User-centred AI-based energy services built on personal preference models

Buildings, neighbourhoods and urban districts must become a more flexible and intelligent component in the renewable energy system of the future. Conventional HVAC systems in buildings are aimed at conditioning room temperature setpoints or similar. Incorrectly set setpoint values, overriding by controllers, etc. lead to inefficient operation and do not take dynamic factors (e.g. user comfort) into account. Current research results emphasise that people must be placed at the centre of energy services if the energy transition is to succeed. The perceived thermal comfort is individual (age, gender, activity, clothing, etc.). Despite existing standards (e.g. ISO 7730:2005, ASHRAE 55-2020), these do not do justice to realistic individual needs, as the basic assumptions were determined under laboratory conditions and are only aimed at the "average thermal sensation" of a group (e.g. PMV Predicted Mean Vote). Energy efficiency measures that are successful in the long term must put people at the centre. A paradigm shift towards personal comfort models (PCMs) is needed in order to realise energy-efficient building performance while at the same time meeting user requirements. The use of AI in combination with IoT technologies will be the backbone and enabler of innovative energy services.