Renewable Energy Forecasting
We provide consulting and build custom software for PV forecasting, integrating the software with our customers hardware and software. Recent commercial project for a major energy supplier involved building and hyperpamaters fine tuning of a custom CNN-LSTM hybrid AI model to forecast power generation. Its input includes power consumption, on site measurements (incl. temperature, cameras feeds) and historical satellite images. Its outputs are 15-mins-ahead predictions on generated PV power and they have proven accurate by 99%. Our code is operational at the energy supplier large PV site.