Reduced-order modeling for control of indoor building airflows
Sunil Ahuja, Eugene Cliff, Satish Narayanan
We present computationally tractable tools for mod- eling indoor airflow that are amenable to control de- sign. The model reduction technique used, called bal- anced truncation, requires data from CFD simulations but results in models that are of much smaller dimension, typ- ically O(100). The method also guarantees accurate and stable models. For indoor building airflows and thermal environment, the low-order models are used to address the control prob- lem of disturbance rejection. The disturbances include heat gains from solar radiation, occupants or lights, and changes in ambient temperature. The modeling tech- niques are expected to be relevant primarily to systems equipped with low-energy consumption HVAC terminal units, such as displacement ventilation (DV) or radiant cooling/heating, where the stratified indoor environment and airflows are unsteady, spatially inhomogeneous and whose control is highly sensitive to disturbances. Finally, the proposed tools are demonstrated using two building spaces of different spatial scales, both modeled with a DV system; one, a small conference room, and the sec- ond having a large floor-to-ceiling height. In both the cases, the model-based controller is implemented in the full-order system, and is shown to reject disturbances to ensure robust indoor comfort control.
- There are currently no refbacks.