Selecting control input type for a building predictive controller integrated in a power grid

Abstract

Application of Model Predictive Control (MPC) for a Smart Building in a Smart Grid environment is studied by using an experimentally validated building thermal model. The goal is to compare three control input types for an office building's Heating, Ventilation, and Air Conditioning (HVAC) system, and select the control type with the best performance. Three types of MPC control input are considered for the HVAC system: (1) solitary control over the supply air temperature, (2) solitary control over the air mass flow rate, and (3) combined control over the supply air temperature and mass flow rate. An objective function is defined based on an introduced Normalized Performance Index (NP-Index) which balances price minimization while maintaining a balanced steady load profile in the grid which benefits customers and the distribution grid utility. The results show that using the combined control approach leads to 20% improvement on NP-Index compared to the solitary mass flow rate control. Additionally, controlling both the supply air temperature and air mass flow rate reduces power consumption by 4% and 13% compared to solitary air temperature control and solitary air mass flow rate control, respectively.