Modelling Room Temperature Dynamics for Efficient HVAC System Energy Consumption

This work performs an study of contemporary methodologies for the modelling and forecasting of heating, ventilation, and air conditioning (HVAC) systems and introduces a novel feature selection approach. This novel feature selection approach is grounded in physical equations and expert knowledge. The use of HVAC system dynamics, facilitates the creation of pivotal variables, essential for enhancing model performance. The methodology employed to construct tendency variables also enables the computation of energy consumption for individual HVAC and accurate temperature predictions for the acclimatized areas and inertia deposits. Finally, the methodology is validated by test, using real data from a real scenario. Results of the case study demonstrate that using the purposed methodology can achieve accurate results. The complete code is available at https://github.com/ferran137/SOCO-HVAC.git .