A comparison between temperature modeling strategies in smart buildings

This paper deals with modeling the temperature of a room in a smart building. Two different approaches are analyzed and compared. The first one is based on a black-box identification procedure, which yields a fairly simple dynamical system that is suitable for real-time control. The second approach is based on first principles, it requires a complex and time consuming calibration procedure but it is capable of describing accurately the physical behavior of the system. The trade-off between the accuracy of the model and its computational complexity is evaluated by using experimental data collected from a smart pilot building located in Almería, Spain.