Identification and Control for Optimizing Building Energy Demand

This talk focuses on developing a dynamic, real-time control and optimization strategy to shift electricity demand in commercial buildings to reduce peak time consumption. Unmeasured real time disturbances are predicted and incorporated in the optimal strategy using the framework of model predictive control (MPC). System identification with intermittent HVAC control is a challenging and critical issue for success. We propose a method to deal with this unique challenge from input signal excitation to setpoint level design. The day-night variations in ambient conditions are considered as well.