Thermal Building Models for Energy Management Systems

This paper discusses the benefits of thermal building models for energy management schemes to improve building energy efficiency and reduce energy costs. The authors describe a gray box approach for modeling buildings using simplified physical models tuned with empirical data. The paper also explores the benefits of online parameter estimation for building energy management, including the ability to detect and correct modeling errors in real-time, improve model reliability, and reduce the need for manual calibration and maintenance. The authors compare the convergence and prediction performance of the Joint Extended Kalman Filter (JEKF) for different building models. They conclude that the linear building models in conjunction with the JEKF provide a good prediction accuracy with relatively low complexity. Furthermore, the model parameters converge reliable to the building parameters for the investigated cases.