Diploma in GNNs for Flows, Energy and Traffic
Online postgraduate diploma of 800 hours in GNNs for flows, energy and traffic, with 8 modules covering physical flow network graphs, core GNN architectures, spatio-temporal traffic models, GNNs in power systems, water and heat networks, physics-informed learning and deployment in network operations, with an assigned technical tutor.
Take the controls.
Run a real engineering trade-off from this field. Move the sliders, watch the model react in real time and beat the challenge.
What you will be able to do
Model power, road and water networks as graphs with physical attributes.
Implement GCN, GAT and GraphSAGE architectures with PyTorch Geometric.
Develop spatio-temporal models for traffic forecasting.
Build power flow surrogate models with physical constraints.
Detect and locate leaks in water networks using GNNs.
Assess model robustness and uncertainty under topology changes.
Deploy graph models integrated with SCADA systems and control centres.
Who it's for
- Machine learning engineers working on infrastructure networks.
- Data scientists in electricity transmission and distribution.
- Mobility and traffic analysts building forecasting tools.
- Smart grid operations engineers.
- Network digital twin specialists.
- Researchers in graph learning looking for real-world applications.
- Engineering or science graduates aiming to join electricity transmission and distribution operators.
What you'll work with
- ISO/IEC 23053:2022 (framework for AI systems using machine learning)
- IEC 61970 (Common Information Model for energy management systems)
- IEEE C37.118.1 (synchrophasor measurements for power systems)
- IEC 61968 (application integration in electricity distribution)
- ISO/IEC 42001:2023 (AI management systems)
- Regulation (EU) 2024/1689 (Artificial Intelligence Act)
- EN 50160 (voltage characteristics of supplied electricity)
- PyTorch Geometric
- DGL (Deep Graph Library)
- pandapower
- EPANET
- SUMO
8 modules
1.1Power networks represented as bus graphs1.2Road networks: intersections, links and origin–destination matrices1.3Water and gas distribution networks1.4Adjacency, incidence and Laplacian matrices1.5Kirchhoff's laws and flow conservation1.6Graph signals and dynamic graphs1.7Data sources: SCADA, PMU and loop detectors1.8Gap imputation in spatio-temporal series1.9MAE, RMSE and MAPE metrics by horizon1.10Workshop: graph of the IEEE 118-bus network
2.1Message passing and aggregation functions2.2Spectral convolution and ChebNet filters2.3Kipf and Welling's GCN2.4GraphSAGE and neighbourhood sampling2.5Graph Attention Networks and attention mechanisms2.6Message Passing Neural Networks with edge attributes2.7Over-smoothing and effective network depth2.8Normalisation, dropout and regularisation on graphs2.9First models with PyTorch Geometric and DGL2.10Practical: node regression with a GCN on a synthetic network
3.1METR-LA and PEMS-BAY datasets3.2DCRNN: diffusion convolutional recurrent networks3.3STGCN and gated temporal convolutions3.4Graph WaveNet and adaptive adjacency matrices3.5ASTGCN and spatio-temporal attention3.6Multi-step forecasting and 15- to 60-minute horizons3.7Incidents, events and exogenous weather variables3.8Comparison with ARIMA, LSTM and historical models3.9Transferring models between cities3.10Lab: speed forecasting with DCRNN
4.1AC power flow and the DC approximation4.2GNNs as surrogate models for load flow4.3State estimation with PMU synchrophasors4.4N-1 contingency screening4.5Nodal and substation demand forecasting4.6Distributed generation and renewable variability4.7Physical constraints in the loss function4.8Scaling to medium-voltage distribution networks4.9Data generation with pandapower and MATPOWER4.10Case study: neural surrogate for load flow
5.1Hydraulic network modelling with EPANET5.2Pressure-based leak detection and localisation5.3Pressure estimation at unmonitored nodes5.4District heating and cooling networks5.5Pump scheduling optimisation with learned models5.6Calibration with sparse and noisy sensors5.7Uncertainty in demand patterns5.8Digital twin of the water supply network5.9Validation against reference hydraulic simulation5.10Practical: leak localisation on the L-Town benchmark
6.1Physics-informed GNNs: principles and formulation6.2Mass and energy conservation as penalty terms6.3Neural operators on meshes and graphs6.4MeshGraphNets for flow simulation6.5Hybridising physical models and data6.6Bayesian uncertainty quantification and ensembles6.7Explainability with GNNExplainer and gradients6.8Robustness to topology changes and reconfigurations6.9Generalisation to unseen topologies6.10Workshop: GNN with a nodal balance penalty
7.1Real-time inference and latency requirements7.2Partitioning and sampling of large-scale graphs7.3Distributed training on multiple GPUs7.4MLOps for spatio-temporal models7.5Data drift and retraining strategies7.6Integration with SCADA and control centres7.7Cybersecurity of models in critical infrastructure7.8Interoperability with the CIM model7.9Monitoring performance in production7.10Case study: real-time traffic forecasting service
8.1Definition of the problem and target network8.2Data acquisition, cleaning and alignment8.3Building the graph and nodal features8.4Selecting and justifying the architecture8.5Training and hyperparameter tuning8.6Validation against baseline models8.7Error analysis and interpretability8.8Operational deployment proposal8.9Report and presentation8.10Final project: final review with the tutor
How you'll learn
100% online training on the SEIUM campus, with downloadable material, recorded classes and an assigned technical tutor. There are no fixed timetables: you progress at your own pace within each intake, so it can be combined with work.
Each module combines theory, guided reading of the original standard and applied exercises. The working software is PyTorch Geometric, DGL (Deep Graph Library), pandapower, EPANET and SUMO, in free, academic or evaluation versions. Real procedures are studied: traffic speed forecasting from detector data, a load flow surrogate model on IEEE test networks and leak localisation in simulated water supply networks. Assessment is continuous by module plus the final integrative project.
Standards covered: ISO/IEC 23053:2022, IEC 61970 (CIM), IEEE C37.118.1, IEC 61968, ISO/IEC 42001:2023, Regulation (EU) 2024/1689 (AI Act), EN 50160.
Build it end to end
Your final project follows the real engineering workflow, step by step, reviewed by your tutor.
Where this takes you
Requirements
- Prior technical training or professional experience in the sector. No university degree is required.
- An internet connection and a computer capable of running the programme's software.
- English sufficient to read technical standards.
Ready to start?
Send your application today. Our admissions team replies within 72 hours with your study plan and funding options.
Frequently asked questions
The programme totals 800 hours over 6–8 months. You set your own pace within each intake, and your tutor helps you plan the workload around your job.
You work with PyTorch Geometric, DGL (Deep Graph Library), pandapower, EPANET and SUMO, using free, academic or evaluation versions. Your tutor indicates the free alternative wherever one exists.
A SEIUM Institute diploma, awarded on passing the continuous assessment and the final project. It is SEIUM's own qualification, not an official state degree; its recognition depends on the programme completed and the regulations of each country.
Yes. Modules passed are recognised in a higher programme in the same area (postgraduate diploma or master's). Ask about the recognition policy when you enrol.
Prior technical training or professional experience in the sector (no university degree required), a computer able to run the software and English sufficient to read technical standards.
Start dates are set by SEIUM's faculty for each intake. Our admissions team will confirm the next available date when you apply.
Related programmes
Diploma in Graph Analytics Pipeline Production
Online postgraduate diploma of 800 hours in production graph analytics pipelines, with 8 modules covering graph platform architecture, ingestion and graph construction, graph analytics algorithms, orchestration…
Diploma in Demand Models and Control Strategies
Online postgraduate diploma of 800 hours in demand models and control strategies: 8 modules covering mobility demand, discrete choice models, activity-based models, short-term demand forecasting, traffic…
Diploma in 1D/3D Coupling and Co-Sim Orchestration
Online postgraduate diploma of 800 hours in 1D/3D coupling and co-simulation orchestration: 8 modules covering co-simulation and multi-domain modelling, Modelica and 1D modelling, the FMI and…
More in Data, AI, Simulation & Analytics
All Data, AI, Simulation & Analytics programmes → · How to apply · Fees & financing · Scholarships