PV-Sp - Improving PV yield prediction for storage control - a pilot project with rooftop systems

In Germany, photovoltaic (PV) systems are increasingly being combined with battery storage systems to enhance self-consumption and relieve the electricity grids of weather-related power peaks. Predictions of solar radiation and expected PV power generation are prerequisites for an optimal battery charging strategy. In the PV-Sp project, a feasibility study is conducted exploring whether a radiation map can be created based on the PV yields of PV battery systems, leading to an improvement of PV power predictions.

An algorithm to derive the solar radiation from measured PV data was developed within the research project "MetPVNet". In the present project, the algorithm is adapted to a test application in five PV plants of the company Solarwatt Innovation GmbH. The plants are equipped with radiation sensors in order to compare the derived  radiation values with on site radiation measurements. The aim of the test is to study the reliability of the derived radiation data. If the test is successful, the software developed in the project can be applied to a larger dataset in the future, with the goal of characterizing the radiation over a larger area.