10 5445ir1000149384 PDF
10 5445ir1000149384 PDF
10 5445ir1000149384 PDF
DOKTOR-INGENIEURS
genehmigte
DISSERTATION
von MSc.
Alexandru N ICHERSU
Karlsruhe 2022
iii
Declaration of Authorship
I, Alexandru N ICHERSU, declare that this thesis titled, “Scale aware modeling and
monitoring of the urban energy chain” and the work presented in it are my own. I
confirm that:
• This work was done wholly or mainly while in candidature for a research de-
gree at this University.
• Where any part of this thesis has previously been submitted for a degree or
any other qualification at this University or any other institution, this has been
clearly stated.
• Where I have consulted the published work of others, this is always clearly
indicated.
• Where I have quoted from the work of others, the source is always given. With
the exception of such quotations, this thesis is entirely my own work.
• I have acknowledged all main sources of help.
• Where the thesis is based on work done by myself jointly with others, I have
made clear exactly what was done by others and what I have contributed my-
self.
Signed:
Date:
v
“If you really want to affect environmental outcomes it’s not shaping a building that matters,
it’s shaping a community”
“We ourselves feel that what we are doing is just a drop in the ocean. But the ocean would
be less because of that missing drop.”
Abstract
Department of Civil Engineering, Geo and Environmental Sciences
Doctor of Engineering
by Alexandru N ICHERSU
Keywords: smart cities, 3D semantical city models, spatial aware energy data, inter-
operability, UBEM, IoT, spatial actors, spatial scales, spatial standards, urban energy
management, urban planning, city-wide energy chain, spatial awareness
With energy modeling at different complexity levels for smart cities and the con-
current data availability revolution from connected devices, a steady surge in de-
mand for spatial knowledge has been observed in the energy sector. This transfor-
mation occurs in population centers focused on efficient energy use and quality of
life. Energy-related services play an essential role in this mix, as they facilitate or
interact with all other city services. This trend is primarily driven by the current
age of the Ger.: Energiewende or energy transition, a worldwide push towards renew-
able energy sources, increased energy use efficiency, and local energy production
that requires precise estimates of local energy demand and production. This shift
in the energy market occurs as the world becomes aware of human-induced climate
change, to which the building stock has a significant contribution (40% in the Euro-
pean Union [71]). At the current rate of refurbishment and building replacement, of
the buildings existing in 2050 in the European Union, 75% would not be classified
as energy-efficient, [70]. That means that substantial structural change in the built
environment and the energy chain is required to achieve EU-wide goals concern-
ing environmental and energy policy. These objectives provide strong motivation
for this thesis’ work and are generally made possible by energy monitoring and
modeling activities that estimate the urban energy needs and quantify the impact of
refurbishment measures.
To this end, a modeling library called aEneAs was developed in the scope of this
thesis that can perform city-wide building energy modeling. The library performs
its tasks at the level of a single building and was a first in its field, using standard-
ized spatial energy data structures that allow for portability from one city to another.
For data input, extensive use was made of digital twins provided from CAD, BIM,
GIS, architectural models, and a plethora of energy data sources. The library first
quantifies primary thermal energy demand and then the impact of refurbishment
measures. Lastly, it estimates the potential of renewable energy production from
solar radiation. aEneAs also includes network modeling components that consider
energy distribution in the given context, showing a path toward data modeling and
viii
simulation required for distributed energy production at the neighborhood and dis-
trict level.
In order to validate modeling activities in solar radiation and green façade and roof
installations, six spatial models were coupled with sensor installations. These dig-
ital twins are included in three experiments that highlight this monitoring side of
the energy chain and portray energy-related use cases that utilize the spatially en-
abled web services SOS-SES-WNS, SensorThingsAPI, and FIWARE. To this author’s
knowledge, this is the first work that surveys the capabilities of these three solutions
in a unifying context, each having its specific design mindset.
The modeling and monitoring activity and their corresponding literature review in-
dicated gaps in scientific knowledge concerning data science in urban energy mod-
eling. First, a lack of standardization regarding the spatial scales at which data is
stored and used in urban energy modeling was observed. In order to identify the
appropriate spatial levels for modeling and data aggregation, scale is explored in-
depth in the given context and defined as a byproduct of resolution and extent,
with ranges provided for both parameters. To that end, a survey of the encountered
spatial scales and actors in six different geographical and cultural settings was per-
formed. The information from this survey was used to put forth a standardized spa-
tial scales definition and create a scale-dependent ontology for use in urban energy
modeling. The ontology also provides spatially enabled persistent identifiers that
resolve issues encountered with object relationships in modeling for inheritance, de-
pendency, and association. The same survey also reveals two significant issues with
data in urban energy modeling. These are data consistency across spatial scales and
urban fabric contiguity. The impact of these issues and different solutions such as
data generalization are explored in the thesis.
Further advancement of scientific knowledge is provided specifically with spatial
standards and spatial data infrastructure in urban energy modeling. A review of
use cases in the urban energy chain and a taxonomy of the standards were carried
out. These provide fundamental input for another piece of this thesis: inclusive soft-
ware architecture methods that promote data integration and allow for external con-
nectivity to modern and legacy systems. In order to reduce time-costly extraction,
transformation, and load processes, databases and web services to ferry data to and
from separate data sources were used. As a result, the spatial models become central
linking elements of the different types of energy-related data in a novel perspective
that differs from the traditional one, where spatial data tends to be non-interoperable
/ not linked with other data types. These distinct data fusion approaches provide
flexibility in an energy chain environment with inconsistent data structures and soft-
ware. Furthermore, the knowledge gathered from the experiments presented in this
thesis is provided as a synopsis of good practices.
ix
Zusammenfassung
Fakultät für Bauingenieur-, Geo- und Umweltwissenschaften
Doktor-Ingineurs
by Alexandru N ICHERSU
Acknowledgements
I want to acknowledge my family’s unwavering support throughout the entire life of
this project. You are my lighthouse; I’d be lost without you. Thank you for guiding
my voyages.
The most significant influence on my development has come from my parents, my
mom, a mechanical engineer, and my dad, a Ph.D. in geodesy. Both provided me
with endless love and persistent stability. Since the very start, they have been a
source of strength and inspiration in my life. They have taught me priceless lessons,
from decency to modesty, geometry to French, and the value of teamwork to the sig-
nificance of family. In addition, my friend, teammate, ally, collaborator, companion,
and twin sister, who has known me ever since I was smaller than a raindrop, was
always there, guided by her outstanding determination. She offered support and
different perspectives, as twins always do, and her career as a Ph.D. Civil Engineer
gifted me with another path of inspiration.
My dear sweetheart, partner, love, companion, and beloved wife provided me with a
constant source of support and encouragement even as she moved to another coun-
try, grappled with recognizing her titles to be allowed to practice Orthodontics in a
new language (not a feat for the fainthearted) and pursued, parallel to my own, her
Ph.D. in Medicine. To myself and all who know her, she is a fountain of perseverance
and integrity and played a large part in bringing this thesis’ to the printer.
We traveled the world, had two kids, and completed our PhDs.
For the duration of my employment at EIFER, I have actively participated in con-
structing a city simulation platform with a broad goal of urban simulation. The
project, named CURTIS (Coupled URban SimulaTIon), financed by the EDF (Élec-
tricité de France) group, successfully built and delivered a city simulation platform.
This has supported activities (meetings, conferences, publications, and student the-
sis supervision) in this Ph.D. For that, the author is ever grateful to the EIFER man-
agement: Dr. Jean Copreaux, Dr. Roman Zorn, Dr. Nurten Avci and Ludmila
Gautier, the EDF management: Fabrice Casciani and Maxime Cassat, the project
management: Dr. Kevin McKoen, David Blin and Dr. Alberto Pasanisi, my project
colleagues: Alexander Simons, Dr. Jochen Wendel, Dr. Sebastien Cajot, Dr. Samuel
Thiriot, Dr. Daniel Fehrenbach, Dr. Jonathan van der Kamp, Francisco Marzabal,
Isaac Boates, Jason Yip, Manfred Wieland, Dr. Maria Sipowicz, Dr. Atom Mirakyan,
Dr. Syed Monjur Murshed, Omar Benhamid, Saed Muhamad and Wanji Zhu, the
partners at ULPGC: Pablo Fernández and Jaisiel Santana and my students: Aleksan-
dra Gabryjalowicz, Chenfeng Liu, Yao Jiacheng and Thibault Morin.
Sincere thanks go to the thesis supervisors for their assistance. The GRACE school
system had a system of three persons helping and guiding, each with their fair share
of work and associate contribution. Prof. Dr. Stefan Hinz, head of the IPF (Institut
für Photogrammetrie und Fernerkundung), is the supervisor in charge and the main
guarantor responsible for the work presented here. On behalf of EIFER, my former
group manager from EIFER, Dr.-Ing. Andreas Koch made this thesis possible from
an administrative and managerial perspective. Both Dr. Koch and Dr. Hinz saw the
potential of collaboration between IPF and EIFER and helped create the structure
and support that this Ph.D. candidate benefited from in his work. For this, I am very
much grateful.
xii
Contents
Abstract vii
Acknowledgements xi
Contents xiii
1 Introduction 1
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Research questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Outline of the thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Why We . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Bibliography 141
xv
List of Figures
List of Tables
4.1 Use cases for 3D city models in the energy wide chain . . . . . . . . . . 45
4.2 BIM data formats used in the urban energy chain . . . . . . . . . . . . . 51
4.3 CityGML 1.0 and 2.0 level of detail concept scale and accuracy re-
quirements, adapted from [91] and [242] . . . . . . . . . . . . . . . . . . 54
4.4 CAD, BIM and GIS standards/methodologies summary . . . . . . . . 54
4.5 Utility network modeling solutions, after [34], [23] and own work,
with three support levels, 1 - poor, 2 - basic, 3 - good . . . . . . . . . . . 61
4.6 Standards/Methodologies information sheet, aggregated from own
work and [79] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.7 Sensors/things web service summary . . . . . . . . . . . . . . . . . . . 71
—————————————————————————————-
xix
List of Abbreviations
List of Symbols
Chapter 1
Introduction
1.1 Motivation
Cities are great centers of human capital concentration. They have always been
places that attract, gather, facilitate and drive innovation. To do this they help peo-
ple congregate and foster companies development. This happens as long as a sum of
factors transpire simultaneously. These factors are generally related to either geogra-
phy or governance. A successful city is one that provides good quality city services
that convince people to move in and stay. According to [175], one century ago 10%
of all people used to live in cities, where as today, that ratio has exceeded 50%. This
trend, slowly decelerated by today’s pandemic, will nevertheless accelerate towards
the mid of this century and is projected to stabilize in 2100 at around 85%. The pre-
viously cited report argues that cities have always existed and expanded because of
the economic growth and services that benefit their residents. These services pro-
vide for a better quality of life at a lower economic cost due to economies of scale.
These can only be achieved in urban areas with a corresponding human settlement
density. On the other side, large numbers of humans bundled up in cramped pop-
ulation centers bring about issues of its own. These are related to quality of life,
city services and space sufficiency which urban planners and managers have to deal
with. The indicators, [175], used for the comparison of cities are:
• Livability – clean, healthy living, digital infrastructure for city services
2 Chapter 1. Introduction
Smart cities
Most often these changes in city energy data management and data handling are
identified as a primordial requirement, which is part of the new generation cities
(often also referred to as smart cities) revolution. In interviews with different city
stakeholders (the people responsible for planning and management of cities), [104],
three depths of the smart city trend setters were identified:
• Tools-economic dimension
• Administrative-procedural dimension
• Governance dimension
From the city perspective, for change to happen there is a clear need to improve and
coordinate the collection, processing and connectivity of existing (and/or available)
information in order to achieve efficient resource management (tool/instrumental-
economic dimension). Also required are technical specifications for objectives and
actions for city development (administrative-procedural dimension) and new city
stakeholder interaction constellations (governance dimension). This thesis tackles
parts of the dimensions mentioned for next generation cities trying to offer an inte-
grated perspective from both a spatial but also an energy thematic viewpoint of the
city wide energy chain.
Both scientists and practitioners of the urban management have shown heightened
interest in the topic of smart cities, being very visible in standardization organiza-
tions. This standardization work is being carried by the OGC (Open Geospatial
Consortium) - with the Smart Cities WG (Working Group) and the ISO (International
Organization for Standardization) [111] with the Smart community infrastructures
1.1. Motivation 3
- Specification of multi-source urban data integration for smart city planning (SCP)
group. The work pushed by these communities is also joined by volunteer based
committees that work on standardization within the intertwining geospatial and en-
ergy domain, namely the Energy ADE, [22] and the Utility Network ADE, [1]. A
definition of the term smart city is provided in Chapter 2 as part of the trends that
ever changing cities are currently experiencing.
Spatial context
As concluded in [62], the spatial context has big weight in the city wide energy chain
and a spatial approach is fundamental. This helps build an exhaustive information
system concerning the energy performances of buildings in the urban context. This
in turn is used in modeling and simulation by energy providers to estimate energy
demand and quantify hourly and spatial variation.
The current proposed solution has been tried out in different geographical contexts.
While performing this analysis different organizational models of cities have been
observed as well as divergent citizens’ expectations. The implementation was tested
in distinct legislation and requirements under which companies operate. In this
thesis examples related to the following cities were provided: Singapore (Singapore),
Berlin (Germany), Geneva (Switzerland), New York (USA), Karlsruhe (Germany),
Torino (Italy) and Bucharest (Romania). These cities have different urban sprawls,
densities, historical and cultural backgrounds. This is reflected in the spatial and
temporal scales at which information is stored and organized.
1.4 Why We
In certain sentences and phrases of this thesis the pronoun We is used to describe
the authors work. The author of this work believes that such a manuscript is never
a single individuals work but it is always team work that helped build, model, di-
rect the methodologies, results, conclusions and findings. A similar example of this
philosophy is Diana Nyad, who swam, alone, from Cuba to Florida - the longest
recorded unassisted swim. She testifies that her work, a single individual in the wa-
ter, in what is an incredible feat of human resilience, is actually nothing else than
team work.
One can argue that even the main structure of support for this PhD student at the
University of Karlsruhe, (KIT - the Karlsruhe Institute of Technology), the GRACE
(GRACE - GRAduate school for Climate and Environment) program has a similar
perspective and recognizes the need for team work in achieving results like the the-
sis presented here. This student received support from three parties, trying to bal-
ance out the conflicting interests affecting the thesis work, be those scientific, work
or personal life related. These were: the supervisor in charge, the direct/personal
advisor and the group manager, and also institutional from three institutes, EIFER
(the European Institute For Energy Research), IPF (Institute for Photogrammetry and
Remote Sensing) and IAI (Institute for Automation and applied Informatics).
7
Chapter 2
With this dissertation study the focus is on a complex mechanism which relates to
public services delivery, namely production, delivery, storage and use of energy that
cities deliver to their inhabitants in an urban spatial environment. The current chap-
ter describes the city wide energy chain from the perspective that it is referred to in
this thesis. This envelope helps prepare the chapters that follow, supports the so-
lutions proposed and confines their scope. In order to define the city wide energy
chain the concepts in it are separated as follows: city, energy and chain. Different
definitions are explored for each of those 3 concepts and explained within the scope
of the methodological approach by defining its viable spatial and thematic context.
The urban sprawl has a minimum density of 1000 persons per square kilometer and
a minimum size set at 5000 individuals. Diversified energy networks are unattain-
able in rural areas due to larger distances and lower population density and, most
importantly, the associated costs. The thesis scope is further narrowed to the ad-
ministrative boundary of the spatial entities defined in the chain subsection be-
low.
• Oil
• Waste
• Air pressure
* when used for energy production ** utility networks usually integrate communi-
cation tools, or use communication networks for their own data transmission
Out of all energy goods delivered in cities, electricity is the most readily available
commodity to all population types. It is also the only one for which world wide
database numbers exist (United Nations, International Energy Agency and World
Bank), see 2.1. Specific to urban environments, according to the IEA 2017 data, [105],
as of 1991, 97% of the world urban population (or 2.28 billion urban dwellers) had
access to electricity. Since then, that percentage decreased for a decade (even though
in absolute numbers the value increase remained), a process most likely caused by
the main communist socialist block disintegration (with large numbers of people
leaving rural communities) and the lack of state investment that followed. However,
considering the accelerated urbanization trend mentioned in 1 and visible in the
numbers presented in table 2.1, the new arrivals to urban environments have moved
to cities that already had this resource readily available.
Moreover, regional and national governments strive to provide cities first with ac-
cess to reliable electricity. The reason is that the main factor in electrical network
development is population density, [24]. These trends and policies are further con-
firmed by numbers from 2017 where the same percentage of 97% of the world urban
population is shown to have access to electricity (while the relative number was con-
stant, the absolute value increased in the 25 years span that the table depicts). In any
case, this numbers make electricity the most readily available energy resource out of
all the types identified in the previous list.
F IGURE 2.2: The energy chain inside urban spaces with its compo-
nents and flows of energy, special notice to newly added flow direc-
tions
national/regional grid input, green the traditional chain components and yellow the
most recent additions to the city.
Energy chains help further define the spatial extent of the study areas. This the-
sis is purposely confined to exclusively study directly (physically) connected enti-
ties in the urban domain. For example, an electrical network connected from local
source to end user, represents the maximum extent of urban space as described in
this manuscript. This extent respects in most cases the administration boundaries
of cities but does not constitute a rule. As such, the approach has to consider both
administrative boundaries and those of energy utilities.
2.2 The ever changing city - trends in cities and the energy
chain
One cannot discuss the city wide energy chain and present solutions to today’s prob-
lems without understanding the current trends and disruption technologies affect-
ing the different fields and actors intertwining in this context. Furthermore, data
from different, intertwining fields is modelled: GIS, energy and IT. Very often do-
main experts propose solutions from their own field of expertise and perspective. In
this thesis a holistic approach that handles the connections in the most efficient and
non-partisan way possible is used.
changing the fabric of cities in real life not just digitally. In [63] a survey of existing
definitions was made. The study also proposed a comprehensive definition. A smart
city is:
• a well defined geographical area,
• governed by a well defined pool of subjects, able to state the rules and policy
for the city government and development.
• using high technologies such as ICT, logistic, energy production, and so on,
• creating benefits for citizens in terms of well being, inclusion and participation,
environmental quality, intelligent development;
Based on interviews and reviews presented in [104], the term Smart City is not differ-
ent in terms of substance that clearly differentiate it from other recognized naming
conventions such as Green Cities, Sustainable Cities or Low-Carbon Cities. What the
concept does achieve is a definition of the conversion process. Purposefully used by
decision makers in the public domain, it has built acceptance within the mainstream
and has thus allowed for funds to be dedicated to smart city investments (e.g. in-
frastructure investments).
In Europe, as far back as 2007, the cities associated together in the Covenant of May-
ors program, [44]. Among their declared aims are the improvement of energy effi-
ciency and increased renewable energy use inside cities. The authors of the previous
study, provide the alternate term intelligent cities and mention citizens QoL as the
most important indicator with which all initiatives should be rated. They look at
cultural, economic and social growth in a healthy, safe, stimulating and dynamic
environment as the components of positive effects in peoples lives from smart city
initiatives. From the authors perspective the central challenge for cities in this ini-
tiatives is the integration of new enabling infrastructures and IoT with pre-existing
structures while exploiting synergies and interoperability between systems. This
facilitates better value services for citizens, and contributes to an improved QOL.
Taking a stakeholder perspective in [16], Batty explains that interviews with deci-
sion makers and interested parties show that they are very much aware of and un-
derstand the smart city concept. They also see their own cities current situation as
one of competition with other cities. That means that in order to stay relevant, cities
need to adopt technologies, integrate them in their decision making process and use
them in order to provide a higher QoL that convinces citizens to reside in their city,
and companies to operate there. This is only possible when manpower and infras-
tructure combine to provide the right recipe for success.
More generally, digital transformation can be understood as the use of new and fre-
quently changing digital technology to solve problems often using cloud computing
while also enhancing capabilities of traditional software products. It is achieved
when the digital enables innovation and creativity and stimulates significant change
within the professional or knowledge domain, [120]. Specifically with regards to the
digital change that cities undergo, these can be structured in 4 pillars, [55], digital:
• infrastructure, the backbone that allows for public digital services to exist in
the first place, including WI-FI, 5G and broadband
• services for citizens, direct online citizen relationship to the public administra-
tion.
14 Chapter 2. Defining the city wide energy chain
• education, the restructuring of peoples skills so that all citizens can have basic
access to digital privileges
• skills valorification, or the building and development of projects and initia-
tives, to help spread the value creation from digital footprints both for compa-
nies and people
As with many profound societal changes that impact citizens lives, it takes time
for the term to be universally adopted and agreed to, so alternatives have emerged
to the term "smart city". The same term was used to describe very different and
sometimes seemingly not converging activities. Parts of the scientific community
choose to distance themselves from it, for example Ursula Eicker, one of the most
cited authors in the UBEM (Urban Building Energy Modeling) field, prefers the term
‘next-generation city’, and defines it by linking it to sustainability - a bridge that
unifies our common purpose with these transformation actions, [155].
For this present thesis a smart city and a next-generation city are one and the same.
They are defined as: a city that has embraced digital amenities and provides its inhabitants
and companies with digitally optimized and innovative public services. This is done in a
way that stimulates and enhances open data initiatives, direct citizen participation in city
management in order to guide resources allocation and change.
Smart districts
Smart districts are the embodiment of a smart city on a district scale. In the European
Union context, a knowledge platform was built to exchange data on the topic of
smart cities. Called the Smart Cities Information System (SCIS), it allows cities to
exchange data, experience and know-how required in order to achieve smart cities,
[68]. On the same platform, a definition of smart districts as spatial units that rely
on three pillars can be found. These are:
• to make efficient use of energy through smart grids,
• the inclusion of district heating and cooling distribution systems,
• to provide new mobility solutions and the use of ICT.
Using the three components mentioned above energy loads can be locally handled,
reduced and shifted from peak to off-peak hours. This maximizes energy load bal-
ancing and allows for the integration of renewable energies. Even though the def-
inition is very much "alive" in the sense that it changes as things progress in the
field, the current one gives us an idea as to the goals of the projects aiming to deploy
smart districts. These are most often projects that do so from an energy perspective
or at least include it in the proposed work. In addition the use of ICT can enable
energy providers and policy makers to use the monitoring data to guarantee a level
of transparency in between all stakeholders. This allows citizens to have an impact
and even take responsibility in the smart management of energy in their district. At
the same time it often leads to changing habits. In addition, largely automatized
systems can decrease administrative loads and diminish the costs associated with
network operation.
Direct energy generation and storage at building and district level is also identified
in [43] as one of the most important trends in local energy management. Benefiting
from lower costs and technological break troughs, direct generation from local re-
newable sources and energy storage continues to gain in presence. The emergence of
2.2. The ever changing city - trends in cities and the energy chain 15
these energy-efficient facilities and their integration into daily lives is what has been
identified as one of the most important tasks in achieving next generation cities. Re-
garding energy transportation, tendencies show that the micro grid and smart-grid
paradigms will become the standard in the electricity transportation domain.
The energy chain components at the level of a smart district can be categorized in 5
categories: producer, prosumer (an entity that both consumes and produces a com-
modity), energy storage, energy source and consumer. These categories and the
distinct entity types that are included in them are presented in 2.3.
domains to use open data. Of course, not all public data can become open data. The
intricate topic of privacy, with different restrictions set in various local, regional and
national contexts, requires constant attention and adjustments to data policy. This
process of data opening can be quite elaborate and time consuming.
However, open city data has its downsides too, and these come usually in terms of
quality and consistency, which affects data processes further down the processing
pipeline and most importantly the duration of the treatment required to make data
useful. In an attempt to tackle this issue, [169] defined and developed intricate ETL
(Extract Transform Load) processes with the help of semantics data and open API
called the CitySDK. However, this approach has still to gain notoriety and critical
implementation mass. This is why classical standard regression methods in combi-
nation with principal component analysis (PCA) used to improve the quality and
amount of predicted values are still a common tool within the data scientists com-
munity.
Keeping the advantages and disadvantages of open data in mind, one must rec-
ognize the increased legitimacy that it provides to city administrators and the fact
that it creates a tangible connection of citizens with their own city’ institutional rep-
resentatives, [27]. That is also concurred to by [177] who go further in analyzing
the impact of open data in Economy, Education, Energy, Environment, Governance,
Tourism and Transportation. Their results show the impact of data initiatives in the
5 smart cities sample they analyzed to be strongest in Governance and Economy.
The authors conclude from their analysis that such initiatives provide an inherent
innovation pattern which is most important in an “open innovation economy” that
brings together citizens, NGOs (Non Governmental Organizations) and companies.
Citizen observatories
The way in which cities are governed is changing in order to adapt to societal changes
in terms of life styles, to competition from other cities and to the integration of the
digital domain in daily lives. Citizens use a very diverse set of digital applications
in their daily lives and expect that their city administrator reacts and provides them
with similar ways of information exchange. This data flow needs to work both ways:
information is provided and feedback returned. For the scientists involved in study-
ing this phenomena, it is recognized as an established field called citizen science
which creates information hubs, called citizen observatories. These observatories,
provide further assistance in intricate endeavours such as energy urban planning
and management via participatory decision making tools that aim to enable citizens
to act as "eyes" of the authorities and policy makers and to monitor land-cover/use
changes through everyday activities, [219].
The success of such observatories is already visible in Horizon 2020 projects like
SCENT Citizen Observatory (CO), where infrastructure, monitoring systems and
legal and administrative frameworks that allow for monitoring and management
have been developed, [168]. The methodology used in this project is based on ex-
pert judgement (for data quality indicators) and ontological analysis (interpretation),
using citizen collected information with the SCENT toolbox. The data quality tests
performed by local experts follow the thinking that citizens that are familiar with the
specific urban context situations have thematic knowledge and possess an under-
standing of possible solutions and their implications. This is in turn complemented
by a set of Key Performance Indicators that the stakeholders have evaluated. To sum
2.2. The ever changing city - trends in cities and the energy chain 17
up, COs can be used to enhance communication opportunities and facilitate partic-
ipatory decision making whilst using local knowledge and human resources rather
than seeking top down measures.
2.2.2 Big data and urban digital infrastructure with the Internet of Things
- IoT
IoT is identified as early as 1999 as a potential game changer in the context of supply
chain management, [7]. IoT relates to one of the aims of this thesis, to show the
feasibility of using standardized integrated spatial-temporal data and standards in
the energy supply chain. As [213] shows, the Internet of Things can be realized in
three paradigms:
• Internet-oriented (middle-ware), Internet is used for data transmission,
• Things oriented (sensors), oriented to measuring devices,
• Semantic oriented (knowledge), linking the measurement and the devices with
the help of semantics to things in citizens’ surrounding.
Internet availability, the requirement of the first paradigm, is another commodity,
usually taken as granted by city citizens and even engraved in citizen rights in some
countries. Finland, [17], has even provided a minimum legal speed for nationwide
internet access at 1Mbps. 4G internet (60Mbps) has already made IoT possible, how-
ever the availability of 5G (22Gbps) is showing to be the tool that allows IoT to
achieve critical implementation mass and create what is called the 5G IoT or the In-
ternet of everyone and everything, [239]. It is precisely this availability of data flows
in near-real time that creates the so called big data perspective.
So how large does a data set need to be in order to be classified as big data? If the
concept of big data is treated from an absolute value perspective it must be men-
tioned that Bill Gates, the largest stakeholder of Microsoft, considered, at the birth of
his Windows operating system, that a data set was considered very large already at
640KB (the big data term did not exist at that date). As of the start of 2019 there is no
standard number that the IT community accepts as the threshold for discussions on
big data, this varies greatly depending on the skills and tools available to the users
in a group. It must also be said that these numbers matter because they form the
basis of discussions in between the departments that ensure continuous and unin-
terrupted data flows (e.g. IT and research groups). As to what exactly qualifies as
"big data", this also varies depending on the capabilities of the users and their tools,
and their available infrastructure. Ever expanding capabilities and better software
make defining big data in terms of information quantity difficult.
At the same time what also matters is the quality of data. Well structured, reliable
and well defined data sets are the general standard of data in relational databases.
Big data is generally defined as the opposite approach to relational databases, large
volumes of poorly structured data that is sometimes provided with gaps and per-
haps affected by errors. What these data sets lack in structure and perhaps quality
is covered however by simple sample size. These provide scientists and city stake-
holders with perspectives that would have otherwise remained unseen and unrep-
resented.
Within the urban energy chain, data flows remain an issue. It is not as if this data
was not produced before, but rather that its flow was and still is largely interrupted.
18 Chapter 2. Defining the city wide energy chain
It is part of this thesis scope to study infrastructure solutions and software that could
take advantage of this freshly freed energy (or energy relevant) data.
19
Chapter 3
Space is a resource.
This chapter focuses on the components and scales at which information is found
in the city wide energy chain. The spatial context is also properly explored, includ-
ing its dimensions of extent and resolution. As with regards to its components, the
use of the word refers to the spatial actors that interact with the energy produc-
tion, transportation and use. The chapter also includes a classification of this actors.
Further, the issue of the right scale for urban building energy modeling is exam-
ined and approached from a perspective of spatial relevance for the data used. The
issue of spatial data consistency for UBEM modeling is explored together with an
assessment of urban fabric contiguity impact. Finally, a scale dependent urban data
ontology with persistent identifiers to help with origin-destination relationships is
proposed for modeling of the urban energy chain.
3.1.2 Extent
Scale has many meanings, but in GIS two are of greatest significance: resolution and
extent, [84]. With regards to the maximum perimeter or spatial extent of the thesis,
this is city wide as defined using the energy chain expanse in chapter 2. That means
an urban area where people live together in a high density urban sprawl, where a
minimum density of 1000 persons per square kilometer and a minimum size of 5000
individuals is found. Those criteria are paramount in order to allow city services to
be delivered in networks (unattainable in rural areas due to larger distances, lower
population density and most importantly associated economic costs). Also, the the-
sis confines itself in the administrative boundary of the spatial entities defined in the
chain subsection below.
This scale dimension is reasoned by [136] as the optimal one at which inclusive
urban-planning projects that consider full energy cycles cutting across all the pre-
sented intervention areas can be performed. Additionally, solitary efforts (e.g. de-
signing and managing independent passive houses and smart buildings) might not
be optimal overall, as these cannot facilitate the planning and management of in-
teractions between energy significant actors and their spatial locations. The scale
issue is also reflected in the work of [44] who evaluates different smart cities initia-
tives (from city, regional and nation levels). The work concludes that the scale at
which current projects in the topics of energy efficiency in buildings, flexible public
transport services, digital infrastructures, smart grids are focused on is still narrow.
This is due to the fact that most projects are proof of concepts rather than funda-
mental change implementation, at city level. As such, building level initiatives are
discarded as actual change, and rather as prof of concept implementations.
3.1.3 Resolution
With regards to resolution, this thesis makes a case for using buildings as the basic
unit for the specific needs in this thesis. The motivations behind this choice stem
rather from a mix of reasons both in the spatial and in the energy domain such as
specific data availability, geographical barriers, statistical sample scale significance,
and are presented and discussed within this chapter for spatial related issues, and
in 6 for UBEM (urban building energy modeling). This choice is significant because
the accuracy of the results of spatial analysis methods is linked to both the scale of
3.2. Spatial actors 21
the data sets involved and the algorithms used. [84] discusses the scale problematic
in GIS and its semantic definition in detail and concludes that raster and vector data
have to be treated differently.
In raster data, resolution has a direct link to the scale in which the data can be used
because of the explicit size of raster cells (pixel or voxel). Spatial urban energy data
is sometimes, albeit seldom, provided in raster format. Vector data has no direct
resolution compared to raster data. There is no simple definition like the resolution
in the raster case.
The vector data’s fitness for a distinct scale range has to be defined during its acqui-
sition process. Often, the metadata provides suggestions, if such data is available
at all. From within the vector data set itself it is difficult if not impossible to guess
scale. Vector data attributes have a scale dependency, too. The detail or number
of distinct features classes has a scale dependency which is directly linked to the
geometric objects involved.
To tackle different use cases requirements that involve discrete changes, data used in
spatial sciences is often provided at different levels of detail (LOD). 3.1 presents the
most typical level of detail with regards to data used in the urban domain. Engaging
different LODs simultaneously may require different data generation or acquisition
strategies. In the urban domain scale, the lowest level of detail may be represented
by a digital surface model, meteorological information, air pollution or geothermal
applications, with pixels in the order of magnitude of dozens to hundreds of meters.
A more detailed LOD with coarse 3D building geometries may be derived by extrud-
ing the 2D geometries of building footprints with corresponding building heights
stored in a different dataset. The features may be taken from a cadastral dataset,
a LIDAR or photogrammetric flight. The higher the level of detail should be, the
harder it is to maintain the consistency between the layers [89]. For new structures
and buildings, parametric approaches help to enable consistency between multiple
scales because they can be generated out of an all digital modeling workflow, [40].
TABLE 3.1: Typical level of detail for data used in the urban domain
information and actors together. Spatial data for the city wide energy chain is stored
and maintained by legal entities who are first of all entitled (or legally obliged) to
do so and whose activity revolves around administration, civil engineering, urban
planning or management.
The following list comes from open administrative data provided within a couple
of geographical areas that the author has extensive experience. These are depicted
as examples in the thesis. These areas are: Singapore (Singapore), Torino (Italy),
Bucharest (Romania), New York (USA) and Karlsruhe (Germany). This list is not
(and due to the different and unique nature of each city cannot be) exhaustive. It
presents the actor and their perspective, first from profession separation:
• Architects and architecture companies. By their profession description, they
design buildings and are sometimes involved in the construction process in
order to ensure correct building standards. Architects store and work with
data in a way that ensures them a functional perspective for the building. Ulti-
mately, an architect aims for beauty and functionality. This often involves the
creation of visualization prior to actual construction. This process often uses
standards that do not pursue the interoperability perspective. That means se-
mantical city models are rarely used, but rather CAD and BIM related stan-
dards with few, if any architects venturing in, and using, GIS domain stan-
dards. Due to the often unique nature of the products used, conversions from
files created by architects are difficult to complete due to the different perspec-
tive taken. For energy use analysis, architects use building scale products such
as EnergyPlus with extensions in their own architect oriented software prod-
ucts which focus on the detailed energy estimates of single buildings.
• Urban planner. An urban planner is a specialist that keeps an architectural,
economic and political perspective on neighborhood / city development with
an ultimate goal of public welfare. As urban planners often work on larger
projects, GIS is a necessity but often they are confronted with BIM and CAD
input which require costly standard conversions. For the energy chain work
this professionals regularly lack proper tools at neighborhood and city scale.
This is why a host of research projects have sprung up in the last years to assist
in mid and long term energy planning and management. In addition compa-
nies like EDF, IBM, Google, Microsoft, Siemens and others have created unique
software tools that allow for urban energy planning, see 6 for a presentation of
some of those tools.
• Civil Engineers. These are professionals which design, build, destroy or main-
tain structures or facilities for living, industry and transportation and also do
environmental/land improvements works. The civil engineer is a general user
of CAD and BIM. A rather large panel of software products exist which sup-
port the construction, development, and monitoring of construction works. By
the nature of their work engineering products tend to respect standards which
facilitates exchanges with the other professionals involved in the city wide en-
ergy chain.
• Public Administration / civil servants. These professionals are very often
building, maintaining or even designing public works. Most public admin-
istrations keep and maintain their own data repositories. These have mostly
been converted to a digital format. However, that is not always the case and
3.2. Spatial actors 23
there are still situations where historical print-ed blueprints are in use, some-
thing especially true with utility networks, which often have a history dating
back to longer than a century. When dealing with the city wide energy chain
these category of professionals are in general the most important data owners
/ actors in the data chain. They most often control or own the most signifi-
cant data. In the rare cases that they don’t they maintain a network of contacts
which facilitates the exchange of information. Standardization in storing this
data is rare due to the fact that data is presented to them in many different
formats and standards. However, there often exist local or institution based
data standards. Standardization of all data in the spatial domain is something
that has attracted the attention of strong institutions in the EU and these can be
seen by standardization drives like the INSPIRE framework. However, the im-
plementation is an ongoing work with a mid and long term completion goals
that this specialists have to accomplish. Regarding software choices both CAD
and GIS software are used by these professionals.
• Facility Managers maintain and support the use of built environments. Re-
garding spatial context they focus on building (rarely also at district) scale.
They go in to great detail to store all information related to the smallest uniquely
identifiable element that belongs to a facility. Due to the laborious nature of
their tasks they tend to be very precise and favor extremely detailed data mod-
els like BIM IFC models. Very often, facility managers are involved in opti-
mization tasks for the facilities they manage. In the energy domain that means
they maintain a good understanding of the material usage, the occupancy and
schedules in usage of their facilities. As shown in [80], building physics and
occupancy play a key role in energy demand, which makes these professionals
key stakeholders in energy modeling. Additionally, some experts of Facility
Management (FM) use sensors and gather data in order to perform detailed
analysis in order to support a sustainable use of their buildings. Facility man-
agers are among the first (together with security companies) to ask for and
develop smart applications that convert building to smart buildings - being
trend setters from this perspective.
• Transportation entities professionals (engineers) are transportation operators.
They transport people or goods in their networks and, after energy demand in
housing, they represent the second highest demand of energy in cities. Due to
the larger spatial context that they focus on it is quite common that they use
GIS both in data modeling and operation of their networks. There are plenty of
software solutions that support transportation planning, maintenance and op-
eration. CAD and BIM is also used when great detail is required, particularly
in transportation hubs, but then we go back to civil engineering and facility
managers.
• Utility network operators. Professionals working in energy utility network op-
erations are usually engineers focused on the delivery of the energy commod-
ity from the production center to the consumers. These professionals use net-
work specific standards and data models. The operation of such tools require
that standardization is respected and maintained, although in most situations
internationally recognized standards get a local flavor with specific extensions
or entities. It is very often the case that they use both CAD and GIS with some
very specific associated data models and software products. The spatial data
they use is rarely found in the open domain due to security concerns as most
24 Chapter 3. One scale too many - Divisions of UBEM space
F IGURE 3.1: Spatial actors classification upon the scale at which they
operate and maintain data
modelers rarely have access to all the above mentioned spatial actors repositories.
As such, most often data sets are completed / repaired with the use of statistical
data and ETL (Extract Transform Load) operations.
F IGURE 3.2: Spatial scales in the city wide energy chain in different
geographical, cultural and administrative contexts
Six cities that the author was able to obtain open data for were selected for a graphi-
cal representation of the spatial divisions of data, see 3.2. These cities were chosen to
26 Chapter 3. One scale too many - Divisions of UBEM space
depict different geographical and cultural contexts and share to different extent the
open data initiatives that facilitate the work carried out in this thesis. These cities are
Singapore (Singapore), Geneva (Switzerland), Shanghai (China), Berlin (Germany),
Bucharest (Romania), New York (USA). The cities chosen for the comparison lie in
terms of size and population in the same category of multi-million metropolis. What
can be observed from these examples is that public administration shares the ten-
dency of organizing data on a similarly called scale, neighborhood. This spatial
level contains anywhere in between a few hundred to a few thousand buildings,
3.2. Neighborhood, sometimes referred to as community district, section, borough,
precinct, block, area, zone, or in other languages as quartier (French) and Stadt-
teil (German), is in fact a common spatial scale (in terms of approximate size) used
for division across different cities and is something that people of all backgrounds
generally refer to when understanding city scale. The same concept can be met with
cities that contain smaller population numbers and have a reduced surface size, with
the difference that smaller cities tend to lack the the 4th administrative level (referred
to as town).
To best illustrate the spatial scales an example in the city of Torino, Italy was cre-
ated using the six spatial scales present in the city: city (città), town (circoscrizione),
neighborhood(foglio), district (isolati), building (edificio) and dwelling / client unit
(abitazione). This can be observed in 3.3.
Largely, the spatial scales identified in urban planning and management involve
a significant number of spatial data stakeholders and can be classified in six levels,
presented in 3.2. Across the world, the keywords used to name the areas vary greatly
in meaning and size, spanning municipalities, subdivisions of municipalities, school
districts, or electoral district. In this table, all names have been translated to English,
using Google provided machine translations. These are deemed correct for English
use in Academic Purposes by [87] for most purposes, except grammar. The numbers
provided in the table serve only as indication of scale, rather than accurate number-
ing. In fact few numerical description for these spatial scales have been found, with
the exception of China, where neighborhoods are estimated to be at around 2000-
10000 families, and quarters at 100 to 600 families, [49], however, numbers like this
are seldom provided and do not constitute a fixed rule, rather an indicative value,
even in the Chinese context. For this manuscript the following definitions were used:
• city, administrative entity where people live together in a high density urban
sprawl with a minimum density of 1000 persons per square kilometer and a
minimum size of 5000 individuals
• town, spatial unit, generally smaller (or equal) than a city, that encompasses
multiple distinct zoning areas (different area uses such as commercial, indus-
trial or residential)
• neighborhood, geographical entity that usually contains a single use zone,
very often residential in nature, with limited tertiary activities present;
• city block, the smallest group of buildings that is surrounded by streets; usu-
ally statistically significant for the city administration, and does not have ther-
mal bridges to other surrounding entities (a distance larger than 0.3m - nar-
rowest street in the world);
• building, structure with a roof and walls standing permanently in a location,
that has it’s own entryway (or multiple);
F IGURE 3.3: Spatial scales in the city of Torino, Italy
28 Chapter 3. One scale too many - Divisions of UBEM space
• energy meter, or the client unit / final consumer, usually having a one to one
relationship with an administrative unit (apartments, building or a couple of
buildings), is for energy companies a unique entity located on the premises of a
residential/tertiary/industrial unit usually within a larger structure/building;
a single building can contain multiple energy meters.
The logical way to describe the definitions and the relationship of scales from above
is as follows: every city has at least one town, and must have at least one neighbor-
hood with a couple of city blocks and a minimum of a thousand buildings. Within
the work presented in chapter 6, the energy data is linked or aggregated at the level
of buildings or city blocks. A graphical depiction of the definitions of city blocks,
buildings, and energy meters presented before is offered in 3.4.
The natural evolution for administrative scales means that often historical data sets
exist for these spatial units, which in turn allows for the performance of data-driven
approaches that support trend and change detection. For example in the Canton of
3.3. A proposal of standardized spatial scales in UBEM 29
3.3.1 Spatial relevance of data unit sample size - scale sensitivity in UBEM
A noteworthy aspect in the city wide energy chain modeling is the relationship be-
tween the extent of what is modeled and the size of the data unit at which informa-
tion is aggregated/modeled. This entire chapter makes a case for the use of build-
ings as data units and neighborhoods as the "right" spatial scale at which aggregated
and disaggregated data facilitates city-wide energy chain modeling and city data
modeling in general.
The lack of standardized input data means that UBEM models often make use of
"building archetypes" in order to reduce the simulation inputs required, [3], [228].
Archetypes are groups of buildings definitions that share similar properties, such as
use, age and belong to the same spatial extent. A review of existing national and
30 Chapter 3. One scale too many - Divisions of UBEM space
local archetypes was presented in [191] and further methods for developing multi-
scale (national, city, county and district) archetype development methodology using
different data-driven approaches can be found in [3], and [46]. The procedure gener-
ally consists of five steps: data collection, segmentation, characterization, quantifica-
tion, and results modeling. Archetypes and the entire effort of the UBEM scientific
community represent another testament to the potency of using buildings as data
units in energy modeling.
This chapter states that city wide modeling should be done at neighborhood level
with buildings as the data aggregation unit. However, one can argue for the possi-
bility of going beyond building data units, and providing more, or less, data aggre-
gation. For example, [121], approaches this issue from the perspective of sensitivity
of the energy model, and converges on the idea that larger data collection processes
for higher details in UBEM result in statistical improvement of the results only with
small sample size calculations (a small number of data units). In his case the data
unit is the household, what is previously identified in 3.2 as the energy meter. The
author compares simulation results at different scales against measured energy use
in the case of single residential units to a multi dwelling building in the area of Karl-
sruhe, Germany. The results depicted in A.6 show the increasing coherence of a
predictive energy demand model with increasing sample size of one, five, ten and
thirty residential units of the same building. The statistical model uses the mean
daily outdoor temperature as single regressor to predict daily heating needs. The
findings are an indirect indication of the dependency of scale in the aggregation of
energy use data. The correlation of both time series at the scale of individual units is
relatively weak and derives strongly from the line of equality, (top left figure). With
increasing scale (larger number of energy meters / consumers / residential units) a
much tighter fit of both samples is reached. This matches well the line of equality (30
samples, bottom right). As the predictive model is not adapted to represent individ-
ual behavior, the results indicate that these are clearly visible at a small sample size.
However, with increasing sample size the results indicate that different behavior is
resulting in an increased as well as diminished energy use so that the results can
be predicted using an uncalibrated model. This corresponds well with the purpose
of UBEM, city and town scale modeling, where large numbers of individual usage
patterns combine to form the total expected demand.
This statistic effect is also underlying the use of statistic load profiles, in which in-
dividual behavior is aggregated to a smooth profile. This perspective was studied
in electric systems, with the aggregation described by [181]. In the case of thermal
demand, [85] used the aggregation methodology for developing synthetic load pro-
files. The aggregation of load profiles allows for accurate prediction of energy use
at larger scales. In the electricity market this results in reliable predictions based on
standard electrical profiles for a large number of similar consumers as well as in the
gas load prediction for regulatory zones, [18].
The results for the scale of urban neighborhoods presented in [121] show similar
statistical effects. This leads to two main conclusions based on the same effect. On
the one hand statistical modelling approaches such as the energy signature model
proved to be applicable at this intermediate urban scale. This conclusion is in line
with [136] and is the main thinking behind UBEM applications that include the pre-
diction of energy needs for district heating systems or common systems for groups
of buildings. On the other hand, the fact that differences in energy use by individ-
ual behavior neutralize each other clearly indicates that the scale is not suitable to
3.4. Spatial consistency 31
can also be required because of data inconsistency with regards to the spatial
levels in between stakeholders / institutions.
– Extrapolated from direct neighbours
– Disaggregation from the larger scale spatial unit to which the building
belongs to
– Aggregation from smaller residential units that belong to the building in
question
– Spatially randomly distributed from statistical data sets
– Clustering using socio-economic indicators
– Measurement based allocation
This procedures ensure consistency and provide an interrupted flow in UBEM method-
ologies. They are required either for providing data as input to the UBEM algorithms
(where they become the simulation backbone), either for the purpose of sharing re-
sults and providing immersive decision-making in the analysis sequence. Providing
data to the simulation is usually the first part in a inclusive and iterative process
where multiple stakeholders contribute input that help energy and urban planners
devise processes that involve multiple steps in the preparation. These processes can
even impact spatial segregation of planning units due to energy resources allocation
(over production in certain areas, energy poverty in others).
To facilitate the theoretical approach proposed in this thesis with regards to build-
ings as data units for energy information aggregation an example is offered in the
case of a neighborhood in the city of Torino, Italy. The energy and socio-economic
relevant data is linked to a 3D CityGML model of the buildings. The buildings are
a 3D data set, while the certificates are point data in 2D. Both data sets originate in
the open domain and prior to handling within UBEM algorithms underwent several
ETL procedures in order to further clean the data (eliminate incomplete, inaccurate
or conflicting entries). To facilitate the merger of such data two GIS related methods
were used: contained within and nearest neighbour assignment. These very potent
operations facilitate the assignment of energy certificates data to the buildings. As
can be graphically observed in figure 3.5, only one in five (19%) of all buildings with
an energy consumption located in the center area of Torino receive a direct assign-
ment from energy certificates. This incomplete data set is a common occurrence in
the UBEM field and is handled on a case by case basis.
In addition to missing or incomplete data, there is also the issue of contradicting
data. For example, energy certificates of dwellings of the same building can present
conflicting indicators (e.g. different year of construction or energy use class). Issues
such as this one can be tackled using automated data validation processes (for ex-
ample data-driven using rule based algorithms, as presented in [47]). However, in
certain cases, manual intervention is required on behalf of the modeler, and this is
the reason that the data processing itself is usually about 40 to 50% of the estimated
work load in the case of UBEM modeling.
34 Chapter 3. One scale too many - Divisions of UBEM space
centroid method, [141]. Prior to processing taking place the number of cluster (k) is
defined and the optimal solution for the given k distributes centroids as far apart
as possible with data observations being associated with the nearest centroid. Clus-
ter centroids are recalculated based on mean distance to all observations in newly
generated clusters. This process is repeated with cluster centroids shifting locations
after each iteration, until the clusters stabilize. SOM, a specific case of ANN (Arti-
ficial Neural Networks), [241], are particularly useful with multi-dimensional data
sets. It projects input variables into a low-dimensional grid which is then used in
visualizing and exploring properties of the data. The SOM grid is user-specified in
terms of size and shape with cell values being initialized randomly. Through mul-
tiple iterations the values are adjusted. A good example of a UBEM publications
using clustering methods is found in [200] where SOM is applied on a dataset of ap-
proximately 40k buildings to identify clusters of similar socio-economic parameters
that can be used for energy retrofitting business cases.
Even though, the spatial scales presented in 3.2 are the results of historical devel-
opments, it is noteworthy the fact that scientists also started recognizing neighbor-
hoods as the optimal scales for energy urban-planning projects, [136]. Due to its
significance, a definition of neighborhoods was sought in existing literature. In gen-
eral the concept is described by sociologists as the place where social interaction
occurs. As such it is essentially a social construct, made of ever-changing and lay-
ered systems of personal and collective links within and across cultural boundaries,
[156], which is a rather similar concept to energy networks. What makes people
share an energy network and an energy source is arguably very simply explained
by their vicinity and accessibility to/of it. This corresponds to Tobler’s first law of
geography in [217]: "everything is related to everything else, but near things are
more related than distant things.".
This is especially true with regards to energy trends presented in chapter 2, specif-
ically the fact that buildings and their inhabitants are now concurrently consumers
and energy producers (prosumers) and there is a lack of energy transportation net-
works able to take that extra energy in order to deliver it. For electrical networks the
issue is not a lack of networks, but rather a design "flaw" as originally networks were
designed with uni-directional flows from large production centers to consumers, see
2.2. For all other energy utility networks, in most cases they simply do not exist, and
represents one of reasons for which building data units are recommended for use
in city wide energy modeling the by this manuscript. Buildings, per se, constitute
contiguous entities. An appropriately scaled framework, in this thesis’s case, the
building unit, on which a more continuous, more spontaneous urban pattern may
be formed.
3.6. A scale dependent urban data ontology with persistent identifiers 37
The unique parent of all entities in a city is the object City. It contains at least one
town. Each town contains at least one neighborhood, which in turn has at least one
district. Districts can contain multiple buildings. Finally, the building meters can ex-
ist in buildings, however, it is often the case that multiple buildings can be allocated
to a single meter. The parent - child relationships are all of type one mandatory to
many optional. Thematic attributes are allocated to each entity and can be aggre-
gated and disaggregated from one scale to another using specific calculation meth-
ods. Certain attributes are present at all scales, for example, number of inhabitants,
while others could be scale specific, such as building physics parameters, as can be
observed in 3.9.
Within the energy chain energy thematic information are often linked to spatial units
building entities. Ergo, buildings host this data as attributes. This further facilitates
the links to the Energy ADE and the Utility Network ADE which also use buildings
for the storage of specific information. The larger scales, that go beyond buildings,
are relevant to energy planners and usually host energy related KPI indicators such
as total energy demand, GHG production, CAPEX (Capital expenditure) or OPEX
(Operating expense).
In order to facilitate modeling activities, the entities host two identifiers as attributes.
This are referred to as PID (Persistent IDentifier) and GUID (Globally Unique IDen-
tifier). No single PID system for urban energy planning was identified during the
writing of this thesis. Typically, in information technology, PIDs are used for the
foundation referencing of digital assets in scientific publications, books, and digi-
tal repositories, [229]. PIDs usually contain metadata, and the same thinking was
3.6. A scale dependent urban data ontology with persistent identifiers 39
applied to the realization of the ones included in the thesis. The method proposed
here embedded spatial levels in the identification, thus facilitating modelers under-
standing of the spatial entity that they are working with. The presence of the GUID,
sometimes referred to as a UUID (Universal Unique IDentifier,) as a second attribute,
is made so that the IT systems to be provided with a GIS typical unique identifica-
tion mechanism (also used in CityGML and the entire GML - Geographic Markup
Language community). Table 3.4 presents the two identifiers used in this proposal.
The number of digits used at each spatial level in PIDs was determined using sam-
ples available for the previously mentioned cities, while also giving future users
flexibility, in case certain spatial scales are non-existent in their own use case. The
length of the string depends on the number of spatial scales that precede it, see tables
3.4 and 3.3.
The flexibility of the proposed method allows for different sized cities to be modeled.
For example in the case of the city of Karlsruhe, Germany, with an approximate pop-
ulation size of 313k and 85k buildings the scale Town is not required (nor can it be
identified within the existing administrative structure). In fact, most UBEM model-
ing situations do not include an entire city, but rather sections of it. Also, multiple
cities can be included, as often data repositories contain multiple applications of the
same UBEM models. The modeled area can be referred to as a study area, in case
that the entire city is not being modeled, and replace one of the spatial scales in the
proposal.
Following the identification and modeling of all spatial scales relevant to the study
area and use case, the existing spatial structure can use the EAV (Entity Attribute
Vale) data model for populating all relevant information. The proposed ontology
helps mitigate issues such as data heterogeneity and consistency in urban domains
by offering a binding platform to all stakeholders from a non-biased perspective.
F IGURE 3.9: Hierarchy based on spatial levels
F IGURE 3.10: Spatial scales based identification system
43
Chapter 4
This chapter focuses on the spatial-temporal data in the digital domains that inter-
act in the urban energy chain of spatially aware cities. This includes the standards,
Spatial Data Infrastructure (SDI) and methodologies used when modeling and mon-
itoring the city wide energy chain domain, as previously defined in chapter 2. The
entities and interactions modeled with the standards included in this chapter are
approached from the perspective of urban energy modeling and planning. Mod-
eling activities require the detailed description of the existing situation in order to
carefully predict the behaviour of existing systems. Ideally, energy modelers are
looking for situations where the modeling input (the entire information required for
the description of the status quo ante) is made available, in uninterrupted workflows,
that allow for contiguous energy chain modeling, from standardized sources. In or-
der to process all this information good infrastructure and consistent data flows are
required. As Hodges and his colleagues explain in [101], declarative modeling lan-
guages already facilitate the definition of semantic relationships with such quality
that it makes them suitable for enabling machine-to-machine (M2M) interoperabil-
ity (interaction and understanding). They continue to identify gaps with regards to
proprietary ontologies and cascading effects from their use, an conclude that these
semantic models are unlikely to play a large part in resolving the pressing horizontal
and vertical interoperability issues currently encountered. These proprietary ontologies
hold the promise of interoperability in name only. This judgement further motivates the
work presented in this section as it focuses on using open semantically organized in-
formation as support for interoperable spatial-temporal data flows and applications
in energy modeling.
value in the use case. In the thesis context of the city wide energy chain, 25 use cases
were distinguished (presented in table 4.1) that help plan and manage the energy
chain in a city.
What can be observed from the table is that the use cases touch all the energy chain
components previously identified in Chapter 2, either directly in production centers,
substations, consumer buildings (residential and tertiary) or networks. With such an
inclusive list in terms of energy chain components the case for modeling the entire
energy chain by using CityGML as a data hub is strengthened. The fact that 3D se-
mantical city models are currently used in at least 25 use cases with ramifications in
the energy chain in cities also provides a justification for cities and mapping agencies
to provide data in this formats.
Specifically in urban energy planning both [183] and [196] highlight the importance
of open data and standard formats for decision support tools in urban planning.
This is essential for three main reasons: (i) to streamline workflows processes, from
input to output, (ii) to give software tools the ability of functioning in different ge-
ographical settings and at different spatial scales, and (iii) to facilitate comparison
and interpretation of results.
Planning of cities for urban infill (greenfield) or redevelopment (brownfield) invest-
ments both requires good data availability for successful design and implementation
decisions. However, brownfield is far more data intensive, as [42] indicates. To han-
dle the large amount and variety of related data in a structured way, this thesis has
adopted the CityGML standard, extending it in the process to cope with the genera-
tion of many urban scenarios through interactive optimization.
However, the scale design and dependency is also linked to the energy commodity
that is transported. For example electrical network data using the CIM standard can
encompass cities, regions and even national and continental networks while district
heating networks standards such as DIN 4747-1 and the associated spatial model are
limited in scope to districts, neighborhoods and cities.
F IGURE 4.2: Spatial model standards and design scales for BIM, CAD
and GIS
The third perspective taken in the taxonomy is energy thematic. First with regards
to the position of a modeled element in the energy chain, these entities can be en-
ergy producers, transportation/utility networks entities, consumers or prosumers,
all described in detail in chapter 2. Standards used for producers often relate to the
technical requirements of the operator, this can be a heating production plant. For
the UBEM modeler the purpose is to be able to model the commodity exchange with
the producer, e.g. heating sent in the network has certain technical parameters de-
scribing it (temperature, flow) and other parameters on the return flow. Secondly
the standards refer to different commodities (electricity, heating, gas, waste water,
storm water, fresh water, communication, process steam, oil, waste, air pressure),
for example DIN 4747-1:2003 or VDI 2036:2009 for DHN components and annexes.
The last perspective included in the taxonomy is based on the entity type using them.
Standards in use by public entities - city administrations, standards in use by compa-
nies or standards in use by citizens. These three are separate because of the different
use cases that they serve. In general public entities are incentivized to maintain fac-
tually correct records of the energy chain components that they are responsible for,
and often use standards with open source data models, such as CityGML or CIM.
As such their data records tend to cover larger areas and use cases, while the ones in
use by companies tend to be narrower, as stipulated to legislation and related to the
specific business case of the company. In the context of the digital revolution dig-
ital devices such as smart homes installations, smart phones and low cost sensors
have also facilitated the emergence of citizen scientists that collect data and often
provide it via open data repositories. These treasure data trove often facilitate the
work performed by UBEM modelers.
48 Chapter 4. Not all data are created equal - elephants in the room
which is a graphical user interface that connects to the EnergyPlus simulation en-
gine, [65]. It allows for the import of DXF files and is used to perform day lighting
analysis, thermal simulation of natural ventilation, visualization of site layouts, so-
lar shading and sizing of HVAC systems, [151]. Most energy simulation software
that allow for the import of CAD files are slowly being discontinued, these include
i.e.: Ecotect, [11]), and Project Vasari (now replaced by REVIT), [10]. Frequently the
software take the form of plugins, such as on AutoCAD products: EnergyPlugged,
e-QUEST, Energy + or Insight 360.
However, due to initial design decisions pure CAD data formats show their limita-
tions when it comes to data models in the built environment. CAD was designed to
cover different and very broad use cases related to computer-aided design. To cover
the emerging demand of complex and integrated engineering design and planning
standardized Building Information Modeling (BIM) has emerged. A BIM model is a
digital representation of physical and functional characteristics of a facility. As such
it serves as a shared knowledge resource for information about a facility forming
a reliable basis for decisions during its lifecycle from inception onward, [142]. It al-
ways contains a 3D graphical representation of the building and additionally focuses
on the information data model regarding the object. Given that, not only geometry,
but also semantic information, such as ownership, construction material, costs and
energy-related data are stored in the BIM model.
Fundamentally, BIM assists with constructing buildings virtually, before their con-
struction in the real world. However, it does not stop there, and further supports the
building use throughout its life cycle, all the way into dismantling and demolition,
[67]. What it brings in addition to CAD systems, is that it enables the creation of sin-
gle and unified representations of the building, so complete that it can generate all
the necessary documentation, [165]. Typical for BIM projects is a high level of detail,
which helps in conflict detection and improves planning and construction. Overall,
it delivers an optimisation in the quality of the entire building related process while
enhancing interoperability when connected with the use of single all-inclusive soft-
ware packages or open data formats, [93]. However, producing very detailed repre-
sentations of the building results in labor-consuming processes. Hence, the usage of
BIM is usually limited in terms of spatial scale to that of single buildings, or at most
district.
Specifically for storing energy-related information, the BIM community has devel-
oped the Green Building XML (gbXML) data format, which enables interoperability
between BIM software and Building Simulation Tools. According to Hardin and
McCool, it supports model-based analysis for the energy simulations of the whole
building, [97]. gbXML is an XML file that includes over 500 categories of elements
and attributes used for the description of all aspects of a building related to its en-
ergy footprint: building physics, occupancy, meteorological parameters and the ge-
ometry. The gbXML schema was designed so as to enable interoperability between
energy analysis tools and BIM software.
Concurrently, significant research effort has been made to provide automatic data
flows that built gbXML data such as presented in [227]. BIM software tools that sup-
port gbXML include: Autodesk Revit, Bentley Architecture and ArchiCAD. gbXML
files can be used in multiple simulation software, for example Open Studio ([94]),
a full-featured software framework that supports rigorous, multidisciplinary build-
ing simulation. It is a collection of software tools that enable whole building energy
50 Chapter 4. Not all data are created equal - elephants in the room
modeling. With regards to the simulation engine used, it is that of EnergyPlus, [60],
as well as advanced daylight analysis using the Radiance software, [230].
The most used and produced data format in the BIM field is the IFC format, [97].
The Industry Foundation Classes (IFC) specification is a neutral data format devel-
oped and maintained by buildingSMART International, an international organiza-
tion which focuses on improving interoperability for software applications used in
the building industry. IFC is developed as an open specification data format for
describing, exchanging and sharing construction and management data across mul-
tiple applications in the building industry. The data model is object-oriented and
contains elements used throughout construction or facility management projects.
The IFC format is also the international standard for openBIM, [109], a collabora-
tive effort that is vendor-neutral with processes that supports seamless teamwork
throughout project lifecycle. This initiative seeks to enhance interoperability by pro-
viding reliable data exchanges and collaboration workflows. The latest version of
the standard is provided as an ISO standard - ISO 16739, [110], and can be described
using either the STEP (.ifc) or the XML (.ifcXML) data structure.
One other potential data source for UBEM practitioners is the BIM Collaboration
Format (BCF), a format used in the BIM community that allows multiple users to
work on the same file concurrently, [13]. It grants multiple BIM applications the
ability to communicate on model-based issues on already shared IFC models. There
are two alternative solutions, one based on file exchanges between software plat-
forms and another one using a web service. BCF is developed by buildingSMART
International, and provided as a standard. What makes BCF valuable in the energy
chain is that it can be easily exchanged or “roundtripped” (sent back and forward
in between users). This stands true as long as all users maintain the integrity and
uniqueness of the shared BCF file, [114].
Both gbXML and IFC formats represent a source of heightened interest for UBEM
experts and provide interoperable data exchange while BFC creates a framework for
data sharing within the same team and among institutions. The shift in the construc-
tions sectors towards standardized BIM formats and the similarity of design spatial
scales with UBEM is creating larger volumes of BIM data, therefor becoming a good
source of information in the modeling of the energy chain. Table 4.2 presents both
file formats and the BCF framework with their main characteristics.
While CAD tools and data may be encountered on each of the energy chain relevant
spatial scales mentioned in chapter 3, BIM’s use cases are narrower in terms of size.
CAD software can be used for the detailed drawings and plans of single buildings
as well as to work on larger regions, i.e. while planning and designing electrical
grid within the city. Due to the fact that BIM models are usually very detailed and
accurate depictions of reality, they are usually used to describe single buildings and
rarely districts. However, with the advent of BIM in industrial workflows, it is ex-
pected that the amount of data available will steadily increase. Figure 4.2 depicts the
spatial scales for which CAD, BIM and GIS technologies are employed.
For the handling of data within the applications related to this manuscript CAD and
BIM data has been converted to GIS relevant formats described in the next chap-
ter, SHP and CityGML. This was done mostly for the purpose of standardization of
input data to the simulation tool aEneAs (Energy Assessment) and various user in-
terfaces that were used. The conversions, energy model, workflow and architectural
solutions used are described in detail in chapter 6.
4.3. Digital domains for spatially aware cities 51
TABLE 4.2: BIM data formats used in the urban energy chain
6, where the energy models library aEneAs - that uses Python/SQL is presented.
The library seeks to avoid issues of standardization by using semantical city mod-
els. In an urban context, for creating and standardizing digital urban twins semantic
city models have been the main solution employed in the field of UBEM, [191]. This
combines both vector and raster data sources and represents a good basis for urban
management and urban planning as [90] testifies. These models include:
• 3D geometry
• Thematic coverage (e.g. building, vegetation, city furniture)
• Topology (spatial relations between connecting or adjacent geometry)
• Semantics (e.g. the meaning of a surface door, window)
• Appearance (surface texture)
• LOD (level of details concept)
• Open standardization (to provide interoperability)
CityGML is the international standard of the Open Geospatial Consortium (OGC)
for the representation and exchange of 3D city models. CityGML is based on XML
and is used for the storage and exchange of such models. It has gained traction over
the years to become a significant part of the open 3D data available in the developed
world. CityGML distinguishes between geometrical and thematic models. A geom-
etry model grants consistent and homogeneous definition of geometrical and topo-
logical properties of spatial objects within 3D city models. The thematic model of
CityGML employs the geometry model for different thematic fields like DTM (Dig-
ital Terrain Models), sites (e.g. buildings; bridges and tunnels), vegetation, water
bodies and transportation facilities.
CityGML 2.0 includes a concept of Levels Of Detail (LOD), presented in table 4.3.
It is a hierarchical detailed approach where objects evolve to include more details
with higher LODs with regards to their geometry and thematic differentiation, [91].
LODs are defined using data/point accuracy, minimal dimensions of objects and
related model scale. Simply put, the LOD0 includes a 2.5 dimensional Digital Ter-
rain Model, a ground mesh. Starting with LOD1, building blocks and classes are
included while LOD2 brings in differentiated structures, textures and thematically
differentiated surfaces. With LOD3 external building architectural models, as well
as high-resolution textures are included (facilitated by the intricate surface descrip-
tion). The highest level, LOD4 finally includes interior structures, and has been sel-
dom used outside of academia, as such models are usually built and managed within
the CAD or BIM domains. CityGML can concurrently include buildings with differ-
ent levels of information, which facilities the merger of multiple source data sets as
well as efficient data visualization and data analysis.
What makes the standard stand out is the fact that it includes not only graphical as-
pects of city models and the representation of the semantic and thematic properties,
taxonomies and aggregations, as seen in [91]. CityGML’s classes and ownership re-
lations for the most relevant topographic objects in cities and regional models with
respect to their geometrical, topological, semantical and appearance properties sup-
port the aforementioned design choices.
National and city level public authorities have started providing CityGML data, as
shown in [148] and [149]. This statement is valid for more than half the territory of
4.3. Digital domains for spatially aware cities 53
the EU. The EU has also adopted CityGML building feature class as a standard with
the INSPIRE directive. In fact the main part of CityGML used in this thesis for de-
scription of the entities in the city revolves around the building class, significant for
modeling information used in the energy chain. This simplifies any ETL operation
that originates in the INSPIRE format. Figure 4.3 presents a diagram of the CityGML
core and the included classes.
When the concepts surrounding the thesis were developed, CityGML had been up-
dated to its second version 2.0. Summarized from [90], CityGML is:
• XML data model for exchanging 3D city models
• Open and international standard from the OGC since 2008
• Containing geometry, topology, semantics, visual appearance
• Based on Geographic Markup Language (GML)
• Describing entities and their spatial and non-spatial attributes, relations, and
their complex hierarchical structures in five levels of detail.
Beside the building class, there are other CityGML classes that did not come into
use in this thesis’ experiments, because the standard is used to describe all entities
in a city. On the other hand, CityGML has its own limitations, which a reference
benchmark study can attest to, [170]. As this manuscript finds its way to the printer,
CityGML is having a new upgrade in terms of content and syntax top, to 3.0; [124].
The scientific community is divided as to whether a new version was required at
this stage or rather increased use and implementation methodologies that facilitate
wider use of the standard. Certain constraints and drawbacks are corrected with
the upcoming version 3.0 of the standard, however, the solutions presented can also
have an enhancement effect with regards to difficulties in implementation.
As a data model’s advantages and disadvantages can only be properly understood
and tested when integrated in data flows, different implementations of the stan-
dard were analyzed. The open database schema version of CityGML, the so called
54 Chapter 4. Not all data are created equal - elephants in the room
TABLE 4.3: CityGML 1.0 and 2.0 level of detail concept scale and ac-
curacy requirements, adapted from [91] and [242]
3DCityDB, offered a great tool for connecting existing information and standards to
spatial entities. The open software tool, described in full detail in [238], offers so-
lutions for storage, analysis, management, interaction, and visualization of the 3D
city models stored in the standard. It works both with PostgreSQL and Oracle, the
second and third most used SQL (Structured Query Language) DB engines on the
planet in 2019, numbers presented in [205].
In the particular case of this manuscript the applications developed to test the re-
search questions use the aforementioned database schema within a spatial relational
database, PostgreSQL with the PostGIS spatial extension. The package has an as-
sociated set of procedures that allow one to import, export and manage virtual 3D
city models according to the CityGML standard. These procedures greatly facilitate
the use of CityGML. 3DCityDB was used in association with other non standard-
ized data sources. For that purpose the persistent identifier PID system presented in
chapter 3 was used and links to the entities in the schema by calling on the UUIDs
of each object were made.
Table 4.4 summarizes the information that regards CAD, BIM and GIS spatial scale,
formats, software, and energy simulation tools presented in this section of the pa-
per. Further information on simulation tools that use GIS is presented in chapter 6.
For all intents and purposes CityGML remains the only GIS related standard with
4.3. Digital domains for spatially aware cities 55
capabilities that support performing energy simulations at city scale using buildings
as the smallest measurement unit. Complete data sets (spatial and energy thematic
data combined) from single sources describing the characteristics of the objects to
be simulated are difficult to come by, a topic discussed in chapter 3. As such, these
models are generated with the help of different data sources from different energy
chain actors.
Energy ADE which facilitates energy simulation at urban level with building scale
results.
Figure 4.5 depicts the way in which the Energy ADE extension is built on top of the
Building class of CityGML and on entities and features defined in the international
standard ISO 19136 GML. Also included is the Timeseries ADE, another extension,
used internally at the KIT. In this specific case it is employed in order to accommo-
date time series data (a common occurrence in the energy domain, e.g. occupancy
profiles or measurement data from meters).
The CityGML Energy ADE enables the calculation of energy flows in buildings,
based on stored and managed data, [79]. For performing detailed energy calcu-
lations such as energy gains and losses, the envelope of the building (including
external installations, e.g. outdoor window shades or solar panels) is considered
the physical boundary of the data model and is stored in the new attributes, Ther-
malOpening, ThermalBoundary and ThermalZone. At the same time, the envelope
is also associating it with the geometry of features in the old Building class.
The data model also allows for energy systems mounted inside buildings to be mod-
eled. A significant design choice was not to include large energy producing instal-
lations that supply multiple buildings and are connected to grid infrastructure. An-
other ADE, the UtilityNetwork ADE, can be used to handle such installations, see
the following section for more details. The Energy ADE allows for buildings energy
input to be modeled within each unit inside the data model. This is also true for
energy output, in case the building houses a prosumer.
4.3. Digital domains for spatially aware cities 57
F IGURE 4.5: CityGML and the Energy ADE linkage, screen capture
in Enterprise Architect
Figure 4.6 presents the internal structure of the EnergyADE v1. It consists of four
modules: Occupant Behaviour, Material and Construction, Energy Systems and Build-
ing Physics, supporting classes and a core model. This structure can be extended by
adding new modules or adding new classes to the modules (a standard ADE fea-
ture). It includes the possibility for further data/concept model extension and in-
cludes interoperability between the modules. The modules in the data model can be
filled on an optional basis, depending on the purpose of the utilization.
F IGURE 4.7: CityGML and the UtilityNetwork ADE link, screen cap-
ture in Enterprise Architect
such modelled systems are already compatible with existing CityGML-based city
models. Figure 4.8 depicts the core module of the UtilityNetwork ADE.
A more recent attempt at comparing standards in use and development for utility
networks, is found in [34]. The conclusion of the study is that the UtilityNetwork
ADE can handle multiple types of networks and features in a common infrastruc-
ture. At the same time the functionalities that are required with modeling complex
utility networks interactions are present. Additionally, a use case for interconnected
water and electrical networks is presented in the aforementioned manuscript. It
leverages the extension for modeling the two intertwined networks with water flow
being used to generate electricity.
Summarized in table 4.5 are the capabilities of different utility network modeling
solutions. These are compared from the different perspectives of the integrated fea-
tures: standardized features definition, different network types capabilities, network
hierarchies, feature hierarchies, topological connectivity, and explicit topography
versus topology. The table also includes work from [23] which focuses its analysis
on the LandXML standard. This standard is recommended by the OGC for use with
utility network modeling and includes a very simple topology model. The stan-
dard is specifically designed for open canals and waterways that transport water,
wastewater, storm water, and other similar products, and the associated properties
are similar in scope. This design decision is the reason for which the relevant pipe
element model structure is only basic (parametric geometrical representation, and a
spatial reference that includes point coordinates).
60 Chapter 4. Not all data are created equal - elephants in the room
«enumeration»
UtilityNetworkADE KIT - Core NetworkClassValue
CableNetwork
ADE «featureType» PipeNetwork
«A DE E lem...
OtherNetwork
_CityObject +consistsOf CityGML_Core::_CityObject
«CodeList»
FunctionValue 0 ..*
«enumeration»
+ feeding S patialQualityValue
+ draining
unknown
+ distribution
surveyed
+ storage
+ venting calculated
+ abstraction interpolated
+ measurement estimated
assumed «FeatureType»
+ shortCircuit
Netw ork
+ branch «FeatureType» +component
+ observer AbstractNetworkFeature
«P roperty»
+ protection 0 ..* 0 ..* + class :NetworkClassValue [0..1]
+ control «P roperty»
+ function :NetworkFunctionValue [0..1]
+ shutOff + function :FunctionV alue [0..1]
+ unknown + yearO fConstruction :Date [0..1] 1 1 1
+ status :S tatusValue [0..1] +subdivision 0 ..* +relatedNetwork
+ locationQ uality :S patialQ ualityV alue [0..1] +belongsTo
+ elevationQ uality :S patialQ ualityV alue [0..1] «featureType»
* 0 ..* NetworkSystem
1 ..
+geoRep 0 ..*
+nodeMember 0 ..*
«dataType» +topoGraph 0..1
GeometricRep «FeatureType»
+geoRep +nodeMember NetworkGraph
Node
«P roperty» 1 ..*
0 ..* 1
+ representationFrame :GeoRepFrameValue
+ representationContext :GeoRepContextValue
+ geometry :GM_Object +start 1 +end 1
+geoRep 0 ..*
«enumeration» +connectedWith 1 ..*
GeoRepContextValue
«featureType»
annotation +linkMember
InterFeatureLink
center
axis 0 ..*
«P roperty»
footprint «enumeration»
+ type :RelNetworkNodesTypeValue
profile GeoRepFrameValue
+networkLinkMember
surface
paper «dataType»
body
geographic Netw orkLink 0 ..*
«enumeration»
«enumeration» RelNetworkNodesTypeValue
NetworkFunctionValue
in let
supply outlet
disposal undefined
communication unknown
F IGURE 4.8: The core class of the UtilityNetwork ADE link, screen
capture in Enterprise Architect
In order to classify the solutions depending on the level of support provided for
different features, each support level was categorized from 1 to 3 in the table (1 -
poor, 2 - basic, 3 - good). This indicates that the most complete solution is provided
by the UtilityNetworkADE, with runners-up being a tie in between LandXML and
IFC. These three solutions have different spatial design scales and scopes, the first
(CityGML coupled) is city oriented, the second (LandXML) directed to use at na-
tional level, while the third (IFC) is aimed at building to district level. As such, this
further indicated that the UtilityNetworkADE matches the use cases presented in
the dedicated section, 4.1. The other two runner-up solutions can facilitate similar
use cases, however, at different spatial scales.
Making a case for comprehensive data models of utility networks, [178] explains
that in the context of utility network renovation processes there is a lack of knowl-
edge of other existing networks. Most often, utility network operators operate their
own networks in close vicinity of other networks. When cities or companies concur-
rently refurbish their networks, each organization is responsible only for their own
network. However, due to to the nature of underground work and possible interac-
tions they should be forced to work together and share responsibility. Cities often
oblige these utilities to make public works at the same time, so as to avoid unnec-
essary repeated long duration construction sites in the same locations. This justifies
case for the further development and deployment of the UtilityNetwork ADE.
TABLE 4.5: Utility network modeling solutions, after [34], [23] and
own work, with three support levels, 1 - poor, 2 - basic, 3 - good
on XML and GML. However, GML encoding has been shown to inhibit develop-
ment, [133], by way of difficult CityGML files parsing, different geometrical data
representations, XLinks, and a lack of implementation support for ADEs. Provid-
ing an easy-to-use alternative to CityGML, CityJSON uses a simpler JSON encoding
that streamlines file parsing operations and allows for 3D models/digital twins in-
side these files to be easily visualised, manipulated, and edited.
Complete (this attribute should be taken with a grain of salt) spatial and energy the-
matic data sets originating in single data sources that describe the characteristics of
the objects to be modeled and simulated are difficult to come by, a topic discussed in
chapter 3. These have rather a speckled and inconsistent character. As such, the digi-
tal real world replicas (models) are generated with the help of different data sources
from disparate and very diverse energy chain actors. Therefor, the creation of ac-
curate digital city models and their following enrichment are time consuming and
expensive ETL data processes. To expedite this process and ensure the usefulness
of data, it must be usable for a wide range of applications. One way to ensure this
4.3. Digital domains for spatially aware cities 63
is the use of open standard formats such as CityGML. That was also the solution
chosen within the applications developed to help test the research questions posed
in this thesis. This aspect is discussed in [231] where different methodologies for the
rapid production of energy modeling prone city models are presented. For the work
performed to test the hypothesis in this manuscript the common feature of all those
formats is their open data structure.
Often, differences arise in between coordinate systems, which are at times local
Cartesian coordinate systems in BIM and CAD formats and geographical or geodesic
for GIS. The second category is most often included in the standardized 4-5 digit
EPSG system. This lack of foresight in data creation with local coordinates creates
further issues in handling combined GIS-BIM-CAD workflows. For this thesis the
conversion issues were handled in software tools such as PostgreSQL (with the ex-
tension PostGIS), FME, FZKViewer, QGIS or ArcGIS, City Editor, gModeller (both of
the last two are software extensions built on top of SketchUp).
Figure 4.9 depicts the funneling system and the tools used in the data conversion.
Most often, data provided is non-standardized. As such it requires extensive quality
testing before any standardized workflow can be applied to it. These processes,
referred to as ETL, Extract Transform Load, are previously discussed in chapter 2,
and can sometimes use more than half of the actual UBEM modeling time.
CityGML spatial models can be generated in multiple ways. In the case of the work
behind this manuscript it was mainly produced with three software solutions. In
the first case BIM or CAD (in IFC, gbXML or DXF format) data are extracted and
converted using FME. The same software can also add other UBEM relevant data
64 Chapter 4. Not all data are created equal - elephants in the room
possibly much more. Web Services generally involve communication over the inter-
net using protocols like Hypertext Transfer Protocol (HTTP), Simple Object Access
Protocol (SOAP), REpresentational State Transfer (REST), and eXtensible Markup
Language-Remote Procedure Call (XML-RPC) as a means of communication.
As [162] demonstrate, being able to easily discover relevant data is a prerequisite
to unlocking the potential of data. Online sharing of spatial data faces similar chal-
lenges to other distribution methods, such as file based. These include provision of
descriptive and structural metadata to enhance data’s discoverability and reusabil-
ity. The W3C (World Wide Web Consortium) recommends the following best prac-
tices for data sharing on the web, [226]:
• valid licensing information,
• transparency regarding data provenance and quality,
• data versioning to deal with data update on a scheduled basis
• provision of unique identifiers,
• interoperability enforcement by way of standardized, locale-neutral and rich
data formats,
• improved data accessibility and usability from flexible and stable API.
Several web service standards, often used with sharing spatial data in the energy
chain, are covered in this section. Easy and stable access to data on the web allows
users to take advantage of the advanced web infrastructure which often means that
by default the HTTP method is used. This provides access to data using atomic
transactions (atomicity is the defining property of database transactions, where each
SQL query gets wrapped in its own transaction). In practice most data providers
or delegated third parties support data access either through bulk download or
through an API, or both of them.
When specifically discussing the issue of sharing spatial data, most often interoper-
ability issues arise due to the data heterogeneity of this data category. More specific,
most data providers might use different data formats with their own data models,
which requires a special and tailored software solution to share their data. These
data-driven implemented web services cause issues for data consumers, as they are
locked to specific vendors and increase the difficulties of sharing of the spatial data.
To overcome the interoperability issues, the OGC has been established. This consor-
tium aims at providing standards both in data model and web services. The purpose
is to make spatial data sharing across any network, application or platform routine.
BCF data is stored in the cloud and enables concurrent users to sync BCF data in one
centralized location. The existence/presence of dedicated third-party BCF servers
are the main purpose for the development of the OpenBIM initiatives and help act
as the hub for such communications enhancing shareability among BIM users. BCF
is a powerful tool, similar to the WFS of the OGC, for BIM related data in the energy
chain and will impact the management of energy related facilities once BIM and BIM
data become part of the modus operandi.
The alternative to using the BCF-API for data sharing is the use of the open source
software BIM server. The Building Information Model server (BIMserver) enables
the handling and stockpiling of the information of a construction project by using
the open data standard IFC, [128]. The BIMserver uses a model-driven architecture
approach where IFC data is stored as objects, similar to a database where operations
such as query, merge and filter are possible. It also contains extra features (e.g. model
checking, versioning, project structures, merging). Other similar software in terms
of capabilities are all commercial solutions, [144].
The SensorThingsAPI was designed for things that sense and equips their users
with an open, geospatial-enabled and unified way to interconnect, thus constitut-
ing a standardized interface to the Internet of Things (IoT), [139]. It is an alterna-
tive solution to the previously presented SOS-SES-WNS, and can be regarded as a
lightweight SWE profile. The OGC SensorThings API architecture can be largely
separated in two parts, the Sensing part and the Tasking part. For this thesis, the
first was used as it provides a standard way to manage and retrieve observations
and metadata from heterogeneous IoT sensor systems. The implementation used in
the experiments was developed in Java EE by Fraunhofer IOSB, and is called the
SensorThingsServer and is of type Server.
FIWARE
FIWARE is a systematized framework, similar in scope to the SWE, provided how-
ever from a different perspective, that of the larger IT domain, and not specifi-
cally from the geospatial community. It does however offer integrated support for
geospatial data attributes. It is being developed by the FIWARE Foundation, also a
non-profit organization that promotes open standards across domains such as Smart
Cities, Smart Energy, Smart AgriFood and Smart Industry. FIWARE provides an
open alternative to existing proprietary internet platforms and it’s software ecosys-
tem is successfully constituting a reference option for developing digital twins, [56].
It’s development has received substantial funding and support from the European
Comission (EC) and has been supported by a large number of academic and private
entities, [201]. The project was started in 2011 and today provides cloud hosting
services based on OpenStack technology. It has a set of open standards APIs that
offer a number of added-value functions “as a service”, the Generic Enablers (GE),
[75]. These mainly facilitate the connection to the Internet of Things, the processing
of data and media in near real-time, and at large scale, the performing of Big Data
analysis. These APIs also incorporate advanced features for user interaction making
FIWARE an open alternative to existing proprietary internet platforms.
Important assets provided by using FIWARE are the abilities to manage spatial data
and multiple sensor networks (a specific use case that usually gives developers and
providers issues with handling different protocols and data structures), [74]. Specif-
ically for spatial data it includes functions for handling and querying (of interest for
this thesis’ use cases). Context information (data that characterizes the state of that
real-world object at a given moment in time) is allocated in the form of attributes as-
sociated with an instance of an object. One of the GE REST APIs, the Context Broker,
is used for that end. It can handle high data volumes, a concept known as Big Data
(or massive data), via a Hadoop framework. Similarly to the previously presented
SWE WNS alert manager, FIWARE contains a Complex Event Processing (CEP) GE.
Technical summary
A short summary of the capabilities of the two sensor frameworks with the three
available implementations included in this thesis’s experiments is presented in ta-
ble 4.7. The characteristics presented in the table purposefully do not include any
performance metrics, such as request response time or CPU use as the applica-
tions in which the software solutions were used are different as per the intended
use/characteristics. Ultimately, a choice of the best solution can only be done when
4.4. Web services lend a hand 71
Chapter 5
Within the energy chain one aspect is most important with regards to energy spatial
data, that of phenomena with both spatial and thematic varying properties. Long
before the term Internet of Things (IoT) or its associated data revolution, the energy
domain was producing large data sets of spaghetti like data sets, or better described
as continuous data bundles. Energy domain experts handling this data have focused
their attention to the thematic content and its functionality (energy relevant) and
have had less of an interest in the spatial character of these data sets. This can be
seen with the specifications of the most used electrical networks domain standard,
the Common Information Model (CIM), also a semantical standard, [59], developed
using UML, [236]. In general, network standards have been thematic oriented and
sensor data repositories had no specific association to location information.
brings added value, see table 4.1. These requests for connectivity are now coming
from a variety of actors including architects, urban planners, facility managers, spa-
tial scientists and policy makers. Often, different actors interact within the data col-
lection, sharing and consumption process, as they act both as a collector/intermediate
data publisher but also as a data consumer (for analysis and decision support) which
in turn produces further data. This emerging trend gives access to spatial data and
byproducts to an increasing number of parties and thus enhances reproducibility,
collaboration and decision making. As such, considering the definitions presented
in 2, energy data with spatial attributes plays a vital role in smart city energy appli-
cations.
As with any nascent field no specific domain can claim ownership till expertise and
workflows are developed. That meant that both the W3C and the OGC have looked
for solutions for the problem of sharing large bundles of data that also holds spatial
characteristics. One of the main sources for this heightened interest is the current In-
ternet of Things (IoT) revolution, already mentioned in the ongoing trends with next
generation cities in 2. IoT requires functional interoperability with other products,
[226], and is interesting for this thesis as large parts of the data produced by this
domain are energy relevant (e.g. accurate occupancy for buildings, detailed weather
information, advanced forecasting for the aforementioned themes, energy use val-
ues, air pollution measurements). To achieve this capability, platform independent
APIs for application developers are needed, and a means for different platforms to
discover how to inter-operate with one another, [226]. The specific needs of this the-
sis requires the present work to focus on standardized frameworks that allow for
measured data (from sensors) and its spatial information to be stored together. The
suitable identified candidates are FIWARE (The Future Internet platform) developed
as a by-product of research paid for and organized by the European Commission,
and the SWE (Sensor Web Enablement) of the OGC with the SensorThings and SOS
implementations.
5.2 Experiments
Within the scope of this thesis was managing the flow of energy related measure-
ments and incorporating them within the digital representation of a city. This is
why solutions were explored that allow things to be spatially located and for their
observations to receive location attributes. Such capabilities help enable distributed
energy production and a quantification of the environmental influences on that pro-
duction. Understanding where a measurement is made can make a world of differ-
ence. To provide a simple example, temperatures measured in the shade or in the
sun provide very different results even if the location is only but a few meters apart.
In order to abide by the good practices (presented in section 5.4) that this thesis seeks
to respect experiments were carried out as to the suitability of the previously pre-
sented web services solutions. The focus lies with the distribution and availability
of open standardized spatial and thematic data with public and private authorities.
A fully functional workflow, that facilitates the flow of information with regards to
UBEM models, in a solution that is purely based on web services was tested. The
implemented workflows aimed at paving the way for a seamless integration of spa-
tial CityGML and sensor data. On behalf of sensor data, SOS, SensorThingsAPI and
FIWARE successfully linked to spatial data and allowed for the linking of these dif-
ferent data sources, while for CityGML four solutions are presented.
5.2. Experiments 75
F IGURE 5.1: Sensor installations used, left hand side presents images
from [95], down-middle EIFER and ULPGC building test site, down-
right from [224] (left and right pictures: green façades on vineyard
house in Karlsruhe, Germany, an multi-family house in Ettlingen,
Germany, middle depicts a green roof installation in Illingen, Ger-
many)
H2O(g), O3, NO, NOx, CO and CO2, in addition to weather indicators like air tem-
perature, RH (relative humidity), P (air pressure) and horizontal wind. Its loca-
tion was measured using an integrated antenna for Global Positioning System (GPS)
measurements, [95, 242]. Data for both of the monitored train lines (S1/S11 and S2)
was offered in support of this research work for the purpose of sensor infrastruc-
ture experiments. Part of the train track runs in the Rhine valley, while a second
part continues up into the Black Forest (German: Schwarzwald). Measurements were
made at short time intervals as the train moved along its track, making the sensed
data a perfect candidate for our study with changing spatial, temporal and thematic
characteristics. The measurement campaign occurred prior to this thesis, and there
was no possibility to test live data feed from the tram onto the software platforms
proposed into this thesis. Data was provided in R format and then converted and
stored into a local database from where it was further used as sensor data for this
thesis’s infrastructure experiments.
Solar installations testing for vertical façades
The second use case includes data from a UV measurement campaign on a static lo-
cation in two buildings. One was located on the university grounds of Las Palmas de
Gran Canaria (ULPGC), , the most populous city in the autonomous community of
the Canary Islands, Spain. The second monitored facility was located in Karlsruhe,
Germany, the second-largest city in the state of Baden-Württemberg, in southwest
Germany, near the French-German border. Both buildings host tertiary activities,
namely research and development. One sensor installation was mounted on each
building facade (four in total for each location). The objective of the sensors instal-
lation was to provide data support for modeling activities. The model’s purpose
was to determine suitability of commercial solar energy installations on vertical fa-
cades in two very different locations, central European, with continental climate and
southwest European/northwest African, with sub-tropical climate. Both implemen-
tations have been realised under similar conditions and used the same equipment.
The sensor installation was designed and built in-house from separate purchased
parts. The sensors were directly situated on a window of each facade of each build-
ing, eight for the entire experiment. For radiation measurements the model SI1145
Digital UV Index/IR/Visible Light Sensor, from SiLabs was used. This equipment
can measure infrared as well as visible light and using a calibrated light sensing al-
gorithm it can provide the UV Index, which provided the input to further data mod-
eling in this experiment. The sensor was attached to a Arduino Uno Board REV3
with a UartSBee V5 - Xbee Adapter attached to it. It picks up the measurements
of the sensor and sends it to a centralising repository by using the Bluetooth capa-
bilities of the Xbee adapter. The central repository unit receiving these data was a
Globmall ABOX Raspberry Pi 3 Model. As such, all data produced was able to flow
in live fashion to the data repositories and make use of a standardized web services
protocol.
Green roofs/façades monitoring
The third use case was a sensor installation developed to quantify the impact of
green roofs and green façades on a building’s energy demand. The sensors were
installed in three buildings located in the region of Baden-Württemberg, in south-
west Germany. One building included a green roof installation and the other two
had partial green façades on the side of the buildings. The objective of the installa-
tions was to measure the impact of green roofs and façades on the building heating
5.2. Experiments 77
and cooling demand. Parallel to the sensor installation a model was developed that
can provide estimates of the impact of such external building extensions on the heat
transfer based on the heat balance principle of foliage, soil, and structural layers,
[224]. The measurement campaign was performed during the summer and vali-
dated the model with regards to the estimates of impact on cooling.
The sensor installation was developed starting from senseBox, an educational sen-
sor toolkit, built on top of an Arduino, [173]. It provides (in this experiment) well
documented, portable interactive sensors for reproducible research based on open
source software. The data produced was released as open data on the platform’s
interactive website. It included multiple off the shelf additions and performed the
following measurements: air temperature and humidity (sensor HDC100x), air pres-
sure (BMP200), visible radiation (TSL45315), wind speed (Aenometer) and concrete
and substrate temperature (DS18B20). In two of the locations internet was available
and the data was able to flow live to the central data repository, located at the KIT,
while electricity was provided from large batteries. In one location (Karlsruhe vine-
yard house) no electricity or live data feed was possible, as such data was stored on
a memory card and then transferred to the repository post measurement campaign.
A summary of the use cases and associated measurements and technologies is pre-
sented in table 5.1. This table indicates the location, the objectives of each experiment
and the measurements that were taken in each location and whether or not live data
feed was achieved within the trial.
78 Chapter 5. Energy relevant coupling of sensors and 3D models
stream the chosen protocol, although the simplest of the three proposed solutions,
offers sufficient capabilities for this use case: data gathering from sensors, well es-
tablished software solutions, with associated documentation and standardized data
storage.
The sensor data flowed from the sensors attached to a Arduino Uno via two Blue-
tooth adapters to a Raspbery Pi which then sent the data further to a central data
repository located on an EIFER PostgreSQL server using the SensorThings API database
structure. With regards to the location without internet access (Karlsruhe vineyard
house use case), the data had to be transferred manually, every second week during
the measurement campaign as the power source (battery) was depleted.
Figure 5.2 encapsulates the concepts presented above. An important mention is that
a real live implementation of the SOS-SES-WNS would not use the R format file as an
in-between the sensor and SOS. The R format was chosen only due to availability on
the side of the data hosting entity (KIT-IMK) and is valid in the presented experiment
as the measurement campaign occurred prior to this thesis.
F IGURE 5.2: Sensor data infrastructure, left SOS-SES-WNS, used with AERO-TRAM, middle SensorThingsAPI used both in the Vertical
solar installations testing and the Green roofs/façades monitoring, right FIWARE used in Vertical solar installations testing
5.2. Experiments 81
Spatial data in 2D that is relevant for the energy chain can be shared via web services
(such as the WMS, WCS or the WFS) in a standardized manner. These are presented
in their dedicated section, 4.4.1. The data can be sent for further use once ETL op-
erations are performed and the quality of data has been checked. This subsection
narrows its focus on the flow of spatial 3D city models. In this thesis’ case, these
were always stored in CityGML, either as files or in a database version of the stan-
dard, the 3DCityDB. This data originated from open data or it was built from open
data and is in general restricted to building models.
Three methods to produce CityGML data are presented in the section 4.3.5. These
methods were used directly in this thesis to produce CityGML files. Four additional
methods to obtain CityGML from open data are discussed in [231]. These are (1)
using building footprints and remote sensing data, (2) converting Open Street Map
(OSM) building data, (3) using LIDAR data sets and open data for building foot-
prints and (4) using geoportals, in case a city has one available, see figure A.7 for a
detailed presentation of each workflow. When CityGML samples could be directly
obtained, extensive testing had to be performed in order to ensure data quality. A
sample was provided by the public authorities of Karlsruhe, while for the building
in Spain it was built from openly available data.
In the case of CityGML data transmission via web services, different working solu-
tions were attempted. Figure 5.3 presents the different implemented solutions. Two
of these were purely CityGML serving oriented (one used the 3DCityDB WFS, and
another one used the GeoServer), while a third looked at a workaround for pro-
viding non-standardized spatial data, that can mimic CityGML data characteristics
(and can later be used to generate 3D city models). All approaches aimed at testing
the viability of such a solution in a completely open-source environment. From the
previously presented web services, the WFS (Web Feature Service) is found by [186]
to be the most suitable OGC standard for supplying real geometry data. Table 5.2
presents the implemented solutions with their associated advantages and disadvan-
tages.
3DCityDB WFS
As CityGML is much more than geometry, semantics and topology are big part of
the reason for which the standard has seen implementation and adoption by pub-
lic authorities. The first implementation of a useful WFS service for CityGML came
with [122], where the 3DCityDB library contained an OGC compliant WFS imple-
mentation working with the 3DCityDB schema. This implementation is referred to
as the 3D City Database WFS. With its open source software version, simple online
access to the 3D city entities is possible directly to the database, [163]. With the cur-
rent version (at the time of this manuscript writing v. 4.1) the following operations
are supported: GetCapabilities (metadata information regarding the server provid-
ing the WFS), DescribeFeatureType (returns a schema of the CityGML features that
are invocable via the WFS), ListStoredQuerries (provides a list of the cached queries),
DescribeStoreQuerry (makes available detailed information regarding the aforemen-
tioned queries) and GetFeature (returns a selection of CityGML features from the
3DCityDB using a query). A commercial solution, the Virtual City Systems Web Fea-
ture Service (VCS WFS), proposes a more complete solution where CityGML objects
82 Chapter 5. Energy relevant coupling of sensors and 3D models
F IGURE 5.3: Four suitable approaches for CityGML web service, after
[242]
can not only be queried from the database, as in the previously mentioned open-
source version, but can be filtered (spatially or thematically), as well as inserted,
edited and deleted, [215]. This first implemented solution provides only partial res-
olution to the research question and only focuses on the retrieval of CityGML data.
GeoServer App Schema
The second technical solution used, GeoServer, is an open source server, known
for its purposely interoperable design, [48], [131]. It was investigated if serving
CityGML via a WFS with advanced functionality is possible, with complete results
presented in [242]. This would allow for data exchanges based on specific fea-
ture types and complex queries to be performed. The implemented solution used
GeoServer (as it also respects the OGC WFS standard and provides all tools: dis-
covery, query, locking, transaction and stored query operations). It was found that
the tested solution can cover almost all CityGML classes with the exception of the
Generic Object one. For modeling the urban energy chain this represents a signif-
icant downside as the Generic Object can handle all objects not described by the
CityGML standard (in this thesis’s specific case, energy relevant data, such as build-
ing characteristics). These issues in between CityGML 2.0 and GML encoding rules
of GML refer to the fact that in the GeoServer Application Schema, all public GML
application schema used must respect the GML “Striping” encoding rule (a com-
plex type cannot be the direct property of another complex type). The generic.xsd
schema does not obey the GML encoding rules. In addition CityGML models often
use complex data types for properties and group features within feature collections
(this generates very useful multi-level hierarchical database structures, see the pro-
posed identifier structure in chapter 3). Another possible limitation stems from the
types of geometry used with CityGML which are more complex than the limited ge-
ometry types supported in relational databases. As such this second tested solution
is also incomplete.
A practical workaround
The third approach stems from a recognition of the data availability reality. This
means that most often available data is made available in SHP file format, it is not
standardized and incomplete, a fact debated broadly in chapter 3. Therefore, a solu-
tion was conceived to work within this reality: SHP file data that was subjected to
5.2. Experiments 83
ETL operations and then stored in local repositories. From there it was either fur-
ther shared via standard WFS, converted to CityGML with the use of Python scripts
or converted to glTF (graphics language Transmission Format) using the CesiumJS
application. Inside the shared data, an extra attribute was offered, that contains
the number of floors for each building. Using this attribute, the interface mimics
a CityGML LOD1 model in which an extrusion of the building is created. This is
useful as it generally replicates the volumes of the buildings in the city without the
shape of the roofs. In the background this data can be used to create 3D models and
perform detailed spatial analysis relevant for UBEM, such as shading analysis.
CityGML RESTful web service
Following an attempt to identify other available solutions, similar in purpose, in
literature, one such example was identified in the work of [186]. They compare
SOAP and REST solutions and propose another methodological approach based on
REST-style architecture. This allows for the acquisition of CityGML data based on
semantics instead of following on the same track as the standard OGC WFS, which
is designed to retrieve, visualize and modify data based on geometry characteristics.
The authors of the aforementioned study focus their efforts on the building class, as
it is the most commonly used with CityGML. They also provide a successful path
for integration into a virtual data hub for all the other CityGML classes. In their
proposed solution one can also query data beyond geometry, using topological and
semantic rules and most importantly the standard extensions, namely the Energy
ADE and the Utility Network ADE. This solution does fully cover the issue at hand,
and offers significant advantages from a perspective of energy relevant data transfer.
However, it is for the moment only a theoretical implementation. It does not include
data management such as update or edit abilities.
Technical summary
TABLE 5.3: Web services used in BIM, CAD and GIS for geospatial
data relevant in UBEM
Table 5.3 presents web services encountered in the energy chain while testing hy-
potheses included in this manuscript. These web services, either used in applica-
tions developed and tested by the author, either serving as input, were integrated
in an infrastructure that was purposely designed to have a GIS centric perspective.
This happens as most data delivered per web service in this domain tends to be
stored in repositories that have GIS capacities. In addition, the funnel system devel-
oped to load data in the central repository of this thesis mainly imports 3D GIS data,
which means that data is converted or obtained a priori to feed it into the system.
Several possible parallel approaches are presented in the tight vs. loose coupling
section, 5.5.
Orion, Cygnus, NoSQL DB. The author finds that the FIWARE solution provides a
significant benefit for smart applications due to the very large capabilities bundled
together. It is a conclusion also shared in the study of Salhofer, who finds that this
statement holds true also for their tested use cases, [201].
What this manuscript’s reader should take away is that a proper use case definition
before any implementation is done is paramount. This will facilitate the choice with
regards to the proper implementation based on the table 5.1.
For the data modeling of the energy chain, one of the main issue is related to the cus-
tom devices from the energy and utility domain. These experiments show that they
can be modeled alternatively either using CityGML extensions (e.g. AbstractDevice
inside the UtilityNetwork ADE) or even inside FIWARE (with the DeviceModel, its
schema allows for the creation of custom devices that monitor ambient conditions
or state of components within a system). Within SensorThings API, the data model
contains a table entitled FeatureOfInterest. This entity was used to describe the build-
ing that housed the sensors. It was then further linked by way of an identifier to the
CityGML object. This information can be then inherited to the CityGML object that
describes the building.
The way in which data connectivity (in terms of the same object belonging to mul-
tiple data sources) was always assured in our implementations was via the unique
identifier method presented in chapter 3.6, that then allows for the same object to ex-
ist in two systems and to simply be referenced from one system to another by means
of XREF links. CityGML, FIWARE, SensorThings API and the SOS data model
specifically allow for this to exist. Similar work in terms of scope to the one pre-
sented in this chapter was submitted by Chiang and his colleagues, [52]. This work
approaches the integration of things and semantical city models by using the OGC
SensorThings API and OGC CityGML via Semantic Web Technology. The authors
present an integration ontology to connect data from these two standards that in-
cludes different perspectives and definitions of a Thing. To connect information from
the CityGML and SensorThingsAPI, SPARQL queries were used. It represents an-
other step forward and potentially a useful lead for developers implementing smart
city apps.
In general, this proposed sensor - urban energy data integration by way of spatial
links can facilitate the integration of IoT and/or energy data with 3D city models to
achieve truly interoperable smart cities. However, this work represents only a step
towards this goal. It can be regarded as a facilitator, if the distributed data/loose
coupling solution is adopted, as presented in the dedicated section on tight vs. loose
data coupling, 5.5.
The presented work includes no performance metrics, a potential drawback, even
though in the experiment phase performance and load tests were conducted. These
metrics cannot be used for comparison, as the services compared operated in differ-
ent use cases. Only SensorThings API and FIWARE could be directly compared in
the second tested constellation (Karlsruhe with the first and Las Palmas with the sec-
ond), however the two services belong to different weight categories when it comes
to the capabilities and support infrastructure and software. This is why the only
technical comparison that is included in this manuscript is to be found in table 4.7.
It should however suffice in helping practitioners orient themselves towards a pre-
ferred choice.
86 Chapter 5. Energy relevant coupling of sensors and 3D models
As can be seen in the presented work GIS has a slight lead on BIM when it comes to
standardized data sharing and availability. This is true when it comes to commercial
use, web services and data availability. As such, it supported the idea that the main
data infrastructure used in the applications of this thesis follows a CityGML centric
approach. This hypothesis was put to the test in the thesis life. A similar proposed
approach that stores all data in GIS and focuses on converting BIM to GIS was pro-
posed and tested by [237]. The authors partake in a GIS focused experiment with
limited conversions of BIM data and conclude that for meaningful city wide impact
to occur communities wide planning units and city scales with building units are re-
quired. In many ways this work draws parallels to the one presented in this chapter.
nation of Estonia. In order to provide a solution for a functional data sharing infras-
tructure the Baltic state of Estonia has built a framework with virtual government
services called e-Estonia, [153]. Their service is built on block chain technology in
order to uniquely identify and secure transactions and offers citizens and companies
more than 4,000 virtual services. In fact 99% of their public services are accessible
online via a one-time login gateway, [174]. All spatial data that are owned by the
Estonian state, local governments and other legal persons governed by public law
are published and made available in a single point of access, [145]. This reliable
and secure way of exchanging information, based on transparency between citizens,
companies and the government, seems to have fostered faith between those actors,
with most IT framework projects started by the government in the late 2000’s (the
geospatial portal and the medical DB for example in 2008) and paying off in public
confidence. According to the 2015 spring Euro-barometer, [180], 51 percent of Esto-
nians trust their state, compared to the average of 29 percent of all Europeans. This is
related to the so-called snowball effect, as once the standardized web services were
put in motion and people started using them, there was an ever increasing depen-
dency on these services and interest in their further development.
The previous literature references focus on public actors. A balancing of this per-
spective is offered by [197]. The author presents solutions for companies seeking to
benefit from the ongoing digital transformation. They recommend that that digital
conscious companies should take part in standardization efforts as this is a mutu-
ally beneficial endeavour. Participating, brings in needed know-how and models
the further development of standards with the firms involvement.
In the energy chain specific spatial extent, all actors need to operate according to the
law, with accountability and transparency guiding political and business practices.
This should go further, using foresight to plan adequately to developing trends, such
as the current energy market transformation. At the same time stakeholders need to
be responsive to the needs of other stakeholders, and be proactive in critical issues
such as land use, pollution, and the fairness and honesty of political and business
practices.
their own SDI to the publicly available one via web services. Further facilitating this
links would be the unique identifiers proposed in chapter 3 and the use of spatial
entity marking and scales (presented in the same chapter mentioned before).
With an adaptable (sometimes referred to as a generic) SDI a city can able to be mod-
ify it for a new use or purpose. Generally, an adaptive system is able to respond to
changes in use or changes in the interacting parts, [5]. Software specific in a generic
system is their ability to support a variety of architectural models and adaptation
mechanisms. One of the features that commonly appears in such systems is a feed-
back loop. It ensures that the system can deal with the current requirements. In
addition such systems often present hierarchical structure.
Lastly, the reusable aspect, a key note in software architecture, as [157] testifies. Soft-
ware reusability is their ability to be reused without requiring further development
and should not be confused with the adaptable characteristic mentioned above. In
the case of a city SDI, whatever development is done of it, it should be made with the
direct intent of it being used more than once, not a one off solution that potentially
wastes the organizations resources.
Even though it would be a major advantage, using only open source SDI is not a
condition per se. It would however facilitate integration with other tools and ac-
tors. What is required of the SDI, is that it uses open web services, can work (im-
port/export) with open standards and has non proprietary entry/exit ways and data
storage capabilities. Architectural choices are further discussed in their dedicated
section 5.5.
in one source, they still very often entail very different characteristics, with a lack
of standardization presenting a large impediment. It creates the need for extensive
ETL processes, as presented in chapter, 3.
Linked to the open data initiatives by ideology open-source software (OSS) is defined
by [132] as a type of computer software whose source code is released under a license
in which the copyright holder grants users the rights to study, change, and distribute
the software to anyone and for any purpose. The success and impact of this software
was being predicted as far back as 2002, [35] and [137] when already OSS packages
were becoming available outside of the scientific community.
There are large variations of the licence models, however, as our spatial extent in-
cludes various actors with different interests. It is not the purpose of this thesis to
evaluate which model suits best what situation but rather to express the fact that
the entire process modeling and simulation of the energy chain can be performed
with OSS, a position also confirmed by [86] with regards to open source energy sys-
tem optimization tools. This approach enhances interoperability but does not ex-
clude connections to other commercial tools due to the open specifications of data
input/output mechanisms, a fact which is especially important as legacy software
exists in many of the entities from the energy chain.
In general, open source licenses can be categorised in two categories: Copyleft and
Permissive Licenses with the main differences occurring in compliance requirements
and the license of products developed on top of the current products. Permissive
licences promote re-use, including further assimilation into commercial products,
without requiring improvements to the code to be released at all. A copy of the
license text and the original copyright notice must be included in the newly licensed
code. Copyleft licenses have the same requirement as the permissive licences, with
the addition that any further development of the code must be made available under
the same license as the original. Table 5.4 presents the most common licences for both
data and code.
TABLE 5.4: Common licence models for code and data, after [185] and
own work
In addition to file based standards, open (as in publicly available standards) web
services are required to standardize the input and output of information from data
owners or data managers to companies and citizens that can produce value out of it.
Because ultimately data is only valuable when it is used.
During the modeling, simulation and monitoring of different use cases in the energy
chain, this manuscript’s author aimed to use open source software and open data
throughout the experiments. Figure 5.5 presents at least one open source solution for
each of the domains that intertwine in the energy chain: spatial analysis, modeling,
92 Chapter 5. Energy relevant coupling of sensors and 3D models
simulation, monitoring and statistical analysis. This point is also made in [167] for
the same given context. QGIS together with PostgreSQL and PostGIS were used for
spatial analysis and data storage, Python for modeling, C++ for simulation, R for
statistical testing and analysis and monitoring was performed with open hardware
solutions built from scratch or on top of senseBox, an educational sensor toolkit,
built on top of an Arduino (with an open source operating system).
The use of open standards and the publication of contiguous open data sets are pre-
requisites that, in addition to building public trust in authorities, increase services
quality and create added value for cities and all other energy chain stakeholders.
managing and modeling the city wide energy chain is the fact that very often, data
is still present only in paper form or basic digital sharing formats, such as PDFs. As
a consequence, there are many great efforts to digitize (creating a digital copy of)
still existing analog content, [147, 83, 204]. However, digitization per se, is only a
step towards the conversion of data and it requires to be done from a perspective
of digitalization, where the digital products are to be integrated with other ones.
This is where the previously mentioned open standards come to assist, a so called
digitizing towards open standards. This would create the premises for a successful
digital transformation of the energy chain.
Digital by default should ensure that the ETL operations discussed in Chapter 3
are greatly reduced. Even though the pyramid presented in figure 5.4 has Digital by
default as a smallest step, it is the spearhead that can further all the other three pillars
of good digital practices.
structure. It facilitates the linking of all the previously mentioned categories of data
and can be easily linked to standardized data structures. The disadvantage is that the
relational data structure is more often than not lacking standardization. In addition,
hosting many different types of data together in a single repository forces the data
owner to use a one size fits all approach in terms of the databases and software. This
bundles in together sensor data, geospatial data and different energy chain thematic
data with various degrees of success.
An alternative to this approach is loose coupling. This method does not centrally
store data. It fact it does the exact opposite, and that is to keep it within different
repositories that have different data owners. A good practice within this approach
is that data is often treated is that from the very beginning and classified, for ex-
ample into databases with nonSQL or SQL characteristics. Figure 5.7 presents the
architectural concepts behind loose coupling.
However, even though separate, this data remains accessible to the outside world
via a virtual data hub. This hub ensures standardization of data available to the
spatial actors. The owners of the actual data repositories data usually have different
perspectives, energy thematic, spatial thematic or urban thematic. For this second
method to be efficient and handled with ease, standardization implementation needs
to be accepted on a wide scale.
As with the first method there are advantages and disadvantages. The main advan-
tage is that data is handled by their thematic experts, who are well aware of existing
standards in their field. Also, these repositories tend to be well maintained as they
are the bread basked of many of these spatial actors. A disadvantage is the lack of
control on the side of the virtual hub owner as to the content of data. This is how-
ever dependent on weather the access granted to the repositories is bi-directional,
meaning Get and Write, or unidirectional, meaning Get Capabilities only. Figure 5.7
presents the architectural concept of loose coupling.
The study [197] finds this second perspective useful and states that data ownership
per se should not be a priority, but rather data accessibility. Ensuring that information
5.5. On data fusion - to merge or not to merge 95
Both of these solutions were implemented when dealing with the experiments pre-
sented in this manuscript, 5.2, and the UBEM modeling activities, 6. The loose cou-
pling presents the most advantages as it shortens ETL processing times, and always
provides access to the latest data available. However, the tight coupling implemen-
tation also provides a working solution, which can, in certain cases be the only viable
way to follow.
The loose architecture also embraces REST principles, allowing for interaction with
Uniform Resource Identifier (URI) available resources (in practical terms, over the
internet). In addition, the loose coupling architecture creates the premises for good
practices, as they include error handling in the inner APIs and prevent error affected
data of spreading in the data produced further down the pipeline. Adding new
components is a fairly straight forward procedure once the workflow has stabilized,
making the architecture scalable and flexible. This is for example how the infrastruc-
ture evolved to include monitoring data, from a first use case oriented only towards
UBEM, and then towards Things and energy monitoring for model validation. These
statements are also in line with the findings of [159], in a specific energy chain use
case, and [138] in a more general perspective.
5.5. On data fusion - to merge or not to merge 97
To underline the fact that one most not simply choose among the two architecture
views, a mixed file/web service solution based on CityGML and sensor web services
is presented in figure 5.8. Similar situations are presented very often to practitioners,
where a purely loose solution is not feasible due to certain restrictions regarding
the data of one energy chain or another. Energy producer data is to be stored in
CityGML and the EnergyADE. These producers can be large entities, but can also
be a single family residence which owns a solar panel and is selling electricity in the
network. Consumers can be represented using the same solution. Networks require
the second extension, the Utility Network ADE. Lastly the monitoring data can be
stored using the EnergyADE, the DynamizerADE or another database solution via
the sensor specific web services mentioned in the dedicated section, 4.4. All types of
data can be either stored as a file or in a database depending on the use case.
Finally, a choice in between the two architecture types should be made once use cases
are defined, user categories have been properly identified and user stories have been
outlined, following the AGILE development logic. Further, all potential data sources
need to be identified, with entry/exit ports definitions and legal requirements estab-
lished (data and code licenses).
99
Chapter 6
At the interaction of the GIS, urban planning and energy modeling worlds lies a
nascent field, [191], called Urban Building Energy Modeling (UBEM). It is main ob-
jective is the modeling of dozens to thousands of buildings using bottom-up ap-
proaches and is somewhat opposite to detailed individual Building Energy Mod-
els (BEM). In BEM, practitioners perform energy use analysis of the building in-
ventory at the level of a single building in high detail. UBEM tools, standards,
and paradigms have been identified in the previously cited manuscript and are of-
ten simplified BEM tools that use streamlined workflows. In addition, UBEM also
provides hybrid methods, merging single building energy models and broad scale
building stock modeling, that survey the energy analysis of districts and cities. In
this context multiple studies show that the emerging science of UBEM (Urban Build-
ing Energy Modeling), can successfully provide support for different design or pol-
icy decisions with acceptable error ranges, [121], [164].
UBEM belongs in general to the Urban Modeling field, together with its sister field,
Urban Energy System Modeling (UESM), [202]. UESM is a discipline, that focuses
more on the energy solution in the energy chain. As depicted by Samadzadgegan
and colleagues, UBEM provides information on the energy demand to UESM, and
together provide for complete solutions in the modelling of urban energy chain.
F IGURE 6.1: UBEM model classification, after [76], [191] and [103]
The first type, physics-based dynamic simulation, tries to mimic a building’s energy
behaviour up to the smallest possible resolution, both spatially as well as tempo-
rally, identifying all sub-components responsible for energy flow, in and out as well
as within of a building. The second type of models, reduces the complexity of the
first models, and thus requires less data input. They are able to offer information
on building energy behaviour for large building numbers in a relatively short time
span. This category is also where the library of models proposed within this chapter
belong to. Lastly, data-driven methods rely on large quantities of statistical and/or
measured data. These usually include standards (ISO, DIN) based classifications of
buildings per different categories with associated occupancy and thermal charac-
teristics. Sensors then provide the ability to determine building characteristics and
match it to existing database types.
An integrated presentation of the two methods for classifying UBEM models is
found in the study [77]. This is presented in figure 6.1. What can be observed in
the figure is that ultimately both top-down and bottom-up methods can be physics-
based, reduced order approaches or data drive. It all depends on the availability of
data and the intended results use.
• Monitoring data.
The first category refers to building geometry and its physical location in the real
world, while the second one regards the physical properties of the materials that
form the building, as heat capacity and thermal transmittance of exterior building
elements (walls, slabs, roofs) and optical properties of windows. In the third cate-
gory all weather related variables are included, outside temperature, solar radiation,
wind speed, humidity, cloud coverage. The fourth category includes variables that
help classify the behavior of building occupants and profiles for heating, lighting
equipment and ventilation. The solutions proposed in this thesis fall in the field of
UBEM, as large city-wide spatial models were produced and combined with build-
ing physics and socio-economic indicators to provide a detailed approximation of
the city. Lastly, monitoring data is used for calibration and validation of results.
When the building inventory is described using the previously mentioned param-
eters a virtual (or digital) twin is created. This can also be referred to as a virtual
prototype - a dynamic digital representation of a physical system and of the spa-
tial entity, [146]. These twins are generally created using a software package, or a
combination of multiple software.
The UBEM workflow can be summarized in five steps, graphically depicted in figure
6.2. These are as following:
(1) First, the objective definition occurs. This entails the narrowing down of large
general national, regional and local policies down to local objectives. This ultimately
drive the energy model and scenario definition. In general results are presented as
Business as usual, where the status quo ante is quantified, and other scenario, where
different implementation measures are applied and their impact is quantified, rela-
tive to the status quo ante.
(2) In the second step, the spatial description of city entities is created. This step
involves extensive ETL processes that clean and validate data. Multiple data sources
(GIS, CAD, BIM) are combined to create uniform spatial data sets that represent
buildings and networks, components of the urban energy chain. Further, UBEM
relevant data (energy parameters) from multiple sources that can be used for energy
modeling is identified. Again, comprehensive ETL processes are employed, that
finally allow this data to then be merged to the previously mentioned spatial data.
These steps are presented in figure 4.9 and correspond to the section dedicated to
the data funnelling system section of this manuscript, 4.3.5. In this thesis manuscript
and in the presented experiments, this data is stored and linked to CityGML building
objects that contain hierarchical relations based on the identification system defined
in section 3.6. Once this step is complete the previously mentioned digital twin is
obtained.
(3) In the following step, simulations are performed on the data, generating demand
estimates under different assumptions (scenarios) and other energy use related KPIs.
This step is performed with certain given assumptions over a predetermined time
duration. The given graphical depiction of figure 6.2 assume that a model library is
already in place that can be used with minimal effort.
(4) After the results are obtained, calibration and validation is performed, either by
comparing against statistical data or by use of the sensor data, see section 5.2.3 with
regards to the types of data, services and storage solutions. Finally the results are
6.2. aEneAs - Trojan hero’s odyssey 103
describes using a report or publication. Often, data is made available on open data
portals.
(5) These results are then used to achieve the goals of the UBEM modeling process,
to manage and improve the urban energy chain. These tasks, required for urban en-
ergy modeling, are prerequisites in all energy planning and management activities.
This is done in order to understand the processes and the relevant scales at which
they happen, as [99] indicates. This is due to the fact that optimal energy decisions
in city/district planning and design help optimize whole-system performance (city
and district level).
Based on the results presented in [149], it has been observed that a fifth of all UBEM
publications encountered work with EnergyPlus as the simulation tool, a physics-
based dynamic simulation engine. This happens even though the tool cannot make
use of CityGML data models, the most often used data set in UBEM, according
to [103], which means that ETL operations are involved. Chapter 4, section 4.3.5
presents and discusses data sets stemming from GIS, CAD, BIM and other relevant
data sources and methods of converting these into CityGML. In addition to these
methods, the ever greater availability of open data sets stored in this format made
it the ideal choice to support UBEM activities. This is why a library working exclu-
sively on such data input was proposed.
As such, it became part of the goal to show the viability of 3D semantic city models
as a data hub for energy modeling. As it is standard with software development
efforts, certain decisions related to the environment of development were made in
the very beginning. In short, CityGML allows for the storage of standardized 3D
semantic city models and has the ability to contain standardized extensions that are
built on top of it. One of the key things that needs to be mentioned here is that be-
yond its imperfections and difficulties with implementations there exists, still, no
viable alternative open standard that competes in terms of scope with CityGML,
at least none that this author has interacted with. In addition there is a community
of practitioners that stands behind it and offers support and develops open-source
tools. Beyond the standard, aEneAs library makes use of the work of standardization
within two consortia that the author of this thesis has been working in, one develop-
ing the Energy ADE and a second one the Utility Network ADE. Both the standards
and the extension concepts are further detailed in the aforementioned chapter.
Earlier versions of this infrastructure were presented and discussed in [166] and
[231]. When comparing to the previous versions of the infrastructure, the one pre-
sented in this manuscript includes additional components required for the coupling
of sensors with 3D spatial models and web services for data flow. The storage of
sensor data is made inside two databases, one PostgreSQL (for SOS, FIWARE and
SensorThingsAPI data) and another in MongoDB (a NoSQL database specifically
integrated in the workflow for use with FIWARE).
On the upper side of the same graphic, 6.3, visualization tools can be observed.
These access data stored in the two DBs via web services that at times permit even
data management inside the repositories. For more details on the web services
please see the dedicated section in 5.2.
The following tools and extensions were used extensively in the work for aEneAs:
• PostgreSQL, an open database system,
• PostGIS, a database extension that enables spatial data storage and calcula-
tions,
• Python, a programming language with a very large user base.
PostgreSQL is a free and open-source relational database management system. It
emphasises extensibility (an IT concept that provides space for further growth) and
SQL compliance. The extensible database management system was originally de-
veloped at the University of California and uses a relational data model (data is
organized with hierarchies, and structures), [195]. The choice for an SQL database in
a city environment where data can be provided in very large quantities, may seem
striking from that perspective, however, it is the urban context that provides hierar-
chies itself, with cities owning towns, who in turn contain neighborhoods, that are
composed of multiple districts with dozens to hundreds of buildings.
Enabling the storage of spatial data, PostGIS is a database extender for PostgreSQL.
It provides more than a thousand geospatial functions that work with both vector
and raster data, [15]. It supports all the open 2D and 3D spatial data types as defined
by the OGC. Additionally, a second extension made the modeling of networks pos-
sible. This is PGRouting, [130], which was used extensively with all network models
developed in this library (integrated water and electrical networks, iterative )
The coding of algorithms, as well as a functional coding interface, is programmed
in Python. This is a programming language that has a very large base of users,
[209] and support libraries. The large number of users means that it is attractive to
other potential users and developers. Additionally, it can connect to the PostgreSQL
database, easily import different types of file standards and provides for mecha-
nisms that allow it to be connected to other stand alone software and platforms.
were used, representing a relational schema of the EnergyADE and the UtilityNet-
work ADE, explained in their dedicated section 4.3.3, and documented further in
[232] and [33]. A series of network analysis functions, implemented in SQL, are
stored within the PostGIS extension. These functions allow for the reading of seman-
tic properties of elements, calculating composite physical parameters of a network,
and performing simple topological routing.
The sections below provide only a bare description of each component, as this the-
sis focuses on binding this models into an UBEM capable library, and not on one
particular model per se. It provides further support to the hypothesis of this the-
sis, that spatial models can provide a solid base for, and act as bridging elements,
in modeling the urban energy chain. The main goal of these models is to assist in
thermal building simulation. As a secondary objective the modeling of urban energy
networks provides a means to the transportation of energy.
In addition to the models presented in this section further development has taken
place. Models were developed for statistical analysis of the building stock and re-
furbishment measures, optimal electric mobility stations placement, PV potential,
city services accessibility and iterative expansion of utility networks using the same
infrastructure and data principles. These models have been used in the production
workflow of the EDF city simulation platform and are not presented in this section.
As these models often required distinct spatial calculations the library was also ex-
tended so as to allow for the calculation of spatial functions. These functions are:
volume calculation, surface classification and CityGML data quality checks. The de-
velopment of these functions was also a time costly endeavour in itself, however,
the work falls outside of the boundaries of the present manuscript and will not be
presented in the section below. See [210], [166], [233], [234], [208], [33], [34], [224]
and [223] for further references with regards to these spatial operations.
Heat losses
Developed in the context of a master thesis [210] under the direct supervision of
this dissertation’s author, the model was partially published in [166]. This model is
designed to calculate heating losses for a building. The depending on the building
geometry and external temperature. The mathematical background is provided by
the Passiv Haus Method developed by the Institute for Passive House, [72], also
available in the form of a software package, [73].
Figure 6.4 depicts a fact sheet of the model with the most important characteristics
and graphic results. On the left hand side of the fact sheet information regarding
the model is depicted. The model’s purpose is defined as estimation of the ther-
mal energy losses in a building. The modeled resolution is a building or a group of
buildings when no distinction could be made with regards to building thermal en-
velope among a group of buildings. Further, the fact sheet specifies that the model
is connected to geometry. This means that certain input variables are calculated by
making use of the 3D spatial model. It also implies that the model uses the unique
identifier system allowing the model to extract information straight from the spatial
database. In the next lines the input, output, relevance and accuracy of the model
is defined. In the lower left hand corner, the spatial scales involved are presented.
The model resolution is presented with red, while yellow depicts the optimal de-
sign scale. When these two coincide, then the colour used is orange. On the right
hand side of the illustration graphical results from the model are depicted. These
6.2. aEneAs - Trojan hero’s odyssey 109
are heating losses estimates for buildings in the study area of Weidenweg, Karlsruhe
and the distributed relative heating demand for five IWU building types. This fact
sheet information pattern is repeated with following models in order to provide a
standardized description of aEneAs’s components.
The spatial input data required by this model are represented by the building model.
All the subsequent work assumes that the model was corrected in previous ETL op-
erations from a topological, semantic and syntax perspective. Spatial calculations
are performed that allow for the building’s inner volume and the further classifica-
tion of external surfaces to be performed (into roof, wall and ground types). Further,
the areas of each type are calculated. Classifying the previously mentioned surfaces
of the building envelope allows for the further energy relevant calculation of gains
and losses. Each of the corresponding building surfaces is allotted a thermal trans-
mittance U value depending on the building age and type. These U-values originate
with the IWU (ger.: Institut Wohnen und Umwelt) institute, [112], which provides a
German national database for building types. The previously calculated parameters
provide input to the calculation of the transmission and ventilation heat loss. Using
additional parameters, namely the temperature zone factor (an indicator that adds
climatic influences to the calculation) and air capacity coefficient the calculation of
the building heat loss (W K−1 ) is performed using the following formula:
Hourly values of temperature data allow for the calculation of a Heating Degree Day
(HDD) value (the difference between the set inside temperature of a building and
the daily average outside temperature). These values are calculated only when the
average day temperature was below 15◦ C for three days in a row (so that the thermal
inertia of the building allowed for the captive heat to be dissipated). Aggregating all
daily values produces the HDD annual value, HDDa (K h):
12
HDDa = ∑(( HD ∗ DM ) ∗ (Ti − T0 ))
1
110 Chapter 6. aEneAs, the scale aware urban energy modelling library
Where:
Finally the energy demand corresponding to the heat loss of the building is calcu-
lated (W h):
EDhl = Phl ∗ HDDa
n
GP = A P ∗ ∑ ∆t ∗ ( BPi + DPi )
i
GS = ∑ ( GP )
GB = ∑( GS )
F IGURE 6.4: Heat loss model fact sheet and associated graphic results, developed after [72], and documented in [166] and [210]
F IGURE 6.5: Solar radiation model fact sheet and associated graphic results, developed after [244], and documented in [234] and [233]
6.2. aEneAs - Trojan hero’s odyssey 113
Where:
The accumulation of the resulting irradiance values for the total time steps produces
the corresponding radiation values per square meter. Radiation values are interpo-
lated linearly for each surface point for each day of the year. Further, the summation
of global radiations value can be performed at the level of month, trimester and
yearly. These summations can be performed at the level of a surface or building
depending on user requirements.
A further application of the solar radiation model is the calculation of photo-voltaic
(PV) potential with different photo voltaic technologies. A model allowing this cal-
culation was proposed for aEneAs, to be built on top of the solar radiation model
presented in paragraph 6.2.2. However, no implementation in this library was made
that past the technical proposal stage. A model developed by colleagues at EIFER,
that follows in the steps of the SolarB and further developed parts of the code, was
published in [160].
Solar gains
Once the amount of solar radiation has been calculated it is then possible to calcu-
late the gains from solar radiation in the buildings heating demand equation. The
summary of this model is depicted in figure 6.6. Most often, the available spatial
data is missing window relevant information, as can be seen with public CityGML
in the federal states of Germany, [81]. It was decided to approach this issue from
a statistical perspective. The actual window surface is estimated using ratios pro-
vided by the IWU institute, in their systematic survey of the German building stock,
[112]. All external building surfaces receive corresponding window ratios according
to the year of construction and building type. In this way, the nominal value of the
window surface is obtained.
The IWU institute also provides a method to calculate window solar gains, [143].
This method is only applied on heating days (as defined with the Heating losses
model, 6.2.2). The value for solar gains is obtained at the relevant resolution desired
and can be determined per surface or per building. The following formula was used:
Qs = ri ∗ gnormal,i ∗ ∑ Gi ∗ A F,i
i
ri = r Frame ∗ rShadow ∗ r Dirt ∗ rnot−normal
114 Chapter 6. aEneAs, the scale aware urban energy modelling library
Where:
PV Paneling
This model is used to provide optimal positioning of the photo-voltaic panels that
are to be installed on the building roof tops. The model was first documented and
published in [208]. The main goal of the implementation is to seek the optimum
spatial position of PV panels on the roof of a building structure. In order to propose
such an installation, the place and orientation of the panel needs to be determined
on the roof of the building. In order to provide this estimate with regards to the
electricity produced by a PV panel, solar irradiation values for a typical year were
produced for the building rooftops. Typical years are meteorological average years
produced with a dedicated software solution, called Meteonorm, that uses stochas-
tic generation, [192]. For the technical details of the standard PV panel dimension,
multiple manufacturer technical specifications were found and averaged to obtain
an estimated surface size of 1.5m2 (1*1.5m).
With regards to the actual positioning and filling of the roof surface two methods
were proposed and tried. In the first method, given limitations to the weight of
the panels or to the cost of the panels are assumed. These are provided by giving
a maximum number of panels that can be installed on a building roof. The solar
irradiation is then calculated for a point grid resolution of 0.5m using the solar ir-
radiation model, 6.2.2. The roof grid points are sorted depending on the recorded
value of total solar irradiation. In the next step, a PV panel is created on the roof
top, in portrait or landscape, around the point with the highest radiation value. The
algorithm then iterates through the following points to determine the optimal posi-
tion for the panel, and if need be restarts the process with a lower ranking point. The
process is repeated with the remaining points so as to include the maximum number
of PV panels.
The second method is designed for use at district, neighborhood, town and city level.
It divides the building’s roof surface into PV panel sized pieces, and then calculates
the amount of solar radiation received by each piece. This implementation adjusts
the grid size definition from the solar irradiation model, 6.2.2 to 1*1.5m, thus remov-
ing the need for further processing time solely for the paneling model. The point
F IGURE 6.6: Solar gains and PV paneling models fact sheet, developed after [112], [143], and documented in [233], [200] and [208]
116 Chapter 6. aEneAs, the scale aware urban energy modelling library
grid dimensions are adjusted for portrait or landscape orientation, in order to en-
sure maximum coverage of the roof surface. For performing city wide PV panel
estimates, this second method was found optimal, even as it is less precise at the
level of a building (lower resolution point grid and less flexible positioning of the
PV panel).
EDheating = EDhl − Qs
Where the two components are described in the Heat losses (6.2.2) and Solar gains
(6.2.2) models.
GHG emission is a prevalent term used to describe greenhouse gases (CO2, CH4,
and N2O). In this model’s case, the results provide an estimate with regards to the
amount of GHG emitted with the generation of thermal energy for building heating.
The method used is presented in [200], and estimates the total amount of CO2 equiv-
alent gasses. It converts other greenhouse gases into CO2 by using the conversion
ratios of the GWPV (Global Warming Potential Values) presented in [161].
The method was applied in Karlsruhe, Germany where information on the heat-
ing sources was made available at the level of a town (Stadtteil) with the help of
a commercial database solution called INFAS, [108]. This solution has been used
previously with other scientific publications, see [107]. Statistical information with
regards to the average GHG emissions (CO2 equivalent was obtained from the IWU
institute, [88]. The formula used in the calculation, presented in [200], is as follows:
Where:
As per [30] and [32] the formulas below were used for the calculation. The formula
for the normalization of building refurbishment costs with help of the indices for
different states (example provided for BW, Baden-Würtemberg, where the city of
Karlsruhe, used in this application is located):
Land Kold
Knew = ∗ k m2 LandBW
k m2 Landi
The following formula was used for the normalization of building refurbishment
costs with help of the historical indices:
Year Kold
Knew = ∗ indexCurrentYear
index Re f urbYear
Finally, the formula for the cost estimation for new buildings provides the estimated
refurbishment costs in € at the level of a building:
Where:
Different refurbishment strategies can be tested for medium and deep refurbish-
ments, and the application can be reduced in implementation for historical build-
ings, whose street facing façades have restrictions with regards to the type of refur-
bishment that can be performed.
fact that for the green façade no soil layer is included in the simulation. The primary
thermal exchange processes of the green façade model is presented in figure A.3.
The following formula represents the heat balance equation and, most importantly,
the heat stored in the wall behind the vegetated façade (Svw ) used in the model:
Where:
The green roof model includes additionally the behaviour of the substrate layer (the
soil layer of the plants), which can be observed in the previous formula. Svw is
computed for both implementations using the temperature behind the vegetation
and the substrate layers using a numerical bisection method, also presented in [214].
The model was validated for using monitoring data from real-time experiments dur-
ing summer measurements at three locations in Germany (see 5.2 and [223] for more
details with regards to the monitoring installations and validation of results).
The model was developed within the framework of a master thesis, [34], and partly
documented (with regards to modeling of the water network) in [33]. This dual net-
work behavioural model simulated 24 hours of network operation at a time step of
30 minutes. The networks each have associated usage profiles sets with higher mea-
surement rates, however, for the model to run smoothly and show the downward
effect of a network event this time step was considered sufficient. It also included
cascading effects, which can be triggered by modifying the set of input parameters.
The integrated dual network acts as a demonstrator for the capabilities of modeling
multi-utility networks using the UtilityNetwork ADE and semantical city models
stored with CityGML.
In order to be able to perform such complex spatial operations a suitable data set
was found in the city of Nanaimo, Canada, where a coupled electric/water network
exists. There, an energy reclamation facility converts the incoming flow of fresh-
water and generates electricity. Thus, the two networks and their data models are
linked at the water turbines. This is presented conceptually in figure A.4. The object
class used to describe the turbine is the CityFurniture from CityGML. This solution
can be replaced by making use of the GenericCityObject class.
The city of Nanaimo with their local Public Works company (and very benevolent
staff) provided open data for their water and electrical networks and diagrams of
the mixed plant, freshwater delivery/power generation facility. The water network
pipes and appurtenances could be produced from the data provided by the city,
while the electrical network had to be fabricated by using the road network (which
was available as open data). The original water network data had to undergo a
lengthy ETL process in order to produce a UtilityNetwork ADE data set. This work
is documented in [33].
The behavioural model represents a typical day of operation on the water and elec-
trical network sample. It includes both water demand and electrical production from
the energy reclamation facility. Water demand is associated with the buildings in the
nearby suburb from the water reservoir. As water from the water reservoir is being
depleted by (building) usage it is replenished. The usage follows a consumption
pattern provided by the City of Nanaimo Public Works staff, [34]. As the re-filling of
reservoir is activated, the water pushing through the turbines produces electricity.
Additional electrical production in the network stems from PV panels installed on
the roofs of some of the buildings. This occurs only when the production exceeds the
electrical demand in the respective building. The electrical output from PV produc-
tion is replicated from that measured in a real local installation that provides open
source data, the Nanaimo Food Share PV power production.
The behaviour of the water and electrical networks is mimicked using a linear time
series model that is stored in a Python script. Values for properties of network ele-
ments are calculated at given locations (usually at appurtenances, bifurcations and
buildings). These values impact the state of other elements in the network, and
are able, for example, to open the valve for re-filling of the water reservoir at the
Nanaimo water and energy reclamation facility.
In order to simulate cascading effects, the topology of the model is altered. This
is depicted in figure A.5. If a water pipe belonging to the network is undergoing
maintenance or an electrical cable fails, the topological link in between the elements
is set to "broken". This change will impact further (connected) nodes, in what would
be a typical cascade effect. The behavioural model that models this behaviour does
6.2. aEneAs - Trojan hero’s odyssey 123
not interfere with the data model, stored in the DB and keeps a separate log to the
changes that occur to elements and their properties, thus maintaining a separation
between the logic and data tiers.
F IGURE 6.9: Fact sheet for integrated water/electrical networks management, documented in [34] and [33]
125
Chapter 7
In modeling the urban energy chain, there is no variable as potent as the fabric of
space and time. It brings the multiple components of the energy chain together and
allows for spatial and temporal dependencies to be correctly modeled collectively.
This embraces Tobler’s first law of geography, [217]: "[E]verything is related to ev-
erything else, but near things are more related than distant things."
Geographers are bound to observe the world while engineers help build it. To em-
brace both sciences, a facilitator is found in modeling and data integration, sup-
ported by spatial data in the form of 3D semantic city models that act as data hubs.
These are stored in spatial databases and use web services that ferry data to and
from one component to another without time-costly ETL processes.
7.1 Conclusion
This thesis’ results endorse the idea that city-wide energy modeling is most efficient
when performed at a neighborhood scale while using buildings as data units for
consumers and production entities. Utility networks require modeling under new
paradigms where both energy and information flows are bi-directional. End con-
sumers are no longer just that; they are transformed into prosumers. This is espe-
cially true for the Ger.: Energiewende concept where precise estimates of local energy
demand and production are required, and local energy production is preferred.
Modeling urban energy problems facilitates solving complex questions and equa-
tions as the spatial models become central linking elements of the different types of
energy-related data. Testing includes urban simulation, executed to provide quan-
titative and qualitative energy site management assessments. Using scenarios, both
green and brownfield urban development can be compared, and the influence of fu-
ture contextual parameters, such as energy prices, warmer climates, and upgraded
technologies, on new designs can be explored. This is also the primary use case
of the EDF city simulation platform project, which has partially financed this the-
sis, for which the author is very grateful. These scenarios and their corresponding
impact/results are then provided via visually intelligible/comprehensible predic-
tions to multiple stakeholders. The proposed open style architecture enables joint
decision making, a participatory approach, in which all actors can work together to
understand the impact of proposed actions.
All buildings and their associated energy systems are unique. Therefore, their digital models
must reflect this reality. Since the 1950s, there has existed a trend in creating standard-
ized building components, for example, the use of identical utility network cabi-
nets or shower/bath modules in sky-scrapers. These come pre-fabricated and are
126 Chapter 7. Outcome - Conclusion and Outlook
only installed on-site, as in large hotels. This is very similar to the proposed digi-
tal standardization approaches taken for modeling and further supports the use of
standardized buildings data models.
in the case of UBEM models used for thermal energy demand provided further in-
sight into the appropriate scale in terms of results’ significance. Calculations were
made using building units that are then aggregated at a neighborhood or a district
level. This method facilitated the provision of accurate data at their respective spa-
tial scales. With the current availability of monitoring data and its use in model
validation and revision, there is a noticeable trend towards achieving high accuracy
at lower scales, going as far as the building unit and the residential unit.
Two significant issues with urban energy modeling are data consistency concerning
spatial scales and urban fabric contiguity. Spatial scale mismatches require exten-
sive ETL operations. An exhaustive list of scale-related ETL operations was pre-
sented and their impact and the potential use of the produced data. These oper-
ations should not be underestimated, as they can occupy a significant amount of
time in an urban energy model project’s life span. Regarding contiguity, the non-
interrupted urban fabric of building envelopes and neighborhoods helps to ensure
energy savings and facilitates the creation of planning units. This issue plays a vital
role in volume calculations, a major regressor in thermal energy calculations. Build-
ings, per se, constitute contiguous entities, and their use as spatial units allows for
continuous urban pattern analysis.
A taxonomy of spatial standards and scales is presented in chapter 4. Formats can
be classified by using four attributes: format/content related (standard type, data
model availability, data content), design scale, energy thematic (energy chain posi-
tion, energy commodity type), and entity type (public/private/citizen). These mul-
tiple characteristics illustrate and provide depth and dimensions to the complexity
of the task at hand, modeling and monitoring the entities, their attributes, and in-
teractions at play in the urban energy chain. Again, the scale appears as a major
separator, as the lack of standardized spatial units in the design scale hinders further
efforts of data integration and adoption of existing standards.
Standardized spatial scales and persistent identifiers
Holistic, integrated energy systems work on different spatial levels, from residential
and tertiary use buildings, which are now active players in energy production, up to
district, neighborhoods, city, regional and national scale. Modeling and simulation
are the sole means of exploring the performance of new designs and concepts in this
rapidly changing urban context. In section 3.3 a proposal of standardized spatial
scales for use in UBEM is made. These unit scales are energy meter, building, district,
neighborhood, town, and city. The scales were tested within different urban energy
chains regarding economic, cultural, and geographical characteristics and are shown
to be adaptable to new locations.
A scale-dependent urban data ontology with persistent identifiers was proposed to
solve the identified problems. This open ontology avoids issues related to propri-
etary ontologies and cascading effects from their use and is built on top of the pro-
posed standardized spatial sales. This human-readable code allows for differently
sized cities to be modeled and mitigates associated issues, such as data heterogene-
ity and inconsistency in urban modeling domains. In this manner, all urban energy
chain stakeholders are bound, within a non-biased perspective, by persistent means
of identification. For a successful implementation of this proposal, it is not necessary
that all energy chain actors use this system, but rather that the data is linked, in a
central repository or a virtual hub, see 5.5. This repository needs to contain a map-
ping system that allows for objects to be identified depending on their ownership
128 Chapter 7. Outcome - Conclusion and Outlook
status (What is the entity in the chain that they belong to?). This will fix the miss-
ing spatial data issue and help increase the quality of recommendations that can be
made on top of the modeling results.
further issues in horizontal and vertical interoperability. The fact that CityGML in-
cludes graphical, semantic, and thematic properties, taxonomies, and aggregations
supports this manuscript’s claim to the suitability of CityGML for modeling the en-
tirety of the city-wide energy chain. These design choices allow for both UBEM input
data and the urban interactions, be they spatial or energy thematic, to be modeled
and simulated.
in the framework of this thesis starts by first quantifying primary thermal energy
demand and then the impact of refurbishment measures. Lastly, it estimates the po-
tential of renewable energy production. aEneAs also includes components that con-
sider energy distribution in the given context, showing a path toward data modeling
and simulation required for distributed energy production at the neighborhood and
district level.
Holistic, integrated energy systems work on different spatial levels, from buildings,
which are now active players in energy production, to district and city-level designs
and national-scale infrastructure. The technologies identified in chapter 2 with re-
gards to the urban energy chain meet at the level of a building. This scale is also
where they will have maximum impact, as this is the scale at which people live and
work and around which they usually plan their life. This is precisely why aEneAs
was developed to work using building units. It performs complex energy modeling
tasks using semantic city models stored in CityGML inside a spatial database.
Context is king
The technical solutions and implementations in the present thesis are designed to
work in an urban setting and allow for the harnessing of the power of the spatial and
temporal character of data in energy modeling. Standardizing the input of models
and binding it to open standards allows for the re-utilization of advanced tools. It
shortens the time to move one solution from one urban center to another.
aEneAs performs urban wide modeling and presents results at a single building unit
level. The model results offer varying accuracy at the standardized spatial scales
used in the thesis. Overall, good results are achieved at the district’s scale and in
some models at the level of a single building. The surrounding environment and its
influence, such as shading and collected solar gains, are also included. This informa-
tion can be used for integrated energy system planning and precise local renewable
energy production estimates.
Good practices
The experience gathered from this thesis’ experiments is provided as a synopsis in
the form of good practices that can be recommended to cities that want to evolve to-
wards next-generation cities. These are: the application of a common, adaptable, and
reusable SDI, the development of reusable tools, the creation of open data and use
of open source code, and the use of a digital by default policy. Once applied, these
work principles should improve future work on the energy chain, both in terms of
modeling and monitoring. They will also help facilitate intelligent cities initiatives,
enhance public trust and promote value creation from public data.
In the thesis’ experiments, a sturdy case is made for monitoring and modeling the
urban energy chain using exclusively open-source software solutions and open stan-
dards for files and web services. A good example lies in the implemented data stor-
age solutions: PostgreSQL and MongoDB. MongoDB is fast becoming an industry-
standard while PostgreSQL is already one, being the top choice for spatial reposito-
ries. Furthermore, PostgreSQL is a 30-year running open-source effort that shows
that open source is not just a trend but a phenomenon that is here to stay. It has
jumped from one generation to the next and supports cities and modelers alike in
many use cases. In addition, public data is a public good, and its availability, even
132 Chapter 7. Outcome - Conclusion and Outlook
limited, creates trust through open standards. This is a fundamental requirement for
a well-governed city.
What could be concluded from this work is that there is no perfect one size fits all
tool. Instead, the proposed good practices should be followed and implemented
as much as possible to allow the entire ecosystem to be made interoperable and
expanded in a sustainable manner.
Limitations
The findings in this thesis should lead to enhanced data handling and data quality in
the urban energy chain. However, this is very much dependent on local know-how
and the willingness to invest in such adaptation. The author of this thesis recognizes
that, for the moment, such complex modeling and monitoring, even though central
to becoming next-generation cities, is an expensive feature. This is currently only
available to first-tier cities with enough capital for dedicated spatial departments
and skilled personnel to deliver such integrated data sets. On the other hand, the
willingness and political tenability to redirect substantial funds is also linked to po-
litical, cultural, and public recognition of environmental phenomena. Modeling of
urban heat island impact on cooling demand, green facades, and green surfaces can
easily find financial support in cities already impacted by climate change where the
general public and state officials follow scientifically supported evidence.
Another significant limitation of the proposed solution is that entities are often mov-
ing in the ever-changing city. The presented modeling focused on buildings alone.
However, a substantial part of the energy chain is not stationary. The modeling of
spatially moving entities, such as buses, trams, and metros, is also possible within
this methodology. This thesis has received basic testing in the AERO-TRAM use
case where moving trams were visualized, and the moving measurement flow per-
formed well while using the SensorThings API. Additionally, maintaining the city
model and its versioning are not easy tasks in the current versions of CityGML,
though its newest version promises to improve this capability.
7.2 Outlook - the future of semantic city models and the ur-
ban energy chain
Digital transformation
Cities, in general, are going through a significant transformation. This shift is part
of a society-wide change and is generally referred to as the digital transformation.
As shown in chapter 5, certain cities and countries are embracing it and will emerge
fundamentally changed out of it. The world pandemic context has accelerated this
change as it forced people and enterprises alike to use more digital services and
enhanced the focus on interoperability.
This shift means that standardization organizations are asked to provide standards
that work in an interconnected world of big data. At the same time, these standards
need to embrace legacy software and data structures that were never designed for
interoperability. This contrast requires a show of patience for the developing actors
and support for the overachievers - the ones who set new paths for others to follow.
7.2. Outlook - the future of semantic city models and the urban energy chain 133
This author expects live or near-real updates of 3D semantical city models in first-
tier cities in the next decade. This will be possible with the advent of very potent,
universally available mobile sensing platforms.
This ever more extensive interoperable infrastructure will require data storing capa-
bilities in organized and non-organized data structures. This is a strong incentive for
the further development of the existing standards, particularly the UtilityNetwork
ADE, which at the time that this manuscript was written is still primarily a work in
progress. Furthermore, links to non-organized data repositories need to be created
and maintained, such as FIWARE already foresees to MongoDB. High data volume
will help improve modeling results and, at times, reduce the need for modeling as
monitoring will help directly illustrate urban energy phenomena.
Initial objections towards XML as a solution in the IT domain regarding its size and
complexity have lost steam as computers kept acquiring more computational power
and internet bandwidth increased. At the same time, wide adoption of the CityGML
standard remains an elusive goal, with such data sets only being part of the spatial
repositories of first-tier cities, or better said, cities with available capital. At times,
as in the case of the EU, national or regional spatial data agencies take it upon them-
selves to provide such data sets. This is a step forward, however, as shown in [81],
data quality is not a given, and even though the format is the same, the content can
differ.
In parallel to this work, the CityGML 3.0 has been developed, [124]. This thesis fo-
cused solely on the second version of the standard for which large open data sets
exist. The new version proposes solutions to some of the points raised in this thesis.
It remains to be seen how successful it will become, an indicator easily measured
with the adoption ratio. However, the discussion surrounding CityGML v3.0 raised
at the standard adoption by the OGC shows that the spatial community at large is
very much interested in the further use and improvement of this standard, regard-
less if an improvement of 2.0 or adoption of 3.0 is the correct way to push ahead.
The circumstances in which CityGML 3.0 came to be are similar, in a way, to the
evolution of the Python programming language to version 3.0. Both Python 2 and
Python 3 continue to exist. Still, slowly the community seems to be inching towards
the later version, which presents a likely pattern for what will eventually happen
with CityGML semantic city models.
Sharing is caring
Specifically, in the energy chain, the residential, tertiary, and industrial urban sites
will be able to communicate and provide energy as a service to one another in the
same way that the car-sharing concept is operating today. Cities and local decision-
makers need to aim for energy sharing concepts at the neighborhoods and district
level first, for a plethora of reasons, some presented in this very thesis. Standard-
ization and data sharing among spatial actors will help locally produced energy be
consumed locally, significantly reducing the need for large distribution networks
and their associated losses, which follow the so-called distance decay law - related
to Tobler’s first law of geography. In a second stage, this can be shifted towards
the following spatial scales of the urban divide so that the urban fabric contiguity
discussed in chapter 3 is respected, and artificial boundaries are not an impediment.
To achieve this, communication standards are required and improved and widely
available utility networks.
134 Chapter 7. Outcome - Conclusion and Outlook
While electrical networks are a given in urban settings, a far-reaching effort to de-
velop heating and cooling networks is still required, with limited, yet significant, ex-
ceptions in mainly major planning-based economies. Existing infrastructure spaces,
such as gas networks, need to be used to provide existing installations with energy
and, potentially, make place for new utility infrastructure as space is a scarce and
precious resource in the urban domain. Development of the various energy net-
works of the urban energy chain presented in chapter 2 can be an expensive long-
term burden. This is why they should be included in comprehensive urban planning
and management, where the broad horizontal and vertical contributions to the ur-
ban energy system can be used to their full potential.
This means that the planning and management of the urban energy chain need ad-
justment, a balancing act of sorts between multi-faceted perspectives. Semantic city
models and web services are there to bolster this balancing act by enforcing inter-
operability, providing links between different actors and their data, and helping to
model and monitor the many types of spatial interactions that cramped urban spaces
see plenty of.
FIN
135
Appendix A
The figures included in this appendix are referenced to in different sections of the
thesis and belong to other publications.
136 Appendix A. Figures from other sources
F IGURE A.2: Technology readiness levels, marked with color are the
levels at which the models produced in this thesis have been imple-
mented, after [69]
Appendix A. Figures from other sources 137
F IGURE A.4: Solution for coupling the water and electrical networks
at a turbine, as presented in section 6.2.2, from [34]. On the left a con-
ceptual representation of the means of coupling between the electric
and water networks via the connectedCityObject attribute and a City-
Furniture object. On the right a graphical representation of the data
models.
138 Appendix A. Figures from other sources
Bibliography
Barry McGaw. Third Edition. Oxford: Elsevier, 2010, pp. 374–377. ISBN: 978-
0-08-044894-7. DOI: 10.1016/B978-0-08-044894-7.01358-0.
[15] Dominik Bartoszewski, Adam Piorkowski, and Michal Lupa. “The Compari-
son of Processing Efficiency of Spatial Data for PostGIS and MongoDB Data-
bases”. In: Beyond Databases, Architectures and Structures. Paving the Road to
Smart Data Processing and Analysis. Ed. by Stanis\law Kozielski, Dariusz Mrozek,
Pawe\l Kasprowski, Bo\.zena Ma\lysiak-Mrozek, and Daniel Kostrzewa.
Cham: Springer International Publishing, 2019, pp. 291–302. ISBN: 978-3-030-
19093-4.
[16] Michael Batty. The new science of cities. English. 2013th ed. Massachusetts In-
stitute of Technology, 2013. ISBN: 978-0-262-01952-1.
[17] BBC. Finland makes broadband a ’legal right’ - BBC News. news. 07/2010. URL:
https://www.bbc.com/news/10461048 (Accessed on: 11/08/2019).
[18] BDEW, VKU, and GEODE. Abwicklung von Standardlastprofilen Gas. Guide.
2016. URL: https://www.bdew.de/media/documents/Leitfaden_20160630_
Abwicklung-Standardlastprofile-Gas.pdf (Accessed on: 01/24/2021).
[19] Jeff de la Beaujardiere. OpenGIS Web Map Server Implementation Specification.
English. 2006. URL: https://portal.ogc.org/files/?artifact_id=14416
(Accessed on: 06/01/2021).
[20] T. Becker, C. Nagel, and T. H. Kolbe. “Integrated 3D Modeling of Multi-utility
Networks and Their Interdependencies for Critical Infrastructure Analysis”.
English. In: Advances in 3D Geo-Information Sciences, Lecture Notes in Geoinfor-
mation and Cartography (2011). Editors: Kolbe, T. H., König, G. & Nagel, C.
Publisher: Springer, Berlin Heidelberg, pp. 1–20. URL: http://mediatum.ub.
tum.de/doc/1145740/358854.pdf.
[21] T. Becker, C. Nagel, and T. H. Kolbe. “Semantic 3D Modeling of Multi-Utility
Networks in Cities for Analysis and 3D Visualization”. English. In: Progress
and New Trends in 3D Geoinformation Sciences, Lecture Notes in Geoinforma-
tion and Cartography. Editors: Pouliot, J., Daniel, S., Hubert, F. & Zamyadi,
A. Quebec City: Springer, Berlin Heidelberg, 2012, pp. 41–62. URL: http://
mediatum.ub.tum.de/doc/1145740/358854.pdf (Accessed on: 05/06/2021).
[22] Joachim Benner. CityGML Energy ADE - CityGML Wiki. 2020. URL: http://
www . citygmlwiki . org / index . php / CityGML _ Energy _ ADE (Accessed on:
09/19/2019).
[23] Joachim Benner. LandXML – Review with respect to the modelling of Utility Net-
works. en. Standardization comitee workshop. Munich, Germany, 2016. URL:
https://en.wiki.utilitynetworks.sig3d.org/images/upload/20160913_-
_Benner_-_LandXML.pdf (Accessed on: 01/23/2022).
[24] Paul Bertheau, Catherina Cader, and Philipp Blechinger. “Electrification Mod-
elling for Nigeria”. en. In: Energy Procedia 93 (08/2016), pp. 108–112. ISSN:
18766102. DOI: 10.1016/j.egypro.2016.07.157. (Accessed on: 10/10/2019).
[25] Filip Biljecki, Kavisha Kumar, and Claus Nagel. “CityGML Application Do-
main Extension (ADE): overview of developments”. In: Open Geospatial Data,
Software and Standards 3.1 (08/2018), p. 13. ISSN: 2363-7501. DOI: 10 . 1186 /
s40965-018-0055-6.
[26] Filip Biljecki, Jantien Stoter, Hugo Ledoux, Sisi Zlatanova, and Arzu Çöltekin.
“Applications of 3D City Models: State of the Art Review”. en. In: ISPRS In-
ternational Journal of Geo-Information 4.4 (12/2015), pp. 2842–2889. ISSN: 2220-
9964. DOI: 10.3390/ijgi4042842. (Accessed on: 09/10/2019).
[27] Stefan Bischof, Christoph Martin, Axel Polleres, and Patrik Schneider. “Col-
lecting, Integrating, Enriching and Republishing Open City Data as Linked
Bibliography 143
Data”. In: The Semantic Web - ISWC 2015. Ed. by Marcelo Arenas, Oscar Cor-
cho, Elena Simperl, Markus Strohmaier, Mathieu d’Aquin, Kavitha Srinivas,
Paul Groth, Michel Dumontier, Jeff Heflin, Krishnaprasad Thirunarayan, and
Steffen Staab. Cham: Springer International Publishing, 2015, pp. 57–75. ISBN:
978-3-319-25010-6.
[28] BKI. BKI Objektdaten Energieeffizientes Bauen E2. Verlagsgesellschaft Rudolf
Müller, 2002. ISBN: 978-3-481-01885-6. URL: https : / / de . book - info . com /
isbn/3-481-01885-1.htm (Accessed on: 03/17/2022).
[29] BKI. BKI Objektdaten Energieeffizientes Bauen E4. Baukosteninformationszen-
trum Deutscher Architektenkammern (BKI), 2011. ISBN: ISBN 978-3-941679-
21-4. URL: https : / / bki . de / objektdaten - energieeffizientes - bauen /
bki - objektdaten - energieeffizientes - bauen - e4 . html (Accessed on:
03/17/2022).
[30] BKI. BKI Objektdaten Energieeffizientes Bauen E5. Baukosteninformationszen-
trum Deutscher Architektenkammern (BKI), 2014. ISBN: 978-3-941679-56-6.
URL : https : / / bki . de / objektdaten - energieeffizientes - bauen / bki -
objektdaten-energieeffizientes-bauen-e5.html (Accessed on: 03/17/2022).
[31] BKI. Energiekonzept: Planung nach DIN 4108 / DIN 18599 | BKI Objektdaten En-
ergieeffizientes Bauen E3. Baukosteninformationszentrum Deutscher Architek-
tenkammern (BKI), 2009. URL: https://tinyurl.com/ze6xvvnb (Accessed
on: 03/17/2022).
[32] BMUV. Bericht der Baukosten-senkungskomission. German. 2015. URL: https://
www.die-wohnraumoffensive.de/fileadmin/user_upload/pdf/buendnis-
bezahlbares - wohnen - baukostensenkungskommission . pdf (Accessed on:
03/12/2022).
[33] I. Boates, G. Agugiaro, and A. Nichersu. “NETWORK MODELLING AND
SEMANTIC 3D CITY MODELS: TESTING THE MATURITY OF THE UTIL-
ITY NETWORK ADE FOR CITYGML WITH A WATER NETWORK TEST
CASE”. en. In: ISPRS Annals of Photogrammetry, Remote Sensing and Spatial In-
formation Sciences IV-4 (09/2018), pp. 13–20. ISSN: 2194-9050. DOI: 10.5194/
isprs-annals-IV-4-13-2018. (Accessed on: 10/12/2018).
[34] Isaac Boates. “Demonstrating Utility Network Interdependency Modelling
using the UtilityNetwork Application Domain Extension for CityGML”. En-
glish. Supervisor: Alexandru Nichersu and Giorgio Agugiaro. MA thesis.
Karlsruhe: Hochschule Karlsruhe, 2018.
[35] Andrea Bonaccorsi and Cristina Rossi. “Why Open Source software can suc-
ceed”. In: Research Policy 32.7 (2003), pp. 1243–1258. ISSN: 0048-7333. DOI:
https://doi.org/10.1016/S0048-7333(03)00051-9.
[36] Mike Botts, George Percivall, Carl Reed, and John Davidson. “OGC® sen-
sor web enablement: Overview and high level architecture”. In: International
conference on GeoSensor Networks. Springer, 2006, pp. 175–190. DOI: https :
//doi.org/10.1007/978-3-540-79996-2_10.
[37] Mike Botts and Alexandre Robin. OpenGIS® Sensor Model Language (SensorML)
Implementation Specification. Version 1.0.0. en. Accepted: 2019-10-29. 2007. DOI:
10.25607/OBP-646. (Accessed on: 11/11/2021).
[38] Mike Botts, Alexandre Robin, and Eric Hirschorn. OGC SensorML: Model and
XML Encoding Standard. English. Publisher: Open Geospatial Consortium.
08/2020. URL: http://docs.ogc.org/is/12- 000r2/12- 000r2.html (Ac-
cessed on: 11/28/2021).
144 Bibliography
[39] Roger Brackin and Pedro Gonçalves. OGC OWS Context Conceptual Model.
English. 2014. URL: https://portal.ogc.org/files/?artifact_id=55182
(Accessed on: 06/01/2021).
[40] M. Breunig, A. Borrmann, E. Rank, S. Hinz, T. Kolbe, M. Schilcher, R.-P. Mun-
dani, J. R. Jubierre, M. Flurl, A. Thomsen, A. Donaubauer, Y. Ji, S. Urban, S.
Laun, S. Vilgertshofer, B. Willenborg, M. Menninghaus, H. Steuer, S. Wursthorn,
J. Leitloff, M. Al-Doori, and N. Mazroobsemnani. “COLLABORATIVE MULTI-
SCALE 3D CITY AND INFRASTRUCTURE MODELING AND SIMULA-
TION”. In: ISPRS - International Archives of the Photogrammetry, Remote Sensing
and Spatial Information Sciences XLII-4/W4 (2017), pp. 341–352. DOI: 10.5194/
isprs-archives-XLII-4-W4-341-2017.
[41] Encyclopædia Britannica. energy | Definition, Types, & Examples. en. 2019. URL:
https://www.britannica.com/science/energy (Accessed on: 10/06/2019).
[42] S. Cajot, M. Peter, J.-M. Bahu, F. Guignet, A. Koch, and F. Maréchal. “Obsta-
cles in energy planning at the urban scale”. In: Sustainable Cities and Society
30 (2017), pp. 223 –236. ISSN: 2210-6707. DOI: https://doi.org/10.1016/j.
scs.2017.02.003.
[43] C. F. Calvillo, A. Sánchez-Miralles, and J. Villar. “Energy management and
planning in smart cities”. In: Renewable and Sustainable Energy Reviews 55 (2016),
pp. 273–287. ISSN: 1364-0321. DOI: 10.1016/j.rser.2015.10.133. (Accessed
on: 09/13/2019).
[44] M Casini. “Green Technology for Smart Cities”. en. In: IOP Conference Series:
Earth and Environmental Science 83 (08/2017), p. 012014. ISSN: 1755-1307, 1755-
1315. DOI: 10.1088/1755-1315/83/1/012014. (Accessed on: 11/02/2019).
[45] Juan Antonio Ruiz Ceniceros, José Alfonso Aguilar-Calderón, Roberto Es-
pinosa, and Carolina Tripp-Barba. “The external and data loose coupling for
the integration of software units: a systematic mapping study”. en. In: PeerJ
Computer Science 7 (12/2021). Publisher: PeerJ Inc., e796. ISSN: 2376-5992. DOI:
10.7717/peerj-cs.796. (Accessed on: 01/22/2022).
[46] Carlos Cerezo, Julia Sokol, Saud AlKhaled, Christoph Reinhart, Adil Al-Mumin,
and Ali Hajiah. “Comparison of four building archetype characterization meth-
ods in urban building energy modeling (UBEM): A residential case study in
Kuwait City”. In: Energy and Buildings 154 (2017), pp. 321 –334. ISSN: 0378-
7788. DOI: https://doi.org/10.1016/j.enbuild.2017.08.029.
[47] Tania Cerquitelli, Evelina Di Corso, Stefano Proto, Paolo Bethaz, Daniele Maz-
zarelli, Alfonso Capozzoli, Elena Baralis, Marco Mellia, Silvia Casagrande,
and Martina Tamburini. “A Data-Driven Energy Platform: From Energy Per-
formance Certificates to Human-Readable Knowledge through Dynamic High-
Resolution Geospatial Maps”. en. In: Electronics 9.12 (12/2020), p. 2132. ISSN:
2079-9292. DOI: 10.3390/electronics9122132. (Accessed on: 02/07/2021).
[48] Debasish Chakraborty. “Web Based GIS Application using Open Source Soft-
ware for Sharing Geospatial Data (IJARSG, pp 1124-1228)”. In: international
journal of advanced remote sensing and gis 4 (09/2015), pp. 1224–1228. DOI: 10.
23953/cloud.ijarsg.109.
[49] Kam Wing Chan. “Misconceptions and Complexities in the Study of China’s
Cities: Definitions, Statistics, and Implications”. English. In: Journal of Eco-
nomic Literature (2007). URL: https://web.archive.org/web/20130115173048/
http://courses.washington.edu/chinageo/ChanCityDefinitionsEGE2007.
pdf# (Accessed on: 01/24/2021).
[50] Kanishk Chaturvedi and Thomas H. Kolbe. “InterSensor Service: Establish-
ing Interoperability over Heterogeneous Sensor Observations and Platforms
Bibliography 145
for Smart Cities”. In: 2018 IEEE International Smart Cities Conference (ISC2).
09/2018, pp. 1–8. DOI: 10.1109/ISC2.2018.8656984.
[51] Yixing Chen, Tianzhen Hong, Xuan Luo, and Barry Hooper. “Development
of city buildings dataset for urban building energy modeling”. In: Energy and
Buildings 183 (2019), pp. 252–265. ISSN: 0378-7788. DOI: https://doi.org/
10.1016/j.enbuild.2018.11.008.
[52] Yao-Hsin Chiang, Chih-Yuan Huang, and Masaaki Fuse. “The Integration of
OGC SensorThings API and OGC CityGML via Semantic Web Technology”.
In: Web and Wireless Geographical Information Systems. Ed. by Sergio Di Mar-
tino, Zhixiang Fang, and Ki-Joune Li. Cham: Springer International Publish-
ing, 2020, pp. 55–67. ISBN: 978-3-030-60952-8.
[53] Ping-Tsai Chung and Sarah H. Chung. “On data integration and data mining
for developing business intelligence”. In: 2013 IEEE Long Island Systems, Ap-
plications and Technology Conference (LISAT). 05/2013, pp. 1–6. DOI: 10.1109/
LISAT.2013.6578235.
[54] Paul Clarke, Rory V. O’Connor, and Murat Yilmaz. “In Search of the Origins
and Enduring Impact of Agile Software Development”. In: Proceedings of the
2018 International Conference on Software and System Process. ICSSP ’18. event-
place: Gothenburg, Sweden. New York, NY, USA: Association for Comput-
ing Machinery, 2018, pp. 142–146. ISBN: 978-1-4503-6459-1. DOI: 10 . 1145 /
3202710.3203162.
[55] Roberta Cocco. Smart City? Milan is leading the game: here are its 4 secrets. en.
04/2019. URL: https://www.morningfuture.com/en/article/2019/04/17/
smart-city-milano-roberta-cocco/596/ (Accessed on: 11/03/2019).
[56] Javier Conde, Andres Munoz-Arcentales, Alvaro Alonso, Sonsoles Lopez-
Pernas, and Joaquin Salvachua. “Modeling Digital Twin Data and Architec-
ture: A Building Guide with FIWARE as Enabling Technology”. In: IEEE In-
ternet Computing (2021), pp. 1–1. DOI: 10.1109/MIC.2021.3056923.
[57] World Wide Web Consortium. W3C XML Schema Definition Language (XSD)
1.1 Part 1: Structures. English. 2012. URL: https://www.w3.org/TR/xmlschema11-
1/.
[58] Simon J D Cox. ISO 19156:2011 - Geographic information – Observations and
measurements. English. 01/2011. DOI: 10.13140/2.1.1142.3042.
[59] Andrew Crapo, Katrina Griffith, Ankesh Khandelwal, John Lizzi, Abha Moitra,
and Wang Xiaofeng. “Overcoming Challenges Using the CIM as a Semantic
Model for Energy Applications”. In: Proc. Grid-Interop Forum 2010. 2010. URL:
www.pointview.com/data/files/3/2433/1730.pdf.
[60] Drury B. Crawley, Linda K. Lawrie, Frederick C. Winkelmann, W.F. Buhl,
Y.Joe Huang, Curtis O. Pedersen, Richard K. Strand, Richard J. Liesen, Daniel
E. Fisher, Michael J. Witte, and Jason Glazer. “EnergyPlus: creating a new-
generation building energy simulation program”. en. In: Energy and Buildings
33.4 (04/2001), pp. 319–331. ISSN: 03787788. DOI: 10.1016/S0378-7788(00)
00114-6. (Accessed on: 02/25/2020).
[61] Felix Creutzig, Giovanni Baiocchi, Robert Bierkandt, Peter-Paul Pichler, and
Karen C. Seto. “Global typology of urban energy use and potentials for an
urbanization mitigation wedge”. en. In: Proceedings of the National Academy
of Sciences 112.20 (05/2015), pp. 6283–6288. ISSN: 0027-8424, 1091-6490. DOI:
10.1073/pnas.1315545112. URL: http://www.pnas.org/lookup/doi/10.
1073/pnas.1315545112 (Accessed on: 11/03/2019).
[62] S. Dalla Costa, E. Roccatello, and M. Rumor. “A CITYGML 3D GEODATABASE
FOR BUILDINGS’ ENERGY EFFICIENCY”. en. In: ISPRS - International Archives
146 Bibliography
International Conference and Exhibition. Vol. 16. Rome: IBPSA, 2019, pp. 3452–
3459. DOI: 10.26868/25222708.2019.210632.
[77] Martina Ferrando, Francesco Causone, Tianzhen Hong, and Yixing Chen.
“Urban building energy modeling (UBEM) tools: A state-of-the-art review
of bottom-up physics-based approaches”. In: Sustainable Cities and Society 62
(2020), p. 102408. ISSN: 2210-6707. DOI: https://doi.org/10.1016/j.scs.
2020.102408.
[78] Giuseppe Fusco and Lerina Aversano. “An approach for semantic integration
of heterogeneous data sources”. In: PeerJ Computer Science 6 (03/2020), e254.
ISSN : 2376-5992. DOI : 10.7717/peerj-cs.254.
[79] Aleksandra Gabryjalowicz. “BIM-GIS integration for energy performance stud-
ies”. English. Supervisor: Alexandru Nichersu. MA thesis. Karlsruhe: KIT,
2016.
[80] A Geiger, A Nichersu, and V Hagenmayer. “Sensitivity of input data in build-
ing heating energy demand simulation”. In: BauSIM 2020 – 8th Conference of
IBPSA Germany and Austria. Graz, Austria, 2020, pp. 450–457.
[81] Andreas Geiger, Alexandru Nichersu, Karl-Heinz Häfele, and Veit Hagen-
mayer. “Usage profile enrichment of CityGML models for urban building
energy modeling”. In: Proceedings of IBPSA BauSIM 2022. Weimar, Germany:
IBPSA, 2022.
[82] Christopher Gold. Spatial context An introduction to fundamental computer algo-
rithms for spatial analysis. ISPRS Book Series 8. Taylor & Francis Group, 2016.
ISBN: 978-1-138-02963-7.
[83] Rachel Gonsenhauser, Katy Hansen, Walker Grimshaw, Jannette Morris, Klaus
Albertin, and Megan Mullin. “Digitizing a Statewide Map of Community Wa-
ter System Service Areas.” In: Journal AWWA - American Water Works Associ-
ation 112.10 (2020), pp. 56–61. DOI: https://doi.org/10.1002/awwa.1595.
(Accessed on: 12/01/2021).
[84] Michael F. Goodchild. “Scale in GIS: An overview”. English (US). In: Geomor-
phology 130.1-2 (07/2011). Publisher: Elsevier, pp. 5–9. ISSN: 0169-555X. DOI:
10.1016/j.geomorph.2010.10.004.
[85] Niklas Grießbaum. “Coupling residential heating systems - Modeling net-
works of distributed micro co-generation units incorporating electrical heat
pumps to satisfy residential heating needs”. English. MA thesis. Karlsruhe:
KIT, 2012.
[86] Markus Groissböck. “Are open source energy system optimization tools ma-
ture enough for serious use?” In: Renewable and Sustainable Energy Reviews
102 (2019), pp. 234–248. ISSN: 1364-0321. DOI: https://doi.org/10.1016/j.
rser.2018.11.020.
[87] Michael Groves and Klaus Mundt. “Friend or foe? Google Translate in lan-
guage for academic purposes”. In: English for Specific Purposes 37 (01/2015),
pp. 112–121. ISSN: 0889-4906. DOI: 10.1016/j.esp.2014.09.001.
[88] Marc Großklos. Kumulierter Energieaufwand und CO2-Emissionsfaktoren verschi-
edener Energieträger und –versorgungen. German. 2011. URL: https : / / www .
iwu.de/fileadmin/tools/kea/kea.pdf (Accessed on: 03/13/2022).
[89] G. Gröger and L. Plümer. “How to achieve consistency for 3D city models”.
In: GeoInformatica 15 (2011), pp. 137–165.
[90] Gerhard Gröger and Lutz Plümer. “CityGML – Interoperable semantic 3D
city models”. en. In: ISPRS Journal of Photogrammetry and Remote Sensing 71
(07/2012), pp. 12–33. ISSN: 09242716. DOI: 10.1016/j.isprsjprs.2012.04.
004. (Accessed on: 10/07/2018).
148 Bibliography
//citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.19.9831&rep=
rep1&type=pdf (Accessed on: 03/13/2022).
[103] Tianzhen Hong, Yixing Chen, Xuan Luo, Na Luo, and Sang Hoon Lee. “Ten
questions on urban building energy modeling”. en. In: Building and Environ-
ment 168 (01/2020), p. 106508. ISSN: 0360-1323. DOI: 10.1016/j.buildenv.
2019.106508. (Accessed on: 11/27/2019).
[104] Andreas Huber and Ines Mayer. “#215-Huber_Mayer-Smart Cities: an emerg-
ing city concept to frame sustainable transitions?” en. In: The Technical Uni-
versity of Denmark, Copenhagen, Denmark, 08/2012, p. 14. URL: shorturl.
at/dBLQT (Accessed on: 10/01/2018).
[105] IEA. Access to electricity, urban (% of urban population) | Data. English. Statisical
Database. 2018. URL: https://data.worldbank.org/indicator/EG.ELC.
ACCS.UR.ZS?end=2017&start=1990 (Accessed on: 10/10/2019).
[106] IEA. Energy Technology Perspectives 2017. Tech. rep. Paris: IEA, 2017. URL:
https://www.iea.org/reports/energy-technology-perspectives-2017.
[107] Ines Imbert, Marie Sevenet, Patrice Nogues, and Jean-Marie Bahu. “Messung
von Energiearmut trotz Datenmangels.” In: Energie und soziale Ungleichheit:
Zur gesellschaftlichen Dimension der Energiewende in Deutschland und Europa.
Ed. by Katrin Großmann, André Schaffrin, and Christian Smigiel. Wiesbaden:
Springer Fachmedien Wiesbaden, 2017, pp. 263–290. ISBN: 978-3-658-11723-8.
DOI : 10.1007/978- 3- 658- 11723- 8_10. URL : https://doi.org/10.1007/
978-3-658-11723-8_10.
[108] infas. infas360. Commercial product, only accessible by paid subscription.
2013. URL: https://www.infas360.de/ (Accessed on: 03/16/2022).
[109] buildingSMART International. openBIM. en-GB. 2021. URL: shorturl . at /
gwCL2 (Accessed on: 05/20/2021).
[110] ISO. ISO 16739-1:2018. en. 2018. URL: https://www.iso.org/cms/render/
live / en / sites / isoorg / contents / data / standard / 07 / 03 / 70303 . html
(Accessed on: 05/20/2021).
[111] ISO. ISO/AWI 37166 Smart community infrastructures —Specification of multi-
source urban data integration for smart city planning(SCP). en. 09/2019. URL:
http://www.iso.org/cms/render/live/en/sites/isoorg/contents/data/
standard/06/92/69252.html (Accessed on: 09/19/2019).
[112] IWU. Systematik und Datensätze Deutsche Gebäudetypologie. German. 2005. URL:
https://www.iwu.de/fileadmin/publikationen/gebaeudebestand/2003_
IWU _ Deutsche - Geb % C3 % A4udetypologie - Systematik - und - Datens % C3 %
A4tze.pdf (Accessed on: 03/12/2022).
[113] Yao Jiacheng, Alexander Simons, and Alexandru Nichersu. Refurbishment cost
model. English. 2015.
[114] Shaohua Jiang and Zheng Wu. “A BIM-Based Code Checking Approach for
Green Construction”. en. In: International Conference on Construction and Real
Estate Management 2017. American Society of Civil Engineers, 11/2017, pp. 156–
163. DOI: 10.1061/9780784481066.016. (Accessed on: 05/24/2021).
[115] Eike Juerrens. SensorEventService < SensorWeb < Wiki. Wiki page. 2015. URL:
https://tinyurl.com/y7whc2rc (Accessed on: 11/28/2021).
[116] Byung Chul Kim, Duhwan Mun, and Soonhung Han. “Retrieval of CAD
model data based on Web Services for collaborative product development
in a distributed environment”. In: The International Journal of Advanced Manu-
facturing Technology 50.9 (10/2010), pp. 1085–1099. ISSN: 1433-3015. DOI: 10.
1007/s00170-010-2571-0.
150 Bibliography
[117] Lawrence A Klein. Sensor and data fusion: a tool for information assessment and
decision making. English. SPIE press, 2004. ISBN: 978-0-8194-5435-5.
[118] Rosemarie Klemm and Dietrich Klemm. Gold and Gold Mining in Ancient Egypt
and Nubia. English. Springer, 2013. ISBN: 978-3-642-22508-6. URL: https://
tinyurl.com/3hpsey7c (Accessed on: 03/06/2021).
[119] K Klobut, T Mäkeläinen, A Huovila, J Hyvärinen, and J Shemeikka. “KPI
framework for energy efficient buildings and neighbourhoods”. en. In: 11th
European Conference on Product and Process Modelling: eWork and eBusiness in
Architecture, Engineering and Construction. Limassol, Cyprus, 2016, p. 7. URL:
https://www.researchgate.net/publication/317872245_KPI_framework_
for _ energy _ efficient _ buildings _ and _ neighbourhoods (Accessed on:
01/04/2021).
[120] Michele Knobel. Digital Literacies: Concepts, Policies and Practices. en. Google-
Books-ID: doVQq67wWSwC. Peter Lang, 2008. ISBN: 978-1-4331-0169-4.
[121] Andreas Koch. “Continuous Simulation for Urban Energy Planning Based on
a Non-Linear Data-Driven Modelling Approach”. en. PhD thesis. KIT, 2017.
[122] Thomas Kolbe, Zhihang Yao, Claus Nagel, Felix Kunde, Philipp Willkomm,
György Hudra, and Arda Müftüoglu. 3D City Database for CityGML v3. En-
glish. 2015. URL: https://tinyurl.com/2sws5w4u (Accessed on: 09/25/2019).
[123] Smita Kumari and Santanu Kumar Rath. “Performance comparison of SOAP
and REST based Web Services for Enterprise Application Integration”. In:
2015 International Conference on Advances in Computing, Communications and
Informatics (ICACCI). Kochi, India, 08/2015, pp. 1656–1660. DOI: 10 . 1109 /
ICACCI.2015.7275851.
[124] Tatjana Kutzner, Kanishk Chaturvedi, and Thomas H. Kolbe. “CityGML 3.0:
New Functions Open Up New Applications”. In: PFG – Journal of Photogram-
metry, Remote Sensing and Geoinformation Science 88.1 (02/2020), pp. 43–61.
ISSN: 2512-2819. DOI : 10.1007/s41064-020-00095-z.
[125] Tatjana Kutzner and Thomas H Kolbe. CityGML UtilityNetwork ADE – Scope,
Concepts, and Applications. en. New York, NY, USA, 2017. URL: https : / /
123dok.com/document/y9gwlvrq-citygml-utility-network-ade-scope-
concepts-applications.html (Accessed on: 04/04/2022).
[126] Tatjana Kutzner and Alexandru Nichersu. Minutes of the UtilityNetwork ADE
Workshop. English. 03/2017. URL: https : / / en . wiki . utilitynetworks .
sig3d.org/images/upload/Minutes_2017_03_UtilityNetworkADE_Workshop.
pdf (Accessed on: 10/06/2019).
[127] Manfred Köhler. “Green facades—a view back and some visions”. In: Urban
Ecosystems 11.4 (2008). Publisher: Springer, pp. 423–436.
[128] Ruben de Laat. opensourceBIM/BIMserver. original-date: 2013-05-08. 05/2021.
URL : https://tinyurl.com/yjy9heme (Accessed on: 05/24/2021).
[129] Ruben de Laat and Léon van Berlo. “Integration of BIM and GIS: The De-
velopment of the CityGML GeoBIM Extension”. en. In: Advances in 3D Geo-
Information Sciences. Ed. by Thomas H. Kolbe, Gerhard König, and Claus
Nagel. Lecture Notes in Geoinformation and Cartography. Berlin, Heidel-
berg: Springer, 2011, pp. 211–225. ISBN: 978-3-642-12670-3. DOI: 10 . 1007 /
978-3-642-12670-3_13. (Accessed on: 05/24/2021).
[130] Julien-Samuel Lacroix. PGRouting for dummies. Nottingham, 2013. URL: https:
//www.mapgears.com/downloads/foss4g2013/foss4g2013-pgrouting-for-
dummies.pdf (Accessed on: 03/24/2022).
[131] Pablo Landeta, Jorge Vásquez, Xavier Rea, and Iván García-Santillán. “Edi-
tion, Publication and Visualization of Geoservices Using Open-Source Tools”.
Bibliography 151
[142] Ziwen Liu, Yujie Lu, and Lu Chang Peh. “A Review and Scientometric Anal-
ysis of Global Building Information Modeling (BIM) Research in the Archi-
tecture, Engineering and Construction (AEC) Industry”. In: Buildings 9.10
(2019). ISSN: 2075-5309. DOI: 10.3390/buildings9100210.
[143] T. Loga and U. Imkeller-Benjes. Energiepaß Heizung/Warmwasser - Energetis-
che Qualität von Baukörper und Heizungssystem. German. 1997. URL: https :
//www.iwu.de/fileadmin/tools/ephw/1997_IWU_LogaImkeller-Benjes_
Energiepass-Heizung-Warmwasser-EPHW.pdf (Accessed on: 03/13/2022).
[144] S. Logothetis, E. Karachaliou, E. Valari, and E. Stylianidis. “OPEN SOURCE
CLOUD-BASED TECHNOLOGIES FOR BIM”. en. In: ISPRS - International
Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
XLII-2 (05/2018), pp. 607–614. ISSN: 2194-9034. DOI: 10.5194/isprs-archives-
XLII-2-607-2018. (Accessed on: 05/24/2021).
[145] Maa-amet. Estonian geoportal. en. 11/2021. URL: https://geoportaal.maaamet.
ee / eng / INSPIRE / index . php ? lang _ id = 2 & page _ id = 742 (Accessed on:
11/30/2021).
[146] Azad Madni, Carla Madni, and Scott Lucero. “Leveraging Digital Twin Tech-
nology in Model-Based Systems Engineering”. en. In: Systems 7.1 (01/2019),
p. 7. ISSN: 2079-8954. DOI: 10.3390/systems7010007. (Accessed on: 09/17/2019).
[147] Prodromos Makris, Dimitrios J. Vergados, Ioannis Mamounakis, Georgios
Tsaousoglou, Konstantinos Steriotis, Nikolaos Efthymiopoulos, and Emmanouel
Varvarigos. “A Novel Research Algorithms and Business Intelligence Tool
for Progressive Utility’s Portfolio Management in Retail Electricity Markets”.
In: 2019 IEEE PES Innovative Smart Grid Technologies Europe (ISGT-Europe).
Bucharest: IEEE, 2019. ISBN: 978-1-5386-8218-0. DOI: 10.1109/ISGTEurope.
2019.8905515.
[148] Avichal Malhotra, Julian Bischof, James Allan, James O’Donnell, Thomas Sch-
wengler, Joachim Benner, and Gerald Schweiger. “City Quarter Information
Modelling for Building Energy - A review”. English. In: BauSIM 2020 : 8th
Conference of IBPSA Germany and Austria, 23-25 September 2020, Graz University
of Technology, Austria : proceedings /. Graz, Austria: Graz University of Tech-
nology, 2020, pp. 543–549. DOI: 10.18154/RWTH- 2020- 10859. (Accessed on:
01/31/2022).
[149] Avichal Malhotra, Julian Bischof, Alexandru Nichersu, Karl-Heinz Häfele, Jo-
hannes Exenberger, Divyanshu Sood, James Allan, Jérôme Frisch, Christoph
van Treeck, James O’Donnell, and Gerald Schweiger. “Information modelling
for urban building energy simulation—A taxonomic review”. In: Building and
Environment (2021), p. 108552. ISSN: 0360-1323. DOI: https://doi.org/10.
1016/j.buildenv.2021.108552.
[150] Vida Maliene, Vytautas Grigonis, Vytautas Palevičius, and Sam Griffiths. “Ge-
ographic information system: Old principles with new capabilities”. In: UR-
BAN DESIGN International 16.1 (01/2011), pp. 1–6. ISSN: 1468-4519. DOI: 10.
1057/udi.2010.25.
[151] Prateek Manke, Yogesh K. Garg, and Vinay M. Das. “Energy simulation tools
and CAD interoperability: A critical review”. In: 2013 International Conference
on Energy Efficient Technologies for Sustainability. 2013, pp. 191–196. DOI: 10.
1109/ICEETS.2013.6533381.
[152] Tathiane Martins, Luc Adolphe, and Marion Bonhomme. “Building Energy
Demand Based on Urban Morphology Analysis:” en. In: PLEA2013. Munich,
Germany: TUM, 2013, p. 6.
Bibliography 153
[153] Priit Martinson. Estonia – the Digital Republic Secured by Blockchain. en. Estonia,
2017. URL: https://www.pwc.com/gx/en/services/legal/tech/assets/
estonia- the- digital- republic- secured- by- blockchain.pdf (Accessed
on: 04/05/2022).
[154] Eoghan McKenna, Ian Richardson, and Murray Thomson. “Smart meter data:
Balancing consumer privacy concerns with legitimate applications”. In: En-
ergy Policy 41 (2012), pp. 807–814. ISSN: 0301-4215. DOI: https://doi.org/
10.1016/j.enpol.2011.11.049.
[155] Boisse Meagan. “Ursula Eicker is Concordia’s new Canada Excellence Re-
search Chair in Smart, Sustainable and Resilient Communities and Cities”.
English. In: Concordia University News (04/2019). URL: https://www.concordia.
ca / content / shared / en / news / stories / 2019 / 04 / 17 / ursula - eicker -
is - concordias - new - canada - excellence - research - chair - in - smart -
sustainable - and - resilient - communities - and - cities . html (Accessed
on: 10/23/2019).
[156] Michael W. Mehaffy, Sergio Porta, and Ombretta Romice. “The "neighbor-
hood unit" on trial: A case study in the impacts of urban morphology”. In:
Journal of Urbanism: International Research on Placemaking and Urban Sustain-
ability 8.2 (2015), pp. 199–217. DOI: https://doi.org/10.1080/17549175.
2014.908786. (Accessed on: 01/24/2021).
[157] Nabor C. Mendonça, David Garlan, Bradley Schmerl, and Javier Cámara.
“Generality vs. Reusability in Architecture-Based Self-Adaptation: The Case
for Self-Adaptive Microservices”. In: Proceedings of the 12th European Confer-
ence on Software Architecture: Companion Proceedings. ECSA ’18. event-place:
Madrid, Spain. New York, NY, USA: Association for Computing Machinery,
2018. ISBN: 978-1-4503-6483-6. DOI: 10.1145/3241403.3241423.
[158] Christopher D. Michaelis and Daniel P. Ames. “Considerations for Imple-
menting OGC WMS and WFS Specifications in a Desktop GIS”. In: Journal of
Geographic Information System Vol.04No.02 (2012). Type: 10.4236/jgis.2012.42021,
p. 7. URL: //www.scirp.org/journal/paperinformation.aspx?paperid=
18782.
[159] Ralf Mosshammer, Friederich Kupzog, Mario Faschang, and Matthias Stifter.
“Loose coupling architecture for co-simulation of heterogeneous components”.
In: IECON 2013 - 39th Annual Conference of the IEEE Industrial Electronics So-
ciety. ISSN: 1553-572X. 11/2013, pp. 7570–7575. DOI: 10.1109/IECON.2013.
6700394.
[160] S.M. Murshed, A. Lindsay, S. Picard, and A. Simons. “PLANTING: Comput-
ing High Spatio-temporal Resolutions of Photovoltaic Potential of 3D City
Models”. In: The Annual International Conference on Geographic Information Sci-
ence. Geospatial Technologies for All. Springer International Publishing, 2018.
ISBN: 978-3-319-78207-2. URL : https : / / doi . org / 10 . 1007 / 978 - 3 - 319 -
78208-9.
[161] Gunnar Myhre, D Shindell, FM Bréon, W Collins, J Fuglestvedt, J Huang, D
Koch, JF Lamarque, D Lee, B Mendoza, and others. Anthropogenic and natural
radiative forcing, climate change 2013: the Physical science basis. Contribution of
Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on
Climate Change, 659–740. 2013.
[162] Renáta Máchová, Miloslav Hub, and Martin Lnenicka. “Usability evalua-
tion of open data portals”. In: Aslib Journal of Information Management 70.3
(01/2018). Publisher: Emerald Publishing Limited, pp. 252–268. ISSN: 2050-
3806. DOI: 10.1108/AJIM-02-2018-0026. (Accessed on: 05/23/2021).
154 Bibliography
[175] OECD. The Metropolitan Century: Understanding Urbanisation and its Conse-
quences. OECD Publishing, Paris, 2015. DOI: https : / / doi . org / 10 . 1787 /
9789264228733-en. (Accessed on: 03/12/2022).
[176] OGC. OGC Web Coverage Service (WCS) 2.1 Interface Standard - Core. English.
2018. URL: http://docs.opengeospatial.org/is/17-089r1/17-089r1.html
(Accessed on: 06/01/2021).
[177] Adegboyega Ojo, Edward Curry, and Fatemeh Ahmadi Zeleti. “A Tale of
Open Data Innovations in Five Smart Cities”. In: 2015 48th Hawaii Interna-
tional Conference on System Sciences. ISSN: 1530-1605. 01/2015, pp. 2326–2335.
DOI : 10.1109/HICSS.2015.280.
[178] Léon L. Olde Scholtenhuis, Timo Hartmann, and André G. Dorée. “Charac-
terizing coordination in both loose and very tightly coupled utility recon-
struction processes”. In: roceedings of International Construction Specialty Con-
ference of the Canadian Society for Civil Engineering (ICSC) 5th edition : 2015.
Vancouver: University of British Columbia Library, 06/2015. DOI: 10.14288/
1.0076361.
[179] OpenStreetMap contributors. Planet dump retrieved from https://planet.osm.org.
Published: https://www.openstreetmap.org. 2017.
[180] TNS opinion. PUBLIC OPINION IN THE EUROPEAN UNION. English. Eu-
robarometer 83. European Comission, 07/2015. (Accessed on: 09/18/2019).
[181] Jukka Paatero. “COMPUTATIONAL STUDIES ON VARIABLE DISTRIBUTED
ENERGY SYSTEMS”. English. Doctoral Dissertation. Helsinki: Helsinki Uni-
versity of Technology, 2009. URL: https : / / core . ac . uk / download / pdf /
301128713.pdf (Accessed on: 01/24/2021).
[182] Neal R. Peirce, W. Johnson Curtis, and Farley M. Peters. Century of the City.
English. The Rockefeller Foundation, 12/2006. URL: https://tinyurl.com/
mu9vz6yx (Accessed on: 09/12/2019).
[183] Christopher Pettit, Ashley Bakelmun, Scott N. Lieske, Stephen Glackin, Karl-
son ‘Charlie’ Hargroves, Giles Thomson, Heather Shearer, Hussein Dia, and
Peter Newman. “Planning support systems for smart cities”. In: City, Culture
and Society 12 (2018), pp. 13–24. ISSN: 1877-9166. DOI: https://doi.org/10.
1016/j.ccs.2017.10.002.
[184] Felipe Pezoa, Juan L. Reutter, Fernando Suarez, Martín Ugarte, and Domagoj
Vrgoč. “Foundations of JSON Schema”. In: Proceedings of the 25th International
Conference on World Wide Web. WWW ’16. event-place: Montréal, Québec,
Canada. Republic and Canton of Geneva, CHE: International World Wide
Web Conferences Steering Committee, 2016, pp. 263–273. ISBN: 978-1-4503-
4143-1. DOI: 10.1145/2872427.2883029.
[185] Stefan Pfenninger, Lion Hirth, Ingmar Schlecht, Eva Schmid, Frauke Wiese,
Tom Brown, Chris Davis, Matthew Gidden, Heidi Heinrichs, Clara Heuberger,
Simon Hilpert, Uwe Krien, Carsten Matke, Arjuna Nebel, Robbie Morrison,
Berit Müller, Guido Pleßmann, Matthias Reeg, Jörn C. Richstein, Abhishek
Shivakumar, Iain Staffell, Tim Tröndle, and Clemens Wingenbach. “Open-
ing the black box of energy modelling: Strategies and lessons learned”. In:
Energy Strategy Reviews 19 (2018), pp. 63–71. ISSN: 2211-467X. DOI: https :
//doi.org/10.1016/j.esr.2017.12.002.
[186] I. Pispidikis and E. Dimopoulou. “CITYGML RESTFUL WEB SERVICE: AU-
TOMATIC RETRIEVAL OF CITYGML DATA BASED ON THEIR SEMAN-
TICS. PRINCIPLES, GUIDELINES AND BLDG CONCEPTUAL DESIGN”.
en. In: ISPRS Annals of Photogrammetry, Remote Sensing and Spatial Information
156 Bibliography
[213] Harald Sundmaeker, Patrick Guillemin, Peter Friess, Sylvie Woelfflé, Euro-
pean Commission, and Directorate-General for the Information Society and
Media. Vision and challenges for realising the Internet of things. en. OCLC: 84735
5368. Luxembourg: EUR-OP, 2010. ISBN: 978-92-79-15088-3.
[214] Irina Susorova, Melissa Angulo, Payam Bahrami, and Brent Stephens. “A
model of vegetated exterior facades for evaluation of wall thermal perfor-
mance”. In: Building and Environment 67 (2013), pp. 1–13. ISSN: 0360-1323.
DOI : https://doi.org/10.1016/j.buildenv.2013.04.027. URL : https:
//www.sciencedirect.com/science/article/pii/S0360132313001388.
[215] VC Systems. VC Database. en-US. Software. 2021. URL: https://vc.systems/
en/products/vc-database/ (Accessed on: 10/20/2021).
[216] Daniel Szepielak, Przemyslaw Tumidajewicz, and Lars Hagge. “Integrating
Information Systems Using Web Oriented Integration Architecture and REST-
ful Web Services”. In: 2010 6th World Congress on Services. 2010, pp. 598–605.
DOI : 10.1109/SERVICES.2010.115.
[217] W. Tobler. “Movie Simulating Urban Growth in the Detroit Region”. In: Eco-
nomic Geography 46 (1970), pp. 234–240. URL: https : / / www . jstor . org /
stable/143141?origin=JSTOR-pdf&seq=1 (Accessed on: 01/24/2021).
[218] Erik H. Trainer, Arun Kalyanasundaram, Chalalai Chaihirunkarn, and James
D. Herbsleb. “How to Hackathon: Socio-Technical Tradeoffs in Brief, Inten-
sive Collocation”. In: Proceedings of the 19th ACM Conference on Computer-
Supported Cooperative Work & Social Computing. CSCW ’16. event-place:
San Francisco, California, USA. New York, NY, USA: Association for Com-
puting Machinery, 2016, pp. 1118–1130. ISBN: 978-1-4503-3592-8. DOI: 10 .
1145/2818048.2819946.
[219] A. Tserstou, A. Jonoski, I. Popescu, T. Herman Asumpcao, G. Athanasiou,
A. Kallioras, and I. Nichersu. “SCENT: Citizen Sourced Data in Support of
Environmental Monitoring”. In: 2017 21st International Conference on Control
Systems and Computer Science (CSCS). ISSN: 2379-0482. 05/2017, pp. 612–616.
DOI : 10.1109/CSCS.2017.93.
[220] UNCTAD. Beyond 20/20 WDS - Table view - Total and urban population, an-
nual. English, French. Statistical Database. 2019. URL: https://unctadstat.
unctad.org/wds/TableViewer/tableView.aspx (Accessed on: 11/01/2019).
[221] Union européenne and Commission européenne. EU energy in figures 2018.
en. OCLC: 1079865467. 2018. ISBN: 978-92-79-88737-6.
[222] Department of Economic and Social Affairs United Nations Population Di-
vision. The World’s Cities in 2016 – Data Booklet. United Nations, 2016. URL:
https://www.un.org/en/development/desa/population/publications/
pdf/urbanization/the_worlds_cities_in_2016_data_booklet.pdf (Ac-
cessed on: 09/28/2019).
[223] Trang Vo. “DEVELOPMENT AND VALIDATION OF A GREEN ROOF /
GREEN FAÇADE MODEL FOR ESTIMATING THEIR IMPACT ON BUILD-
INGS ENERGY DEMAND”. English. Supervisor: Alexandru Nichersu M.Sc.
and Dr. Jochen Wendel. MA thesis. Darmstadt: Technische Universität Darm-
stadt, 2018.
[224] Trang Thuy Vo, Alexandru Nichersu, and Jochen Wendel. “Modeling, Moni-
toring, and Validating Green Roof and Green Facade Solutions with Semantic
City Models Using Low Cost Sensors and Open Software Infrastructures”. In:
Urban Science 3.2 (2019). ISSN: 2413-8851. DOI: 10.3390/urbansci3020039.
Bibliography 159
[225] Panagiotis (Peter) A. Vretanos. OGC® Web Feature Service 2.0 Interface Stan-
dard – With Corrigendum. English. 2014. URL: http://docs.opengeospatial.
org/is/09-025r2/09-025r2.html (Accessed on: 05/31/2021).
[226] W3C. Data on the Web Best Practices. en. 2021. URL: https://www.w3.org/TR/
dwbp/ (Accessed on: 05/23/2021).
[227] Chao Wang and Yong K. Cho. “Automated gbXML-Based Building Model
Creation for Thermal Building Simulation”. In: 2014 2nd International Confer-
ence on 3D Vision. Vol. 2. 2014, pp. 111–117. DOI: 10.1109/3DV.2014.109.
[228] Cheng-Kai Wang, Simon Tindemans, Clayton Miller, Giorgio Agugiaro, and
Jantien Stoter. “Bayesian calibration at the urban scale: a case study on a large
residential heating demand application in Amsterdam”. In: Journal of Building
Performance Simulation 13.3 (2020). Publisher: Taylor & Francis, pp. 347–361.
DOI : 10.1080/19401493.2020.1729862.
[229] O. Wannenwetsch and T. A. Majchrzak. “On constructing persistent identi-
fiers with persistent resolution targets”. In: 2016 Federated Conference on Com-
puter Science and Information Systems (FedCSIS). 09/2016, pp. 1031–1040.
[230] Gregory J. Ward. “The RADIANCE Lighting Simulation and Rendering Sys-
tem”. In: Proceedings of the 21st Annual Conference on Computer Graphics and
Interactive Techniques. SIGGRAPH ’94. New York, NY, USA: Association for
Computing Machinery, 1994, pp. 459–472. ISBN: 0-89791-667-0. DOI: 10.1145/
192161.192286.
[231] Jochen Wendel, Alexander Simons, Alexandru Nichersu, and Syed Monjur
Murshed. “Rapid development of semantic 3D city models for urban energy
analysis based on free and open data sources and software”. en. In: Proceed-
ings of the 3rd ACM SIGSPATIAL Workshop on Smart Cities and Urban Analytics -
UrbanGIS’17. Redondo Beach, CA, USA: ACM Press, 2017, pp. 1–7. ISBN: 978-
1-4503-5495-0. DOI: 10.1145/3152178.3152193. (Accessed on: 10/08/2018).
[232] E. Widl, C. Agugiaro, and P. Puerto. “FIRST STEPS TOWARDS LINKING SE-
MANTIC 3D CITY MODELLING AND MULTI-DOMAIN CO-SIMULATION
FOR ENERGY MODELLING AT URBAN SCALE”. en. In: ISPRS Annals of
Photogrammetry, Remote Sensing and Spatial Information Sciences IV-4 (09/2018),
pp. 227–234. ISSN: 2194-9050. DOI: 10.5194/isprs-annals-IV-4-227-2018.
URL : https://www.isprs- ann- photogramm- remote- sens- spatial- inf-
sci.net/IV-4/227/2018/ (Accessed on: 10/01/2018).
[233] Manfred Wieland. “Computing Solar Radiation on Buildings at an Urban
Scale”. English. Supervisor: Alexandru Nichersu. MA thesis. Karlsruhe: Hoch
schule Karlsruhe, 2014.
[234] Manfred Wieland, A. Nichersu, S.M. Murshed, and Jochen Wendel. “Com-
puting Solar Radiation on CityGML Building Data”. In: Proceedings of the 18th
AGILE conference on Geographic Information Science. Issue: June. 2015, pp. 2–5.
URL : https://agile- online.org/conference_paper/cds/agile_2015/
shortpapers/107/107_Paper_in_PDF.pdf.
[235] Mark D Wilkinson, Michel Dumontier, I Jsbrand Jan Aalbersberg, Gabrielle
Appleton, Myles Axton, Arie Baak, Niklas Blomberg, Jan-Willem Boiten, Luiz
Bonino da Silva Santos, Philip E Bourne, Jildau Bouwman, Anthony J Brookes,
Tim Clark, Mercè Crosas, Ingrid Dillo, Olivier Dumon, Scott Edmunds, Chris
T Evelo, Richard Finkers, Alejandra Gonzalez-Beltran, Alasdair J G Gray,
Paul Groth, Carole Goble, Jeffrey S Grethe, Jaap Heringa, Peter A C ’t Hoen,
Rob Hooft, Tobias Kuhn, Ruben Kok, Joost Kok, Scott J Lusher, Maryann E
Martone, Albert Mons, Abel L Packer, Bengt Persson, Philippe Rocca-Serra,
Marco Roos, Rene van Schaik, Susanna-Assunta Sansone, Erik Schultes, Thierry
160 Bibliography
Sengstag, Ted Slater, George Strawn, Morris A Swertz, Mark Thompson, Jo-
han van der Lei, Erik van Mulligen, Jan Velterop, Andra Waagmeester, Peter
Wittenburg, Katherine Wolstencroft, Jun Zhao, and Barend Mons. “The FAIR
Guiding Principles for scientific data management and stewardship”. eng. In:
Scientific data 3 (03/2016). Publisher: Nature Publishing Group, pp. 160018–
160018. ISSN: 2052-4463. DOI: 10.1038/sdata.2016.18.
[236] Lixian Wu, Yujie Lin, and Weilin Pang. “Distribution Network Topology Mod-
elling and Automatic Mapping Based on CIM and GIS”. In: 2018 IEEE 4th In-
formation Technology and Mechatronics Engineering Conference (ITOEC). 12/2018,
pp. 1–5. DOI: 10.1109/ITOEC.2018.8740469.
[237] Shinji Yamamura, Liyang Fan, and Yoshiyasu Suzuki. “Assessment of Urban
Energy Performance through Integration of BIM and GIS for Smart City Plan-
ning”. In: Procedia Engineering 180 (2017), pp. 1462–1472. ISSN: 1877-7058. DOI:
https://doi.org/10.1016/j.proeng.2017.04.309.
[238] Zhihang Yao, Claus Nagel, Felix Kunde, György Hudra, Philipp Willkomm,
Andreas Donaubauer, Thomas Adolphi, and Thomas H. Kolbe. “3DCityDB
- a 3D geodatabase solution for the management, analysis, and visualization
of semantic 3D city models based on CityGML”. en. In: Open Geospatial Data,
Software and Standards 3.1 (12/2018). ISSN: 2363-7501. DOI: 10.1186/s40965-
018-0046-7.
[239] Di Zhang, Zhenyu Zhou, Shahid Mumtaz, Jonathan Rodriguez, and Takuro
Sato. “One Integrated Energy Efficiency Proposal for 5G IoT Communica-
tions”. In: IEEE Internet of Things Journal 3.6 (12/2016), pp. 1346–1354. ISSN:
2327-4662, 2372-2541. DOI: 10.1109/JIOT.2016.2599852.
[240] Qing Zhao, Hoh In, Xiaowei Wu, and Garng Huang. “Transforming legacy
energy management system (EMS) modules into reusable components: a case
study”. In: Proceedings 24th Annual International Computer Software and Ap-
plications Conference. COMPSAC2000. ISSN: 0730-3157. 10/2000, pp. 105–110.
DOI : 10.1109/CMPSAC.2000.884698.
[241] Junlin Zhou and Yan Fu. “Clustering High-Dimensional Data Using Growing
SOM”. In: Proceedings of the Second International Conference on Advances in Neu-
ral Networks - Volume Part II. ISNN’05. event-place: Chongqing, China. Berlin,
Heidelberg: Springer-Verlag, 2005, pp. 63–68. ISBN: 3-540-25913-9.
[242] Wanji Zhu, Alexander Simons, Sven Wursthorn, and Alexandru Nichersu.
“Integration of CityGML and Air Quality Spatio-Temporal Data Series via
OGC SOS”. en. In: Geospatial Sensor Webs. Münster, Germany: 52°North Ini-
tiative for Geospatial Open Source Software GmbH, 08/2016, p. 6. URL: http:
//ceur-ws.org/Vol-1762/Zhu.pdf (Accessed on: 09/25/2019).
[243] Dimitrios Zissis, Dimitrios Lekkas, Philip Azariadis, Paraskevas Papanikos,
and Elias Xidias. “Collaborative CAD/CAE as a cloud service”. In: Interna-
tional Journal of Systems Science: Operations & Logistics 4.4 (2017). Publisher:
Taylor & Francis, pp. 339–355. DOI: 10.1080/23302674.2016.1186237.
[244] Marcel Šúri and Jaroslav Hofierka. “A New GIS-based Solar Radiation Model
and Its Application to Photovoltaic Assessments”. In: Transactions in GIS 8.2
(2004), pp. 175–190. DOI: https://doi.org/10.1111/j.1467- 9671.2004.
00174.x.