Context
As renewable-energy penetration increases, reliable wind-power forecasting becomes increasingly important for energy planning, storage optimisation and demand-side flexibility.
Traditional point forecasts provide limited information about uncertainty and risk.
Discovery & framing
- Decision-making under uncertainty
- Risk-aware planning
- Digital energy
- Platform integration considerations
- Decision support rather than model accuracy alone
Data & method
- Real wind and weather measurements from Sweden (SE1, SE3)
- Station selection near wind parks
- u/v wind-vector features
- Cyclical time features
- End-to-end ML pipeline with preprocessing and evaluation
- Temporal Fusion Transformer (TFT) with quantile regression
Product / platform relevance
By combining forecasts with uncertainty ranges rather than a single predicted value, planners can make decisions with a clearer understanding of risk.
- Demand-side flexibility
- EV charging
- Smart-building load optimisation
- Industrial load shifting
- Storage optimisation
- Planning and load balancing
- Risk-aware decisions
Images
No illustrative or simulated result curves are used — only real material from the thesis work.
Image to be added
Reflection
Uncertainty becomes useful decision-support information rather than simply prediction error.