Waste heat valorisation at multiple scales: focus on in-

building waste water and regional heat recovery

THÈSE NO 7927 (2017)

PRÉSENTÉE LE 15 SEPTEMBRE 2017
À LA FACULTÉ DES SCIENCES ET TECHNIQUES DE L'INGÉNIEUR
GROUPE SCI STI FM
PROGRAMME DOCTORAL EN ENERGIE

ÉCOLE POLYTECHNIQUE FÉDÉRALE DE LAUSANNE

POUR L'OBTENTION DU GRADE DE DOCTEUR ÈS SCIENCES

PAR

Alexandre Suzanne Jean-Marie BERTRAND

acceptée sur proposition du jury:

Dr P. Ott, président du jury

Prof. F. Maréchal, Dr R. Aggoune, directeurs de thèse
Prof. M. Patel, rapporteur
Prof. V. Verda, rapporteur
Prof. T. Larsen, rapporteuse

Suisse
2017
 to my grandparents,
who introduced me to the world of physics
by showing me the infeasibility of a perpetuum mobile

to my parents and my sister,

who introduced me to the world of art
Acknowledgements

First of all, I would like to thank my thesis supervisors Dr. Riad Aggoune and Prof. François
Maréchal for their support. I really appreciated that you let me try things out and provided
relevant insights when I missed important issues. Some of our discussions were intense, but
always fruitful and in the sense to make a contribution to science. I really appreciated our
informal relationship.

This work is a compilation of three articles, elaborated with my co-authors Ivan Kantor (addi-
tionnal thanks for the proof reading and language improvements), Alberto Mian (your support
was perfect), Alessio Mastrucci and Nils Schüler. Your feedback was really appreciated and
greatly helped improve the publications - my thanks for that! Also thanks to Alya Bolowich for
the great proof reading and critical review.

My gratitude goes to Brigitte, Sylvie and Cécile for your kindness and efficient support in the
administrative tasks - thank you for being so easy going!

I would like also to thank Cedric Auveiler from Arcelor Mittal, Martin Kox, Gaston Demuth and
Nadine Tornambé-Duchamp of the city of Esch-sur-Alzette, Stefano Beni, Fernand Jungmann
and Georges Liesch from the city of Differdange, Olivier Thunus and Liliane Reichmann from
the national institute of statistics and economic studies (Statec) of Luxembourg, Olivier Auchet
an Laurent Magi from Enovos, Jeannot Behm of Sudstroum, Ingo Both and Anne-Marie Solvi
of Paul Wurth, Fabian Schröer from Wagner Solar GmbH and Sebastian Drefenstedt for their
valuable input concerning the various case studies.

A great source of moral support were my two teams in Switzerland and Luxembourg. My
thanks to the IPESE team: Stefano (thanks for the relax exchanges and the discussion on
integer variables!), Sophia (keep on rockin’), Raluca, Theodoros, Luc, Jessen, Ema, Nasibeh,
Jean-Loup, Shivom, Ligang, Hür, Sébastien, Dilan (anymore dancing?), Victor (the only one
getting my jokes...), Zlatina, Daniel, Maziar, Xun, Katie, Elfie, Raman, Paul, Joël, Araz, Nicolas,
Leandro, Stephane, Adriano, Juliana, Diego, Tivi, Mathias (no more comments about the stuff
on my plate), Sam, Laurence, Samira, Jakob, Priscilla, Manuel, Thierry, Hossein. You folks
always made me feel welcome when I came around, it was really cool to see you and exchange
with you.

i
Acknowledgements

Thanks to the LiSRA/ERIN team, especially for the cool and supportive discussions we already
had and hopefully will continue having: Thomas, Thomas, Thomas and Thomas (spot the
difference), Javier, Daniel, Frank, Olli, Bianca, Pol, Paula, Arno, Sylvain, Benoît, Benedetto
(thanks again for the pasta), Viktor, Emil, Florin, Alya, Antonino, Paul, Rodolphe, Elorri (I know
that this section will be the only part you’ll read but go on, maybe there are some Easter eggs
hidden further on...), Enrico, Seb and Mélanie (let’s make this economy circular now, shall we
?).

My friends regularly checked on how my thesis was doing and even dared to ask what I was
actually doing. That’s really cool, so a huge thanks for your interest and kind support: Let, Alain,
Véro, Ken, Ol, Claude, Tessy, Sandra, Ralph, Martine, Marc, Anik, Ruppert, Béa, Win, Noah,
Enya, Gabi, Christian, Liz, Nathalie, Marta, Cyril, Klôdine, Pierre, Josiane, Angelo, Danielle
(flatmate), Marc (boyfriend/husband), Catherine, Kathy, Véro S., Andrey, Véro P., Paulo, Kim,
Hase, Odile, Laurent.

Jos, Marianne, Marc, Robert, Paul, Christophe: thank you too for your interest in my work and
your warm welcom and inclusion in the Thill family.

This work would not have been possible without the support of my parents, my sister and my
grandmother. You large interest for so many different topics is a source of inspiration to me
and your love is a strong support. Thank you.

Peggy, merci fier déng Ënnerstëtzung, déng Gedold an dain Interess. Ouni dëch hätt ech daat
heiten définitiv nët esou gepackt. Ech frée mech op eis gemeinsam Zukunft :¬).

This work was supported by the Luxembourgish National

Research Fund, grant agreement AFR - 5775018 “Opti-
Heat”.

ii
Abstract

Environmental issues like climate change, resources depletion and the pollution of air, water
and soil represent major challenges for the sustainable development of our society. Several
of these issues are related to energy consumption, and energy efficient solutions must be
proposed to address these. In that framework, this thesis focuses on waste heat valorisation
by proposing methods for the data characterisation, assessment and optimisation of heat
recovery at the building, city and regional scales.

As a first contribution, a method for the modelling of domestic hot water streams (shower,
bath, washing up, etc.) in hotels, nursing homes and households at the building scale is
presented. This method provides specific data on energy consumption and temperature level
for each stream, taking into account the number of inhabitants and households, the end-use
occurrence, as well as the use frequency and duration. The energy consumption can be put in
relation to the total heating demand of the building and, by spatial allocation and bottom-up
data aggregation, of a district, a city or a region. It is demonstrated in a case study that with
the construction of near-zero energy buildings and the improvement of the thermal envelope
of existing ones, domestic hot water will represent in the future between 30 and 50% of the
residential building heating demand.

Considering these findings, measures to improve the energy efficiency related to domestic hot
water use must be addressed. Residential waste water streams are therefore complementarily
characterised, and energy consumption and investment cost calculations methods at the
building level are presented. Different in-building waste water heat recovery configurations
are then assessed using pinch analysis. The residential heating savings of shower and grey
water heat recovery systems of a real case study range between 1 and 12%, while in high effi-
ciency residential buildings savings between 6 and 22% are obtained. An integrated approach
combining heat recovery, temperature optimisation and heat pump use leads to heating
savings ranging between 28 and 41% in high efficiency single family and multifamily buildings,
respectively, therefore demonstrating the relevance of such a holistic method.

Concerning regional waste heat valorisation, an integrated optimisation method, based on a

mixed integer linear programming model and maximising profits for energy service companies,
is proposed. The model takes into account the specific energy price of the heat sinks, the

iii
Abstract

distances between sources and potential users, the energy losses due to heat transportation
as well as the related investment and operating costs. The model also includes the optimal
selection of the backup heating technology. With the energy prices of 2015 and considering
the Southern region of Luxembourg, the production of electricity proves to be viable for waste
heat prices below 10 e/MWh, while waste heat could still be valorised for heating demand
at prices up to 25 e/MWh. At that price, profits of more than 10 Me/a from the transport of
waste heat and the electricity production of combined heat and power systems are obtained.

By proposing improved and novel methods related to the valorisation of waste heat at multiple
scales, this thesis supports the development of energy efficiency solutions in and across the
building, commercial and industrial sectors.

Key words: domestic hot and waste water characterisation, in-building waste water heat
recovery, MILP-based regional waste heat valorisation, energy service companies.

iv
Résumé

Les problèmes environnementaux comme le changement climatique, l’épuisement des ressou-

rces ainsi que la pollution des milieux naturels représentent des défis majeurs pour le développ-
ement durable de notre société. Plusieurs de ces impacts environnementaux sont liés à la
consommation d’énergie et des solutions à haute performance énergétique doivent être
développées pour les réduire. Dans ce cadre, cette thèse porte sur la valorisation de la
chaleur excédentaire en proposant des méthodes de caractérisation de donnée, d’évaluation
et d’optimisation de la récupération de chaleur à l’échelle du bâtiment, de la ville et de la
région.

En premier lieu, une méthode de caractérisation des flux d’eau chaude sanitaire (douche,
bain, etc.) dans les hôtels, les maisons de soins et les ménages à l’échelle du bâtiment est
proposée. Cette méthode fournit des données spécifiques sur la consommation d’énergie en
fonction du nombre d’habitants et des ménages, de la fréquence et de la durée d’utilisation.
La consommation énergétique peut ainsi être mise en relation avec la demande totale de
chauffage du bâtiment et, par localisation et agrégation des données, d’un quartier, d’une
ville ou d’une région. Il est démontré qu’avec la construction de bâtiments à très faible
consommation en énergie et l’amélioration de l’enveloppe thermique des immeubles existants,
l’eau chaude sanitaire représentera à l’avenir entre 30 et 50% de la demande résidentiel en
chauffage.

Compte tenu de ces résultats, il convient de prendre des mesures pour améliorer l’efficacité
énergétique liée à l’eau chaude sanitaire. Les eaux usées résidentielles sont donc caractérisées
de manière complémentaire et des méthodes de calcul de la consommation d’énergie et
des coûts d’investissement sont proposées. Différentes configurations de récupération de
chaleur des eaux usées sont évaluées à l’aide d’une analyse par pincement. Une approche
intégrée combinant la récupération de chaleur, l’optimisation de la température et l’utilisation
d’une pompe à chaleur conduit à des économies de chauffage entre 28 et 41% dans les
bâtiments unifamiliaux et multifamiliaux à haute performance énergétique, démontrant ainsi
la pertinence d’une telle méthode holistique.

En ce qui concerne la valorisation régionale de la chaleur résiduelle, une méthode d’optim-

isation intégrée maximisant les bénéfices, basée sur un modèle de programmation linéaire

v
Résumé

en nombres entiers mixtes, est proposée pour les entreprises de services énergétiques. Le
modèle prend en compte le prix de l’énergie spécifique des utilisateurs, les distances, les
pertes de chaleur dues au transport, les coûts d’investissement et d’exploitation ainsi que la
sélection optimale du système de chauffage de support. Considérant les prix de l’énergie de
2015 dans un cas d’étude réel, la production d’électricité s’avère viable pour les prix de chaleur
excédentaire en dessous de 10 e/MWh, et cette chaleur pourrait encore être valorisée pour
des besoins de chauffage jusqu’à un prix de 25 e/MWh. A ce prix, des bénéfices de plus de 10
Me/a sont obtenus par le transport de la chaleur et la production d’électricité des systèmes
de cogénération.

En proposant des méthodes améliorées et novatrices liées à la valorisation de la chaleur

excédentaire à des échelles multiples, cette thèse soutient le développement de solutions
améliorant l’efficacité énergétique à travers les secteurs du bâtiment, du commerce et de
l’industrie.

Mots clefs: caractérisation des flux d’eau chaude sanitaire et des eaux usées, récupération de
chaleur sur les eaux usées dans les bâtiments, valorisation régionale de la chaleur excédentaire
sur base d’une approche de programmation linéaire en nombres entiers mixtes, entreprises
de services énergétiques.

vi
Contents

Acknowledgements i

Abstract (English/Français) iii

List of Figures xiii

List of Tables xv

1 Introduction 1

1.1 Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Scientific background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.2.1 Domestic hot water streams characterisation . . . . . . . . . . . . . . . . 4

1.2.2 In-building waste water heat recovery . . . . . . . . . . . . . . . . . . . . . 5

1.2.3 Regional waste heat valorisation . . . . . . . . . . . . . . . . . . . . . . . . 6

1.3 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

1.4 Outlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2 Domestic hot water characterisation 9

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.2 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.2.1 Characterisation methodology . . . . . . . . . . . . . . . . . . . . . . . . . 13

vii
Contents

[Link] DHW end-use load . . . . . . . . . . . . . . . . . . . . . . . . . . 13

[Link] DHW and space heating energy demand . . . . . . . . . . . . . . 13

[Link] Utility load calculation . . . . . . . . . . . . . . . . . . . . . . . . 15

2.2.2 Input data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

[Link] Temperatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

[Link] Volumetric flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

[Link] Use duration and frequency . . . . . . . . . . . . . . . . . . . . . 18

[Link] Streams occurrence and spatial allocation . . . . . . . . . . . . . 18

[Link] Occupant and unit spatial allocations . . . . . . . . . . . . . . . 22

2.3 Case study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

2.3.1 Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

[Link] DHW energy consumption . . . . . . . . . . . . . . . . . . . . . . 24

[Link] Total heat demand using linear regression . . . . . . . . . . . . . 26

[Link] Space heating energy consumption . . . . . . . . . . . . . . . . . 28

[Link] Heat exchanger load for district heating systems . . . . . . . . . 28

[Link] Energy integration of decentralised heat pump considering vari-

ous hot water temperature levels . . . . . . . . . . . . . . . . . . 29

2.3.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

[Link] Validation of outcomes with literature values . . . . . . . . . . . 29

[Link] Contribution of DHW streams to household energy consumption 33

[Link] DHW temperature level requirements . . . . . . . . . . . . . . . 33

[Link] Heat exchanger load for district heating systems . . . . . . . . . 34

[Link] Energy integration of decentralised heat pump considering vari-

ous hot water temperature levels . . . . . . . . . . . . . . . . . . 35

2.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

viii
 Contents

3 In-building waste water heat recovery assessment 39

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

3.2 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

3.2.1 Domestic grey water streams characterisation . . . . . . . . . . . . . . . . 44

[Link] Bathroom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

[Link] Kitchen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

[Link] Laundry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

3.2.2 Energy assessment of heat recovery configurations . . . . . . . . . . . . . 47

[Link] Shower heat recovery configurations . . . . . . . . . . . . . . . . 48

[Link] Grey water heat recovery configurations . . . . . . . . . . . . . . 51

3.2.3 Costs calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

3.3 Case studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

3.3.1 Common input data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

3.3.2 Retrofit solutions at urban scale . . . . . . . . . . . . . . . . . . . . . . . . 56

[Link] Specific input data . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

[Link] Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

3.3.3 Energy optimisation of high efficiency residential buildings . . . . . . . . 61

[Link] Specific input data . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

[Link] Energy integration considering WWHR . . . . . . . . . . . . . . . 63

[Link] Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

ix
Contents

4 Optimised regional waste heat valorisation 69

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

4.2 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

4.2.1 Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

4.2.2 Economic constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

[Link] Objective function and global economic constraints . . . . . . . 78

[Link] Revenues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

[Link] Operating costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

[Link] Investment costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

4.2.3 Energy constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

[Link] Heat cascade formulation . . . . . . . . . . . . . . . . . . . . . . . 83

[Link] Heat losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

4.3 Case study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

4.3.1 Input data description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

[Link] General data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

[Link] Process data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

[Link] Economic data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

4.3.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

[Link] Regional valorisation considering different waste heat selling

prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

[Link] Regional waste heat valorisation at 25 e/MWh . . . . . . . . . . 92

4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

5 Conclusions and future developments 99

5.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

5.2 Future developments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

x
 Contents

A In-building waste water heat recovery models 103

A.1 Mass flows and temperature levels for shower heat recovery configurations . . 103

A.2 Mass and temperature levels for grey heat recovery configurations . . . . . . . . 104

B Regional waste heat valorisation input data 107

B.1 Distances and road ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

B.2 Pipe data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

B.3 Heat capacity of transportation fluid . . . . . . . . . . . . . . . . . . . . . . . . . . 110

B.4 Piecewise linearisation of heat exchanger investment costs . . . . . . . . . . . . 110

Bibliography 111

CV 137

xi
List of Figures

1.1 World delivered energy according to sector, excluding electricity losses (U.S.
Energy Information Administration [223]) . . . . . . . . . . . . . . . . . . . . . . 2

1.2 World delivered energy type according to sector in 2012 (U.S. Energy Information
Administration [223]) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.3 Waste heat potential according to sector (Forman et al. [78]) . . . . . . . . . . . 3

2.1 Simultaneity factors (Thorsen and Kristjansson [214], Gaderer [84]) . . . . . . . 17

2.2 Districts of Esch-sur-Alzette . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

2.3 Total household yearly heat consumption at district level . . . . . . . . . . . . . 30

2.4 DHW-related energy demand and surface per inhabitant (blue column) of house-
holds at district level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

2.5 DHW energy demand of lodgings . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

2.6 Specific thermal energy demand of households (SFB - single family building,
MFB - multifamily building, MUB: mixed-use building) . . . . . . . . . . . . . . 32

2.7 DHW energy demand contribution in households (SFB - single family building,
MFB - multifamily building, MUB: mixed-use building) . . . . . . . . . . . . . . 33

2.8 End-use contribution to household DHW energy demand (SFB - single family
building, MFB - multifamily building, MUB: mixed-use building) . . . . . . . . 34

2.9 Temperature level contribution to household DHW energy demand (SFB - single
family building, MFB - multifamily building, MUB: mixed-use building) . . . . 35

2.10 Utility heating load at district level . . . . . . . . . . . . . . . . . . . . . . . . . . 35

xiii
List of Figures

2.11 Composite curves of a low energy, single family building . . . . . . . . . . . . . . 36

3.1 In-building WWHR assessment method at the urban scale . . . . . . . . . . . . . 44

3.2 Pinch analysis diagram - balanced conditions . . . . . . . . . . . . . . . . . . . . 48

3.3 Configuration 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

3.4 Configuration 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

3.5 Configuration 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

3.6 Pinch analysis diagram - DHW and WW streams . . . . . . . . . . . . . . . . . . . 53

3.7 Configuration 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

3.8 Configuration 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

3.9 Configuration 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

3.10 Scenario 1.1: horizontal shower heat exchanger - relative energy savings (SFB –
single family building, MFB – multifamily building, MUB - multi-use building) 58

3.11 Scenario 1.2: grey water heat recovery - relative energy savings (SFB – single
family building, MFB – multifamily building, MUB - multi-use building) . . . . 58

3.12 Average payback time for one household buildings . . . . . . . . . . . . . . . . . 59

3.13 Average payback time for five households buildings . . . . . . . . . . . . . . . . 60

3.14 Energy savings per districts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

3.15 Scenario 2.3 - Pinch analysis diagram . . . . . . . . . . . . . . . . . . . . . . . . . 64

3.16 Case study 2: results (LE - low energy, P - passive, SFB – single family building,
MFB – multifamily building) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

4.1 Global waste heat valorisation method . . . . . . . . . . . . . . . . . . . . . . . . 77

4.2 Regional heat cascade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

4.3 Regional waste heat sources and sinks . . . . . . . . . . . . . . . . . . . . . . . . . 87

4.4 Profits, revenues and energy quantities in function of waste heat prices . . . . 92

4.5 Waste heat valorisation and selected utility technologies for a waste heat price
of 25 e/MWh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

B.1 Linear regression of pipe costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

B.2 Piecewise linearisation of heat exchanger costs, based on Tarres-Font [213] . . 110

xiv
List of Tables

2.1 DHW end-use temperatures (Yao and Steemers [242], Schramek [192]) . . . . . 18

2.2 Domestic DHW volumetric flows, in l/s . . . . . . . . . . . . . . . . . . . . . . . . 19

2.3 Lodging DHW volumetric flows, in l/s . . . . . . . . . . . . . . . . . . . . . . . . 20

2.4 DHW use frequencies, in capita*day-1 . . . . . . . . . . . . . . . . . . . . . . . . 20

2.5 DHW use durations, in s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

2.6 Dishwasher occurrence rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

2.7 Hotels and nursing homes data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

2.8 Average household surface for mixed-use buildings in Esch (Statec [199]) . . . 26

2.9 Statistical summary of the multiple linear regression for residential buildings . 27

3.1 Bathtub waste water temperatures . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

3.2 Dishwasher grey water streams characterisation, according to Saker et al. [185] 47

3.3 Domestic hot and grey water streams (Bertrand et al. [19]) . . . . . . . . . . . . . 52

3.4 Cost parameters for grey water heat recovery . . . . . . . . . . . . . . . . . . . . 57

3.5 Scenario 1.1 and 1.2 energy savings at city scale, considering 2015 energy price 61

3.6 Building characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

3.7 Operating period characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

4.1 Heat source data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

xv
List of Tables

4.2 Industrial heat sink data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

4.3 Urban heat sink data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

4.4 Revenues and costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

4.5 Investment costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

4.6 Transferred waste heat loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

4.7 Heat losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

B.1 Distances / road ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

B.2 Pipe data (Isoplus [105],Nussbaumer and Thalmann [151]) . . . . . . . . . . . . 109

B.3 Heat capacity of Dowtherm A (Dow Chemical Company [50]) . . . . . . . . . . . 110

xvi
1 Introduction

1.1 Context
Climate change, natural resource depletion and anthropogenic pollution of air, water and
soil represent major environmental challenges for our society. Following the United Nations
Millennium Development Goals, 194 countries adopted the United Nations Sustainable De-
velopment Goals in 2015 addressing these challenges. Among the 17 Goals, one particularly
focuses on the access to affordable, reliable, modern as well as sustainable energy, with the
objective to double the improvement rate of energy efficiency by 2030 (United Nations [221]).
The use of fossil fuels is particularly targeted by this objective, as their availability is limited,
and their combustion is the source of pollutants, like SOx and NOx , leading to negative im-
pacts on human health, flora and fauna (Bilgen [21]). Their consumption is also an important
contributor to greenhouse gas emissions, the cause for climate change. In 2010, 25% of these
greenhouse gas emissions were related to the indirect emissions related to electricity and
heat production (Victor et al. [229]). Energy saving measures therefore address several of the
United Nations Sustainable Development Goals simultaneously.

According to the U.S. Energy Information Administration [223], the global delivered energy,
excluding electricity losses, amounted to 120 PWh in 2012, with the industrial sector making
up 54% of this use, followed by the transport sector with 31%, the residential sector with 13%
and the commercial sector with 7% (Fig. 1.1). The business-as-usual projection does not
foresee major changes in this cross-sectoral distribution. However, the amount of delivered
energy is expected to further increase by 44% until 2040 among all sectors, unless major energy
efficient solutions are implemented.

In terms of energy use, liquid fuels represented 44% of the delivered energy in 2012, natural
gas 20%, electricity 16%, coal 15% and renewable energies 5%. When excluding the transport
sector, where the demand was covered almost 100% by liquid fuels (Fig. 1.2), the delivered
energy was 26% as natural gas and liquid fuels , 22% as electricity, 20% as coal and 6% as
renewable energies.

1
Chapter 1. Introduction

Figure 1.1: World delivered energy according to sector, excluding electricity losses (U.S. Energy
Information Administration [223])

Figure 1.2: World delivered energy type according to sector in 2012 (U.S. Energy Information
Administration [223])

One opportunity to improve energy efficiency in and across the residence, commerce, trans-
port and industry sectors consists in the recovery of waste heat. Forman et al. [78] estimated
the global, theoretical, waste heat potential in 2012 to reach 68 PWh, with 21% of this potential
available at temperatures above 300°C, 16% between 300 and 100°C and 63% below 100°C (Fig.

2
 1.1. Context

1.3). The largest waste heat potential would be obtained from electricity production (44%),
followed by transport (25%), industry (13%), residence (12%) and commerce (6%).

Figure 1.3: Waste heat potential according to sector (Forman et al. [78])

While these outcomes highlight an important potential for waste heat recovery, especially
when only energy quantities are considered, its actual valorisation has so far been limited
by technical, organisational and economic constraints (McKenna and Norman [136], Walsh
and Thornley [235], Chew et al. [41], Brueckner et al. [33]). Financial barriers (high invest-
ment costs, high payback time, financial risks) are among the major hurdles to waste heat
valorisation (Brückner et al. [30], Oluleye et al. [153], Broberg-Viklund [31]). Furthermore, the
detection and selection of adequate heat sinks are often indicated as important constraints to
the reuse of waste heat (Viklund and Johansson [231], Miró et al. [142], Oluleye et al. [154]).

Scientists have been addressing these financial and selection issues for over 40 years with the
development of process integration methods (Klemeš [113]). Process integration consists of
an holistic approach to the optimal combination of operations in a process or between several
processes with the objective to decrease resource consumption and pollutant emissions
(Varbanov [224]). These methods are generally based on pinch analysis (Linnhoff and Flower
[122], Linnhoff et al. [124]), operations research formulations (Papoulias and Grossmann
[159, 160, 161]) or a combination of the two approaches (Maréchal and Kalitventzeff [134]).
While the initial focus of process integration methods was on industrial process and factory
optimisation (e.g. Klemeš et al. [114], Varbanov et al. [226], Amon et al. [6]), the scope of
process integration methods was enlarged to encompass industrial zones to further profit from
waste heat transfer and utility sharing opportunities (e.g. Stijepovic and Linke [207], Hackl
et al. [91], Stijepovic et al. [208]). Two further fields of application are still relatively new
to process integration: urban waste water heat recovery (WWHR) and regional waste heat
valorisation (RWHV).

3
Chapter 1. Introduction

Waste water heat recovery consists in the reuse of the heat contained in the waste water,
either at the building level (in-building solution), or at the sewage level (Hepbasli et al. [96]).
Considering that domestic hot water demand will represent up to 50% of the heating demand
in high efficiency buildings (Frijns et al. [80]), WWHR is expected to become a relevant topic
in the future. The valorised heat can be used either for hot water preheating or as a heat
source for a heat pump to cover the total heat demand of the building. While the potential
of in-sewer heat recovery has been addressed (Abdel-Aal et al. [1], Cipolla and Maglionico
[43], Durrenmatt and Wanner [51]), the energy saving impact of in-building solutions has yet
to be analysed in detail at a building or city level.

Concerning regional waste heat valorisation, this development is a further increase in scope
of process integration methods. As waste heat can not always be valorised on-site or in the
industrial zone due to the above-mentioned constraints, it may be recovered for other users.
Persson and Werner [170] estimated that approximately 90% of the EU residential and service
sectors heating demand could theoretically be covered by excess heat from manufacturing
industries, power production and waste incineration. Regional waste heat valorisation opens
the opportunity to create new energy services related to waste heat management, a subject
that has been barely addressed from the point of view of energy service companies (ESCos).

While assessment and optimisation methods addressing waste water heat recovery and re-
gional waste heat valorisation are applied at different scales, they share the necessity for
high quality input data on building heat requirements. Heating demand characterisation has
been addressed in the past, with a major focus on the modelling of space heating. However,
domestic hot water (DHW) energy modelling, similar to what has already been done in the
water field (Blokker et al. [22], Pieterse-Quirijns et al. [172]), has so far not been considered in
detail.

This thesis therefore targets these three topics by proposing data characterisation, assessment
and optimisation methods supporting the valorisation of waste heat at building, city and
regional scale.

1.2 Scientific background

The scientific background covered in this thesis is presented below, while specific literature
reviews are conducted in the respective Chapters 2, 3 and 4.

1.2.1 Domestic hot water streams characterisation

A large number of publications addressed the quantification and spatial allocation of heating
in urban systems, as demonstrated in the reviews of Keirstead et al. [109], Allegrini et al. [3]
and Reinhart and Davila [182]. The latter authors highlighted the fact that bottom-up urban
system modelling will grow in importance due to their higher level of detail. This modelling

4
 1.2. Scientific background

approach is often coupled to geographical information systems (GIS) data on building foot-
print, height, occupant number etc., which can be used to characterise the building energy
demand. Multiple linear regression analysis combined with GIS data was recurrently used
for the calculation of the space heating demand (e.g. Mastrucci et al. [135], Schüler et al.
[189], Nouvel et al. [150]). However, the reliability of data still needs to be improved to limit the
level of uncertainty of the results (Reinhart and Davila [182]). In particular, a lack of methods
to characterise in detail the energy demand of specific domestic hot water streams (showering,
bathing, washing up, etc.) is observed.

Domestic hot water energy demand in buildings has usually been calculated as a function
of the building surface and considered as a single stream (Girardin et al. [85], Jennings et al.
[106], Fonseca and Schlueter [77], Nouvel et al. [150]). This approach, while reflecting the
centralised hot water production at the heating utility level, fails to acknowledge that hot
water demand depends on the end-use occurrence and occupant number (Blokker et al.
[22]). Also, DHW streams have different temperature level requirements (Yao and Steemers
[242], Schramek [192]), that can influence the heating system design, in particular when
low temperature floor heating is involved (Hesaraki et al. [97]). A more specific bottom-up
approach to DHW characterisation is therefore necessary to better quantify the energy demand
at building and urban scale, thus improving the outcomes of regional waste heat valorisation
methods. Also, the energy consumption of the various DHW streams relative to space heating
has not been specifically quantified yet, an obvious hurdle for the assessment of the relevance
of WWHR systems as energy efficiency measure in the residential sector.

1.2.2 In-building waste water heat recovery

In the field of in-building WWHR, several works focused on horizontal and vertical shower heat
exchangers (e.g. Meggers and Leibundgut [138], Kordana et al. [117], Torras et al. [217]), also
in combination with heat pumps and/or solar energy (Chen et al. [39], Dong et al. [49], Tanha
et al. [212]). Other authors analysed waste water collecting and recovery systems as to their
energy and economic savings (Hepbasli et al. [96], Postrioti et al. [176]). The assessment
methods were based either on experimental setups (Wong et al. [240], Dong et al. [49], Torras
et al. [217], Postrioti et al. [176]), simulation (Liu et al. [125], Chen et al. [39], Torras et al.
[217]) or exergy calculations (Meggers and Leibundgut [138]). As waste water energy data
is still very scarce (Meggers and Leibundgut [138]), a few authors proposed various waste
water models, e.g. for showers (Guo et al. [89], Dong et al. [49]) or dishwashers (Persson and
Werner [170], Bengtsson et al. [14]). Ni et al. [148] proposed a method to derive grey water
temperatures as a function of hot and fresh water temperatures, without, however, providing
further information on temperature loss coefficients used in their approach.

No detailed stream characterising methods considering inhabitant and household numbers

at the building scale have been proposed so far. Furthermore, as past assessments focused
particularly on the heating system, the energy savings were usually related to the energy

5
Chapter 1. Introduction

demand of DHW. These savings were thus not put in relation to the building type (which
constrains the actual implementation of such solution - McNabola and Shields [137]) or the
total heating demand and therefore its energy efficiency. Furthermore, the energy saving
impact of such solutions have barely been considered at large scale. The few publications
addressing this topic only used a simplified top-down approach, considering a fixed energy
saving ratio per building to quantify the impact of WWHR at the level of a district, a city or a
region (Leidl and Lubitz [118], Ni et al. [148], Deng et al. [46]). The specific quantification of
energy savings based on a bottom-up approach has yet to be proposed.

1.2.3 Regional waste heat valorisation

By better characterising the energy demand for domestic hot water, and therefore of urban
systems, the outcomes of RWHV assessments are also improved. The valorisation of industrial
waste heat at a regional scale has been addressed by Perry et al. [166] in the framework of their
locally integrated energy sector approach. The concept of a locally integrated energy sector
is based on total site heat integration (Dhole and Linnhoff [48], Klemeš et al. [114]), which
targets maximal energy savings of industrial processes by process integration. According
to the review of Liew et al. [119], the methodological developments on locally integrated
energy sector approaches mostly focused on the inclusion of the aspect of demand variability,
e.g. with the multi-period approach of Nemet et al. [145]. Waste heat valorisation at large
scale has been considered particularly for urban systems, with the works of e.g. Tveit et al.
[219], Fang et al. [71], Xia et al. [241]. The economic profitability of such concepts was also
recurrently addressed, with the objective function often being the minimisation of investment
and operating costs of the heat user (Svensson et al. [210], Oh et al. [152], Eriksson et al.
[57], Sandvall et al. [188], Sameti and Haghighat [186]). Concerning the actual implementation
of such concepts, Ammar et al. [5], Brueckner et al. [33] and Päivärinne and Lindahl [173]
highlighted the reluctance of industries (both as heat source and sink) and municipalities alike
to actually engage in the planning, investment and management tasks of RWHV projects, as
these are not part of their core activities. The inclusion of an energy service company (ESCo)
as third actor fulfilling these tasks is therefore necessary to implement waste heat valorisation
at the regional scale.

However, specific RWHV optimisation methods for energy service companies have, so far,
not been formulated. While the minimisation of costs are very relevant aspects to such an
approach, the key issue is profitability (Deng et al. [45]). These profits depend also on the
optimal selection of the most appropriate heat sinks, which is influenced by the sink initial
energy price, an aspect, with the exception of Oluleye et al. [153], so far not taken into account.
Furthermore, as heat supply must be guaranteed in case of waste heat supply failure, a backup
heating system must also be included in the system design. The selection of the optimal
technology would further maximise the profits and therefore improve even more the service
value of the ESCo (Brady et al. [28]).

6
 1.3. Objectives

1.3 Objectives

Considering the above-mentioned shortcomings, the objective of this thesis is threefold:

1. Propose a characterisation method for domestic hot water and waste water streams as a
function of specific building characteristics, e.g. occupant number, end-use occurrence,
use duration and frequency.

2. Develop a method to assess in detail the relevance, in terms of energy savings and costs,
of various waste water heat recovery systems in buildings and, by data aggregation, at
the level of a city or a region.

3. Formulate an ESCo-specific optimisation model for the valorisation of waste heat at a

regional scale and for the selection of a heating technology used as backup system.

By characterising DHW according to their temperature levels, thermal loads, end-use occur-
rence as well as inhabitant and household numbers, the quantification of the energy demand
of buildings and of urban systems is improved. With the description of the related waste water
streams complementary to the DHW modelling approach, the opportunity to assess in detail
in-building WWHR systems at the building, city and regional level is taken. Finally, a novel,
integrated, optimisation method on regional waste heat valorisation specifically formulated
for energy service companies is proposed.

1.4 Outlines

The first development of this thesis consists in the characterisation of domestic hot water
streams (Chapter 2). The methodology first covers the equations necessary to determine the
energy consumption of the various DHW streams (Section 2.2). A review of typical European
values for mass flows, temperature levels, occurrence rate, etc. as inputs to these equations is
then conducted. The characterisation method is demonstrated in a real case study applied to
the city of Esch-sur-Alzette (Grand-Duchy of Luxembourg - section 2.3). The resulting out-
comes are compared with literature values. The relevance of the various DHW stream energy
consumption as a function of the building age and thus its thermal efficiency is presented. In
addition, the impact of an integrated approach combining the use of a heat pump and hot
water temperature reduction in a low energy building is demonstrated. The final section sees
the advantages and shortcomings of the proposed method discussed (Section 2.4).

Waste water stream characterisation and their valorisation at the building and city scale are
addressed in Chapter 3. Based on the method and data described in the preceding chapter,
the waste water streams are characterised as to their drain temperature (Section 3.2). Methods
for the calculation of the investment costs and the energy savings using pinch analysis are
proposed in the same section. Furthermore, various shower and grey water heat recovery

7
Chapter 1. Introduction

configurations are assessed as to their energy savings. Two case studies are then presented
(Section 3.3). First, the impact of the implementation of a horizontal shower heat exchanger
and a grey water heat recovery system in the residential buildings of Esch-sur-Alzette is as-
sessed. In the second case study, low energy and high efficiency single family and multifamily
buildings are considered for the application of the same technologies as well as for an inte-
grated approach combining hot water temperature reduction, heat recovery and the use of a
heat pump. The proposed approach is then discussed as to its strengths and disadvantages
(Section 3.4).

A regional waste heat and utility optimisation approach for energy service companies is
described in Chapter 4. Following the literature review, the multi-period, mixed integer linear
programming model is presented (Section 4.2). The economic constraints are formulated
to calculate the profits generated from the revenues of waste heat supply and electricity
production, taking into account operating costs related to energy consumption as well as
annualised investment costs linked to the acquisition of the heat exchangers, pipes and backup
heating systems. The energy constraints are based on the heat cascade formulation, adapted
to consider heat losses from transportation. The method is demonstrated in a case study set
in the Southern region of Luxembourg, with two steel plants as heat sources as well as three
factories and nine towns as heat sinks (Section 4.3). The method is first deployed using various
waste heat prices at the sources. In a second step, the optimisation outcomes, considering a
waste heat price generating the highest revenues for the steel plants, are presented in detail.
The chapter closes with a discussion on the methodology (Section 4.4).

In Chapter 5, conclusions as to the proposed methodological developments and outcomes

of the case studies are drawn. This thesis closes with some perspectives as to future develop-
ments.

8
2 Domestic hot water characterisation

This chapter is the updated postprint version of an article published in Applied Energy (Bertrand et al.
[19]). A preliminary version was presented at the Biennial International Workshop Advances in Energy
Studies (BIWAES) 2015 in Stockholm (Bertrand et al. [15]). This work was elaborated by the author of this
thesis, in collaboration with François Maréchal and Nils Schüler from the Ecole Polytechnique Fédérale
de Lausanne (Switzerland) as well as Riad Aggoune and Alessio Mastrucci from the Luxembourg Institute
of Science and Technology (Luxembourg). The latter co-author contributed the section on total heat
demand calculation using regression analysis (Section [Link]). The outcomes of the case study will be
used by the national institute of statistics and economic studies of Luxembourg (Statec) in its reporting
to the EU concerning the energy consumption of households, as required by the regulation on energy
statistics ( European Parliament [61], Eurostat [64]).

9
Chapter 2. Domestic hot water characterisation

Abbreviations and symbols

Abbreviations

COP coefficient of performance

DH district heating

DHW domestic hot water

GIS geographical information system

MFB multi family building

MUB mixed-use building

OLS ordinary least squares

SFB single family building

SH space heating

Symbols

b specific bed number per room [room-1 ]

cp heat capacity [kJ/kg*K]

d use duration [s]

f use frequency [capita*day]-1 , [household*day]-1

m room occupancy [%]

p bed place occupancy [%]

Q̇ thermal load [kW]

Q energy [kWh]

R2 coefficient of determination [-]

r room number r [-]

Sn DHW simultaneity factor [-]

s stay duration [day]

T temperature [°C]

t time t [h]

V̇ volumetric flow [m3 /s]

x f l oor useful floor surface [m2 ]

x occ number of occupants [occupants]

10
 2.1. Introduction

Symbols

x t y pe building type [-]

x per i od ,a construction period a of buildings

y dwelling energy consumption [kWh/a]

ε random error term [-]

ηutility utility efficiency [-]

ρ density [kg/m3 ]

σ end-use occurrence [%]

υ random value for end-use spatial allocation [-]

Subscripts

b building b

e end-use e

h household h

o occupant o

t time t

u unit u

2.1 Introduction

Cities are responsible for 75 % of the global energy consumption, as well as between 50 to 60%
of the green house gas emissions (UN Habitat [220]) and therefore play a central role in the
global improvement of energy efficiency. In that sense, new buildings shall be designed in the
EU to be near-zero energy until 2020 (European Parliament [62]). In 2013, space heating (SH)
household final energy consumption amounted to 2.33 PWh, while domestic hot water (DHW)
reached 0.44 PWh - representing 16% of the total heat requirements - in the EU28 (Enerdata
[54]). However, with improved thermal insulation, the relevance of DHW consumption is
increasing. In the Netherlands, 23% of the household gas consumption is already related to
DHW use, and new buildings see this contribution reach 50% of the energy consumption
(Frijns et al. [80]).

So far, integrated urban energy assessments and optimisation mostly focused on the charac-
terisation of space heating, and did not differentiate between the various DHW end-uses like

11
Chapter 2. Domestic hot water characterisation

showering, dish washing, etc. (Jennings et al. [106], Fonseca and Schlueter [77], Nouvel et al.
[150]). Nevertheless, such a differentiation is conducted by Yao and Steemers [242], although
relevant streams (shower and bath) are still aggregated as one stream, and by Aydinalp et al.
[8], Widen et al. [239], Beal et al. [9], where the specific temperature levels of the various
end-uses are not considered specifically. However, the differentiation and characterisation
of DHW streams is of importance for an integrated approach to urban energy assessment
and optimisation. With the distinction of the multiple domestic hot water end-uses, various
energy saving measures can be more specifically addressed in an integrated way at urban
scale (building blocks, streets, districts, city). In particular, various grey water heat recovery
configurations, as described by Schmid [190], McNabola and Shields [137], Dong et al. [49] for
buildings or by Abdel-Aal et al. [1], Elias-Maxil et al. [52], Hepbasli et al. [96] for sewer systems,
requires the characterisation of the specific DHW streams. The characterisation of the various
end-use temperatures also improves the assessment of the heating utility temperature level
of buildings with low temperature space heating systems, as this level is determined by the
DHW demand (Brand et al. [29]). So far, the hot water temperature level has generally been
assumed at 60°C (Perry et al. [166], Girardin et al. [85], Kordana et al. [117]), but with more
specific DHW streams, this level can actually be lowered. This again influences the selection of
the optimal utility configuration, where heat pumps and low temperature waste heat recovery
can become more competitive compared to biomass of fossil fuels. Finally, urban energy
integration focusing on the optimisation of district heating (DH) systems only considered
buildings equipped with hot water storage system (Weber [237], Fazlollahi [72], Elmegaard
et al. [53]), although Thorsen and Kristjansson [214], Christiansen et al. [42], Rosa et al. [184]
showed that a configuration without storage (stand-alone heat exchanger) can be equivalent
or even better in terms of costs and energy efficiency. The modelling of such a configuration,
however, requires the characterisation of the DHW loads according to the main end-uses.

Considering the above-mentioned shortcomings, the objective of this work is to propose a

detailed urban DHW characterisation method differentiating between the various end-uses
encountered in domestic buildings (households) and lodgings (hotels and nursing homes),
and describing them as to their temperature levels. The input data as well as the outcomes are
geoallocated to benefit from the spatial differentiation, which allows the detection of locations
with specific energy consumption, the aggregation and representation of the data according
to building blocks, streets or districts as well as the planning of centralised energy utilities like
district heating [Link] proposed method can therefore be applied, when the required
data (inhabitant and household numbers, building type, etc.) is available, at the level of cities,
districts, streets and building blocks. Based on these DHW models, a complementary method
to calculate the load of DH stand-alone heat exchangers is formulated.

The main contributions of the proposed method are therefore to increase the detail level of
urban thermal energy assessments, as well as to improve integrated urban energy optimisation
with the consideration of DHW temperature for optimal heating utility type selection and an
additional district heating system configuration.

12
 2.2. Method

To this end, the methodology to characterise the energy demand and load requirement of
various DHW end-uses is presented in Section 2.2. The equations for households, hotels and
nursing homes are developed, and a review on European domestic water use is conducted. In
Section 2.3, the proposed characterisation method is applied to the households, hotels and
nursing homes of the city of Esch-sur-Alzette (Grand-Duchy of Luxembourg), where the DHW
energy demands are calculated and compared to calculated or measured values mentioned in
scientific publications and technical literature addressing DHW demand. The energy use of
the various DHW streams are also put in relation to the total and space heating energy demand
in this case study. Temperature level and utility load requirements as well as the impact of the
proposed DHW characterisation on the energy integration of heat pumps are finally addressed
in this section. Considering the outcomes of the case study, Section 2.4 covers the advantages,
shortcomings and main contributions of the proposed method.

2.2 Method

This section describes the proposed method for urban DHW characterisation. In Section
2.2.1, equations for the DHW-related energy consumption as a function of occupant and units
numbers (e.g. a household in multifamily buildings, a room in a hotel), space heating energy
demand as well as the utility load of stand-alone heat exchangers are developed. A review
of European literature on the characterisation of domestic water use, considering typical
flows, use frequency and duration, temperature and occurrence is conducted in Section 2.2.2.
The spatial allocation of the various DHW end-uses according to inhabitant and household
numbers, allowing the spatial differentiation of the results, is also presented in that section.

2.2.1 Characterisation methodology

[Link] DHW end-use load

The thermal power requirement Q̇ e of a DHW end-use e is calculated considering its density
ρ, its heat capacity c p , the water volumetric flow V˙e and the difference between hot and fresh
water temperatures (Te − T f r esh ) (Eq. 2.1). Both volumetric flow and hot water temperature
are specific to the considered end-use e.

Q̇ e = ρ × c p × V̇e × (Te − T f r esh ) (2.1)

[Link] DHW and space heating energy demand

Domestic buildings The daily energy demand Q eD HW of a DHW end-use e is the product of
the thermal power Q̇ eD HW with its use duration d e (when necessary, corrected to the appropri-
ate time unit) and daily use frequency f e (Eq. 2.2).

13
Chapter 2. Domestic hot water characterisation

Q eD HW = Q̇ eD HW × d e × f e (2.2)

In a household, some of the DHW end-uses e are related to the activities of the occupant
o (e.g. showering, bathing, washing and shaving), while other streams (e.g. dish washing)
are directly linked to the household h. The total DHW-related yearly energy demand of the
household Q D HW,househol d , expressed in kWh/a, is therefore obtained by summing the daily
D HW
energy consumption Q e,o,t of the various DHW streams e of occupant o for time t , with the
D HW
energy use Q e,h,t required at household level h for time t (Eq. 2.3).

365 µ ¶
Q D HW,househol d = D HW D HW
X X XX
Q e,h,t + Q e,o,t (2.3)
t =1 e o e

For multifamily buildings, some DHW end-uses (e.g. cleaning of common spaces) are required
at building b level. The yearly DHW energy demand Q D HW,bui l d i ng of a building is therefore
D HW D HW
obtained by summing up the DHW energy demand Q e,o,h,b,t of the occupants, Q e,h,b,t of the
D HW
households and Q e,b,t of the streams attributed to the building common areas over 365 days
(Eq. 2.4).

" ¶#
365 µ
D HW,bui l d i ng D HW D HW D HW
X X X X XX
Q = Q e,b,t + Q e,h,b,t + Q e,o,h,b,t (2.4)
t =1 e h e o e

Lodgings While the equations above can also be applied to hotels and nursing homes (the
room replacing the household), the necessary input data are scarce. An alternative to Eq. 2.4
can be used, considering several parameters related to room and bed occupancies:

• the total room number r ,

• the yearly room occupancy m, expressed in percentage,

• the average bed number per room β,

• the yearly bed place occupancy p, expressed in percentage.

Using Eq. 2.5, the energy demand of the various end-uses e related to the occupant
P D HW,occupant P D HW,bui l d i ng
, room e Q eD HW,r oom and building e Q e are summed up ac-
P
e Qe
cording to the number of occupants (product of bed number per room and bed place occu-
pancy) and the number of occupied rooms (product of number of rooms and room occupancy)
for the whole year.

14
 2.2. Method

X D HW,bui l d i ng
Q D HW,l od g i ng = 365 × [ Q e +
e
µ ¶
X D HW,occupant
r × m × Q eD HW,r oom + β × p × Q e
X
] (2.5)
e e

Space heating energy demand With the characterisation of the DHW energy consumption,
the space heating energy consumption is calculated considering the total fuel consumption
and the efficiency of the heating utility η ut i l i t y (Eq. 2.6).

³ ´
Q SH ,bui l d i ng = Q t ot al × η ut i l i t y −Q D HW,bui l d i ng (2.6)

In case measured data are not available, the total heat consumption can be conveniently
determined using multiple regression analysis out of a sample of measured consumption data.
Multiple linear regression is one of the different techniques available to predict the energy
consumption of buildings based on measured consumption data as well as several other
independent variables, instead of only one as with simple linear regression. This technique is
commonly used by other authors in literature for similar studies (Wahlstrom and Harsman
[233], Schüler et al. [189]). Guerra Santin et al. [88] highlighted that building characteristics de-
termine a large part of the energy use in dwellings. Howard et al. [101] used the total floor area
of each included building type as a predictor of fuel consumption. Mastrucci et al. [135] used
floor surface, and the combination of type of dwelling and period of construction to predict
the natural gas consumption of residential buildings. Compared to other techniques, linear
regression is particularly promising for this goal due to reasonable accuracy and relatively
simple implementation (Fumo and Biswas [83]).

[Link] Utility load calculation

Utility with thermal storage The large majority of heating utilities in buildings cover both
domestic hot water and space heating demands (Schramek [192])), and are usually used in
combination with a hot water storage system. With the space heating energy consumption
and data on outdoor temperature, the space heating load requirements can be obtained using
the heating signature method (Girardin et al. [85]). The decentralised utility load (Eq. 2.7) is
SH
obtained by summing the maximal space heating load Q̇ max and the continuous, averaged
over 8’760 hours, DHW load (Schramek [192], Girardin et al. [85]). It is referred to the work of
Becker and Maréchal [12], Fazlollahi et al. [73, 75] for the optimised design of thermal storage
tanks.

Q D HW,bui l d i ng
Q̇ ut i l i t y,[Link] or ag e = Q̇ max
SH
+ (2.7)
8760

15
Chapter 2. Domestic hot water characterisation

Utility without hot water storage Buildings connected to a district heating network are not
necessarily equipped with a local DHW storage system (Christiansen et al. [42], Rosa et al.
[184], Tol and Svendsen [215]). In literature, several methods to design the heat exchanger
(HE) are proposed: while Gaderer [84], Tol and Svendsen [215] add up both SH and DHW load,
Rosa et al. [184] considers only DHW, and Thorsen and Kristjansson [214] proposes to design
the HE according to the highest load between DHW and SH. Considering building thermal
inertia and the short DHW pulse duration, the approach of Thorsen and Kristjansson [214] is
used for this work (Eq. 2.8).


Q̇ ut i l i t y,D HW i f Q̇ ut i l i t y,SH < Q̇ ut i l i t y,D HW
Q̇ ut i l i t y,no st or ag e = (2.8)
Q̇ ut i l i t y,SH i f Q̇ ut i l i t y,SH > Q̇ ut i l i t y,D HW

To determine the DHW power requirements of one household, the load of the DHW end-use
with the highest value is selected and multiplied by a simultaneity factor S =1.15 (Schramek
[192]). It is considered that the various large end-uses (e.g. bathtub, dish washing and shower-
ing in households) are available once and are not used simultaneously, but that smaller
end-uses can be required at the same time than a large DHW appliance (Eq. 2.9).

Q̇ D HW,househol d = 1.15 × Q̇ max

D HW
(2.9)

For multifamily buildings, mixed-use buildings and lodgings, the DHW thermal power re-
quirement at building level is not obtained by summing up the loads of the single end-uses of
the u units, as not all hot water demands occur at the same time (Thorsen and Kristjansson
[214], Schramek [192]). Instead, in order to avoid an over-sizing of the utility, and thus higher
investment costs, a simultaneity factor S u is considered, which is multiplied by the sum of the
uni t
single largest hot water end-use Q̇ u,max of each unit u (Eq. 2.10).

Q̇ D HW,bui l d i ng = S u × uni t
X
Q̇ u,max (2.10)
u

The simultaneity factor S u is determined according to the number of units u in the building.
S u values are based on empirical data, and several models have been proposed to describe
its behaviour (see Gaderer [84] and Christiansen et al. [42] for comparisons of simultaneity
factor models). The simultaneity factors S u using the equations provided by Thorsen and
Kristjansson [214] and Gaderer [84] are represented in Fig. 2.1.

Considering that the equation of Thorsen and Kristjansson [214] has been specifically designed
for Danish hot water utility design conditions (32.3 kW) and that the results are still overestim-
ated when compared to measured data (Thorsen and Kristjansson [214], Christiansen et al.
[42]), the equation of Gaderer [84] is selected (Eq. 2.11).

16
 2.2. Method

Figure 2.1: Simultaneity factors (Thorsen and Kristjansson [214], Gaderer [84])

S u = 0.02 + 0.92u (−0.58) (2.11)

2.2.2 Input data

In order to apply the proposed method, the following input data are required:

• End-use temperature,

• Volumetric flow for load calculation,

• Use frequency and duration,

• Geographical allocation of end-uses, occupants and unit numbers.

[Link] Temperatures

f r esh
The inlet temperature of fresh water Ti n varies over the seasons between 5°C and 15°C, but
is, for yearly assessments, generally assumed at 10°C (Spur et al. [197], Widen et al. [239]). End-
use temperature data are generally scarce. Yao and Steemers [242] and Schramek [192] give
identical end-use temperatures for bath and shower, hand washing and hand dish washing
(Tab. 2.1), while Wong et al. [240] mentions an average temperature of 40.9°C for showers.

17
Chapter 2. Domestic hot water characterisation

Table 2.1: DHW end-use temperatures (Yao and Steemers [242], Schramek [192])

End-use type [-] Temperature [°C]

Hand washing 35
Washing and shaving 35
Hand dish washing 55
Showering 40
Bath filling 40

[Link] Volumetric flows

Volumetric flow data for European DHW end-uses are summarised for domestic and lodging
buildings in Tab. 2.2 and Tab. 2.3, respectively. Widen et al. [239] provides water consumption
of 25 different DHW streams, of which three, indicated in Tab. 2.2, have a particularly high
energy consumption. For non-European data, it is referred to DeOreo et al. [47], Hendron
and Burch [95], Neunteufel et al. [147], Beal et al. [10], Cahill et al. [37], Kenway et al. [110],
Rathnayaka et al. [181], Makki et al. [132].

Concerning household DHW streams, a value around 0.13-0.14 l/s for showers appears to
be common. For bathtubs, the volumetric flow depends of the considered volume. The
volumetric flow of kitchen sinks revolves around 0.1 l/s, although lower and higher values are
reported. Bathroom sink volumetric flow data show the largest variance, with values between
0.03 and 0.1 l/s.

[Link] Use duration and frequency

While data on volumetric flow of DHW are available in the literature, European data on use
frequency and duration (Tab. 2.4 and Tab. 2.5, respectively) are scarce. Blokker et al. [22]
mentions a total kitchen tap use frequency of 12.6 (household day)-1 , which is subdivided
between hand washing and hand dish washing at one fourth each, according to the penetration
rate mentioned in the publication (the remaining 50% water use is for drinking, cooking, etc.).
The tap use frequency of Neunteufel et al. [147] is distributed evenly between kitchen and
bathroom sinks. Data on bathtub and shower use frequencies in non-European countries are
given by Hokoi et al. [100], Kenway et al. [110].

[Link] Streams occurrence and spatial allocation

The occurrence of a dish washing machine or a bathtub in a household or a room is not

automatically given (Blokker et al. [22]). As DHW end-use occurrence data are unlikely to
be available at building level when conducting an urban energy assessment, it is proposed

18
 Table 2.2: Domestic DHW volumetric flows, in l/s

Reference Koiv and Thorsen and Schramek Widen Blokker et al. Neunteufel Neunteufel Gutierrez-
Toode Kristjansson [192] et al. [239] [22] et al. [147] et al. [147] Escolar et al.
[115] [214] [90]

Pub. year 2006 2006 2007 2009 2010 2012 2012 2014
Country Estonia Denmark Germany Sweden Netherlands Austria EU Spain
a f
Bathroom - - 0.05 / 0.083 - 0.04 0.03 - 0.07 *, 0.1
sink
Kitchen sink 0.2 0.10 0.10 / 0.17 b 39 l d 0.08 / 0.13 g 0.03 - 0.10 *, 0.13
Shower 0.2 0.14 0.14 0.13 / 0.20 0.12 *, 0.14 0.13 0.13 0.17 *, 0.25
/ 0.58 e
Bathtub 0.3 0.21 0.11 / 0.17 / 100 l/bath 0.20 h 76 l/bath 150 l/bath 250 l/bath
0.21 c
a
small /large sink, b single / double sink, c size: 100 / 160 / 180, d dish washing, mix of tub and running water, e assumed value / modern tap / trad. tap,f washing and shaving,
g
hand washing / dish washing, h capacity: 120 l, * water-saving end-use
2.2. Method

19
Chapter 2. Domestic hot water characterisation

Table 2.3: Lodging DHW volumetric flows, in l/s

Reference Cobacho et al. [44] Blokker et al. [23]

Pub. year 2005 2011

Country Spain Netherlands
Building type Hotel Hotel, nursing home,
Bathroom sink 15.26 l/day*guest 0.08 a
Showering 13.03 l/day*guest 0.12 / 0.14 / 0.37 b
Bath filling - 0.20 c
a
hand washing, washing and shaving,b water-saving/normal/comfort,c capacity: 120 l

Table 2.4: DHW use frequencies, in capita*day-1

Reference Cobacho et al. Blokker et al. Blokker et al. Neunteufel Neunteufel

[44] [22] [23] et al. [147] et al. [147]
(EU)

Building Hotel Household Offices, hotel, Household Household

type nursing home,
restaurant
Bathroom 5.36 1.35 a 4.50 / 1.98 c 11.50 -
sink /day*guest
Kitchen - 3.150 b - 11.50 -
sink
Showering 0.45 0.70 0.20 /0.80 d 0.70 0.8
/day*guest
Bath filling - 0.04 0.20 e 0.03 0.1
a b c
washing and shaving, hand and dish washing, per household, hand washing / washing and shaving,
d
nursing homes / hotels,e capacity: 120 l

to distribute the occurrence according to the socio-economic level of the household or the
district, depending of the data available. Geographically Weighted Regression would strongly
support the spatial allocation, as it allows the disaggregation of the occurrence into the specific
spatial unit while considering socio-economic predictors. However, this approach requires
data as to the occurrence of the end-uses according to the socio-economic level, as well as
the spatial distribution of the various levels, which is not always given locally. As simplified
alternative, the end-use occurrences can be randomly distributed geographically as a function
of the appliance occurrences o e and by attributing random values εe between 0 and 1 to each
end-uses in the household, e.g. using the RANDOM function in PostgreSQL [175]. The end-use
is considered installed for εe < o e (Eq. 2.12).

20
 2.2. Method

Table 2.5: DHW use durations, in s

Reference Cobacho Schramek Blokker Blokker et al. Neunteufel Neunteufel

et al. [44] [192] et al. [22] [23] et al. et al.
[147] [147]
(EU)

Building Hotel Household Household Hotel, nursing Household Household

type home
Bathroom - 90 / 120 a 40 c 16 / 40 e 59 -
sink
Kitchen - 300 15 / 48 d - 59 -
sink
Showering 270 360 510 510 288 474
Bath - 900 / 1200 b 600 600 - -
filling
a
small / large sink,b size: 100 & 160 / 180,c washing and shaving,d hand and dish washing, e hand washing /
washing and shaving


Q̇ i f εe < o e
e
Q̇ eD HW = (2.12)
0 i f εe > o e

Values on the occurrence of dishwasher odishwasher in households can be obtained from the
respective national statistic agency (Tab. 2.6). Hand dish washing DHW demand is excluded
for households equipped with a dishwasher.

Table 2.6: Dishwasher occurrence rates

Country Occurrence Year of survey Source [-]

[%] [year]

Austria 74 2012 Statistik Austria [205]

France 56 2014 INSEE [104]
Germany 68 2011 Statistisches Bundesamt [206]
Luxembourg 79 2011 Statec [200]
Switzerland 85 2014 Morgenthaler et al. [144]

Concerning the occurrence o bat h of household bathtubs, data is rare. 97.5% of the EU27
households in 2013 were equipped with a shower or a bathtub (Eurostat [65]) but a further
breakdown between these two end-uses is almost non-existing in European and national
household or production statistics. The French national statistic agency indicates that 74.4
% of the households were equipped with a bathtub in 2006, and 24.0% with only a shower

21
Chapter 2. Domestic hot water characterisation

(INSEE [103]). A study on the water consumption in Austrian households showed that at least
25% of the 24 assessed households only used a shower (Neunteufel et al. [147]). Blokker et al.
[22] indicate a dissemination rate of 36% for household bathtubs in the Netherlands. A survey
conducted in China showed that between 20 to 100% of the households in Nanjing and around
70% of the Hefei households were only equipped with a shower (Hokoi et al. [100]).

Information on DHW appliance of non-residential buildings are also very scarce (Pieterse-
Quirijns et al. [171], Blokker et al. [24]). Data on lodging bathtubs availability can be obtained
from the site manager, from the site homepage or reservation service homepage. While
Pieterse-Quirijns et al. [171] considers both showers and bathtubs available in Dutch hotel
rooms, the number of hotel bathtubs in the USA has strongly decreased with newer construc-
tions. In 2001, 95% of the Holiday Inn hotels were equipped with bathtubs, but hotels have
recently been built either without (Indigo hotel) or only with a much smaller ratio varying
between 25% (Marriott) to 55% (Holiday Inn) (Jones [108]).

[Link] Occupant and unit spatial allocations

Inhabitant and household numbers can be obtained from the municipality, preferably as a
Geographical Information System (GIS) data set or as a table including the address of the
building. While the spatial allocation is not strictly necessary for the proposed assessment at
city scale, it allows a spatial differentiation of the DHW characterisation data, e.g. useful for
the planning of district heating systems, or to present the outcomes at the level of building
blocks, streets , districts, etc.

For lodgings, the best option to obtain the required data is to contact the facility manager,
a solution which is nevertheless time-consuming and yields only a limited number of re-
turns (Neunteufel et al. [147]). Should the detailed data not be available, public sources and
statistical averages can be used to estimate the energy consumption. For hotels, national or
regional tourism offices as well as reservation service homepage can provide the number of
rooms r (e.g. Vienna Tourist Board [230], ONT [156], Switzerland Tourism [211]). Eurostat
provides national and regional data on bedroom and bed place numbers, so that a specific
bed number per room β (Eurostat [67]) as well as the net bed places occupancy p (Eurostat
[66]) are available. Nursing home room and bed numbers can be obtained from dedicated
internet sites (e.g. Bundesministerium fur Arbeit, Soziales und Kosumentenschutz [36], Cap-
geris [38], Privatinstitut für Transparenz im Gesundheitswesen GmbH [177], Haederli et al.
[92], Luxsenior [130]). Occupancy rate in Europe is very high, with values in Luxembourg, Italy
and France reaching 95.4% (Statec [201]), 93 - 98% and 98%, respectively (Evans et al. [70]).

2.3 Case study

The proposed DHW characterisation method is applied to the residential buildings of the
city of Esch-sur-Alzette (Esch), comprising the domestic buildings, the five hotels and the

22
 2.3. Case study

two nursing homes. The city has a population of 33’487 inhabitants and 14’321 households
(Service des travaux municipaux [194]), distributed over 18 districts (Fig. 2.2). The main data
source is the city GIS data set (status: June 2015), imported into a PostgreSQL geospatial
database (PostgreSQL [175]) using the PostGIS extension (PostGIS Project [174]). The data set
contains data on inhabitant and household numbers, year of construction, building footprint
and floor number.

In order to assess the impact of DHW energy consumption on low energy and passive buildings
(not included in the city assessment due to insufficient data on recent constructions from
Statec [202]), these building types are considered specifically. An average inhabitant number
for single family buildings (SFB) and multifamily buildings (MFB) of, respectively, 2.98 and 12
as well as an average household number of 1 and 5.48 is used.

The calculations deployed for the case study are described in Section 2.3.1. The results are
presented in Section 2.3.2.

Figure 2.2: Districts of Esch-sur-Alzette

2.3.1 Calculations

In Section [Link], the calculation of the various DHW energy demands of the households and
lodgings of Esch is set up with the proposed characterisation method. In order to validate
the outcomes of the method, DHW energy demand using a surface-related approach is also
used. The method to quantify this surface is included in this section. To put the impact of
the various DHW streams in relation to the urban heat demand, the total heat consumption

23
Chapter 2. Domestic hot water characterisation

using linear regression is described in Section [Link]. Section [Link] describes how space
heating energy requirements are obtained from total heat and DHW energy consumption. The
calculation of space heating and utility load of stand-alone heat exchangers in district heating
systems are formulated in Section [Link]. Finally, heat pump energy integration according to
various hot water temperature levels is addressed in Section [Link].

[Link] DHW energy consumption

DHW occupant-related method The end-use types, volumetric flows V̇e , use frequency f e
and duration d e of Blokker et al. [22] are selected for the present case study, as the data are
relatively recent and similar DHW use between the Netherlands and Luxembourg can be
assumed as both countries are geographically close and have similar living standards. The
end-use types retained are also those generally mentioned in DHW-related publications.

A distribution between efficient and normal shower of 50% is considered. Energy consumption
related to shower, bathtub (when available) and bathroom sink use is calculated as a function
of inhabitant number. Kitchen sink use (hand washing and hand dish washing) is related
to the household number in the building, as indicated by Blokker et al. [22]. A dishwasher
occurrence rate of 79% is used (Statec [200]), with the assumption that this type of equipement
has its own heating system. 21% of the households are thus doing their dish washing by hand.
The French bathtub occurrence rate of 74.4% is applied here. Dishwashers and bathtubs are
geoallocated using the RANDOM function in PostgreSQL [175], as more detailed information
are not available. The resulting percentage of dishwasher and bathtub occurrence in Esch
reaches 78.8% for the dishwasher, and 74.3% for the bathtubs. For the theoretical low energy
and passive single family and multifamily buildings , it is assumed that they are equipped with
a dishwasher and bathtub.

In hotels and nursing homes, the DHW streams related to the rooms - shower, bathroom
sink (washing and shaving) and bathtubs (when occurring) - are considered. The volumetric
flows V̇e , use frequency f e and duration d e values of Blokker et al. [23] are used. A normal
shower type having a volumetric flow of 0.14 l/s is considered. While in hotel five the rooms
are equipped with a bathtub and hotels two and three are not, detailed information for hotel
one and four are not available. An occurrence rate of 50% is therefore assumed. Eurostat [67]
indicates for the Grand-Duchy of Luxembourg, in 2014 for hotels and similar accommodations,
a bed-per-room number of 1.9. The number of hotel customers is calculated considering a
yearly average bed place occupancy of 35,1% (Eurostat [66]). The year of construction of hotel
two is not given in the GIS data set, but as it is a very recent building, it is set to 2012. Nursing
homes are only equipped with showers. The nursing homes are also equipped with small
kitchen units, but it is not expected that the inhabitants do manual dish washing. The number
of occupants is taken from the GIS data set. The values for the various sites are summarised in
Tab. 2.7.

24
 2.3. Case study

Table 2.7: Hotels and nursing homes data

Building Number of Daily customer- Bathtub Average room

rooms [-] /patients number [-] surface [m2 ]
[-]

Hotel 1 23 15.34 12 18.3

Hotel 2 110 73.36 0 18.0
Hotel 3 22 14.67 0 18.3*
Hotel 4 15 10.00 15 35.0
Hotel 5 20 13.34 10 27.8
Nursing home 1 168 157.00 0 30.5
Nursing home 2 46 32.00 0 30.5*
*
assumption

DHW surface-related method and surface quantification In order to compare the out-
comes of the proposed DHW occupant-related model, the DHW energy consumption accord-
ing to the surface is calculated. These values are generally based on hot water use assessments,
where surface data is available Fuentes et al. [81]. For this case study, specific energy con-
sumption values for domestic hot water production of 13.9 and 20.8 kWh/m2 a for single and
multifamily buildings are assumed (Luxemburgish Parliament [128]). The latter value is also
used for mixed-use buildings (MUB). For lodgings, values of 88 kWh/m2 for nursing homes
and 153 kWh/m2 for hotels are used (Luxemburgish Parliament [126]).

The household floor surfaces of domestic use buildings are obtained by multiplying the
footprint area, obtained from the ST_AREA function of PostGIS (PostGIS Project [174]), with
the building floor number (including the attic). This method was selected as its outcomes
fits best with average household surface values for Esch mentioned in national population
census data (Statec [199]). The average surface of single family and multifamily buildings in
Esch, 165.65 m2 and 511.72 m2 respectively, are considered too for the low energy and passive
buildings. For mixed-use buildings and lodgings, this method is not specific enough, as
surfaces not related to domestic use would also be included. The reference surface is therefore
calculated using the average household surface, obtained from the national population survey
(Statec [199]) for Esch, multiplied by the household number (Tab. 2.8).

For hotels and nursing homes, average room surfaces and room numbers are used to calculate
the relevant area (Tab. 2.7). Data for hotels are obtained from tourism sites: ONT [156] for
room numbers and [Link] [27] for the calculation of average room surface. As surface
data for hotel 3 are not available, the value of hotel 1 is assumed, because both hotels are
similar in size and type. Values between 22 to 32 m2 for 114 rooms and 34 to 42 m2 for 54
rooms are mentioned for nursing home 1 (Servior [195]), leading to an average room surface
of 30.5 m2 . This value is equally used for nursing home 2, as no further data are available. The

25
Chapter 2. Domestic hot water characterisation

Table 2.8: Average household surface for mixed-use buildings in Esch (Statec [199])

Building type, Statec

Building type, GIS typology Households Average
typology
per building surface [m2 ]

1 136.55
Collective building,
Building, mixed usage
mixed usage 2-4 66.59
>4 75.52
1 74.88
Home for adults Hosting structure
2-4 25.00
>4 85.00
1 108.71
Collective building, for
Student home
living purpose 2-4 79.67
>4 71.81
Public usage, industry 1 93.33
Other dwelling building, commerce or
2-4 64.80
service industry
>4 54.13

number of rooms for this type of building is taken from the GIS data set of the municipality.

[Link] Total heat demand using linear regression

A multiple linear regression model is developed to estimate the energy consumption of resid-
ential buildings based on measured consumption data. Household budget survey data are
obtained from the national statistic agency STATEC (Statec [202]). The model is implemen-
ted in the software R (R Core Team [179]) and fitted using the Ordinary Least Squares (OLS)
method.

Data from the year 2011 are selected to fit the model due to meteorological condition similar
to the average of the region. Only dwellings having natural gas or fuel oil as main fuel are
selected. The energy consumption in kWh/a is obtained by multiplying the amount of fuel
consumed (m3 of gas and litres of oil) by suitable calorific values. A distribution of heating
systems (traditional and condensing) is assumed based on statistical data to obtain an average
calorific value based on national values (Luxemburgish Parliament [128]).

The final sample of observations used to fit the model consists of 794 records (of a total of
1’142) and is obtained by excluding the following items from the original data set: records with
missing values; use of other fuels than gas and oil; presence of solar panels; not realistic ratio
between energy expenditure and energy consumption, index of errors in the compilation of
the survey.

26
 2.3. Case study

The formulation of the outcome of the linear regression analysis is given by the following
equation:

5
X
l n(y) = β0 +l n(x f l oor )·β f l oor +l n(x occ )·βocc +x t y pe ·βt y pe + (x per i od ,i ·βper i od ,i )+ε (2.13)
i =1

where y represents the energy consumption of the dwelling in kWh/a, x f l oor the useful floor
surface in m2 , x occ the number of occupants, x t y pe the type of building (single family building
= 1, multifamily building = 0), x per i od ,i the construction period of buildings (factorial variables)
and ε the random error term. Some of the variables are logarithmically transformed as they
present right skewness and other authors showed how this transformation can substantially
improve the performance of the method (Kolter and Ferreira [116]).

Results of multiple linear regression are reported in Tab. 2.9. The coefficient of determination
R 2 shows that 54.1 % of the variance is taken into account by the model and it is comparable
with the ones obtained by similar studies (Guerra Santin et al. [88]: 42%). The model as-
sumptions were carefully verified and not significant heteroskedasticity and multi-collinearity
problems were detected. The equation for multifamily buildings is also applied to mixed-use
buildings.

Table 2.9: Statistical summary of the multiple linear regression for residential buildings

Coefficients Estimate Std. Error t value Significance

Intercept 6.809 0.206 33.080 <0.0001 ***

Floor surface (ln) 0.568 0.047 12.074 <0.0001 ***
Number of 0.106 0.030 3.555 0.0004 ***
occupants (ln)
Type 1: single 0.297 0.044 6.675 <0.0001 ***
family house
Period 1: <1919 0.220 0.068 3.235 0.00127 **
Period 2: 1919-45 0.148 0.054 2.768 0.0058 **
Period 3: 1946-60 0.159 0.048 3.278 0.0011 **
Period 4: 1961-80 0.204 0.041 4.979 <0.0001 ***
Period 5: 1981-95 0.149 0.044 3.388 0.0007 ***
Signific. codes: *** < 0.001, ** < 0.01, * < 0.05, . < 0.1, R² = 0.541, Adjusted R² = 0.536
Residual standard error: 0.404 on 785 degrees of freedom, p − v al ue < 2.2 · 10−16
Notes: Variables "T2: Multi-family house" and "P6: >1995" assumed as reference.
Coefficients marked with (ln) have been logarithmically transformed.

27
Chapter 2. Domestic hot water characterisation

[Link] Space heating energy consumption

The space heating energy consumption is calculated by subtracting the DHW energy con-
sumption from the total heat consumption, considering an utility efficiency η ut i l i t y of 90%.
For low energy and passive single family and multifamily buildings, specific SH values of 43/22
kWh/m2 and 27/14 kWh/m2 are used (Luxemburgish Parliament [128]). As the equations are
not applicable to hotels or nursing homes due to difference in user behaviour, the total and
space heating energy consumption is only calculated for domestic buildings.

[Link] Heat exchanger load for district heating systems

The SH load of domestic buildings is calculated using the heating signature calculation (Gir-
ardin et al. [85]). As input data, the outdoor temperature Tout d oor behaviour as a function of
hour t , based on data stemming from the air transport service in Luxembourg (Administration
de la navigation aérienne [2]) and structured for the period of Septembre 2011 to August
2011 (to represent the cold season as one continuous curve), is modelled for this case study
using polynomial regression analysis, as it yielded the highest R 2 (Eq. 2.14). t = 1 thus repres-
ents midnight on the first of Septembre, while t = 8760 is for the 31st of August at 23:[Link]
coefficient of determination R 2 reaches 64%.

Tout d oor = (2 ∗ 10−17 )t 5 − (4 ∗ 10−13 )t 4 + (3 ∗ 10−9 )t 3 − (8 ∗ 10−6 )t 2 + 0, 0003t + 16, 771 (2.14)

A minimum outdoor temperature of -10°C and a threshold temperature of 15°C are considered
for the use of the heating signature and therefore the calculation of the space heating load.
Equation 2.14 is integrated between the time intervals of 550 and 5’940 hours, where the
outdoor temperature is below and above 15°C, respectively. The space heating load of each
building is finally obtained by multiplying its reference area with the specific SH load obtained
from the heating signature.

With the selected DHW household streams, three types of end-uses are to be considered for
the maximal DHW load: bathtub, hand dish washing or shower. The determination of the
relevant end-use depends of the occurrence of dishwasher and bathtub in the household.
In case a bathtub is available, a load of 25.08 kW is retained. If a dishwasher is not installed,
a maximum DHW load of 24.45 kW is considered. Else, the load of the shower is prevailing
with 16.30 kW. Following Eq. 2.8 as well as Eq. 2.9 or Eq. 2.10 for households and multifamily
buildings, respectively, the load of stand-alone heat exchangers is finally obtained.

In order to compare the results, Eq. 2.7 is applied to calculate the load of an utility combined
with a hot water storage system.

28
 2.3. Case study

[Link] Energy integration of decentralised heat pump considering various hot water tem-
perature levels

Former works on integrated urban energy optimisation have so far modelled DHW demand
as hot water stream at 60°C (Perry et al. [166], Weber [237], Girardin et al. [85], Varbanov et al.
[226], Fazlollahi et al. [74]). However, as stated by Girardin et al. [85], various heating utilities
providing heat at close space heating and domestic hot water temperatures, like heat pumps
and low-grade waste heat recovery, are particularly sensitive to temperature level requirements.
Therefore, the impact of the proposed DHW models on the integration of heat pump as heating
utility is assessed as example, using the energy integration approach described by Weber [237].

The above-mentioned low energy single family building, equipped with bathtub and dish-
washer, is selected as low-temperature SH case study. The space heating load of the building
is calculated using the heating signature for the coldest day of the year, with a floor heating
system with supply and return temperatures of 30/25°C (Hesaraki et al. [97]). For the first
case, a hot water production at 60°C in the storage tank is considered, the load being calcu-
lated by dividing the annual DHW energy consumption by 8670 hours. In the second case, a
temperature of 50°C is used, as it has been showed that for new or renovated buildings with
DHW system volumes below 3 litres and individual DHW feeding pipes, this level is sufficient
to avoid Legionella proliferation (Brand et al. [29]). The temperature difference between the
condensing side of the heat pump and the hot water is maximum 5K. The evaporation side has
a temperature of 5°C (5K below the average of 10°C outdoor temperature), and the electricity
consumption is calculated considering a Carnot factor of 55% (Becker et al. [13]). Finally, a
price of 0.141 e/kWh is used to calculate the operating costs (Enovos [55]).

2.3.2 Results

As measured DHW energy consumption data are not available to validate the proposed meth-
odology, the results obtained are compared, in Section [Link], to values indicated in scientific
publications and technical literature addressing DHW. In Section [Link] and [Link], the DHW-
related energy demand is assessed as to the main end-uses and temperature levels. The heat
exchanger loads, considered with and without hot water storage, are displayed at district level
in Section [Link]. Finally, the outcomes of the energy integration of heat pumps at different
hot water temperatures are presented in Section [Link].

[Link] Validation of outcomes with literature values

The total yearly heat demand of the households in Esch amounts to 189.2 GWh (Fig. 2.3 and
Fig. 2.4).

The household DHW energy demand of the proposed methodology is, at city level, 17 % lower
than that of the surface-related model (Fig. 2.4). The districts Dellheicht (5), Schlassgoart

29
Chapter 2. Domestic hot water characterisation

Figure 2.3: Total household yearly heat consumption at district level

(13) and Universiteit (16) have particularly high negative differences, with 48-54% less energy
demand. On the other hand, the districts Belval (2) and Raemerich (12) show the opposite
behaviour, with the proposed methodology generating energy use values 13-21% higher than
the surface-related method. Part of these differences is due to the non-consideration of the
heat losses in the proposed characterisation method. Another reason for these differences
is probably also due to the reference surface used for the calculations. While using GIS data
provides specific information as to floor number and surface per building, areas not relevant
to domestic hot water demand (e.g. garages, staircases) are also included, thus leading to an
overestimation of the relevant surface and therefore of the energy demand. As represented with
the light blue bars, the higher the surface per inhabitant, the higher the difference between
the two DHW models. In case the average surface per inhabitant is particularly low, the
occupant-related method obtains higher energy consumption values than the surface-related
model.

Concerning the hotels and nursing homes, the difference between the two methods is even
more important, with the detailed characterisation method reaching only 20% and 9% of the
surface-related energy consumption, respectively (Fig. 2.5). Part of this difference can be
explained by the fact that the specific DHW energy consumption values are generic and might
include additional streams (e.g. room cleaning, bathtub) that are not considered with the
proposed model. In addition, an occupancy rate of 35.1% is considered for the hotels, while
the surface-related approach assumes an occupancy rate of 100% of the surface.

Considering a hot water temperature of 60°C, a hot water consumption of 33.1 l/capita*day is

30
 2.3. Case study

Figure 2.4: DHW-related energy demand and surface per inhabitant (blue column) of house-
holds at district level

Figure 2.5: DHW energy demand of lodgings

obtained using the proposed DHW streams. Maas et al. [131], Schramek [192] and Girardin
et al. [85] mention values of 40 l/capita*day (Luxembourg), 30-60 l/capita*day (Germany) and
50-70 l/capita*day (Switzerland), respectively.

Single family (SFB in Fig. 2.6), multifamily (MFB) and mixed-use (MUB) buildings have a

31
Chapter 2. Domestic hot water characterisation

specific energy demand between 124-172 kWh/m2 , 60-122 kWh/m2 and 80-123 kWh/m2 ,
respectively. The outcomes for single family buildings build after 1995 confirms the findings
of Maas et al. [131], who obtained an average energy consumption of 131 kWh/m2 over a
sample of 54 buildings built between 1997 and 2007. However, the generated values are lower
than those measured by Merzkirch et al. [140], who mentions values of 170 kWh/m2 and
120 kWh/m2 for single and multifamily buildings built after 1994 on. This difference is most
probably due to the uncertainties of the regression analysis results.

Fuel consumption for heating is, as expected, mostly related to space heating (Fig. 2.6), and
decreases for more recent buildings. Concerning existing single family buildings, the fuel
conversion losses of the heating utility are higher than the fuel consumption related to DHW.
For multifamily and mixed-use buildings, DHW fuel consumption is almost twice as high than
the utility losses.

Figure 2.6: Specific thermal energy demand of households (SFB - single family building, MFB -
multifamily building, MUB: mixed-use building)

The contribution of the DHW streams to the total heat demand (including utility losses),
amounts at city level to 12.3%, while it varies between 6-9% (single family building), 17-23%
(multifamily building) and 16-24% (mixed-use) for the considered periods of construction
(Fig. 2.7). The relative DHW contribution at city scale should only be compared with other
urban systems of very similar number and types of buildings, as this contribution varies
according to these parameters. However, the order of magnitude is coherent with the European
average of 15.9% (Enerdata [54]). Moreover, Frijns et al. [80] mention an average value of 23%
for the Netherlands, and Tooke et al. [216] 22% for Canada. The relevance of DHW energy
consumption is 20% and 33% for low-energy and passive SFB and 34% and 49% for MFB. Esch
single family and multifamily buildings (mixed usage buildings included) have a specific DHW

32
 2.3. Case study

energy consumption of 11.75 kWh/m2 and 17.33 kWh/m2 , while values of 13.9 kWh/m2 and
20.8 kWh/m2 are considered at national level (Luxemburgish Parliament [128]).

Figure 2.7: DHW energy demand contribution in households (SFB - single family building,
MFB - multifamily building, MUB: mixed-use building)

[Link] Contribution of DHW streams to household energy consumption

In terms of energy consumption, showering represents by far the most relevant DHW stream,
which is confirmed by Elias-Maxil et al. [52], contributing between 5-8% (SFB), 14-18% (MFB)
and 13-19% (MUB) to the total heat consumption of the buildings (Fig. 2.7). For low energy
and passive buildings, this value reaches between 17 and 28 % for single family and between
29 and 41% for multifamily buildings. The bathtub makes up between 2 to 5% of the total heat
consumption, and the other streams around 1%.

Related to the total DHW energy consumption, showers contribute between 80 and 84% (Fig.
2.8). The energy consumption of the other DHW streams is comparatively small, with the
bathtub contributing to 7%, the manual dish washing between 3 to 7%, the bathroom sink to
3% and hand washing to 2 to 3%. In low energy and passive buildings, shower makes up to
85% of the DHW energy consumption, bathtub 10%, with the remaining streams totalling 5%.

[Link] DHW temperature level requirements

Considering temperature levels, 0-1% of the total heat consumption is used for domestic
hot water use at 35°C (all types of buildings), while 40°C streams represent 5-8%, 15-20%
and 14-21% of the heat consumption of single family, multifamily and mixed-use buildings.

33
Chapter 2. Domestic hot water characterisation

Figure 2.8: End-use contribution to household DHW energy demand (SFB - single family
building, MFB - multifamily building, MUB: mixed-use building)

DHW demand at 55°C (hand dish washing) lies between 0-2% of the total household heat
consumption. For low energy and passive buildings equipped with dishwashers, 1-3 % of
the DHW energy consumption is related to 35°C streams, while 19 to 46% of the energy
consumption is for 40°C domestic hot water.

Set in relation to the DHW energy consumption, dish washing at 55°C makes between 5 to
7 % of the energy use (Fig. 2.9). 93% of the DHW energy consumption is therefore related to
streams at or below 40°C, with 35°C end-use streams (hand wash and washing and shaving)
representing around 5%. In low energy and passive buildings, 35°C streams amounts to
approximately 6% of DHW energy consumption, while, with the assumption that dishwasher
are installed and therefore no 55°C hot water is required, the rest is used for 40°C streams.

[Link] Heat exchanger load for district heating systems

With the characterisation of the various DHW streams, the load of DH stand-alone heat
exchangers of 99% of the single family and 100% of the multifamily and mixed-use buildings
is designed according to the DHW demand. Due to the high occurrence of dishwashers
and bathtubs, 59% of the utility load of households are sized according to the bathtub load
requirement. The stand-alone heat exchanger load of buildings, aggregated at district level
(Fig. 2.10a), is between 1.9 and 2.9 times higher compared to a configuration with a hot water
storage system (Fig. 2.10b).

34
 2.3. Case study

Figure 2.9: Temperature level contribution to household DHW energy demand (SFB - single
family building, MFB - multifamily building, MUB: mixed-use building)

(a) Stand-alone heat exchanger (b) Heat exchanger with hot water storage

Figure 2.10: Utility heating load at district level

[Link] Energy integration of decentralised heat pump considering various hot water tem-
perature levels

The results of the energy integration of the heat pump in the low energy single family building
is represented as cold (blue line) and hot (red line) composite curves, with the former repres-
enting the heating requirements of the building, and the latter the heat pump heating load. As
the curves are based on pinch analysis theory (Linnhoff et al. [124]), they are both shifted by

35
Chapter 2. Domestic hot water characterisation

half of the assumed minimum temperature difference of 5 K, which explains e.g. the heating
requirements temperature of 62.5°C in the figure on the left.

Figure 2.11a represents the configuration with 60°C hot water production. Considering an
evaporation temperature of 10°C, the heat pump has a temperature lift of 50 K, which leads to
a Coefficient Of Power (COP ) of 3.10 and electricity costs of 419 e/year. By reducing the hot
water temperature level (Fig. 2.11b) to the minimum required by space heating, DHW end-use
temperature levels and hygienic constraints, the COP of the heat pump is increased by 14 % to
a value of 3.61, reducing the costs to 359 e/year.

(a) Hot water at 60°C (b) Hot water at 50°C

Figure 2.11: Composite curves of a low energy, single family building

2.4 Discussion

A method to characterise and assess DHW streams at city level, down to districts, streets or
building blocks, is presented in this work. Although measured data on DHW energy consump-
tion are not available to validate the outcomes of the case study and, therefore, the proposed
characterisation method, the related energy demand results are nevertheless confirmed by
previous works focusing on DHW urban energy consumption.

The lower DHW energy consumption values of the proposed method, compared to the surface-
related method, and the lower specific energy consumption per surface, compared to legisla-
tion values, actually highlight an advantage of the exposed method. The observed differences
of the case study are most probably due to the inaccuracies in surface estimations, generated
with geographical data and therefore including non-heated surfaces (e.g. garages) or actually
non-existing areas for buildings with different levels. As an alternative to these estimation,
the proposed characterisation approach instead relies on occupant and household or room
numbers, data that are usually available from the municipality or other sources. In addition,

36
 2.4. Discussion

as particularly highlighted in the case of hotels in Section [Link], using building occupant
number instead of surface data as input parameter also allows to reflect the actual level of
occupancy of a building. An allocation of DHW energy use to empty or less-occupied build-
ings is therefore avoided. By relating the DHW consumption to the occupants, the proposed
method generates more accurate results, in particular when surface data can only be roughly
estimated. The outcomes of space heating energy demand, when calculated at district, street
or building block levels, is therefore more precise. Due to the use of a randomising function
for the spatial allocation of various end-uses, the proposed method should not be applied at
building level. However, this is actually not an issue, as the specific characterisation of the
actual DHW end-uses can then be addressed in detail anyway.

One drawback of the proposed DHW characterisation method is the limited input data avail-
ability. DHW data specific to the country should be considered, as water use differs according
to geographical location and living standards. But information on volumetric or mass flow,
use duration, use frequency and end-use occurrence related to the socio-economic level are
generally scarce, in particular for non-domestic buildings. Specific data for other types of
buildings with relevant DHW demand (e.g. hospitals, sport facilities) are not even available
or sufficiently detailed out. This data scarcity also hinders the use of certain methods like
Geographically Weighted Regression, which allows a good spatial distribution of end-use
occurrence according to living standards, as well as a more precise validation of the proposed
characterisation method. Furthermore, the heat losses occuring between the heating utility
and the various end-uses are currently not taken into account. An additional weak point is the
use of national statistical values to calculate unit occupancy of lodgings, as an equal distri-
bution of customers across the country is assumed. Moreover, the detailed characterisation
of the DHW streams implies the risk that some end-uses might be neglected, thus distorting
the DHW and SH energy demand outcomes. Finally, the actual user water consumption
behaviour, which differs according to age and occupation during the day (Blokker et al. [22]),
is not reflected, as this work focuses on yearly assessments.

Further DHW data, covering additional building types, are therefore necessary for a complete
application of the proposed characterisation method. Specific DHW-related energy data
would also allow a more precise assessment of the validity of the presented work. It is also
proposed to include regional or city-level information on hotels and nursing homes directly
in the input data set. To avoid the risk of omitting relevant streams, users are encouraged to
at least consider the four end-uses described in Section 2.2.2, which are commonly referred
to in DHW-related publications. Furthermore, with the proposed detailed characterisation
approach, the relevance of heat losses, especially considering large buildings with important
distribution lenthgs, should be included to improve the accuracy of the outcomes. A validation
of the aggregated DHW demand results with national values is also recommended. Daily user
behaviour patterns have already been addressed in former publications and can be referred to
in order to conduct assessments with smaller time scale. Globally, as household DHW energy
consumption is due to more than 80% to showering only, integrated urban energy assessments
and optimisation shall particularly focus on the characterisation on this specific end-use.

37
Chapter 2. Domestic hot water characterisation

The main significance of the proposed DHW characterisation method lies in the simultaneous
differentiation and temperature characterisation of various DHW streams at urban scale,
where data on inhabitant and household numbers and building types is available. This leads
to three main contributions to current integrated urban energy assessments and optimisation
methods.

First, the various heating demands of urban systems are better differentiated. This allows
to address the impact of specific optimisation measures, like water saving techniques or
in-shower, in-building or in-sewer heat recovery solutions. Moreover, the proposed method
already generates part of the necessary DHW data (stream types, flows, temperature, etc.) for
the modelling and integration of these optimisation measures at urban scale.

Second, the technological scope of integrated energy optimisation of heating utilities in build-
ings with low temperature space heating is increased. The utility temperature of this type of
building is defined by the DHW requirements, as space heating temperature lies below that
of DHW end-uses (Brand et al. [29]). These requirements are precisely characterised with
the use of the proposed DHW characterisation method, the only limit remaining Legionella
proliferation, which, under certain DHW system configurations, can still be avoided with hot
water production temperature below 50°C. Therefore, with the decrease of the temperature
level requirement from typical 60°C to below 50°C, low temperature utilities, like heat pumps
or low temperature waste heat, have a stronger impact in the integrated selection of optimal
heating utilities. The reduction in temperature lift profits heat pumping solutions, as their effi-
ciency is improved with lower condensing temperature, and waste heat at a low temperature
level can be further valorised as the demand of fitting low temperature heat users is better
characterised.

Finally, with the characterisation of the main DHW end-use loads, an additional configuration
for district heating transfer stations is available for integrated energy optimisation of urban
systems. The characterisation of these loads allows to model stand-alone heat exchangers, an
alternative to the heat exchanger and hot water storage unit configuration considered so far in
integrated energy optimisation. Previous works have showed that these stand-alone systems
can yield equivalent or even better costs and energy efficiency results. The optimal selection
between these two configurations should therefore be addressed in future integrated urban
energy optimisation works.

38
3 In-building waste water heat recovery
assessment

This chapter is the updated postprint version of an article published in Applied Energy (Bertrand
et al. [18]). The initial developments were presented at the 29th international conference on efficiency,
cost, optimisation simulation and environemental impact of energy systems (ECOS) 2016 in Slovenia
(Bertrand et al. [17]). This work was fully developped by the author of this thesis, with the support of
Riad Aggoune from the Luxembourg Institute of Science and Technology (Luxembourg) and François
Maréchal from the Ecole Polytechnique Fédérale de Lausanne (Switzerland).

39
Chapter 3. In-building waste water heat recovery assessment

Abbreviations and symbols

Abbreviations

COP coefficient of performance

DHW domestic hot water

FW fresh water

GIS geographical information system

GW grey water

HE heat exchanger

HW hot water

MFB multifamily building

MUB mixed-use building

SFB single family building

SH space heating

WH waste heat

WW waste water

WWHR waste water heat recovery

Symbols

A heat exchanger surface [m2 ]

C costs [e]

cp heat capacity [kJ/kg*K]

d use duration [s]

d Tm logarithmic mean temperature difference [K]

d Tmi n minimum temperature difference [K]

f use frequency [capita*day]-1 , [household*day]-1

I investment costs [e]

ṁ mass flow rate [kg/s]

m mass [kg]

p fuel price [e/unit]

PT payback time [a]

40
Symbols

Q̇ thermal load [kW]

Q energy [kWh]

S yearly operational financial savings [e]

Su subsidies [e]

T temperature [°C], [K]

t time [s]

U heat transfer coefficient [W/m2 K]

ε heat exchanger efficiency [-]

η utility efficiency [-]

x occ number of occupants [occupants]

∆ savings [-]

Super- and subscripts

b building

cond condensation

e end-use

evap evaporation

h household

ins installation

m material

o occupant

ph preheated

su start-up

t time

to total

u household

41
Chapter 3. In-building waste water heat recovery assessment

3.1 Introduction
With a total of 3’441 TWh, 26.8% of the EU28 final energy consumption in 2013 originated
from the household sector, coming only second to transport (31.6%) (European Commission
[59]). Residential domestic hot water (DHW) consumption represented, with 442 TWh, ap-
proximately 16% of the EU household heating demand (Enerdata [54]), energy that is currently
lost to the environment with its transfer to the sewers. With the improvement of the build-
ing envelope, DHW will have an increasingly important role in energy consumption, with a
contribution to total heating demand between 20 and 32% in high efficiency single family
buildings and between 35 to almost 50% in multifamily buildings (Meggers and Leibundgut
[138], Alnahhal and Spremberg [4], Bertrand et al. [19]).

An option to reduce DHW-related energy consumption in the building, among water flow
reduction devices and temperature level decrease, is to recover the heat from the various
waste water (WW) streams (in-building solution). Frijns et al. [80] estimated the theoretical
maximum waste water heat potential of an Dutch household to 2.16 MWh/a. As showering
is the DHW stream with the highest energy demand in a household (Elias-Maxil et al. [52]),
most of the works focused on the energy saving and cost impacts of shower heat exchangers
(Eslami-Nejad and Bernier [58], Wong et al. [240], Meggers and Leibundgut [138], Guo et al.
[89], McNabola and Shields [137], Kordana et al. [117], Torras et al. [217], Deng et al. [46])
or in combination with heat pumps (Liu et al. [125], Chen et al. [39], Wallin and Claesson
[234], Dong et al. [49], Gou et al. [86]) and solar energy (Liu et al. [125], Tanha et al. [212]).
Works on the use of grey water (GW, waste water not loaded with urine and faeces) streams
combined with heat pumps were reviewed by Hepbasli et al. [96]. Recently, Postrioti et al. [176]
developed an experimental setup for the assessment of building waste water heat recovery
potential, demonstrating that the heat pump could be almost independent from external
conditions in case of sufficient waste water rejections. Alnahhal and Spremberg [4] estimated,
based on measured data of residential waste water streams, that 30% of the DHW energy
demand could potentially be covered by heat recovery in combination with a heat pump.
Other publications focused on specific applications, like dishwashers (Paepe et al. [157], Lin
et al. [121], Persson [167], Persson and Roennelid [168], Hoak et al. [99], Hauer and Fischer
[94], Jeong and Lee [107], Bengtsson et al. [14], Saker et al. [185]), washing machines (Persson
[167], Persson and Roennelid [168], Pakula and Stamminger [158], Saker et al. [185]) and barber
shops (Sun et al. [209]) or on the experimental evaluation of waste water streams (Ramadan
et al. [180]).

However, specific data on mass flow and temperature level of the various residential WW
streams is not given, a fact already observed by Meggers and Leibundgut [138]. With the
exception of Meggers and Leibundgut [138] and Kordana et al. [117], who focused solely on
showers, characterisation methods applied to waste water streams considering inhabitant
and household number or end-use occurrences have also not been explored. In addition, the
energy saving or cost impacts were not related to the total heating demand while considering
varying building characteristics (size, period of construction, etc.). The relevance of these heat

42
 3.2. Method

recovery systems, both in terms of financial and energy saving impacts, nevertheless changes
according to the specificities of the building (e.g. conventional compared to high efficiency
buildings). Moreover, with the exception of Leidl and Lubitz [118], Ni et al. [148], Deng et al.
[46], the impact of in-building waste water heat recovery (WWHR) at the level of building
blocks, districts or a city has so far not been explored. However, Leidl and Lubitz [118] and
Ni et al. [148] applied only used a simplified ‘flat-rate’ energy saving value or the result from
one building type, respectively, instead of aggregating specific building results to the urban
scale. Deng et al. [46] calculated the energy savings from shower heat recovery considering
the population number, but did not consider other waste water streams. Finally, a preliminary
version of this work covered only certain WWHR configurations, and costs as well as the impact
of energy integration were not considered (Bertrand et al. [17]).

The objective of this work is to propose a new method for the detailed energy saving and
cost assessments of grey water heat recovery systems from the building level and, by data
geoallocation and aggregation, to the urban scale. Based on pinch analysis, the method
addresses (i) the WW streams characterisation necessary as input to (ii) the thermal load
calculation methods, taking into account the number of occupants, the number of households,
the appliance occurrences, the building type and the WWHR configuration.

Considering the current methodological shortcomings, the proposed method contributes

to existing works with a novel waste water characterisation approach. By geoallocating and
defining the waste water streams as a function of occupant and household number, end-use
occurrence and building type, the impact of WWHR on the energy consumption is better
quantified, which improves the comparison with other energy saving measures. By consid-
ering the building age and thus global energy performance, the importance of WWHR in
high efficiency residential buildings is demonstrated in detail for the first time. Finally, by
aggregating building-specific data, the relevance of grey water heat recovery at the urban scale
is more precisely calculated.

The methodology is described in Section 3.2. It is then deployed in two case studies in Section
3.3. Section 3.4 discusses the advantages, shortcomings and contributions of the presented
work.

3.2 Method
The energy savings from in-building WWHR systems are quantified in this work using pinch
analysis, a recognised assessment method for process integration and more particularly heat
recovery (Klemeš [112], Turton et al. [218]). Pinch analysis and its algorithmic formulation
(problem table method), initially developed by Linnhoff and Flower [122], assess the heat
recovery potential obtained from cooling down hot streams in order to preheat cold streams.
As an input, they require the thermal load as well as start and end temperatures of hot and
waste water streams. In the detailed assessment being conducted at building scale, tem-
peratures and loads must be characterised according to occupant and household numbers

43
Chapter 3. In-building waste water heat recovery assessment

as well as appliance occurrence (Meggers and Leibundgut [138]). Also, in order to generate
realistic load values, actual WWHR systems available on the market are considered. Due
to their configuration, the mass flows and temperature levels of the water streams of these
systems are constrained, which influence the final heat recovery potential. Their selection
is partially limited by space availability and thus dependent on the building type (McNabola
and Shields [137]). The operating cost savings are then quantified and the payback time of the
various WWHR systems are determined at the building level. By aggregating the geoallocated
outcomes to the required scale, the impact of such energy saving measures are finally obtained
at the level of a building block, a district or a city (Fig. 3.1).

Figure 3.1: In-building WWHR assessment method at the urban scale

3.2.1 Domestic grey water streams characterisation

In order to determine the heat recovery potential of grey water streams for residential DHW
end-uses heating, mass flow, duration and frequency of use per capita must be characterised
to calculate their thermal load. It is also important to define typical temperature levels and to
geographically allocate the various end-uses.

The characterisation and spatial allocation of European DHW streams were addressed in
the former chapter (Bertrand et al. [19]). A review of DHW end-use providing reference
data on temperature, mass flow, use frequency and duration, considering the inhabitant
and household numbers as well as end-use occurrence, was conducted. Similar data for
waste water streams is generally very scarce (Meggers and Leibundgut [138]). An equation
characterising grey water temperatures as a function of hot and fresh water (FW) temperatures
was proposed by Ni et al. [148]. However, the quantification approach of the temperature
loss coefficient required by that method is not given. Therefore, grey water streams are
characterised based on the work of Bertrand et al. [19], which builds on the characterisation
of residential water streams of Blokker et al. [22]. The thermal load of a stream e is obtained
from the heat capacity c p , the mass flow ṁ e and the water temperature difference between
exit from the end-use and rejection to the sewer (Teexi t − T r e j ect ) (Eq. 3.1)

44
 3.2. Method

Q̇ e = ṁ e × c p × (Teexi t − T r e j ec t ) (3.1)

With the use duration d e and daily use frequency f e , the related daily rejected heat is then
computed (Eq. 3.2).

Q e = Q̇ e × d e × f e (3.2)

At the building scale, the multiple grey water streams e are attributed to the occupant o (e.g.
shower use, washing), the household h (e.g. cloth washing) and the building b (e.g. cleaning
WW WW
of the common areas), with the related rejected heat Q e,o,h,b,t of the occupants, Q e,h,b,t of
WW
the households and Q e,b,t of the building. By summing up all the waste water streams e as a
function of the occupant and household numbers in the building over a considered period
of time (e.g. 365 days for a yearly assessment), the rejected waste heat at building level is
determined (Eq. 3.3).
" ¶#
365 µ
W W,bui l d i ng WW WW WW
X X X X XX
Q = Q e,b,t + Q e,h,b,t + Q e,o,h,b,t (3.3)
t =1 e h e o e

Using Geographical Information System (GIS) data containing geoallocated information on

occupant and household numbers for each building, the rejected waste water heat can be
calculated at the urban scale.

Regional data on mass flows, durations and use frequencies of water streams are indicated
in Chapter 2. Only the data on exit grey water temperature from most of the end-uses, as
well as the data from some appliances not connected to the HW circuit (washing machine,
dishwasher), are missing for the pinch analysis. A complementary literature review on the few
publications mentioning such information is conducted to determine this information.

This approach implies that only reference values are used for the characterisation of the
streams. A differentiation between age and gender of the occupants (Blokker et al. [22]), as
well as the socio-economic level influencing end-use occurrence (Bertrand et al. [19]) are
currently not considered.

[Link] Bathroom

Concerning shower streams, Wong et al. [240] provided an equation correlating the drain
temperature with outdoor temperatures. However, the method was deployed for Hong-Kong,
a humid sub-tropical city, with outdoor temperatures of 15°C in winter. This correlation might
not be applicable to other climates as it can be expected that in colder climates, the bathroom
temperatures remain constant over the year. A limited number of publications nevertheless

45
Chapter 3. In-building waste water heat recovery assessment

provides values for shower temperature differences. Eslami-Nejad and Bernier [58] mention a
difference of 4 K for Canada, while Wong et al. [240], Guo et al. [89], Dong et al. [49] indicate
ranges of 2-5, 5-8 and 6-8 K for China. A difference of 5 K is used for the heat exchanger (HE)
certification in Germany (Passivhaus Institut [162]).

No waste water temperature data stemming from bathtubs is indicated in the literature. To
obtain an order of magnitude of the temperature decrease, a difference between 0.5 and
1.5 K was measured under different conditions before and after bathing using a mercury
thermometer (Tab. 3.1).

Table 3.1: Bathtub waste water temperatures

Bath duration Room temp. Start temp. End temp. [°C] Temp.
[min] [°C] [°C] difference [K]

20 21 37.0 36.0 1
23 22 43.0 42.5 0.5
26 22 40.5 39.0 1.5
29 22 38.0 36.5 1.5
35 22 39.0 38.0 1

It can be assumed that the grey water temperature of the bathroom sink corresponds to that
of the DHW stream, as the distance and retention time in the sink are too short to induce a
relevant temperature decrease.

[Link] Kitchen

Concerning dishwashers, not all of the grey water is rejected at high temperature (Saker
et al. [185]). For the prewash phase, the water retains its initial, cold, temperature. Grey water
temperature levels varying between 34 and 61°C for the different washing phases (washing, hot
rinsing, cold rinsing, etc.) were given by Paepe et al. [157]. However, this publication may be
outdated, as a water consumption of approximately 33 l per washing cycle was mentioned by
the authors, which is more than twice the water use indicated in other works, e.g. Blokker et al.
[22]. Temperature profiles varying between 55 and 60°C are provided by other several authors
(Hoak et al. [99], Hauer and Fischer [94], Persson and Werner [170], Bengtsson et al. [14]),
while Jeong and Lee [107] presented the GW energy profile as a function of time. Information
according to waste water volumes was not provided.

Temperature and water volume profiles for an A rated, 12 place dishwasher from Blomberg/Beko
were presented by Saker et al. [185] (Tab. 3.2). The water volumes indicated in the table are
averaged to simplify the heat recovery calculations. The prewash water is assumed to be trans-
mitted at ambient temperature to the sewer. The washing phase temperature is confirmed by
the findings of Richter [183], who observed that 52% of users select cleaning temperatures at

46
 3.2. Method

65°C and higher. The energy of the condensing water (drying phase) is negligible compared to
the other phases and is therefore not further considered (Jeong and Lee [107]).

Table 3.2: Dishwasher grey water streams characterisation, according to Saker et al. [185]

Phase [-] Waste water Waste water

quantity [kg] temperature [°C]

Wash 5.0 65
Cold rinse 3.5 50
Hot rinse 4.0 45

Concerning dish washing by hand, a certain temperature decrease needs to be considered as

the plates are initially at room temperature. As no data is available, it is proposed to consider
a difference of 5 K. Short uses of the kitchen sink (e.g. hand washing) can be assumed as
inducing no temperature losses due to the low duration time.

[Link] Laundry

Concerning washing machines, Pakula and Stamminger [158] provided data on ownership rate,
wash cycle number, water consumption per wash cycle and most frequent wash temperature
for several countries and continents (e.g. 40°C in Western Europe). As not all of the machine
water is heated up (Saker et al. [185]), national household water consumption statistics to
quantify hot grey water volumes must therefore be avoided.

The temperature profile of a washing machine was provided by Persson [167], but the grey
water volumes were not given. Ni et al. [148] indicated a hot water (HW) temperature of 49°C
for this type of equipment. Saker et al. [185] provided water volume and temperature profiles
for a mid-range, A-rated, 7 kg washing machine manufactured by Blomberg/Beko. Of the 65
l water used, 10 l are rejected to the sewer at around 37°C, while the remaining grey water
remains generally cold.

3.2.2 Energy assessment of heat recovery configurations

In order to introduce the current shower and grey water heat recovery configurations and to
demonstrate their impact on the energy savings, pinch analysis is used in this section for a
single family building of 2.98 inhabitants (average number of inhabitants per household from
case study 1). The maximum heat recovery is obtained by defining a minimum temperature
difference d Tmi n between the hot and the cold streams. The results are represented in a
load/temperature diagram as represented in Fig. 3.2, with the cold composite curve as a blue,
bottom curve and the hot composite curve as a red, top curve. The point where both curves
are at the distance of the d Tmi n value is the pinch point. The plots are presented here in

47
Chapter 3. In-building waste water heat recovery assessment

corrected temperatures, deducing from the real temperature the d Tmi n contributions of
the two streams (here d Tmi n/2). The overlapping segment of the two curves indicates the
heat recovery potential, while the non-overlapping segments represent the remaining cooling
(left segment) and heating (right segment) requirements. As complementary outcomes, the
exergetic efficiencies of the various configurations are also presented. It should be noted that
the parameters used below are standard values (Schramek [192]), and that user behaviour and
single component optimisation (e.g. temperature optimisation), out of scope of the present
work, could still lead to further energy and water savings (Cahill et al. [37], Gutierrez-Escolar
et al. [90]). In addition, the effects of fouling of the heat exchangers and therefore efficiency
losses (Deng et al. [46]) are currently not considered. Finally, due to the urban scope, the
calculations are conducted under steady-state conditions.

[Link] Shower heat recovery configurations

As an input to the pinch analysis calculation, we consider fresh water and sewer temperatures
of 10°C (Spur et al. [197], Widen et al. [239]), shower head and tray temperatures of 40°C
(Bertrand et al. [19]) and 35°C (Passivhaus Institut [162]), respectively, and a mean water flow
of 0.13 kg/s (Widen et al. [239], Neunteufel et al. [147], Bertrand et al. [19]). A theoretical dTmin
of 3 K is assumed, which therefore shifts the hot and cold stream curves by 1.5 K.

Using mass balance and pinch analysis, the DHW shower production could be covered, in
theory, to 74% by heat recovery when considering fresh and waste water with identical mass
flows (balanced flow), with the waste water exiting at 11.5°C (corrected temperature at the
bottom left point where the two curves do not overlap, Fig. 3.2). However, the fresh water
temperature would reach 32°C instead of the required 40°C, which implies a further mixing
with hot water. This is not feasible, as the maximum fresh water mass flow is already reached.

Figure 3.2: Pinch analysis diagram - balanced conditions

Three concrete configurations are usually considered for the production of shower hot water
(Slys and Kordana [196]). The first system (configuration 1) uses the waste water to preheat
the hot water flow of the shower (Fig. 3.3a). The temperature of 40°C is then realised by mixing

48
 3.2. Method

with cold water. Due to mass balance, the heat recovery in this configuration implies a lower
hot water flow of 0.09 kg/s. In this case the heat recovery reaches 49%, with the grey water
leaving the heat exchanger at the corrected temperature 19°C instead of 11 °C (Fig. 3.3b).

(a) System (b) Pinch analysis diagram

Figure 3.3: Configuration 1

A second option (configuration 2) is to preheat the full shower mass flow of 0.13 kg/s, and to
split it between a preheated ‘cold’ water stream and a preheated water stream used for hot
water production (Fig. 3.4a). The heat recovery reaches 74% and the waste water exits at the
corrected temperature 11.5°C (Fig. 3.4b).

(a) System (b) Pinch analysis diagram

Figure 3.4: Configuration 2

In configuration 3, only the cold water is preheated (Fig. 3.5a). The hot water demand is
reduced, as the ’cold’ water has a higher temperature. Similarly to configuration 1, this config-
uration is constrained by the cold and hot water mixing, as the mass flow and temperature
after mixing must match the shower flow and temperature requirements. After heat recovery,
pinch analysis indicates that the waste water is at corrected temperature 19 °C and the heat
recovery corresponds to 48% of the heating load (Fig. 3.5b).

A small minimum temperature difference dTmin of 3 K and immediate heat transfer are
assumed above, in order to compare the efficiency of the various configurations. However,

49
Chapter 3. In-building waste water heat recovery assessment

(a) System (b) Pinch analysis diagram

Figure 3.5: Configuration 3

the energy savings by shower heat recovery are also dependent on the heat exchanger type,
which influences both heat transfer and duration of the exchange. Actual shower HE are either
mounted horizontally in the shower tray or vertically as an element of the waste water piping.
Horizontal heat exchangers have a short start-up phase (period where heat exchange does not
occur yet) of 5 seconds for the HE (Passivhaus Institut [163]), while additional 10 seconds for
the circulation duration through the shower pipe and tray is further assumed for the present
calculations. However, due to their small surface leading to a d Tmi n between 12 and 15 K,
their heat transfer efficiency is low (balanced flows of 0.13 kg/s). Vertical HE yield a higher
heat recovery efficiency due to a larger exchange surface but the d Tmi n is still between 9 and
10 K (source: Wagner Solar GmbH, [Link]). Tanha et al. [212] measured a start-up phase of
90 seconds.

The implementation of the different HE and heat recovery configurations is also constrained
by space availability imposed by the building type and by the location of the heating utility.
Vertical HE require one to two meters of space below the shower tray, which limits their instal-
lation as a retrofit solution in multifamily buildings or single family houses with showers at
the ground floor (McNabola and Shields [137]). They are mostly combined with configuration
1 and 2, as the component can be installed close to the heating utility for hot water preheating.
Horizontal heat exchangers have lower space requirements and are easily installed, even in
existing buildings (Schnieders [191]). They are mostly intended for systems where preheated
fresh water is mixed with hot water (configuration 3) and a direct connection to the heating
utility is not available.

Taking into account the actual HE efficiencies and start-up durations as well as implementa-
tion constraints, the daily shower energy savings ∆Q shower , expressed in kWh, can therefore
be determined at building level (Eq. 3.4):

ṁ ph × c p × (T ph − T F W ) × (t t o − t su ) × f × x occ
∆Q shower = (3.4)
3600

50
 3.2. Method

with ṁ ph the preheated mass flow in kg/s, c p the heat capacity in kJ/kg*K, T ph and T F W the
preheated and fresh water temperatures in °C, the duration (t t o − t su ) in s, with t t o being the
total and t su the start-up durations, the daily shower frequency f per person in (day*capita)-1
and the number of inhabitants x occ . Values for shower mass flow, duration and frequency
in various EU countries are given in Bertrand et al. [19]. The preheated mass flow and tem-
perature variables of the various configurations are obtained by energy and mass balances
depending of the considered configuration and heat exchanger type (Appendix A.1).

Using Eq. 3.4, and considering a household of 2.98 inhabitants, a vertical heat exchanger
implemented in configuration 1 and 2 (with estimated T ph of 21 and 26.5°C) yields daily
energy savings of 1.0 and 2.2 kWh/day, respectively, which represents 21 and 45% of the
daily shower energy requirements. The use of a horizontal heat exchanger combined with
configuration 3 would result in savings of 402 kWh/a, which is identical to the energy savings
indicated by McNabola and Shields [137] when assuming the same use frequency. The exergy
efficiency of the systems, based on the exergy values on waste water and preheated streams
with a reference temperature of 10°C and considering the various start-up durations, is 11%,
37% and 15%, respectively.

[Link] Grey water heat recovery configurations

The pinch analysis at building level is conducted with the streams described in Tab. 3.3, where
data from Bertrand et al. [19] and Section 3.2.1 are compiled. It is assumed that the building is
equipped with a bathtub and a dishwasher with an own internal heating system.

Pinch analysis was initially developed for industrial processes considering continuous opera-
tion. For the heat recovery analysis conducted here, it is necessary to consider mean flows of
water representing the mean power over its time of use. The DHW, grey water and hot water
stream loads are therefore obtained by summing the daily energy values and averaging these
over one hour, as expressed in Eq. 3.5-3.7.

x occ (ṁ e × d e × f e ) × c p × (Te − T F W )

P P
x hhol d
Q˙e = (3.5)
3600

GW
x occ (ṁ e × d e × f e ) × c p × (T GW − T sewer )
P P
GW x hhol d
Q̇ = (3.6)
3600

HW
x occ (ṁ e × d e × f e ) × c p × (T HW − T F W )
P P
HW x hhol d
Q̇ = (3.7)
3600

with x occ and x hhol d the inhabitant and household numbers, Te , ṁ e , d e and f e the tempera-
ture, mass flow, duration and frequency of use of the end-uses e, c p the heat capacity and the

51
Chapter 3. In-building waste water heat recovery assessment

Table 3.3: Domestic hot and grey water streams (Bertrand et al. [19])

Stream [-] Appliance Use level End-use Drain Mass day] Frequency
[-] [-] temp. temp. flow [1/capita*day]
[°C] [°C] [kg/s] [1/h-
hold*day]

Hand wash Kitchen Household 35 35 0.08 15 3.15

sink
Washing Bathroom Inhabitant 35 35 0.04 40 1.35
and shaving sink
Shower Shower Inhabitant 40 35 0.13 510 0.70
Bath Bath Inhabitant 40 39 0.20 600 0.044
Wash Dishwasher Inhabitant n.a. 65 0.08 60 0.30
(5kg)
Cold rinse Dishwasher Inhabitant n.a. 50 0.06 60 0.30
(3.5kg)
Hot rinse Dishwasher Inhabitant n.a. 45 0.07 60 0.30
(4kg)
Cloth Washing Household n.a. 37 0.17 60 0.45
washing machine (10kg)
n.a. - not applicable

various fresh water, grey water, sewer and hot water temperatures T F W , T GW , T sewer , T HW .
Complementary equations specific to the various configurations assessed below are given in
Appendix A.2.

When considering the immediate heat transfer between the specific DHW and WW streams,
represented in the pinch analysis of Fig. 3.6, 80% of the DHW heating load would be covered
by heat recovery, with the grey water being rejected at 14°C. However, DHW demand and
WW rejection do not occur simultaneously and a storage (and according control) systems are
necessary for heat recovery.

One option is to use a grey water storage tank for hot water production (configuration 4, Fig.
3.7a). This system reduces the DHW-related energy consumption by 52%. The water is stored
at 37°C and rejected to the sewer at a corrected temperature of 21°C (Fig. 3.7b). The exergy
efficiency of the system, with 10°C as a reference temperature, reaches 58%.

Another option is to use a grey water tank for fresh water preheating and a utility producing the
DHW only at the required temperature (configuration 5, Fig. 3.8a). The necessity of hot water
production at 55°C is linked to hygiene constraints (limitation of Legionella proliferation), but
can be avoided in buildings where the volume of the DHW distribution system does not exceed
3 l and individual pipes are installed (Brand et al. [29]). With a dishwasher, the actual DHW
end-use temperatures do not exceed 40°C (Bertrand et al. [19]). The DHW energy consumption
can be reduced by 80%, with the grey water rejected at 14°C and an exergy efficiency of 88%

52
 3.2. Method

Figure 3.6: Pinch analysis diagram - DHW and WW streams

(a) System (b) Pinch analysis diagram

Figure 3.7: Configuration 4

(Fig. 3.8b). While yielding a high efficiency, this configuration is difficult to implement, as
it requires direct connections between utility and DHW end-uses, which would drastically
increase installation and equipment costs.

Finally, grey water streams can also be used for hot water production and storage (configura-
tion 6, Fig. 3.9a). Using mass balance and pinch analysis, while considering a HW temperature
of 55°C, heat recovery would cover 55% of the DHW heating, with the grey water streams
leaving at a corrected temperature of 20°C and a exergy efficiency of the system of 63% (Fig.
3.9b).

For urban assessments comprising several buildings, the daily energy savings from grey water
heat recovery ∆Q g r e y is computed for each building using the problem table method (Linnhoff
and Flower [122]), the algorithmic form of the pinch analysis. The results are then aggregated
to the required scale (building block, street, district, city).

53
Chapter 3. In-building waste water heat recovery assessment

(a) System (b) Pinch analysis diagram

Figure 3.8: Configuration 5

(a) System (b) Pinch analysis diagram

Figure 3.9: Configuration 6

3.2.3 Costs calculations

The investment costs I are obtained as the sum of material and installation costs C m and C i ns
(Eq. 3.8).

I = C m +C i ns (3.8)

For shower heat exchangers, only the additional costs, compared to a normal shower tray or
inline drain system, should be considered, to avoid including the cost of the normal drain
system.

To calculate the investment costs of the heat exchanger of the grey water configuration, its
power is calculated considering the grey water stream with the highest power (usually linked
to the bath or the dishwasher) multiplied with a simultaneity factor of 1.15 for single family
buildings (Schramek [192]). Concerning multifamily buildings, the simultaneity factor given
by Gaderer [84] for DHW demand is multiplied with the sum of the maximum grey water load
P WW
u Q̇ u,max , with u the number of households in the building (Eq. 3.9).

54
 3.3. Case studies

Q̇ W W,bui l d i ng = [0.02 + 0.92u (−0.58) ] × WW

X
Q̇ u,max (3.9)
u

The yearly operation savings S are proportional to the energy savings ∆Q, the utility efficiency
η and the fuel price p (Eq. 3.10). As raised by Bianco et al. [20], energy prices can strongly
increase or decrease according to market changes (e.g. increased demand or oversupply),
thus making short to mid-term price estimation a difficult task. The current formulation of
the method allows the user to select its own energy price definition, i.e. an average price
taking into account a potential fluctuation (Kordana et al. [117]) or the energy price of a given
reference year (Tanha et al. [212]).

S = ∆Q × η × p (3.10)

The payback time P T in years is the ratio of investment costs I and operating savings S (Eq.
3.11).

P T = I /S (3.11)

The subsidies Su necessary to reach a given payback time P T is finally obtained with Eq. 3.12.

Su = I − (S × P T ) (3.12)

3.3 Case studies

In order to highlight the diversity of results encountered at city scale (compared to the single
family building-related assessments conducted in Section 3.2), two case studies, subdivided
into several scenarios to assess different optimisation configurations as to their impact on
the total heating demand, are deployed in this work. The characterisation as well as energy
savings and cost calculation methods are first applied to the existing residential buildings
of the city of Esch-sur-Alzette (case study 1). As the necessary data for the quantification of
heating demand of the low energy and passive (high efficiency) residential buildings of the
city are not available, and as grey water heat recovery could be of particular relevance for these
buildings, these are specifically assessed as to potential energy savings in a second case study
(case study 2).

3.3.1 Common input data

Temperatures of 10°C and 55°C are assumed for the fresh and hot water, respectively (Spur et al.
[197], Schramek [192], Widen et al. [239]). The grey water streams are characterised according

55
Chapter 3. In-building waste water heat recovery assessment

to Section 3.2.1 and summarised in Tab. 3.3. Use frequencies of dishwashers and washing
machines are taken from Blokker et al. [22] and Pakula and Stamminger [158], respectively.
The waste water mass of the streams of these two utilities are considered to be rejected within
one minute (Saker et al. [185]).

To limit the scope of the case studies, two types of heat recovery systems are deployed: a
horizontal shower heat exchanger (configuration 3) for scenario 1.1 and 2.1 and a grey water
heat recovery system for hot water preheating (configuration 6) for scenario 1.2, 2.2 and 2.3,
as the majority of the heating systems are equipped with a hot water storage tank (Schramek
[192]). The shower heat exchanger is a Ecoshower 900/DSS showerdrain channel WWHR
model 900/4, with an efficiency of 54% under stead-state conditions (Passivhaus Institut
[163]). With a pipe length of 6.8 m and an external diameter of 0.016 m, the power under
unbalanced conditions is 4,86 kW (source: Wagner Solar GmbH). Using energy and mass
balance, the fresh water exits the heat exchanger at a temperature of 27°C with a mass flow of
0.07 kg/s. The energy savings related to the grey water heat recovery system are calculated at
building level using the problem table method (Linnhoff and Hindmarsh [123]). A minimum
temperature difference of 5K is considered for the heat exchanger.

3.3.2 Retrofit solutions at urban scale

[Link] Specific input data

The DHW requirements of the city, based on GIS data (Service des travaux municipaux [194])
converted into a PostgreSQL database (PostgreSQL [175]), have been characterised in a former
work (Bertrand et al. [19]). The occurrence of the various waste water streams are related to
the use of the multiple DHW end-uses in each building. 78.8% of the households are equipped
with a dishwasher (Statec [200]). It is assumed that these dishwashers have an internal heating
system and are thus not connected to the hot water system. The remaining 21.2% do the dish
washing manually, with an end-use temperature of 55°C, a mass flow of 0.13 kg/s, a duration
of 48 s and a frequency of 3.15 uses per household per day (Blokker et al. [22], Schramek [192]).
The waste water is assumed to be emitted to the sewer at a temperature of 50°C (5K heat
losses).

The additional costs of a horizontal shower heat exchanger (scenario 1.1), compared to a
normal drain system, are between 150 and 300 e. Average investment costs of 225 e are
therefore considered for horizontal heat exchangers. Typical installation costs are around 100-
300 e; an average value of 200 e is used here (source: Wagner Solar GmbH). The investment
costs for the grey water heat recovery system (scenario 1.2) are summarised in Tab. 3.4. The
costs for the prefilter, the 3-way valve to avoid cold streams in the storage tank and the sensor
are market prices (source: Dehoust). The specific price of heat exchangers considers a price
increase of 50% to reflect the necessity of a double-wall construction as safety measure to avoid
a mixing of grey water with fresh water (source: Wiltec). Additional piping and installation

56
 3.3. Case studies

costs are estimated at 50 e and 200 e, respectively (source: Wagner Solar GmbH). A utility
efficiency of 90% is applied. Following the input of local partners regarding short-term energy
price fluctuations and the need for clear communication of the outcomes to a larger public, it
was decided to use the reference energy price of 2015, which reached 0.043 e/kWh (including
VAT, source: Sudgaz). All other values are without VAT (3% for construction projects).

Table 3.4: Cost parameters for grey water heat recovery

Component [-] Unitary price, excluding VAT [e]

Prefilter 330
Heat exchanger 45 e/kW
3 way valve 200
Sensor 70
Piping 50
Installation costs 200

[Link] Results

Energy savings Figure 3.10 (scenario 1.1) and Fig. 3.11 (scenario 1.2) represent the relative
energy savings (light blue) and remaining DHW energy requirements (dark blue), related to
the total fuel consumption for heating (covering space heating, DHW and utility inefficiency
losses). The percentages indicated are the savings, relative to the total fuel consumption,
obtained from the implementation of the HR systems.

With a horizontal heat exchanger, energy savings between 1.4 and 2.2% in single family
buildings, as well as between 3.8 and 5.7% in multifamily and mixed-use buildings can be
reached (Fig. 3.10). For a three inhabitants, single family house, the energy savings would
reach 402 kWh/a. An average multifamily building of 12 inhabitants, 5.5 households, would
save 1’617 kWh/a. 23% of the showering energy would thus be saved. These outcomes are in
line with the values indicated by McNabola and Shields [137] when recalculated for a shower
frequency of 0.7 1/day. Wong et al. [240] mention savings between 4 and 16% for a single-
pass HE and Frijns et al. [80] indicate natural gas savings between 30 and 40%. Guo et al.
[89] reached shower energy savings of 50% with a novel design for the heat exchanger and
conducting the assessment under laboratory conditions.

With grey water heat recovery for hot water preheating, savings between 3.0 and 4.5% in single
family buildings and between 8.0 and 11.9% for multifamily and mixed-use buildings are
obtained (Fig. 3.11). A three inhabitants, single family house, would save in average 1’059
kWh/a, while 4’282 kWh/a would be saved in a multifamily building of 12 inhabitants, which
represents approximately 50% of the DHW energy demand. Compared to the theoretical
maximum heat recovery potential of 2’160 kWh/a indicated by Frijns et al. [80] for households

57
Chapter 3. In-building waste water heat recovery assessment

Figure 3.10: Scenario 1.1: horizontal shower heat exchanger - relative energy savings (SFB –
single family building, MFB – multifamily building, MUB - multi-use building)

in the Netherlands, it appears that, by using real input data and considering actual WWHR
system configurations, 49% of these theoretical savings are achieved.

Figure 3.11: Scenario 1.2: grey water heat recovery - relative energy savings (SFB – single family
building, MFB – multifamily building, MUB - multi-use building)

58
 3.3. Case studies

Payback time For the assessed heat recovery systems, the investment costs and cost savings
depend on the energy price, the numbers of inhabitants and the number of households. For
one-household buildings (Fig. 3.12), considering the reference natural gas price indicated
above, shower heat recovery for singles or couples leads to payback times above 50 years,
while the average household (three inhabitants), would see a payback time of almost 18 years.
From 6 inhabitants on, the payback time falls below 10 years. For the grey water heat recovery
of scenario 1.2, the payback time is almost twice as high as scenario 1.1. A payback time of
10 years is reached for households of at least 13 inhabitants. Single family buildings with 12
inhabitants are not occurring in the city and are therefore not displayed. With increasing
household numbers per building, the average payback time of shower HR does not change for
the households, as displayed for a five households building (Fig. 3.13). The payback time for
three inhabitants per 5 households remains at 18 years. The payback time of grey water heat
recovery falls below 10 years for 18 occupants.

Figure 3.12: Average payback time for one household buildings

The direct comparison of these outcomes with the results of similar works is generally not
recommended due to strongly varying conditions (e.g. energy prices, equipment and/or
installation costs, use frequency and duration). Nevertheless, it should be mentioned that
Kordana et al. [117] obtained, with an energy price of 0.16 e/kWh, a payback time for vertical
shower heat exchangers of 7, 5 and 4 years for a Polish household with 3, 4 and 5 inhabitants,
respectively. Values between 4 and 5 years for a 4 persons household are calculated by Slys
and Kordana [196]. Tanha et al. [212] estimated the payback time of a vertical shower heat
exchanger system for a four inhabitant house installed in Canada to be 5.6 years (with an
electrical heater at 0.136 $/kWh) and 17 years (with a natural gas boiler at an energy price of
0.046 $/kWh). McNabola and Shields [137] indicated a paypack time of 5 years, with an energy
price of 0.18 e/kWh. The much lower payback times can, at least partially, be explained by the

59
Chapter 3. In-building waste water heat recovery assessment

Figure 3.13: Average payback time for five households buildings

considered energy price, which is three times higher than the price used in the present case
study. On the other hand, the payback time of 13 years obtained here for a four inhabitant
single family building is in the same order of magnitude than the 17 years of Tanha et al. [212],
who used a similar energy price.

With the considered investment costs and actual natural gas price, shower or grey water heat
recovery are currently not an economically viable solution in buildings with low inhabitant
numbers. Subsidies from the state or the municipality would therefore be necessary as
incentive for the implementation of such energy saving solutions.

Assessment at city level The yearly total heating demand of the residential sector of Esch-
sur-Alzette amounts to 189.2 GWh, of which 23.8 GWh is for domestic hot water demand. By
aggregating the energy saving and cost results of the various residential buildings to the level
of the city, the absolute and relative savings, investment costs, subsidies necessary to reach
a payback time of 10 years (the acceptance limit for the implementation of environmental
technologies (Leidl and Lubitz [118])), and the specific cost per saved kWh of energy are
determined for the reference year 2015 (Tab. 3.5). The values in brackets are the percentages
of subsidies related to the investment costs. With the 2015 natural gas price, subsidies of
approximately 60% of the investment costs are required to reach a payback time of 10 years by
the inhabitants.

The city energy savings of 3.1% are in line with Deng et al. [46], who estimated the savings of
shower heat recovery systems in the city of Amsterdam to be 4%. The relative energy savings of
24.3% in scenario 1.1 come close to the 35% estimated by Leidl and Lubitz [118], considering

60
 3.3. Case studies

Table 3.5: Scenario 1.1 and 1.2 energy savings at city scale, considering 2015 energy price

Scenario Energy Energy Energy Investment Subsidies to Specific costs

sav- savings, savings, costs [e] reach 10 per saved
ings related to related to years payback energy
[GWh] total heating DHW time [e] [e/kWh]
demand [%] demand [%]

1.1 5.8 3.1% 24.3% 6’269’018 3’559’414 1.08

(57%)
1.2 12.0 6.3% 50.6% 14’239’324 8’771’601 1.19
(62%)

that they assumed the application of a vertical shower heat exchanger, which yields a higher
energy savings than the horizontal HE used in this work.

The geoallocated energy savings per district are represented for the two scenarios in Fig. 3.14a
and b.

(a) Scenario 1.1: horizontal shower heat recovery (b) Scenario 1.2: grey water heat recovery

Figure 3.14: Energy savings per districts

3.3.3 Energy optimisation of high efficiency residential buildings

[Link] Specific input data

Four scenarios focusing on the energy savings in low energy and passive single and multifamily
buildings in Luxembourg are considered in this second case study. The reference scenario

61
Chapter 3. In-building waste water heat recovery assessment

2.0 includes space heating and hot water demand at 55°C. The same streams are used for
scenario 2.1, to which the shower waste water stream for heat recovery with an horizontal
heat exchanger, is added. In scenario 2.2, space heating at 28/35°C and hot water streams at
55°C and all grey water streams listed in Tab. 3.3 are used as inputs. Scenario 2.3 presents an
integrated approach to heating optimisation. Space heating remains identical, while DHW
is characterised as 45°C hot water (Meggers et al. [139]) and the waste water streams can be
cooled down further than 10°C. These streams are used as input to the pinch analysis to design
the optimal heat recovery and utility system. It is important to note that in this case, due to
the low temperature level, it is assumed that the hot water storage and distribution systems
are equipped with an adequat protection against the proliferation of Legionella (e.g. regular
thermal desinfection), as indicated in technical standards (Schramek [192], Brand et al. [29]).
The energy consumption and investment costs related to this protection are not considered in
the following calculations.

The specificities of the considered buildings are summarised in Tab. 3.6. The average inhabi-
tant and household numbers as well as average surface area are taken from the GIS database
from Esch-sur-Alzette (Service des travaux municipaux [194]), while the specific space heating
energy consumption is from the Luxembourgish legislation on energy efficient buildings (Lux-
emburgish Parliament [128]). It is assumed that 100% of the households are equipped with
a washing machine equipped with an own heating circuit. The space heating nominal load
is calculated using the heating signature as described by Girardin et al. [85] and the monthly
temperature profile indicated in Luxemburgish Parliament [128] for a theoretical coldest day
(-10°C) and average day (7°C), up to an outdoor temperature of 15°C.

The electricity consumption and savings are calculated for three specific periods: winter, inter-
mediate and summer (Tab. 3.7). The period durations have been determined by considering
the space heating load and energy requirements of Tab. 3.6, assuming a short duration of
10 hours at minimal outdoor temperature. Floor heating temperature is set to 28/35°C. The
temperature of the grey water streams to the sewer does not go below 10°C in scenario 2.2.
and 4°C (winter), 7°C (intermediate period) and 17°C (summer) for scenario 2.3.

The heating utility considered is a two-stage air/water heat pump. The first stage covers
water temperature of 35°C for space heating and hot water preheating, with a condensation
temperature Tcond of 38°C, and a second stage for hot water (scenarios 2.0, 2.1 and 2.2: 55°C,
scenario 2.3: 45°C), with condensing temperatures of 58°C (scenarios 2.0 and 2.1), 60°C
(scenario 2.2, considering a dTmin of 5 K for the grey water heat exchanger) and 48°C (scenario
2.3). For scenarios 2.0., 2.1 and 2.2, the evaporation temperature Tev ap is 3K below outdoor
temperature. For scenario 2.3, the remaining heat going to the sewer is used as partial heat
source for the evaporation side of the heat pump (see Tab. 3.7 for the considered temperature
levels). The remaining heat source for evaporation is air. Finally, the coefficient of power COP ,
to calculate the electricity consumption, is obtained with Eq. 3.13, considering an exergy
efficiency rate η of 34% (Girardin et al. [85]).

62
 3.3. Case studies

Table 3.6: Building characteristics

Building type Single family building Multifamily building

Total surface [m2 ] 166 512

Inhabitant number 3 12
[inhabitant]
Household number 1 5.5
[household]
DHW energy demand 1’128 4’584
[kWh/a]
Efficiency type Low energy Passive Low energy Passive
Specific SH energy 43 22 27 14
demand [kWh/m2 ]
Total SH energy 7’123 3’644 13’816 7’164
demand [kWh]
SH maximal load [kW] 3.6 1.8 7.0 3.6
SH intermediate load 1.2 0.6 2.2 1.2
[kW]

Table 3.7: Operating period characteristics

Period Winter Intermediate Summer

Duration [hrs] 10 6’163 2’587

Average outdoor temp. [°C] -10 7 17
Scenario 2.3 evaporation temp. [°C] 4 10 17

Tcond
COP = η × (3.13)
Tcond − Tev ap

[Link] Energy integration considering WWHR

By applying energy integration design rules based on pinch analysis (no heat exchangers
across the pinch point, heat pumps must have their evaporation and condensation elements
below and above the pinch point (Becker [11])) to detect optimal heat recovery and utility
design configuration, the energy consumption for the optimised scenario 2.3 is calculated.
The hot and cold composite curves of the energy integration for the single family, low-energy
building are represented in Fig. 3.15 for the three periods. The horizontal segments of the hot
stream curves represent the condensation loads of the heat pump. The horizontal segment
of the cold streams is the evaporation load. The pinch point, for the four building types, is

63
Chapter 3. In-building waste water heat recovery assessment

situated at 29.5°C for the winter and intermediate periods, and 33.5°C for the summer period.
Heat recovery must therefore be designed with two heat exchangers: one to preheat the cold
streams with the hot streams below and one for heat transfer between the streams above the
pinch point.

(a) Winter

(b) Intermediate

(c) Summer

Figure 3.15: Scenario 2.3 - Pinch analysis diagram

64
 3.3. Case studies

The heat recovery potential in winter is rather limited (4% of the total power) due to the
relevance of the space heating requirements. However, the waste heat can be fully used for
hot water preheating until a temperature of 11.5°C, then valorised as heat source for the heat
pump. During the intermediate period, the grey water is led to the heat pump at a corrected
temperature of 14°C, with waste heat recovery contributing to 11% of the load. In summer,
where no space heating demand occurs, 55% of the heating load is covered by heat recovery,
with the grey water cooled down to 21°C before being used by the heat pump. Due to the small
load, using the remaining grey water as heat source for the heat pump reduces the electricity
consumption only by 1-2%.

[Link] Results

The results of the various scenarios of case study 2 are summarised in Fig. 3.16. The percent-
ages indicate the electricity savings related to the electrical consumption for heating.

Figure 3.16: Case study 2: results (LE - low energy, P - passive, SFB – single family building,
MFB – multifamily building)

With the implementation of a horizontal shower heat exchanger, the total electricity consump-
tion related to heating can be reduced between 6 and 14% according to the building type.
The single family building would save 191 kWh/a, while 772 kWh/a would be avoided in the
multifamily building. This corresponds to a reduction of 25% of DHW-related electricity con-
sumption. The difference with the savings of 402 kWh/a indicated in Section [Link] are due to
the heat pump, which produces 2.1 units of heat per unit of electricity for the conditions cited
above. The impact of grey water heat recovery on electricity consumption is between 10 and
22%. This represents savings of 38% on the DHW energy requirements with reductions of 291
kWh/a and 1’179 kWh/a for the single family house and the multifamily building, respectively.

65
Chapter 3. In-building waste water heat recovery assessment

Finally, the integrated approach, where hot water production, heat recovery and utility design
are optimised, reduces the electricity consumption between 18 and 41%. The single family
building would save 544 kWh/a and the multifamily building would reduce its electricity use
by 2’213 kWh/a. As already observed in the first case study, waste water heat recovery systems
have a larger impact in multifamily buildings than in single family houses.

Considering these outcomes, the implementation of grey water heat recovery systems should
be included in an energy integration approach to further optimise the energy savings. More-
over, it is demonstrated that energy integration approaches for high efficiency residential
buildings (as deployed by e.g. Fazlollahi et al. [73], Jennings et al. [106] and applied in scenario
2.3) should also include waste water streams and hot water demand optimisation to increase
the energy optimisation potential.

3.4 Discussion
A new, detailed method to assess the energy savings and costs at an urban scale from in-
building grey water heat recovery in residential buildings is proposed.

One of the main strengths of the deployed work is the detailed characterisation and spatial
allocation of residential grey water streams as to mass flow and temperature level, as a function
of inhabitant and household numbers. This characterisation allows a more precise assessment
of WWHR potential at urban scale. In addition, WWHR energy savings can be related to
buildings specificities, e.g. end-use occurrence, building type, age and energy efficiency. Their
impact can therefore be calculated in reference to the total heating demand, thus supporting
decision processes as to the selection and design of appropriate energy saving measures in
buildings. Furthermore, the energy savings and costs are attributed to each specific building
in the considered geographical scope. Results can easily be generated at specific spatial levels
(building blocks, streets, districts city). The outcomes of urban energy assessments considering
WWHR are thus improved, as the large-scale results are obtained by data aggregation. Finally, it
is also demonstrated that an integrated approach to heating system selection and design must
include hot water demand and grey water streams to further optimise energy consumption.

One limitation of the exposed work consists in the low availability and poor technical, geo-
graphical and socio-economic detail level of the input data. Knowledge of the occurrence of
retrofitting constraints, which influences configuration selection at the urban level, is also
limited. The proposed calculation methods are also simplified to accommodate the problem
of scale and do not reflect thermal losses by distribution, energy requirements of desinfection
processes related to low temperature solutions (protection against Legionella proliferation),
transient conditions of storage systems, or a long-term efficiency drop of the heat exchangers,
which further reduce the energy saving potential. Moreover, the urban assessment of case
study 1 and the integrated solution deployed in case study 2 do not include other systems (e.g.
sewer heat recovery, solar thermal collectors), which could, potentially, further improve the
optimisation potential.

66
 3.4. Discussion

Concerning data availability, mass flow and temperature data as a function of building type
and socio-economic level of the household must be further gathered. The use of geographically
weighted regression would also improve the quality of the assessment as the spatial allocation
of certain end-uses would better reflect socio-economic conditions. Also, sensitivity analysis
shall be applied to the proposed method in order to quantify the uncertainty of the outcomes,
as reflected in case study 1 by the comparison of the payback time values. In addition, the
occurrence rate of the implementation constraints, as a function of building type, must
be better characterised in order to improve the assessment of WWHR systems at an urban
scale. More detailed calculation methods must also be developed for the considered urban
scale, although resolution time might become an issue when assessing very large systems.
Finally, the competition between in-building and sewer heat recovery configurations must be
assessed at an urban scale in order to select adequate solutions according to district/city age
and infrastructure.

The main significances of the present work are the characterisation method of grey water
streams and the detailed energy saving and cost assessments methods, considering building
specificities and various grey water streams, of residential WWHR potential at the urban scale.
The exposed methods lead to several contributions in the field of building and urban energy
analysis and optimisation.

At building level, residential grey water streams are more specifically characterised by reflecting
DHW end-use occurrence as well as inhabitant and household numbers. The assessment
for grey water heat recovery potential is therefore qualitatively improved, independently of
the configuration (in-building, in-sewer or at waste water treatment plants), which allows a
better comparison with other energy saving measures. In addition, the integrated optimal
selection of heating utility configurations is extended with the characterisation of the grey
water streams as additional source for heat recovery or heat pumps.

At the urban scale, energy and cost assessments at building block, district or city levels are
qualitatively improved and spatially better differentiated, as the outcomes are generated by
results aggregation of the single buildings. Energy assessments and optimisation, focusing so
far mostly on thermal insulation and heating utility selection, are also expanded to include
detailed grey water heat recovery as additional optimisation measure.

With the proposed calculation and the related waste water streams characterisation methods,
consulting companies conducting urban, municipal or even national energy assessments can
increase the scope of their work and better rank and select optimisation scenarios. This will
result in a improved decision-making process by local and national politicians regarding the
implementation of energy saving measures in the residential sector.

67
4 Optimised regional waste heat valori-
sation

A preliminary version of this work was presented at the Strategic Energy Technology Plan conference
(SETPlan) 2015 in Luxembourg (Bertrand et al. [16]). This work, fully developped by the author of
this thesis, was supported by Riad Aggoune from the Luxembourg Institute of Science and Technology
(Luxembourg) as well as Alberto Mian, Ivan Kantor and François Maréchal from the Ecole Polytechnique
Fédérale de Lausanne (Switzerland).

69
Chapter 4. Optimised regional waste heat valorisation

Abbreviations and symbols

Abbreviations

CHP combined heat and power

DHW domestic hot water

EAF electric arc furnace

ESCo energy service company

HP heat pump

ICE internal combustion engine

IPPC integrated pollution and prevention control

MILP mixed integer linear programming

MINLP mixed integer non linear programming

WH waste heat

Continuous variables

annualised investment cost of equipments e (turbine, heat exchangers,

Ce
pipelines and heating utility) [e/a]

HE HEj annualised investment cost of heat exchanger at source i and sink j of

C i j i ,C i j
connection i j [e/a]

C jHU annualised investment cost for hot utility at sink j [e/a]

C iI nv annualised investment costs of electricity production at source i [e/a]

C iI jnv annualised investment costs of heat supply at sink j [e/a]

Op operating costs of electricity production at source i in temperature interval k

C i ,k,t
and period t [e/h]

Op operating costs of heat supply at sink j in temperature interval k and period

C j ,k,t
t [e/h]
Pi p
Ci j annualised investment cost of transport pipes of connection i j [e/a]

70
Continuous variables

I iTur b turbine investment cost at source i [e]

HE HEj heat exchanger investment cost at source i and sink j for connection
Ii j i , Ii j
i j [e ]

I jHU heating utility investment cost at sink j [e]

investment cost at source i or sink j of equipment e (turbine, heat

Ie
exchangers, pipelines and heating utility) [e]
Pi p
Ii j pipe investment cost for connection i j [e]

ṁ i j highest mass flow of connection i j [kg/s]

mass flow of connection i j in temperature interval k and period t

ṁ i j ,k,t
[kg/s]

Pi profits from electricity production at source i [e/a]

Pj profits from heat supply at electricity production at sink j [e/a]

specific heat losses of connection i j in temperature interval k and

q̇ iLoss
j ,k,t
period t [kW/m]

Q̇ Bj o , Q̇ Cj H P , Q̇ H
j
P
highest thermal load of boiler, CHP and heat pumps of sink j [kW]

thermal load of boiler, CHP unit and heat pump of sink j in

Q̇ Bj ,k,t
o
, Q̇ Cj ,k,t
HP
, Q̇ H P
j ,k,t
temperature interval k and period t [kW]

waste heat load transferred to cooling utility of source i in

Q̇ iCU
,k,t
temperature interval k and period t [kW]

thermal load for electricity production at source i in temperature

Q̇ iEl,k,t
ec

interval k and period t [kW]

HE HEj
Q̇ i j i , Q̇ i j heat exchanger loads at source i and sink j for connection i j [kW]

hot utility load of sink j in temperature interval k and period t , as the

Q̇ HU
j ,k,t ,d em ,l oss
sum of Q̇ HU
j ,k,t
and Q̇ HU
j ,k,t
[kW]

,d em
hot utility load for the demand of sink j in temperature interval k and
Q̇ iHU
j ,k,t
period t [kW]

71
Chapter 4. Optimised regional waste heat valorisation

Continuous variables

,l oss
hot utility load for the compensation of the transportation losses of
Q̇ iHU
j ,k,t
connection i j in temperature interval k and period t [kW]

heat load of transportation losses of connection i j in temperature

Q̇ iLoss
j ,k,t
interval k and period t [kW]

waste heat load for heat demand and heat losses compensation from
Q̇ iWj ,k,t
H

source i for sink j in temperature interval k and period t [kW]

H ,d em
waste heat load for heat demand transferred from source i for sink j in
Q̇ iWj ,k,t
temperature interval k and period t [kW]

waste heat load for heat losses compensation related to heat transport
H ,l oss W H ,l oss
Q̇ iWj ,k−1,t Q̇ i j ,k,t from source i for sink j at period t from temperature interval k − 1 to k

[kW]

electricity production revenues at source i in temperature interval k and

R iEl,k,t
ec

period t [e]

electricity production revenues from utility at sink j in temperature

R El ec
j ,k,t
interval k and period t [e]

R i ,k−1,t , R i ,k,t heat cascaded from interval k − 1 to the interval below [kW]

Binary variables

P i p,i nvest use of pipe type with standard size dn of connection i j , for investment
y i j ,d n
cost calculations

P i p,l oss use of pipe type with standard size dn of connection i j in temperature
y i j ,d n,k,t
interval k and period t , for heat loss calculations

existence of heating utility technology, either boiler, CHP or heat pump at

y Bj o , y Cj H P , y H
j
P

sink j

Parameters

cpi,t , cpj,t heat capacity of hot stream i and cold stream j at period t [J/kg*K]

dij distance between source i and sink j [m]

72
Parameters

dt duration of period t [h]

fjCHP
,t electricity conversion factor of CHP of sink j at period t [-]

fijground underground type for connection i j [open field: 0, road: 1]

HIn In
i , t , H j ,t , inlet enthalpy of hot stream i and cold stream j at period t [J/kg]

HOut Out
i , t , H j ,t outlet enthalpy of hot stream i and cold stream j at period t [J/kg]

ITurb HE
ref , Iref reference investment cost for turbines and heat exchangers [e]

investment cost for boilers, CHP units, heat pumps and internal
IBo CHP , IHP ,IICE
j , Ij j j
combustion engine specific to sink j [e]

M large value [-]

ṁi,t , ṁj,t mass flow of hot stream i and cold stream j at period t [kg/s]

ṁmax
dn maximum mass flow allowed in pipe diameter dn [kg/s]

ne expected lifetime of equipment e [a]

pIC E specific investment price for an internal combustion engine [e/kWel ec ]

,s Elec,s
piElec
,t
, p j ,t electricity selling prices for turbine at source i and CHP at sink j [e/kWh]
,b Heat,s
piHeat
,t
, p j ,t heat purchase price from source i and selling price to sink j [e/kWh]
Elec,b
pGas
j , t , p j ,t natural gas and electricity purchase prices of sink j at period t [e/kWh]

pField Road
dn , pdn specific pipe cost according to pipe type, for fields and roads [e/m]

PEl ec Electric load of an internal combustion engine [kW]

q̇Loss
dn,k,t specific heat losses of standardised pipe type [kW/m]

available waste heat load at source i in temperature interval k and period

Q̇Avail
i , k, t
t [kW]

Q̇Dem
j,max maximal heat load of sink j for the design of the heating utility [kW]

Q̇Dem
j , k, t heat demand of sink j in temperature interval k and period t [kW]

reference turbine and heat exchanger thermal loads for investment cost
Q̇turb HE
ref , Q̇ref
calculations [kW]

r interest rate [-]

rp , rins internal and external, insulated, pipe radius [m]

73
Chapter 4. Optimised regional waste heat valorisation

Parameters

heat supply revenues at sink j in temperature interval k and period

RHeat
j , k, t
t [e ]

corrected inlet temperature of hot stream i and cold stream j at

TIn In
i ,t ,T j , t
period t [K]

TOut
t outdoor temperature at period t [K]

corrected outlet temperature of hot stream i and cold stream j at

TOut Out
i ,t , T j , t
period t [K]

upper and lower temperature level of temperature interval k at

TUp
k, t
, TLow
k, t
period t [K]

α1 contingencies correction factor [-]

κi thermal conductivity of pipe insulation material [W/m*K]

ηBo CHP ,ηHP ,ηTurb

j ,η j j boiler, CHP thermal, heat pump and turbine efficiency [-]

γ cost scaling factor for investment cost calculations [-]

Subscripts

dn standard pipe diameter

i hot stream

j cold stream

k temperature interval

t time period

Sets

CS set of cold streams j

DN set of standard pipe types d n

HS set of hot streams i

I Tt set of temperature intervals k in period t

TP set of periods t

74
 4.1. Introduction

4.1 Introduction
In 2012, the European Union emphasized, with the directive on energy efficiency (European
Parliament [63]), the necessity to valorise waste heat (WH) from manufacturing, electricity
production and waste incineration industries at the regional scale. Project proposals for new
or extensions of industrial plants and district heating networks must assess, through cost
benefit analysis, opportunities to valorise nearby industrial waste heat sources.

While a real interest for regional waste heat valorisation is expressed by industries and muni-
cipalities alike, these actors are often reluctant to engage in the actual planning, investment
and management tasks as they are not part of their core business activities (Ammar et al.
[5], Brueckner et al. [33], Päivärinne and Lindahl [173]). The involvement of a third party, e.g.
an energy service company (ESCo), that would bear the risks of the initial investment and
manage the waste heat, is therefore necessary to pursue a realistic WH valorisation project.
However, regional waste heat recovery represents a challenging engineering task where mul-
tiple aspects need to be considered and the economic viability, in competition with other
heating alternatives, must be assured for an ESCo to engage in such an enterprise (Päivärinne
and Lindahl [173], Deng et al. [45]).

A large number of publications focused on urban waste heat use (Svensson et al. [210], Sandvall
et al. [187], Fujii et al. [82], Xia et al. [241]), particularly assessing the economic viability of such
projects (Kim et al. [111], Fang et al. [71], Morandin et al. [143], Eriksson et al. [57], Sandvall
et al. [188]). In their network design approach, Haikarainen et al. [93] included standard
pipe diameters for the calculation of the investment costs. Due to their scope, these works
did not formulate waste heat valorisation methods for higher temperature demands, e.g.
electricity production or industrial heat demand, and thus missed potential valorisation
opportunities (Ammar et al. [5]). Perry et al. [166], Varbanov and Klemeš [225], Čuček et al.
[227], Oh et al. [152] adapted the Pinch Analysis and Total Site Targeting methods, initially
developed for industrial processes, to large scale energy systems comprised of industries
and buildings. Electricity production and transportation heat losses were not considered,
despite the fact that the latter aspect is highly relevant for large scale waste heat valorisation
(Stijepovic and Linke [207], Wang et al. [236], Nemet et al. [146]). Tveit et al. [219], Stijepovic
and Linke [207], Stijepovic et al. [208], Nemet et al. [146] combined detailed Mixed Integer
Linear and Non Linear Programming (MILP, MINLP) approaches for waste heat valorisation
applicable to industrial zones and beyond, selecting waste heat connection matches according
to temperature levels. In these works, the connection selection was generally formulated as
linear problem, while the detailed design of the heat exchange network was formulated as a
non-linear problem. Oluleye et al. [153] developed an hierarchy-based method for waste heat
on and off-site valorisation. However, these works were not formulated for ESCos and did not
consider (with the exception of Oluleye et al. [153]) that sinks can have different energy prices
for the same fuel type and time period (Eurostat [69]), which influences the WH valorisation
opportunities. In addition, transportation heat losses were calculated without taking into
account standard pipe sizes/types, which would have improved the quality of the results.

75
Chapter 4. Optimised regional waste heat valorisation

Finally, the selection of the optimal heating utility technology, necessary as a backup system
in case of waste heat supply interruption, was not considered. Integrating optimal heating
technology selection to the waste heat valorisation would allow the ESCo to improve their
service value (Brady et al. [28]).

The objective of this work is to propose a multi-period optimisation method for regional
waste heat valorisation and utility selection formulated for ESCos. Potential connections
and thermal loads (including heat losses) between industrial and urban systems as well as to
electricity production units, considering specific energy prices, are determined according to
temperature matches and generated profits. Due to the high number of variables, the method,
based on the works of Papoulias and Grossmann [160], Maréchal [133], Mian et al. [141], is
formulated as an MILP problem to obtain a single, global, solution (Hillier and Lieberman
[98], Lin et al. [120]).

Considering the methodological shortcomings identified in the literature, the proposed work
contributes to existing developments with a novel ESCo-specific formulation targeting regional
waste heat valorisation. By considering specific sink energy prices in addition to the thermal
constraints obtained from the heat cascade, the competition of waste heat valorisation with
other energy sources is more thoroughly assessed. With the combined optimal waste heat
valorisation and heating technology selection, regional heat valorisation is conducted using
an integrated systems approach. Finally, by including standard pipe sizes in the method, the
calculation of the heat losses accounts for more realistic constraints on valorisation viability
which are often overlooked.

The optimisation problem is formulated in Section 4.2. Its application to the south-west region
of the Grand-Duchy of Luxembourg, composed of two steel plants as heat sources as well as
three industries and nine towns as heat sinks, is demonstrated in Section 4.3 and Appendix C.
The advantages, shortcomings and impacts are discussed in Section 4.4.

4.2 Method

4.2.1 Scope

Similar to the waste heat recovery method for industrial zones of Stijepovic and Linke [207],
Stijepovic et al. [208], the design of regional waste heat valorisation system can be decomposed
into several steps (Fig. 4.1).

The first step consists of acquiring and processing the necessary data describing the waste
heat sources and sinks: heat loads, temperature levels, availability over time, geographical
position, energy prices, etc. Several publications focused on the issue of data generation
(Persson and Werner [170], Brueckner et al. [33], Forman et al. [78], Persson and Münster
[169]), while Grönkvist and Sandberg [87], Chertow [40], Upham and Jones [222], Päivärinne

76
 4.2. Method

Figure 4.1: Global waste heat valorisation method

and Lindahl [173] assessed operational and financial barriers limiting the valorisation of waste
heat. The second step consists of detecting relevant potential connections between source and
sink and quantifying the transferable waste heat load, taking into account temperature level
matches, heat losses over distance, availability over time, and investment costs. The third step
consists of designing the network, which normally leads to a non-linear problem formulation
(Stijepovic et al. [208]). In particular, the heat exchanger network should be optimised from
a thermo-economic point of view, taking into account the trade-offs between number of
exchangers, heat exchange area and operating costs, as has been shown by the sequential
approach for multi-period heat exchanger network synthesis (Floudas et al. [76], Mian et al.
[141]). Additionally, the transportation network layout should be optimised by combining
common network segments (Wang et al. [236]), thus reducing the pipe investment costs. The
outcomes of the network optimisation are then used to replace the initial design data of the
previous step. The network design is finalised once the results of the second step cannot be
further improved.

The scope of the present work focuses on the second step described above, with the selection
of profitable connections between hot sources i and cold sinks j and the quantification of the
transferred heat load between them. The mathematical model is subdivided into two main
parts. First, the economic constraints covering the calculation of profits, revenues, as well as
investment and operating costs are given in Section 4.2.2. Second, the technical constraints
describing the energy cascade, the energy loss and the energy balance calculations used as
input for the investment and operating costs are indicated in Section 4.2.3.

Due to the important decrease in physical strength of common materials for operating tem-
peratures above 400°C (Turton et al. [218]), the application of the method is constrained to
streams below this temperature level. This limitation is also applied in the energy auditing tool
EINSTEIN (Brunner et al. [34]). Higher temperatures would imply more exotic (and therefore
more expensive) materials, which must be specifically accounted for in the economic viability
assessment. The method also focuses on the selection of waste heat transportation and not
distribution networks, as the latter are independent of the initial heat source. Finally, it is
assumed that electricity production from waste heat occurs on the source site.

To ease the equation readability, parameters are in normal font, while dependent and inde-

77
Chapter 4. Optimised regional waste heat valorisation

pendent variables are represented in italics. In terms of sets, hot streams i representing
industrial waste heat to be valorised are included in the set H S, while cold streams j are in the
set C S. As the proposed method is formulated as a multi-period problem, the time periods t
are included in the set T P . I T t represents the set of intervals of temperature k necessary for
the heat cascade formulation, which are dependent of period t . Finally, the set D N regroups
the pipes of standard diameter d n.

4.2.2 Economic constraints

[Link] Objective function and global economic constraints

The objective function maximises the annual profits obtained from electricity production P i
at source i , and from the supply of heat and electricity production from the utility P j of sink j
(Eq. 4.1).

X X
max Pi + Pj (4.1)
i ∈H S j ∈C S

Profits at source i are generated from the difference between revenues from the electricity
Op
production with a turbine R iEl,k,t
ec
, the combined operating costs C i ,k,t at each temperature
interval k and period t and the annualised investment cost C iI nv (Eq. 4.2). The definition of
the temperature intervals is given in the work of Papoulias and Grossmann [160].

Op
(R iEl,k,t
ec
−C i ,k,t ) −C iI nv
X X
Pi = ∀ i ∈ HS (4.2)
t =1,...T P k∈I T t

The annual profit P j of a sink j is obtained from the heat supply and electricity production
Op
revenues RHeat El ec
j,k,t and R j ,k,t , operating costs C j ,k,t of each temperature intervals k and period
t and the annualised investment costs C iI jnv related to the connection between source i and
sink j (Eq. 4.3).

Op
(RHeat El ec
C iI jnv
X X X
Pj = j,k,t + R j ,k,t −C j ,k,t ) − ∀ j ∈ CS (4.3)
t =1,...T P k∈I T t i ∈H S

[Link] Revenues

In case waste heat is used for electricity generation, assumed here to be implemented at source
i , the electricity revenues R iEl,k,t
ec
are determined from the related waste heat load Q̇ iEl,k,t
ec
obtained
from the heat load distribution (Section 4.2.3), the turbine efficiency η Turb , the duration dt of
Elec,s
period t and the electricity selling price pi,t specific for this type of valorisation (Eq. 4.4).

78
 4.2. Method

,s
Elec
Ri,k,t Elec
= Q̇i,k,t ∗ ηTurb ∗ dt ∗ piElec
,t
∀ j ∈ C S, k ∈ I T, t = 1, ...T P (4.4)

The heat supply revenues RHeat Dem

j,k,t are obtained from the heat demand Q̇j,k,t for the duration dt
considering a heat selling price pHeat
j ,t specific to sink j (Eq. 4.5). These revenues are considered
as independent of the primary energy form and are therefore expressed as parameter.

Heat,s
RHeat Dem
j,k,t = Q̇j,k,t ∗ dt ∗ pj,t ∀ j ∈ C S, k ∈ I T, t = 1, ...T P (4.5)

Finally, as the heating utility selection can also include combined heat and power (CHP)
plants, it is also necessary to include the electricity production revenues R El ec
j ,k,t
. In addition
,s
to the duration dt and electricity price pElec
j ,t , these revenues depend on the utility thermal
capacity Q̇ Cj ,k,t
HP
. Finally, the electricity conversion factor fjCHP
,t
calculating the electrical load
based on Q̇ Cj ,k,t
HP
specific to period t is also included, as CHP plants are composed of an internal
combustion engine (ICE) producing electricity during base load and a natural gas boiler for
peak loads (Eq. 4.6).

Elec,s
R El ec C HP
j ,k,t = Q̇ j ,k,t ∗ dt ∗ pj,t ∗ fjCHP
,t ∀ j ∈ C S, k ∈ I T, t = 1, ...T P (4.6)

[Link] Operating costs

Op
The operating costs C i ,k,t of electricity production at source i is proportional to the valorised
,b
waste heat load Q̇ iEl,k,t
ec
, the heat purchase price piHeat
,t
and the duration dt of t (Eq. 4.7).

Op ,b
C i ,k,t = Q̇ iEl,k,t
ec
∗ piHeat
,t
∗dt ∀ i ∈ H S, k ∈ I T, t = 1, ...T P (4.7)

Op
The operating costs C j ,k,t at j (Eq. 4.8) include the energy costs of transferred waste heat load
H ,d em
Q̇ iWj ,k,t
H
(the sum of waste heat load for heat demand Q̇ iWj ,k,t and waste heat load for loss
H ,l oss
compensation Q̇ iWj ,k,t , Eq. 4.9) from the various sources i in temperature intervals k and
the according waste heat source price pHeat
i ,t in period t . The operating cost also includes the
fuel costs obtained from the loads of the boiler Q̇ Bj ,k,t
o
, cogeneration unit Q̇ Cj ,k,t
HP
and heat pump
(HP) Q̇ H P
j ,k,t
, identical to the heating utility variable Q̇ HU
j ,k,t
generated from the energy balance
Elec,b
described in Section 4.2.3 (Eq. 4.31), multiplied with the energy prices pGas
j,t and pj,t , the
according efficiency rate ηBo CHP and ηHP as well as the duration d .
j , ηj j t

Op ,b Elec,b
(Q̇ iWj ,k,t
H
∗ piHeat ) + (Q̇ Bj ,k,t
o
∗ ηBo Gas HP HP
X
C j ,k,t = [ ,t j ∗pj,t ) + (Q̇ j ,k,t ∗ η j ∗pj,t )] ∗ dt
i ∈Hk
∀ i ∈ H S, j ∈ C S, k ∈ I T, t = 1, ...T P (4.8)

79
Chapter 4. Optimised regional waste heat valorisation

H ,d em H ,l oss
Q̇ iWj ,k,t
H
= Q̇ iWj ,k,t + Q̇ iWj ,k,t ∀ i ∈ H S, j ∈ C S, k ∈ I T, t = 1, ...T P (4.9)

[Link] Investment costs

As profitability is calculated on an annual basis, the investment cost I ej of equipment e must

be annualised to be comparable to the revenues and operational costs, considering the ex-
pected equipment lifetime ne and the annualisation interest rate r (Eq. 4.10). Considering
contingencies due to unforeseen circumstances, a correction factor α1 , usually 18% (Turton
et al. [218]), is also included.

e
e r(1 + r)n
e
C =I ∗ e ∗ (1 + α1 ) ∀ e = H E i , H E j , P i p, HU (4.10)
(1 + r)n − 1

The annualised investment costs C iI nv for electricity production at source i are only related to
the acquisition of a turbine, while the costs C jI nv for sink j are composed of the acquisition
H Ei HEj Pi p
costs C i j and C i j of the two heat exchangers at i and j , the transportation pipe cost C i j
and the heating utility costs C jHU (Eq. 4.11).

H Ei HEj Pi p
C jI nv = C i j +C i j +C i j +C jHU ∀ i ∈ H S, j ∈ C S (4.11)

Turbine investment costs The turbine investment costs should be calculated proportionally
to the sum of the waste heat load supplied for electricity production Q̇ iEl,k,t ec
. As described by
he
Turton et al. [218], the relation is not directly linear to a reference value Iref , but depends also
on a capacity exponent γ (typical value between 0.6 and 0.8). Maintaining linearity for this
problem requires treating this non-linear equation using piecewise linearisation as described
by e.g. Söderman and Ahtila [193].

!γ
Q̇ iEl,k,t
ec
ÃP P
t =1,...T P k∈I T t
I iTur b
= Iturb
ref ∗ ∀ i ∈ HS (4.12)
H
Q̇turb
ref

HE HEj
Heat exchanger investment costs The investment costs I i j i and I i j of the heat exchangers
of i and j for the connection i j should be calculated as a function of the heat exchanger area
(Turton et al. [218]). However, this would lead to a non-linear problem, as the area is obtained
from the division of the load by the logarithmic mean temperature difference, which are both
variables. The detailed design of the heat transfer network (especially the thermo-economic
optimisation of the heat exchanger area), and therefore the final calculation of the investment

80
 4.2. Method

costs, is out of scope of the present method. Therefore, it is proposed to estimate the invest-
HE HEj HE HEj
ment costs I i j i and I i j as a function of the highest load Q̇ i j i and Q̇ i j obtained from
the energy balance constraints, reference investment cost IHE HE
ref and load Q̇ref with the same
minimum temperature difference value assumed in the heat cascade constraints, and the
capacity exponent γ (Eq. 4.13 for heat exchanger in i ). This expression can then be piece-wise
linearised, as demonstrated for the case study in Appendix B.4.

 H Ei
γ
Q̇ i j
H Ei
Ii j = IHE
ref ∗
  ∀ i ∈ H S, j ∈ C S (4.13)
Q̇HE
ref

Pi p
Pipe investment costs The investment costs I i j of the transportation pipes is calculated
P i p,i nvest
considering the binary variable y i j ,d n for the selection of the adequate pipe diameter, the
underground type parameter fij Ground Ground
(fij = 1 for road, 0 for open fields), the specific pipe
Road Field
costs pdn and pdn (in e/m) related to standard pipe size and underground type and the
distance dij (Eq. 4.14).

Pi p X P i p,i nvest
Ii j = (y i j ,d n ∗(fijGround ∗ pRoad Ground
dn + (1 − fij ) ∗ pField
dn )) ∗ dij
dn
∀ i ∈ H S, j ∈ C S, d n ∈ D N (4.14)

P i p,i nvest
To exclude pipes with insufficient diameter, the binary variable y i j ,d n is declared to
describe the existence of the connection from i to j of size d n. The existence of the connection
of such a size is related to the highest mass flow ṁmax dn allowed in the pipe dn and the highest
mass flow ṁ i j of connection i j over all periods t as obtained from the heat cascade constraints
(Eq. 4.15-4.16).

X
ṁ i j ≥ ṁ i j ,k,t ∀ i ∈ H S, j ∈ C S, t = 1, ...T P (4.15)
k∈I T

P i p,i nvest
∗ ṁmax
X
y i j ,d n dn ≥ ṁ i j ∀ i ∈ H S, j ∈ C S, d n ∈ D N (4.16)
dn

Pi p
To avoid the selection of more than one diameter, the sum of the binaries y i j ,d n cannot exceed
one (Eq. 4.17).

X P i p,i nvest
y i j ,d n ≤1 ∀ i ∈ H S, j ∈ C S, d n ∈ D N (4.17)
dn

81
Chapter 4. Optimised regional waste heat valorisation

Heating utility investment costs The investment cost I jHU of the heating utility, currently a
natural gas boiler, a heat pump or a CHP unit with peak boiler, is determined according to the
binary variables y Bj o , y Cj H P and y H
j
P
selecting the technology type with the investment prices
C

IBo CHP and IHP specific to each sink (Eq. 4.18).

j , Ij j

I jHU = y Bj o ∗ IBo C HP
j + yj ∗ ICHP
j + yH
j
P
∗ IHP
j ∀ j ∈ CS (4.18)

The investment costs IBo CHP and IHP are input parameters determined by considering the
j , Ij j
highest load demand Q̇Demj,max . Concerning CHP plants, the investment cost ICHP
j includes the
cost of internal combustion engines designed to meet a certain heating base load determined
by the user as well as the peak boilers required to reach Q̇Dem
j,max (Eq. 4.19).

ICHP
j = IICE
j + IBo
j ∀ j ∈ CS (4.19)

Existing central heating utilities are reflected by setting the investment costs IBo CHP and IHP
j , Ij j
to zero.

With Eq. 4.20, only one technology can be selected at one sink.

y Bj o + y Cj H P + y H
j
P
=1 ∀ j ∈ CS (4.20)

The relation between these variables and the utility heat loads Q̇ Bj o , Q̇ Cj H P and Q̇ H
j
P
provided
to sink j are defined with M a large value (Eq. 4.21-4.23).

Q̇ Bj o − y Bj o ∗ M ≤ 1 ∀ j ∈ CS (4.21)

Q̇ Cj H P − y Cj H P ∗ M ≤ 1 ∀ j ∈ CS (4.22)

Q̇ H
j
P
− yH
j
P
∗M ≤ 1 ∀ j ∈ CS (4.23)

The utility heat loads Q̇ Bj o , Q̇ Cj H P and Q̇ H

j
P
are the highest load of all periods t , obtained from
the sum of the loads over all temperature intervals k constrained by the operating costs of Eq.
4.8 (Eq. 4.24-4.26).

Q̇ Bj o ≥ Q̇ Bj ,k,t
o
X
∀ j ∈ C S, t = 1, ...T P (4.24)
k∈I T t

82
 4.2. Method

Q̇ Cj H P ≥ Q̇ Cj ,k,t
HP
X
∀ j ∈ C S, t = 1, ...T P (4.25)
k∈I T t

Q̇ H P
Q̇ H P
X
j ≥ j ,k,t ∀ j ∈ C S, t = 1, ...T P (4.26)
k∈I T t

4.2.3 Energy constraints

The geoallocated hot and cold streams are characterised by their mass flow rate ṁi,t , ṁj,t ,
inlet enthalpy HIn In Out Out
i,t , Hj,t , outlet enthalpy Hi,t , Hj,t , heat capacity cpi,t , cpj,t , corrected inlet
temperature Ti,t , Tj,t and corrected outlet temperature TOut
In In Out
i,t , Tj,t for period t . Identically to
the pinch analysis method, the temperatures are corrected in order to consider the minimum
temperature difference d Tmi n of the heat exchanger: hot stream temperatures are decreased
by d Tmi n/2, while cold stream temperatures are increased by d Tmi n/2.

[Link] Heat cascade formulation

The energy balance of a source i and a sink j in the temperature intervals k at period t is
represented in Fig. 4.2.

For each source i in temperature interval k and period t , the energy balance is composed of
the available waste heat load Q̇Avail
i,k,t , the waste heat cascaded from the interval above R i ,k−1,t
and to the one below R i ,k,t , the waste heat valorised for electricity production Q̇ iEl,k,t
ec
, the
H ,d em
load of the cooling utility Q̇ iCU
,k,t
, the transferred waste heat Q̇ iWj ,k,t for the demand of the
H ,l oss
various sinks j and finally the transferred waste heat Q̇ iWj ,k,t to compensate the temperature
H ,d em
decrease related to the losses Q̇ iLoss
j ,k,t
of Q̇ iWj ,k+1,t (Eq. 4.27).

W H ,d em H ,l oss
R i ,k−1,t + Q̇Avail j ∈C S (Q̇ i j ,k,t + Q̇ iWj ,k+1,t ) − Q̇ iEl,k,t
ec
− Q̇ iCU − R i ,k,t = 0
P
i , k, t − ,k,t
∀ i ∈ H S, k ∈ I T, t = 1, ...T P (4.27)

The heating demand Q̇Dem

j,k,t of sink j in temperature interval k and period t is covered by the
H ,d em
sum of the transferred waste heat Q̇ iWj ,k,t and, in case waste heat is not available or too
,d em
expensive, the heating utility Q̇ HU
j ,k,t
(Eq. 4.28).

H ,d em ,d em
Q̇Dem Q̇ iWj ,k,t − Q̇ HU
X
j , k, t − j ,k,t
=0 ∀ j ∈ C S, k ∈ I T, t = 1, ...T P (4.28)
i ∈H S

83
Chapter 4. Optimised regional waste heat valorisation

Figure 4.2: Regional heat cascade

H ,l oss
The heat losses Q̇ iLoss
j ,k,t
of connection i j are compensated by waste heat Q̇ iWj ,k−1,t from the
temperature interval k − 1 above, in order to reach the upper temperature level of k, and, if
,l oss
necessary, by the heating utility Q̇ iHU
j ,k,t
(Eq. 4.29).

W H ,l oss HU ,l oss
Q̇ iLoss
j ,k,t − Q̇ i j ,k−1,t − Q̇ i j ,k,t =0 ∀ i ∈ H S, j ∈ C S, k ∈ I T, t = 1, ...T P (4.29)

The selection of the hot utility technology is completed with the sum of the load per technology
type, thus making the link with the Equations 4.24-4.26 of the investment costs constraints
(Eq. 4.30-4.31).

HU ,d em ,l oss
Q̇ HU Q̇ iHU
X
j ,k,t = Q̇ j ,k,t + j ,k,t
∀ j ∈ C S, k ∈ I T, t = 1, ...T P (4.30)
i ∈H S

Q̇ HU Bo C HP HP
j ,k,t = Q̇ j ,k,t + Q̇ j ,k,t + Q̇ j ,k,t ∀ j ∈ C S, k ∈ I T, t = 1, ...T P (4.31)

84
 4.2. Method

[Link] Heat losses

The heat losses Q̇ iLoss

j ,k,t
are calculated as the product of the distance dij and the specific heat
losses per pipe length q̇ iLoss
j ,k,t
for connection i j in interval k and period t , thus reflecting the
varying outdoor temperature over time (Eq. 4.32).

Q̇ iLoss Loss
j ,k,t = dij ∗ q̇ i j ,k,t ∀ i ∈ H S, j ∈ C S, k ∈ I T, t = 1, ...T P (4.32)

As a consequence of the energy balance formulations of Eq. 4.27 and Eq. 4.29, the heat losses
q̇ iLoss
j ,k,t
are determined for each temperature interval k instead of considering the final pipe
P i p,i nvest
diameter calculated for the investment costs with y i j ,d n . Therefore, the losses q̇ iLoss
j ,k,t
are
P i p,l oss
calculated using a binary variable y i j ,d n,k,t selecting the pipe diameter for connection i j in
interval k and period t and the specific heat losses q̇Loss
dn,k,t for the diameter d n (Eq. 4.33).

P i p,l oss
q̇ iLoss (y i j ,d n,k,t ∗ q̇Loss
X
j ,k,t = dn,k,t ) ∀ i ∈ H S, j ∈ C S, k ∈ I T, t = 1, ...T P (4.33)
d n∈D N

P i p,l oss
The binary variable y i j ,d n,k,t must be constrained to exclude diameters where the maximum
permitted mass flow ṁmax dn is smaller than the actual mass flow ṁ i j ,k,t of connection i j , in
interval k and period t (Eq. 4.34). The smallest of the permitted diameters is automatically
selected as it leads to the lowest heat losses and investment costs.

P i p,l oss
(y i j ,d n,k,t ∗ ṁmax
X
ṁ i j ,k,t ≤ dn ) ∀ i ∈ H S, j ∈ C S, k ∈ I T, t = 1, ...T P (4.34)
d n∈D N

To ensure that only one diameter is selected, Eq. 4.33 must be further constrained (Eq. 4.35).

X P i p,l oss
y i j ,d n,k,t ≤ 1 ∀ i ∈ H S, j ∈ C S, k ∈ I T, t = 1, ...T P (4.35)
d n∈D N

The variable mass flow ṁ i j ,k,t of connection i j is obtained from the transferred waste heat
H ,d em
Q̇ iWj ,k,t from Eq. 4.27 and Eq. 4.28 as well as the enthalpy or the heat capacity and upper
Up
and lower temperatures Tk,t , TLow
k,t of interval k at period t (Eq. 4.36). This approach maintains
linearity of the formulation, as the temperature difference of the interval k is a parameter. The
use of the actual temperature difference of connection i j aggregated over all temperature
intervals k would lead to non-linearity and thus must be treated in this way. The decision
to formulate the method as MILP thus implies an overestimation of the heat losses for a
connection across several temperature intervals.

85
Chapter 4. Optimised regional waste heat valorisation

H ,d em
Q̇ iWj ,k,t
ṁ i j ,k,t = ∀ i ∈ H S, j ∈ C S, k ∈ I T, t = 1, ...T P (4.36)
cpk,t ∗ (TUp
k, t
− TLow
k, t
)

The parameter q̇Loss

dn,k,t is calculated with the equation given by Çomakli et al. [155] (Eq. 4.37).
Out
Tt is the average outdoor temperature for period t . κi is the thermal conductivity of the
pipe insulation material (in W/m*K), while rp and rins are the internal and external, insulated,
pipe radius, respectively.

Up
Tk,t +TLow
2
k, t
− TOut
t
q̇Loss
dn,k,t = 2∗ 1 rins ∀ d n ∈ D N , k ∈ I T, t = 1, ...T P (4.37)
2∗π∗κi ∗ln( rp )

4.3 Case study

The proposed method is applied to the south-western region of the Grand-Duchy of Lux-
embourg (Western Europe). The case study encompasses two steel production sites as heat
sources (star symbol in Fig. 4.3). As potential heat sinks, three industries (triangle symbol) in
the sectors of asphalt, hard metals and plastic manufacturing as well as the residential users
of the City of Esch-sur-Alzette (second-largest city in the country) and eight nearby towns
(circle symbols) are considered, comprising 62’540 inhabitants, an area of 73.5 km2 and a
population density of 851 inhabitants/km2 . For comparison, the metropolitan area of Porto
(Portugal) in 2014 had a density of 851 inhabitants/km2 and the area of Zürich (Switzerland)
865 inhabitants/km2 , while the European average was 117 inhabitants/km2 (Eurostat [68]).

4.3.1 Input data description

[Link] General data

The case study is built considering energy prices in Luxembourg from 2015. The data is organ-
ized in five mutually exclusive periods covering in total 8’760 hours, with t1 = 1hr, t2 = 6’036hrs,
t3 = 168hrs, t4 = 2’381hrs and t5 =168hrs. t1 represents the coldest day of the year, t2 and t3
the intermediate period where space heating (Tout t ≤ 15°C) and domestic hot water (DHW)
are required, and t4 and t5 the periods where only DHW heating is used by the residential
users. Waste heat is available for t1 , t2 and t4 , while one week production interruptions of
the steel plants are assumed in t3 and t5 . The average outdoor temperatures for the differ-
ent periods are Tt1 = −7°C, Tt2 ,t3 = 9°C, Tt4 ,t5 = 15°C. The transportation distance between
sources and sinks, within 4 km and thus supposedly viable (Ammar et al. [5]), is obtained
with the line measurement tool of QGIS Development Team [178], by manually drawing a
connection between the source and the closest building of the considered sink. By overlaying

86
 4.3. Case study

Figure 4.3: Regional waste heat sources and sinks

OpenStreetMap ([Link]) terrain information, data on the ground type (road

or field) is obtained (Appendix B.1).

The problem, comprising 64’764 variables, is formulated in AMPL (Fourer et al. [79]) and
solved with CPLEX (IBM [102]) on an Intel Core 2 Duo T9400 2.53 GHz Processor and 8 GB
RAM.

[Link] Process data

The hot streams of the steel plant sites, representing an energy quantity of 1’297 GWh/a, are
summarised in Tab. 4.1. The selection of the process stream types and the characterisation of
their physical properties are based on Tarres-Font [213], who assessed the site of Differdange
in detail. The author mentioned that waste heat recovery from the Electric Arc Furnace (EAF)
off-gases after post-combustion and from the water jacket is technically difficult due to the
dust separation system limiting the formation of dioxin but also constraining the temperature
levels of the hot streams. These heat sources are nevertheless considered here, as they present
extremely high load and temperature levels, thus potentially generating high enough revenues
to justify a modification of the processes. The inlet temperature levels were set to a maximum
of 400°C. An initial cooling end temperature of 40°C is considered, to which a dTmin of 10K
is applied to obtain a final system temperature of 50°C. The loads of the site of Belval were
proportionally adapted as a function of the production mass flow estimated from nominal
data of the Integrated Pollution and Prevention Control (IPPC) authorisation, considering a

87
Chapter 4. Optimised regional waste heat valorisation

24 hours, 6 days/week production schedule and a correction factor of 75% to reflect reduced
production activity. All sources are assumed to be located at the same place in the plants.

Table 4.1: Heat source data

Site Stream Load [kW] Temp. Tin Temp.

[°C] Tout [°C]
EAF off gases after post 43’750 400 50
combustion
Steel_
Differdange EAF off gases end of water jacket 40’625 400 50
EAF off gases end after 7’500 400 50
quenching
WBF 1 off gases 3’250 400 50
WBF 2 off gases 4’125 400 50
EAF off gases after post 24’306 400 50
combustion
Steel_Belval EAF off gases end of water jacket 13’542 400 50
EAF off gases end after 4’167 400 50
quenching
WBF 1 off gases 1’806 400 50
WBF 2 off gases 2’292 400 50

The main thermal streams of the three industrial sink sites, totaling 74 GWh/a of energy
demand, are described in Tab. 4.2. According to the heat cascade approach, the temperatures
were corrected by a temperature difference of 10K to consider the temperature difference from
the heat exchanger, except for the streams of the non-ferrous manufacturing plant, which are
set to 400°C. The temperatures and specific heat demands are taken from Stadler [198], based
on data from Brown et al. [32], Peinado et al. [164] for the asphalt industry, Vlachopoulos and
Wagner [232], Wen [238] for the plastic manufacturing industry and Zawrah [243], NIST [149]
for the hard metal manufacturing plants. The mass flow data is based on information from
the IPPC authorisation of the asphalt plant and estimated as 1’000kg/h and 5’000 kg/h for the
two other industries, considering a 12 hours, 5 days/week production schedule.

The residential heating demand data (Tab. 4.3) is generated using the regression analysis
information of Bertrand et al. [19], combined with georeferenced data indicating residential
building type and location (Entreprise des Postes et Télécommunications [56], Service des
travaux municipaux [194]) as well as averaged building age, inhabitant and household number
data from the 2011 population census (Statec [199]). The total energy demand amounts to
455 GWh/a. The temperature levels are those of the district heating systems of the city of
Luxembourg (70/95°C), to which a correction temperature difference of 10K is added. Non-
residential buildings are not included, as the necessary data is not available.

88
 4.3. Case study

Table 4.2: Industrial heat sink data

Site Stream Load [kW] Temp. Tin Temp.

[°C] Tout [°C]
Asphalt Aggregate heating 5’079 30.0 185.0
manufacturing
Evaporation 2’521 109.9 110.1
Heating 65 30.0 174.9
Plastic
manufacturing Melting 65 174.9 175.1
Superheating 26 175.1 220.0
Hard metal Oxide reduction heating 417 30.0 400.0
manufacturing
Carbonisation heating 102 30.0 400.0

Table 4.3: Urban heat sink data

Site Stream Load [kW], Temp. Tin Temp.

t 1 / t2 & t 3 / t 4 & t 5 [°C] Tout [°C]
Belvaux Space heating and 23’380 / 7’329 / 503 80 105
DHW
Differdange Space heating and 34’499 / 11’006 / 1’015 80 105
DHW
Ehlerange Space heating and 3’850 / 1’196 / 67 80 105
DHW
Esch-sur- Space heating and 81’981 / 26’333 / 2’714 80 105
Alzette DHW
Mondercange Space heating and 15’072 / 4’688 / 272 80 105
DHW
Niederkorn Space heating and 19’023 / 6’022 / 493 80 105
DHW
Oberkorn Space heating and 12’734 / 4’014 / 305 80 105
DHW
Sanem Space heating and 11’203 / 3’487 / 206 80 105
DHW
Soleuvre Space heating and 21’456 / 6’703 / 430 80 105
DHW

The CHP units are composed of an internal combustion engine and a peak boiler. The ICE load
corresponds to the sink load at period t 2 and t 3 , while the boiler load make up the remaining
power demand at t 1 . With a thermal efficiency of 47% and an electrical efficiency of 39%
(ASUE [7]), the electricity conversion factor fjCHP
,t
necessary for the revenue calculations of

89
Chapter 4. Optimised regional waste heat valorisation

Eq. 4.6 is set to 26% for t1 and 83% for the other periods, thus reflecting the use of the ICE as
base-load unit producing electricity, while the remaining heating demand is covered by the
natural gas boiler. The efficiency of the electricity production turbine on the steel plant sites is
considered to reach 24% (Tarres-Font [213]). The boiler is attributed an efficiency of 90%. The
efficiency of the heat pumps are calculated considering a Carnot efficiency of 55% (Becker
et al. [13]), an average outdoor temperature of -1°C and the uncorrected process temperatures
indicated in Tab. 4.2 and Tab. 4.3.

The pipe internal and external diameters, maximum mass flows per size and heat transfer rate
ki (0.0255 W/m*K) are taken from the design manual of a pipe manufacturer (Isoplus [105]
- Appendix B.2). To cover the full range of temperature intervals of this case study, the heat
transfer fluid considered is Dowtherm A thermal oil, which can be used at temperatures as low
as 12°C and up to 425°C which makes it an acceptable fluid for this study. The heat capacity
data, given in Appendix B.3, is taken from Dow Chemical Company [50].

The expected lifetime of heating utilities and heat exchangers are assumed to be 15 years and
30 years for the pipes.

[Link] Economic data

For the annualisation calculation, an interest rate r of 2% is considered (2015 average interest
rate for Luxembourgish non financial corporations, [Link]). Cost data not from 2015 are
recalculated for the reference year using the correction indexing factor from the Chemical
Engineering Plant Cost Index ([Link]

For a 1’090 kW boiler, Buderus [35] indicates a price of 76’668 CHF, excluding VAT (17% in
Luxembourg). A 2015 average exchange rate of 0.94 e/CHF is considered. A specific price
of 300 e/kW for large heat pumps is taken from Boissavy [26]. Concerning ICEs price, ASUE
[7] provided for 2011 a relation between electrical power and specific price in e/kWel ec (Eq.
4.38). By multiplying the thermal load of the ICE with the electricity conversion factor fjCHP
,t
,
the specific price related to the thermal load at period t is obtained.

pICE = 9′ 332.6 ∗ (PElec )−0.4611 (4.38)

The specific pipe costs pRoad

dn and pField
dn , listed in Appendix B.2, are taken from Nussbaumer
and Thalmann [151]. The heat exchanger costs are based on data from Tarres-Font [213] and
were recalculated for the assumed minimum temperature difference of dTmin=10K. The curve
has been piece-wise linearised in 5 segments, and is represented in Appendix B.4.

Concerning the heat selling price, natural gas prices of 48.50 e/MWh for the domestic sector
and 38.19 e/MWh for industrial sinks are assumed (Statec [203]). One of the largest district
heating systems indicates that its tariff, which is aligned with the largest heat supplier of the

90
 4.3. Case study

country, will not be above that of an autonomous heat production ([Link]/distribution,

in French). The values for natural gas to provide the heat are therefore used as a proxy
which would be what sinks would be paying otherwise. The gas price for municipal boiler
and CHP plants, 43.36 e/MWh, is an average of residential and industrial gas prices. The
electricity purchase price for the heat pumps, 146.75 e/MWh, is an average of residential and
industrial prices with an additional 25% for grid costs (Statec [204]). These energy prices have
been confirmed by local actors (energy providers and a municipality). The selling price of
internal combustion engines is set at 74 e/MWh (Luxemburgish Parliament [127]). Electricity
production from waste heat valorisation is not currently addressed by national regulations.
The price would have to follow market trends, around 30 e/MWh ([Link]
which is not sufficient to be viable. Considering that this electricity generation would have a
lower impact on the environment than the current gas and coal-based production, reduce
exergy losses, improve the overall energy efficiency of the region and be provided by a local
source, it is therefore argued that it should be considered as a form of renewable energy. As
reference, the price for hydroelectric power plants of 125 e/MWh (Luxemburgish Parliament
[129]) is used. This is an intermediate price when considering various renewable energy
sources (wind, photovoltaic, biomass use, etc.) and it is assumed that the production profile
of hydropower and waste heat electricity production are similar in terms of stability and
prediction, compared to solar or wind energy.

As raised by Svensson et al. [210] and Morandin et al. [143], the waste heat selling price depends
on several factors (e.g. minimal and maximal prices, generated revenues for the source, etc.).
Therefore, the optimisation scenarios are considered with values ranging from 0 to 30 e/MWh
in steps of 5 e.

4.3.2 Results

[Link] Regional valorisation considering different waste heat selling prices

The profits generated by the ESCo for the supply of heat from industrial waste heat sources and
CHP units as well as from production of electricity, the revenues of the steel manufacturing
company, the quantity of valorised waste heat and the produced electricity under various
waste heat selling prices are represented in Fig. 4.4.

Given the Luxembourgish context, if the steel plants were to provide their waste heat at no
cost, the profits of the ESCo would reach a maximum of 25 Me, with more than 1’000 GWh of
waste heat valorised for electricity production (210 GWh) and for the heating demand of all
the considered towns (441 GWh). Due to their constant heat demand over the year, it is more
economically attractive to install CHP plants for the industrial sinks (additional electricity
production of 60 GWh). The same valorisation configuration is obtained with a waste heat
price of 5 e/MWh, with the ESCo receiving 20 Me and the steel plants revenues reaching 5
Me. With a waste heat price above 5 e/MWh, the electricity production by turbine is no more

91
Chapter 4. Optimised regional waste heat valorisation

Figure 4.4: Profits, revenues and energy quantities in function of waste heat prices

profitable, an outcome also observed by Tveit et al. [219]. At 10 e/MWh, all towns are still
sustained with waste heat, while connections to Ehlerange and Mondercange are not viable at
a price above 10 e/MWh, Sanem above a price of 15 e/MWh and Belvaux above a price of
20 e/MWh. The heat demand is then covered by CHP plants, again increasing the electricity
production in the region. Natural gas boilers and heat pumps are not retained as heating
technology when no WH is valorised.

With a price of 25 e/MWh, and although the steel plants are supplying less heat (338.3 GWh to
cover heat demand, 1.4 GWh for the heat losses), their revenues are reaching their maximum
at 8.5 Me, while 10.7 Me would be obtained by the ESCo. These profits are allocated as
70% from waste heat transportation and to 30% from the CHP plants. Above this WH price,
CHP plants are economically more viable in this case study and the steel plants would not
generate any revenues. The profits of the ESCo would also reach their lowest value with 9.9
Me. Detailed data on revenues, investment costs, payback time, valorised heat and energy
losses at this waste heat price is provided in the next section.

[Link] Regional waste heat valorisation at 25 e/MWh

The resulting waste heat transportation networks and selected utilities, considering a waste
heat price of 25 e/MWh, are represented in Fig 4.5.

The total investment costs for heat exchange and transport, the operating costs, the revenues
from heat supply and electricity production as well as the payback time are indicated in Tab.
4.4 in detail for the waste heat valorisation sinks and in aggregated form for the remaining

92
 4.3. Case study

Figure 4.5: Waste heat valorisation and selected utility technologies for a waste heat price of
25 e/MWh

sites. The investment costs for the waste heat valorisation equipments are given in Tab. 4.5.
The payback time reaches 3 years for waste heat valorisation in nearby sinks (Differdange,
Esch-sur-Alzette), and increases up to 5.6 years for more distant sinks. With an average of 2.4
years, it should be mentioned that, under the prices considered above, the actual payback
time of the CHP plants strongly varies according to the amount of heat delivered, with values
ranging between 1 and 13 years.

Table 4.4: Revenues and costs

Sink [-] Total in- Yearly Heat Electricity Payback

vestment operating supply production time [a]
costs [Me] costs revenues revenues
[Me/a] [Me/a] [Me/a]

Differdange 5.9 1.8 3.8 - 3.0

Esch-sur-Alzette 14.4 4.4 9.2 - 3.0
Niederkorn 4.6 1.0 2.1 - 4.2
Oberkorn 3.9 0.7 1.4 - 5.4
Soleuvre 6.7 1.1 2.3 - 5.6
Sum of sinks with 9.9 15.7 8.8 11.0 2.4
CHPs

93
Chapter 4. Optimised regional waste heat valorisation

Table 4.5: Investment costs

Sink [-] Total in- Utility Pipe [Me] Heat Heat

vestment [Me] exchanger exchanger
costs [Me] at source at sink
[Me] [Me]
Differdange 5.9 2.5 0.6 1.4 1.4
Esch-sur-Alzette 14.4 6.0 2.3 3.0 3.0
Niederkorn 4.6 1.4 1.4 0.9 0.9
Oberkorn 3.9 1.0 1.6 0.7 0.7
Soleuvre 6.7 1.6 3.2 1.0 1.0

The absolute and relative (transfer to demand ratio, in parenthesis) waste heat loads for time
periods t1 , t2 and t4 , as well as the pipe diameter per connection are given in Tab. 4.6. Instead
of selecting the pipe with the largest diameter to cover the highest load demand in t1 , the
diameter related to the load of period t2 is selected to avoid excessive investment costs for the
short duration of t1 . Due to the important heat demand, the heat supplied to Esch-sur-Alzette
is provided by two different processes, the off-gases after post-combustion and the off-gases
of the cooling water jacket.

Table 4.6: Transferred waste heat loads

Sink [-] Source [-] Load at Load at Load at Pipe type

t1 t2 t4 [-]
[kW / %] [kW / %] [kW / %]

Esch-sur- Steel_Belval - EAF off gases 22’194 22’199 2’714 DN500

Alzette after post combustion (27%) (84%) (100%)
Steel_Belval - EAF off gases 4’130 4’134 - DN250
end of water jacket (5%) (16%)
Differdange Steel_Differdange - EAF off 11’004 11’006 1’015 DN400
gases end of water jacket (32%) (100%) (100%)
Niederkorn Steel_Differdange - EAF off 6’016 6’022 493 DN300
gases end after quenching (32%) (100%) (100%)
Oberkorn Steel_Differdange - EAF off 4’007 4’014 305 DN250
gases after post combustion (31%) (100%) (100%)
Soleuvre Steel_Differdange - EAF off 6’690 6’703 430 DN300
gases after post combustion (31%) (100%) (100%)

The transportation heat losses, in absolute value and as percentage of the transferred load, are
summarised in Tab. 4.7. They range between 43 and 55 W/m, which is in line with studies on
specific pipe losses (Boehm [25]: 60 W/m, Perpar et al. [165]: 49 - 58 W/m). The heat losses are

94
 4.3. Case study

compensated in this case study by additional waste heat taken from the upper temperature
interval and not from the heating utility. The remaining gas consumption, due to the design of
the pipe size and the production interruptions in t3 and t5 are given in the same table. They
amount to 3% of the total energy demand for heating.

Table 4.7: Heat losses

Sink [-] Source [-] Load at Load at Load at Remaining

t1 t2 t4 gas con-
[kW] [kW] [kW] sumption
[MWh]

Esch-sur- Steel_Belval - EAF off gases 28 24 18

4’936
Alzette after post combustion (0.1%) (0.1%) (0.7%)
Steel_Belval - EAF off gases 23 19 -
end of water jacket (0.6%) (0.5%)
Differdange Steel_Differdange - EAF off 13 11 8 (0.8%) 2’043
gases end of water jacket (0.1%) (0.1%)
Niederkorn Steel_Differdange - EAF off 33 28 21 1’108
gases end after quenching (0.6%) (0.5%) (4.3%)
Oberkorn Steel_Differdange - EAF off 40 33 24 734
gases after post combustion (1.0%) (0.8%) (7.8%)
Soleuvre Steel_Differdange - EAF off 80 67 51 1’213
gases after post combustion (1.2%) (1.0%) (12.0%)

For this case study, the valorisation of the waste heat does not need further complex network
design and optimisation efforts, as the measures are rather straightforward. In the case of
waste heat valorisation from the steel plant in Differdange, several optimisation opportunities
arise. First, considering the size of the plant, it should be checked with the steel plant manager
if one of the sources with a load above 6’022 kW, the heating demand of Niederkorn, is closer to
the town than the clustering point assumed. This would allow for a reduction of the pipe costs
(20% of the annualised costs). Second, due to the transportation centralised in Differdange,
there is the opportunity to use one common heating utility instead of four separate ones.
This would further reduce the specific investment costs of the utility (36% of the annualised
costs). Finally, the optimisation method selected various processes as heat sources, although
either the loads of the EAF off-gases after post-combustion or at the end of the water jacket
are sufficiently high to cover the heating demand of the four towns. The costs of the heat
exchangers at the source (22% of the annualised costs) would then be reduced. The waste
heat valorisation in Esch-sur-Alzette can be improved by joining the two heat sources of
the steel plant at the level of the transportation pipes, reducing the pipe investment costs.
However, before the detailed network layouts and utility selection is finalised, the design
selection should be further discussed with the local actors to ensure the inclusion of potential
constraints influencing the layout of the networks (e.g. ease of access of the heat sources)

95
Chapter 4. Optimised regional waste heat valorisation

or additional opportunities (e.g. the existence of service tunnels from the steel plants to the
towns).

The low payback times, relevant profits for the ESCo and important revenues for the waste heat
transport demonstrate the economic relevance and competitiveness of regional waste heat
valorisation for the Luxembourgish domestic sector in 2015, given an adequate heat price. It is
also relevant to note that, while 41% of the available waste heat could theoretically be valorised
to cover the considered heat demand, only 26% of the excess heat can be economically
valorised when constraints of fuel prices, investment costs, energy losses, etc. are included.
By considering aspects of economic viability, as well as real population and energy demand
densities, this method generates more realistic estimations than other large scale waste heat
potential assessments (e.g. Persson and Werner [170]).

4.4 Discussion

A multi-period, MILP-based, method for the optimal regional valorisation of industrial waste
heat by ESCos is proposed in this work. It maximises the profits stemming from electricity
production and the supply of heat, while considering infrastructure investment and operating
costs.

One of the main strengths of the proposed method is the integrated approach to waste heat
valorisation. It does not focus solely on the actual WH valorisation, but also includes the
possibility of electricity production as well as the optimal selection of the heating utility type
necessary as a backup or main heating equipment. Another major contribution is that the
formulation allows the consideration of a potential closure of the heat source plant within
a few years. Instead of using the typical lifetime of the technical equipment, the lifetime n e
of the annualisation factor of Eq. 4.10 can be used to calculate the investment costs over the
prospective remaining lifetime of the plant. Potential investors have the opportunity to assess
the annualised investment costs compared to the waste heat price set by the industry, and
either require a supply guarantee or a decrease of the waste heat price, thus reducing the
investment risk. Finally, while the targeted users are mainly ESCos, potential waste heat supply
sites can also use this approach to determine the optimal heat selling price generating the
highest revenues. This aspect is of particular importance in case the waste heat valorisation
implies changes on manufacturing processes. Normally, industries are resistant to such
changes but sufficiently high revenues might prove to be adequate incentive for process
modification.

A drawback of the optimisation formulation is that the global heat losses over several temper-
ature intervals k are overestimated, as these are determined independently for each interval.
Calculating the losses over all intervals would have led to a non-linear problem, which the
authors made efforts to avoid to take advantage of desirable characteristics of MILP formula-
tions (e.g. generation of a global solution). Another current shortcoming is the temperature

96
 4.4. Discussion

limitation to 400°C reflecting material constraints, thus excluding high-temperature waste

heat valorisation. This could be avoided by including material type selection as a variable in
the problem, although this would further increase its complexity. Finally, the fact that several
heat sources or sinks can be situated at the same location is not reflected in the method,
although it would potentially reduce utility, heat exchanger and pipe investment costs. This
decision is intentional, as the sharing of equipment depends on the temperature level of
sources and sinks and mixing of streams cannot be automatically assumed. This issue is,
however, solved in the network design and optimisation stage which would be completed in
more detail based on the analysis presented here.

The main significance of this work lies in the development of a new method for energy service
companies focusing on the optimal valorisation of regional waste heat and the simultaneous
selection of the best heating utility technology. The optimisation method expressed here leads
to two main contributions in the field of integrated large-scale energy supply networks.

First, a novel waste heat valorisation problem is formulated for ESCos instead of industries
or municipalities. The outcomes differ from similar works insofar as the configuration selec-
tion is based on highest profitability instead of lowest operating and investment costs. The
method also considers that competing sinks can have different energy prices, an aspect rarely
considered which, however, influences the selection of the valorisation opportunities.

Second, the method improves regional waste heat valorisation methods by also considering op-
timal heating technology and standard pipe diameter selection. With this integrated approach,
the solutions regarding energy provision proposed by the ESCo are more thoroughly assessed,
as potential interactions between these aspects are taken into account simultaneously.

97
5 Conclusions and future developments

5.1 Conclusions
Considering current scientific shortcomings in the field of waste heat recovery, this thesis
focused on the modelling of domestic hot and waste water streams and on the assessment
and optimisation of building, urban and regional excess heat valorisation.

First, a method characterising the energy-related parameters of domestic hot water streams
(shower, bath, washing up, etc.) in households at building and urban scale, together with the
necessary input data, was presented. The DHW energy demand of hotels and nursing homes
was also addressed. The proposed model provided precise data on energy consumption and
temperature level requirements for each stream at the building level, taking into account the
number of inhabitants and households, the end-use occurrence, as well as use frequency
and duration, based on data from the water domain. The related energy consumption can
be put in relation with the total heating demand of the building and, by data aggregation,
of a district, a city or a region. It was shown in a real case study that the global outcomes
(e.g. specific hot water demand per capita) were in line with similar works. Furthermore it
was confirmed that domestic hot water currently only plays a minor role in terms of energy
consumption at urban scale, i.e. 12.3% of the residential heat demand of the city considered
in the case study. However, with the construction of near-zero buildings and the improvement
of the thermal envelope of existing ones, it was also highlighted that domestic hot water will
represent between 30 and 50% of the residential building heating demand in the future.

Considering these findings, measures to improve the energy efficiency related to domestic hot
water use must be addressed. Residential waste water streams were therefore characterised
based on the method proposed above. Using pinch analysis, six different in-building waste
water heat recovery configurations were assessed as to their energy savings. Investment cost
calculations were also included in the method. As a demonstration, the proposed charac-
terisation and assessment approaches were deployed to the buildings and districts of the
first case study, considering shower and grey water heat recovery systems. With the energy

99
Chapter 5. Conclusions and future developments

prices of 2015 leading to payback times generally above 10 years, such solutions are barely
economically viable in existing residential buildings in Luxembourg. The heating savings in
existing buildings ranged between 1 and 12%, while in high efficiency residential buildings
these savings ranged between 6 and 22%. An integrated approach, combining WWHR, hot
water temperature optimisation and heat pump deployment, led to much more relevant
heating savings ranging between 28 and 41% in high efficiency single family and multifamily
buildings, respectively, thus demonstrating the relevance of such a holistic method.

With the detailed domestic hot water demand models addressed above, the related energy
demand was calculated to be 17% lower than the results obtained from a surface-related
approach. The characterisation and spatial allocation of urban system heat demand data were
therefore improved with the proposed DHW characterisation method. This is a relevant step in
the decision-taking process for regional waste heat valorisation, as it reduces the risks related
to uncertain and incomplete data, particularly when investors, like energy service companies,
are concerned. An optimisation method maximising profits and based on a mixed integer
linear programming model was proposed for this type of actor. The model took into account
the specific energy price of the heat sinks, the distances between sources and potential users,
the energy losses due to heat transportation as well as the related investment and operating
costs. Due to the integrated approach, the proposed model also included the optimal selection
of the backup heating utility. The method was applied to the Southern region of Luxembourg,
encompassing the same city considered in the former chapters. First, a sensitivity analysis
concerning the impact of the waste heat acquisition price was conducted. The production
of electricity proved to be viable for prices below 10 e/MWh, while waste heat could still be
valorised for heating demand at prices up to 25 e/MWh. Second, the detailed results of the
optimisation at a price of 25 e/MWh were assessed. At that price, profits of more than 10
Me/a from the transport of waste heat and the electricity production of combined heat and
power systems were obtained.

The various developments of this thesis lead to several contributions in the field of building,
urban and regional heat recovery assessment and optimisation, described in detail in the
respective chapters. These contributions are globally summarised as follows:

1. The specific domestic hot water streams characterisation method improves the energy
assessment of buildings and urban systems. With a better understanding of the energy
and temperature requirements of domestic hot water, energy efficiency objectives will
be more easily reached as priorities will be better defined. Risks related to data of poor
quality in decision-making processes are also reduced.

2. With the complementary characterisation of residential waste water streams, the energy
saving impact of various in-building waste water heat recovery systems on the total
heat demand is quantified specifically for the considered building, an aspect not yet
fully addressed. With the spatial allocation of data and aggregation of the results, the

100
 5.2. Future developments

impact can also be quantified at district, city or regional scale with a bottom-up ap-
proach. The outcomes are thus more precise than with methods proposed so far. The
relevance of such solutions can be assessed in relation to other energy saving measures.
The decision-making process, both by the building owner or local decision-takers, is
therefore improved.

3. The impact of an integrated approach including waste water heat recovery in high energy
efficiency buildings in Western Europe is quantified for the first time in a practical case
study. With savings ranging between 28 and 41%, the importance of a holistic approach
to energy optimisation in buildings considering WWHR is concretely demonstrated.

4. An optimisation method for regional waste heat valorisation is formulated specifically

for energy service companies, a point of view so far not considered in regional process
integration problems. In particular, the selection of potential heat sinks is not only
done by considering constraints of temperature levels, distances and costs, but also as a
function of the energy price specific to the sink type. This is an aspect barely considered
so far and not yet implemented in a mixed integer linear programming formulation
targeting regional waste heat valorisation.

By proposing novel or improved methods related to the valorisation of waste heat, this thesis
supports the development of energy efficiency solutions in and across the building, commer-
cial and industrial sectors. These solutions will contribute to address global environmental
challenges that are constraining the sustainable development of our society.

5.2 Future developments

With the discussions conducted in the various chapters, several issues and potential develop-
ments shall be further addressed in the future.

When considering the characterisation of domestic hot water and waste water streams,
country-specific data is still too scarce to apply this approach globally. In the short term,
typical data concerning the frequency and duration of use, mass flows, etc. must be further
acquired. In parallel, the uncertainty behind these models must be addressed. Considering the
proposal for an amendment of the EU directive on energy efficiency ([60]), it can be expected
that new and more specific data on domestic hot water will soon be available for European
countries. This data will also allow for a better calibration of the proposed domestic hot
water models. In the mid-term perspective, considering that spatially allocated data related to
socio-economic level and age of inhabitants is expected to be available, the models shall be
expanded to include these aspects, thus further improving the outcomes of the related energy
demand quantification.

Concerning waste water heat recovery, an interesting prospect is, in the short term, to assess
the effect of the combination with user behaviour and single component optimisation. In

101
Chapter 5. Conclusions and future developments

addition, in-sewer and waste water treatment plant heat recovery configurations should
be included in the urban energy and cost assessment methods in order to compare the
advantages and disadvantages of centralised and decentralised waste water heat recovery
systems. Considering the large number of potential design parameters and waste water
(in-building and sewer) heat recovery configurations, it is also suggested that an integrated
optimisation method is, in the mid-term, developed to design optimal heat recovery, storage
and heating utility systems.

Finally, the proposed regional waste heat optimisation method could be further improved
regarding heat loss calculations. An alternative formulation considering the compensation
from a single, optimal, temperature interval should therefore be developed. In that frame-
work, a reconciliation of the binary variables concerning the consideration of standard pipe
diameters for the investment cost and heat losses calculations should be targeted. In order to
further demonstrate the usefulness of the proposed method, the detailed effect of waste heat
selling and electricity production prices on the outcomes should also be analysed in future
work. In the mid-term, the method will also be extended to cover further practical aspects
usually considered in real projects and so far excluded from this work to limit the complexity
of the addressed problem. In particular, the investment and operating costs of distribution
network, heat transfer fluid, pump station and maintenance work must be integrated in the
model. In order to make a full use of regional excess heat, the problem formulation must
be updated to include the possibility of low-temperature heat valorisation combined with
heat pumps. This formulation should also be expanded to include in-sewer waste water heat
recovery systems. In addition to the classical heat transfer media like steam and thermal oil,
more innovative solutions like CO2 networks must be further included. Also the transportation
technology should be expanded to include the possibility to select mobile heat transport
systems, e.g. via truck or train, which would allow for further flexibility in the management
of RWHV scenarios. On a more practical note, data characterisation methods (including the
proposed DHW models) should be combined with the optimisation problem formulation to
create a tool for energy service companies aiming to valorise available waste heat at a regional
scale.

As long term perspective of this thesis, and in view of the relevance of building energy consump-
tion, the integrated optimisation of domestic hot water demand must be further addressed
as it yields an important potential for energy savings, therefore supporting energy efficiency
objectives set locally, nationally and globally.

102
A In-building waste water heat recovery
models

A.1 Mass flows and temperature levels for shower heat recovery con-
figurations

Configuration 1

The preheated mass flow ṁ ph corresponds to the hot water mass flow ṁ HW , obtained by
considering mass (Eq. A.1) and energy (Eq. A.2) conservation equations, with the temperatures
expressed in K (Eq. A.3).

ṁ shower = ṁ HW + ṁ F W (A.1)

T shower × ṁ shower = T HW × ṁ HW + T F W × ṁ F W (A.2)

(T shower − T F W )
ṁ HW = ṁ shower × (A.3)
(T HW − T F W )

The preheated fresh water temperature T ph of unbalanced flows must be calculated iteratively
based on the fresh water mass flow and using the relations between mass flow, temperature
difference and heat transfer coefficient U in W/m2 *K, heat exchanger surface A in m2 and the
logarithmic mean temperature difference d Tm, expressed in K (Eq. A.4).

Q̇ H E = ṁ F W × c p × (T ph − T F W ) = U × A × d Tm (A.4)

The heat exchanger surface A can be obtained from the manufacturer. It is referred to the
literature for the calculation procedure of the logarithmic mean temperature difference d Tm
and the heat transfer coefficient U (e.g. VDI Gesellschaft [228]), as the detailed description
would be out of scope of the current urban-scale work.

103
Appendix A. In-building waste water heat recovery models

Configuration 2

The preheated mass flow of configuration 2 corresponds to the shower mass flow. For balanced
flows, the preheated water temperature T ph is obtained from heat exchanger efficiency ε pro-
vided by certification institutions, e.g. KIWA in the Netherlands ([Link]) or Passivhaus
Institut in Germany ([Link]), following Eq. A.5.

T ph − T F W
ε= (A.5)
T W W − T FW

Configuration 3

Concerning configuration 3, the preheated mass flow corresponds to the fresh water flow, a
function of the shower mass flow and the system temperatures (Eq. A.6).

(T HW − T shower )
ṁ ph = ṁ shower × (A.6)
(T HW − T ph )

The preheated water temperature T ph is obtained by subtracting the minimum temperature

difference d Tmi n from the waste water temperature (Eq. A.7).

T ph = T W W − d Tmi n (A.7)

A.2 Mass and temperature levels for grey heat recovery configura-
tions

Configuration 4

The hot water mass flow ṁ eHW is proportional to the temperatures of the end-use Te , hot water
T HW and fresh water T F W as well as the end-use mass flow ṁ e (Eq. A.8).

(Te − T F W )
ṁ eHW = ṁ e × (A.8)
(T HW − T F W )

The grey water tank temperature is obtained from the energy conservation equation consider-
ing the sum of the products between temperature and mass of the various grey water streams
and the tank water mass (Eq. A.9).

GW
GW (T × mGW )
P
t ank
T = (A.9)
m t ank

104
 A.2. Mass and temperature levels for grey heat recovery configurations

The grey water thermal power is calculated with the tank energy content potentially rejected
at sewer temperature, over a period of one hour (Eq. A.10).

m t ank × c p × (T t ank − T sewer )

Q̇ t ank = (A.10)
3600

Configuration 5

The thermal load Q˙e of the various DHW end-uses e is obtained by aggregating the daily energy
requirements considering the occupant x occ or household numbers of the building. Eq. A.11
is given as example for the load of DHW end-uses related to occupant use (e.g. showering,
bathing).

x occ (ṁ e × d e × f e ) × c p × (Te − T F W )

P
Q˙e = (A.11)
3600

Configuration 6

The thermal power of a grey water stream Q̇ GW is calculated with Eq. A.12, considering the
sewer temperature T sewer as final temperature.

GW
x occ (ṁ × d e × f e ) × c p × (T GW − T sewer )
P
GW
Q̇ = (A.12)
3600

105
B Regional waste heat valorisation in-
put data

B.1 Distances and road ratio

Table B.1: Distances / road ratio

Site Steel_Differdange [km]/[-] Steel_Belval [km]/[-]

Asphalt 8.89 / 1.00 4.02 / 0.13

Plastic 4.00 / 0.23 7.66 / 0.80
Hard metal 3.54 / 0.20 7.44 / 0.80
Belvaux 3.15 / 1.00 1.39 / 1.00
Differdange 0.20 / 1.00 4.35 / 0.65
Ehlerange 5.91 / 0.95 2.12 / 0.50
Esch-sur-Alzette 4.95 / 1.00 0.44 / 0.20
Mondercange 8.57 / 1.00 3.66 / 0.50
Niederkorn 0.60 / 1.00 6.25 / 0.50
Oberkorn 0.75 / 1.00 4.81 / 0.90
Sanem 2.38 / 0.25 5.74 / 0.50
Soleuvre 1.44 / 0.50 1.88 / 1.00

107
Appendix B. Regional waste heat valorisation input data

B.2 Pipe data

The specific costs for DN300 to DN600 have been extrapolated from smaller diameters using a
linear regression approach (Fig. B.1), with R 2 =98.7% for roads (Eq. B.1) and 98.4% for open
fields (Eq. B.2). The initial data is taken from Nussbaumer and Thalmann [151].

p = 207.9 + 5.8392 ∗ d i amet er (B.1)

p = 139.15 + 5.3625 ∗ d i amet er (B.2)

Figure B.1: Linear regression of pipe costs

108
 B.2. Pipe data

Table B.2: Pipe data (Isoplus [105],Nussbaumer and Thalmann [151])

Type Internal External Maximal Cost (road) Cost (open

radius [m] radius, incl. mass flow [e/m] field) [e/m]
insulation [kg/s]
[m]

DN20 0.014 0.063 0.14 391 308

DN25 0.017 0.063 0.28 396 313
DN32 0.021 0.070 0.56 422 340
DN40 0.024 0.070 0.83 437 355
DN50 0.030 0.080 1.53 495 400
DN60 0.038 0.090 3.06 537 442
DN80 0.045 0.100 4.58 616 500
DN100 0.057 0.125 9.17 760 645
DN125 0.070 0.140 16.11 913 798
DN150 0.084 0.158 26.39 1’101 956
DN200 0.110 0.200 53.61 1’311 1’141
DN250 0.137 0.250 96.67 1’755 1’569
DN300 0.162 0.280 151.94 1’960 1’748
DN350 0.178 0.315 195.83 2’252 2’016
DN400 0.203 0.335 277.78 2’544 2’284
DN450 0.229 0.355 380.56 2’836 2’552
DN500 0.254 0.400 505.56 3’128 2’820
DN600 0.305 0.500 811.11 3’711 3’357

109
Bibliography

[1] M. Abdel-Aal, R. Smits, M. Mohamed, K. De Gussem, A. Schellart, and S. Tait. Modelling

the viability of heat recovery from combined sewers. Water Science and Technology,
70(2):297–306, 2014. ISSN 0273-1223. doi: 10.2166/wst.2014.218. URL [Link]
[Link]/content/70/2/297.

[2] Administration de la navigation aérienne. Hourly outdoor temperature findel-airport,

2015. In French, obtained 06/2015.

[3] J. Allegrini, K. Orehounig, G. Mavromatidis, F. Ruesch, V. Dorer, and R. Evins. A review

of modelling approaches and tools for the simulation of district-scale energy systems.
Renewable and Sustainable Energy Reviews, 52:1391 – 1404, 2015. ISSN 1364-0321.
doi: [Link] URL [Link]
science/article/pii/S1364032115007704.

[4] S. Alnahhal and E. Spremberg. Contribution to exemplary in-house wastewater heat

recovery in Berlin, Germany. Procedia CIRP, 40:35 – 40, 2016. ISSN 2212-8271. doi:
[Link] URL [Link]
science/article/pii/S2212827116000615.

[5] Y. Ammar, S. Joyce, R. Norman, Y. Wang, and A. P. Roskilly. Low grade thermal energy
sources and uses from the process industry in the {UK}. Applied Energy, 89(1):3 – 20,
2012. ISSN 0306-2619. doi: [Link] URL
[Link]

[6] R. Amon, M. Maulhardt, T. Wong, D. Kazama, and C. W. Simmons. Waste heat and
water recovery opportunities in california tomato paste processing. Applied Thermal
Engineering, 78:525 – 532, 2015. ISSN 1359-4311. doi: [Link]
applthermaleng.2014.11.081. URL [Link]
S1359431114011284.

[7] ASUE. Bhkw-kenndaten 2011. Technical report, Arbeitsgemeinschaft für

sparsamen und umweltfreundlichen Energieverbrauch e.V., 2011. URL
[Link]
broschueren/05_07_11_asue-[Link]. Last accessed: 14/11/2016.

111
Bibliography

[8] M. Aydinalp, V. I. Ugursal, and A. S. Fung. Modeling of the space and domestic hot-
water heating energy-consumption in the residential sector using neural networks.
Applied Energy, 79(2):159 – 178, 2004. ISSN 0306-2619. doi: [Link]
1016/[Link].2003.12.006. URL [Link]
S0306261903002344.

[9] C. D. Beal, E. Bertone, and R. A. Stewart. Evaluating the energy and carbon reductions
resulting from resource-efficient household stock. Energy and Buildings, 55:422 – 432,
2012. ISSN 0378-7788. doi: [Link] URL
[Link]

[10] C. D. Beal, R. A. Stewart, and K. Fielding. A novel mixed method smart metering ap-
proach to reconciling differences between perceived and actual residential end use
water consumption. Journal of Cleaner Production, 60:116 – 128, 2013. ISSN 0959-6526.
doi: [Link] URL [Link]
com/science/article/pii/S0959652611003386.

[11] H. Becker. Methodology and thermo-economic optimization for integration of industrial

heat pumps. PhD thesis, EPFL, Lausanne, Switzerland, 2012. URL [Link]
[Link]/record/175150/files/EPFL_TH5341.pdf.

[12] H. Becker and F. Maréchal. Targeting industrial heat pump integration in multi-period
problems. In I. A. Karimi and R. Srinivasan, editors, 11th International Symposium on
Process Systems Engineering, volume 31 of Computer Aided Chemical Engineering, pages
415 – 419. Elsevier, 2012. doi: [Link]
URL [Link]

[13] H. Becker, F. Maréchal, and A. Vuillermoz. Process integration and opportunities for
heat pumps in industrial processes. International Journal of Thermodynamics, 14(2):
59–70, 2011. doi: 10.5541/ijot.260.

[14] P. Bengtsson, J. Berghel, and R. Renstroem. A household dishwasher heated by a heat

pump system using an energy storage unit with water as the heat source. International
Journal of Refrigeration, 49:19 – 27, 2015. ISSN 0140-7007. doi: [Link]
10.1016/[Link].2014.10.012. URL [Link]
S0140700714002849.

[15] A. Bertrand, R. Aggoune, and F. Maréchal. Residential and non-residential hot water use
models for urban and regional energy integration. In Biennial International Workshop
Advances in Energy Studies - Energy and urban systems, Stockholm, Sweden, May 4-7
2015.

[16] A. Bertrand, R. Aggoune, and F. Maréchal. Multi-objective, integrated, regional energy

system optimisation. In Strategic Energy Technology Plan, Luxembourg, Luxembourg,
September 21-22 2015.

112
 Bibliography

[17] A. Bertrand, R. Aggoune, and F. Maréchal. In-building waste water heat recovery: an
urban scale method for energy saving assessments. In 29th international conference on
Efficiency, Cost, Optimization, Simulation and Environmental Impact of Energy Systems
(ECOS), Portoroz, Slovenia, June 19-23 2016.

[18] A. Bertrand, R. Aggoune, and F. Maréchal. In-building waste water heat recovery: an
urban-scale method for the characterisation of water streams and the assessment of
energy savings and costs. Applied Energy, 192:110 – 125, 2017. ISSN 0306-2619. doi:
[Link] URL [Link]
science/article/pii/S0306261917301137.

[19] A. Bertrand, A. Mastrucci, N. Schüler, R. Aggoune, and F. Maréchal. Characterisation

of domestic hot water end-uses for integrated urban thermal energy assessment and
optimisation. Applied Energy, 186, Part 2:152 – 166, 2017. ISSN 0306-2619. doi: [Link]
[Link]/10.1016/[Link].2016.02.107. URL [Link]
article/pii/S030626191630263X.

[20] V. Bianco, F. Scarpa, and L. A. Tagliafico. Analysis and future outlook of natural gas
consumption in the italian residential sector. Energy Conversion and Management, 87:
754 – 764, 2014. ISSN 0196-8904. doi: [Link]
081. URL [Link]

[21] S. Bilgen. Structure and environmental impact of global energy consumption. Renew-
able and Sustainable Energy Reviews, 38:890 – 902, 2014. ISSN 1364-0321. doi: http:
//[Link]/10.1016/[Link].2014.07.004. URL [Link]
article/pii/S1364032114004560.

[22] E. J. M. Blokker, J. H. G. Vreeburg, and J. C. Van Dijk. Simulating residential water demand
with a stochastic end-use model. Journal of Water Resources Planning & Management,
136(1):19 – 26, 2010. ISSN 07339496. URL [Link]
(ASCE)WR.1943-5452.0000002.

[23] E. J. M. Blokker, E. J. Pieterse-Quirijns, J. H. G. Vreeburg, and J. C. van Dijk. Sim-

ulating nonresidential water demand with a stochastic end-use model. Journal
of Water Resources Planning & Management, 137(6):511 – 520, 2011. ISSN
07339496. URL [Link]
0000146?journalCode=jwrmd5.

[24] E. J. M. Blokker, E. J. Pieterse-Quirijns, J. H. G. Vreeburg, and J. C. van Dijk. Closure

to ’simulating nonresidential water demand with a stochastic end-use model’ by e. j.
m. blokker, e. j. pieterse-quirijns, j. h. g. vreeburg, and j. c. van dijk. Journal of Water
Resources Planning & Management, 139(3):347 – 348, 2013. ISSN 07339496. URL http:
//[Link]/doi/abs/10.1061/(ASCE)WR.1943-5452.0000346.

113
Bibliography

[25] B. Boehm. Experimental determination of heat losses from buried district heating
pipes in normal operation. Heat Transfer Engineering, 22(3):41–51, 2001. doi: 10.1080/
01457630118274. URL [Link]

[26] C. Boissavy. Cost and return on investment for geothermal heat pump systems in france.
In World Geothermal Congress 2015 Melbourne, Australia, April 19-25, 2015, 2015. URL
[Link]

[27] [Link], 2015. URL [Link]. Last accessed: 23/10/2015.

[28] T. Brady, A. Davies, and D. M. Gann. Creating value by delivering integrated solutions.
International Journal of Project Management, 23(5):360 – 365, 2005. ISSN 0263-7863.
doi: [Link] URL [Link]
com/science/article/pii/S0263786305000025.

[29] M. Brand, J. E. Thorsen, and S. Svendsen. Numerical modelling and experimental meas-
urements for a low-temperature district heating substation for instantaneous prepara-
tion of dhw with respect to service pipes. Energy, 41(1):392 – 400, 2012. ISSN 0360-5442.
doi: [Link] URL [Link]
com/science/article/pii/S0360544212001752.

[30] S. Brückner, S. Liu, L. Miró, M. Radspieler, L. F. Cabeza, and E. Lävemann. In-

dustrial waste heat recovery technologies: An economic analysis of heat transform-
ation technologies. Applied Energy, 151:157 – 167, 2015. ISSN 0306-2619. doi:
[Link] URL [Link]
science/article/pii/S0306261915004584.

[31] S. Broberg-Viklund. Energy efficiency through industrial excess heat recovery—

policy impacts. Energy Efficiency, 8(1):19–35, 2015. ISSN 1570-6478. doi: 10.1007/
s12053-014-9277-3. URL [Link]

[32] H. L. Brown, B. B. Hamel, B. A. Hedman, M. Koluch, B. C. Gajanana, and P. Troy. Energy

Analysis of 108 Industrial Processes. Fairmont Press, Inc., 1996.

[33] S. Brueckner, L. Miró, L. F. Cabeza, M. Pehnt, and E. Laevemann. Methods to estim-

ate the industrial waste heat potential of regions - a categorization and literature re-
view. Renewable and Sustainable Energy Reviews, 38:164 – 171, 2014. ISSN 1364-0321.
doi: [Link] URL [Link]
science/article/pii/S1364032114003165.

[34] C. Brunner, B. Muster-Slawitsch, E. Heigl, H. Schweiger, and C. Vannoni. Einstein:

Expert system for an intelligent supply of thermal energy in industry: audit methodology
and software tool. In Eurosun 2010, Graz, Austria, Sept. 28 - Oct.1, 2010, 2010. URL
[Link]

114
 Bibliography

[35] Buderus. Preise + technik wärmeerzeugung - gasheizkessel. Technical report, Buderus,

2016. URL [Link] Last
accessed: 20/10/2016.

[36] Bundesministerium fur Arbeit, Soziales und Kosumentenschutz. Alten-/pflegeheime,

2015. URL [Link] In German,
last accessed: 17/09/2015.

[37] R. Cahill, J. R. Lund, B. DeOreo, and J. Medellin-Azuara. Household water use

and conservation models using monte carlo techniques. Hydrology and Earth Sys-
tem Sciences, 17(10):3957–3967, 2013. doi: 10.5194/hess-17-3957-2013. URL http:
//[Link]/17/3957/2013/.

[38] Capgeris. Hebergement personnes agees, 2015. URL [Link] In

French, last accessed: 17/09/2015.

[39] W. Chen, S. Liang, Y. Guo, K. Cheng, X. Gui, and D. Tang. Investigation on the thermal
performance and optimization of a heat pump water heater assisted by shower waste
water. Energy and Buildings, 64:172 – 181, 2013. ISSN 0378-7788. doi: [Link]
10.1016/[Link].2013.04.021. URL [Link]
S0378778813002740.

[40] M. R. Chertow. "Uncovering" industrial symbiosis. Journal of Industrial Ecology, 11(1):

11–30, 2007. ISSN 1530-9290. doi: 10.1162/jiec.2007.1110. URL [Link]
1162/jiec.2007.1110.

[41] K. Chew, J. J. Klemeš, S. Rafidah Wan Alwi, and Z. Manan. Industrial implementation
issues of total site heat integration. Applied Thermal Engineering, 61(1):17 – 25, 2013.
ISSN 1359-4311. doi: [Link] URL
[Link]

[42] C. Christiansen, A. Rosa, M. Brand, P. Olsen, and J. Thorsen. Results and experiences
from a 2 - year study with measurements on a low - temperature dh system for low energy
buildings. In 13th International Symposium on District Heating and Cooling 2012, Copen-
hagen, Denmark, Sept. 3-4, 2012, 2012. URL [Link]
sites/default/files/project/docs/VFHZC102_results-and-experiences_lores_0.pdf.

[43] S. S. Cipolla and M. Maglionico. Heat recovery from urban wastewater: Analysis of the
variability of flow rate and temperature in the sewer of bologna, italy. Energy Procedia,
45:288 – 297, 2014. ISSN 1876-6102. doi: [Link]
URL [Link]

[44] R. Cobacho, F. Arregui, J. Parra, and E. Cabrera Jr. Improving efficiency in water use and
conservation in spanish hotels. Water Science and Technology: Water Supply, 5(3-4):
273–279, 2005. URL [Link]

115
Bibliography

[45] Q. Deng, X. Jiang, L. Zhang, and Q. Cui. Making optimal investment decisions for
energy service companies under uncertainty: A case study. Energy, 88:234 – 243, 2015.
ISSN 0360-5442. doi: [Link] URL [Link]
[Link]/science/article/pii/S0360544215005381.

[46] Z. Deng, S. Mol, and J. P. van der Hoek. Shower heat exchanger: reuse of energy from
heated drinking water for co2 reduction. Drinking Water Engineering and Science, 9(1):
1–8, 2016. doi: 10.5194/dwes-9-1-2016. URL [Link]
2016/.

[47] W. DeOreo, J. Heaney, and P. Mayer. Flow trace analysis to assess water use. Journal of
the American Water Works Association, 88(1):79–90, Jan 1996.

[48] V. R. Dhole and B. Linnhoff. Total site targets for fuel, co-generation, emissions, and
cooling. Computers and Chemical Engineering, 17, Supplement 1(0):S101 – S109, 1993.
ISSN 0098-1354. doi: [Link] URL http:
//[Link]/science/article/pii/0098135493802148.

[49] J. Dong, Z. Zhang, Y. Yao, Y. Jiang, and B. Lei. Experimental performance evaluation of a
novel heat pump water heater assisted with shower drain water. Applied Energy, 154:
842 – 850, 2015. ISSN 0306-2619. doi: [Link]
URL [Link]

[50] Dow Chemical Company. Dowtherm a heat transfer fluid - product technical
data. Technical report, Dow Chemical Company, 1997. URL [Link].
com/PublishedLiteratureDOWCOM/dh_0030/[Link]. Last accessed:
26/10/2016.

[51] D. J. Durrenmatt and O. Wanner. A mathematical model to predict the effect of heat
recovery on the wastewater temperature in sewers. Water Research, 48:548 – 558, 2014.
ISSN 0043-1354. doi: [Link] URL [Link]
[Link]/science/article/pii/S0043135413007951.

[52] J. Elias-Maxil, J. P. van der Hoek, J. Hofman, and L. Rietveld. Energy in the urban water
cycle: Actions to reduce the total expenditure of fossil fuels with emphasis on heat
reclamation from urban water. Renewable and Sustainable Energy Reviews, 30:808 –
820, 2014. ISSN 1364-0321. doi: [Link] URL
[Link]

[53] B. Elmegaard, T. S. Ommen, M. Markussen, and J. Iversen. Integration of space heating

and hot water supply in low temperature district heating. Energy and Buildings, pages –
, 2015. ISSN 0378-7788. doi: [Link] URL
[Link]

[54] Enerdata. Odyssee database on energy efficiency data & indicators, 2016. URL http:
//[Link]/[Link]. Last accessed:
08/02/2016.

116
 Bibliography

[55] Enovos. Electricity price naturstroum, 2015. URL [Link]

electricity/naturstroum. Last accessed: 25/08/2015.

[56] Entreprise des Postes et Télécommunications. GIS data set, 2013. Obtained 07/2013.

[57] L. Eriksson, M. Morandin, and S. Harvey. Targeting capital cost of excess heat collection
systems in complex industrial sites for district heating applications. Energy, 91:465 –
478, 2015. ISSN 0360-5442. doi: [Link] URL
[Link]

[58] P. Eslami-Nejad and M. Bernier. Impact of grey water heat recovery on the electrical
demand of domestic hot water heaters. In 11th International Building Performance
Simulation Association Conference, Glasgow, Scotland, July 27-30, 2009, 2009.

[59] European Commission. Statistical pocketbook 2015, 2015. URL [Link]

energy/sites/ener/files/documents/PocketBook_ENERGY_2015PDFfinal.pdf.

[60] European Commission. Proposal for a directive of the european parliament and of
the council amending directive 2012/27/eu on energy efficiency, 2016. URL [Link]
[Link]/energy/sites/ener/files/documents/1_en_act_part1_v16.pdf. Last accessed:
10/02/2017.

[61] European Parliament. Regulation (ec) no 1099/2008 of the european parliament and of
the council of 22 october 2008 on energy statistics, 2008.

[62] European Parliament. Directive 2010/31/eu of the european parliament and of the
council of 19 may 2010 on the energy performance of buildings, 2010.

[63] European Parliament. Directive 2012/27/eu of the european parliament and of the
council of 25 october 2012 on energy efficiency, amending directives 2009/125/ec and
2010/30/eu and repealing directives 2004/8/ec and 2006/32/ec, 2012.

[64] Eurostat. Manual for statistics on energy consumption in households, 2013.

URL [Link]
PDF/baa96509-3f4b-4c7a-94dd-feb1a31c7291. Last accessed: 13/03/2017.

[65] Eurostat. Share of total population having neither a bath, nor a shower in their dwelling
(ilc-mdho02), 2015. URL [Link]
ilc_mdho02. Last accessed: 07/09/2015.

[66] Eurostat. Net occupancy rate of bed-places and bedrooms in hotels and similar
accommodation(tour-occ-mnor), 2015. URL [Link]
products-datasets/-/TOUR_OCC_MNOR. Last accessed: 15/09/2015.

[67] Eurostat. Number of establishments, bedrooms and bed-places by nuts 2 regions (tour-
cap-nat)), 2015. URL [Link]
cap_nat. Last accessed: 15/09/2015.

117
Bibliography

[68] Eurostat. Population density by nuts 3 region, 2016. URL [Link]

[Link]/nui/[Link]?dataset=demo_r_d3dens&lang=en. Last accessed: 06/12/2016.

[69] Eurostat. Energy price statistics, 2016. URL [Link]

statistics-explained/[Link]/Energy_price_statistics. Last accessed: 08/12/2016.

[70] J. Evans, P. Evans, and C. Babatope. European healthcare - care homes report 2014.
Technical report, Knight Frank, 2014. URL [Link]
656/documents/en/[Link]. Last accessed: 15/09/2015.

[71] H. Fang, J. Xia, K. Zhu, Y. Su, and Y. Jiang. Industrial waste heat utilization for low
temperature district heating. Energy Policy, 62:236 – 246, 2013. ISSN 0301-4215. doi:
[Link] URL [Link]
science/article/pii/S0301421513006113.

[72] S. Fazlollahi. Decomposition optimization strategy for the design and operation of district
energy system. PhD thesis, EPFL, Lausanne, Switzerland, 2014.

[73] S. Fazlollahi, G. Becker, and F. Maréchal. Multi-objectives, multi-period optimiza-

tion of district energy systems: II - daily thermal storage. Computers & Chemical
Engineering, 71:648 – 662, 2014. ISSN 0098-1354. doi: [Link]
compchemeng.2013.10.016. URL [Link]
S0098135413003384.

[74] S. Fazlollahi, G. Becker, and F. Maréchal. Multi-objectives, multi-period optimiza-

tion of district energy systems: III. distribution networks. Computers & Chemical
Engineering, 66(0):82 – 97, 2014. ISSN 0098-1354. doi: [Link]
compchemeng.2014.02.018. URL [Link]
S0098135414000507.

[75] S. Fazlollahi, N. Schueler, and F. Maréchal. A solid thermal storage model for the
optimization of buildings operation strategy. Energy, 88:209 – 222, 2015. ISSN 0360-5442.
doi: [Link] URL [Link]
com/science/article/pii/S0360544215005526.

[76] C. A. Floudas, A. R. Ciric, and I. E. Grossmann. Automatic synthesis of optimum heat

exchanger network configurations. AIChE Journal, 32(2):276–290, 1986. ISSN 1547-5905.
doi: 10.1002/aic.690320215. URL [Link]

[77] J. A. Fonseca and A. Schlueter. Integrated model for characterization of spatiotem-

poral building energy consumption patterns in neighborhoods and city districts. Ap-
plied Energy, 142(0):247 – 265, 2015. ISSN 0306-2619. doi: [Link]
1016/[Link].2014.12.068. URL [Link]
S0306261914013257.

118
 Bibliography

[78] C. Forman, I. K. Muritala, R. Pardemann, and B. Meyer. Estimating the global waste heat
potential. Renewable and Sustainable Energy Reviews, 57:1568 – 1579, 2016. ISSN 1364-
0321. doi: [Link] URL [Link]
com/science/article/pii/S1364032115015750.

[79] R. Fourer, D. Gay, and B. Kernighan. AMPL: A Modeling Language for Mathematical Pro-
gramming. Scientific Press series. Thomson/Brooks/Cole, 2003. ISBN 9780534388096.

[80] J. Frijns, J. Hofman, and M. Nederlof. The potential of (waste)water as energy carrier.
Energy Conversion and Management, 65:357 – 363, 2013. ISSN 0196-8904. doi: http:
//[Link]/10.1016/[Link].2012.08.023. URL [Link]
science/article/pii/S0196890412003512.

[81] E. Fuentes, L. Arce, and J. Salom. A review of domestic hot water consumption profiles
for application in systems and buildings energy performance analysis. Renewable
and Sustainable Energy Reviews, pages –, 2017. ISSN 1364-0321. doi: [Link]
10.1016/[Link].2017.05.229. URL [Link]
S1364032117308614.

[82] M. Fujii, Y. Dou, T. Togawa, L. Dong, S. Ohnishi, H. Tanikawa, and T. Fujita. Innovative
planning and evaluation system for district heating using waste heat considering spatial
configuration: A case in fukushima, japan. Resources, Conservation and Recycling, 128:
406 – 416, 2016. ISSN 0921-3449. doi: [Link]
URL [Link]

[83] N. Fumo and M. R. Biswas. Regression analysis for prediction of residential energy
consumption. Renewable and Sustainable Energy Reviews, 47:332–343, 2015. ISSN
13640321. doi: 10.1016/[Link].2015.03.035. URL [Link]
pii/S1364032115001884.

[84] M. Gaderer. Wärmeversorgung mit fester Biomasse bei kleiner Leistung. PhD thesis,
Technische Universtiät München, Munich, Germany, 2007. URL [Link]
[Link]/doc/630149/[Link]. Last accessed: 11/03/2015.

[85] L. Girardin, F. Maréchal, M. Dubuis, N. Calame-Darbellay, and D. Favrat. Energis: A

geographical information based system for the evaluation of integrated energy con-
version systems in urban areas. Energy, 35(2):830 – 840, 2010. ISSN 0360-5442. doi:
[Link] URL [Link]
science/article/pii/S0360544209003582.

[86] X. Gou, Y. Fu, I. Shah, Y. Li, G. Xu, Y. Yang, E. Wang, L. Liu, and J. Wu. Research on a
household dual heat source heat pump water heater with preheater based on ASPEN
PLUS. Energies, 9(12), 2016. doi: [Link]

[87] S. Grönkvist and P. Sandberg. Driving forces and obstacles with regard to co-operation
between municipal energy companies and process industries in Sweden. Energy Policy,

119
Bibliography

34(13):1508 – 1519, 2006. ISSN 0301-4215. doi: [Link]

11.001. URL [Link]

[88] O. Guerra Santin, L. Itard, and H. Visscher. The effect of occupancy and building
characteristics on energy use for space and water heating in Dutch residential stock.
Energy and Buildings, 41(11):1223–1232, 2009. ISSN 03787788. doi: 10.1016/[Link].
2009.07.002.

[89] Y. X. Guo, Y. L. Cai, S. Q. Liang, and W. Chen. Experimental study on a shower waste
water heat recovery device in buildings. Applied Mechanics and Materials, 226-228:
2402–2406, November 2012. doi: 10.4028/[Link]/AMM.226-228.2402.

[90] A. Gutierrez-Escolar, A. Castillo-Martinez, J. M. Gomez-Pulido, J.-M. Gutierrez-Martinez,

and E. Garcia-Lopez. A new system for households in spain to evaluate and reduce their
water consumption. Water, 6(1):181, 2014. ISSN 2073-4441. doi: 10.3390/w6010181.
URL [Link]

[91] R. Hackl, E. Andersson, and S. Harvey. Targeting for energy efficiency and improved en-
ergy collaboration between different companies using total site analysis (TSA) . Energy,
36(8):4609 – 4615, 2011. ISSN 0360-5442. doi: [Link]
03.023. URL [Link]

[92] A. Haederli, D. Balmer, C. Boss, O. Zihlmann, and J. Schmidli. Atlas der Altersheime,
2015. URL [Link] In German, last accessed: 17/09/2015.

[93] C. Haikarainen, F. Pettersson, and H. Saxén. A model for structural and operational
optimization of distributed energy systems. Applied Thermal Engineering, 70(1):211–
218, 2014. ISSN 1359-4311. doi: 10.1016/[Link].2014.04.049. URL http:
//[Link]/science/article/pii/S1359431114003135.

[94] A. Hauer and F. Fischer. Open adsorption system for an energy efficient dishwasher.
Chemie Ingenieur Technik, 83(1-2):61–66, 2011. ISSN 1522-2640. doi: 10.1002/cite.
201000197. URL [Link]

[95] R. Hendron and J. Burch. Development of standardized domestic hot water event sched-
ules for residential buildings. In Energy Sustainability 2007, Long Beach, California, Jun.
27 - 30, 2007, 2007. URL [Link] Last accessed:
17/12/2014.

[96] A. Hepbasli, E. Biyik, O. Ekren, H. Gunerhan, and M. Araz. A key review of wastewater
source heat pump (WWSHP) systems. Energy Conversion and Management, 88:700 –
722, 2014. ISSN 0196-8904. doi: [Link]
URL [Link]

[97] A. Hesaraki, E. Bourdakis, A. Ploskic, and S. Holmberg. Experimental study of energy

performance in low-temperature hydronic heating systems. Energy and Buildings, 109:

120
 Bibliography

108 – 114, 2015. ISSN 0378-7788. doi: [Link]

URL [Link]

[98] F. S. Hillier and G. J. Lieberman. Introduction to Operations Research. McGraw-Hill, New

York, USA, eigth edition, 2005.

[99] D. Hoak, D. Parker, and A. Hermelink. How energy efficient are modern dishwashers.
In American Council for an Energy Efficient Economy 2008 - Summer Study on Energy
Efficiency in Buildings, Washington, U.S.A., August 2008, August 2008. URL [Link]
[Link]/en/publications/pdf/[Link]. Last accessed: 18/01/2016.

[100] S. Hokoi, D. Ogura, X. Fu, and Y. Rao. Field survey on energy consumption due to hot wa-
ter supply and cooking in Nanjing and Hefei, China. Frontiers of Architectural Research,
2(2):134 – 146, 2013. ISSN 2095-2635. doi: [Link]
URL [Link]

[101] B. Howard, L. Parshall, J. Thompson, S. Hammer, J. Dickinson, and V. Modi. Spatial

distribution of urban building energy consumption by end use. Energy and Buildings,
45(0):141 – 151, 2012. ISSN 0378-7788. doi: [Link]
061. URL [Link]

[102] IBM. IBM ILOG CPLEX Optimization Studio, 2016. URL [Link]
software/products/en/ibmilogcpleoptistud.

[103] INSEE. Confort sanitaire et chauffage des résidences principales en 2006, 2006. URL
[Link] In French,
last accessed: 11/09/2015.

[104] INSEE. Équipement des ménages en biens durables selon le type de ménage, 2012. URL
[Link] In French,
last accessed: 11/09/2015.

[105] Isoplus. Planungshandbuch. Technical report, Isoplus, 2015. URL http:

//[Link]/fileadmin/data/downloads/documents/germany/Catalogue_
German/Kapitel_2_Starre_Verbundsysteme.pdf. Last accessed: 11/11/2016.

[106] M. Jennings, D. Fisk, and N. Shah. Modelling and optimization of retrofitting residential
energy systems at the urban scale. Energy, 64:220 – 233, 2014. ISSN 0360-5442. doi:
[Link] URL [Link]
science/article/pii/S0360544213009432.

[107] S. W. Jeong and D. H. Lee. Drying performance of a dishwasher with internal air circula-
tion. Korean Journal of Chemical Engineering, 31(9):1518–1521, 2014. ISSN 0256-1115.
doi: 10.1007/s11814-014-0194-0. URL [Link]

[108] C. Jones. Hotels take the bath out of bathrooms, 2011. URL [Link]
MONEY/usaedition/2011-05-20-hotel1tubs19_ST_U.htm. Last accessed: 11/09/2015.

121
Bibliography

[109] J. Keirstead, M. Jennings, and A. Sivakumar. A review of urban energy system models:
Approaches, challenges and opportunities. Renewable and Sustainable Energy Reviews,
16(6):3847 – 3866, 2012. ISSN 1364-0321. doi: [Link]
047. URL [Link]

[110] S. J. Kenway, R. Scheidegger, T. A. Larsen, P. Lant, and H.-P. Bader. Water-related energy
in households: A model designed to understand the current state and simulate possible
measures. Energy and Buildings, 58:378 – 389, 2013. ISSN 0378-7788. doi: [Link]
[Link]/10.1016/[Link].2012.08.035. URL [Link]
article/pii/S0378778812004392.

[111] J.-K. Kim, A. Kapil, I. Bulatov, and R. Smith. Process integration of low grade heat in pro-
cess industry with district heating networks. Energy, 44(1):11 – 19, 2012. ISSN 0360-5442.
doi: [Link] URL [Link]
com/science/article/pii/S0360544211008097.

[112] J. J. Klemeš, editor. Handbook of Process Integration (PI). Number 61. Woodheat Pub-
lishing Limited, 2013.

[113] J. J. Klemeš. 1 - process integration (pi): An introduction. In J. J. Kleme, editor,

Handbook of Process Integration (PI), Woodhead Publishing Series in Energy, pages
3 – 27. Woodhead Publishing, 2013. ISBN 978-0-85709-593-0. doi: [Link]
org/10.1533/9780857097255.1.3. URL //[Link]/science/article/pii/
B9780857095930500010.

[114] J. J. Klemeš, V. Dhole, K. Raissi, S. Perry, and L. Puigjaner. Targeting and design methodo-
logy for reduction of fuel, power and {CO2} on total sites. Applied Thermal Engineering,
17:993 – 1003, 1997. ISSN 1359-4311. doi: [Link]
00087-7. URL [Link]

[115] T.-A. Koiv and A. Toode. Trends in domestic hot water consumption in Estonian apart-
ment buildings. Proc. Estonian Acad. Sci. Eng., 12(1):72–80, 2006.

[116] J. Z. Kolter and J. Ferreira. A Large-scale Study on Predicting and Contextualizing

Building Energy Usage. Proceedings of the Conference on Artificial Intelligence (AAAI),
Special Track on Computational Sustainability and AI, 2011, page 8, 2011.

[117] S. Kordana, D. Slys, and J. Dziopak. Rationalization of water and energy consumption in
shower systems of single-family dwelling houses. Journal of Cleaner Production, 82:58 –
69, 2014. ISSN 0959-6526. doi: [Link] URL
[Link]

[118] C. M. Leidl and W. D. Lubitz. Comparing domestic water heating technologies. Tech-
nology in Society, 31(3):244 – 256, 2009. ISSN 0160-791X. doi: [Link]
1016/[Link].2009.06.005. URL [Link]
S0160791X09000591.

122
 Bibliography

[119] P. Y. Liew, W. L. Theo, S. R. W. Alwi, J. S. Lim, Z. A. Manan, J. J. Klemeš, and P. S. Varbanov.

Total site heat integration planning and design for industrial, urban and renewable
systems. Renewable and Sustainable Energy Reviews, 68, Part 2:964 – 985, 2017. ISSN
1364-0321. doi: [Link] URL //[Link].
com/science/article/pii/S1364032116301988.

[120] M.-H. Lin, J. G. Carlsson, D. Ge, J. Shi, and J.-F. Tsai. A review of piecewise linearization
methods. Mathematical Problems in Engineering, 13:1 – 8, 2013. doi: [Link]
10.1155/2013/101376. URL [Link]

[121] S. Lin, J. Broadbent, and R. McGlen. Numerical study of heat pipe application in
heat recovery systems. Applied Thermal Engineering, 25(1):127 – 133, 2005. ISSN
1359-4311. doi: [Link] URL http:
//[Link]/science/article/pii/S1359431104000651.

[122] B. Linnhoff and J. R. Flower. Synthesis of heat exchanger networks: I. systematic genera-
tion of energy optimal networks. AIChE Journal, 24(4):633–642, 1978. ISSN 1547-5905.
doi: 10.1002/aic.690240411. URL [Link]

[123] B. Linnhoff and E. Hindmarsh. The pinch design method for heat exchanger networks.
Chemical Engineering Science, 38(5):745 – 763, 1983. ISSN 0009-2509. doi: [Link]
[Link]/10.1016/0009-2509(83)80185-7. URL [Link]
article/pii/0009250983801857.

[124] B. Linnhoff, D. Townsend, D. Boland, G. Hewitt, B. Thomas, A. Guy, and R. Marsland.

User guide on process integration for the efficient use of energy. Technical report,
Institution of chemical engineers, Rugby, U.K., 1982.

[125] L. Liu, L. Fu, and Y. Jiang. Application of an exhaust heat recovery system for domestic
hot water. Energy, 35(3):1476 – 1481, 2010. ISSN 0360-5442. doi: [Link]
10.1016/[Link].2009.12.004. URL [Link]
S0360544209005246.

[126] Luxemburgish Parliament. Règlement grand-ducal du 31 août 2010 concernant la

performance énergétique des bâtiments fonctionnels, 2010.

[127] Luxemburgish Parliament. Règlement grand-ducal du 26 décembre 2012 relatif à la

production d électricité basée sur la cogénération à haut rendement, 2012.

[128] Luxemburgish Parliament. Règlement grand-ducal du 26 mai 2014 modifiant 1. le

règlement grand-ducal modifié du 30 novembre 2007 concernant la performance éner-
gétique des bâtiments d’habitation; 2. le règlement grand-ducal modifié du 31 août 2010
concernant la performance énergétique des bâtiments fonctionnels; et 3. le règlement
grand-ducal du 27 février 2010 concernant les installations à gaz, 2014.

[129] Luxemburgish Parliament. Règlement grand-ducal du 1er août 2014 relatif à la pro-
duction d’électricité basée sur les sources d’énergie renouvelables et modifiant: 1. le

123
Bibliography

règlement grand-ducal du 31 mars 2010 relatif au mécanisme de compensation dans

le cadre de l’organisation du marché de l’électricité; 2. le règlement grandducal du
15 décembre 2011 relatif à la production, la rémunération et la commercialisation de
biogaz., 2014.

[130] Luxsenior. Releve des services pour personnes agees au luxembourg. Technical report,
Luxsenior, 2015. URL [Link]

[131] S. Maas, D. Waldmann, A. Zuerbes, and J. Scheuren. Wie viel der luxemburg wirklich
verbraucht... Revue technique Luxembourgeoise, 4:211 – 216, 2007. URL [Link]
lu/attachments/article/40/Artikel_WasverbrauchteinLuxemburgerwirklich.pdf.

[132] A. A. Makki, R. A. Stewart, C. D. Beal, and K. Panuwatwanich. Novel bottom-up urban

water demand forecasting model: Revealing the determinants, drivers and predictors of
residential indoor end-use consumption. Resources, Conservation and Recycling, 95:15 –
37, 2015. ISSN 0921-3449. doi: [Link] URL
[Link]

[133] F. Maréchal. Méthode d’analyse et de synthèse énergétique des procédés industriels. PhD
thesis, 1997. In French.

[134] F. Maréchal and B. Kalitventzeff. Synep1 : A methodology for energy integration and
optimal heat exchanger network synthesis. Computers and Chemical Engineering, 13:
603 – 610, 1989. ISSN 0098-1354. doi: [Link]
URL [Link]

[135] A. Mastrucci, O. Baume, F. Stazi, and U. Leopold. Estimating energy savings for the
residential building stock of an entire city: A GIS-based statistical downscaling approach
applied to rotterdam. Energy and Buildings, 75(0):358 – 367, 2014. ISSN 0378-7788. doi:
[Link] URL [Link]
science/article/pii/S0378778814001467.

[136] R. McKenna and J. Norman. Spatial modelling of industrial heat loads and recovery
potentials in the {UK}. Energy Policy, 38(10):5878 – 5891, 2010. ISSN 0301-4215. doi:
[Link] URL //[Link]/science/
article/pii/S0301421510004131.

[137] A. McNabola and K. Shields. Efficient drain water heat recovery in horizontal domestic
shower drains. Energy and Buildings, 59:44 – 49, 2013. ISSN 0378-7788. doi: [Link]
[Link]/10.1016/[Link].2012.12.026. URL [Link]
article/pii/S0378778812006755.

[138] F. Meggers and H. Leibundgut. The potential of wastewater heat and exergy: Decentral-
ized high-temperature recovery with a heat pump. Energy and Buildings, 43(4):879 –
886, 2011. ISSN 0378-7788. doi: [Link] URL
[Link]

124
 Bibliography

[139] F. Meggers, V. Ritter, P. Goffin, M. Baetschmann, and H. Leibundgut. Low exergy building
systems implementation. Energy, 41(1):48 – 55, 2012. ISSN 0360-5442. doi: http://
[Link]/10.1016/[Link].2011.07.031. URL [Link]
article/pii/S0360544211004798.

[140] A. Merzkirch, T. Hoos, F. Maas, F. Scholzen, and D. Waldmann. Wie genau sind unsere
Energiepaesse? Bauphysik, 36(1):40–43, 2014. ISSN 1437-0980. doi: 10.1002/bapi.
201410007. URL [Link]

[141] A. Mian, E. Martelli, and F. Maréchal. Framework for the multiperiod sequential syn-
thesis of heat exchanger networks with selection, design, and scheduling of multiple
utilities. Industrial and Engineering Chemistry Research, 55(1):168–186, 2016. doi:
[Link] URL [Link]
[Link].5b02104.

[142] L. Miró, S. Brückner, and L. F. Cabeza. Mapping and discussing industrial waste heat
(iwh) potentials for different countries. Renewable and Sustainable Energy Reviews, 51:
847 – 855, 2015. ISSN 1364-0321. doi: [Link] URL
//[Link]/science/article/pii/S1364032115006073.

[143] M. Morandin, R. Hackl, and S. Harvey. Economic feasibility of district heating delivery
from industrial excess heat: A case study of a Swedish petrochemical cluster. Energy, 65:
209 – 220, 2014. ISSN 0360-5442. doi: [Link]
URL [Link]

[144] B. Morgenthaler, A. Grossenbacher, and E. Burnier. Statistical data on switzerland 2015.

Technical report, Federal Statistical Office, 2015. URL [Link]
portal/en/index/news/[Link].

[145] A. Nemet, J. J. Klemeš, and P. S. Varbanov. Methodology for maximising the use of
renewables with variable availability. In M. G. E.N. Pistikopoulos and A. Kokossis,
editors, 21st European Symposium on Computer Aided Process Engineering, volume 29
of Computer Aided Chemical Engineering, pages 1944 – 1948. Elsevier, 2011. doi: http:
//[Link]/10.1016/B978-0-444-54298-4.50167-7. URL [Link]
com/science/article/pii/B9780444542984501677.

[146] A. Nemet, J. J. Klemeš, and Z. Kravanja. Designing a total site for an entire lifetime
under fluctuating utility prices. Computers & Chemical Engineering, 72:159 – 182, 2015.
ISSN 0098-1354. doi: [Link] URL
[Link]

[147] R. Neunteufel, L. Richard, and R. Perfler. Wasserverbrauch und Wasserbedarf - Aus-

wertung empirischer Daten zum Wasserverbrauch. Technical report, Bundesminis-
terium für Land- und Forstwirtschaft, Umwelt und Wasserwirtschaft, 2012. URL http:
//[Link]/uploads/wasserverbrauch_und_wasserbedarf.pdf. In Ger-
man, last accessed: 11/09/2015.

125
Bibliography

[148] L. Ni, S. Lau, H. Li, T. Zhang, J. Stansbury, J. Shi, and J. Neal. Feasibility study of a localized
residential grey water energy-recovery system. Applied Thermal Engineering, 39:53 – 62,
2012. ISSN 1359-4311. doi: [Link]
URL [Link]

[149] NIST. National institute of standards and technology chemistry webbook.

[Link] U.S. Department of Commerce, 2011.

[150] R. Nouvel, A. Mastrucci, U. Leopold, O. Baume, V. Coors, and U. Eicker. Combining

GIS-based statistical and engineering urban heat consumption models: Towards a new
framework for multi-scale policy support. Energy and Buildings, pages –, 2015. ISSN
0378-7788. doi: [Link] URL [Link]
[Link]/science/article/pii/S0378778815302061.

[151] T. Nussbaumer and S. Thalmann. Influence of system design on heat distribution costs
in district heating. Energy, 101:496 – 505, 2016. ISSN 0360-5442. doi: [Link]
10.1016/[Link].2016.02.062. URL [Link]
S036054421630113X.

[152] S.-Y. Oh, M. Binns, Y.-K. Yeo, and J.-K. Kim. Improving energy efficiency for local energy
systems. Applied Energy, 131:26 – 39, 2014. ISSN 0306-2619. doi: [Link]
1016/[Link].2014.06.007. URL [Link]
S0306261914005819.

[153] G. Oluleye, M. Jobson, and R. Smith. A hierarchical approach for evaluating and selecting
waste heat utilization opportunities. Energy, 90, Part 1:5 – 23, 2015. ISSN 0360-5442. doi:
[Link] URL [Link]
science/article/pii/S0360544215006672.

[154] G. Oluleye, M. Jobson, R. Smith, and S. J. Perry. Evaluating the potential of process
sites for waste heat recovery. Applied Energy, 161:627 – 646, 2016. ISSN 0306-2619.
doi: [Link] URL //[Link]/
science/article/pii/S0306261915008405.

[155] K. Çomakli, B. Yüksel, and Ömer Çomakli. Evaluation of energy and exergy losses
in district heating network. Applied Thermal Engineering, 24(7):1009 – 1017, 2004.
ISSN 1359-4311. doi: [Link] URL
[Link]

[156] ONT. [Link], 2015. URL [Link]

eat-sleep/hotels.

[157] M. D. Paepe, E. Theuns, S. Lenaers, and J. V. Loon. Heat recovery system for dishwashers.
Applied Thermal Engineering, 23(6):743 – 756, 2003. ISSN 1359-4311. doi: [Link]
[Link]/10.1016/S1359-4311(03)00016-4. URL [Link]
article/pii/S1359431103000164.

126
 Bibliography

[158] C. Pakula and R. Stamminger. Electricity and water consumption for laundry washing
by washing machine worldwide. Energy Efficiency, 3(4):365–382, 11 2010. doi: http:
//[Link]/10.1007/s12053-009-9072-8.

[159] S. A. Papoulias and I. E. Grossmann. A structural optimization approach in process

synthesis - i: Utility systems. Computers & Chemical Engineering, 7(6):695 – 706, 1983.
ISSN 0098-1354. doi: [Link] URL http:
//[Link]/science/article/pii/0098135483850224.

[160] S. A. Papoulias and I. E. Grossmann. A structural optimization approach in process

synthesis - ii: Heat recovery networks. Computers & Chemical Engineering, 7(6):707 –
721, 1983. ISSN 0098-1354. doi: [Link] URL
[Link]

[161] S. A. Papoulias and I. E. Grossmann. A structural optimization approach in process

synthesis - iii: Total processing systems. Computers & Chemical Engineering, 7(6):723 –
734, 1983. ISSN 0098-1354. doi: [Link] URL
[Link]

[162] Passivhaus Institut. Kriterien und Algorithmen für die Zertifizierte Passivhaus-
Komponente DuschwasserWärmerückgewinnung. Technical report, Passivhaus Institut,
2014. URL [Link]
pdf. In German, last accessed: 09/12/2015.

[163] Passivhaus Institut. Component database - drain water heat recovery, 12 2016. URL
[Link] Last
accessed: 18/04/2016.

[164] D. Peinado, M. de Vega, N. Garcia-Hernando, and C. Marugán-Cruz. Energy and exergy

analysis in an asphalt plant’s rotary dryer. Applied Thermal Engineering, 31(6 - 7):1039 –
1049, 2011. ISSN 1359 - 4311. doi: [Link]
029. URL [Link]

[165] M. Perpar, Z. Rek, S. Bajric, and I. Zun. Soil thermal conductivity prediction for dis-
trict heating pre-insulated pipeline in operation. Energy, 44(1):197 – 210, 2012. ISSN
0360-5442. doi: [Link] URL [Link]
[Link]/science/article/pii/S0360544212004884.

[166] S. Perry, J. J. Klemeš, and I. Bulatov. Integrating waste and renewable energy to reduce
the carbon footprint of locally integrated energy sectors. Energy, 33(10):1489 – 1497,
2008. ISSN 0360-5442. doi: [Link] URL http:
//[Link]/science/article/pii/S0360544208000911.

[167] T. Persson. Dishwasher and washing machine heated by a hot water circulation loop.
Applied Thermal Engineering, 27(1):120 – 128, 2007. ISSN 1359-4311. doi: [Link]
[Link]/10.1016/[Link].2006.05.005. URL [Link]
science/article/pii/S1359431106001694.

127
Bibliography

[168] T. Persson and M. Roennelid. Increasing solar gains by using hot water to heat dish-
washers and washing machines. Applied Thermal Engineering, 27(2 - 3):646 – 657, 2007.
ISSN 1359-4311. doi: [Link] URL
[Link]

[169] U. Persson and M. Münster. Current and future prospects for heat recovery from waste
in european district heating systems: A literature and data review. Energy, 110:116 –
128, 2016. ISSN 0360-5442. doi: [Link] URL
[Link]

[170] U. Persson and S. Werner. District heating in sequential energy supply. Applied Energy,
95(0):123 – 131, 2012. ISSN 0306-2619. doi: [Link]
02.021. URL [Link]

[171] E. Pieterse-Quirijns, E. Blokker, E. Van Der Blom, and J. c. Vreeburg. Modelling charac-
teristic values for non-residential water demand. In Water Distribution System Analysis
2010, Tucson, Arizona, USA, Sept. 12-15, 2010, 2010.

[172] E. Pieterse-Quirijns, E. Blokker, E. Van Der Blom, and J. c. Vreeburg. Non-residential

water demand model validated with extensive measurements and surveys. Drinking
Water Engineering and Science, 6(2):99–114, 2013. doi: [Link]
dwes-6-99-2013. URL [Link]

[173] S. Päivärinne and M. Lindahl. Exploratory study of combining integrated product and
services offerings with industrial symbiosis in order to improve excess heat utiliza-
tion. Procedia {CIRP}, 30:167 – 172, 2015. ISSN 2212-8271. doi: [Link]
10.1016/[Link].2015.02.101. URL [Link]
S2212827115001778.

[174] PostGIS Project. PostGIS, 2015. URL [Link]

[175] PostgreSQL. PostgreSQL: an open source object-relational database system, 2015. URL
[Link]

[176] L. Postrioti, G. Baldinelli, F. Bianchi, G. Buitoni, F. D. Maria, and F. Asdrubali. An

experimental setup for the analysis of an energy recovery system from wastewater
for heat pumps in civil buildings. Applied Thermal Engineering, 102:961 – 971, 2016.
ISSN 1359-4311. doi: [Link] URL
[Link]

[177] Privatinstitut für Transparenz im Gesundheitswesen GmbH. Deutsches Senioren-

portal, 2015. URL [Link] In German, last accessed:
17/09/2015.

[178] QGIS Development Team. QGIS Geographic Information System. Open Source Geospa-
tial Foundation, 2009. URL [Link]

128
 Bibliography

[179] R Core Team. R: A Language and Environment for Statistical Computing. R Foundation
for Statistical Computing, Vienna, Austria, 2015. URL [Link]

[180] M. Ramadan, T. lemenand, and M. Khaled. Recovering heat from hot drain water -
experimental evaluation, parametric analysis and new calculation procedure. 128:575 –
582, 2016. ISSN 0378-7788. doi: [Link] URL
[Link]

[181] K. Rathnayaka, H. Malano, S. Maheepala, B. c. George, B. Nawarathna,

M. Arora, and P. Roberts. Seasonal demand dynamics of residential water end-
uses. Water (Switzerland), 7(1):202–216, 2015. doi: 10.3390/w7010202. URL
[Link]
md5=299b6245f833ffaa8a4267f17896868a.

[182] C. F. Reinhart and C. C. Davila. Urban building energy modeling Ű a review of a nascent
field. Building and Environment, 97:196 – 202, 2016. ISSN 0360-1323. doi: [Link]
[Link]/10.1016/[Link].2015.12.001. URL [Link]
article/pii/S0360132315003248.

[183] C. P. Richter. Usage of dishwashers: observation of consumer habits in the domestic

environment. International Journal of Consumer Studies, 35(2):180–186, 2011. ISSN
1470-6431. doi: 10.1111/j.1470-6431.2010.00973.x. URL [Link]
1470-6431.2010.00973.x.

[184] A. D. Rosa, R. Boulter, K. Church, and S. Svendsen. District heating (dh) network design
and operation toward a system-wide methodology for optimizing renewable energy
solutions (smores) in canada: A case study. Energy, 45(1):960 – 974, 2012. ISSN 0360-5442.
doi: [Link] URL [Link]
com/science/article/pii/S0360544212005142.

[185] D. Saker, M. Vahdati, P. Coker, and S. Millward. Assessing the benefits of domestic hot
fill washing appliances. Energy and Buildings, 93:282 – 294, 2015. ISSN 0378-7788. doi:
[Link] URL [Link]
science/article/pii/S0378778815001322.

[186] M. Sameti and F. Haghighat. Optimization approaches in district heating and cooling
thermal network. Energy and Buildings, 140:121 – 130, 2017. ISSN 0378-7788. doi:
[Link] URL [Link]
science/article/pii/S0378778816312245.

[187] A. F. Sandvall, M. Börjesson, T. Ekvall, and E. O. Ahlgren. Modelling environmental

and energy system impacts of large-scale excess heat utilisation - a regional case study.
Energy, 79:68 – 79, 2015. ISSN 0360-5442. doi: [Link]
10.049. URL [Link]

129
Bibliography

[188] A. F. Sandvall, E. O. Ahlgren, and T. Ekvall. System profitability of excess heat utilisation
- a case-based modelling analysis. Energy, 97:424 – 434, 2016. ISSN 0360-5442. doi:
[Link] URL [Link]
science/article/pii/S0360544215016795.

[189] N. Schüler, A. Mastrucci, A. Bertrand, J. Page, and F. Maréchal. Heat demand estimation
for different building types at regional scale considering building parameters and urban
topography. Energy Procedia, 78:3403 – 3409, 2015. ISSN 1876-6102. doi: [Link]
org/10.1016/[Link].2015.11.758. URL [Link]
pii/S187661021502490X.

[190] F. Schmid. Sewage water: Interesting heat source for heat pumps and
chillers. SwissEnergy Agency for Infrastructure Plants, 2008. URL http:
//[Link]/sites/default/files/2008_EnergieinInfrastrukturanlagen_
SewageWaterInterestingheatsourceforheatpumpsandchillers_0.pdf.

[191] J. Schnieders. D5.1.12 drain water heat recovery in retrofits. Technical report, Passive
House Institute, 2015. URL [Link]
EuroPHit_D5.1.12_DWHR_PHI_b.pdf. Last accessed: 18/04/2016.

[192] E.-R. Schramek, editor. Taschenbuch für Heizung und Klimatechnik 07/08. Oldenbourg
Industrieverlag München, 73 edition, 2007.

[193] J. Söderman and P. Ahtila. Optimisation model for integration of cooling and heating
systems in large industrial plants. Applied Thermal Engineering, 30(1):15 – 22, 2010.
ISSN 1359-4311. doi: [Link] URL
[Link]

[194] Service des travaux municipaux. GIS data set, 2015. Obtained 06/2015.

[195] Servior. [Link], 2015. URL [Link] In German,

last accessed: 15/09/2015.

[196] D. Slys and S. Kordana. Financial analysis of the implementation of a drain water
heat recovery unit in residential housing. Energy and Buildings, 71:1 – 11, 2014. ISSN
0378-7788. doi: [Link] URL [Link]
[Link]/science/article/pii/S0378778813008062.

[197] R. Spur, D. Fiala, D. Nevrala, and D. Probert. Influence of the domestic hot-water daily
draw-off profile on the performance of a hot-water store. Applied Energy, 83(7):749 –
773, 2006. ISSN 0306-2619. doi: [Link] URL
[Link]

[198] P. Stadler. Assessment of energy requirements of commercial and industrial facilities in

Luxembourg. Master thesis, 2014.

130
 Bibliography

[199] Statec. Population census 2011, 2011. URL [Link]

lu/en/methodology/methodes/population-employment/Pop-logement/
recensement-population/[Link]?highlight=census. Last accessed: 09/01/2015.

[200] Statec. Population census: household appliances on the february 1st , 2011, 2011.
URL [Link]
8604. Last accessed: 09/01/2015.

[201] Statec. Infrastructures pour personnes agées 2002 - 2012, 2013. URL [Link]
[Link]. In French, last accessed: 15/09/2015.

[202] Statec. Household budget survey 2008-2013, 2015. URL [Link]

[Link]/en/surveys/espace-households/budget-household/[Link]. Last ac-
cessed: 11/12/2014.

[203] Statec. Composition of the natural gas price 2008-2015, 2016. URL [Link]
[Link]. Last accessed: 14/11/2016.

[204] Statec. Composition of the electricity price 2008-2015, 2016. URL [Link]
[Link]. Last accessed: 15/12/2016.

[205] Statistik Austria. Ausstattungsgrad der Haushalte - Bundesländerergebnisse, 2011.

URL [Link]
ausstattung_privater_haushalte/[Link]. In German, last accessed: 11/09/2015.

[206] Statistisches Bundesamt. Ausstattung privater Haushalte mit elektrischen Haushalts-

und sonstigen Geräten - Deutschland, 2015. URL [Link]
DE/ZahlenFakten/GesellschaftStaat/EinkommenKonsumLebensbedingungen/
AusstattungGebrauchsguetern/Tabellen/Haushaltsgeraete_D.html. In German, last
accessed: 08/09/2015.

[207] M. Z. Stijepovic and P. Linke. Optimal waste heat recovery and reuse in industrial
zones. Energy, 36(7):4019 – 4031, 2011. ISSN 0360-5442. doi: [Link]
1016/[Link].2011.04.048. URL [Link]
S0360544211003136.

[208] V. Stijepovic, P. Linke, M. Stijepovic, M. Kijevčanin, and S. Šerbanović. Targeting and

design of industrial zone waste heat reuse for combined heat andăpower generation. En-
ergy, 47(1):302 – 313, 2012. ISSN 0360-5442. doi: [Link]
09.018. URL [Link]
Asia-Pacific Forum on Renewable Energy 2011.

[209] F. T. Sun, C. J. Li, R. X. Wang, and D. Y. Li. Waste heat recovery systems of shampoo drain
water from barbershops. Applied Mechanics and Materials, 577:564–567, July 2014. doi:
10.4028/[Link]/AMM.577.564.

131
Bibliography

[210] I.-L. Svensson, J. Jönsson, T. Berntsson, and B. Moshfegh. Excess heat from kraft pulp
mills: Trade-offs between internal and external use in the case of Sweden - part 1:
Methodology. Energy Policy, 36(11):4178 – 4185, 2008. ISSN 0301-4215. doi: http:
//[Link]/10.1016/[Link].2008.07.017. URL [Link]
article/pii/S0301421508003637.

[211] Switzerland Tourism. [Link], 2015. URL [Link]

com. Last accessed: 15/09/2015.

[212] K. Tanha, A. S. Fung, and R. Kumar. Performance of two domestic solar water heat-
ers with drain water heat recovery units: Simulation and experimental investiga-
tion. Applied Thermal Engineering, 90:444 – 459, 2015. ISSN 1359-4311. doi: http:
//[Link]/10.1016/[Link].2015.07.038. URL [Link]
com/science/article/pii/S135943111500722X.

[213] J. Tarres-Font. Energy efficiency in electric steelmaking: a detailed energy audit of an

entire plant focusing on waste heat recovery. PhD thesis, 2013.

[214] J. Thorsen and H. Kristjansson. Cost considerations on storage tank versus heat
exchanger for hot water preparation. In 10th International Symposium on District
Heating and Cooling, 2006, Hanover, Germany, Sept. 3-5, 2006, 2006. URL http:
//[Link]/PCMPDF/VFKQA202_Cost_considerations.pdf. Last accessed:
09/03/2015.

[215] H. Tol and S. Svendsen. Improving the dimensioning of piping networks and network
layouts in low-energy district heating systems connected to low-energy buildings: A
case study in roskilde, denmark. Energy, 38(1):276 – 290, 2012. ISSN 0360-5442. doi:
[Link] URL [Link]
science/article/pii/S0360544211007961.

[216] T. R. Tooke, M. van der Laan, and N. C. Coops. Mapping demand for residential building
thermal energy services using airborne lidar. Applied Energy, 127(0):125 – 134, 2014.
ISSN 0306-2619. doi: [Link] URL http:
//[Link]/science/article/pii/S0306261914002657.

[217] S. Torras, C. Oliet, J. Rigola, and A. Oliva. Drain water heat recovery storage-type
unit for residential housing. Applied Thermal Engineering, 103:670 – 683, 2016. ISSN
1359-4311. doi: [Link] URL http:
//[Link]/science/article/pii/S1359431116305749.

[218] R. Turton, R. Bailie, W. Whiting, J. Shaeiwitz, and D. Bhattacharyya. Analysis, synthesis,

and design of chemical process, Fourth edition. Pearson Education, 2013.

[219] T.-M. Tveit, J. Aaltola, T. Laukkanen, M. Laihanen, and C.-J. Fogelholm. A framework for
local and regional energy system integration between industry and municipalities - case
study upm-kymmene kaukas. Energy, 31(12):2162 – 2175, 2006. ISSN 0360-5442. doi:

132
 Bibliography

[Link] URL [Link]

science/article/pii/S0360544205001830.

[220] UN Habitat. Cities and climate change initiative - launch and conference report. 2009.
URL [Link] Last
accessed: 12/11/2015.

[221] United Nations. Transforming our world: The 2030 agenda for sustainable devel-
opment, 2015. URL [Link]
[Link]. Last accessed: 20/01/2017.

[222] P. Upham and C. Jones. Don’t lock me in: Public opinion on the prospective use of waste
process heat for district heating. Applied Energy, 89(1):21 – 29, 2012. ISSN 0306-2619.
doi: [Link] URL [Link]
com/science/article/pii/S0306261911001309.

[223] U.S. Energy Information Administration. International energy outlook 2016, 2016. Last
accessed: 23/01/2017.

[224] P. S. Varbanov. 2 - basic process integration terminology. In J. J. Klemeš, editor,

Handbook of Process Integration (PI), Woodhead Publishing Series in Energy, pages
28 – 78. Woodhead Publishing, 2013. ISBN 978-0-85709-593-0. doi: [Link]
org/10.1533/9780857097255.1.28. URL //[Link]/science/article/pii/
B9780857095930500022.

[225] P. S. Varbanov and J. J. Klemeš. Integration and management of renewables into total
sites with variable supply and demand. Computers and Chemical Engineering, 35(9):
1815 – 1826, 2011. ISSN 0098-1354. doi: [Link]
02.009. URL [Link]

[226] P. S. Varbanov, Z. Fodor, and J. J. Klemeš. Total site targeting with process specific
minimum temperature difference (dtmin). Energy, 44(1):20 – 28, 2012. ISSN 0360-5442.
doi: [Link] URL [Link]
com/science/article/pii/S0360544211008504.

[227] L. Čuček, P. Varbanov, J. J. Klemeš, and Z. Kravanja. Multi-objective regional total site
integration. Chemical Engineering Transactions, 35:97–102, 2013. ISSN 1974-9791. doi:
[Link] URL [Link]/cet/13/35/[Link].

[228] VDI Gesellschaft. VDI-Wärmeatlas. Springer Berlin Heidelberg, 10th edition, 2006.

[229] D. G. Victor, D. Zhou, E. H. M. Ahmed, P. K. Dadhich, J. G. J. Olivier, H.-H. Rogner,

K. Sheikho, and M. Yamaguchi. Climate Change 2014: Mitigation of Climate Change.
Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental
Panel on Climate Change: Introductory Chapter. Cambridge University Press, 2014. URL
[Link]

133
Bibliography

[230] Vienna Tourist Board. List of all accommodations, 2015. URL [Link]
en/hotel/suche. Last accessed: 17/09/2015.

[231] S. B. Viklund and M. T. Johansson. Technologies for utilization of industrial excess

heat: Potentials for energy recovery and {CO2} emission reduction. Energy Conver-
sion and Management, 77:369 – 379, 2014. ISSN 0196-8904. doi: [Link]
10.1016/[Link].2013.09.052. URL //[Link]/science/article/pii/
S0196890413006134.

[232] J. Vlachopoulos and J. R. Wagner. The SPE Guide on Extrusion Technology and
Troubleshooting. Society of Plastics Engineers, Brookfield, CT, 2001.

[233] M. H. Wahlstrom and B. Harsman. Residential energy consumption and conservation.

Energy and Buildings, 102:58–66, 2015. ISSN 03787788. doi: 10.1016/[Link].2015.05.
008. URL [Link]

[234] J. Wallin and J. Claesson. Analyzing the efficiency of a heat pump assisted drain water
heat recovery system that uses a vertical inline heat exchanger. Sustainable Energy
Technologies and Assessments, 8:109 – 119, 2014. ISSN 2213-1388. doi: [Link]
org/10.1016/[Link].2014.08.003. URL [Link]
pii/S2213138814000733.

[235] C. Walsh and P. Thornley. Barriers to improving energy efficiency within the process
industries with a focus on low grade heat utilisation. Journal of Cleaner Production, 23
(1):138 – 146, 2012. ISSN 0959-6526. doi: [Link]
URL //[Link]/science/article/pii/S0959652611004240.

[236] Y. Wang, X. Feng, and K. H. Chu. Trade-off between energy and distance related costs
for different connection patterns in heat integration across plants. Applied Thermal
Engineering, 70(1):857 – 866, 2014. ISSN 1359-4311. doi: [Link]
applthermaleng.2014.06.012. URL [Link]
S135943111400489X.

[237] C. Weber. Multi-objective design and optimization of district energy systems including
polygeneration energy conversion technologies. PhD thesis, EPFL, Lausanne, Switzerland,
2008. URL [Link]

[238] J. Wen. Heat Capacities of Polymers, chapter 9, pages 145–154. Springer, 2 edition, 2007.

[239] J. Widen, M. Lundh, I. Vassileva, E. Dahlquist, K. Ellegard, and E. Wackelgard. Con-

structing load profiles for household electricity and hot water from time-use data
- modelling approach and validation. Energy and Buildings, 41(7):753 – 768, 2009.
ISSN 0378-7788. doi: [Link] URL http:
//[Link]/science/article/pii/S0378778809000413.

[240] L. Wong, K. Mui, and Y. Guan. Shower water heat recovery in high-rise residential
buildings of hong kong. Applied Energy, 87(2):703 – 709, 2010. ISSN 0306-2619. doi:

134
 Bibliography

[Link] URL [Link]

science/article/pii/S0306261909003225.

[241] J. Xia, Y. Li, H. Fang, Y. Su, and Y. Jiang. Case study on industrial surplus heat of
steel plants for district heating in Northern China. Energy, 102:397 – 405, 2016. ISSN
0360-5442. doi: [Link] URL [Link]
[Link]/science/article/pii/S0360544216301670.

[242] R. Yao and K. Steemers. A method of formulating energy load profile for domestic
buildings in the {UK}. Energy and Buildings, 37(6):663 – 671, 2005. ISSN 0378-7788. doi:
[Link] URL [Link]
science/article/pii/S037877880400307X.

[243] M. Zawrah. Synthesis and characterization of wc - co nanocomposites by novel chemical

method. Ceramics International, 33(2):155 – 161, 2007. ISSN 0272-8842. doi: [Link]
[Link]/10.1016/[Link].2005.09.010. URL [Link]
article/pii/S0272884205002567.

135
Bibliography

136
 Alexandre Bertrand
Grand-Duchy of Luxembourg
[Link]

Key skills

Scientific and technical skills:

- Integrated material and energy assessment - Consulting and training in the field of
& optimisation energy
- Database management - Project and team management
- Mathematical optimisation

Language skills: French (native), fluent in written and spoken English, German and Luxem-
burgish

Software skills: AMPL, Postgres, QGIS, Lyx, MS and Open Office

Professional experience

2017 Luxembourg Institute of Science and Technology, Luxembourg, Research and

technology associate:
- R&D in the field of energy efficiency, industrial symbiosis and circular economy,
- Project definition and management.

2013-2016 Luxembourg Institute of Science and Technology, Luxembourg / Ecole Polytech-

nique Fédérale de Lausanne, Switzerland, PhD candidate:
- Development of data characterisation, assessment and optimisation methods
supporting the valorisation of waste heat at building, city and regional scale
(Luxembourgish National Research Fund AFR grant).

2003-2013 Luxembourg Institute of Science and Technology (former Public Research Centre
Henri Tudor), Luxembourg, Senior R&D engineer:
- European and national R&D projects in the field of energy assessment meth-
odology development ([Link], [Link], [Link]-
[Link], [Link] – project manager on SmartHeatFlow, EINSTEIN2,
SEMS, C2C-Bizz),
- Consulting of Luxemburgish ministries and companies in the field of energy
assessment, energy optimisation and best available technologies,
 - Trainings of process engineers, consultants and students in the field of energy
audits and optimisation in industries.

2002-2003 Bureau Goblet Lavandier et associés, Luxembourg, Engineer:

- Energy concepts for municipalities in Luxembourg,
- Integrated Pollution Prevention Control (IPPC) authorisation requests for com-
panies.

Education

2004-2011 Degree in mechanical engineering (Dip. Ing. TU), specialisation in energy,

Dresden University of Technology, Germany (distance courses, exams with
normal students). Main courses: thermodynamics, thermal energy plants,
cooling systems, heat exchangers, compressors and pumps. Final thesis topic:
exergy assessment of a trigeneration plant in Esch/Alzette (LU)

1997-2001 Degree in electrical engineering (Ing. Ind.), specialisation in energy and sys-
tem automation, Luxembourg University of applied sciences, Grand-Duchy
of Luxembourg. Main courses: electrical energy, electricity distribution, sys-
tem automation, electrical motors. Final thesis topic: energy concept of 7
municipalities in Luxembourg

Recent publications

2017 A. Bertrand, R. Aggoune, and F. Maréchal. In-building waste water heat recov-
ery: an urban-scale method for the characterisation of water streams and the
assessment of energy savings and costs. Applied Energy, 192:110 – 125, 2017.
[Link]
A. Bertrand, A. Mastrucci, N. Schüler, R. Aggoune, and F. Maréchal. Character-
isation of domestic hot water end-uses for integrated urban thermal energy
assessment and optimisation. Applied Energy, 186, Part 2:152 – 166, 2017.
[Link]

2016 A. Bertrand, R. Aggoune, and F. Maréchal. In-building waste water heat re-
covery: an urban scale method for energy saving assessments. In 29th inter-
national conference on Efficiency, Cost, Optimization, Simulation and Envir-
onmental Impact of Energy Systems (ECOS), Portoroz, Slovenia, June 19-23
2016.

2015 N. Schüler, A. Mastrucci, A. Bertrand, J. Page, and F. Maréchal. Heat demand

estimation for different building types at regional scale considering building
 parameters and urban topography. Energy Procedia, 78:3403 – 3409, 2015.
[Link]
A. Bertrand, R. Aggoune, and F. Maréchal. Residential and non-residential hot
water use models for urban and regional energy integration. In Biennial Inter-
national Workshop Advances in Energy Studies - Energy and urban systems,
Stockholm, Sweden, May 4-7 2015.

Award

Winner of the 2014 Secolux award, supporting research and innovation in the field of construc-
tion, for the elaboration of a heat map of Luxembourg on industrial excess heat potential and
household heat demand in the framework of the ERDF SmartHeatFlow project.

Personal interests

Member of the local theatre association, cooking, climbing, running, reading, art.