Decarbonisation of cities through data-driven intelligence
BEE Group-CIMNE
BEE Group Research Lines
Statistical learning methods for energy assessment in buildings with applications at different geographic levels
Presentation content
BEE Group Research Lines
-
Artificial intelligence and big data at the BEE Group
-
Big data analytics for energy efficiency in buildings
-
Smart grids and DR at the district level
-
Energy transition and climate adaptation in communities
-
Tubular biodigesters
AI & big data at the BEE Group
BEE Group Research Lines
Incubation of SIE
Big data analytics for energy efficiency in buildings
Big data architecture
BEE Group Research Lines
Big data technologies
BEE Group Research Lines
Linked data framework
BEE Group Research Lines
Ontologies
Ontologies
Linked-data
Linked data framework
BEE Group Research Lines
Reuse of W3C ontologies (SAREF...)
- Including additional concepts (KPIs, EEM, etc.).
BEE Group Research Lines
Data valorization
Harmonized energy data in our architecture:
- 5,000 buildings of Generalitat
- 400 remotely controlled buildings from infraestructures.cat
- 1,000 buildings in Bulgaria
- 377 public buildings in the Czech Republic
- 2,000 public buildings in Greece
- 4,800 CUPS of public facilities in 67 municipalities in the Province of Girona
- 70,000 monthly energy bills residential buildings BCN
Data analytics and AI
BEE Group Research Lines
Data analytics and AI
BEE Group Research Lines
BIGGR package to standardize:
- Data cleaning processes
- Abnormal periods detection
- Outliers, vacation periods...
-
Data manipulation
- Interpolations
- Clusters
- Climate dependence...
- Baseline modeling
- Prediction models (day ahead)
Data analytics and AI
BEE Group Research Lines
What do we make with the data?
- Energy consumption prediction
- Model Predictive Control (MPC)
- PV Fault detection
- Building benchmarking
- Energy and economic impact evaluation of building Energy Efficiency Measures
Generic MPC scheme example
Smart grids and DR at the district level
DR at district level
BEE Group Research Lines
- User consumption modelization (aggregated)
- RLS Models with autoregresives and weather inference
- Low voltage electric grid modeling
- Electric models (detection of non-technical losses
- Optimization and DR in energy communities with centralized batteries
- Demand respond services to increase the flexibility of the low voltage grid
DR services to the Low Voltage grid
BEE Group Research Lines
Energy transition and climate adaptation in communities
Optimization of REC via energy allocation
BEE Group Research Lines
Motivation
Development of a tool to assist in the planning and operation of Renewable Energy Communities (ECs) seeking to:
1. Extract maximum potential of RES
2. Economic profitability for all participants
Optimization of REC via energy allocation
BEE Group Research Lines
Methodology
Optimization of REC via energy allocation
BEE Group Research Lines
How? GA with ordering and simplified inputs
Multi criteria optimization
Optimization of REC via energy allocation
BEE Group Research Lines
Case study: public buildings in BCN
128 buildings as participants; 7 solar PV installations
Comparison among 3 criteria:
1. Profitable: The selection of participants is random
The energy allocation is based on investment
1. Sustainable: The selection of participants is random
The energy allocation is based on investment
1. Optimized The selection of participants & energy
allocation is made using the optimization
Optimization of REC via energy allocation
BEE Group Research Lines
Conclusions
Climate vulnerability map of BCN
BEE Group Research Lines
Knowledge graph of the whole city
1. building level
2. census tract level
3. postal code level
Climate vulnerability map of BCN
BEE Group Research Lines
Heterogenous temporal graph
Train a Spatial-temporal Heterogeneous Graph Convolutional NN based on the knowledge graph
Estimate node atributes & predict new node attributes
Thanks for your attention
BEE Group