The production of renewable energy, based on solar panels and wind turbines, is intrinsically based on weather conditions. These must be forecasted. This makes it difficult to trade energy with long term contracts. As a result, renewable energy is almost exclusively traded on the spot markets, which become more important in most European countries.
Forecast errors result in imbalance positions. The large correlation in forecast errors between different parties, results often in large imbalance costs. This could on the long term lower investments in renewable energy sources. Our objective is to safeguard the profitability of renewable energy sources.
A first task in this project is to improve the trading strategies used. In particular, biddings on the sport market can be altered to increase revenues on other markets. These strategies do not interfere with the operations of wind turbines or solar panels and are called, for that reason, non-intrusive energy management methods.
A trend in contemporary operations of renewable resources is to steer production and consumption. A second task consists of improving trading strategies, taking into account these steering options.
These tasks are performed in close collaboration with industrial partners.
Tasks
- Forecasting energy prices.
- Development of non-intrusive energy management methods.
- Development of machine learning algorithms for intrusive control.
Expected Results
- Market-ready non-intrusive algorithms that can be rolled-out at a balance responsible party.
- Machine learning algorithm for intrusive control that can cope with an internal energy management system/ operator
Early Stage Researcher:
Supervisor:
Fjo De Ridder
Gowri Suryanarayana
Host institution:
