In 2021 the buildings sector accounted for 30% of global final energy consumption and 27% of total energy sector CO2 emissions [1]. To reach the clean energy transition towards the Net Zero Scenario by 2050, the carbon footprint of buildings needs to be more than halved by 2030. Thus, against the 2019 European Green Deal, the European Union (EU) has presented the “Renovation Wave” aimed to enhance the ecological transition of 35 million building units by 2030. 

In this framework, neighborhood and community-scale approaches are pivotal pillars of the Action Plan 3.2 of the European Strategic Energy Technology Plan (SET-Plan). This recognizes the district level as the ideal field of application for promoting a clean, circular and sustainable future economy. 

To perform this transition, it is necessary to exploit different renewable energy sources (RESs) in existing districts in conjunction with improving energy efficiency and refurbishment actions. In particular, the integration of RESs in PEDs poses critical challenges to energy balances facing the intermittent nature of non-programmable sources. It is possible to deal with these issues by the introduction of thermal and electric storage and through the connection of different energy sectors, such as the power and mobility sectors by means of vehicle-to-grid/to-buildings, and the power and heating sectors through power-to-heat/to-cold technologies. 

These are the main challenges to reach the flexibility target in PEDs on the supply side. The matching of supply and demand relies on the proper assessment and forecasting of load profiles and on the adoption of strategies for peak shaving and shifting. Information about current and future buildings operating conditions can be obtained by applying data-driven procedures  on the massive amount of data collected using Advanced Metering Infrastructure, which is more and more available in our country. 

Moreover, great attention needs to be paid to archetypes since Urban Energy Modelling approach requires a significant computational time which adversely affects the models developed. By adopting pattern recognition techniques, the characteristic load profiles of homogeneous clusters of buildings can be obtained. The latter can be applied in combination with data-driven approaches.