Framework for the utilization of forward and inverse airflow models in systematic sensor system design for indoor air

The building industry faces many challenges, including maximizing energy efficiency and improving the health and safety of building occupants. Indoor air sensor systems can be used to regulate ventilation rates, monitor indoor air quality, and detect and eliminate harmful contaminants. However, current sensor system design is intuitively-based. This research presents the framework for utilizing airflow modeling techniques in systematic indoor air sensor system design.;Airflow modeling is used to simulate the contaminant data necessary for systematic sensor system design. Forward airflow models require detailed information about an indoor space in order to simulate airflow and contaminant transport. If this information is not available, inverse airflow models can estimate airflow patterns using measurements from commonly installed sensor systems. Thus, this research was divided into three parts: develop a framework for utilizing (1) forward airflow models in systematic sensor system design; (2) inverse airflow models for estimating airflow patterns in a single zone; and (3) inverse airflow models for estimating airflow patterns in a whole building.;In Part 1, it was found that data from simple airflow models could be used to design sensor systems that performed just as well as those designed using more complex airflow models for sensor systems with more than one sensor. Thus, saving modeling time without compromising sensor system performance. In Part 2, it was found that velocity sensors placed on the wall closest to the exhaust in a single zone most improved the airflow estimation accuracy of the developed inverse model. Thus, offering practicality in experimental setup without sacrificing estimation accuracy. In Part 3, it was found that the proposed building airflow network inverse model could be applied to a building of any size if the rank of the known-information matrix was greater than or equal to the number of unknown airflow rates. The estimated building airflow network was in good agreement with a synthetic building airflow network, and its contaminant concentration prediction ability comparable to similar studies published in the literature. The proposed building airflow network inverse model also offered computational advantages over methods published in the literature.