performer analysis and scenario tool to
assist the design phases and also later -after the realization of the headquarter- the real-time
scheduling with the Total Green “Building Automation System (BAS). The integrated BAS
synthesized intelligent control feedback loops and self-learning algorithms for constant
optimized building operation and bench-marking. It was clear at this early level of collaborative
design that the virtual reality tool could be simultaneously a medium, prototype, and rapid
modeling tool to visualize, manipulate data and analyze complex relationships intuitively. The
Fraunhofer Visual Processing of 3-D Content (VRfx) format has similarities with other 3-D
game engines, such as Quest3D® fed by 3-D-data resulting from multiple scenario inputs of
built environment modeling (BEM) tools, to create navigable, immersive, integrated 3-D
architecture models or urban environments using false color. According to Prof. Guenter Wenzel
“these virtual models enhance designers and clients” perception by allowing them to explore
experientially hundreds of potential multidisciplinary design combinations, both qualitatively
and quantitatively, in a practical, collaborative and cost effective way. Participatory processes
with accessible data exchange are essential for getting necessary change and radical design
innovations adopted. Today, designers have the opportunity to create interactive models with
interrelated data that are immersive—including a mix of real and symbolic representation—that
facilitate experiential design and ‘what if’-scenarios’ (Figure 1)
© ASCE
Sustainable Human–Building Ecosystems 170
Design and
a Producction of the intelligen nt multifuncctional Sun n/Daylight Façade Coontrol
System. One of the major
m innov
vative Q1 systems are thhe solar facaade controlleed system annd the
Q1-total green buildding automaation (DESIG GO). The German company Siem mens specifi
fically
developeed DESIGO to operate ass an integratted Building Automationn System (BA AS). This allso so
called Grreen Buildin
ng Monitor™ ™ is created to supervisee in real-tim
me the buildiing operationn and
assemblees continuouusly consum mption data for the exppert and straategy analyssis. For exaample
DESIGO O includes a specially developed
d en
nergy-efficieent operatinng control cooncept for 1,133
rooms. Sensors
S record daylight and detect the
t presencee of people, which allow ws the systeem to
provide optimal
o worrkspace therrmal and lighting condittions based on those reeal-time readdings.
The system displayss accurately to the hourr – how mucch power annd water is consumed iin the
building,, the tempeerature, and greenhousee gas emisssions. BAS monitors uuser humann and
building systems beehaviour, compares, meeasures and visualizes rreal-time progress data with
© ASCE
Sustainable Human–Building Ecosystems 171
various performance
p benchmarkss for carbon neutrality oor the nearly zero-fossil eenergy standdards.
Other feaatures includ de integrated control fo or electrical and securitty technologgy tailored tto the
specific and
a dynamicc needs of th he ThyssenK Krupp campuus buildings ((Fig. 3).
The
T Sun and daylight con ntrol system
ms are equippped with suittable glare reemediation w while
still givin
ng the buildiing occupannts sufficientt outviews annd visibilityy require an integrated ddesign
approach h. The archiitects design ned successffully a direcct link to thhe dynamicc and perpettually
evolving patterns of outdoor illu umination an nd direct solaar radiation with a uniqque controlleed but
dynamic and adaptiv ve sensor controlled day ylighting faç ade system. It is beneficial clearly helps
to create a visually sttimulating annd productivve environmment for the bbuilding occuupants whilee also
reducing significantly the total building
b enerrgy costs annd greenhousse gas emisssions (Fig. 4). All
the Q1 office
o m steel profilles and operrable verticaal lamellas. In total therre are
façadees have slim
400,000 stainless steeel slats coveering an areaa of approxim
mately 8,0000 m² to contrrol sun and
Figuree 3: Images ab
bove: Q1 Th
hyssenKrupp p Quarter witth automated d solar and d
daylight conttrol
system, JSWD Arch hitekten and Chaix & Mo orel et Associiés, 2006. (Im
mage: Thomaas Spiegelhallter,
2013.) Images beloow: Q1 lamellla performance angles peerpendicularr to the glasss façade. Imaage
right, bottom:
b The virtual
v anim
mation of the façade
f by Frrener + Reifeer with digitaal 3-D mock-ups
(So
ource: Freneer + Reifer, S
Sept. 2013.)
© ASCE
Sustainable Human–Building Ecosystems 172
fully functional even in high wind speeds of 70 km/hour, which is unusual for highrise buildings.
Most outside shading and daylight systems in similar building types around the world turn off
during the high-speed wind season, and as a result their energy consumption increases
dramatically due higher cooling loads. The solar and daylighting control concept was
schematically designed by JSWD Architekten and Chaix & Morel et Associés and
algorithmically detailed and optimized through multiple simulations and digital mock-ups
(DMUs) by the Fraunhofer Institute for Solar Energy Systems (FhG) in Freiburg, Germany in
collaboration with the special façade contractor Frener+Reifer from South Tyrol in South Tirol,
Italy. The DMUs included a parametric algorithmic dynamic weather data-driven solar radiation
and illumination analysis with multizone simulations to engineer a constant horizontal overhang
for summer sun protection and daylight control. All systems were combined with a vertical set of
twisting fins. The matte underside of the system was further designed by the architects and
engineers and Frener+Reifer after the FhG modeled and optimized various property behaviors
such as reflection, shadow, daylight, glare, and heat protection (Fig. 4). All the shading elements
are sensor controlled through a separate EIBSKNX- Bus system and embedded for data
recording in the SIEMENS BAS with the assistance of 1,600 electric motors to dynamically
perform to the following settings: Closed or parallel to the glass façade Moving with the sun /
variable and perpendicular to the sun Open: the horizontal lamellas move perpendicular to the
glass façade (Fig. 4).
The preliminary design-to-factory-file process of the innovative façade system was
simulated and a 3-D mock-up tested by the Fraunhofer Institute of Solar Energy (ISE) and the
special façade contractor Frener and Reifer from South Tyrol in Italy under the supervision of
General Planner ECE and the client ThyssenKrupp Real Estate. Then the different analog and
digital file formats for the Q1 value chain control batch of 22,000 lamellas were tested, cleaned,
calibrated, and digitally processed by the Simultaneous System Engineering group of
ThyssenKrupp Umformtechnik and ThyssenKrupp Nirosta, the group’s automotive
manufacturing units.
In general, the ThyssenKrupp System Engineering workflow contains an integrated
project process and system management for automotive systems and parts. This includes all
project phases beginning with the parametric concept study through to production design,
physical and digital mock-up (DMU) testing, assembly, and production planning—in this case—
to the façade system and component delivery for Q1. The integration of testing within the early
phase of the development is predominantly important Frener+Reifer, Thyssen Krupp
Umformtechnik, and Nirosta used for -controlled just-in-time (JIT) supply chain production such
as the integrated KANBAN supermarket system and optimization protocols with RFID radio
frequency identification (RFID).
The programming of the sensor-driven control system includes real-time daylight
provision in the interior, glare control, seasonal sun position, and the weather patterns assisted
through the weather station on the roof. The fully automated façade system is separated with a
parallel the logistics from fabrication, and transportation to assembly a barcode for the
correlation of multidimensional simulation models through finite element methods and
contributes to the verification of the computation results before any fabrication planning takes
place and bus system from the overall Siemens BAS and energy control system, coupled in a
central control gate to minimize the risk of collisions of the sun protection lamella components
and potential data package overloads that could occur with direct decentralize data package
integration into the overall BAS control system. Based on the previous algorithmically
© ASCE
Sustainable Human–Building Ecosystems 173
The
T German n Thyssen-KKrupp headq quarter’s reaal-time SIEM MENS totaal green buiilding
automation system (BAS) performs with intelligent control feeedback looops and leaarning
algorithm
ms for constaantly optimized building g performannce, security,, and user coomfort operaation.
This systtem also inclludes a wireeless environ
nmental mannagement sysstem to ensuure trend anaalysis
and optiimizations toward yeaarly mandattory net-zerro-energy ccertificationss based onn the
mandatorry yearly European Bu uilding Enerrgy Performmance Certifi fication systeem. The BA AS is
centralizeed, interlinkked, and a sensor driiven humann-computer-iinterface (H HCI) networrk of
hardwaree and softwaare. The leaarning autom mation systeem ensures the operatioonal perform mance
(transporrtation, light, water, HVAAC, energy , storage annd distributioon, etc.) of thhe Q1-8 faciilities
as well as the comforrt and safety
y of building occupants. TThe BAS opperates with intelligent aagents
(IAs) annd machine learning algorithms
a by
b identifyinng patterns for real-tiime optimizzation
potential including time
t schedulling and trend logging and verificaation of buillding autom mation
process.
Figure 4:
4 Room au or solar shading, dayligght control, artificial ligghting, heatting,
utomation fo
and cooling,
c dew
w point control, benchm
marked with h Energy Effficiency Claassification A
accoording to th
he BACS Eu uropean Norrm (EN) 152232. Diagraam adapted by Thomass
Spiegellhalter. (Sou
urce: Huberrt Dierkes, Siemens
S Inffrastructuree & Cities S Sector Build ding
Technoloogies Division, Essen, G
Germany, 2013.)
© ASCE
Sustainable Human–Building Ecosystems 174
Figure 5: Overa
all systems topology of the
t buildingg services w
with calculatting and selff-
learnin
ng algorithm
ms. Diagramm adapted an nd translateed by Thom
mas Spiegelhhalter. (Sourrce:
Hubert Dieerkes, Siemens Infrastrructure & C
Cities, Essen
n, Germanyy, 2013.)
BENCH
HMARKINGG RESULTTS THROU
UGH THE
E Q1 TOT
TAL GREE
EN BUILD
DING
AUTOM
MATION (D
DESIGO)
LUSION
CONCL
The
T author iss still in neg
gotiation witth SIEMENS
S and Thysssenkrupp to receive the three
years (2
2012-2015) real-time operating
o Q1
Q Total G Green Buildiing Automaation (DESIGO)
© ASCE
Sustainable Human–Building Ecosystems 175
benchmarking data. Unfortunately, the benchmarked real-time data was not yet permitted for
release. It would have helped to analyze the efficiency of how much carbon has been reduced
and energy costs for revenue growth with return on capital employed (ROCE). These pending
results will definitely feed into a future research case study with new post-occupancy interviews
and benchmarking.
In summary, one can argue that already today and in the future, automated green building
design-to-factory-file manufacturing and building performance automation will go naturally
together with faster and more flexible customization and corporate sustain ability strategies to
reduce carbon footprints and energy costs for revenue growth with return on capital employed
(ROCE). Over the next couple of decades, we will see major enhancements in automated
scenario network planning and in high-speed cloud computing that will further improve resource
innovations and flexibility. On the side of architects and engineers the next generation will
design and operate within a type of Green Automation, where designers and engineers deal with
graphical descriptions of system and complex cloud software with machine learning algorithms
that automatically translate repeatedly new models into optimized executable software. This new
era of green automated design to-virtual-to-real manufacturing, building operation and
automated benchmarking and scenario reporting will reorder the global AEC business for
decades.
REFERENCES
© ASCE
Sustainable Human–Building Ecosystems 176
Abstract
Sustainability and energy efficiency are crucial to many sectors and all
stakeholders at local, municipal, national, and global scales. According to the city of
Vancouver, buildings contribute more than 50 percent of the city’s total GHG
emissions. With the introduction of standards such as LEED the hope has been
significant reductions in the footprint; however, change has been much lower than
predicted. Much of this can be attributed to how people actually use the buildings
they inhabit. Preliminary results of semi-structured interviews with residents of a
Vancouver multi-residential LEED Gold building show lack of feedback the most
common barrier among inhabitants who intend to save energy. Lack of appropriate
affordances and constraints follow respectively.
INTRODUCTION
Conservation and efficient use of energy are now major items for discussion
and action by governments, public companies, private groups, and individual people.
Various technical and non-technical sustainable solutions have been proposed at both
macro scales (urban fabric), and micro scales (individual buildings). However, it is
mostly only in the design of the urban fabric that we see evidence of people involved
in the design process (Haas, 2012). In the case of individual buildings, strategies
ranging from passive, active, renewable energies, renewable and non-renewable
materials, efficient building management system, and water and waste management
have been researched and discussed in detail (Fuad-Luke, 2002; Kibert, 2012;
Lechner, 2009). What is lacking in all these approaches is the effects buildings have
on the people using them, and, conversely, the effects of inhabitants on the built
environment. Furthermore, there is an extremely limited amount of literature that
looks at sustainable, green building design with the lens of inhabitants as the main
focus (Nielsen & E-Types (Firm), 2010; Strengers, 2013).
It is useful to think of buildings in terms of how people can affect them and
visa versa. Though this notion might seem complex and abstract, it is evident in
architectural history. Taking a careful look at the development of architecture and the
built environment throughout the years, we will see that as the interpretation and
perception of life changes, architecture has been transformed. The notion of
‘Functionalism’, which first appeared in 1932 in a letter addressed to Alberto Sartori,
was meant as a substitute to ‘rational’. Later, the meaning of functionalism morphed
© ASCE
Sustainable Human–Building Ecosystems 177
Research Goal & Research Questions. Our goal is to view current frameworks of
green architecture and engineering through the lens of human-computer interaction
(HCI) in order to define new approaches to sustainable design thinking for architects.
HCI focuses on how people use technological systems and information and the
implications of designing and deploying those systems; buildings themselves are
complex combinations of systems. We re-frame the concepts of usability and design
from the HCI field to building design and propose a new framework for architectural
design interventions where our re-framing provides the main lens for both analyzing
sustainable architecture and providing guidelines and methods for architects to better
design for effective inhabitant use.
Our general research questions are: How can we devise and apply a new
interaction design lens and methods to the current green building frameworks to
augment their performance? What can designers and architects of the built
environment do differently in their process and approach to design?
In this short paper we start by describing the key principles and frameworks
for the study, followed by details of the semi-structured open-ended interviews, a
discussion on the results of the study, and finally a brief on future work.
This paper aims at bringing a human centered lens to the field of green
architecture and engineering by reframing the way in which these two fields are
connected. Hence, we propose figure 2 by adapting figure 1 in the context of green
building design. Here, we identify specific concepts from HCI that are particularly
applicable. Similarly, from the social sciences, we identify key concepts for
understanding behavior in the context of green buildings for enhanced energy
efficiency. The largest change combines the nodes of engineering and design into a
single node, structured by the building artifact itself, not by the decision-making
© ASCE
Sustainable Human–Building Ecosystems 178
structure by which it comes to be. The following section elaborates concepts from the
now three areas: HCI, social sciences and buildings.
Social Sciences
Barriers
Social HCI
Structure
Skin
Service
Space
Stuff
Figures 1 & 2 from left to right: (1) Human Computer Interaction draws from
social sciences as well as engineering and design disciplines (Shneiderman, 1992).
(2) Reframing figure 1 to account for architecture and engineering specifically in
the realm of energy saving.
HCI Common Engineering & Design Principles. In the widely cited popular book
of The Design of Everyday Things, Norman identifies seven design principles that
should guide every object we interact with in our lives (Norman, 2002). The
principles defined relevant to the field of HCI are:
1. Affordances: This concept provides a different perspective to the design of
environments and objects. It highlights the relationship between design intentions
and the user. For example, the form of an object or building and the behaviour of
the user/occupant as these link in practice (Maier, Fadel, & Battisto, 2009). For
example, a glass can be for seeing through and breaking. Adding the notion of
breaking to the actual function of glass is part of its affordance. This principle can
be essentially described as the functions or possibilities that are offered to the end
user through design.
2. Feedback: Systems should be able to give instant feedback that is “immediate and
synchronized with the user action” (Kai, 2010, p. 1). For example, playing a piano
provides immediate sound and key “feel”. There is a huge body of research on the
cognitive basis of feedback in interactive systems (Cialdini, 1993; Fogg, 2002).
3. Constrain: Lacing restrictions to some interactions can reduce the chance of mis-
use or errors. If we combine constraints with persuasion and feedback and make
sure each approach is selected based on the user’s behaviour at that point of time;
we can achieve more than either alone (Norman, 2002).
Barriers to Energy Saving. Lifestyle and living conditions affect how each person
faces barriers and impediments to saving energy. There are impediments that can be
modulated by small interventions, while other barriers need change at a larger level,
© ASCE
Sustainable Human–Building Ecosystems 179
for example th hrough policiies. Althoug gh the latter has demonsstrated influuence in the
red
duction of en nergy consu umption, thee former hoolds significaant value, ssince it can
happpen in a shoorter time fraame and may y be easier tto achieve. BBased on an online semi
stru
uctured survvey conducted among 274 2 people residing in British Collumbia, we
disccovered the most domin nant barrier to
t energy saaving is lackk of informattion or mis-
info
ormation, foollowed by cost,
c attitudee and preferrence, technnical and spaatial design
bassed barriers, living condiitions and lasstly perceiveed lack of aggency.
Arcchitectural Design Fra amework. A very comm mon framewoork for desiggn practices
fromm 1994 is to t look at buildings
b wiith the lens of what theey are madee and what
happpens to them m after they are construccted (Brand,, 1994). Builldings compprise several
layers, with wh hich the inh
habitants con nstantly interract and whhich change at different
ratees. In How Buildings
B Leearn, Brand discusses hoow buildinggs adapt to change. One
of the
t key argu uments he makes
m is in terms
t of thee different shhearing layeers (“shear”
refeers to the rellative longev
vity of buildiing componeents) in eachh building. H He proposes
thatt buildings can
c be seen as comprisiing six layerrs (figure 3).. He definess Site as the
settting and surrrounding off a structure,, Structure aas the founddation and looad bearing
elemments of a building, Skkin as the laayer that coovers the strructure, Servvices as the
commmunication n hub of the building, HVAC
H systemms, electricaal wiring, eleevators, etc,
Spaace plan ass the interio or layout off a buildingg, and Stuff ff as the furrniture and
belongings of occupants.
o
Figure
F 3: Bra
and's six sh
hearing layers of changee, and the n
number of yyears they
can exisst (Brand, 1994)
INH
HABITANT
T STUDY
Tallking to Reesidents. Baased on figuure 2 we carrried out a sstudy that eengaged the
inh
habitants of “green” buiildings, baseed on semi--structured, open-endedd interviews
witth 12 residen
nts of a LEE
ED Gold mullti-residentiaal apartment unit in Vancouver-BC.
Alll residents were
w owners.. As such, th
he lifestyle ffactors that iinfluenced ttheir energy
con
nsumption were
w similarr. We used the conceppts and structures of thhe previous
section in designing the interview questions.
q Inn particularr, we took a “human-
com nding these ppatterns: in essence, we considered
mputer interaaction” lens to understan
the inhabitant as a user and the buiilding as a complex syystem with which the
© ASCE
© ASCE
Elements of Opportunities- Opportunities- Opportunities- Barriers
Architecture Affordances Feedback Constraints
bottom.
Structure
in the space.28
Odd placement of air vent.29
Stuff Larger sink in kitchen.6 Beeping sound from dryer, Predetermined wash No feedback from the stove.
Dual Flush Toilet. dishwasher and cycles .14 Marble floors in the
Task lighting and light fridge/freezer.8 Smart Dryers.15 bathroom.30
controls.7 Controlling blinds.9 Smaller showerheads .16 Nice rain shower.31
Bedroom power fuse Smaller load washer Lack of dimmers on light
popping if too much /dryer.17 switches.32
electricity used at one point Smaller fridge.18 Lack of lights in the den
points are elaborated in the order of left to right, and within each column from top to
of time.10 Faucets with less water and bedrooms.33
Table 1 briefly summarizes the main points that came out of the study. Below the key
Results. We categorized data from discussions with residents using the specific
concepts from figure 2, and mapping the HCI and social science concepts against
inhabitant interacts. We discussed living experience, thermal comfort, lighting,
pressure.19
Door handles.20
LED Lights only
180
Sustainable Human–Building Ecosystems 181
Opportunities - Affordances:
1
Building is located in an area with close proximity to transit and general services.
2
Floor to ceiling windows afford the transmission of light, and hence illumination of
the interior environment as well as a view of the exterior environment. “Even on
rainy days apartment feels cozy”.
3
Operable windows may afford the exchange of air.
4
Apartment size is much smaller than their large house.
5
High ceiling make the space look bigger, without having to actually add to the size
of the apartment.
6
Larger sink in kitchen allows for more water to fill up, hence more flexibility of
choice.
7
Toilet affords dual flush for less water use.
Opportunities - Feedback:
8
Beeping sound from dryer, dishwasher and fridge/freezer.
9
Controlling the blinds gives you direct experience of something that can indirectly
control energy use. But it affords privacy.
10
Bedroom power fuse popping every half an hour if all plug outlets used at the same
time (ie, working from home and plugging in computers and printers, or electric
heaters).
Opportunities - Constraints:
11
No yard means less maintenance and water use.
12
Lack of garbage shoot causing more recycling habits, because neighbours might see
what each household is throwing out in their trash bags.
13
Bedroom power fuse popping every half an hour.
14
Predetermined wash cycles afford use of less water and energy.
15
Smart dryers stop when detecting the clothes are dry, affording less energy use.
16
Smaller showerheads allow less water out in each unit of time.
17
Smaller load capacity in washer/ drier allows frequent washing with less waste.
18
Smaller fridge means less groceries, more frequent shopping, and less rotting
vegetables and waste.
19
Faucets are water savers cause they have less pressure.
20
Door handles that do not afford the understanding of whether they should be pushed
or pulled, hence leaving them open at all times. This way there is visibility of whether
any lights have been left on when leaving the apartment.
Barriers:
21
Immediate areas around the large windows feel cold in winter and warm in summer.
22
Windows in the bathroom, make bathrooms very cold, hence residents turn heat up.
23
Lack of separate zones for heating and cooling in the home cause residents to
heat/cool up all the apartment even if they only intend to heat/cool up one room/space.
24
Extremely complex thermostat and settings cause mis-use of the thermostat.
25
Pre programmed bathroom fan, which is not visible but automatically turns on and
off creates confusion and extra waste of electricity since inhabitants also turn the fan
on and off when required.
26
Bathroom drainage has a bad odor coming out of it causing residents to constantly
pour boiling water in it, hence wasting hot water.
© ASCE
Sustainable Human–Building Ecosystems 182
27
Pot lights are on the perimeter of the living room, and light junction boxes in some
places but never in the right location. So, all the lights have to be on, or furniture has
to be moved around to fit the junction box.
28
Placement of light switches in the hallway leading to bedrooms is way too far from
the bedroom itself.
29
Odd placement of air vent, giving hot air in summer and cold air in winter
constrains the possibility of sitting beside the window.
30
Marble floors in bathrooms feel cold to walk on causing residents to increase heat.
31
In some bathrooms with nice rain showers inhabitants tend to take longer showers.
32
Lack of dimmers on light switches causes one to waste more electricity.
33
Lack of light in the bedroom and den causes residents to try to brighten up these
rooms through turning on the light in other areas of the home excessively.
DISCUSSION
Results from talking to residents indicate that inhabitants perceive that the
structure layer has little influence. Though the LEED point sheet in this building
clearly indicated the skin and the site to have obtained the highest points, inhabitants
saw constraints in stuff to be the most dominant opportunity while saving energy.
This once again refers back to our initial goal to focus on the way inhabitants interact
with their surroundings, because despite the number of points a green building might
score, it ultimately comes down to the actions inhabitants take. In addition to the
categorical insights from residents we have included a set of direct quotes below:
Inhabitant#2: “LEED means luxury. It is a buzzword that adds to the resale value”.
Inhabitant#3: “I never got to see where the LEED standard is happening. How can I
know it is LEED?”. Inhabitant#6: “If LEED is what you say, then the building is
taking care or energy saving itself. Energy saving is automatically in place”.
Inhabitant#8: “In general I don’t think about energy saving. If I saw the LEED thing
it would be a great learning tool for both kids, and adults. I absolutely would love to
have a list of LEED points”.
These comments all point to the fact that lack of information is the most
dominant barrier inhabitants face to saving energy. Though this wasn’t initially part
of our reframed notion of architectural design, but it provided additional valuable
insight that can be useful for the multiple players involved in the design and
execution of green buildings.
© ASCE
Sustainable Human–Building Ecosystems 183
REFERENCES
Alexander, C., Ishikawa, S., & Silverstein, M. (1977). Pattern languages. Center for
Environmental Structure, 2.
Bartram, L., & Woodbury, R. (2011). Smart Homes or Smart Occupants? Reframing
Computational Design Models for the Green Home. Proceedings of the AAAI
Spring Symposium on Artifical Intelligence and Sustainable Design, 1–8.
Brand, S. (1994). How buildings learn: what happens after they’re built.
Cialdini, R. B. (1993). Influence: The psychology of persuasion. New York, Morrow.
Fogg, B. J. (2002). Persuasive technology: using computers to change what we think
and do. Ubiquity, 2002(December), 5.
Fuad-Luke, A. (2002). The Eco-design handbook. Thames & Hudson London.
Haas, T. (2012). Sustainable Urbanism and Beyond: Rethinking Cities for the Future.
Kai, C. (2010). Design Principle for Ubiquitous Computing.
Kibert, C. J. (2012). Sustainable construction: green building design and delivery.
Wiley. com.
Lechner, N. (2009). Heating, cooling, lighting: sustainable design methods for
architects. John wiley & sons.
Lockton, D., Harrison, D., & Stanton, N. (2008). Making the user more efficient:
Design for sustainable behaviour. International Journal of Sustainable
Engineering, 1(1), 3–8.
Maier, J. R., Fadel, G. M., & Battisto, D. G. (2009). An affordance-based approach to
architectural theory, design, and practice. Design Studies, 30(4), 393–414.
© ASCE
Sustainable Human–Building Ecosystems 184
Nielsen, K. H., & E-Types (Firm), D. arkitekturcenter. (2010). Mind your behaviour:
how architecture shapes behaviour. [Copenhagen]: 3XN.
Norman, D. A. (2002). The design of everyday things. Basic books.
Shneiderman, B. (1992). Designing the user interface: strategies for effective human-
computer interaction (Vol. 3). Addison-Wesley Reading, MA.
Strengers, Y. (2013). Smart Energy Technologies in Everyday Life: Smart Utopia?.
Palgrave Macmillan.
© ASCE
Sustainable Human–Building Ecosystems 185
Abstract
Ever since the first global oil crisis in the seventies, the shortage of fossil
fuels, which is considered as the number one resource for energy carrier, has
emphasized the need of modern society on cheap energy and resources. Over
short or long term, that circumstance is compelling people to consider global
energy strategies and accordingly take proper measures. Along with a limited
supply of carbon based fuels worldwide, the impact of the carbon footprint— the
effect of an increasing carbon dioxide (CO2) concentration in the earth’s
atmosphere and its consequence on the global climate — has become indubitably
clear. Recent studies show that building sector consumes significant amounts of
energy and is responsible for a considerable amount of greenhouse gas (GHG)
emissions. Studies show that amongst several factors contributing to the building
energy performance, occupant behavior is one of the driving factors. According to
these studies building characteristics and the quality of indoor environment
significantly influence the level of satisfaction and efficiency of occupants in
buildings. Therefore, a proper design is needed to increase building occupant’s
satisfaction and efficiency.
To determine the best design strategy, several building envelope materials
as well as occupant behaviors should be considered as design variables in the
design process. Therefore, the objective of this study is to develop a proper
weighting system for different strategies to secure occupants’ satisfaction as well
as constructors’ consent. To achieve this objective, a Quality Function
Deployment (QFD) method is utilized to identify, translate, and implement the
end-user requirements into the design process. In addition, the utility theory is
used to reflect the constructors’ perception on usefulness and practicality of end
user requirements. The insights from this study will aid in developing better
decision making system and more viable alternatives.
© ASCE
Sustainable Human–Building Ecosystems 186
INTRODUCTION
Ever since the first global oil crisis in the seventies, the shortage of fossil
fuels has emphasized energy efficiency in industries such as construction and
transportation and the use of cheap energy and resources. Along with a limited
supply of carbon based fuels worldwide, the impact of the carbon footprint (i.e.
the effect of an increasing carbon dioxide (CO2) concentration in the earth’s
atmosphere and its consequence on the global climate) has become indubitably
clear. Recent studies show that building sector consumes significant amounts of
energy and is responsible for a considerable amount of greenhouse gas (GHG)
emissions. According to the United Nations (UN), the building sector consumes
40% of global energy consumption and has a crucial role in mitigating the impacts
and risks of global warming [1] .
Optimal energy-efficient building design should consider multiple
objectives (e.g. energy consumption, financial costs and environmental impacts)
and thus multi-objective optimization is a promising direction to follow. There are
numerous optimization methods to find the tradeoff between different design
variables and optimum strategy for construction of an energy efficient, low cost,
and environmentally sustainable building [2][3][4][5][6]. In a study performed by
Weimin [7], the environmental impacts of buildings were investigated. The
objective of this study was to minimize the life cycle cost (LCC) and life cycle
environmental impact (LCEI). To achieve the objective, the genetic algorithm was
used to develop the multi objective model. They demonstrated that the utility
structure has a significant impact on the building’s environmental performance. In
other study conducted by Hamdy [6], a three-phase genetic algorithm multi-
objective optimization model (PA-GA-RF) was developed and utilized to
minimize the CO2 emission and investment cost of a two story building and its
HVAC system.. It was found that the optimal combinations or Pareto front could
reduce the carbon dioxide equivalent (CO2-eq) emission by 32% and initial
investment by 26%. In another study, Fesanghary et al. [8] developed a multi
objective optimization based on harmony search algorithm (HSA) to minimize the
LCC and CO2-eq. They considered a number of different building envelops as
design variable parameters in optimization process. A comprehensive life cycle
assessment including pre-use (extraction, raw material processing, transportation,
and construction), use (all emissions during building life), and end-of-life
(demolishing, and landfilling) was considered in this study. According to defined
model, authors suggested series of solutions for economically and
environmentally efficient design.
While these optimization models are capable of determining the best
construction strategy, implementing the end-user satisfaction preferences into the
optimization process is still a missing puzzle piece. Satisfaction metrics — the
state of fulfilment—are generally qualitative and depend on many different
factors. They can be recorded by short surveys and questionnaires of participants’
attitudes toward a given design or issue. For example, Brokelman et al.[9], applied
visual analogue scale (VAS) to assess patients’ satisfaction in the health care
system. In this method, patients’ satisfaction rates are marked on vertical line with
length of 100 units. VAS is the horizontal line having two descriptors
demonstrating extreme satisfaction levels (lowest and highest satisfaction level) at
both ends. The validation performed by measuring Spearman correlation test
which showed that pain is the most important patient satisfaction factor. In
© ASCE
Sustainable Human–Building Ecosystems 187
another study, Castillo et al. [10] identified and modeled users’ satisfaction of
public transportation system. Satisfaction level of different parameters of
transportation system comfort were determined through a survey asking
individuals to rate their satisfaction level. Three models including average based
model, multivariate discrete distribution and a generalized linear model were used
to model the overall satisfaction level of individuals as a function of specific
satisfactions.
There are numerous clear evidences denoting that building characteristics
and the quality of indoor environment significantly influence the level of
satisfaction and efficiency of occupants in buildings. Therefore, to fill this gap and
increase the practicality and satisfactory level of the proposed strategy the
objective of this study is to develop a proper weighting system for different
possible strategies to assure occupants’ satisfaction as well as constructors’
consent.
METHODOLOGY
© ASCE
Sustainable Human–Building Ecosystems 188
=1÷ −
RT can be estimated by the following equation:
( )= 0.5 × + 0.5 × × (− )
The single utility function can be easily derived if the RT value is known.
The RT value can be computed from the certainty equivalent (CE) by inversely
calculating the RT value from the CE. Thus, the first step to obtain the single
attribute function can be the determination of the CE (which is equal to 3) for
each attribute. The CE value of each attribute is decided by the decision problem
such that the decision maker is indifferent between the CE and a lottery between
the best and worst in the corresponding attribute. The detailed procedure and
calculation is shown in the result section.
In the effort to create a useful and practical strategy to increase occupants’
satisfaction, 25 different parameters having the most influence on the occupants’
© ASCE
Sustainable Human–Building Ecosystems 189
RESULTS
Quality of furniture
□ - Medium (3) Amount of wiring
No. of Equipment
Temp Insulation
∆ - Low (1)
Meet standards
AC insulation
Having windows □ ∆ ○ ○
Enough lighting □ ○ □
Individual desk lamps □ ○ □
Proper Temperature
IAQ
(Summer) □ ∆ □ ∆
Proper Temperature (Winter) □ ∆ □ ∆
Noise Level (Instruments) ∆ ○ □
Noise Level (Coworkers) ∆ ○
Noise Level (outdoor source) □ ○
Having Private Space ○ ∆ ∆ ∆ ∆
Spaces
Switches ○
Temperature Adjusting ○ ○
Wireless Internet □ □
Equipment
Printer ∆ □
Copy machine ∆ □
Scanner ∆ □
Desks ○
Furniture
Chairs ○
Bookshelves ○
Indoor plants □
Texture
Flooring texture ○
wall texture ○
© ASCE
Sustainable Human–Building Ecosystems 190
( ) = −0.758 + 0.193 × .
Table 2 shows the effectiveness and practicality utility of each parameter.
As it can be seen in this table, proper temperature in summers and winters are
having the highest effectiveness utility and providing offices with better chairs
and desks are having the highest practicality utility. On the other hand providing
enough storing space and having access to copy machine are having the least
effectiveness utility while having windows in offices and proper noise level due to
coworkers are having the least practicality utility.
Experts’ Attitude. Based on the given utility functions for usefulness and
practicality of occupants’ satisfaction parameters, the following utility graphs can
be plotted. As Figure 1 shows, constructors are having negative attitude
(downward curvature, risk averse attitude) toward usefulness of the proposed
© ASCE
Sustainable Human–Building Ecosystems 191
parameters, however they have positive attitude (upward curvature, risk seeker
attitude) toward practicality of this parameters. Therefore utilizing the strategies
with higher utility of usefulness can be an effective motivation for selecting the
optimum strategy.
1
Utility 0.8
0.6
0.4
0.2
0
2 2.5 3 3.5 4
Usefulness
1
0.8
Utility
0.6
0.4
0.2
0
2 2.5 3 3.5 4
Practicality
The list of top 15 parameters and associated rankings based on the constructors’
usefulness utility and associated assigned weighting are given in Table 3. These
weightings are calculated using the following equation:
( )
( ) = ∑
( )
Where W(X) is the weight of each strategy, U(X) is the utility associated to each
of the strategies.
© ASCE
Sustainable Human–Building Ecosystems 192
CONCLUSION
REFERENCES
© ASCE
Sustainable Human–Building Ecosystems 193
Users Surveys,” Procedia - Soc. Behav. Sci., vol. 54, pp. 1104–1114, Oct.
2012.
[11] D. Bernoulli, “Exposition of a new theory on the measurement of risk,” J.
Econom. Soc., pp. 23–26, 1738.
[12] D. L. Thurston and A. Locascio, “Decision theory for design economics,”
Eng. Econ., 1994.
© ASCE
Sustainable Human–Building Ecosystems 194
Abstract
A key factor to promote energy efficiency across campus and buildings is
through education and clarity of information. Energy data is complex to
understand and difficult to comprehend for the majority of users. Higher
education institutions consists of building groups that usually operate according to
strategies pre-set by the operation teams that respond to a specific energy demand
condition. The purpose of this research is to design a data friendly interface to
simplify and assist in the implementation of energy literacy across campus. The
research method utilized available energy metered and simulated data at the
campus level provided by the university operations. A web based data
visualization interface is developed and an energy map is created to inform the
campus community on the status of energy consumption. In addition to
visualization, the proposed interface helps operators to benchmark campus
buildings, promote energy efficiency across campus and disseminate the resulting
information in a simple, clear and concise way to the campus community. The
interface is a great step towards change of behavior through education, a
questionnaire to measure user actions and understanding is distributed to measure
the effectiveness of this new approach and the results are compiled to advance the
interface design and user experience.
INTRODUCTION
© ASCE
Sustainable Human–Building Ecosystems 195
Malkawi et. al. developed a web-based user interface that employed simulated
thermal data of buildings to provide an energy-based decision tool for windows by
predicting its performance (Malkawi, 2007). However, as energy performance
change over the lifetime of the building, energy-based decisions concerning
systems and components should be included as well. Using only simulated data
for buildings will hardly give a real picture of how a building is actually
performing, hence in day-to-day operations, constant data feed and continuous
monitoring and analysis of the actual energy data is crucial to improve energy
performance and retain savings.
Domínguez et al. proposed a tool to collect energy data from university
buildings through data mining, monitoring and statistical techniques that led to
economic savings of around 15%. The three-layer structure of the developed
system started with acquiring data from campus meters, this data was stored and
processed then the client was able to access it online as a monitoring interface and
conduct a number of relevant statistical analyses (Domínguez, 2013). The
measured energy data from campus buildings was used to calibrate simulated
energy models. The calibration of simulated energy models often result in better
prediction for the building’s performance and accordingly better decisions were
made regarding the different building systems, therefore allowing for streamlined
calculation of the impacts was done with higher precision. For example, it was
possible to use collected real-time campus data, the way Dong et al. did, to
develop a building energy model and calibrate it to get performance predictions
within ±15% error band. Furthermore, the model was used to adjust energy
consuming such as the HVAC system and achieved a high savings rate of 39%
(Dong, 2014).
Recently, Yarbrough et al. utilized a university campus energy data
visualization and comparisons for a more specific analysis. The focus was
directed to the peak energy, tracking the buildings contribution and assessing the
type of activities and uses that add-up to that peak value (Yarbrough, 2015). In
addition, the monitored energy data was visualized and made into a tool for
comparative analyses ultimately to assist occupants’ with the required knowledge
that can lead to change in behavior and accordingly in saving even more energy.
A six week empirical study in a residential building on the campus of
Columbia University by Jain et al. demonstrated how the interface components of
the eco-feedback system influence energy consumption. The study correlated
users’ engagement with different interface tools to the reduction in energy
consumption (Jain, 2012). Gulbinas et al. experimented the effect of informing
commercial building occupants with their energy consumption through the
development of a modular socio-technical energy management system, BizWatts.
This system however, only provided the real-time data of plug-load appliances.
Nevertheless, the study concluded that providing the occupants with a network
level eco-feedback resulted in a statistically significant energy savings (Gulbinas,
2014). Another study was conducted by Emeakaroha et al. on students residential
halls, using IPTED (Integration of Persuasive Technology and Energy Delegate)
and real time energy feedback energy consumption reduction in 8 halls reached
37% amounting to savings of 1360.49 kWh, and 713.71 kg of CO2 compared to
the baseline (Emeakaroha, 2014).
In summary, University settings represented a convenient medium to
initiate energy performance tool prototype demonstration through actual energy
data monitoring and visualization. Data visualization provided comparison among
© ASCE
Sustainable Human–Building Ecosystems 196
the different campus buildings, and as a result, it led to a more energy conscious
behavior from students, faculty members, staff, and most importantly better
understanding of the actual energy performance of building assets, thus improved
the operation process and promoted better ways to manage resources. The
emerging need by universities for clear data recently used a commercial data
visualization tool such as the Building Dashboard by Lucid Design Group
(Oakland, California), Energy Efficient Education Dashboard by Quality
Automation Graphics (Ankeny, Iowa), or conducted studies and developed their
own data analysis and visualization interface tools. The following (Table 1) is a
summary of selected studies relevant to this research:
(Emeakaroha, 2014)
(Domínguez, 2013)
(Yarbrough, 2015)
(Gulbinas, 2014)
(Malkawi, 2007)
(Dong, 2014)
(Jain, 2012)
Actual energy data ● ● ● ● ● ●
Simulated energy data ● ●
Peak energy analysis
● ●
and optimization
Calibrate energy models ●
Predicting future
● ●
consumption
Adjust the performance
of building systems/
● ● ● ●
Operations & facility
Management
Detect power faults and
●
deviations
Energy/economic
● ● ● ● ● ●
savings
Data
visualization/statistical ● ● ● ● ●
analysis
Online tool ● ●
Visualized data
availability for ● ● ●
occupants
Energy savings
possibility by
● ● ●
influencing occupants’
behavior
© ASCE
Sustainable Human–Building Ecosystems 197
FR
RAMEWO
ORK
Thhis paper explains
e thee initial worrk on the developmen
d nt of a dataa visualization
daashboard thhat can bee of assistaance to booth campus operators and camppus
coommunity. The structuural framew work of the data visuallization boaard is dividded
innto three coomponents, the monitooring system componeent, the daata processing
coomponent, and a the dataa visualization component (see Fig gure 1).
Inn the first coomponent “Layer 1: Monitoring”,
M the data is collected frrom two maain
soources, the first sourcce is the university’s
u energy monitoring
m c
center, whiich
m
monitors thee electricityy consumptiion data onn the parceels level allong with tthe
ennergy used forf the chillled water inn the HVAC C systems. AsA for the water
w data, itt is
coollected man nually by thhe universitty facilities office whicch providess daily data of
w
water consum mption on thet campus and buildinng levels. In the secon nd componeent
“LLayer 2: Daata Processsing” the daata sorting and filterin ng and repoort generation
taakes place. InI the third componentt “Layer 3: Visualizatioon”, the datta is uploadded
onn the web online
o throuugh a data importer
i annd is compoosed of the interface anda
daata importerr.
Figu
ure 1. Fram
mework dessign of the data visuallization pro
ocess
Thhe raw dataa is presentted in user friendly deesign with comparisonn ability. TheT
tyype of data visualized
v o the website are simpple bar graphhs, pie charrts and gaugges
on
thhat contain energy annd water daata for each building or cluster of buildinngs
“pparcel”. Thee electrical energy connsumption isi presented d in kilowattt hour (kW
Wh)
annd the wateer is presennted in liteers. The tottal energy ofo the whoole campus is
caalculated ussing the colllected enerrgy data. Foor better beenchmarkin
ng, the enerrgy
usse intensity (EUI) is pllaced at thee main hommepage of th he website referenced
r by
yeear, where the
t EUI is thet total energy consum mption of a building divided
d by the
t
Wh/m2). Thhis kind of information
tootal area (kW i n gives the user
u the abiility to realiize
caampus energgy consumpption compaared to otheer campuses and buildinngs.
G
GRAPHICA
AL USER INTERFAC
I CE (GUI) DESIGN
D
Thhe user intterface for the system
m was devveloped usin ng the Sym mphony PH HP
framework, combined
c w the usee of the Boootstrap CSS
with S, jQuery JaavaScript, and
a
G
Google Charrts librariess. The userr-facing weebsite was divided intto three maain
seections: Ho
ome, Energgy Breakdoown, and Energy Map; M in adddition to tthe
© ASCE
Sustainable Human–Building Ecosystems 198
Figure
F 2. Bar w averagee consumption display
B chart with y
The homepage
h aalso contain
ns a control that is com
mmon acrosss most pagges
off the websitte, showingg the currentt month’s usage
u for a single buildding, for eaach
off the energyy types. Thhis data is displayed
d a a color-ccoded dial gauge,
as g withh a
seelector that can be usedd to switch between different builldings, or piick a building
att random. Every
E time a page is displayed to t the userr, a differennt building is
seelected at raandom and its
i data is displayed in this controll. The energ
gy breakdow wn
© ASCE
Sustainable Human–Building Ecosystems 199
paage contain ns a similar dial gaugee control foor each buillding, with an additionnal
abbility to sellect which month
m to display
d fromm the data. This allowss the users to
coompare the historic ussage data forf this buiilding acrosss differentt months (ssee
Fiigure 4).
In adddition to thhe dial gauuge control, the energgy breakdow wn page allso
coontains a bar chart shhowing the energy connsumption of the seleected building
accross all thee months off a selectedd year, withh the option
n for the ussers to chooose
w
which year to t display. This
T gives the users ana overall view
v of the trends in the
t
buuildings’ ussage for eacch energy ty ype over thhe span of a year, and the ability to
coompare betw ween these trends
t acrosss different years (see Figure
F 5).
© ASCE
Sustainable Human–Building Ecosystems 200
Figurre 6. Cumu
ulative enerrgy consum
mption
Fiigure 7. Bu
uilding by building
b eneergy use in
ntensity (EU
UI)
The website
w inttegrated muultiple techhnologies too reach a high
h level of
fleexibility annd usability, clad in a modern useer interfacee design. Thhe Symphony
PH HP framew work was ussed as a baasis for the code, whicch offers a modular and a
m
maintainable way to devvelop the website
w efficciently and allow
a easy modificatioons
inn the future when needded. The in nterface waas implemennted using the Bootstrrap
CSS framew work, which offers flexxible and responsive usser interfacee capabilitiees,
cooupled withh the jQueery JavaScrript library,, which offfers easy accessa to tthe
poowerful funnctionality ofo JavaScriipt. The chaarts displayyed by the website weere
geenerated ussing the Gooogle Chartts (gChartss) chart librrary, which h enabled the
t
w
website to display
d cleaan-looking and dynam mic charts that allow the users to
innteract withh the data as
a they brow wse throughh the webssite and vieew the enerrgy
coonsumptionn at differentt times or fo
or different energy typees.
A
ASSESSME
ENT
© ASCE
Sustainable Human–Building Ecosystems 201
As seeen in (seee Figure 8 above), moost users foound the chhart easier to
innterpret, unllike the spreead sheet. The
T weighteed average level
l of undderstandingg is
9 for the charrt and 6 for the tabulateed data.
© ASCE
Sustainable Human–Building Ecosystems 202
F
Figure 10. Buildings
B haaving an ab
bove average energy consumptio
c on accordin
ng
to users’ choiices
CONCLUSIION
In thiis paper, wee have demmonstrated a new energgy visualizaation tool thhat
caan serve as educationaal tool for the
t campus communityy and potenntially impaact
beehavior by promoting better wayys to learn the energy y performannce issues on
caampus. Th he early feedback
fe frrom the campus
c ommunity support daata
co
viisualization. The futurre developm ment of thhe dashboarrd will be to includee a
caampus operaation decisiion support interface thhat integratee simulated building
b daata,
acctual and prredicted dataa per academmic year as shown in (see Figure 11). The errror
inn predictionn should be b minimizzed in the future as a result of o continuoous
reefinement off the simulaated modelss of campus buildings.
R
REFERENC
CES
M
Malkawi, A.,, and Srinivvasan, R. (20007). “Enerrgy based deecision supp
port system
for faacilities mannagement: inntegration of
o data/web mining, kn nowledge baase
© ASCE
Sustainable Human–Building Ecosystems 203
© ASCE
Sustainable Human–Building Ecosystems 204
Abstract
The production of sustainable housing requires the cooperation of a variety of
participants with different goals, needs, levels of commitment and cultures. To
achieve mainstream net zero energy housing objectives, there is arguably a need for a
non-linear network of collaboration between all the stakeholders. In order to create
and improve such collaborative networks between stakeholders, we first need to map
stakeholders’ relationships, processes, and practices. This paper discusses compares
and contrasts maps of the sustainable housing production life-cycle in Australia,
developed from different perspectives. The paper highlights the strengths and
weaknesses of each visualisation, clarifying where gaps in connectivity exist within
existing industry networks. Understanding these gaps will help researchers and
practitioners identify how to improve the collaboration between participants in the
housing industry. This in turn may improve decision making across all stakeholder
groups, leading to mainstream implementation of sustainability into the housing
industry.
INTRODUCTION
Globally the key barriers inhibiting the residential construction industry’s
ability to implement sustainability requirements include technical, economic and
regulatory issues, the culture and practices of the industry, the lack of feedback loops
and poor levels of stakeholder engagement and communication (Holloway and
Bunker 2006; Williams and Dair 2006; Osmani and O'Reilly 2009). The realization
of efficient sustainable housing arguably requires the integration of functionality,
cultural sensitivity and local climatic conditions into a long-lived product (Larsson
2004), enabled by the implementation of sustainability practices throughout the
product life-cycle, from green field development and infrastructure provision,
through design, manufacture and construction, to ultimate deconstruction, re-cycling
or demolition (Romero et al. 2009; Miller and Buys 2012). This life cycle involves
© ASCE
Sustainable Human–Building Ecosystems 205
© ASCE
Sustainable Human–Building Ecosystems 206
EVIEW OF
RE F AUSTRAL
LIAN HOU
USING NE
ETWORK MAPS
M
Thiis section prresents threee housing industry
i rellationship maps,
m from the perspecctive
of building innformation flows, connstruction industry reelationships and end-uuser
expperiences.
Maap1: Buildin ng Informaation Flowss and Stakeeholder Reelationshipss. This netw work
mapp (Figure 1)) depicts cuurrent comm mon practicees relating tot the creatiion and floww of
infoormation ab bout an inddividual dw welling from m site deveelopment thhrough to first
resaale to a sub T stages oof the housse production process are
bsequent puurchaser. The
represented onn the x axis (e.g. site developmeent, land saale, planning, construcction
etc)), whilst thee key playerrs involved in the proceesses are prresented in thet y axis. The
diaggram attem mpts to preesent, for eache stage of develop pment, which stakehoolder
creaates informmation abouut that speciific dwellinng, and to whom w this information is
passsed. This mapping
m shoows that noot all informmation abouut a dwellin ng is passedd on
throough the suupply chainn, arguablyy impactingg on decisio on making of subsequuent
stakkeholders. ItI identifies four key reelationships that impact on inform mation flowss: (i)
devveloper / in nfrastructuree provider (circular
( doot-dash linees); (ii) salees / valuatioon /
gulation / inndustry (squuare dotted lines); and (iv)
finaance (ellipse dotted linnes); (iii) reg
inittial purchaseer / subsequuent purchassers (ellipsee dot-dash liines).
© ASCE
Sustainable Human–Building Ecosystems 207
Figure 2.
2 Life Cycle Stages of New and Existing Su
ustainable Houses ma
ap.
© ASCE
Sustainable Human–Building Ecosystems 208
© ASCE
Sustainable Human–Building Ecosystems 209
of information exchange between stakeholders within the construction life cycle (e.g.
between the real-estate market and designer or occupant and builder).
Three key gaps are revealed in these industry-focused maps: (i) where
information flows (or should flow); (ii) the direction of information flows; and (iii)
the (lack of) active involvement participation of end-users. Map 1 (Figure 1) revealed
that information created about a dwelling at different stages of its production is often
not passed on from one set of stakeholders to subsequent stakeholders. For example,
very little information is passed from the developer / infrastructure provider
relationship to other parties, while no information at all is passed from the sales /
valuation / finance relationship to the designer. Another example is the little
information passed to the subsequent owner, whose only source of information is the
sales agent who is likely not aware of most of the information. This lack of
information exchange leads to un-informed decision making or additional costs in the
re-creation of the information. One way to overcome this gap is to create a ‘building
file’ which holds all data sets that could be utilised by different members throughout
the building life-cycle. The building file concept is already being developed in
several EU-countries. It provides sets of information related to a particular dwelling,
such as description of the construction’s materials and elements, legal information,
condition report, quality reference, maintenance guide and energy label. These sets
of information support the buyers, occupants and policymakers in their decision
making process (Van der Bos and Meijer 2005). The building file concept could be
expanded to include sets of information that are distributed to all stake holders since
the inception of the project, using a variety of information generation, storage and
collaboration tools, such as building information model, energy simulation software,
building diagnostic tools, and post occupancy evaluation. Such a building file acts as
a storage hub that stores all the building data for future reference, and is updated
throughout the building life cycle to enhance transparency, so that the impact of
every decision is shown to all members.
Map 2 depicts many of the production stages as uni-directional (e.g. transition
from design to consultant to construction to marketing to new home buyer). A
bidirectional exchange of information would arguably enable each stage to provide
feedback loops to previous stages to enhance efficient and sustainable construction
outcomes. Robust building documentation systems and feedback loops, combined
with a ‘living lab’ approach, are needed in order to achieve the user-centered
aspirational view depicted in Map 3. The living lab involves the user in the creation,
exploration, experimentation and evaluation of innovations through active
involvement in all stages, including post-occupancy feedback (Romero, et al.,
2009).The living lab could be the link between maps 1, 2 and 3 in integrating
© ASCE
Sustainable Human–Building Ecosystems 210
reseearch, user participatiion and thee on-going cycle of innformation flow betw
ween
reseearch and practice, occcupant and the
t housing industry.
TO
OWARDS A COLLAB
BORATIVE
E NETWO
ORK
Map 5 (Figure 5)) is a theorretical reprresentation of how gaaps in comm
mon
praactice miight be filled,
enhhancing thee bidirectional flow
of inform
mation between
stakkeholders during
d the different
houusing construction lifecycle
phaases througgh a netw work that
direectly conneect each staakeholder
to all the oth her stakehoolders. In
thiss map, eache key member
(inccluding e
energy e
efficiency
evaaluator that was added as a key
mem mber in the connstruction
phaase), has access to an
“infformation hub”. Thhis hub
© ASCE
Sustainable Human–Building Ecosystems 211
© ASCE
Sustainable Human–Building Ecosystems 212
CONCLUSION
The paper examined two maps that represented existing common practice and
stakeholder relationships in the Australian housing industry and a third map, centered
on the end-users’ perspectives. From the analysis of these maps it became evident
that there are gaps in the information network that connect the stakeholders to each
other throughout the housing production life-cycle. These gaps could impact the
sustainable housing outcomes. The paper then represented a map which reflects the
need to treat sustainable housing as an integrated system, where collaborative
processes are made to reach solutions that maximize sustainable housing outcomes.
A further development of this paper could be the integration of maps three and five,
where the information from each member’s related “information hub” in Map 5 is
used to make informed decisions that contribute to the fulfilment of a centeral goal
which is the sustainable house aspirations of the occupants.
REFERENCES
Cheung, Yan Ki Fiona and Steve Rowlinson. 2011. "Supply chain sustainability: a
relationship management approach." International journal of managing projects in
business 4 (3): 480-497.
Epstein, Marc J. and Sally K. Widener. 2011. "Facilitating sustainable development
decisions: measuring stakeholder reactions." Business strategy and the environment
20 (2): 107-123.
© ASCE
Sustainable Human–Building Ecosystems 213
Heugens, P. P. M. A. R., van den F. A. J. Bosch and van C. B. M. Riel. 2002. "Stakeholder
integration." Business & Society: a journal of interdisciplinary exploration 41 (1):
36.
Holloway, Darren and Raymond Bunker. 2006. "Practice Reviews: Planning, Housing and
Energy Use: A Review." Urban Policy and Research 24 (1): 115-126.
Larsson, Nils. 2004. "The Integrated Design Process." Accessed 14 December 2010.
http://www.iisbe.org/down/gbc2005/Other_presentations/IDP_overview.pdf.
Miller, Wendy. 2011. "Greening Australia's residential built environment: Part A":
SBENRC.
Miller, Wendy and Laurie Buys. 2012. "Positive Energy Homes: Impacts on, and
Implications for, Ecologically Sustainable Urban Design." Urban Design
International 17 (1): 45-61.
Miller, Wendy and Laurie Buys. 2013. "Factors influencing sustainability outcomes of
housing in subtropical Australia." Smart and Sustainable Built Environment 2 (1):
60.
Osmani, Mohamed and Alistair O'Reilly. 2009. "Feasibility of zero carbon homes in
England by 2016: A house builder's perspective." Building and Environment 4 (9):
1917-1924.
Rohracher, Harald. 2001. "Managing the Technological Transition to Sustainable
Construction of Buildings: A Socio-Technical Perspective." Technology Analysis &
Strategic Management 13 (1): 137-150. Accessed 2015/05/22.
Romero, David, Myrna Flores, Carlos Vallejo and Arturo Molina. 2009. "Towards a Novel
Living Lab Model for Sustainable Innovation in the Construction Industry." In 15th
International Conference on Concurrent Enterprising (ICE 2009), Leiden, The
Netherlands, 22-24 June 2009, edited.
Tzortzatou, Eleni Penelope. 2007. "Culture impact in construction supply chain
management." Dissertation/Thesis, ProQuest, UMI Dissertations Publishing.
Williams, Kate and Carol Dair. 2006. "What is stopping sustainable building in England?
Barriers experience by stakeholders in delivering sustainable developments."
Sustainable Development 15 (3): 135-147.
© ASCE
Sustainable Human–Building Ecosystems 214
J. N. Terman1
1
George Mason University, School of Policy, Government and International Affairs,
4400 University Dr., Robinson B, 3F4, Fairfax, VA 22030. E-mail:
Jterman@gmu.edu
Abstract
INTRODUCTION
Countries all over the world have been increasingly sensitive to their energy
usage and costs. Building energy costs account for 41% of total energy consumption
costs (DOE, 2010). One way to reduce these costs is by engaging in energy retrofits
and weatherization to ensure building stocks are using energy as efficiently as
possible (Dietz, et al., 2009). This is particularly important for low-income
households, who dedicate a larger proportion of their incomes to energy costs
(ORNL, 2010). Thus, energy efficient housing is not only an energy usage issue of
energy efficiency, but it is also a social policy concern. The Weatherization
Assistance Program (WAP) demonstrates this duality of purpose. The WAP is a
federally funded, state implemented program designed to reduce the energy
consumption costs of low-income households through energy audits and retrofits.
While seemingly straightforward, some states view the program as a social services
program aimed at the welfare of the target population. Other states view it as a
vehicle for primarily energy efficiency (EE) or economic development. This study
tests whether these program framing choices influence WAP performance.
© ASCE
Sustainable Human–Building Ecosystems 215
© ASCE
Sustainable Human–Building Ecosystems 216
Governments make choices about how they frame their policy responsibilities
(Krause, Feiock, and Hawkins, 2014). These choices have substantive consequences
in that an important causal mechanism for program performance is the
implementation structures that are shaped by these choices. For example, agencies
that have a substantive focus on economic development will be more likely to
prioritize the programs that have heightened results in terms of economic
development.
Applying these ideas to the WAP, the question becomes whether state
governments frame the program in terms of social policy? This is important because,
as an EE-oriented policy, the return on investment (ROI) for the program is fairly
low. While programs within DOE’s Energy Efficiency and Renewable Energy
(EERE) Office portfolio have historically had a 20 to 1 benefit to cost ratio, WAP has
a benefit to cost ratio of approximately 1.53 to 1 (DOE, 2009a). In the EE policy
world, this ROI helps determine program and policy prioritization. Programs with
lower ROIs become less of a priority. Thus, agencies focused on energy issues may
be less likely to ambitiously implement low-ROI programs such as the WAP.
ROI in social programs can be defined more broadly because the program
intervention has goals that are often more focused on societal value sets. For
example, the American Public Human Services Association (APHS, 2013) has
developed the idea of a social return on investment (SROI). SROI is calculated as a
blended value of the financial and social benefits of particular policies and programs.
These broader social goals are more likely to be taken into consideration and valued
by agencies focused on social services and administering to vulnerable populations –
in comparison to agencies focused on energy or economic development. Thus, the
expectation is that, when the WAP is nested in agencies that have a social service
orientation, program performance will be higher (H1).
Agency assignment is not the only way that states can frame their policies;
they can also engage in staff assignment. They may dedicate staff with particular
backgrounds to administer programs. For example, governments may choose WAP
staff with expertise in energy efficient housing stocks that would see the program in a
different light than those with a background in social services or anti-poverty. In the
WAP, governments can engage in staff assignment by having the program directed
by administrators involved in the delivery of other social services within those
agencies. One way that they do this is by leveraging their administrators from the
Community Services and Block Grant Program (CSBG). CSBG is touted as one of
the only federal programs that is exclusively dedicated to the reduction of poverty
(NASCP, 2012). The professional network of administrators associated with CSBG
programs is robust and well-coordinated with active lobbying and training arms
(NASCP, 2009).
States have the opportunity to use CSBG trained administrators to implement
the WAP. This is the case even if these administrators are not in social service-
oriented agencies. When CSBG administrators direct the WAP, these individuals will
be more focused on the social service aspect of the program – rather than EE –
thereby heightening program performance. The expectation is that, when the WAP is
directed by CSBG administrators, program performance will be higher (H2).
© ASCE
Sustainable Human–Building Ecosystems 217
The larger causal mechanism behind agency and staff assignment is the
substantive prioritization of particular policy goals (and the programs associated with
those policy goals). One indicator of this prioritization is the ancillary support
associated with the program. As a social program, an overarching goal of the WAP is
to not only reduce energy costs but to also train the hard-to-employ with
weatherization skills that are then utilized to provide program weatherization services
(OWIP, 2009). To complete this training, states were mandated to spend 20% (up
from 10%) of their ARRA-allocated WAP dollars on training and technical assistance
(T&TA). However, this training comes in different forms and strengths.
T&TA centers are considered one of the most vigorous training methods
available for building the weatherization workforce of the hard-to-employ (OWIP,
2009). They are specifically intended to train the hard-to-employ in weatherization
centers that consist of classrooms and laboratory space set up for long-term training.
DOE documentation suggests that training centers are the most robust method
through which states can meet their weatherization goals because they are more likely
to reach the hard-to-employ. Attendance can be mandated and training recipients can
be held accountable, tracked and supported in terms of job placement (OWIP, 2009).
While T&TA centers are an extremely robust way to train the hard-to-employ,
there are other less permanent educational methods for training. States may also
choose to invest T&TA dollars in training WAP enrollees, as opposed to curriculum
and website development, facility preparation and overhead. While the latter are
important investments, they do not go directly to training the hard-to-employ. These
investments are indicative of the broader view that the WAP is a social-services
oriented program with a broader scope to reduce the sources of poverty – not only
high energy costs, but also unemployment and lack of job skills. As such, the
expectation is that, when states choose to invest T&TA dollars on activities that focus
on training the hard-to-employ, as opposed to support services, program performance
will be higher (H3).
RESEARCH DESIGN
The dependent variable in the study comes from DOE and State Energy
Program (SEP) offices, which are responsible for maintaining records on the WAP’s
performance goals (and achievement of those goals). As a condition of receiving
ARRA dollars, states had to set performance goals – the number of weatherizations
planned – with their service providers. Performance goal achievement is measured
here as the proportion of weatherizations planned which were completed.
The independent variables of interest include (1) whether the WAP is
administered through an agency that is centered on social services, (2) whether the
WAP is directed by CSBG administrators, (3) whether the state has chosen to invest
in one or more T&TA centers and (4) the proportion of T&TA dollars that are
allocated to direct training of WAP enrollees. The former was operationalized by
examining the state agencies administering the WAP. There are three types of
agencies that administer the WAP: social services-oriented agencies, economic-
development focused agencies and energy focused agencies. Since agency
assignment varies so much across states, observations were coded by whether the
© ASCE
Sustainable Human–Building Ecosystems 218
WAP was part of an agency whose target population is persons in poverty. If the
WAP was outside of the primary agency assigned to service vulnerable populations,
then the agency was not coded as a social services-oriented. The data for the second
independent variable of interest was obtained from the 2009 CSBG annual report
while the source for the third and fourth variables of interest were obtained from
Department of Energy performance reports for T&TA resource use.
Service area characteristics and state partisanship, size, resource availability
and professionalism were controlled for in the analysis. State characteristics were
particularly important because of the potential for them to influence investment a
T&TA center and propensity for goal achievement. Table 1 includes a complete
listing of the control variables, operationalizations and descriptive statistics.
The unit of analysis in this study is the WAP service area; service areas are
nested within states. Thus, a hierarchical linear model is employed. After
considerable experimentation with model specification, the decision was made to
estimate a model with a random intercept and random slopes for T&TA investments
and agency assignment. When modeled as random slopes, the random effects for staff
assignment and the proportion of WAP training dollars allocated to direct training
were not statistically different from zero and were removed from the model
specification.
RESULTS
© ASCE
Sustainable Human–Building Ecosystems 219
© ASCE
Sustainable Human–Building Ecosystems 220
The study has yielded interesting findings. First, the differing effects between
the types of agencies administering the WAP are considerable. The essential
difference between social services and economic development-oriented agencies and
those focused on energy is that the former types deal with social problems. This is
fundamentally different than energy-oriented agencies, which often focus on natural
resource issues. The infrastructure and skill sets required to deal with the issue of
poverty are fundamentally different, and will lead to performance differences.
Table 2: Performance Goal Achievement
β SE
Social Services Orientation
Social services-oriented agency 0.339*** 0.096
Economic development-oriented agency 0.325*** 0.093
WAP directed by CSBG administrators 0.091 0.098
T&TA center 0.184** 0.086
WAP job training activities 0.034 0.1
Service-Area Characteristics
% Eligible -0.01*** 0.004
Housing density 0.007 0.007
WAP allocation
For-profit provider -0.213** 0.109
Government provider 0.123*** 0.42
State Characteristics
Democratic governor 0.067 0.082
% Legislative Democrats -0.009* 0.005
Population density -0.003 0.003
Per capita income < 0.00 < 0.00
% Unemployment -0.012 0.02
Government Performance Project score -.048* 0.025
Random Effects
Social services-oriented agency 0.001 < 0.00
Economic development oriented agency 0.261 0.092
T&TA center 0.23 0.081
Constant 1.03 0.444
Service area effects 0.456 0.011
Service areas/State Service areas = 962 Wald χ2 (16) = 65.23***
© ASCE
Sustainable Human–Building Ecosystems 221
The results of having at least one T&TA center are more intuitive. With
investment in stand-alone T&TA centers, there may a greater commitment to the
broader program purpose of poverty reduction. However, this result may be tenuous
because there may be other reasons for investing in a T&TA center in a given state.
For example, some states may choose not to invest in a T&TA center because a
neighboring state has one (i.e. Texas and Arizona do not have a T&TA center but
New Mexico does). Although, T&TA centers are intended for training the hard-to-
employ from inside the state – unless they are regional centers – states may send
individuals across state lines for training.
The results here are relatively persuasive. Service areas in states whose WAPs
are served by social services or economic-development-oriented agencies, on
average, have one-third higher goal achievement. Services areas in states with T&TA
centers have nearly one-fifth higher performance goal achievement. Nonetheless,
these results highlight some important challenges to the development and
implementation of energy efficiency and sustainability programs. These programs
have environmentally collective benefits but depend on a concentrated population
that is traditionally difficult to administer to. Appropriate program design requires
expertise in both energy policy and social services. The development of
weatherization techniques and training – by experts in the energy and building
industry – is essential to the services provided by the program. However, program
services and overall success cannot occur without considerable knowledge of the
target population, to whom and by whom services are rendered. Perhaps, if energy-
oriented agencies are involved in higher ROI energy programs, it makes sense for
social services-oriented agencies to have the responsibility for program
implementation so that the former can use their expertise to focus on programs with
broader target populations and higher return. However, in the context of federally
funded, locally implemented programs, lower level governments make these choices,
which is what accounts for this variation.
Future research can build on these ideas and address some of the limitations in
this study. The operationalizations used here are somewhat compelling in terms of
viewing the WAP as a social program, however, they need to be broadened and the
causal mechanism behind them needs to be better understood. For example, what is it
about the agency program implementation structures that affect performance? Is this
an issue of heightened staff motivation to achieve program goals? With a low ROI in
energy-related program such as the WAP, some agencies are simply not incentivized
to dedicate significant resources toward its implementation.
REFERENCES
© ASCE
Sustainable Human–Building Ecosystems 222
© ASCE
Sustainable Human–Building Ecosystems 223
Abstract
INTRODUCTION
© ASCE
Sustainable Human–Building Ecosystems 224
© ASCE
Sustainable Human–Building Ecosystems 225
METHODOLOGY
© ASCE
Sustainable Human–Building Ecosystems 226
distribution of solutions in the Pareto Optimal Set provides more distinct choices for
designers. In NSGA-II, existing diversity calculation is based on crowding (Deb et al.
2002). In our proposed research, for multiple objectives, building thermal and
daylighting performances will be used as sample quantifiable objectives and the
architectural aesthetics concerning building forms will be used as a sample non-
quantifiable objective. The thermal and daylighting performances will be evaluated
by simulation tools and the aesthetic form needs to be evaluated by architects through
visual inspection.
Form-diversity will help architects focus on the drastically diverse forms to
choose for further design development. In ArchMOO, building form’s shape distance
will be calculated and will work with crowding distances for diversity ranking. Shape
distance calculation will be based on 3D Shape Similarity assessment algorithms
(Shum et al. 1996; Cardone et al. 2003; Bustos et al. 2005; Tangelder & Veltkamp
2008), which compare the 3D shape similarity between closed surfaces. In our
project, since different building design options are generated using parametric
modeling, the causes of the shape changes – parameter and the model graph changes
are known to our system. This information will be used to improve the performance
of 3D shape similarity assessment algorithms. Figure 1 illustrates ideas of Pareto
Optimal solutions with diverse building forms.
Figure 1. ArchMOO ideas illustrated. Left: Pareto Optimal solutions with diverse
building forms (shown as building masses) but similar building forms eliminated
(shown as dark dots); Right: four sample diverse building forms in the Pareto
Optimal solutions displayed in BIM for architects’ inspection and decision-making.
© ASCE
Sustainable Human–Building Ecosystems 227
Fig
gure 2. Priorr work on BIM
B and Clo
oud-based peerformance optimizationn (Rahmani
Asll et al. 2013 and 2014 4). Left: BIM
M of a testt building. Middle: 2DD graph for
parrametric BIMM. Right: Cloud-based
C thermal annd daylightiing simulatiion runs in
Greeen Buildingg Studio. Th
he project op
ptimizes twoo windows’ design for tthermal and
day
ylighting perrformances.
Fig
gure 3. Optiimo version 0.1.2 graph h for the Scchaffer test ffunction (left), and the
gen
nerated Paretto Optimal set
s (right) (R
Rahmani Asl et al. 2015)..
CO
ONCLUSIONS AND FU
UTURE WO
ORK
© ASCE
Sustainable Human–Building Ecosystems 228
REFERENCES
Altabtabai, J. and Yan, W. (2015). "A User Interface for Parametric Architectural
Design Reviews." Proceedings of the 20th Conference of the Association for
Computer-Aided Architectural Design Research in Asia (CAADRIA).
Kyungpook National University, Daegu, Republic of Korea.
ANSI/ASHRAE (2007). Standard Method of Test for the Evaluation of Building
Energy Analysis Computer Programs.
Bazjanac, V. (2001). "Acquisition of building geometry in the simulation of energy
performance". Building Simulation, 305–311. Rio de Janeiro, Brazil, August
13.
Bradner E, Davis M. (2013). "Design creativity: using Pareto analysis and genetic
algorithms to generate and evaluate design alternatives." Available at:
http://ecologylab.net/workshops/creativity/papers/bradner.pdf
Caldas, L. (2006). "GENE_ARCH: an evolution-based generative design system for
sustainable architecture." Intelligent Computing in Engineering and Architecture.
Springer Berlin Heidelberg. pp.109-118.
© ASCE
Sustainable Human–Building Ecosystems 229
Cardone, A., R. K. Gupta, and M. Karnik (2003). "A survey of shape similarity
assessment algorithms for product design and manufacturing applications."
Journal of Computing and Information Science in Engineering.
Coello, C. A. Coello, G. B. Lamont, and D. Van Veldhuisen (2007). Evolutionary
algorithms for solving multi-objective problems, Springer.
Deb, Kalyanmoy, et al. (2002). "A fast and elitist multiobjective genetic algorithm:
NSGA-II. Evolutionary Computation." IEEE Transactions on 6.2 (2002):
182-197.
Gerber, D. J., Lin, S. H. E., Pan, B. P., & Solmaz, A. S. (2012). "Design optioneering:
Multi-disciplinary design optimization through parameterization, domain
integration and automation of a genetic algorithm." Proceedings of the 2012
Symposium on Simulation for Architecture and Urban Design (p. 11). Society
for Computer Simulation International.
Holm, I. (2006). Ideas and Beliefs in Architecture and Industrial design: How
attitudes, orientations, and underlying assumptions shape the built
environment. Oslo School of Architecture and Design. ISBN 82-547-0174-1
Kellert, S. R., J. Heerwagen, and M. Mador (2011). Biophilic design: the theory,
science and practice of bringing buildings to life. John Wiley & Sons.
Konak, A., D. W. Coit, and A. E. Smith (2006). "Multi-objective optimization using genetic
algorithms: A tutorial." Reliability Engineering & System Safety 91.9 (2006): 992-
1007.
LaHood, S., and M. Vanden Brink (2009). "Aesthetics and new product
development." Evidence-based Design for Healthcare Facilities (2009): 19.
Lotov, A. V., Bushenkov, V. A., & Kamenev, G. K. (2004). Interactive decision
maps: Approximation and visualization of Pareto frontier (Vol. 89). Springer.
Pollio, V. (1914). Vitruvius: The Ten Books on Architecture (Vol. 645). Courier
Dover Publications.
Radford, A. D., and J. S. Gero (1987). Design by optimization in architecture,
building, and construction. John Wiley & Sons, Inc.
Rahmani Asl, M., Zarrinmehr, S., and Yan, W. (2013). "Towards BIM-based
Parametric Building Energy Performance Optimization." Proceedings of The
Association for Computer Aided Design in Architecture (ACADIA), October
24-27, 2013. Cambridge, Ontario, Canada.
Rahmani Asl, M., Bergin, M., Menter, A., and Yan, W. (2014). "BIM-based
Parametric Building Energy Performance Multi-Objective Optimization."
Proceedings of the Conference of Education and Research in Computer Aided
Architectural Design in Europe (eCAADe), Northumbria University,
Newcastle Upon Tyne, UK.
Rahmani Asl, M., Stoupine, A., Zarrinmehr, S., and Yan, W. (2015). "Optimo: A
BIM-based Multi-Objective Optimization Tool Utilizing Visual Programing
for High Performance Building Design." Proceedings of the Conference of
Education and Research in Computer Aided Architectural Design in Europe
(eCAADe), Vienna, Austria.
Shum, H., M. Hebert, and K. Ikeuchi, (1996). "On 3D shape similarity." Proceedings
of Computer Vision and Pattern Recognition, CVPR'96, IEEE Computer
Society Conference on. IEEE.
© ASCE
Sustainable Human–Building Ecosystems 230
Su, Z., Yan, W. (2014). "Improving Genetic Algorithm for Design Optimization
Using Architectural Domain Knowledge." Proceedings of the Annual
Conference of the Association for Computer Aided Design in Architecture
(ACADIA), October 23-25, 2014. Los Angeles, California.
Su, Z. and Yan, W. (2015). "A Fast Genetic Algorithm for Solving Architectural
Design Optimization Problems." Artificial Intelligence for Engineering
Design, Analysis and Manufacturing (AIEDAM), Cambridge University Press.
Accepted.
Tangelder, J. W., & Veltkamp, R. C. (2008). "A survey of content based 3D shape
retrieval methods." Multimedia tools and applications, 39(3), 441-471.
Torcellini, P., Pless, S., Deru, M., Griffith, B., Long, N., & Judkoff, R. (2006).
Lessons learned from case studies of six high-performance buildings, National
Renewable Energy Laboratory.
Welle, B. M. Fischer, J. Haymaker, and V. Bazjanac (2012). "CAD-Centric
Attribution Methodology for Multidisciplinary Optimization (CAMMO):
Enabling Designers to Efficiently Formulate and Evaluate Large Design
Spaces." CIFE Technical Report #195, Center for Integrated Facility
Engineering, Stanford, CA.
© ASCE
Sustainable Human–Building Ecosystems 231
Abstract
INTRODUCTION
© ASCE
Sustainable Human–Building Ecosystems 232
policy adoptions by local governments. The next section explains extant theory and
research on green policy tools, measurements of social equity, and presents
hypotheses for their association. The third section details our data and analytic
approach. The fourth section presents results, and the concluding section speaks
briefly about extensions for this research.
Green Building Policy Tools. Green building policies are often pursued by
communities with the administrative and material means to do so. This presents a
green policy paradox of sorts. Communities which may be in greatest need of green
investment – lower-income, blighted and higher minority communities with
deteriorating infrastructure and services – should generally be less capable of
inducing development interest to make the investments necessary to spur green policy
tool utilization. In other words, cities with fewer resources and less development
demand for innovative building designs and energy efficiencies are generally less
© ASCE
Sustainable Human–Building Ecosystems 233
© ASCE
Sustainable Human–Building Ecosystems 234
These two wealth indicator propositons lead to the first interaction hypothesis:
Since more informaton on the long-term benefits would also alleviate such
cost concerns, communities with higher levels of education among its population
could also be more likely to adopt them.
Proposition 3: The marginal effect of community education levels will be
positively associated with the adoption of greater numbers of green building policy
tools when income inequality is at its lowest level.
Proposition 4: The marginal effect of community education levels will be
positively associated with the adoption of greater numbers of green building policy
tools when income inequality is at its highest level.
Public administration and political science scholars also have a long tradition
of considering how the strucutre of local government decision-making influences
policy choices (Carr 2015; Clingermayer and Feiock 2001). A central argument for
why the Council-Manager form of government may be more active in pushing to
adopt green-building policies is that they represent “win-win” actions which seek to
grow and green local economies, advancing the career objectives of city managers
which seek to establish track records of innovation as well as environmental
sustainability and growth. We therefore expect that local governments in which
managers hold more functional control over executive authority of the government
will be more actively pursuing such green-building policy tools.
Social Inclusion Policy Tools. Although equity has been considered one of the
“three E’s” of sustainability, the ramifications of urban sustainability for the urban
disadvantaged has received little attention from the broader research community
(Berry 2001; Curley 2010). While sustainability proponents cite the need for equity to
be considered along side other efficiency or environmental concerns, it is often
consiered only in certain contexts or subject to tradeoffs. Growing local governments
have been cogniscent of the need to provide workforce housing, advocates have
pushed for socal inclusionary land uses and housing support for the disabled, seniors,
the homeless, while urban reinvestment strategies often tout inclusionary efforts.
A predominant sociological approach holds that material resource dispersal
may be the most noticeable consequence of inequality, and this lack of resources
© ASCE
Sustainable Human–Building Ecosystems 235
© ASCE
Sustainable Human–Building Ecosystems 236
demographics and sustainability practices. The survey response rate was 25.4%, with
2,176 cities and counties responding. We restrict the responses to cities only, yielding
1,867 observations about adoption of 56 questions pertaining to adoption of
greenhouse-gas emissions, smart growth, energy efficiency, and social
inclusion/equity policy tools. Our measures of socio-economic, demographic and
governmental form are derived from the 2010 U.S. Census and the survey responses.
Standard practice in sustainability research to date has been to create single
additive indexes of all sustainability tools or to attempt to differentiate between types
of tools with distinctive political “marketplaces” of demanders and suppliers within
communities. In their 2013 study using the same ICMA survey, Opp and Saunders
(2012) utilized 84 “indicators” of environmental policy, sustainable development, and
social equity tools to create a single measure (the Opp-Suanders Sustainability
Practices Index, or OSSPI) which attempted to capture the three “E’s” within a single
weighted index. This index is composed of three underlying subindices which proport
to capture the willingness of local governments to support environmental, economic
development, and equity causes. Thus, the higher a city scored on an range of 0 to
100, the more “sustainable” it would be considered.
We adopt a different approach to the question of whether the “Three E’s” are
arbitrary classifications of policy choices. First, we use factor analysis to disentangle
five distinct types of green policy tools which tend to be lumped together:
Greenhouse gas reduction targets; energy efficiency; green building; land use; and
social inclusion. We then fit bootstrapped zero-inflated negative binomial models for
the green building and social inclusion indexes to test the conditional hypothesis that
concentrations of wealth, education, fiscal resources and form of government within
cities interact with regional inequality to influence their tool adoptions.
This analysis first compared the performance of these concepts to factor
scores which attempt to reduce the number of variables and capture the underlying
latent structures of the policy tool space. We utilized exploratory factor analysis
(EFA) to determine whether a multivariate strategy for variable reduction can account
for more of the covariance in policy tool adoption and more accurately identify the
underlying latent concepts of sustainable development and equity which are of
theoretical interest. The goal of factor analysis to identify latent variables which may
cause manifest variables to co-vary (Floyd and Widaman 1995). Because the vast
majority of the policy tool responses are not normally distributed, we utilized a
principal factor extraction method and orthogonal Varimax rotation, which identified
five factors through a scree test which had eigenvalues > 1 and minimal, if any,
crossloadings (Costello and Osborne 2005). The five retained factors roughly
correspond to the five functional sustainability policy categories included in the
survey: greenhouse-gas reduction goals; energy efficiency measures; land-use; green-
building; and social inclusion tools. This suggests there are likely underlying distinct
causal factors for their adoption. Rather than merging all tools, we examine the count
of polices adopted for the dimensions of interest: green-building and social-inclusion
tools.
© ASCE
Sustainable Human–Building Ecosystems 237
© ASCE
Sustainable Human–Building Ecosystems 238
We fit four models – two measuring the additive influence, and two
interaction models which account for regional inequality -- to gauge the drivers of
green building and social equity-related policy tool adoptions nationally. Tables 2 and
3 in the following section display the results.
Explanatory Variables
Median Age 2010 U.S. Census 1,844 38.25 6.23 21 70.9
% White 2010 U.S. Census 1,844 80.7 17.38 3.3 99.3
% W/Bachelor's or higher 2010 U.S. Census 1,633 27.97 16.11 3.3 86.1
Median home value 2010 U.S. Census 1,633 $221,016 $180,816 $28,200 $1,000,001
Homeownership 2010 U.S. Census 1,844 66.18 13.37 20.3 97.5
Median Income 210 U.S. Census 1,633 $66,549 $28,607 $23,690 $250,001
GINI Average U.S. Census 730 0.464 0.018 0.426 0.505
Per Capita Property tax U.S. Census 1,808 $428 $1,233 $0 $47,981
Manager ICMA Survey 1,867 0.574 0.495 0 1
Commission ICMA Survey 1,867 0.013 0.117 0 1
age*GINI U.S. Census 722 18 2.94 9.26 33.25
White*GINI U.S. Census 722 36.44 8.34 1.44 48.98
Education*GINI U.S. Census 627 15.84 8.39 2.04 41.62
Home Value*GINI U.S. Census 627 139974 101734 21929 504450
Income*GINI U.S. Census 627 37156 15430 12734 112627
Homeowner*GINI U.S. Census 722 32.45 6.73 10.86 47.87
Population (logged) 2010 U.S. Census 1,844 9.49 1.16 4.3 15.92
% Pop Change 2000, 2010 Censu 1,833 16.26 100.45 -71.5 4081
Pop per sq mile (logged) U.S. Census 1,844 7.29 0.937 0.742 10.2
RESULTS
Hypothesis 1: The marginal effect of wealth indicators on green building policy tool
adoption is positive at lower levels levels of metropolitan inequality; this effect
decreases as metro inequality increases and eventually becomes negative.
In the green-building model, we find evidence that median home values have
a positive direct influence on green building policy tool adoption but that this effect
lessens and actually turns negative when regional inequality is higher. The results
generally conform to our expectations that more affluent communities are more likely
© ASCE
Sustainable Human–Building Ecosystems 239
© ASCE
Sustainable Human–Building Ecosystems 240
demand are more likely to lead to their adoption. We now turn to the social inclusion
model.
© ASCE
Sustainable Human–Building Ecosystems 241
The largest positive driver of adoption in the model is the direct measurement
of income inequality, which suggests metro areas which experience higher degrees of
housing stratification based on income are more likely to generally adopt higher
numbers of inclusionary policies. This suggestion makes sense intuitively, but
without modeling inequality as a hierarchical fixed or random effect, we hesitate to
draw such a conclusion based of this analysis. In other words, we find evidence only
that larger cities with greater property-tax bases – cities which are also more likely to
have larger populations in need of these services -- are more likely to adopt larger
numbers of these policies.
CONCLUSIONS
We are only beginning to understand the role that social cohesion across city-
limits plays in the adoption of policies which benefit human building ecosystems.
Environmental protection, economic development and social equity are distinct terms
co-branded as “sustainability” in the academic literature and often treated as if they
are similar investments. But disentangling the “why?” of these separate dimensions of
governmental action requires unraveling the distinct political economies and drivers
© ASCE
Sustainable Human–Building Ecosystems 242
of differently weighted policy tool adoption, as well as the lnkages between cities
within metropolitian regions.
This paper offers one step in that goal by demonstrating first that the basket of
“sustainability” tools frequently lumped together with equal weight in extant
scholarship manifest from different objectives and should be modeled with great care
– and likely individually from one another. Green building tools are likely valued
differently across local governments, sectors of the economy and citizen subgroups
from socially inclusive policy goals even when they also promote sustainability.
This research has much room for advancement. Besides the causation
problem, there are likely omitted variables such as interest-group pressure which our
models for now leave out. Better theory and modeling of why region-wide social
disparities might influence intra-governmental decisions is needed. As a second
phase, we intend to conduct a comparative case analysis of these drivers as well as
gathering richer health-related outcome and building condition data across cities in
the Chicago and Orlando metropolitan regions.
BIBLIOGRAPHY
Berry, William D, Matt Golder, and Daniel Milton. 2012. "Improving Tests of
Theories Positing Interaction." Journal of Politics 1-19.
Bickers, Kenneth N, Lapo Salucci, and Robert M Stein. 2006. "Assessing the Micro-
Foundations of the Tiebout Model." Urban Affairs Review 42-57.
Boyle, Robin, and Rayman Mohamed. 2007. "State Growth Management, Smart
Growth and Urban Containment: A Review of the US and a Study of the
Heartland." Journal of Environmental Planning and Management 677-697.
Bromley, Daniel W. 1991. Environment and the Economy: Property Rights and
Public Policy. Cambridge, MA: Basil Blackwell.
Brown, Trevor L., and Matthew Potoski. 2003. "Transaction Costs and Insitutional
Explanations for Government Service Production Decisions." Journal of
Public Administration Research and Theory 441-468.
Carr, Jared B, and Antonio Tavares. 2014. "City Size and Political Particiaption in
Local Government: Reassessing the Contingent Effects of Residential
Location Decisions within Urban Regions ." Urban Affairs Review 269-302.
Carr, Jared. 2015. "What Have we learned about the Performance of Council-
Manager Governent: A Review and Synthesis of the Research Literature."
Public Administration Review Forthcoming.
Chapin, Timothy. 2012. "From Growth Controls, to Comprehensive Planning, to
Smart Growth: Planning’s Emerging Fourth Wave." Journal of the American
Planning Association 5-15.
Clingermayer, James, and Richard Feiock. 2001. Institutional Constraints and Policy
Choice: An Exploration of Local Governance. Albany: State University of
New York Press.
Curley, Alexandria M. 2010. "Relocating the Poor: Social Capital and Neighborhood
Resources." Journal of Urban Affairs 79-103.
deHaven-Smith, Lance. 2000. "Facing Up to the Political Realities of Growth
Management." Florida Planning 14-17.
© ASCE
Sustainable Human–Building Ecosystems 243
Downs, Anthony. 1999. "Some realities about sprawl and urban decline." Housing
Policy Debate 955-74.
Feiock, Richard C., Antonio F. Tavares, and Mark Lubell. 2008. "Policy Instrument
Choices for Growth Management and Land Use Regulation." The Policy
Studies Journal 461-480.
Finseraas, Henning. 2009. "Income Inequality and Demand for Redistribution: A
Multilevel Analysis of European Public Opinion." Scandinavian Political
Studies 94-119.
Fiorino, Daniel. 2006. The New Environmental Regulation. Cambridge: MIT Press.
Fischel, William A. 2006. The Tiebout Model at Fifty. Cambridge: The Lincoln
Institute of Land Policy.
Floyd, F.J., and K.F. Widaman. 1995. "Factor analysis in the development and
refinement of clinical assessment instruments." Psychological Asessment 286-
299.
Gerber, Elisabeth R., Adam Douglas Henry, and Mark and Lubell. 2010. "The
Political Logic of Local Collaboration in Regional Planning in California.
Paper 35." Duke Networks Conference. Accessed 1 30, 2014.
http://opensiuc.lib.siu.edu/pnconfs_2010/35.
Hardman, Anna, and Yannis M. Ioannides. 2004. "Neighbors’ Income Distribution:
Economic Segregation and Mixing in US Urban Neighborhoods." Journal of
Housing Economics 368–82.
Hawkins, Christopher V. 2011. "Smart Growth Policy Choice: A Resource
Dependency and Local Governance Explanation." Policy Studies Journal 679-
707.
Heyward, M. 2007. "Equity and international climate change negotiations: A matter
of perspective." Climate Policy 518–534.
Howell-Moroney, Michael. 2007. "Studying the Effects of the Intensity of the U.S.
State Growth Management Approaches on Land Development Outcomes."
Urban Studies 2163-2178.
Howell-Moroney, Michael. 2008. "The Tiebout Hypothesis 50 Years Later: Lessons
and Lingering Challenges for Metropolitan Governance in the 21st Century."
Public Administration Review 97-109.
Ihlanfeldt, Keith R, and Benjamin Scafidi. 2004. "Whites' neighbourhood racial
preferences and neighbourhood racial composition in the United States:
evidence from the multi‐city study of urban inequality." Housing Studies 325-
359.
IPCC. 2014. Intergovernmental Panel on Climate Change Working Group III
contribution to Fifth Assessment Report. United Nations, New York: IPCC.
Jun, Hee-Jung. 2013. "Determinants of Neighborhood Change: A Multilevel
Analysis." Urban Affairs Review 319-352.
Kang, In-Sung, and Richard C Feiock. 2006. "Implementation of Growth
Management Policy in Florida Cities: Zoning Approval and Regulatory Policy
Enforcement." International Review of Public Administration 85-98.
Keohane, N.O., R.L. Revesz, and R.N. Stavins. 1998. "The Choice of Regulatory
Instruments in Environmental Policy." Harvard Environmental Law Review
313-68.
© ASCE
Sustainable Human–Building Ecosystems 244
Klinsky, Sonja, and Harald Winkler. 2014. "Equity, sustainable development and
climate policy." Climate Policy 1-7.
Krause, Rachel M. 2011. "Symbolic or Substantive policy? Measuring the extent of
local commitment to climate protection." Environment and Planning C:
Government and Policy 46-62.
© ASCE