A Thermal Control Methodology Based on a Multiple Linear Regression Predictive Model for Indoor Heating

As part of the energy transition, and in order to take advantage of the data generated by buildings, this study aims to develop indoor temperature control to improve the thermal comfort and energy efficiency of buildings. A temperature control methodology based on a multiple regression model has been developed to control heating systems in a predictive and anticipatory manner. Based on the data measured in the building and the weather data, an iterative statistical approach was set up, allowing to obtain a high-precision indoor temperature forecast model. Using this model, the brute force optimization method is used to find a new heating strategy to improve the thermal comfort and energy efficiency of the building. Applying the developed methodology, using data generated from a simulated building, a multiple linear regression model is set up forecasting the indoor temperature with a mean absolute percentage error less than 2%. The analysis of the results revealed that the heating control based on the developed methodology improved the thermal comfort of the building compared to conventional control.