Ancillary Services

In our production cost model, ancillary services are inherently co-optimized with energy prices, allowing generation and storage units to switch seamlessly between energy and ancillary markets in response to prevailing signals. While we use a reserve requirement price that reflects overall ancillary costs, the model can’t produce separate time series for each service.

To fill that gap, we run a secondary model that projects individual ancillary‐service prices into the future. It takes the production model’s energy‐price outputs as features in a gradient boosting model and applies the historical relationships between energy and ancillary markets. It also uses time-varying weights to capture how growing battery supply—by saturating ancillary-services volumes—puts downward pressure on those service prices.