A Hybrid Physics-Based and Data Driven Approach to Optimal Control of Building Cooling/Heating Systems

This work integrates a physics-based model with a data driven time-series model to forecast and optimally manage building energy. Physical characterization of the building is partially captured by a collection of zonal energy balance equations with parameters estimated using a least squares estimation (LSE) technique and data initially generated from the EnergyPlus building model. A generalized Cochran-Orcutt estimation technique is adopted to describe the data generated from these simulations. The combined forecast model is then used in a model predictive control (MPC) framework to manage heating and cooling set points. This work is motivated by the practical limitations of simulation-based optimizations. Once the forecast model is established capturing sufficient statistical variability and physical behavior of the building, there will be no more need to run EnergyPlus in the optimization routine. The proposed methodology lends itself for real-life implementation of building energy management systems where predictive control is desired to reduce energy use and avoid demand charges and occupant discomfort. At each time step, it determines the optimal set point values of all building's zones and updates these values over time. In practice, the proposed control strategy can be implemented in commercial smart energy boxes to optimally control total daily energy-use costs.