Learnings

Outcomes

The Project delivered and validated a suite of dynamic regional reserve forecasting models covering twelve zones and four operational horizons. Validation testing confirmed that the models perform at a level consistent with operational risk requirements and demonstrate improved reserve efficiency relative to static approaches. 

The Project also delivered enhanced analytical and visualisation capability, including explainability metrics and comparison against historic static methodologies. These outputs provide a robust evidence base to inform future decisions on operational trials and potential integration with locational procurement arrangements, subject to further governance and assurance.

Lessons Learnt

The Project demonstrated that both national‑level system features and region‑specific variables are material drivers of regional reserve requirements. National‑level features were shown to capture broader system dynamics that influence local conditions, while region‑specific predictors such as demand forecasts and errors, embedded wind forecasts, and air temperature provided complementary explanatory power. These findings are consistent with related regional modelling activities, including constraint forecasting, and highlight the importance of forecast accuracy and data consistency in enabling effective reserve setting.

Differences between national and aggregated regional demand outturns were identified due to differing data sources and metering points. Appropriate mapping and adjustments were applied to ensure consistency.