Building simple mathematical models to calculate the energy requirements of buildings

Abstract

This work looks to build a predictive mathematical model that can provide a first approach to a building’s energy requirement (ER) value in a temperate continental climate. The aim is to contribute to the theoretical knowledge of energy assessment tools. To do this, parametric simulations were run and processed using the EnergyPlus 9.5 and JePlus programs. The results were then used as a dataset to build different mathematical models, using the SageMath program to run equations that predicted the ER of each scenario. Work was done with the models, scaling their complexity with the methods and the number of parameters used. Finally, a model with a low error (0.08) and 15 parameters was chosen. It was noted that, although increasing the number of parameters brought the models closer to a 0.02 error, there was a risk of overfitting. The chosen model seeks to incorporate dynamic simulations' accuracy and validity into a simple prediction tool that construction professionals can apply.