Search for a command to run...
Deliverable D2.2 presents the development and proof of concept of four complementary methods designed to support active consumers and energy communities (ECs) in defining and optimising their operational strategies. These methods, Machine-Learning based Forecasting, Data Analytics, Decision Support, and Distributed Energy Resource (DER) Valorisation, form a coherent framework enabling data-driven participation in local energy sharing and peer-to-peer (P2P), market interaction, and flexibility services. The deliverable, developed under Task 2.2 of the U2Demo project, aims to demonstrate the technical feasibility and complementarity of the four methods, validating their capability to support informed, optimised, and customised decision-making processes for consumers and ECs. The Machine-Learning based Forecasting methods, developed by EIFER, provide predictive tools for electricity demand, renewable generation (PV and wind), electric vehicle charging needs, and market price signals. By incorporating exogenous variables such as weather data, time features, and historical patterns, these models deliver accurate short- and medium-term forecasts tailored to different levels of granularity (individual, EC, or regional). The forecasting results constitute a key input for the remaining methods, as they enable proactive planning and flexibility assessment. The Data Analytics methods, developed by KU Leuven, evaluate consumer and community behaviour across technical, economic, social, and environmental dimensions. They quantify how energy strategies influence self-consumption, self-sufficiency, cost savings, fairness, and carbon reduction. The analytics framework combines descriptive, predictive, and prescriptive approaches to create a unified evaluation structure. In this deliverable, the methods have been applied using reference datasets and synthetic energy community profiles to establish baselines and assess the expected impact of different operational strategies. Importantly, Data Analytics also provide validation feedback for forecasting models and serve as a foundation for defining meaningful KPIs used in subsequent optimisation tasks. The Decision Support methods, developed by INESC ID, offer a structured optimisation-based framework for energy community management. Implemented in the PyECOM platform, these methods evaluate operational strategies under multiple objectives, such as reducing operational cost, electricity cost, or grid import, while improving comfort, battery longevity, and environmental impact. Multi-objective optimisation techniques are employed to assess trade-offs between conflicting objectives and identify balanced operational strategies. Results demonstrate the capacity of the decision support system to align community-wide resource coordination with diverse member preferences, providing actionable recommendations to EC managers and supporting participation in P2P trading and flexibility markets. The DER Valorisation methods, developed by R&D Nester, address peer-level optimisation and the quantification of flexibility potential from individual distributed resources such as PV systems, stationary batteries, Electrical Vehicles (EVs), and Heating, Ventilation and Air Conditioning (HVAC) loads. The approach is formulated as a multi-period optimisation problem that determines optimal DER scheduling while respecting technical constraints and comfort conditions. The tool evaluates flexibility margins and economic value of DER operation, aiming to improve self-consumption, minimise curtailment, and create added value through potential participation in market or P2P schemes. The results demonstrate that coordinated operation can significantly reduce grid dependency, increase the use of renewable energy, and add economic return by DER participation. Although each method has been implemented and validated independently, their design follows an integrated methodological framework that ensures data and functional interoperability. Forecasting provides the predictive signals required by the other methods, Data Analytics validates model behaviour and defines performance indicators, and both Decision Support and DER Valorisation use these inputs to generate optimised operational strategies. This framework establishes the logical flow of information and retroactive feedback loops necessary for future integration within the U2Demo platform. eliverable D2.2 thus represents the proof of concept of the four methods, serving as a foundation for subsequent project activities. In Task 4.2, the focus will shift towards increasing the Technology Readiness Level (TRL) of these methods through their implementation and integration in the U2Demo operational platform. The outputs defined in D2.2 will also serve as inputs to Task 2.3, where energy sharing and P2P models are being developed, and to Task 2.4, which addresses market and flexibility participation mechanisms, including system services for Transmission and Distribution System Operators (TSOs and DSOs). In summary, this deliverable establishes a solid methodological basis for empowering energy communities and active consumers through data-driven decision-making. The four methods complement each other by linking forecasting accuracy, behavioural insight, optimisation capability, and flexibility valorisation. Together, they form an integrated set of tools that will support the transition towards interoperable, transparent, and participatory energy systems under the U2Demo framework.