Stoll, M.M.Stoll2022-03-082022-03-081991https://publica.fraunhofer.de/handle/publica/318890Neural networks are an emerging computional technology which mimic the computional abilities of biological systems by using a large number of simple, interconnected artificial neurons. Because of their specific processing advantages, neural networks are the technology of choice for an approach to the energy management problem. The progress in performance that neural networks could offer, is demonstrated on the important subtask "state-of-charge (SOC) estimate" of a lead-acid battery in a renewable enrgy supply system (RES). The results of theoretical investigations encourage practical testing and further application of different kinds of neural networks to other subtasks of the energy management problem.enEnergiemanagementenergy managementInselsystemLadezustandschätzungneural networkneuronales Netzwerkstand-alone systemstate-of-charge estimate621697conference paper