Applying Model Predictive Control to a LEED Silver-certified Building

AbstractThis paper models a LEED Silver institutional building and applies a model predictive control (MPC) strategy. The building is modelled using EnergyPlus, which is then compared to measured consumption data for model calibration. The calibrated model is then exercised with various control strategies to generate data used to create and train a statistical black box model. The black box model is then used as the base model for the model predictive controller, while EnergyPlus is used as a virtual building to determine savings. The resulting MPC strategy results are compared to the existing rule-based control applied to the building to highlight the energy savings. Initial results indicate a 10% HVAC energy reduction based 2013 climatic data for the site. While the energy reduction is small, it is achieved on a new LEED certified building and requires only software changes to the equipment