Neural Network Based Assessment of Delivered Energy for Heating

University North is the sole higher education institution in Croatia that took part in the Energy Renovation Program for Public Sector Buildings between 2014 and 2020. After a thorough renovation and the installation of measurement equipment in the buildings, a study was launched to explore how user behavior affects energy consumption. This was done by monitoring meteorological conditions and indoor comfort parameters such as temperature, relative humidity (RH), carbon dioxide (CO2) levels, and volatile organic compounds (VOC). The objective was to enhance the existing energy performance contracting model (ESCO model). Data collected throughout the 2022/2023 heating season (from October to May) was used as input for training and evaluating a feedforward artificial neural network (ANN). The ANN was designed to estimate the reference energy consumption for heating and to forecast energy use under varying future meteorological and occupancy scenarios.