New method reduces wind farm power imbalance costs

A transparent control policy can lower imbalance costs by 82% using real data from the Belgian grid.

A new control policy for wind farms can reduce imbalance costs by 82%. This method uses real data from an offshore wind farm and the Belgian grid.

The system uses policy distillation with explainable boosting machines (EBMs). EBMs are a variant of generalized additive models used to obtain a transparent control policy. The researchers also include expert knowledge during model design to make the strategy justifiable to stakeholders.

The framework uses a value-oriented perspective inspired by energy forecasting. This approach helps increase the value captured by the EBM model.

Wind farm power control helps the system maintain balance and avoid blackouts. In Europe, implicit balancing allows a wind farm to adjust its power setpoint using information from the Transmission System Operator.

Implicit balancing is a challenging problem. The settlement price is unknown in advance and can be impacted by the control strategy used. Reinforcement learning methods could help but lack the interpretability needed for real wind farms.

Key facts

  • The controller can reduce imbalance costs by 82%.
  • The method outperforms a price threshold controller and a decision tree baseline.
  • The framework uses explainable boosting machines (EBMs) to create a transparent policy.

All sections

Edition