D4.1 Enhanced, user-friendly EMS for residential/building flexibility Discovery and delivery (v1)

This Deliverable presents the progress made so far regarding the Building Flexibility Management System (BFMS) and the Non-Intrusive Load Monitoring (NILM) components of the OPENTUNITY project. These two components concern the interaction between Home Energy Management Systems (HEMS), Building Energy Management Systems (BEMS), and various flexibility market stakeholders, leading to disaggregation of energy loads and forecasting of baseline and flexibility power consumption. The deliverable starts with a description of the HEMS/BEMS that are available at the four pilot sites of the project. These systems are necessary tools for the calculation and dispatch of flexibility, as they are responsible for the seamless communication of data from the pilots, as well as the dispatch of control signal based on flexibility forecast. The main functionalities of the available HEMS/BEMS are presented, focusing on hardware/software setup, performed data collection, supporting IoT devices and interfaces (if available). BFMS collects energy usage data, sensor information, and occupant preferences to forecast baseline consumption and available flexibility. By analyzing this data and using machine learning algorithms, BFMS determines how much energy can be adjusted and when, enabling buildings to participate in Demand Response (DR) programs. It then dispatches control signals to connected devices, ensuring energy is used efficiently while respecting user constraints. The DR Initialization Service is a novel module of the BFMS, used to collect occupant feedback on available assets and comfort preferences. It ensures a user-friendly experience, allowing consumers to participate in DR programs without compromising comfort or altering their behaviors. The NILM component provides meaningful energy related insights by analyzing total power consumption and disaggregating it into individual appliance usage using variables like active power, reactive power, current, and voltage. Unsupervised and semi-supervised methods like k-means clustering have been explored, offering a trade-off between reduced accuracy and the benefit of not requiring intrusive monitoring. In this way, real-time appliance power consumption can be predicted, creating generalized models that can be applied to any household with metering data. By generating accurate consumption curves for flexible assets, NILM supports demand response programs, enabling more effective load management and energy market participation. Together, BFMS and NILM create a powerful tool for the monitoring and optimization of energy consumption, enabling greater engagement in demand response initiatives and flexibility calculation and dispatch.