NeuralPower – A holistic approach to the allocation and forecasting of energy demands as well as the control of consumption using artificial intelligence

The energy system is characterized by high dynamics and complexity, due to many factors such as the growing share of renewable energies and the increasing number of electric cars. The increasing spread of IoT hardware and intelligent metering systems enables a networked control of the energy system that goes beyond classic virtual power plant concepts and also includes consumers.

The project by Neural Power aims to increase automation in the creation of new virtual power plants. By using AI systems that analyze large amounts of data, heterogeneous micro-capacities can be economically utilized, contributing to the stabilization of supply and demand and supporting the energy transition. Building on the concept of virtual power plants, in which decentralized producing or storing actors participate, the project aims to ensure that consuming actors - who only reach a relevant size of marketable, flexible electricity load through bundling - can also be intelligently tapped, forecasted, and controlled. In this context, three pilot scenarios (servers and cooling systems in data centers, cooling furniture in consumer markets, and bundled micro-consumers in private households) are being explored and their potentials determined.

adelphi's role in this endeavor was to develop a catalog of political and regulatory action options to promote digital business models for the flexibilization of electricity demand, with recommendations in the area of electricity and regulation market design, as well as to identify factors that promote and hinder acceptance in the design of business models. Additionally, the regulatory and acceptance-related barriers that have so far prevented market diffusion were analyzed. Furthermore, an investigation, mapping, and summary of the status quo in the area of already existing business models and the respective framework conditions were carried out, including and in feedback with the investigated pilot scenarios.