Blog

Based on 12 years of academic research, Calibsun has developed an machine learning algorithm capable of analyzing weather data in real time to provide optimal forecasts. In order to test the performance of its forecasting models, Calibsun has performed backtests. This evaluation method consists of simulating the forecasts that would have been generated by the model over a past historical period.

Emissions reduction targets for the energy industry, consumers and industry alike, are accelerating the development of renewable energy sources (RE). Photovoltaics (PV) has established itself as a cutting-edge, cost-effective and sustainable technology. The growth in solar energy production has highlighted the importance of accurate forecasting to estimate the power of installations over several time horizons.

To increase their energy sovereignety while accelerating their energy transition, many industries are integrating solar energy into their energy mix through self-consumption.

Intermittent, variable and subject to the uncertain weather, solar energy remains a highly promising solution for meeting decarbonization targets and reducing greenhouse gas emissions.

Unlike deterministic methods, which produce a single result, probabilistic approaches provide a range of possible outcomes with associated probabilities. These methods take into account the uncertainty inherent in the various factors influencing forecasts, such as data variability.

Solar resource forecasting has become an essential aspect of the operation of photovoltaic (PV) power plants to increase the production of renewable energy from the sun. Solar resource forecasting is essentially a data story, where vast quantities of data are collected, analyzed and transformed into actionable information by grid operators, power plant operators, energy trading aggregators and developers of future installations.

According to the International Energy Agency (IEA), photovoltaics should account for 2,350 GW of power worldwide within 4 years, surpassing hydraulic in 2024, natural gas in 2026 and coal in 2027 in electricity generation.