Advanced Machine Learning to support dispatching
Keeping the balance in the grid is essential: the injection of power must always be equal to the consumption in order to keep the frequency at 50Hz. Due to renewable energy integration, the system is getting more and more complex and it is becoming essential to be able to forecast imbalances at very short term.
We are creating a model to detect the correlation between the different parameters influencing the System Imbalance and forecasting the System Imbalance over the next 15 min to hour.
Every TSO in Europe (coordinated under the umbrella organisation ENTSO-E) is responsible for maintaining the balance between injection and offtake (including import and export) in their control area. Discrepancies between injection and offtake create imbalances that have to be resolved by the system operator. Elia has established a balancing mechanism that uses power reserves, i.e. back-up capacity supplied by grid users, to correct such imbalances. In addition, market parties are given an ‘imbalance tariff’ as an incentive to balance their own portfolio or help to correct system imbalances.
Even after the activation of reserves, an imbalance may still remain: the area control error (ACE), which is the difference between the system imbalance and the net volume of activated reserves or net regulation volume (NRV).
The aim of this project is to create a model based on the analysis of historical data to:
- detect correlations between the different parameters influencing system imbalances;
- forecast system imbalances over the next 15 minutes to an hour.
The objective is to demonstrate that the model can help to formalise decision-making rules and facilitate post-decision evaluation by providing better descriptions of actual situations, visibly displaying correlations between different parameters to raise awareness about the causes of imbalances.
Supporting decision-making with AI
Whenever a large system imbalance arises, the system engineer needs to take steps to maintain the frequency at 50 Hz and ensure the grid’s safe operation.
Of course any action taken by Elia to maintain the balance between injection and offtake will impact on costs. For instance, any improvement leading to more efficient balancing will cut costs for consumers, as well as enabling safer grid operation.
Elia hopes to first test and then demonstrate the usefulness of artificial intelligence in heightening control centre engineers’ awareness of the situations they face and supporting their decision-making.
- a working prototype, including a model to forecast system imbalances and an interface allowing users to understand the correlation between parameters;
- a list of requirements for integrating prototypes into the business process as an off-line tool facilitating the decision-making by system operators;
- a first analysis of the model’s potential as a supportive decision-making tool for system operators.