Urban energy vulnerability index assessment at building level

2025

Presenter: G. Mor

Collaborators: J.M. Broto, M. Sellart, A. Pariente, A. Deserteaux, A. Comas, A. Marí, J. Cipriano

Urban energy vulnerability index assessment at building level

Introduction

Urban energy vulnerability index assessment at building level

Energy vulnerability is a subjective combination of risk factors

in homes that can lead to energy poverty:

Technological
Social
Economical
Physical

What is energy vulnerability?

/ Introduction

Urban energy vulnerability index assessment at building level

Technological factors:

Issues related to outdated or inefficient technology within households, such as old appliances or inadequate insulation, contribute to energy poverty.

(Typology of the HVAC systems, domestic appliances, and their usage)

What is energy vulnerability?

/ Introduction

Urban energy vulnerability index assessment at building level

Social Factors:

Socioeconomic circumstances, including income levels, household size, and social support networks, play a significant role in determining vulnerability to energy poverty.

(Population and households characteristics, health conditions)

What is energy vulnerability?

/ Introduction

Urban energy vulnerability index assessment at building level

Economic Factors:

Financial constraints, such as high energy costs relative to income or inability to afford energy-efficient upgrades, exacerbate the risk of energy poverty for households.

(Household prices, salary indexes, energy costs, CPI)

What is energy vulnerability?

/ Introduction

Urban energy vulnerability index assessment at building level

Physical Factors:

Extreme weather events, seasonal variations in energy demand, and inadequate climate control measures can increase vulnerability to energy poverty, especially in regions prone to harsh climates.

Physical attributes of the building, such as its age, construction quality, and energy efficiency features, influence energy consumption and vulnerability to energy poverty among occupants.

(Building envelope characteristics, weather conditions, air conditions)

What is energy vulnerability?

/ Introduction

How to assess the vulnerability of citizens?

Through the implementation of a methodology that integrates:
- Energy thermal building-stock modelling
- Weather model to upscale mesoscale resolution to microlocal
- Integration of interdisciplinary and highly heterogeneous data sources
- Harmonisation of the datasets in a common database framework
- Calculation and estimation of Key Performance Indicators (KPIs) related to each building
- Visualization of KPIs over a web interface

Urban energy vulnerability index assessment at building level

/ Introduction

What is CLIMATE READY BARCELONA?

ICLEI Action Fund 2.0 project (Budget: 1,000,000 €)

Support citizens and public authorities to anticipate and adapt to climate-change effects and the energy crisis

Implemented from July 2023 to June 2025
Partners: Ecoserveis, ABD, BitGenoma, CIMNE, Barcelona City Council

Urban energy vulnerability index assessment at building level

General architecture

Urban energy vulnerability index assessment at building level

Data platform for the vulnerability map

/ General architecture

Urban energy vulnerability index assessment at building level

Data

Urban energy vulnerability index assessment at building level

Cadaster:

Attributes: Geometry, built areas, number of floors, number of dwellings, main usage
Geolocation level: Building
Source: National Cadaster datasets harmonised to INSPIRE (link) and CAT file (link)
Format: Vectorial file (GML)

Datasets considered

/ Data

Attributes: Energy demand, consumption, performance qualification, thermal systems and envelope characteristics
Geolocation level: Dweling/Building
Source: Catalonia Open Data Platform (link)
Format: Parquet

Energy Performance Certificates:

Urban energy vulnerability index assessment at building level

Attributes: 2017 consumption, number of metering devices, street name and street number
Geolocation level: Building (when more than 3 metering devices)
Source: Barcelona Metropolitan Area
Format: Excel file (XLSX)

Datasets considered

/ Data

Attributes: 2022-2023 monthly consumption, number of metering devices, street name and street number
Geolocation level: Building (when more than 3 metering devices) located in specific postal codes. (~7000/69000 buildings)
Source: e-distribución
Format: Excel file (XLSX)

Annual electricity consumption:

Monthly electricity consumption:

Urban energy vulnerability index assessment at building level

Attributes: 2017-2023 annual consumption, number of metering devices, street name and street number
Geolocation level: Building
Source: Barcelona Metropolitan Area and Nedgia
Format: Excel file (XLSX)

Datasets considered

/ Data

Attributes: 2019-2024 average hourly consumption, economic sector, tariff
Geolocation level: postal code
Source: Datadis
Format: JSON through API

Annual gas consumption:

Aggregated daily load pattern of electricity consumption:

Urban energy vulnerability index assessment at building level

Attributes: 2008-2024 temperature, humidity, wind speed, GHI...
Geolocation level: ~10 km grid
Source: COPERNICUS (Dataset: ERA5Land)
Format: GRIB

Datasets considered

/ Data

Attributes: 2008-2024 temperature, humidity, wind speed, GHI...
Geolocation level: ~5 km grid
Source: MeteoGalicia (Thredds server)
Format: NetCDF

Mesoscale historical weather data:

Mesoscale forecasted weather data:

Urban energy vulnerability index assessment at building level

Attributes: 2008-2017 temperature, humidity, wind speed
Geolocation level: 100m grid
Source: COPERNICUS (Dataset: UrbClim)
Format: NetCDF

Datasets considered

/ Data

Attributes: incomes, incomes sources, demography stats, population, ...
Geolocation level: Census tract, district or municipal
Source: INE (Spanish National Statistics Institute)
Format: Parquet

Hyper-detailed historical weather data:

Socio-economic indicators:

Urban energy vulnerability index assessment at building level

Climate Shelters:

Attributes: name, type of building, time schedule, creation date...
Geolocation level: Building
Source: OpenDataBCN (link)
Format: Comma-separated Values file (CSV)

Datasets considered

/ Data

Postal code administrative layer:

Attributes: Geometry, postal code
Geolocation level: postal code
Source: Spanish postal codes (link)
Format: Vectorial file (KML)

Urban energy vulnerability index assessment at building level

Census tracts administrative layer:

Attributes: Geometry, census tract code
Geolocation level: Census tract
Source: INE (link)
Format: Vectorial file (GPKG)

Datasets considered

/ Data

Districts administrative layer:

Attributes: Geometry, district names and code
Geolocation level: District
Source: INE (link)
Format: Vectorial file (SHP)

Urban energy vulnerability index assessment at building level

Municipalities administrative layer:

Attributes: Municipality name and code
Geolocation level: Municipality
Source: INE (link)
Format: Vectorial file (SHP)

Datasets considered

/ Data

Neighbourhood administrative layer:

Attributes: Neighbourhood name and code
Geolocation level: Neighbourhood
Source: OpenDataBCN (link)
Format: Comma-separated Values file (CSV)

Urban energy vulnerability index assessment at building level

Attributes: Vegetation Index
Geolocation level: 1m-grid
Source: ICGC (link)
Format: Raster File (TIFF)

Datasets considered

/ Data

Tourism-related establishments:

Attributes: Location, type of establishment, category, total capacity and stat
Geolocation level: Building
Source: Catalan Government
Format: Excel file (XLSX)

Normalised Difference Vegetation Index (NDVI):

Urban energy vulnerability index assessment at building level

Digital Elevation Model (DEM):

Attributes: Average sea-level
Geolocation level: 30m-grid
Source: COPERNICUS
Format: Raster File (TIFF)

Datasets considered

/ Data

Attributes: Type of establishment, activity and status
Geolocation level: Building
Source: OpenDataBCN (link)
Format: JSON through API

Ground-floor premises census for economic activity:

Urban energy vulnerability index assessment at building level

Buildings Technical Inspections:

Attributes: Description of the building, equipment, defects and deficiencies
Geolocation level: Building
Source: Gencat
Format: XML file

Datasets considered

/ Data

Mortality and morbidity due to extreme heat events:

Attributes: 2018-2023 mortality indexes, number of people attended in primary care...
Geolocation level: Health basic area (Similar to districts)
Source: PADRIS office of the Catalan Government
Format: Comma-separated Values file (CSV)

Urban energy vulnerability index assessment at building level

Datasets considered

/ Data

Attributes: Information gathered through the vulnerability surveys in LaMevaEnergia APP
Geolocation level: Building
Source: La Meva Energia APP
Format: JSON file through API

Building-aggregated vulnerability surveys:

Urban energy vulnerability index assessment at building level

All these highly heterogenous datasets conform the Knowledge Graph of the project.

Ingestion processes

Manually or periodically executed

Reading from webs, files, external databases or APIs

Implemented in Python scripts

Harmonisation processes

All ingested datasets go through a transformation process to align them to the BIGG ontology

Store the data to the databases

Implemented in Python and using RML.io functionalities.

Ingestion to Database

/ Data

Urban energy vulnerability index assessment at building level

BIGG: an ontology to rule them all

/ Data

Urban energy vulnerability index assessment at building level

BIGG: Reusing Ontologies

/ Data

saref, s4blg, s4city, s4agri:
- buildings, devices
ssn:
- systems, deployments
geosp:
- geolocation
qudt:
- units

Challenges and Benefits of Reusability and Inference in the Urban Modeling Context

Urban energy vulnerability index assessment at building level

BIGG: Reusability

/ Data

Challenges and Benefits of Reusability and Inference in the Urban Modeling Context

We realized that we need reused ontology to take profit of the advantages that offer ontologies.
It also allow us to use it in many projects like: Bigg, Entrack, Eplanet, ClimateReady, Dedalus, Climres, infraestructuras, SIME... thanks to the flexibility of a well defined ontology.

Urban energy vulnerability index assessment at building level

Graph databases are not inherently mandatory when using data ontologies, but they can be highly beneficial in certain scenarios.

Ontologies represent structured knowledge, emphasizing relationships between entities.

Graph databases excel in modeling and querying complex relationships, making them a natural fit for ontological data.

Graph database

/ Data

Urban energy vulnerability index assessment at building level

Flexible Schema:

Graph databases allow for adaptable data models, accommodating evolving ontologies.

Efficient Relationship Representation:

Graph databases naturally model complex relationships, facilitating querying and traversal.

Scalability:

Graph databases efficiently handle large-scale ontologies, offering scalability as data grows.

Powerful Querying:

Graph databases provide robust querying capabilities, including traversing relationships effectively.

Semantic Reasoning:

Some graph databases support semantic reasoning, enabling inference and deduction based on

ontological data.

Graph database

/ Data

Urban energy vulnerability index assessment at building level

Neo4J for graph database management:

It offers a flexible schema, efficient relationship representation
Scalable and powerful querying capabilities
It aligns perfectly with ontological requirements.

SHACL for data validation:

It enables us to specify constraints on our ontological data, ensuring
- Integrity
- Consistency
- Compliance with predefined rules.

Graph database

/ Data

Urban energy vulnerability index assessment at building level

Our research adopts a specialized time series database instead of integrating time series data directly into the graph database

Time series database

/ Data

Performance:

Finely tuned for storing and querying time-stamped data

They have indexing strategies for time-based queries, optimizing storage and retrieval.

Scalability:

Designed for handling massive time series data volumes efficiently, ensuring future scalability.

Build-in Functionalities:

Offers built-in functionalities like downsampling, aggregation, and windowing operations

Urban energy vulnerability index assessment at building level

Apache Druid for storing time series data:

Open-source distributed data store for real-time analytics.
Specialized for handling large volumes of time-stamped data.
Features include real-time ingestion, columnar storage, and scalability.
Provides fast query performance with SQL-like query language.
Widely used for monitoring, ad hoc analysis, and operational intelligence.

Time series database

/ Data

Urban energy vulnerability index assessment at building level

By segregating time series from other type of data, we maintain the graph database's focus on managing ontological relationships and complex network structures.

This architectural choice ensures optimal performance, scalability, and maintainability, harnessing each database technology's strengths for its intended role.

Combination of graph and time series databases

/ Data

Urban energy vulnerability index assessment at building level

Modelling

Urban energy vulnerability index assessment at building level

Buildings Thermal Energy model

Simulation of the energy demand of buildings in the urban area, based on archetypes, construction types, local weather data and user behaviour patterns.

Weather upscaling resolution model

Prediction model to upscale meteorological data from mesoscale to microscale.

Graph Neural Network

General model to predict indicators at building level based on real measurements, location of buildings and their relation among several aggregation layers.

Multifaceted Models: Diverse Objectives

/ Modelling

Urban energy vulnerability index assessment at building level

Thermal demand energy model (using IREC's RC models)

/ Modelling

61670 buildings with main or partial residential use (88.7% all BCN buildings)
- Buildings with only residential use:
  - Plurifamiliar ≥ 32 dwellings: 1512 (2.45%)
  - Plurifamiliar when 32 > dwellings ≥ 12: 8474 (13.75%)
  - Plurifamiliar when dwellings < 12: 23345 (37.85%)
  - Unifamiliar: 2694 (4.37%)
- Buildings with partial residential use:
  - Plurifamiliar ≥ 32 dwellings: 4278 (6.93%)
  - Plurifamiliar when 32 > dwellings ≥ 12: 11418 (18.51%)
  - Plurifamiliar when dwellings < 12: 9781 (15.87%)
  - Unifamiliar: 168 (0.27%)

SHORT TERM

Cooling and heating demand simulation based on a calibrated Resistance and Capacitance Model per each building archetype, which is an electric circuit that imitates in a simplified manner the thermal dynamics of a building.

Urban energy vulnerability index assessment at building level

Thermal demand energy model (using IREC's RC models)

/ Modelling

Building characteristics:
- Context of nearby buildings
- Number of dwellings
- Built area per uses
- Effective year of construction
- Category / economic value
- Typology
Thermal energy systems
- Most feasible system

Weather conditions:
- Microlocal weather file
Demographics and socioeconomics*:
- People per household
- Incomes (sources and annual amount)
- Age range
- Rental Housing Price Index

(*aggregated to census tract or district)

SHORT TERM

Align each building to one of RC model archetypes

Urban energy vulnerability index assessment at building level

Thermal demand energy model (using IREC's RC models)

/ Modelling

Model objective
- Characterise actual situation of buildings,
  - Any retrofitting scenario will be evaluated
- Calculate Monthly KPIs:
  - Energy demand
    - Set-point heating temperatures: 19, 20, 21, 22ºC.
    - Set-points cooling temperatures: 26, 27, 28, 29ºC
  - Final energy consumption (assuming most feasible HVAC system)
    - Energy demand * System performance
  - % Indoor Overheating Hours (during summer in FF conditions, natural ventilating during nights)

SHORT TERM

Urban energy vulnerability index assessment at building level

Thermal demand energy model (using UConcordia simulation framework)

/ Modelling

1 - Select a subset of real buildings and their context

2 - Define building envelopes archetypes according to building code

3 - Define user behaviour patterns according to demographics and socioeconomic profiles

4 - Define building systems archetypes according to EPC and cadastral data

5 - Define microlocal weather input files

LONG TERM

Urban energy vulnerability index assessment at building level

Weather upscaling model

/ Modelling

Urban energy vulnerability index assessment at building level

Weather upscaling model

/ Modelling

Urban energy vulnerability index assessment at building level

SKG: Simplified Knowledge Graph

/ Modelling

Link (section 5.3)

Urban energy vulnerability index assessment at building level

Graph Neural Network: prediction and simulation

/ Modelling

Data communication and visualisation

Urban energy vulnerability index assessment at building level

Interoperability: DataAPI

/ Data communication and visualisation

Urban energy vulnerability index assessment at building level

- Tailored endpoints for adaptation to wide-range of application use cases

- Authenticated access

- Optimized for quick and reliable data access.

- It runs on kubernetes:

DataAPI diagram endpoints:
- get_all_indicators
- get_indicator
- post_model_request
- get_model_request
- get_model_request_predictions
- login

Interoperability: DataAPI

/ Data communication and visualisation

Urban energy vulnerability index assessment at building level

A Python library (streetmaps_ES) was developed to easily update a database with the Spanish municipalities street maps.
- Based on the latest street map version of the national cadaster (Link).
- Adding interesting metadata to each address: related postal code, census tract, neighborhood, district, parcel island, main usage of the building, and number of dwellings and floors in the building.
This library is used across several project implementations:
- La Meva Energia: Geolocation of the buildings related to app users.
- Vulnerability map: In harmonisation processes that need to geolocate data

Interoperability: Common geolocation format

/ Data communication and visualisation

Urban energy vulnerability index assessment at building level

Data sharing with La Meva Energia

/ Data communication and visualisation

Urban energy vulnerability index assessment at building level

User interface

/ Data communication and visualisation

Urban energy vulnerability index assessment at building level

User interface

/ Data communication and visualisation

Urban energy vulnerability index assessment at building level

User interface

/ Data communication and visualisation

Urban energy vulnerability index assessment at building level