Urban building energy modeling

A systematic evaluation of modeling

and simulation approaches

Fatemeh Johari




  
Abstract
Urban energy system planning can play a pivotal role in the transition of urban areas to-
wards energy efficiency and carbon neutrality. With the building sector being one of the main
components of the urban energy system, there is a great opportunity for improving energy effi-
ciency in cities if the spatio-temporal patterns of energy use in the building sector are accurately
identified.
A bottom-up engineering energy model of buildings, known as urban building energy model
(UBEM), is an analytical tool for modeling buildings on city-levels and evaluating scenarios
for an energy-efficient built environment, not only on the building-level but also on the district
and city-level. Methods for developing an UBEM vary, yet, the majority of existing models
use the same approach to incorporating already established building energy simulation software
into the main core of the model. Due to difficulties in accessing building-specific information
on the one hand, and the computational cost of UBEMs on the other hand, simplified building
modeling is the most common method to make the modeling procedure more efficient.
This thesis contributes to the state-of-the-art and advancement of the field of urban build-
ing energy modeling by analyzing the capabilities of conventional building simulation tools to
handle an UBEM and suggesting modeling guidelines on the zoning configuration and levels of
detail of the building models.
According to the results from this thesis, it is concluded that with 16% relative difference
from the annual measurements, EnergyPlus is the most suitable software that can handle large-
scale building energy models efficiently. The results also show that on the individual building-
level, a simplified single-zone model results in 6% mean absolute percentage deviation (MAPD)
from a detailed multi-zone model. This thesis proposes that on the aggregated levels, simplified
building models could contribute to the development of a fast but still accurate UBEM.
If not us, who? If not now, when?
John F. Kennedy
List of papers

This thesis is based on the following papers, which are referred to in the text
by their Roman numerals.

I Johari, F., Peronato, G., Sadeghian, P., Zhao, X., Widén, J. (2020).
"Urban building energy modeling: State of the art and future
prospects", Renewable and Sustainable Energy Reviews, Vol. 128,
article id 109902.

II Johari, F., Nilsson, A., Åberg, M., Widén, J. (2019). "Towards urban
building energy modelling: a comparison of available tools". In
Proceedings of eceee 2019 Summer Study on energy efficiency: Is
efficient sufficient?, 3-8 June, Presqu’île de Giens, Hyéres, France,
1515-1524.

III Johari, F., Munkhammar, J., Widén, J. "Validation of simplified

building energy models for urban-scale energy analysis of buildings",
Manuscript.

Reprints were made with permission from the publishers.

Publications not included in the thesis

V Psimopoulos, E., Johari, F. , Bales, C., Widén, J.(2020). "Impact of
boundary conditions on the performance enhancement of advanced con-
trol strategies for a residential building with a heat pump and PV system
with energy storage.", Energies , Vol. 13, no 6, article id 1413.
Notes on my contribution
I contributed the following to the appended papers:

Paper I, I did the literature survey and wrote all the paper except Sections:
2.1.1, 2.2.1.2, 2.3, 3.3.

Paper II, I developed the building models in IDA ICE, TRNSYS and Ener-
gyPlus, analysed the results and wrote the paper.

Paper III, I developed all the building models and wrote the paper.
Contents

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Aim of this thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Overview of thesis and appended papers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1 Urban energy systems in transition to sustainability . . . . . . . . . . . . . . . . . . . . . 5
2.1.1 Urban energy systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1.2 Building energy policies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.3 The need for a decision-making tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 What is urban building energy modeling? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2.1 City-scale energy modeling of buildings . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2.2 UBEM workflow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2.3 Definition of building archetypes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2.4 UBEM simulation engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2.5 Model calibration and validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3 Previous works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.3.1 UBEM simulation tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.3.2 Advancements in the field of UBEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.4 Research gaps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3 Methodology and data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.1 Overview of available data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.2 Overview of BEM and simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.2.1 Outside heat balance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.2.2 Heat conduction through building envelope . . . . . . . . . . . . . . . . . . 25
3.2.3 Inside heat balance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2.4 Comparison of simulation tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.3 Model complexity and zoning configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.4 Overview of case studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.4.1 Case study 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.4.2 Case study 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.1 Comparison of simulation tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.2 Model complexity and zoning configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.2.1 Zoning configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.2.2 Level of detail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
5 Discussion and future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.1 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.2 Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
6 Conclusion ................................................................................................. 49
References ........................................................................................................ 51
List of abbreviations

BEM Building energy modeling

CEA City Energy Analyst
CTF Conduction transfer function
EPC Energy performance certificate
GHG Greenhouse gas
GUI Graphical user interface
HVAC Heating ventilation and air conditioning
LiDAR Light detection and ranging
LoD Level of detail
MAPD Mean absolute percentage deviation
R-C Resistance-capacitance
RMSE Root mean squared error
UBEM Urban building energy model
WWR Window to wall ratio
1. Introduction

Undoubtedly, cities are one of the main contributors to climate change prob-
lems. More than 75% of greenhouse gas emissions in the world originate from
urban activities and their associated energy use. With globally increased ur-
banization at an unprecedented rate, from 56% in 2020 to 68% in 2050 [1],
there is less chances of reducing global emissions to a safe level unless new
and expanding cities are based on low-carbon urban development paradigms
[2, 3]. Indeed, cities are a key part of the solution for climate mitigation and
adoption.
Worldwide, cities adopt different approaches to the style and structure of
their climate actions [2]. Many cities and municipalities undertake these ac-
tions even in the absence of national policies [4]. Interestingly, these am-
bitious city-level climate-conscious policies can considerably reduce carbon
emissions globally while delivering enormous benefits to the cities [2, 5].
From another perspective, cities can serve as policy laboratories for actions
on climate change that help national governments understand the opportuni-
ties while designing effective policies [6]. In summary, urban or local policies
can contribute to the global climate agenda to a large extent.
Addressing the complexity of climate change, local governments and mu-
nicipalities require a systematic approach that identifies problems, formulates
possible strategies, and evaluates resulting benefits [7]. In this context, urban
planning and, in particular, urban energy planning typically seems as a pivotal
approach that can be moved to the forefront of the transition to sustainability
and carbon neutrality [8, 9]. Urban energy planning is an inclusive concept that
targets many individual components of urban energy systems contributing to
the interactive process of using and supplying energy [10]. Thus, in managing
the transition of urban energy systems, it is required to target these different
components, from generation to transmission and consumption, individually
or in unison with each other [11].
With buildings being one of the most energy intensive components of the
urban energy system, opportunities for accelerated transformation towards
sustainability are enormous and can be realized if best practices in energy
efficiency and integrated renewable energy technologies in buildings are ef-
ficiently used [12]. Statistics show that by construction of more energy ef-
ficient buildings and renovation of existing dwellings up to 35% increase in
energy efficiency of the European household sector could be achieved during
1991-2016 [13]. However, in order to plan for a more resource efficient built

1
environment, understanding of flows of energy in buildings and synergies be-
tween buildings and the other components of the urban energy system is cru-
cial. Historically, translation of physical systems into mathematical equations,
i.e., mathematical models, has paved the way for an improved understand-
ing of systems and have given precision to formulating ideas and identifying
solutions [14]. In other words, mathematical models, if properly calibrated,
are key tools in understanding and explaining the function and dynamics of
energy systems and predicting their behaviour in response to internal and ex-
ternal changes.
An Urban building energy model (UBEM) is a bottom-up engineering-
based (or physic-based) model of energy use in large sets of buildings within a
specific geographic area such as an urban district or a whole city [15]. This is
a new concept that has been developing during the last two decades [16]. An
UBEM is an analytical tool that simulates and visualizes patterns of energy
use in buildings and gives insight into urban energy system behavior related
to buildings [17]. Using a broader definition, the UBEM includes not only
the use but also the supply of energy and provides an estimation of spatio-
temporal patterns of energy flows, both demand and supply, over the whole
city. In particular, UBEMs can be applied when designing and investigating
both new and existing urban areas and systems which makes them an attractive
tool for city planners and policy makers [18]. Overall, the extent and applica-
bility of UBEMs motivate the need and the growing trend for development of
such models to aid in improving sustainability and energy efficiency in cities.
It is generally recognized that the development of an UBEM is a challeng-
ing task that requires handling big data, automated procedures for modeling,
simulation and calibration of many buildings as well as high computational
power [16, 19, 20]. To overcome these challenges, choosing the right simula-
tion engine, and deciding on a suitable level of thermal model complexity, are
key points in UBEM development that have not been systematically addressed
and scrutinized previously, but will be so in this thesis.

1.1 Aim of this thesis

The principal aim of this thesis is to shed new light on the field of urban build-
ing energy modeling and contribute to its advancements. More precisely, the
intention is to provide a foundation for the development of an accurate UBEM,
systematically considering the choice of simulation engine, and complexity of
the model. The following precise goals of the thesis were formulated in order
to reach this aim:

i Undertake an extensive survey of the existing scientific literature on UBEMs

in order to identify the state-of-the-art and best practices in the field.

2
 ii Evaluate the applicability and accuracy of existing building energy model-
ing (BEM) software for potential use as a simulation engine in the UBEM.

iii Determine a suitable level of complexity for the thermal building models
in the UBEM.

1.2 Overview of thesis and appended papers

This licentiate thesis is structured as follows: In Chapter 2, a theoretical back-
ground and an overview of important aspects of the field of urban building
energy modeling are provided. A review of existing research and identified
research gaps are also presented in this chapter. Chapter 3 is a summary of the
methods used in the appended papers, including simulation tools, developed
models, data and case studies. This is followed by Chapter 4, in which the
results of the thesis are presented. Chapter 5 includes further discussion on
the findings and an outlook towards further work. Finally, Chapter 6 draws
the final conclusions. Overall, this thesis summarizes the work that has been
done in the following appended papers:

I Paper I provides a comprehensive and up-to-date state-of-the-art litera-

ture review of multi-scale bottom-up engineering-based UBEMs. This
paper aims to highlight the main approaches, persistent challenges and
possible opportunities for the current research. Besides, it suggests a new
perspective on integrated modeling that includes different elements of ur-
ban energy systems, more specifically buildings and their energy systems,
urban microclimate, district energy systems and, most importantly, urban
human mobility, in one interactive model. This paper fulfills the first aim
(Aim i) of this thesis.

II Paper II is a systematic comparison of four simulation tools that could

potentially work as UBEM simulation engines, namely, the indoor and
climate energy simulation software IDA ICE, the transient system simu-
lation software TRNSYS, and the two building energy simulation tools
EnergyPlus and VIP-Energy. This paper compares modeling procedures,
inputs, outputs and accuracy of these tools. The main focus of this paper
is to investigate the accuracy and suitability of these tools for large-scale
application in UBEM simulation. According to the obtained results from
this paper, it is possible to reach the second aim (Aim ii) in this thesis.

III Paper III investigates the trade-off between complexity and accuracy of
the intended building models for the urban building energy modeling. By
evaluating the most common zoning configurations and levels of detail in

3
 the building envelope components, this paper aims at finding a suitable
level of complexity for the building models, which should be simplified
but still accurate enough for the scope of the urban building energy mod-
eling. The last aim (Aim iii) of this thesis is pursued with this paper.

4
2. Background

In this chapter, the background of this thesis and the state-of-the-art of the
field of urban building energy modeling are presented as follows. In Section
2.1, urban energy systems and the necessity of their transition towards sus-
tainability are discussed. The aim of this section is mainly to shed light on
urban building energy models as a decision-making tool for sustainable urban
development of our future cities. Section 2.2 presents an introduction to the
field of urban building energy modeling and discusses possible opportunities,
and persistent challenges during the development of a reliable model. Section
2.3 reviews the latest advances of the field and introduces some of the notable
models that have been developed so far. Finally, in Section 2.4 a number of
research gaps are identified and it is explained how these gaps are intended to
be filled by the thesis and its appended papers.

2.1 Urban energy systems in transition to sustainability

Considering the unique position of local governments to address causes and
effects of climate change, examples of effective local or municipal climate
actions are abundant [21]. However, there is no clear action plan for urban
energy transition towards sustainability that can be delivered for city-specific
conditions [22]. Urban energy system is a broad concept that includes many
components interactively contribute to the process of supplying and using en-
ergy [10]. Understanding urban energy systems may provide a useful per-
spective for informed and inclusive policies for urban energy transition with
respect to city-specific conditions.

2.1.1 Urban energy systems

There is no doubt that the future of sustainable urban development is tied up
with urban energy systems and their characterizations [22, 23]. In the lit-
erature, the urban energy system is given different definitions. Grubler et al.
[24], define it as a "composition of all components related to the use and provi-
sion of energy services associated with a functional urban system, irrespective
where the associated energy use and conversion are located in space" [24].
Unlike this, Castán Broto [22] emphasises the spatial organization of urban
energy systems and describes them as "spatial organisation of multiple energy

5
services depending on how people use energy (for lighting, thermal comfort,
communications, cooking, transportation), and how energy services are pro-
vided (whether this is for the generation of electricity, gas provision or for the
direct use of fuels for heat or mechanical power)" [22]. However, the most
inclusive definition of urban energy systems are provided by Keirstead et al.
[10, 25] who define the urban energy system as "a formal system that rep-
resents the combined processes of acquiring and using energy to satisfy the
energy service demands of a given urban area".
According to the definition given by Keirstead et al. [10], it can be con-
cluded that the transition towards more sustainable urban energy systems re-
quires coordinated energy planning including the whole chain from primary
energy extraction, through energy conversion, transmission, distribution and
use [26].
Given the existing urban and regional energy planning, Asarpota and Nadin
[27] highlight the components (or the areas) of urban energy systems that can
be instrumental in urban energy transitions into sustainability and carbon neu-
trality. Transport and accessibility is one of the main components of the urban
energy systems that show a potential role in spatial planning and low emission
energy strategies, as can be found in the studies from [28]. Energy infrastruc-
ture is the other key component of the urban energy systems which includes
the technologies on district energy systems such as district heating and cool-
ing networks, multi-energy systems and energy hubs, and renewable energy
generation at district scale [19] as well as electricity network and smart grids
[29]. The other energy related component which plays an important role both
on supply and use of energy is the building stock [27]. From design to sys-
tems, buildings suggest a wide range of opportunities for energy efficiency and
integrated renewable energy strategies [10].

2.1.2 Building energy policies

Among the different elements of the urban energy systems, the building stock
has the greatest contribution to energy use and thus, greenhouse gas (GHG)
emissions. It is estimated that buildings account for 40% of energy use and
36% of GHG emissions in the EU [30]. At the same time, the building stock
provides a great opportunity in renewable energy resource integration, e.g.,
building integrated solar [31]. In this respect, to cope with future impacts of
climate change and to move towards a more sustainable future, national, re-
gional, and local authorities deploy a wide range of building oriented policies
and building codes which mainly aim at follows:

• Improved energy efficiency in buildings through design, construction,

and renovation of buildings and their heating ventilation and air condi-

6
 tioning (HVAC) systems.

• Increase the share of renewable energy resources through building inte-

grated and decentralized energy systems, such as building-integrated or
building-applied solar technologies.

National building codes and regulations are primarily means for increasing
energy efficiency in buildings [32], yet their effectiveness is often constrained
in several ways, among them difficulties and differences in compliance and en-
forcement of the regulations [33]. Nonetheless, a combination of these regula-
tions with local policies, and improvement of the same based on city planning
and development strategies, can be one of the most effective ways of achieving
the target goals for climate mitigation and sustainable development [34]. In
many cities, building standards and installations of energy-efficient technolo-
gies have become widely adopted into local energy and climate action plans
[35]. As an example, in the U.S., Massachusetts municipalities have intro-
duced an above-code appendix to the "base" building energy codes, so called
"stretch-code" [36]. The stretch code emphasises the adoption and enforce-
ment of energy efficient technologies in buildings, is designed to make the
new constructions more energy-efficient than the base energy codes. There-
fore, the building code in Massachusetts is required to become updated every
few years according to the new stretch-code. This eventually leads to incorpo-
ration of stretch-code to the base code [36]. However, although the concept of
localization of building standards seems to be successful, it requires substan-
tial funding and a structured local government with sufficient means [37].

2.1.3 The need for a decision-making tool

Local governments require proper tools and knowledge on how and where the
building policies should be implemented. Traditionally, benchmarking and
certification systems have been common frameworks for identification and im-
plementation of the energy efficiency improvements, and energy use reduction
practices in buildings [38]. LEED (Leadership in Energy and Environmental
Design) [39] and BREEAM (Building Research Establishment Environmen-
tal Assessment Method) [40] are the two recognized examples of a such green
certification systems that have been commonly used by decision-makers, and
stakeholders internationally. When benchmarking is applied to buildings en-
ergy use, it serves as a mechanism to measure the energy performance of a sin-
gle building over time, relative to other similar buildings or to simulations of a
reference building [41]. For years, dynamic energy modeling and simulation
of individual buildings have been widely used in benchmarking of buildings
for planning, demonstration, and evaluation of energy conservation measures
and thermal comfort improvement in individual buildings [42]. However, con-

7
sidering interactions between buildings and the urban environment, their role
in renewable resource envelope solution and dynamic influences of buildings
energy use on district energy systems, the focus has begun to shift from bench-
marking and individual building energy studies to district and city-level solu-
tions. For instance, in the latest action plans towards a carbon free Boston, the
urban building energy modeling is the only measure for examination of energy
efficiency improvement in buildings [43].

2.2 What is urban building energy modeling?

In this section an overview of city-scale energy modeling of buildings and,
more specifically, bottom-up engineering energy models of building, referred
to as urban building energy models (UBEMs), is presented. This section is a
short summary of the most important findings from the review of the field in
Paper I.

2.2.1 City-scale energy modeling of buildings

City-scale dynamic energy modeling of buildings is highly dependant on the
availability and granularity of input data and can vary from top-down to bottom-
up models [42, 44]. Top-down models approach the aggregated energy use
data and tend to find its interconnections with end-use related variables. Based
on this terminology, top-down models determine the long-term transitions in
urban energy data and do not focus on individual end-users [45]. As re-
gards the emphasis on socio-econometric and socio-technical factors, most
top-down models are primarily based on the correlations between energy and
variables such as income, employment rate, energy price, population, house-
hold size, and appliance ownership [44]. However, despite being a straight-
forward method for analyzing the overall urban energy use, top-down models
are inherently unable to capture the dynamics in individual-level data. They
also lack any technological and physical details. Thus, they are less suitable
for the identification of improvement areas in existing buildings and scenario
planning of future buildings [44, 46].
Bottom-up models, on the other hand, are developed based on disaggre-
gated data which are then used for aggregation or extrapolation to the district
or city level. In bottom-up models, the description of the individual building
is based on type of input data, i.e., dwelling properties, building physics and
energy use [42]. Accordingly, bottom-up models can be categorized into three
distinct methods: statistical, engineering (or physical) and hybrid models [20].
Statistical or data-driven bottom-up models rely on the analysis of time-series
or cross-sectional actual energy use data with respect to end-use information,
and give an estimation of energy demand. This means that similar to top-down

8
models, statistical models are also capable of capturing the consumption pat-
terns based on the end-use related variables. However, the dependency of
statistical models on historical measurement data makes them less useful for
studying technological changes and future developments [47, 48].
In engineering (or physical) models, however, the approach is to establish a
close to reality description of a building and its HVAC system using math-
ematical modeling. This means that in engineering techniques, individual
building-level energy use is solely estimated from physical and technological
characteristics of buildings and, thus, no previous knowledge on consumption
patterns or demographic factors is necessary. However, the required level of
detail in the data on physical properties of the building and its systems is quite
extensive [16, 46].
Although these engineering models are practical representations of build-
ings, they cannot be reflective of the uncertain variables of the building mod-
els, such as occupants’ behavior. In addition, they are unable to handle the sys-
tematic uncertainty of the simplified modeling techniques. Thus, the third type
of models, so-called hybrid models, has gained increasing popularity among
model developers. A hybrid model is a collection of both statistical and en-
gineering models with all their respective advantages. In the hybrid models,
while the building is modeled according to the engineering methods, the un-
certainty of the input variables as well as the simplified model is approached
by statistics and statistical models [16]. In the existing literature, bottom-up
models that benefit from engineering models of buildings, is commonly refer
to as "urban building energy models" [16]. Table 2.1 summarizes the strengths
and weaknesses of these three modeling techniques in brief.

9
10
Table 2.1. Comparison of bottom-up city-scale energy models of buildings: statistical, engineering and hybrid models.
Models Strengths Weaknesses References
Statistical models Torabi Moghadam et
• End-user information and variability in • Extensive data and a large number of
al. [12], Nutkiewicz
occupant behaviour is covered. samples are required.
et al. [49], Yang et al.
• No detailed technological information is • Highly dependant on historical data.
[50], Lo et al. [51]
needed. • Not applicable for development
• Demographic factors can be considered. studies.

Engineering models Nageler et al. [52],

• Estimation of energy use for different • Detailed technical information on
Cerezo Davila et al.
spatio-temporal resolutions is possible. buildings and systems is required.
[53]
• Detailed technical information and sys- • Extensive computational effort is re-
tems are considered. quired.
• Development studies, e.g., efficiency • No information on end-user as well
measure and urban development, are as demographic information is in-
possible. cluded.

Hybrid models Nouvel et al. [54],

• Estimation of energy use for different • Detailed technical information on
Ghiassi et al. [55]
spatio-temporal resolutions is possible. buildings and systems is required.
• End-user information and variability in • Extensive data is required.
occupant behaviour is covered. • Extensive computational effort is re-
• Detailed technical information and sys- quired.
tems are considered.
• Results in the most accurate urban en-
ergy development tool.
2.2.2 UBEM workflow
As in individual building energy modeling (BEM), in UBEMs every part of
the model is shaped around geometrical and non-geometrical information, i.e.,
construction, materials, systems and occupancy, which are then imported to a
simulation engine where the energy performance of buildings under specific
weather conditions is calculated and then analyzed. Nonetheless, considering
the scope of urban building energy modeling, following the same procedure
as for BEM is impractical. Successive modeling of hundreds or thousands
of buildings with the same level of detail and model complexity as for BEM
requires endless effort and large sources of information that are not available
[16, 17]. To overcome these issues, UBEM developers rely on a multi-step
procedure, as illustrated in Figure 2.1, with steps that are conducted in se-
quence or simultaneously.
A description of the geometries of buildings and their surrounding objects is
first obtained from a pre-processing step for generating 3D models of buildings
in a city, referred to in the following as a 3D city model [56]. In the 3D city
model, building footprints (2D polygons) are extruded from buildings’ height
and elevation information acquired from national geodatasets, photogramme-
try or laser scanning, i.e., light detection and ranging (LiDAR) data [57]. This
virtual extrusion of buildings based on their height (2.5D massing) results in
shoe-box models of buildings corresponding to what is called Level of Detail
1 (LoD1) [17], as seen in Fig 2.2. However, to give an accurate estimation of
the thermal energy performance of buildings and to investigate the solar po-

Figure 2.1. Illustration of the overall urban building energy modeling workflow.

11
Figure 2.2. 3D city model with LoD1 for a district in Uppsala based on the data
obtained from the Swedish mapping, cadastral and land registration authority, Lant-
mäteriet [59].

tential on rooftops, a higher level of detail, e.g., LoD2, is used in many studies
[54, 57, 58]. In addition to building geometries, the other advantage of a 3D
city model is that it can be used to identify adjacent buildings, which generally
escapes the attention of model developers in the absence of any 3D city model.
Adjacencies of buildings can impact the thermal energy balance of buildings
to a large extent [60]; adjacent walls of buildings influence the heat transmis-
sion and adjacent buildings can be used in shading analysis and calculation of
solar irradiance availability on buildings.
Acquisition of non-geometrical information on buildings, e.g., material and
construction, HVAC systems, and occupancy profiles, is a challenging task,
especially when accessible sources of information are limited. On the other
hand, transforming a large amounts of information into distinctive building
models involves considerable amount of time and resources that is beyond
the scope of UBEMs [61]. As an alternative, model developers commonly
focus on reference or representative buildings instead. These representative
buildings, known as building archetypes, summarize the building stock into a
smaller numbers of buildings while maintaining their diversity. In this respect,
they are expected to limit the amount of effort and information put into the
models without compromising accuracy. Identification of building archetypes
and finding their most important characterization is still one of the most im-
portant parts of UBEMs, elaborated upon further in Section 2.2.3.
Once the 3D city model and the building archetypes are available, they, to-
gether with the prevalent weather data, are imported into an UBEM simulation
engine where the building models are implemented and the energy demand is
calculated and output. This simulation engine is capable of executing simple
to complex models of buildings and their energy systems. In some cases, it
can also handle interconnections of buildings with their urban environment
directly, or through co-simulation with other tools [52].

12
 Nevertheless, not every UBEM simulation engine is comprehensive enough
to include different models of not just buildings but also energy systems and
their components such as district heating and electricity distribution. In this
respect, co-simulation [52] and modular [54] approaches seem to open new
opportunities for more advanced UBEMs.
To increase the accuracy and reliability of the results, as in BEM, the UBEMs
also need to be calibrated and validated against actual energy use data. De-
pending on the availability of data, the calibration and validation of the results
can be done for different temporal and spatial scales. For instance, Cerezo
Davilla et al. [53] compared simulation results with annual and hourly energy
use data on a district scale, Moreover, using Bayesian calibration techniques,
Wang et al. [62] validated the postcode level simulation results with 2 years
of measurement data. Methods for calibration and validation of UBEMs are
elaborated upon in Section 2.2.5.

2.2.3 Definition of building archetypes

As mentioned, collecting and handling data are some of the main challenges
of large-scale energy modeling of buildings and, in particular, UBEMs. How-
ever, the introduction of building archetypes, i.e., reference buildings that are
representative of a group of buildings with similar characteristics, is expected
to moderate modeling complexities to some extent.
To define building archetypes, the building stock is first classified based
on common characteristics that are likely to affect the energy performance of
buildings. With a deterministic approach, it suffices to classify the buildings
based on a collection of features such as buildings’ use, type, year of construc-
tion, HVAC system, or heated floor area [63, 64]. However, the simplified
deterministic approach may lead to a misconceived classification of buildings.
More precisely, by systematic classification of buildings based on generalized
features and unrelated to energy use values, distinguish between variations or
similarities in energy performance of buildings could be hard, specially when
the building performance does not follow its characterizations such as type of
use and year of construction [53]. This stresses the need for the second method
of building classification in which adopting influential features and classifying
buildings are conducted with respect to their energy use [65]. Applications of
probability classification using probability distribution of a set of building’s
common features and supervised learning have been proven to be successful
[66, 67]. Recently, attention has moved to a third classification method using
unsupervised learning and cluster analysis. The advantage of clustering over
the other methods is its approach in not only classifying but also finding the
most representative buildings, to be the building archetype, from each class or
cluster of buildings [55, 68].

13
 After the classification of buildings and identification of building archetypes
are completed, non-geometrical characteristics are collected either for a real or
a virtual building. Using this information, building models can be developed
not only for the building archetypes but also for all similar buildings in every
class. Nevertheless, due to the diversity of buildings and their characteristics,
calibration of the models for building archetypes is often necessary [63, 69].
Methods for calibration and validation of models and archetypes are presented
in Section 2.2.5.

2.2.4 UBEM simulation engine

In the BEM, a thermal model of a building based on geometrical and non-
geometrical information is simulated under prevailing weather conditions. This
results in the thermal energy performance of the building, e.g., energy demand
for comfort cooling and heating, and room temperature [70].
In urban building energy modeling, the principal approach is to model and
simulate the buildings as in the BEM. However, considering the scale of a city,
the simulation needs to be extended over hundreds or thousands of buildings.
For this purpose, traditional procedures and step-by-step modeling techniques,
as in BEM, seem not to be working efficiently. Yet, many model developers
rely on the validity of the BEM and intend to upscale its capability for urban
building energy modeling. In this case, a BEM software is placed in the main
core of the model in order to conduct the simulation while a set of algorithms
handle the automated procedure of simulating the whole city. The approach
of making use of BEM-based simulations is found in the majority of studies
such as the models developed by Cerezo Davila et al. [53], Chen et al. [71],
Nageler et al. [52] and Wang et al. [72]. Although these types of models
deploy the resources and capabilities of the BEM simulation software to ac-
curately estimate the energy use in buildings, their use comes with increased
complexity and computation cost of the model. For this reason, tailor-made
simulation engines might be more appropriate.
Describing buildings through the resistance-capacitance (R-C) analogy was
introduces around early 1970s when the very first energy models of buildings
came into being. The BEM simulation tools also use this approach in their
internal calculations, although the level of complexity might be different. De-
veloping tailor-made algorithms based on the R-C analogy and on simplified
heat balance equations helps modelers to considerably reduce the computation
time as well as the complexity and level of detail in UBEMs. The models de-
veloped by Robinson et al. [73] and Fonseca et al. [74], are some well known
examples of this modeling approach.
In addition to the type of simulation core, i.e., BEM software or tailor-
made algorithms, the UBEMs may differ in considering single or multi-zone
building models. Traditionally, depending on the boundary conditions, e.g.,

14
exposure to the ambient condition, and variations in the internal heating, ven-
tilating and air conditioning of a building, the thermal model is designed to
have one zone or multiple zones. However, in urban building energy model-
ing, in the absence of detailed building-level information, it is assumed that
all buildings can be defined through similar zoning configurations. Thus, the
UBEM simulation engine is also responsible for the implementation of the
predefined zoning configuration.

2.2.5 Model calibration and validation

UBEMs heavily rely on physical and operational characteristics of buildings.
Yet, accessibility to such data is restricted in many cases and existing methods
of generalization from discrete sources of data add to systematic uncertainties
of the model [62]. Furthermore, some of the model parameters, such as occu-
pancy profiles, are inherently uncertain and no model is able to fully capture
their variations in time [75]. On the other hand, to reduce associated complex-
ities of UBEMs, simplification of the model is a common method that also
increases the performance gap between the model and reality. To reduce the
input data uncertainty and optimize the performance of simplified models, cal-
ibration of the model becomes an important part of the UBEMs [62, 76, 77].
The approach to calibrate and refine the model varies in existing research,
but two common methods can be distinguished. An iterative process of ad-
justing the model parameters and comparing the results with energy use data
is commonly used in many studies. Due to the simplicity of the method and
its flexibility in adapting data with various spatial and temporal resolutions,
it seems a straightforward approach to use. This method can be found in the
studies conducted by Heiple and sailor [78], and Leroy et al. [76]. The sec-
ond approach applies statistical methods, in particular, Bayesian statistics and
Bayes’ theorem to predict the uncertain parameters and calibrate the model ac-
cordingly [63, 64, 77, 79]. In an attempt to infer the parameter values from the
posterior distributions of uncertain parameters, Nagpal et al. [80] suggests an
auto-calibrated model to reduce the manual effort in calibrating the UBEMs,
while Kristensen et al. [69] implements a hierarchical setting to propose a
multilevel parameter assessment which forms an optimal solution to infer the
uncertain parameters. However, all the methods of calibrating UBEMs cannot
be fit into these two categories; other different methods can be found in [81]
and [82].
Nevertheless, model calibration methods do not suffice for reflecting on
overall accuracy of an UBEM. In this respect, validation of the model against
actual energy use data seems to be the only way to confirm the result of a
model. However, except in a few examples, such as in [83] and [53], no ex-
plicit information on model validation is provided in existing studies.

15
2.3 Previous works
As regards the large number of studies aimed at developing improved UBEMs,
summarizing all the relevant studies is beyond a few pages. Thus, in this
section, it is intended to only review the most notable studies that led to the
introduction of a new tool. For further reading, the reader is directed to Paper
I. Furthermore, since the year 2019 when the review paper was written, there
are a few additional studies that are deserved to be noted here.

2.3.1 UBEM simulation tools

The very first studies on building energy models and dynamic simulation of
buildings emerged during the 70s and early 80s, such as Clarke [84]. Nonethe-
less, by doing a systematic review it is noted that the modern urban-scale
building energy studies based on engineering methods can not be found until
the early 21st century. Some of the examples of these early-stage engineering
models are found to be published by Huang and Broderick [85] in which dy-
namic energy simulation of prototype buildings (or building archetypes) are
conducted by DOE [86] and extrapolated over the whole stock, and Parekh
[87] that specifically focused on establishing certain criteria in defining build-
ing archetypes. However, it was only after the development of SUNtool [88]
and its successor CitySIM [73] that the attention moved to the usability of
urban models of buildings as a tool to support the increasing demand for sus-
tainable urban planning. In this respect, SUNtool and CitySIM [88] can be
regarded as pioneering models (or tools) in the field of urban building energy
modeling.

CitySIM
Developed at the Swiss Federal Institute of Technology of Lausanne (EPFL),
CitySIM is a simulation tool for analysis of energy demand in buildings with
respect to occupants’ behavior, HVAC systems and urban microclimate and
at different scales, from building to district and city. By receiving benefit
from the geographical user interface (GUI) in Java, CitySIM calculates the
building-related energy flows using a C++ solver in the background. Thermal
models of buildings in CitySIM are originated from the R-C network analogy
where the conducting walls transfer the heat between temperature nodes. On
external surfaces, the temperature nodes are affected by microclimatic condi-
tions, in particular solar radiation, and at the internal nodes, are designed to
be reflective of the occupants’ behavior and its stochastic nature. Based on the
availability of geometrical information and individual building characteriza-
tion, CitySIM seems to produce reasonable results [73, 89, 90].

16
SimStadt
Generally speaking, CitySIM can be regarded as a simulation platform that
only handles dynamic simulations of buildings when the required information
is given as input. Yet, this data is not always available. To round this problem,
SimStadt [54], developed at the University of Applied Science Stuttgart, pro-
poses a new modular workflow based on third party software, e.g., CitySIM,
for conducting a multi-scale urban energy and environmental (CO2 emission)
analysis. To solve the complexities of handling data, using its pre-processing
modules, the missing data is deduced from available information on typology
and usage (archetype) or from probabilistic methods of interpolating from ag-
gregated level data. In other words, it utilizes CitySIM while suggesting new
methods to overcome its shortcomings. Furthermore, SimStadt makes use of
the other already established tools to consider not only the buildings but also
the energy systems. For instance, it makes use of the network analysis tool,
Stenet [91] to evaluate the district heating and cooling networks and associ-
ated distribution losses. With the same approach, PV potential and renewable
system integration are also considered in the tool. However, the main novelty
of SimStadt, as compared to the similar studies of the time, is its approach in
using a modern GUI for conducting fast and parallel calculations when compu-
tation power and power limitation was a big issue for many model developers.

CEA
However, SimStadt is not the only successful example of UBEMs that have
been developed based on modular workflows. The integrated framework for
analysis and optimization of buildings, developed by [74, 92], known as City
Energy Analyst (CEA) is another example of such terminology that incorpo-
rates six different modules for building demand forecasting, resource avail-
ability assessment, simulation of conversion, storage and distribution tech-
nologies, bi-level optimization, and multi-criteria assessment. In terms of
building-specific calculations, CEA takes an analytical approach in the phys-
ical description of dynamic heat and mass transfer along with buildings, sys-
tems, users, and the surrounding environment which is then corrected through
statistical analysis with annual specific values for consumption in buildings,
and classified using k-means clustering and illustrated in the output. This tool
utilizes an innovative 4D interface in ArcGIS to facilitate visualization and
dissemination of the results. Unlike SimStadt, CEA is developed in a single
interface and a series of tailor-made models in Python.

umi and Boston UBEM

Not every UBEM is based on simplified mathematical models. Some UBEMs
take full advantage of the validity and reliability of building simulation soft-
ware in calculating energy demand in buildings. With a similar approach as

17
SUNtool, the urban modeling design platform called umi [93] is an UBEM
with capabilities to evaluate operational building energy use, sustainable trans-
portation choices, day-lighting and outdoor comfort at the neighborhood and
city level. In this tool, Rhinoceros 3D CAD environment and its integrated
visual programming environment, Grasshopper, are used as the modeling plat-
form while EnergyPlus handles the subsequent dynamic simulation of build-
ings in the background. umi forms the basis for the UBEM for the city of
Boston [53]. The Boston UBEM is a city-wide model that captures energy
flows of more than 83000 buildings in the city of Boston. Due to the spatial
scale of the model, it is not comparable with the other similar models. Using
available datasets on building information and specific definitions of building
archetypes, the building stock is modeled from characteristics of 52 use/age
archetypes. As mentioned, umi [93] handles building-by-building modeling
and simulation of thousands of buildings in 60 hours.

A summary of the main features of these models (or tools) is presented in

Table 2.2.

2.3.2 Advancements in the field of UBEM

The Boston UBEM is not the only UBEM that applied the method in sum-
marizing the building stock into building archetypes. In fact, most UBEMs
employ various methods of finding a representative building, as described in
Section 2.2.3. Evolving from purely deterministic approaches , recent studies
leaning towards the application of machine learning for probabilistic classi-
fication or unsupervised clustering of building archetypes. Nonetheless, no
machine learning or statistical method is applicable when not enough data is
available. This implies the current trend in making use of available national
databases and building stock surveys in defining the archetypes. For instance,
Ali et al. [94] suggests a multi-scale archetype development for residential
buildings from multi-level sources of data, i.e., national, city, regional, and
district. Torabi Moghadam et al. [95] refers to geospatial data on building and
household information when categorizing the buildings based on their main
attributes.
As regards the availability of empirical national data, research interests have
been turning to the Energy Performance Certificates (EPCs) and their valu-
able role in gathering buildings’ information in one unified database [96, 45].
The energy performance certificate was first established in order to increase
the awareness of energy performances of buildings in the EU [97] although
nowadays, its strength in informing the important characterization of build-
ings receives increasing attention, particularly in the process of archetype de-
velopment for an UBEM study. Ahren et al. [98] examine the Irish EPCs for
the identification of 35 reference buildings representing the Irish predominant

18
housing typology. Using Swedish EPC data Österbring et al. [99] classifies
the building stock in the city of Gothenburg into use-type reference buildings
that can be used for further analysis of the energy demand for heating and hot
water use. Pasichnyi et al. [67] adopt a statistical approach in processing the
EPCs and defining the building archetypes and suggest strategic analysis and
planning for building energy retrofitting accordingly [100]. However, despite
the usefulness of the EPCs, the associated uncertainties and proven deviations
from real measurements [101] emphasize the necessity of a renewed certifica-
tion system [102] that may benefit from BIM technology, big data techniques,
and use of building smart-readiness indicators to increase accuracy, reliability
and applicability of data [103].
Given the uncertainties of occupants’ behavior, at the building level, a va-
riety of occupancy models, from deterministic to stochastic have been devel-
oped that can generate a close to reality occupancy profiles, as found in Widén
et al. [104], Fisher et al. [105, 106] and McKenna et al. [107]. For urban
level studies, however, the urban occupancy modeling is still uncertain. Based
on the review article that is written by Happle et al. [108], clearly in almost
all existing UBEMs, the urban building occupancy is assumed to be similar
to that of individual building archetypes. Besides, due to the complexity of
stochastic models, it can be seen that the deterministic models are dominant.
To solve the uncertainties of simplified treatments on urban building occu-
pancy, the new generation of studies started to evolve from urban mobility
models. At the time when the review article was written by the author, except
[10, 109], and [73] in which the utilization of agent-based transport models
in capturing the diversity of individual activities was conceptualized, no re-
markable example has been found in this area. However, since then, several
studies are taking the approach to integrate transportation and human mobil-
ity models into UBEMs. Barbour et al. [110] and Wu et al. [111] estimate
absence or presence of building’s occupants from cellphones and mobile po-
sitioning data. Based on overall information about individuals, e.g., students
and teachers, Mosteiro-Romero [112] suggests a population-based model and
assign a daily schedule to individuals in order to reach occupants’ presence.
Happle et al. [113] makes use of web mapping services, i.e., Google Maps
and Facebook, to statistically prepare a schedule for occupancy in commer-
cial buildings, e.g, retails and restaurants, at their locations. There is no doubt
that integrating mobility models into UBEMs can solve the question of urban
building occupancy, yet, all these studies still struggle to approach the stochas-
tic nature of human activities at buildings. As suggested in Paper I, it is still a
research gap in addressing occupancy profiles in UBEMs.

19
Table 2.2. Review of the main features of some of the notable UBEMs.
Tool Building type 3D city model Archetype Thermal model
SUNtool Residential R-C
CitySIM Residential R-C
CEA All types 2 R-C
SimStadt All types 1 CitySIM
umi All types EnergyPlus
Boston UBEM All types 3 umi
1 CityGML LoD1 and LoD2.
2 LoD1 from Open Street Maps
3 LoD1 from available geometrical data

Occupancy Microclimate Energy system Mobility Scenario planning

2 4
3 4
2 4
3 4
3 4 5
3 4 5

2 Stochastic.
3 Deterministic.
4 Radiation model.
5 Walkability.

Novelty Platform GUI Year Reference

First UBEM tools JAVA 2007 [88]
First UBEM tools JAVA 2009 [73]
Parallel computation JAVA 2015 [54]
Modular but single interface ArcGIS 2015-16 [74, 92]
Daylighting and walkability models Rhinoceros 2013 [93]
Geo-spatioal extent Rhinoceros 2016 [53]

2.4 Research gaps

In an attempt to achieve a close to reality estimation of spatio-temporal en-
ergy use in cities, the field of urban building energy modeling has experienced
considerable improvements, during the last decade. Yet, there is still a great
deal of uncertainty about the choice of building simulation tools to be used
for large-scale studies, the level of complexity of the thermal models, and the
availability of good quality data on which to build the model. For this reason,
the existing research gaps in the field have been identified as follows:

20
• Relying on the maturity of individual building energy models and the re-
liability of established building energy simulation tools, a large number
of UBEM studies make use of one of the common BEM tools in their
models. Although these BEM tools should be possible to use in large-
scale studies, there is no comprehensive study that investigates their ad-
vantages and disadvantages as UBEM simulators. Paper II, therefore,
aims to present a clear response to these questions by comparing some
of the most common simulation tools with each other, all of them ap-
plied to the same case, and validating them against measured data.

• Due to the large number of buildings included in UBEMs, traditional

multi-zone building models that are used in BEM studies are not appli-
cable in UBEMs. While some model developers considerably reduce
the complexity of the models to simplified single-zone models, others
follow the ASHRAE guidelines in multi-zoning configuration, i.e., one
core zone and several perimeter zones for each floor. Yet, there is no
clear prescription on the proper level of complexity of the building mod-
els in an UBEM. This issue is the basis for Paper III, in which different
zoning configurations and levels of model complexity are simulated and
their effect on the resulting energy use is analysed and compared in the
most common simulation tools.

21
3. Methodology and data

This chapter introduces the data, the case studies, and the method used in pa-
pers II, and III. Section 3.2 describes the fundamentals of the BEM simulation
tools that are applied in the UBEM studies in papers II and III. In Section 3.2.4
the method suggested for comparison and validation of these BEM simulation
tools for use in UBEM simulation is discussed, as in Paper II. Section 3.3 anal-
yses the complexity and thermal zoning configuration of building models with
respect to the scope of the UBEM, which is a summary of what is presented
in Paper III.

3.1 Overview of available data

As mentioned, in UBEM, having access to good quality data is the main chal-
lenge for many model developers. In the appended papers to this thesis, var-
ious sources of data are exploited to overcome this issue. Depending on the
data source, the accuracy, integrity and completeness of data can be different
between datasets.
The data used in paper II comes from a previous study by Åberg et al.
[114]. Detailed information on building construction and material as well as
energy systems and secondary distribution systems were acquired from the
private housing association HSB 53 brf Gräslöken [115] in the city of Uppsala,
Sweden. The hourly heat use on substation level was also obtained from the
district heating supplier company Vattenfall Heat AB [116]. This data was
measured on the supply side and includes the total thermal energy used, flow
rate, and inlet and outlet temperatures for the year 2015.
In Paper III, however, there was limited access to detailed data. Detailed
information on building construction and blue prints have been provided by
the municipal housing company Uppsalahem [117], while the assumptions on
construction material were based on previous studies and available literature
on the typology of Swedish buildings [118, 119].

3.2 Overview of BEM and simulation

As previously discussed in Section 3.2, the BEM-based urban building energy
modeling forms the major part of the existing UBEM research. Traditionally,

23
Figure 3.1. Illustration of the heat balance over building surfaces exposed to the am-
bient.

the BEM tools such as EnergyPlus, TRNSYS or IDA ICE, take full advan-
tage of physical modeling and numerical simulation of heat and mass transfer
throughout buildings. To be more focused on the thermal performance of a
building, it can be stated that the core of these simulations is based on heat
balance principals for every surface of the building. Figure 3.1 gives a simpli-
fied illustration of the components of the heat balance on a exposed surface of
the building.
In general, the heat balance of the building is summarized into three parts,
including heat balance of the external surfaces, heat conduction through the
building envelope and heat balance on the internal surfaces.

3.2.1 Outside heat balance

An external surface, i.e., external surface of a wall or roof, basically incorpo-
rates the measures for heat exchange with the ambient,
Qo = Qabs + Qconv + Qlw , (3.1)
where Qo is the conduction heat into the wall, Qabs is the absorbed solar ra-
diation on the external surface, Qconv is the representative for convective heat
exchange with the air, and Qlw is the long wave radiation exchanges between

24
the wall and the surroundings.

3.2.2 Heat conduction through building envelope

Fourier’s law specifies the heat conduction proportional to the magnitude of
the temperature gradient,
∂ (x, T )
q = −k , (3.2)
∂x
where q is the conductive heat flux, k is the conductivity of the material and
∂ (x,T )
∂ x is the spatial gradient of temperature in one-dimensional form [120].
In order to capture the heat conduction through the building envelope, e.g.,
walls, and solve the partial equations, conduction transfer function (CTF) and
response factor methods are the dominant methods for estimation of transient
heat transfer in most BEM tools [121]. In response factor, the material ther-
mal response is a linear system that is related to time series of current and
past temperature and heat flux. In CTF, additionally, the current temperature
and heat flux is connected to the past outputs which considerably reduces the
computation time [122].

3.2.3 Inside heat balance

As for the outside, the inside heat balance is calculated on the internal surface
of the walls, roofs, and floors as

Qi = Qabs + Qconv + Qlw , (3.3)

where Qi is the conduction though the wall. Due to contribution of internal

sources of energy, e.g., lighting, Qabs is the absorbed short-wave radiation
from diffused solar as well as internal sources. In this equation, Qlw refers
to the long-wave radiation exchange between internal surfaces of the building
as well as internal sources of energy, e.g, occupants,equipment and lighting.
Finally, Qconv is the heat convection to the air flows in the building.

Methods for calculating each components of the heat balance in buildings

is different from one tool to the other. Besides systematic differences in the
way that a simulation tool conducts the energy model of a building, these
differences in the fundamental heat balance equations contribute to variations
in the results that can be obtained from each tool.

25
3.2.4 Comparison of simulation tools
Among various BEM software, only some are compatible with the scope of
the UBEM. In general, there are technical barriers for the application of some
of the tools that limit them to individual building modeling and make them
unable to cope with urban energy studies. An overview of BEM tools that
could potentially be used for UBEMs and their most relevant features from an
UBEM point of view was presented in [19, 123]. Nevertheless, it seems that
only a few are actually capable of handling UBEMs with all their complexities.
An overview of the most feasible tools, IDA ICE, TRNSYS and EnergyPlus,
that have been used in Papers II and III, is given as follows.
EnergyPlus [124] is an open-source building energy simulation software
that estimates the need for heating, ventilation, and air conditioning of build-
ings using a variety of systems and resources. TRNSYS [125] is a dynamic
simulation software that enables users to study the behaviour of transient and
dynamic systems. Having an extensive library and an organized simulation
environment makes TRNSYS a flexible tool that can be used in many ap-
plications from building energy modeling to system energy analysis and solar
energy studies. IDA Indoor Climate and Energy (IDA ICE) [126] is a dynamic
multi-zone simulation software which accurately models buildings, systems,
and controllers in order to maintain the thermal comfort of the building occu-
pants.
Despite the fact that these tools have many similar features, they can be
very different in their fundamentals, especially when it comes to calculation
of the heat balance and its components.

Calculations of outside heat balance

Absorbed solar radiation, Qabs , as a factor of incoming solar radiation, is de-
fined as,
Qabs = αA(Ib + Id ), (3.4)
where α is the solar absorptance of the surface, A is the wall area, and Ib and Id
are beam and diffuse solar radiation, respectively, on the wall surface, in IDA
ICE and EnergyPlus, direct and diffuse incident radiations are mainly assessed
based on ASHRAE guidelines [127] and Perez model [128], while TRNSYS
gives the possibility of choosing between sets of common methods such as,
the Perez model [129] or the Hay and Davies model [129].
Convective heat transfer on external surfaces is determined based on the
convective heat transfer coefficient as
Qconv = hc A(Tair − Twall ), (3.5)
where hc is the convective heat transfer coefficient, A is wall area, Tair is the
ambient air temperature and Twall is the wall surface temperature. In IDA ICE,
the convective heat transfer coefficient for external surfaces is calculated based

26
on the heat transfer coefficient as a factor of local wind velocity. As default,
EnergyPlus calculates the total convective heat transfer coefficient as the sum
of coefficients for forced and natural convection with respect to the local wind
speed and wind direction. In TRNSYS, on the other hand, the convective heat
transfer coefficient is assumed as either a user-defined variable to be static or
time-dependant or to be calculated internally. The internal algorithms calcu-
late the convective coefficient with respect to the surface inclination and heat
flux.
The long-wave radiation exchange with the surrounding environment can
be generally written as
Qlw = Qlw,sky + Qlw,ground + Qlw,air , (3.6)
with Qlw,sky , Qlw,ground and Qlw,air being the components of radiation exchange
with sky, ground and the air respectively. Generally, the radiation exchange
with the surrounding is estimated using
Qlw,x = εσ Fx A(Tx4 − Twall
4
), (3.7)
where ε is the long-wave emittance of the surface, σ is the Stefan-Boltzmann
constant, A is the surface area, Fx is the view factor between the sky, air or
ground and Tx represents the temperature of the sky, air or ground, while Twall
is the surface temperature. Using Equations 3.6 and 3.7, the long-wave radia-
tion exchange with the surrounding is calculated accurately in EnergyPlus. In
IDA ICE, however, the long-wave radiation exchange with the air, Qlw,air , is
ignored from these equations. Calculation of the radiation exchange is con-
ducted slightly different in TRNSYS, with Tx in Equation 3.7 being the fictive
sky temperature and representing not only the sky temperature but also the
ground temperature and the view factor between the sky and the surface. In
this method, based on the definition of the fictive sky temperature, both ground
and sky long-wave radiation exchange with the surface are summarized under
one term.

Calculation of heat conduction through building envelope

Calculation of the heat transfer through opaque and transparent surfaces of
the building envelope, e.g., walls, and windows, plays a principal role in the
differences that can be seen later in the results from Paper II and III.
In IDA ICE, the calculations of the CTF function is simplified through des-
ignating fewer thermal nodes to the heat transfer surfaces. Using an optimized
thermal-electrical analogy of heat conduction for a given thermal resistance-
capacitance (R-C) network with only three capacitance, IDA ICE conducts an
optimization method to estimate the model parameters. On the other hand,
TRNSYS and EnergyPlus take full advantage of CTF methods and response
factor in calculating the heat conduction through the opaque surfaces. How-
ever, each uses a distinctive solver. TRNSYS utilizes the direct root finding

27
methods [130] and treats the surfaces as black-boxes with no need for informa-
tion on temperature variations inside the surface. EnergyPlus, however, makes
use of the state space methods in solving the analytical CTF models [122]. By
assigning multiple temperature nodes to the heat transfer surface, in Energy-
Plus the heat flux through the wall is calculated from one node to the other.
Descretizing the wall into multiple nodes is expected to reduce the length of
CTF series and thus the computation cost to some extent. More information
on differences in calculation and solving the heat conduction transfer through
the building envelope is found in Paper III.

Calculation of the inside heat balance

As for calculation of the outside heat balance, the components of the heat bal-
ance equations on internal surfaces, Equation 3.3, are estimated differently in
each tool. Convective heat transfer to the zone air and convective heat trans-
fer coefficient follows almost the same terminology as outside convective heat
transfer in TRNSYS and EnergyPlus. In IDA ICE, however, the convective
heat transfer is estimated from the slope of the surface, and the temperature
gradient between the internal air and the internal surfaces.
The absorbed short-wave radiation is determined as a function of transmit-
ted solar radiation through windows

Qabs = XIdi f , (3.8)

with X being the absorption matrix in IDA ICE, the transposed of Gebhart ma-
trix [131] in TRNSYS and the relative absorption of the surfaces multiplied to
the window area in EnergyPlus and Idi f the solar radiation transmitted through
windows.
In calculation of the internal long-wave radiation exchange, EnergyPlus as-
sumes the inside air to be as completely transparent to long-wave radiations,
the internal surfaces of the buildings to be grey-bodies and all the radiation to
be diffuse. Then rely on a unique coefficient for all reflections, absorptions
and re-emissions from other surfaces in the zone, EnergyPlus calculates the
rediative exchange in the building. In TRNSYS, and based on its standard-
level calculation methods, the radiation exchange is approximated from the
star method [122], for a hypothetical temperature node in middle of the room.
In this methods, not only the radiative exchange but also the convective heat is
also contributed to the temperature node. In IDA ICE, the long-wave radiation
is proportional to the radiosity and the properties of the black-body surfaces.

In addition to these differences in the principal methods of calculating heat

transfer, there are systematic differences in how the modeling procedure is
conducted.

28
Ground coupling
To model the heat transfer from the building to the ground below the slab or
the basement in IDA ICE, it is suffice to determine the average annual ground
temperature as well as slab and soil properties. According to ISO 13370 [132],
IDA ICE calculates the heat resistance of layers below the building construc-
tion and, accordingly, the heat transfer through them. If the standard weather
datasets are used, the respective annual ground temperature is assumed auto-
matically. In TRNSYS, the procedure is not as effortless as in IDA ICE. For
this purpose, the external slab components from the TRNSYS library should
be connected to the building. Depending on the building model and desired ac-
curacy, ground coupling models with different level of complexity can be used,
e.g., Type 49 for approximation of the slab on grade, or Type 77 for calculation
the soil temperature. More components for calculation of both slab and base-
ment heat transfer are found in the TESS library [125]. However, computing
time is greatly influenced by choosing alternative components. In EnergyPlus,
different ground coupling concepts exist. The two associated auxiliary tools
for basement and slab modelling in EnergyPlus, conduct an accurate calcula-
tion of heat transfer and boundary temperatures [124].

Shading analysis
Analysis of shading from nearby buildings or surrounding obstacles and its ef-
fects on building performance is one of the sources of differences between the
tools. Using the SkechUp [133] plugin for building’s form and external shad-
ings, TRNSYS acquires the geometrical information of the shading obstacle.
However, implementation of shading analysis from SkechUp file generator is
not feasible for TNSYS 17. In TRNSYS 18, this feature has been included,
yet, the building model does not consider the influences of shading obstacle
on the heat performance of the building as effectively. Instead, it is suggested
to add the external components for shading models provided in the TRNSYS
library. With application of a similar SketchUp Plugin for EnergyPlus, or
by determination of the geometry of the shading obstacle in the site location,
shading analysis in EnergyPlus is easily conducted. Due to the increase in
complexity of the model, the computation time increases slightly. IDA ICE
does the shadings analysis differently. By importing the site plan in 3D format
or by drawing the obstacles, IDA ICE includes the shadings and performs the
simulation automatically.

Occupancy and use profile

In calculation of the metabolic heat gain of occupants, all four tools have
the same methodology using ASHRAE standards [127] and taking just the
convective heat into consideration. However, the main challenge is to define

29
annual schedules for hourly based activities and presence of buildings’ occu-
pants. IDA ICE can accept hourly schedules for week days or weekends as
well as months. However, while specifying the hourly or even daily profiles
seems impractical, importing any external file to its database is impossible.
Thus, the only way is either to use constant or predefined schedules or just
rely on the low resolution profile development. In EnergyPlus and TRNSYS,
the procedure brings no difficulties for the user and it would be possible to im-
port the schedules or load profiles by just importing a text file to the programs.

Number of buildings
In TRNSYS and EnergyPlus it is impossible to model and analyze one build-
ing at a time. However, IDA ICE as a BEM tool is capable of handling the
energy modelling of buildings at larger scales e.g., district level up to a limited
number of thermal zones (300 simple zones and 70 detailed zone).

Time step and simulation

In terms of accepting sub hourly data both in input and output, IDA ICE and
TRNSYS have no limitations, even for considerably small time steps e.g., sec-
onds. Similarly, EnergyPlus can handle minute-based time steps although for
receiving the best output from the tool hourly or at least 10-minute resolution
is suggested.

Given the importance of BEM simulation tools and their usefulness in UBEM,
in Paper II, a comparative study on the validity of the dominant BEM tools,
i.e., IDA ICE, TRNSYS, and EnergyPlus, for being used as the main simu-
lation core of the UBEM is determined. In order to explore their advantages
and disadvantages in large-scale studies, a district-level energy model of 32
district heated multifamily buildings was developed in the three tools.
In this study, it has been tried to make the models as similar as possible in
order to be comparable. Some of the assumptions and conducted methods in
Paper II, are linked to this point. For instance, due to limitations of IDA ICE
in handling hourly load profiles input to the model, the occupancy profile, i.e,
occupant presence, household electricity and lighting as well as DHW, and its
contribution to internal heat gain was calculated separately from the models
and added to the final results later. In this study a stochastic occupancy load
generator developed in [104] was used.

3.3 Model complexity and zoning configuration

As mentioned, to overcome the complexity of simulating a large number of
building models, in the UBEM the aim is to abstract the buildings and their

30
 (a) (b)

(c) (d)

Figure 3.2. Thermal zoning configurations in UBEM: (a) single-zone model (1 zone
per building), (b) multi-zone model (1 zone per floor), (c) multi-zone model (5 zone
per floor) and (d) multi-zone model (3 zones per building: one for bottom floor, one
for top floor and one for all middle floors).

systems, and estimate their performance without having to deal with all their
details. Some examples of such simplification are reducing the number of
thermal zones or reducing the level of detail in the building models. However,
the exact effectiveness of theses alternative methods is not clear in existing
research. The risk of defining too simple models is to increase the uncertainty
and decrease the accuracy.
To untangle these questions about model complexity in an UBEM, in Paper
III, some of the most common zoning configurations and levels of detail are
systematically compared. The single-zone model is traditionally one of the
simplest ways of describing a building by a set of equations. Differences in
heating and air conditioning as well as influences of inner walls of the build-
ing are neglected and it is assumed that the whole building can be described as
one thermal zone. The windows are also modelled based on the corresponding
window-to-wall ratio (WWR) on every external wall. Figure 3.2(a) illustrates
the single-zone model. Although a single-zone model (1 zone per building)
cannot, in most cases, be an accurate representation of an individual build-
ing, it improves the computation cost to a large extent, and might be accurate
enough on an aggregated urban scale.
The second approach is to assign one thermal zone for every floor of the
building, as shown in Figure 3.2(b) (1 zone per floor). Defining a simpli-
fied multi-zone model is expected to slightly capture the contribution of in-
ternal building elements, e.g., adjacent floors, but still run the simulation in
an acceptable time. The third and very common method towards multi-zone
modeling follows the ASHRAE guidelines for envelope setting and zoning

31
 (a) (b) (c)

Figure 3.3. Level of detail in building modeling: (a) LoD1, Shoe-box model of the
building with 25% window area on long walls, (b) LoD2? , Detailed building model
excluding the shading components, (c) LoD3, Detailed building model with precise
geometry of windows and shading components

configuration. According to ASHRAE 90.1, Appendix G [127], different ther-

mal zones are assumed for perimeter and interior spaces that are at least 4.6 m
away from the exterior or semi-exterior walls. This means, for a square base
building, that at least 5 thermal zones, i.e., one core and four perimeter zones,
are required for every floor of the building (5 zones per floor), as visualized in
Figure 3.2(c) (5 zones per floor).
In addition to these three configurations, when tall buildings are of interest,
there is anther alternative in which the building is divided into three main
thermal zones. One thermal zone is assigned to the ground-coupled or the
bottom floor of the building while a second one represents the top floor that is
adjacent to the exposed roof. All the intermediate floors are considered as one
big zone (3 zones per building). Figure 3.2(d) illustrates this.
In most UBEM studies, the description of the building geometry is rela-
tively simplified and the building construction is defined through shoe-box
models corresponding to Level of Detail 1 (LoD1). This level of simplifica-
tion, however, may lead to inaccuracy of the results and, therefore, further cal-
ibration might be needed to compensate for the deviations from the measured
data. Thus, in Paper III, the LoD of the building construction and its impacts
on the simulated energy demand is analyzed in depth. In this study, three lev-
els of detail, namely LoD3, LoD2? , and LoD1, are considered, as shown in
Figure 3.3. Accordingly, LoD3 implies a detailed building model with precise
components and geometries. In LoD2? , the geometry of windows is kept as in
LoD3 but the balconies and shading components are excluded from the model.
Finally, LoD1 represents a simplified shoe-box model with 25% WWR 1 .
Finding the most suitable zoning configuration depends not only on the
scale of the study but also the simulation tool that is used as a main core of
the UBEM simulation engine. In this study, to validate the already mentioned
configurations, three different case study buildings are modeled and analyzed
in the three simulation tools introduced previously, IDA ICE, TRNSYS and
EnergyPlus.
1 The WWR varies between 20% and 25% as a typical value for the window area in buildings of

similar type.

32
Figure 3.4. Overview of the area with 32 buildings modeled in Paper II. Different
building types colored in green, representing building type A and blue showing build-
ing type B.

3.4 Overview of case studies

3.4.1 Case study 1
In Paper II, a neighborhood in Uppsala, Sweden, with 32 buildings of two dif-
ferent types (here defined as types A and B) is considered for modeling and
validation of the respective BEM tools when applied to UBEM simulation,
see Figure 3.4. These buildings are connected to the city’s district heating sys-
tem through a local substation and a secondary heat distribution system, from
which the thermal energy is transferred to the buildings. A detailed description
of the buildings and the distribution heating systems are given in [114].
The buildings in this area are multifamily buildings with two heated floors
and an unheated attic. However, for the sake of simplicity necessary in UBEMs,
in this study the unheated attic was not considered as a part of the building
model and rather merged to the external roof. The buildings of type A have
an average heated floor area of 906 m2 and window area of 102 m2 , while the
buildings of type B have 1019 m2 heated floor area and 109 m2 window area.

3.4.2 Case study 2

In Paper III, three different buildings chosen to be representative of the typical
Swedish district heated multifamily buildings constructed during early 1970s
were considered for further study. Detailed characteristics of these buildings
are can be found in Paper III.

33
 • Building A: Multifamily building with 2 floors.
Building type A is a residential multifamily building with 2 floors and
699 m2 heated area. According to the latest EPC, total energy use (in-
cluding the energy used for space heating, domestic hot water and oper-
ational electricity) in this building is 109780 kWh or 157 kWh/m2 y, for
a normal year.

• Building B: Multifamily building with 3 floors

As previously mentioned, building type B is also a district heated mul-
tifamily building built in 1972. This building has 3 heated floors with
1498 m2 heated floor area. Similarly, overall energy use for a normal
year is calculated to be 235209 kWh or 157 kWh/m2 y, in this building.

• Building C: Multifamily building with 8 floors

In addition to the low-rise multifamily buildings, in this study a high-rise
building with 8 heated floors and an unheated basement is also analyzed.
The normal year energy consumption for this building is reported to be
896738 kWh/y which equals to 173 kWh/m2 y. Its total heated floor area
is 5197 m2 .

34
4. Results

In this chapter, concluding results from Papers II, and III are presented in brief.
Section 4.1 summarizes the results from the comparison and validation of the
most common BEM simulation software to be potentially used as the main
simulation core of the intended UBEM. Section 4.2 provides a summary of the
investigation of zoning configurations and level of detail in thermal models of
buildings and their impact on simulated energy demand.

4.1 Comparison of simulation tools

This section summarizes the most important results from modeling and simu-
lation of a case study area, in the BEM simulation tools IDA ICE, TRNSYS,
and EnergyPlus and reflects on their validity and reliability in large-scale stud-
ies. The full results are presented in Paper II. Here there are slight changes
in measures and indicators for analyzing the results, as compared to Paper
II. Furthermore, the results on the performance of the BEM software VIP-
Energy are excluded here. As stated in Paper II, the functionality of this tool
for large-scale studies is limited and, thus, despite its strong correlation with
the measurement data, it is excluded from this summary and from subsequent
analyses.
Following the suggested methodology in Section 3.2.4, the energy model
of a neighborhood with 32 district heated multifamily buildings was imple-
mented in IDA ICE 4.8, TRNSYS 17, and EnergyPlus 8-7-0. IDA ICE is
capable of handling more than one building model in every simulation stu-
dio and gives the user the opportunity to easily move from individual-level
to district-level studies. This is in contrast with TRNSYS and EnergyPlus, in
which every building model is treated individually. This means that, for the
latter simulation tools, in order to reach the district-level energy performance
of buildings, the obtained results from individual buildings need to be aggre-
gated externally.
The final results from modeling and simulation of the buildings with two
heated thermal zones and an ideal heating system1 maintaining the room tem-
perature constantly at 22◦ C are presented in the following.
Figure 4.1(a) shows the annual heat use calculated in the three tools, IDA
ICE, TRNSYS and EnergyPlus, as compared to the measured data for the year
1 An ideal heating system is a room unit that maintains the room temperature at the set temper-
ature without giving any information on actual physical specifications of the room unit.

35
2015. Overall, the simulation results from all three tools seems to be in a
reasonable approximation to the measurements. It is observed that IDA ICE
and TRNSYS tend to overestimate the demand by 18% and 15% respectively,
whereas EnergyPlus results in 13% lower energy use. To be more precise,
while the actual heat use over the whole neighborhood is 3.48 GWh/y (115
kWh/m2 y), it is calculated to be 4.1 GWh/y (136 kWh/m2 y) for IDA ICE,
4.03 GWh/y (133 kWh/m2 y) for TRNSYS, and 3.03 GWh/y (100 kWh/m2 y)
for EnergyPlus.
The same conclusion is drawn from the analysis of the correlations of the
hourly simulated results form TRNSYS and EnergyPlus versus measured data,
as seen in Figure 4.1(b) and presented in Table 4.1. Yet, IDA ICE indi-
cates weaker correlation with measurements, in particular when the hourly
heat demand is higher than 600 kWh/h. Clearly, during heating hours, IDA
ICE highly overestimates the demand, while at low heating hours it gives a
scattered profile with an overall tendency of underestimating the demand. Al-
though on an annual basis IDA ICE and TRNSYS result in very similar num-
bers, on hourly basis it can be seen that TRNSYS has a more consistent profile
with lower variability as compared to IDA ICE. However, it has a positive bias
in estimating the heat use.
Table 4.1. Correlation of the simulation results with measurement data.
Tool MAE RMSE R2
IDA ICE 72 203 0.80
TRNSYS 61 130 0.83
EnergyPlus 52 117 0.83

Figure 4.2 demonstrates the heat use profile that is simulated in the three
tools as compared to the measurements over the course of a year. Figure
4.2(a) demonstrates the deviation of the results from the measurements (ε),
calculated as
ε = Qs − Qm , (4.1)
where Qs is the simulated heat use and Qm is the actual heat use that is mea-
sured at the district heating substation. This figure is presented with hourly
time resolution. Following the results presented in Figure 4.1(b), here, the
high diurnal and seasonal variability of the simulated results in IDA ICE is
also evident. On the other hand, in TRNSYS and EnergyPlus the variability
is less extreme and more consistent throughout the whole year. While the de-
viation of the results from measurement data (ε) can go beyond ±0.5 MWh/h
(±0.016 kWh/m2 h) in IDA ICE, it mainly fluctuates between ±0.3 MWh/h
(±0.009 kWh/m2 h) for TRNSYS and EnergyPlus. In Figure 4.2(b), compar-
ison of the daily heat demand profiles from the tools with respect to actual
use per each day is presented. Clearly, all three tools follow the same trend as
the measurements, with EnergyPlus being closest to the actual data. Despite
differences in magnitude, EnergyPlus and TRNSYS behave very similarly.

36
 (a)

(b)

Figure 4.1. Validation of the simulation results from IDA ICE, TRNSYS and Energy-
Plus against measured heat use data: (a) is the total heat used in the year 2015, and (b)
shows the scatter plot that indicates the correlations between simulated and measured
values for all 8760 hours of the year.

37
 (a)

(b)

(c)

Figure 4.2. Validation of hourly heat use profile from IDA ICE, TRNSYS and En-
ergyPlus against hourly measurement data: (a) is the hourly deviation profile of the
simulation results from measurements, and (b) is the daily heat use profiles with sub-
plot (c) zooming in on 200 hours of the profile with higher time resolution (hourly
resolution).

38
 Considering the analogy of the simulation tools in calculating the heat bal-
ance in the building (Section 3.2), the main reasons in observed discrepancies
between the results lie in the contribution of solar energy heat gain and ther-
mal losses through building envelope model. As compared to TRNSYS and
EnergyPlus, in IDA ICE, simplified thermal zone models result in higher sen-
sitivity of the values to the ambient condition, particularly when there is solar
radiation. The same observation is reported by Nageler et al. [134]. The sec-
ond reasoning is related to the influence of the thermal mass of the building
envelope and thermal responses to short term variations in the ambient condi-
tion. As can be seen in the results from Paper III, reduced thermal mass of the
building in IDA ICE increases the total heat use in the building and at the same
time increases the variability of the hourly heat use profile in response to di-
urnal and hourly changes in the ambient conditions. Finally, ground coupling
and limitations in proper calculation of the heat exchange with the ground lead
to higher energy use in TRNSYS and IDA ICE as compared to EnergyPlus in
which the use of an auxiliary slab model improves the results.

4.2 Model complexity and zoning configuration

Following the suggested method in Paper III, analysis of the different zon-
ing configurations and model complexity was conducted using the same three
BEM simulation tools IDA ICE 4.8, TRNSYS 18 and EnergyPlus 9.1. The
simulations were conducted for a normal year 2 and with hourly time steps.

4.2.1 Zoning configuration

Figure 4.3 illustrates the results from the energy modeling of building type
A, multifamily building with two floors, with different zoning configurations
in IDA ICE, TRNSYS and EnergyPlus. Figures 4.3(a)-(c) present the daily
average heat demand for 365 days of the year with a close-up of hourly heat
use for an example day in winter. In these figures, daily and seasonal variations
of the results with respect to the ambient conditions can be clearly observed.
Finally, Figures 4.3(d) shows the total annual heat demand that is aggregated
from dynamic simulation of the models in the different tools.
Considering Figure 4.3(a)-(c), it can be concluded that EnergyPlus gives no
major changes in the annual results when the zoning configuration is changed.
On the contrary, IDA ICE is more influenced by the total number of thermal
zones in the building. As a result, for the 1 zone per building model it cal-
culates 8% higher annual thermal energy demand as compared to the 5 zones
per floor model. The same trend can be seen in the case of TRNSYS. This
means that in comparison to EnergyPlus, in IDA ICE and TRNSYS, there is
2A normal year describes the average climate for a given place and time period, i.e., one year.

39
 (a) IDA ICE

(b) TRNSYS

(c) EnergyPlus

(d) Annual results

Figure 4.3. Comparison of zoning configuration of the case study building type A in
different tools: subplots (a), (b), and (c) represent the results for daily average heat
use over the coarse of a year with a zoom-in on hourly heat use for an example day in
winter. Subplot (d) compares the results of annual heat use.

40
a systematic tendency to over-estimate the heat use when the simplest zoning
configuration is applied.
Refer to 4.3(d), on the annual basis with more complex zoning configura-
tion, i.e., 5 zones per floor, all three tools tend to produce very similar results,
where the deviation between the annual values from each tools is less than 1%.
When reducing the number of thermal zones to 1 zone per floor and 1 zone
per building, the heat use profiles start diverging from each other to the extent
that in the simplest model, i.e., 1 zone per floor, IDA ICE gives 9% and 5%
higher annual values as compare to EnergyPlys and TRNSYS, respectively.
As mentioned in Section 3.2, in all three tools, estimations of the heat de-
mand in a building is conducted using numerical heat transfer equations. How-
ever, the differences in underlying methods for calculations of the heat transfer
components, leads to observed deviations in the simulated results from each
tool. On account of these differences, in IDA ICE, exclusion of the thermal
mass of the building from the building envelope leads to higher sensitivity of
the building envelope to diurnal and hourly variations in the ambient condition
and thus, higher variability in the simulated result. On overall, this could re-
sult in higher annual thermal energy demand in IDA ICE when a single-zone
building is modelled. With increase in the number of the thermal zones, and
including internal building components, e.g., internal floors, ceiling and walls,
the effects of thermal mass of the building can be covered to some extent.
This is the main reason why with 1 zone per floor model and 5 zones per floor
model, the estimated heat demand from IDA ICE becomes closer to that of
TRNSYS and EnergyPlus.
For building type B, multifamily building with three floors, the results are
very similar to the ones for building type A and hence are not shown here,
and the same conclusions can be drawn. EnergyPlus gives the most robust
results with respect to changed zoning configuration. However, IDA ICE still
estimates the heat use for 1 zone per building model to be 15% higher than
EnergyPlus and TRNSYS.
Figure 4.4 shows a comparison of the simulation results for the energy
model of building type C, multifamily house with 8 floors, with different zon-
ing configurations in the three simulation tools IDA ICE, TRNSYS and En-
ergyPlus. Here, it can also be noticed that with the increase in the number of
the floors to 8, the results of the 1 zone per floor model get closer to those of
the 5 zones per floor model. In this case, the deviation between the 1 zone
per building and 5 zones per building models reaches almost 0% in all three
tools. However, while thermal zoning plays a negligible role in EnergyPlus
and TRNSYS, in IDA ICE it still has a considerable impact on the results. As
seen in Figures 4.4(a) and 4.4(d), the calculated total space heating demand in
IDA ICE is equal to 128.85 MWh/y or 212 kWh/m2 y with 1 zone per build-
ing, while it reduces to 113.62 MWh/y or 187 kWh/m2 y with having 5 zones
per floor. This means that there is a 14% difference in the results for detailed
compared to simplified zoning configuration in IDA ICE. For the case study

41
 (a) IDA ICE

(b) TRNSYS

(c) EnergyPlus

(d) Annual results

Figure 4.4. Comparison of zoning configuration of the case study building type C in
different tools: subplots (a), (b), and (c) represent the results for daily average heat
use over the coarse of a year with a zoom-in on hourly heat use for an example day in
winter. Subplot (d) compares the results of annual heat use.

42
building type C, EnergyPlus gives smaller values as compared to IDA ICE and
TRNSYS.
Following the same justification for Building A and as a result of increased
thermal mass of the building, the simulated heat demand for 1 zone per floor
gets closer to the values for 5 zones per floor model. In addition to the influ-
ence of the thermal capacity of the building on stability of the heat demand,
the other clear trend here in this figure is the differences in the estimated heat
demand in EnergyPlus as compared to IDA ICE and TRNSYS. The motiva-
tion for this can be provided from the discussion on ground-coupling and use
of different methods in calculation of the heat flux to the ground. In EnrgyPlus
the auxiliary basement tool is used which can result in accurate calculations of
the heat transfer from the basement to the ground. In IDA ICE, the approxima-
tion of the heat conduction to the ground follows ISO 13370 guidelines [132].
In the case of TRNSYS, rely on standard component library and unavailability
of a proper standard component for modeling the ground-coupled basement,
the slab model Type 49 has been used instead. With this, some deviations from
the accuracy in the result is expected.
As mentioned, in high-rise buildings, zoning configuration can be different
from low-rise buildings. As regards the complexity of the model and the com-
putation time associated with the large number of thermal zones, in addition to
the previous configurations, the 3 zone per building is also evaluated for this
case, which is visualized in Figure 4.4(a)-(d). It can be seen that the perfor-
mance of a 3 zone per building model is quite similar to a 1 zone per building
model, which might suggest that even a single-zone building model could be
used. However, a major advantage of having 3 zones per building instead of 1
zone per building is its applicability in mixed-use buildings where lower floors
have a different use profile from the upper floors.

4.2.2 Level of detail

Figure 4.5 shows simulation results for a building model with three different
levels of detail, i.e., LoD3, LoD2? , and LoD1, in IDA ICE, TRNSYS and
EnergyPlus. The illustration of the annual heat use, Figure 4.5(d), proves
that on overall, the mean absolute percentage deviation (MAPD) of LoD2?
and LoD1 from LoD3 is 25% and 21% for IDA ICE and 25% and 20% for
EnergyPlus, while for TRNSYS it is only 8% and 3% respectively.
Figure 4.5(a)-(c) suggests that these slight deviations between the LoD2?
and LoD1 with LoD3 mainly occur during the warmer periods of the year
when the solar altitude is higher and the incident solar radiation through win-
dows can be reduced by the shading objects considerably. However, in case
of TRNSYS, as been already mentioned in Section 3.2, the software does not
consider the impact of the shading obstacles on heat performance of the build-

43
 (a) IDA ICE (b) TRNSYS

(c) EnergyPlus

(d) Annual results

Figure 4.5. Results from comparison of the levels of detail in building energy model
in different tools. Subplots (a), (b) and (c) illustrate the monthly results for IDA ICE,
TRNSYS and EnergyPlus. Subplot (g) shows the annual heat all three LoDs and
simulation tools.

44
ing effectively and as a result, no distinguished differences is seen in the result
from simulation of the building with three LoDs.
Overall, based on the definition of the level of detail given in Paper III,
LoD1, i.e., the shoe-box model of a building, can deviate by 25% from the
detailed building models annually. However, this deviation between the results
is basically seen when the heating demand is less.

45
5. Discussion and future work

This chapter contributes to the discussion around the results from Paper II
and Paper III. Furthermore, a future outlook and further suggestions on future
work is presented in this chapter.

5.1 Discussion
In this thesis a review of the state-of-the-art of the field of urban building en-
ergy modeling is summarized in Paper I. From conception to implementation,
Paper I contributes to comprehensive survey of the existing UBEMs, identifies
challenges and opportunities, and presents an outlook for future developments.
In Papers II and III, the overriding aim is to provide the basis for development
of a simple but still accurate UBEM.
According to the findings from Papers II, and III, it can be observed that
not every BEM tool handles the complexities involved in of urban building
energy modeling. Even though most validated and commercialized BEM tools
can generate very accurate and close-to-reality energy models of individual
buildings, when it comes to urban building energy modeling their capabilities
become restricted. From the three building energy simulation tools that have
been evaluated in Paper II, and later in Paper III, only EnergyPlus could meet
almost all the requirements for development of a simple yet accurate UBEM.
Considering all the features that have been discussed in Paper II and Paper
III, including modeling approach, and calculations around heat balance of a
building, EnergyPlus seems a most suitable tool to be used in simplified, but
rather accurate UBEMs. The short computation time in EnergyPlus is also
another merit. While it only takes a few seconds to conduct a simple single-
zone model simulation in EnergyPlus, it reaches to some minutes for IDA
ICE and TRNSYS. Thus, the overall computation time for large number of
buildings increases considerably.
Finally, as presented in Paper III the deviation of single-zone building mod-
els from detailed multi-zone models can be negligible on building level stud-
ies. This means that, on the aggregated level and for city-scale energy mod-
eling, the simplified zoning configuration does not adversely influence the ac-
curacy of the result. With this being said, it is more likely to have even more
simplifications in the models and yet reach the results with an acceptable accu-
racy. In this respect, the the level of detail (LoD) of the building model should
not be an issue on the aggregated level although it is still possible to make
use of correction factors and calibration methods in considering the shading
objects adjacent to the building, such as balconies.

47
5.2 Future work
The field of urban building energy modeling is still under further development
and this opens up significant opportunities for contributions to the advance-
ment of UBEMs. Here follow a few possible directions for future work.

• Data collection is one of the biggest challenges in UBEM. In the ab-

sence of detailed information on buildings, the availability of national
databases for geographical and property information together with infor-
mation on energy performance certifications can be expected to facilitate
the modeling procedure considerably. For this reason, the future outlook
to this work is to analyze the possibility of developing an UBEM using
available national databases and determine their validity and reliability
in UBEM studies.

• Abstracting the building stock into a few building archetypes seems to be

an optimal solution for reducing the complexity and extent of UBEMs.
Except for a small number of studies, a majority have relied on deter-
ministic classification of buildings based on variables such as type and
year of construction. However, deterministic methods cannot represent
the diversity of buildings and thus the second outlook to this work is to
make use of machine learning techniques in probability classification of
building stock and finding the most representative building archetypes.

• As inferred from the review of the state-of-the-art, most UBEMs are

lacking a comprehensive perspective in reflecting on urban energy sys-
tems. Existing models focus more on buildings with some including
the energy systems and microclimate. Yet, no UBEM is found that in-
tegrates different elements of urban energy system in one framework.
This means that there is a need for a more comprehensive urban model
that extends the definition of an UBEM and encompasses not only the
buildings but also the urban human mobility and transport, urban energy
systems (both renewable and conventional), and distribution (both heat
and electricity) as well as urban microclimate. Therefore, an outlook
for future work is to develop a more comprehensive model that can be
referred to as urban and decentralized energy model (UDEM).

48
6. Conclusion

This thesis contributes to the state-of-the-are of the urban building energy

models. UBEMs have emerged from the need for city-wide energy plan-
ning for a low carbon and energy efficient built environment. From individual
buildings to district and city levels, an UBEM provides insight into the energy
performance of buildings, identifies effective energy efficiency measures and
building integrated renewable energy technology solutions.
Among different methods of developing an UBEM, a BEM-based model,
which benefits from already established building energy simulation techniques
in conducting city-scale studies, seems promising. The results of this thesis
provide simulation evidence that on aggregated levels (district- or city-level),
the building energy simulation software, EnergyPlus, gives the least deviations
from measured data. According to the simulation results of an UBEM for a
case study neighbourhood, it is concluded that the root mean squared error
(RMSE) from hourly measurements is calculated to be 117 kWh for Energy-
Plus, while it reaches to 130 kWh and 203 kWh for TRNSYS and IDA ICE,
respectively. In terms of applicability, capability and accuracy, EnergyPlus
can also have more advantages over the others when used in UBEMs.
From the results of this thesis, it is also concluded that for the scope of urban
building energy modeling, simplified zoning configurations and lower levels
of detail, i.e., 1 zone per building model with LoD1 (or single-zone shoe-
box models), can generate accurate estimations of the heat use on individual
building-level. It is also expected that in an UBEM and on aggregated levels
the deviations between estimated results and actual data become smaller as the
differences from simplified building models would be averaged out over.

49
Acknowledgement

I would like to express my sincere gratitude to my supervisor Joakim Widén

and my co-supervisor Joakim Munkhammar for their endless support. Your in-
sightful feedback and constant encouragement helped me a lot throughout this
process. I also wish to thank my co-authors–beside Joakim Widén and Joakim
Munkhammar– Annica Nilsson, Magnus Åberg, and Guiseppe Peronato, for
their valuable contribution to my work. It has been a pleasure working with
you. I also appreciate working and collaborating with my master thesis su-
pervisors and later co-authors, Chris Bales and Emmanouil Psimopoulos. I
gratefully acknowledge the members of UBMEM work group, and all my col-
leagues at the division of Civil Engineering and Built Environment. Thank
you for interesting discussions and practical guidance.
And finally, to my dearest husband and my best friend forever, thanks for
believing in me and supporting me unconditionally.

50
References

[1] H. Ritchie and M. Roster, “Urbanization,”

http://precog.iiitd.edu.in/people/anupama, May 2018.
[2] “A global opportunity for cities to lead.” [Online]. Available:
https://www.c40.org/why_cities
[3] W. F. Lamb, M. W. Callaghan, F. Creutzig, R. Khosla, and J. C. Minx, “The
literature landscape on 1.5◦ C climate change and cities,” Current Opinion in
Environmental Sustainability, vol. 30, pp. 26–34, Feb. 2018. [Online].
Available:
http://www.sciencedirect.com/science/article/pii/S1877343517301434
[4] D. Reckien, M. Salvia, O. Heidrich, J. M. Church, F. Pietrapertosa,
S. De Gregorio-Hurtado, V. D’Alonzo, A. Foley, S. G. Simoes,
E. Krkoška Lorencová, H. Orru, K. Orru, A. Wejs, J. Flacke, M. Olazabal,
D. Geneletti, E. Feliu, S. Vasilie, C. Nador, A. Krook-Riekkola,
M. MatosoviÄ, P. A. Fokaides, B. I. Ioannou, A. Flamos, N.-A. Spyridaki,
M. V. Balzan, O. Fülöp, I. Paspaldzhiev, S. Grafakos, and R. Dawson, “How
are cities planning to respond to climate change? Assessment of local climate
plans from 885 cities in the EU-28,” Journal of Cleaner Production, vol. 191,
pp. 207–219, Aug. 2018. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0959652618308977
[5] R. Freeman and M. Yearworth, “Climate change and cities: problem
structuring methods and critical perspectives on low-carbon districts,” Energy
Research & Social Science, vol. 25, pp. 48–64, Mar. 2017. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S2214629616302936
[6] J. R. Carreón and E. Worrell, “Urban energy systems within the transition to
sustainable development. a research agenda for urban metabolism,” Resources,
Conservation and Recycling, vol. 132, pp. 258 – 266, 2018. [Online].
Available:
http://www.sciencedirect.com/science/article/pii/S0921344917302434
[7] U. Eicker, Urban Energy Systems for Low Carbon Cities. Oxford: Academic
Press [Imprint], 2018.
[8] R. Madlener and Y. Sunak, “Impacts of urbanization on urban structures and
energy demand: What can we learn for urban energy planning and
urbanization management?” vol. 1, no. 1, pp. 45–53. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S2210670710000077
[9] S. Cajot, M. Peter, J. M. Bahu, A. Koch, and F. Maréchal, “Energy planning in
the urban context: Challenges and perspectives,” vol. 78, pp. 3366–3371.
[Online]. Available:
http://www.sciencedirect.com/science/article/pii/S1876610215024844
[10] J. Keirstead, M. Jennings, and A. Sivakumar, “A review of urban energy
system models: Approaches, challenges and opportunities,” Renewable and

51
 Sustainable Energy Reviews, vol. 16, no. 6, pp. 3847 – 3866, 2012. [Online].
Available:
http://www.sciencedirect.com/science/article/pii/S1364032112001414
[11] R. V. Walker, D. Poponi, and B. Lefevre, “Advancing toward a more
sustainable urban energy system,” p. 10.
[12] S. Torabi Moghadam, J. Toniolo, G. Mutani, and P. Lombardi, “A gis-statistical
approach for assessing built environment energy use at urban scale,”
Sustainable Cities and Society, vol. 37, pp. 70 – 84, 2018. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S2210670717303311
[13] “Progress on energy efficiency in Europe â European Environment Agency.”
[Online]. Available: https://www.eea.europa.eu/data-and-maps/indicators/
progress-on-energy-efficiency-in-europe-3/assessment
[14] X. Lu, D. Clements-Croome, and M. Viljanen, “Past, present and future
mathematical models for buildings: Focus on intelligent buildings (Part 1),”
Intelligent Buildings International, vol. 1, pp. 23–38, Jan. 2009.
[15] A. Sola, C. Corchero, J. Salom, and M. Sanmarti, “Simulation Tools to Build
Urban-Scale Energy Models: A Review,” Energies, vol. 11, no. 12, p. 3269,
Nov. 2018. [Online]. Available: http://www.mdpi.com/1996-1073/11/12/3269
[16] C. F. Reinhart and C. Cerezo Davila, “Urban building energy modeling â A
review of a nascent field,” Building and Environment, vol. 97, pp. 196–202,
Feb. 2016. [Online]. Available:
http://linkinghub.elsevier.com/retrieve/pii/S0360132315003248
[17] T. Hong, Y. Chen, X. Luo, N. Luo, and S. Lee, “Ten questions on urban
building energy modeling,” Building and Environment, vol. 168, 2020.
[18] Y. Q. Ang, Z. M. Berzolla, and C. F. Reinhart, “From concept to application: A
review of use cases in urban building energy modeling,” Applied Energy, vol.
279, p. 115738, Dec. 2020. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0306261920312289
[19] J. Allegrini, K. Orehounig, G. Mavromatidis, F. Ruesch, V. Dorer, and
R. Evins, “A review of modelling approaches and tools for the simulation of
district-scale energy systems,” Renewable and Sustainable Energy Reviews,
vol. 52, pp. 1391–1404, Dec. 2015. [Online]. Available:
http://linkinghub.elsevier.com/retrieve/pii/S1364032115007704
[20] F. Johari, G. Peronato, P. Sadeghian, X. Zhao, and J. Widén, “Urban building
energy modeling: State of the art and future prospects,” vol. 128, p. 109902.
[Online]. Available:
http://www.sciencedirect.com/science/article/pii/S1364032120301933
[21] V. Castán Broto and H. Bulkeley, “A survey of urban climate change
experiments in 100 cities,” Global Environmental Change, vol. 23, no. 1, pp.
92 – 102, 2013. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0959378012000891
[22] V. C. Broto, “Energy landscapes and urban trajectories towards sustainability,”
Energy Policy, vol. 108, pp. 755 – 764, 2017. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0301421517300095
[23] A. Grubler and D. Fisk, Energizing Sustainable Cities: Assessing Urban
Energy. Routledge, Nov. 2012.
[24] A. Grubler, X. Bai, T. Buettner, S. Dhakal, D. J. Fisk, T. Ichinose, J. E.

52
 Keirstead, G. Sammer, D. Satterthwaite, N. B. Schulz, N. Shah, J. Steinberger,
H. Weisz, G. Ahamer, T. Baynes, D. Curtis, M. Doherty, N. Eyre, J. Fujino,
K. Hanaki, M. Kainuma, S. Kaneko, M. Lenzen, J. Meyers, H. Nakanishi,
V. Novikova, K. S. Rajan, S. Seo, R. M. Shrestha, P. R. Shukla, A. Sverdlik,
and J. Sathaye, “Urban Energy Systems,” in Global Energy Assessment (GEA),
T. B. Johansson, N. Nakicenovic, A. Patwardhan, and L. Gomez-Echeverri,
Eds. Cambridge: Cambridge University Press, 2012, pp. 1307–1400.
[Online]. Available: https://www.cambridge.org/core/product/identifier/
CBO9780511793677A037/type/book_part
[25] J. Keirstead and N. Shah, Urban Energy Systems: An Integrated Approach.
Routledge, Mar. 2013.
[26] C. Brandoni and F. Polonara, “The role of municipal energy planning in the
regional energy-planning process,” Energy, vol. 48, no. 1, pp. 323–338, Dec.
2012. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0360544212005130
[27] K. Asarpota and V. Nadin, “Energy Strategies, the Urban Dimension, and
Spatial Planning,” Energies, vol. 13, no. 14, p. 3642, Jul. 2020. [Online].
Available: https://www.mdpi.com/1996-1073/13/14/3642
[28] M. Shepero and J. Munkhammar, “Spatial Markov chain model for electric
vehicle charging in cities using geographical information system (GIS) data,”
Applied Energy, vol. 231, pp. 1089–1099, Dec. 2018. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0306261918314922
[29] X. Zhang, M. Lovati, I. Vigna, J. Widén, M. Han, C. Gal, and T. Feng, “A
review of urban energy systems at building cluster level incorporating
renewable-energy-source (RES) envelope solutions,” Applied Energy, vol. 230,
pp. 1034–1056, Nov. 2018. [Online]. Available:
https://linkinghub.elsevier.com/retrieve/pii/S0306261918313503
[30] “Sustainable buildings for Europe’s climate-neutral future,” Jun. 2019.
[Online]. Available: https://ec.europa.eu/easme/en/news/
sustainable-buildings-europe-s-climate-neutral-future
[31] S. A. Kalogirou, “Building integration of solar renewable energy systems
towards zero or nearly zero energy buildings,” International Journal of
Low-Carbon Technologies, vol. 10, no. 4, pp. 379–385, Dec. 2015, publisher:
Oxford Academic. [Online]. Available:
https://academic.oup.com/ijlct/article/10/4/379/2363478
[32] International Energy Agency, “In support of the G8 Plan of Action,” in
Deploying Renewables: Principles for Effective Policies. OECD, Oct. 2008,
pp. 15–198. [Online]. Available: https://www.oecd-ilibrary.org/energy/
deploying-renewables-principles-for-effective-policies/
in-support-of-the-g8-plan-of-action_9789264042216-1-en
[33] E. Vaughan and J. Turner, “The Value and Impact of Building Codes | White
Papers | EESI,” 2013. [Online]. Available:
https://www.eesi.org/papers/view/the-value-and-impact-of-building-codes
[34] R. A. Enker and G. M. Morrison, “The potential contribution of building codes
to climate change response policies for the built environment,” Energy
Efficiency, vol. 13, no. 4, pp. 789–807, Apr. 2020. [Online]. Available:
https://doi.org/10.1007/s12053-020-09871-7

53
[35] K. Andreanidou, P. Barbosa, P. Bertoldi, M. Follador, R. Glancy, S. Kilkis,
A. Kona, N. Labanca, O. Lah, V. Palermo, S. Rivas, T. Serrenho, P. Zancanella,
P. Zangheri, European Commission, and Joint Research Centre, Guidebook
’How to develop a Sustainable Energy and Climate Action Plan (SECAP)’ Part
3, Part 3,, 2018, oCLC: 1111200056. [Online]. Available: http://publications.
europa.eu/publication/manifestation_identifier/PUB_KJNC29412ENN
[36] E. Harmon, “Improving Building Codes to Prepare for Climate Impacts,” Jan.
2019. [Online]. Available:
https://www.mapc.org/resource-library/building-codes-climate-2/
[37] A. Gouldson, S. Colenbrander, A. Sudmant, F. McAnulla, N. Kerr, P. Sakai,
S. Hall, E. Papargyropoulou, and J. Kuylenstierna, “Exploring the economic
case for climate action in cities,” Global Environmental Change, vol. 35, pp.
93–105, Nov. 2015. [Online]. Available:
https://linkinghub.elsevier.com/retrieve/pii/S0959378015300169
[38] L. Pérez-Lombard, J. Ortiz, R. González, and I. R. Maestre, “A review of
benchmarking, rating and labelling concepts within the framework of building
energy certification schemes,” Energy and Buildings, vol. 41, no. 3, pp.
272–278, Mar. 2009. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S037877880800220X
[39] “LEED: Better Buildings Are Our Legacy.” [Online]. Available:
http://leed.usgbc.org/
[40] “BREEAM: the worldâs leading sustainability assessment method for
masterplanning projects, infrastructure and buildings.” [Online]. Available:
https://www.breeam.com/
[41] W. Chung, “Review of building energy-use performance benchmarking
methodologies,” Applied Energy, vol. 88, no. 5, pp. 1470–1479, May 2011.
[Online]. Available:
http://www.sciencedirect.com/science/article/pii/S030626191000485X
[42] M. Kavgic, A. Mavrogianni, D. Mumovic, A. Summerfield, Z. Stevanovic, and
M. Djurovic-Petrovic, “A review of bottom-up building stock models for
energy consumption in the residential sector,” Building and Environment,
vol. 45, no. 7, pp. 1683–1697, Jul. 2010. [Online]. Available:
http://linkinghub.elsevier.com/retrieve/pii/S0360132310000338
[43] C. J. Cleveland, P. Fox-Penner, M. J. Walsh, M. Cherne-Hendrick, S. Gopal,
J. R. Castigliego, T. Perez, A. Pollack, K. Zheng, R. Perry, L. Hurley,
O. Simpson, J. Atkins, V. Gupta, M. Hatfield, and R. McGuinness, “Institute
for Sustainable Energy (ISE), Boston University Project Team,” p. 120.
[44] L. G. Swan and V. I. Ugursal, “Modeling of end-use energy consumption in
the residential sector: A review of modeling techniques,” Renewable
sustainable energy reviews, vol. 13, no. 8, pp. 1819–1835, 2009.
[45] K. Fabbri and V. Tarabusi, “Top-down and Bottom-up Methodologies for
Energy Building Performance Evaluation at Meso-scale Level-A Literature
Review,” Journal of Civil Engineering and Architecture Research, vol. 1, no.
5, 2014, p. 17.
[46] W. Li, Y. Zhou, K. Cetin, J. Eom, Y. Wang, G. Chen, and X. Zhang, “Modeling
urban building energy use: A review of modeling approaches and procedures,”
Energy, vol. 141, pp. 2445–2457, Dec. 2017. [Online]. Available:

54
 https://linkinghub.elsevier.com/retrieve/pii/S0360544217319291
[47] S. Fathi, R. Srinivasan, A. Fenner, and S. Fathi, “Machine learning
applications in urban building energy performance forecasting: A systematic
review,” Renewable and Sustainable Energy Reviews, vol. 133, 2020.
[48] N. Abbasabadi and M. Ashayeri, “Urban energy use modeling methods and
tools: A review and an outlook,” Building and Environment, vol. 161, p.
106270, Aug. 2019. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0360132319304809
[49] A. Nutkiewicz, Z. Yang, and R. K. Jain, “Data-driven urban energy simulation
(due-s): A framework for integrating engineering simulation and machine
learning methods in a multi-scale urban energy modeling workflow,” Applied
Energy, vol. 225, pp. 1176 – 1189, 2018. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0306261918307165
[50] Z. Yang, J. Roth, and R. K. Jain, “DUE-b: Data-driven urban energy
benchmarking of buildings using recursive partitioning and stochastic frontier
analysis,” vol. 163, pp. 58–69. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778817320868
[51] S. Lo, B. Norton, and A. Mannis, “Domestic energy use and air quality; a case
study of the city of Belfast,” Applied Energy, vol. 68, no. 1, pp. 1–18, Jan.
2001. [Online]. Available:
http://linkinghub.elsevier.com/retrieve/pii/S0306261900000441
[52] P. Nageler, G. Schweiger, H. Schranzhofer, T. Mach, R. Heimrath, and
C. Hochenauer, “Novel method to simulate large-scale thermal city models,”
Energy, vol. 157, pp. 633 – 646, 2018. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0360544218310363
[53] C. Cerezo Davila, C. F. Reinhart, and J. L. Bemis, “Modeling boston: A
workflow for the efficient generation and maintenance of urban building
energy models from existing geospatial datasets,” Energy, vol. 117, pp. 237 –
250, 2016. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0360544216314918
[54] R. Nouvel, A. Mastrucci, U. Leopold, O. Baume, V. Coors, and U. Eicker,
“Combining gis-based statistical and engineering urban heat consumption
models: Towards a new framework for multi-scale policy support,” Energy and
Buildings, vol. 107, pp. 204 – 212, 2015. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778815302061
[55] N. Ghiassi and A. Mahdavi, “Reductive bottom-up urban energy computing
supported by multivariate cluster analysis,” Energy and Buildings, vol. 144,
pp. 372–386, Jun. 2017. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778817307508
[56] F. Biljecki, J. Stoter, H. Ledoux, S. Zlatanova, and A. Çöltekin, “Applications
of 3D City Models: State of the Art Review,” ISPRS International Journal of
Geo-Information, vol. 4, no. 4, pp. 2842–2889, Dec. 2015. [Online]. Available:
https://www.mdpi.com/2220-9964/4/4/2842
[57] G. Peronato, J. Kämpf, E. Rey, and M. Andersen, “Integrating urban energy
simulation in a parametric environment: a Grasshopper interface for CitySim,”
in Proceedings of PLEA 2017, Edinburgh, 2017. [Online]. Available:
https://infoscience.epfl.ch/record/228832?ln=en

55
[58] R. Kaden and T. H. Kolbe, “City-wide total energy demand estimation of
buildings using semantic 3d city models and statistical data,” ISPRS Annals of
Photogrammetry, Remote Sensing and Spatial Information Sciences, vol.
II-2/W1, pp. 163–171, Sep. 2013. [Online]. Available: http://www.
isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/II-2-W1/163/2013/
[59] “Lantmäteriet.” [Online]. Available: https://www.lantmateriet.se/
[60] R. Ma, C. Geng, Z. Yu, J. Chen, and X. Luo, “Modeling city-scale building
energy dynamics through inter-connected distributed adjacency blocks,”
Energy and Buildings, vol. 202, p. 109391, 2019. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778819311831
[61] A. Schaefer and E. Ghisi, “Method for obtaining reference buildings,” Energy
and Buildings, vol. 128, pp. 660 – 672, 2016. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778816305977
[62] C.-K. Wang, S. Tindemans, C. Miller, G. Agugiaro, and J. Stoter, “Bayesian
calibration at the urban scale: a case study on a large residential heating
demand application in Amsterdam,” Journal of Building Performance
Simulation, vol. 13, no. 3, pp. 347–361, May 2020. [Online]. Available:
https://www.tandfonline.com/doi/citedby/10.1080/19401493.2020.1729862
[63] A. T. Booth, R. Choudhary, and D. J. Spiegelhalter, “Handling uncertainty in
housing stock models,” Building and Environment, vol. 48, pp. 35–47, Feb.
2012. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0360132311002599
[64] J. Sokol, C. Cerezo Davila, and C. F. Reinhart, “Validation of a Bayesian-based
method for defining residential archetypes in urban building energy models,”
Energy and Buildings, vol. 134, pp. 11–24, Jan. 2017. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S037877881631372X
[65] C. S. Monteiro, A. Pina, C. Cerezo, C. Reinhart, and P. Ferrão, “The Use of
Multi-detail Building Archetypes in Urban Energy Modelling,” Energy
Procedia, vol. 111, pp. 817–825, Mar. 2017. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S1876610217302771
[66] A. A. Famuyibo, A. Duffy, and P. Strachan, “Developing archetypes for
domestic dwellings-An Irish case study,” Energy and Buildings, vol. 50, pp.
150–157, Jul. 2012. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778812001818
[67] O. Pasichnyi, J. Wallin, and O. Kordas, “Data-driven building archetypes for
urban building energy modelling,” Energy, vol. 181, pp. 360–377, 2019.
[68] G. Tardioli, R. Kerrigan, M. Oates, J. O’Donnell, and D. P. Finn,
“Identification of representative buildings and building groups in urban
datasets using a novel pre-processing, classification, clustering and predictive
modelling approach,” Building and Environment, vol. 140, Aug. 2018.
[Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0360132318303019
[69] M. H. Kristensen, R. E. Hedegaard, and S. Petersen, “Hierarchical calibration
of archetypes for urban building energy modeling,” Energy and Buildings, vol.
175, pp. 219–234, Sep. 2018. [Online]. Available:
https://linkinghub.elsevier.com/retrieve/pii/S0378778818312532
[70] V. S. K. V. Harish and A. Kumar, “A review on modeling and simulation of

56
 building energy systems,” Renewable and Sustainable Energy Reviews,
vol. 56, Apr. 2016. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S1364032115014239
[71] Y. Chen, T. Hong, and M. A. Piette, “City-Scale Building Retrofit Analysis: A
Case Study using CityBES,” Building Simulation, p. 8, 2017.
[72] D. Wang, J. Landolt, G. Mavromatidis, K. Orehounig, and J. Carmeliet,
“CESAR: A bottom-up building stock modelling tool for Switzerland to
address sustainable energy transformation strategies,” Energy and Buildings,
vol. 169, pp. 9–26, Jun. 2018. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778817337696
[73] D. Robinson, F. Haldi, J. Kämpf, P. Leroux, D. Perez, A. Rasheed, and
U. Wilke, “CITYSIM: Comprehensive micro-simulation of resource flows for
sustainable urban planning,” p. 8.
[74] J. A. Fonseca and A. Schlueter, “Integrated model for characterization of
spatiotemporal building energy consumption patterns in neighborhoods and
city districts,” Applied Energy, vol. 142, pp. 247–265, Mar. 2015. [Online].
Available:
http://www.sciencedirect.com/science/article/pii/S0306261914013257
[75] E. Delzendeh, S. Wu, A. Lee, and Y. Zhou, “The impact of occupantsâ
behaviours on building energy analysis: A research review,” Renewable and
Sustainable Energy Reviews, vol. 80, pp. 1061–1071, Dec. 2017. [Online].
Available:
http://www.sciencedirect.com/science/article/pii/S1364032117309061
[76] L. Leroy, S. Letellier-Duchesne, and M. Kummert, “Using Model Calibration
To Improve Urban Modeling,” Rome, Italy, pp. 3531–3539. [Online].
Available: http://www.ibpsa.org/proceedings/BS2019/BS2019_210707.pdf
[77] X. Fu, W. Tian, Y. Sun, C. Zhu, and B. Yin, “Uncertainty Analysis of Urban
Building Energy Based on Two-Dimensional Monte Carlo Method,” in
Proceedings of the 11th International Symposium on Heating, Ventilation and
Air Conditioning (ISHVAC 2019), ser. Environmental Science and
Engineering, Z. Wang, Y. Zhu, F. Wang, P. Wang, C. Shen, and J. Liu, Eds.
Singapore: Springer, 2020, pp. 1315–1323.
[78] S. Heiple and D. J. Sailor, “Using building energy simulation and geospatial
modeling techniques to determine high resolution building sector energy
consumption profiles,” Energy and Buildings, vol. 40, no. 8, pp. 1426–1436,
Jan. 2008. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778808000200
[79] M. H. Kristensen, R. Choudhary, R. H. Pedersen, and S. Petersen, “Bayesian
Calibration Of Residential Building Clusters Using A Single Geometric
Building Representation,” p. 10, 2017.
[80] S. Nagpal, J. Hanson, and C. Reinhart, “Auto-Calibrated Urban Building
Energy Models as Continuous Planning Tools,” p. 5.
[81] S. Nagpal, C. Mueller, A. Aijazi, and C. F. Reinhart, “A methodology for
auto-calibrating urban building energy models using surrogate modeling
techniques,” Journal of Building Performance Simulation, pp. 1–16, Apr.
2018. [Online]. Available:
https://www.tandfonline.com/doi/full/10.1080/19401493.2018.1457722

57
[82] K. Sun, T. Hong, S. C. Taylor-Lange, and M. A. Piette, “A pattern-based
automated approach to building energy model calibration,” Applied Energy,
vol. 165, pp. 214–224, Mar. 2016. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0306261915015998
[83] P. Remmen, M. Lauster, M. Mans, M. Fuchs, T. Osterhage, and D. MÃ 41 ller,
“TEASER: an open tool for urban energy modelling of building stocks,”
Journal of Building Performance Simulation, vol. 11, no. 1, pp. 84–98, Jan.
2018. [Online]. Available:
https://www.tandfonline.com/doi/full/10.1080/19401493.2017.1283539
[84] J. Clarke, Energy Simulation in Building Design, 1st ed., 1985.
[85] Y. J. Huang, J. Broderick, and R. International, “A Bottom-Up Engineering
Estimate of the Aggregate Heating and Cooling Loads of the Entire U.S.
Building Stock,” p. 15.
[86] “DOE2.com Home Page.” [Online]. Available: http://www.doe2.com/
[87] A. Parekh, “Development of archetypes of building characteristcs libraries for
simplified energy use evalaution of houses,” p. 8.
[88] D. Robinson, N. Campbell, W. Gaiser, K. Kabel, A. Le-Mouel, N. Morel,
J. Page, S. Stankovic, and A. Stone, “Suntool - a new modelling paradigm for
simulating and optimising urban sustainability,” vol. 81, no. 9, pp. 1196–1211.
[Online]. Available:
http://linkinghub.elsevier.com/retrieve/pii/S0038092X0700120X
[89] J. H. Kämpf and D. Robinson, “Optimisation of urban energy demand using an
evolutionary algorithm,” p. 6.
[90] D. Thomas, C. Miller, J. Kampf, and A. Schlueter, “Multiscale co-simulation
of energyplus and citysim models derived from a building information model,”
p. 8.
[91] “HOME.” [Online]. Available: http://stafu.de/en/home.html
[92] J. A. Fonseca, T.-A. Nguyen, A. Schlueter, and F. Marechal, “City Energy
Analyst (CEA): Integrated framework for analysis and optimization of
building energy systems in neighborhoods and city districts,” Energy and
Buildings, vol. 113, pp. 202–226, Feb. 2016. [Online]. Available:
https://linkinghub.elsevier.com/retrieve/pii/S0378778815304199
[93] C. F. Reinhart, T. Dogan, J. A. Jakubiec, T. Rakha, and A. Sang, “Umi - An
urban simulation environment for building energy use, daylighting and
walkability,” 2013. [Online]. Available:
http://www.ibpsa.org/proceedings/BS2013/p_1404.pdf
[94] U. Ali, M. H. Shamsi, C. Hoare, E. Mangina, and J. O’Donnell, “A data-driven
approach for multi-scale building archetypes development,” Energy and
Buildings, vol. 202, p. 109364, Nov. 2019. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778819306553
[95] S. Torabi Moghadam, S. Coccolo, G. Mutani, P. Lombardi, J.-L. Scartezzini,
and D. Mauree, “A new clustering and visualization method to evaluate urban
heat energy planning scenarios,” Cities, vol. 88, pp. 19–36, May 2019.
[Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0264275118307625
[96] O. Pasichnyi, J. Wallin, F. Levihn, H. Shahrokni, and O. Kordas, “Energy
performance certificates - New opportunities for data-enabled urban energy

58
 policy instruments?” Energy Policy, vol. 127, pp. 486–499, Apr. 2019.
[Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0301421518307894
[97] I. Nolte, “Energy Performance Certificates across Europe - From design to
implementation,” Dec. 2010. [Online]. Available:
https://www.buildup.eu/en/practices/publications/
energy-performance-certificates-across-europe-design-implementation
[98] C. Ahern and B. Norton, “A generalisable bottom-up methodology for
deriving a residential stock model from large empirical databases,” Energy and
Buildings, vol. 215, p. 109886, May 2020. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778819316652
[99] M. Österbring, E. Mata, L. Thuvander, M. Mangold, F. Johnsson, and
H. Wallbaum, “A differentiated description of building-stocks for a
georeferenced urban bottom-up building-stock model,” Energy and Buildings,
vol. 120, pp. 78–84, May 2016. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S037877881630216X
[100] O. Pasichnyi, F. Levihn, H. Shahrokni, J. Wallin, and O. Kordas, “Data-driven
strategic planning of building energy retrofitting: The case of Stockholm,”
Journal of Cleaner Production, vol. 233, pp. 546–560, 2019.
[101] M. Mangold, M. Österbring, and H. Wallbaum, “Handling data uncertainties
when using Swedish energy performance certificate data to describe energy
usage in the building stock,” Energy and Buildings, vol. 102, pp. 328–336,
Sep. 2015. [Online]. Available:
https://linkinghub.elsevier.com/retrieve/pii/S0378778815300207
[102] J. von Platten, C. Holmberg, M. Mangold, T. Johansson, and K. Mjörnell, “The
renewing of Energy Performance Certificates-Reaching comparability between
decade-apart energy records,” Applied Energy, vol. 255, 2019.
[103] Y. Li, S. Kubicki, A. Guerriero, and Y. Rezgui, “Review of building energy
performance certification schemes towards future improvement,” Renewable
and Sustainable Energy Reviews, vol. 113, 2019.
[104] J. Widén and E. Wäckelgård, “A high-resolution stochastic model of domestic
activity patterns and electricity demand,” Applied Energy, vol. 87, no. 6, Jun.
2010. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0306261909004930
[105] D. Fischer, A. Härtl, and B. Wille-Haussmann, “Model for electric load
profiles with high time resolution for German households,” Energy and
Buildings, vol. 92, pp. 170–179, Apr. 2015. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778815000845
[106] D. Fischer, T. Wolf, J. Scherer, and B. Wille-Haussmann, “A stochastic
bottom-up model for space heating and domestic hot water load profiles for
German households,” Energy and Buildings, vol. 124, pp. 120–128, Jul. 2016.
[Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778816303358
[107] E. McKenna and M. Thomson, “High-resolution stochastic integrated
thermal-electrical domestic demand model,” Applied Energy, vol. 165, pp.
445–461, Mar. 2016. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0306261915016621

59
[108] G. Happle, J. A. Fonseca, and A. Schlueter, “A review on occupant behavior in
urban building energy models,” Energy and Buildings, vol. 174, pp. 276–292,
Sep. 2018. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778817333583
[109] J. Keirstead and A. Sivakumar, “Using Activity-Based Modeling to Simulate
Urban Resource Demands at High Spatial and Temporal Resolutions,” Journal
of Industrial Ecology, vol. 16, no. 6, pp. 889–900, 2012. [Online]. Available:
https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1530-9290.2012.00486.x
[110] E. Barbour, C. C. Davila, S. Gupta, C. Reinhart, J. Kaur, and M. C. González,
“Planning for sustainable cities by estimating building occupancy with mobile
phones,” Nature Communications, vol. 10, no. 1, p. 3736, Aug. 2019. [Online].
Available: https://www.nature.com/articles/s41467-019-11685-w
[111] W. Wu, B. Dong, Q. R. Wang, M. Kong, D. Yan, J. An, and Y. Liu, “A novel
mobility-based approach to derive urban-scale building occupant profiles and
analyze impacts on building energy consumption,” Applied Energy, vol. 278,
p. 115656, Nov. 2020. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0306261920311545
[112] M. Mosteiro Romero, I. Hischier, J. Fonseca, and A. Schlueter, “A novel
population-based occupancy modeling approach for district-scale simulations
compared to standard-based methods,” Building and Environment, vol. 181, p.
107084, Aug. 2020.
[113] G. Happle, J. A. Fonseca, and A. Schlueter, “Context-specific urban occupancy
modeling using location-based services data,” Building and Environment, vol.
175, p. 106803, May 2020. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S036013232030161X
[114] M. Åberg, J. Carlsson, and A. M. Nilsson, “Improved efficiency for
distribution and use of district heating: A simulation study of retrofitting a
Swedish apartment complex from the 1970’s,” Journal of Building
Engineering, vol. 20, pp. 559–568, Nov. 2018. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S2352710218307307
[115] “HSB 53 Gräslöken.” [Online]. Available:
https://www.hsb.se/uppsala/brf/grasloken/
[116] “Vattenfall Heat Home Page.” [Online]. Available:
https://www.vattenfall.se/fjarrvarme/
[117] “Uppsalahem.” [Online]. Available: https://www.uppsalahem.se/
[118] C. Björk, L. Reppen, and P. Kallstenius, Så byggdes husen 1880-2000:
arkitektur, konstruktion och material i våra flerbostadshus under 120 år, ny
utg. ed. Stockholm: Svensk Byggtjänst, 2013.
[119] “TABULA WebTool.” [Online]. Available:
http://webtool.building-typology.eu/#bm
[120] J. P. Holman and P. R. S. White, Heat transfer. London: McGraw-Hill, 1992,
oCLC: 28171203.
[121] X. Q. Li, Y. Chen, J. D. Spitler, and D. Fisher, “Applicability of calculation
methods for conduction transfer function of building constructions,”
International Journal of Thermal Sciences, vol. 48, no. 7, pp. 1441–1451, Jul.
2009. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S1290072908002299

60
[122] J. E. Seem, “Modeling of Heat Transfer in Buildings,” 1987. [Online].
Available: https://minds.wisconsin.edu/handle/1793/7916
[123] M. Manfren, P. Caputo, and G. Costa, “Paradigm shift in urban energy systems
through distributed generation: Methods and models,” Applied Energy, vol. 88,
no. 4, pp. 1032–1048, Apr. 2011. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0306261910004204
[124] “EnergyPlus | EnergyPlus.” [Online]. Available: https://energyplus.net/
[125] “TRNSYS : Transient System Simulation Tool.” [Online]. Available:
http://www.trnsys.com/
[126] “IDA ICE - Simulation Software | EQUA.” [Online]. Available:
https://www.equa.se/en/ida-ice
[127] “ASHRAE Handbook Online.” [Online]. Available: https://www.ashrae.org/
technical-resources/ashrae-handbook/ashrae-handbook-online
[128] R. Perez, P. Ineichen, R. Seals, J. Michalsky, and R. Stewart, “Modeling
daylight availability and irradiance components from direct and global
irradiance,” Solar Energy, vol. 44, no. 5, pp. 271–289, Jan. 1990. [Online].
Available:
http://www.sciencedirect.com/science/article/pii/0038092X9090055H
[129] J. A. Duffie and W. A. Beckman, Solar Engineering of Thermal Processes, 4th
Edition, 4th ed. Wiley, 2013.
[130] G. P. Mitalas and J. G. Arseneault, “Fortran IV program to calculate z-transfer
functions for the calculation of transient heat transfer through walls and roofs.”
p. 41.
[131] B. Gebhart, Heat Transfer. McGraw-Hill, 1971.
[132] “ISO 13370:2017.” [Online]. Available: https://www.iso.org/cms/render/live/
en/sites/isoorg/contents/data/standard/06/57/65716.html
[133] “3D Design Software | 3D Modeling on the Web.” [Online]. Available:
https://www.sketchup.com/page/homepage
[134] P. Nageler, G. Schweiger, M. Pichler, D. Brandl, T. Mach, R. Heimrath,
H. Schranzhofer, and C. Hochenauer, “Validation of dynamic building energy
simulation tools based on a real test-box with thermally activated building
systems (TABS),” Energy and Buildings, vol. 168, pp. 42–55, Jun. 2018.
[Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778818307424
[135] “MIT Sustainable Design Lab.” [Online]. Available:
http://web.mit.edu/sustainabledesignlab/index.html
[136] “Boverket.” [Online]. Available: https://www.boverket.se/
[137] J. Widén, M. Lundh, I. Vassileva, E. Dahlquist, K. Ellegård, and
E. Wäckelgård, “Constructing load profiles for household electricity and hot
water from time-use data-Modelling approach and validation,” Energy and
Buildings, vol. 41, no. 7, pp. 753–768, Jul. 2009. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778809000413
[138] J. Widén, A. M. Nilsson, and E. Wäckelgård, “A combined Markov-chain and
bottom-up approach to modelling of domestic lighting demand,” Energy and
Buildings, vol. 41, no. 10, pp. 1001–1012, Oct. 2009. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778809000978
[139] M. H. Kristensen, A. Brun, and S. Petersen, “Predicting Danish residential

61
 heating energy use from publicly available building characteristics,” Energy
and Buildings, vol. 173, pp. 28–37, Aug. 2018. [Online]. Available:
https://linkinghub.elsevier.com/retrieve/pii/S0378778818300951
[140] G. Dall’O’, A. Galante, and M. Torri, “A methodology for the energy
performance classification of residential building stock on an urban scale,”
Energy and Buildings, vol. 48, pp. 211–219, May 2012. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778812000515
[141] L. Frayssinet, F. Kuznik, J.-L. Hubert, M. Milliez, and J.-J. Roux, “Adaptation
of building envelope models for energy simulation at district scale,” Energy
Procedia, vol. 122, pp. 307–312, Sep. 2017. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S1876610217329302
[142] R. Nouvel, A. Mastrucci, U. Leopold, O. Baume, V. Coors, and U. Eicker,
“Combining GIS-based statistical and engineering urban heat consumption
models: Towards a new framework for multi-scale policy support,” Energy and
Buildings, vol. 107, pp. 204–212, Nov. 2015. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778815302061
[143] R. Buffat, A. Froemelt, N. Heeren, M. Raubal, and S. Hellweg, “Big data GIS
analysis for novel approaches in building stock modelling,” Applied Energy,
vol. 208, pp. 277–290, Dec. 2017. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S030626191731454X
[144] J. H. Kämpf and D. Robinson, “A simplified thermal model to support analysis
of urban resource flows,” Energy and Buildings, vol. 39, no. 4, pp. 445–453,
Apr. 2007. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778806002192
[145] M. P. Guillaumet, P. Borges, M. Rosas-Casals, and O. Travesset-Baro,
“Building archetypes in urban energy models. A comparative case study of
deterministic and statistical methods in Andorra.” p. 10.

62