According to the Global Market Outlook for Solar Power 2023 – 2027 report published by SolarPower Europe, 239 GW of new solar capacity was added worldwide, reaching a total capacity of 1.2 TW and experiencing a 25% increase compared to 2021.

ACCIONA Energía is present throughout the entire value chain of the energy sector, including project development and structuring, engineering, construction, supply, operation, maintenance, asset management, and the management and sale of clean energy. In the photovoltaic sector, ACCIONA Energía has accumulated 21 years of experience. The company builds, operates, and maintains photovoltaic parks in 9 countries worldwide with an installed capacity of 2081 MWp, including more than 20 large plants within its portfolio.

One of the greatest challenges in photovoltaic generation is its intermittency and dependence on atmospheric conditions regarding cloud presence and thickness for electricity generation. Increasingly, accurate short-term forecasts of intraday irradiance play a crucial role in optimizing market operations (primarily for participation in adjustment markets) and enhancing the efficiency of utility-scale plants. Improved predictions of energy production will facilitate the greater integration of photovoltaic generation into the grid, considering its variability and progressive hybridization with other technologies such as storage.

Short-term predictions (NowCasting) are techniques for forecasting generation in the near future, typically within the range of minutes or a few hours from the present moment. To feed mathematical models and algorithms, real-time data from various sources such as satellite images, meteorological stations, sky imaging systems, ground irradiance sensors, and cloud cover information are used to make highly accurate and localized predictions of solar irradiance.

From ACCIONA Energía, we are seeking solutions that provide very short-term production data (15 minutes) for utility-scale photovoltaic plants. These solutions should be based on prediction models using satellite image processing and other technologies capable of anticipating sudden decreases or increases in resource availability caused by clouds not detected by conventional prediction models.

Key specifications for the solution include:

• Temporal resolution of the prediction: Minute-level
• Prediction update frequency: Every 15 minutes
• Prediction horizon time: 180 minutes

The pilot project will be developed in the Extremadura I PV plant (Spain).