Preface

Abstract

Keywords: Urban systems; Energy optimisation; Clustering; Geographical database;

The global context of climate change is pushing towards energy savings. Cities are home to most of the world’s population, consume 2/3 of total energy consumption and are responsible for 70% of greenhouse gas emissions. In this context, interest in Urban Systems has been growing due to the necessary decarbonisation of city energy systems. It is expected to be driven by high electrification and the emergence of prosumers, both taking and providing electricity to the grid. Decision tools are developed using mathematical optimisation to help the transition. However, those tools need coherent input and a lot of computational power. This work aims to answer those two problems on the level of Switzerland. First, a generic and coherent database containing building energy-related data is developed. Second, it proposes a method to typify the districts of a city. The districts are clustered according to their energetic characteristics, i.e. the energy demands (heating, hot water, electrical) and the endogenous resources (solar, biomass, geothermal). By grouping districts accordingly, a cluster of districts can be optimised by solving one from the group and considering the solution is similar enough for the rest. The method is applied to the canton of Geneva and two typifications are proposed. One divides the canton into two types of districts: residential buildings dominated districts, and non-residential buildings dominated ones. The second takes into account more parameters such as the endogenous resources and results in 22 clusters.

Acknowledgements

This master thesis has been done under the supervision of Mr. Dorsan Lepour, who has been not only of great help, but also a great partner with whom to work and share discussions. Thank you as well to Mr. Cédric Terrier for the proof reading.

I also want to thank Pr. François Maréchal, who catch me on the train for a intense but illuminating discussion on the project, giving me a direction for the remaining half of the project.
Thanks to all my colleagues at the IPESE laboratory making it an enjoyable place to work. Especially, thank you Raphaël Briguet for your help on QBuildings.

Finally, thank you to Benjamin, great friend of mine, for welcoming me at your place at Savièse and for the good times.

This master thesis also concludes my bachelor and master studies at EPFL and as such I would like to thank the people whom accompanied me during this journey.
Thank you to my first friends of the first year, the people of the Environmental Engineering years, the Bussinight group and to the Brosse team.
Thank you to my family. First of all, to my parents for guiding me through and from who I have learnt me so much; to my brother and sister for what we share; to my aunt, uncle and cousins in Geneva who welcomed me as a full family member.
Finally, thank you of course to Elisa, who shared most of the journey with me.

Glossary

AIC
Akaike Information Criterion
BIC
Bayesian Information Criterion
CH
Calinski-Harabasz
DHW
Domestic Hot Water
EUD
End Use Demand
ERA
Energetic Reference Area
GIS
Geographical Information System
GMM
Gaussian Mixture Model
KPI
Key Performance Indicator
MILP
Mixed Integer Linear Programming
MOO
Multi-Objective Optimisation
PCA
Principal Components Analysis
REHO
Renewable Energy Hub Optimiser
RegBL
Register des bâtiments et des logements { Buildings and accommodations register }
SH
Space Heating
UUID
Universally Unique Identifier
REA
Reference Energy Area

Note for the reader
This report is built to be an interactive format. The reader is welcome to manipulate Figures and Tables to its convenience. For the format to be fully functional, an internet connection is required, otherwise some features would not be able to display properly. If no internet connection is at your disposition, you may read the report on its pdf version, even though it is not the format in which it is meant to be read.

© EPFL-IPESE 2022

Master thesis, Spring 2022

Joseph Loustau