Positive energy districts (PEDs) are seen as a promising pathway to facilitating energy transition. PEDs are urban areas composed of different buildings and public spaces with local energy production, where the total annual energy balance must be positive. Urban areas consist of a mix of different buildings, such as households and service sector consumers (offices, restaurants, shops, cafes, supermarkets), which have a different annual energy demand and production, as well as a different consumption profile. This paper presents a data modeling approach to estimating the annual energy balance of different types of consumer categories in urban areas and proposes a methodology to extrapolate energy demands from specific building types to the aggregated level of an urban area and vice versa. By dividing an urban area into clusters of different consumer categories, depending on parameters such as surface area, building type and energy interventions, energy demands are estimated. The presented modeling approach is used to model and calculate the energy balance and CO2 emissions in two PED areas of the City of Groningen (The Netherlands) proposed in the Smart City H2020 MAKING CITY project.

High penetration of renewable energy sources brings both opportunities and challenges for Smart Grid operation. Due to their high contribution to energy consumption, aggregated load flexibility of small residential and service sector consumers has a potential to address the intermittency challenge of distributed generation. Predicting aggregated load flexibility of this consumer sector involves access to sensitive smart meter data, raising data collection and sharing concerns. Federated Learning, a decentralized machine learning technique that uses data distributed on user devices to construct an aggregated, global model, offers potential solutions to tackling this challenge. This paper explores the potential of using Federated Learning for flexibility prediction in Smart Grids through an analysis of its opportunities and implications for different stakeholders involved, as well as the challenges faced. The analysis shows that Federated Learning is a promising approach for building privacy-preserving energy portfolios of aggregated demand data.

ABSTRACT Resilience of power systems is highly impacted by factors such as increasing severity and frequency of weather events, but also smart grid advances that introduce major operational changes in power systems. Rapidly adapting to these changing circumstances and harnessing the potential of technological advances is the key to ensuring that power systems stay operational during disturbances, thereby improving resilience. This paper addresses this challenge by presenting an approach for improving resilience through local energy resource sharing across multiple distribution systems. The approach brings together the physical and the ICT layer of power systems through a self-organization approach that automatically alters the physical grid topology and forms local energy groups in order to mitigate the effects of widespread outages. Thereby, supply and demand are locally matched, and demand met is maximized during an outage. The results demonstrate that using the proposed approach, operational resilience of impacted distribution systems is improved.

Centralized management of power systems is becoming more challenging due to the increased introduction of distributed renewable energy resources, along with demand increase and aging infrastructures. To address these challenges, this paper proposes new mechanisms for decentralized energy management. Based on self-organization of consumers, prosumers and producers into virtual groups, called clusters, supply and demand of electricity is locally matched. Distributed multi-agent systems are used as a way to represent virtual cluster members. The mechanisms are illustrated, and static and dynamic virtual clusters are compared. Dynamic reconfiguration is achieved by varying the time periods for which clustering is performed. The proposed clustering mechanisms demonstrate that large-scale centralized energy systems can operate in a decentralized fashion when only local information is available.