The integration of RES is vital for agro communities seeking energy autonomy and resilience against rising operational costs. To make decentralised RES profitable, farms need precise tools to predict energy output, enabling smarter resource planning. HarvRESt tackled this by developing advanced forecasting solutions across three horizons, namely:

• Short-term forecasts allow farmers to schedule high energy-intensive operations, like pumping or cooling, to directly align with peak solar production, maximising self-consumption.
  
  
• Mid-term forecasts provide a multi-day outlook, offering strategic visibility for planning major agricultural operations, maintenance, and energy market interactions.
  
  
• Finally, long-term annual energy forecasts are essential for the foundational strategic planning and correct sizing of new solar, wind, or biogas investments.

Our work focused on designing and training sophisticated AI analytics pipelines to deliver this required accuracy. We systematically identified several analytics problems and outlined their business relevance and technical configuration, covering short- and mid-term applications focused on PV energy, and long-term applications covering PV, Wind, and Biogas generation.

Our solutions leverage a blended approach: we employed ML/DL algorithms, using historical and experimental data that includes weather patterns and solar irradiance, for highly responsive short- and mid-term forecasting. For complex long-term energy generation planning, we defined a dedicated simulation-based approach based on the creation of the Typical Meteorological Year (TMY). All pipelines included performance evaluation to ensure maximum accuracy across the project's demonstration sites.

The culmination of this effort is the Generation Analytics Catalogue, a practical project output. This living resource is designed for continuous improvement; it shall be enriched with additional solutions as needed, including the upcoming development of robust short-term wind energy forecasting analytics. This ensures the HarvRESt system remains a comprehensive and evolving tool for supporting energy planning in the agri-food sector.