A machine learning based tool for heating needs forecasting in Campus Leonardo buildings

According to studies done by the IEA (International Energy Agency), buildings constitute to 40 percent of the total primary energy consumption around the world. The fact that 90 percent of our time (human beings) are spent indoors (schools, university, offices), it becomes imperative for us to look at the energy utilization of buildings in terms of efficiency and production. Hence, using the European energy signature as a benchmark and Politecnico Di Milano, Leonardo campus as a case study, we will be looking into the heating energy needs of different building applications. This study involves introducing an additional variable for analysis i.e solar irradiance (Wh/m2) to the existing dry bulb temperature (°C) in the European energy signature. This improvement would help us understand how energy needs of buildings change with respect to Irradiance.With this in mind we have predicted future heating energy needs of the buildings taking different boundary conditions in mind. Here, it was realized that using this additional variable we have an obtained a co-efficient of determination of 0,95 which gives us the good insight of the model we have developed. This method of analysis being adaptable, makes it a good tool for future analysis and forecasting of heating energy needs in buildings.