Model predictive control with feedforward structure for chilled water temperature in HVAC system

Being an optimizing technology, model predictive control (MPC) can now be found in a wide variety of application fields. The main and most obvious control goal to be achieved in a heating, ventilation and air conditioning (HVAC) system is to improve the control accuracy and reduce energy consumption in recent years on the premise of comfort. In this article, modeling variables including chilled water temperature, weather conditions, occupancy and electricity are divided into two categories: manipulated variables and random variables. For modeling the system, step response method is applied under the former variables, and power spectral density (PSD) method is used under the latter variables in this study. After modeling, a MPC strategy with feedforward control structure which utilizes the known dynamic characteristics of random variables and effectively compensate the errors caused by these variables. A chiller plant in a medium-sized commercial building is utilized to evaluate the strategy. The result shows that the proposed MPC with feedforward control structure achieves a considerable amount of energy saving.