From solar generation to balancing: Data-driven forecasting for Swiss grid operations
Loading...
Author
Achi, D.
Tissier, C.
Carrillo, R.
Alet, P-J
DOI
Abstract
The rapid growth of photovoltaic (PV) generation in Switzerland has increased short-term variability in the Swiss control area, challenging grid operations and raising energy imbalance costs. To address this challenge, CSEM has developed two complementary forecasting solutions targeting imbalance mitigation at different system levels. First, a balancing-group-level PV generation forecasting model aims to reduce individual imbalance contributions, achieving a Normalised Root Mean Square Error (NRMSE) of 9.6% over a two-day horizon and 7.7% in the first two hours. Second, a short-term forecasting model for the Swiss control area imbalance to support intraday decision-making, achieving an area under curve (AUC) of 0.92, anticipating 33% of long-short transitions, and reducing next-step Root Mean Square Error (RMSE) by 25%, with robustness for forecast horizons of up to six hours.
Publication Reference
CSEM Scientific and Technical Report 2025, p. 85–86
Year
2025