Incorporating predictive control into heating, ventilation and air conditioning (HVAC) systems has the potential to improve occupancy comfort and reduce energy use. This paper simulates the novel use of carbon dioxide (CO2) concentration inputs to augment temperature prediction and control. An artificial neural network (ANN) model and a least mean squares (LMS) filtering algorithm are used to simulate the temperature and control of a classroom in a high performance academic building with hydronic radiant heating and cooling panels. Numerical models are populated with variables that affect the heat energy entering, leaving, and being generated in a classroom. These variables include indoor and outdoor air temperature, radiant water and supply air temperatures, and classroom CO2 concentrations. The models are compared and then used to simulate the effect of a new control system that inputs CO2 measurements to account for the heat being generated by occupants of the controlled space. Simulation results suggest that augmenting HVAC control systems with CO2 measurements has the potential to improve temperature regulation by anticipating heating and cooling demand fluctuations in spaces with abrupt changes in occupancy.

This content is only available via PDF.
You do not currently have access to this content.