Simulation and experimental demonstration of model predictive control in a building HVAC system

Abstract

This article presents the framework and results of implementing optimization-based control algorithm for building HVAC systems and demonstrates its benefits through reduced building energy consumption as well as improved thermal comfort along with lessons learned. In particular, a practically effective and computationally efficient model predictive control algorithm is proposed to optimize building energy usage while maintaining thermal comfort in a multi-zone medium-sized commercial building. This article has two themes. Driven by the challenge of fully evaluating the benefit of the proposed model predictive controller against baseline control, a model predictive control design framework is first presented with its performance benchmarked based on a high-fidelity building HVAC simulation environment to verify its effectiveness and feasibility. Following the same model predictive control design framework, the experimental results from the same building located at the Philadelphia Navy Yard are then presented. For the simulation study, the performance of the model predictive control algorithm was estimated relative to baseline days with exactly the same internal loads and outdoor conditions, and it was estimated that model predictive control reduced the total electrical energy consumption by around 17.5%. For the subsequent experimental demonstration, the performance of the model predictive control algorithm was estimated relative to baseline days with similar outdoor air temperature patterns during the cooling and shoulder seasons, and it was concluded that model predictive control reduced the total electrical energy consumption by more than 20% on average while improving thermal comfort in terms of zone air temperature.