Characterization of a Constant Air Volume (CAV) Box Based on Measurements

Abstract

HVAC engineers are frequently challenged to design and operate ventilation systems with a high standard of performance considering comfort, energy efficiency, and indoor air quality. However, currently, most design, commissioning and control processes of ventilation systems rely on rules of thumb and engineers' experience. A simulation-based framework for informed decision-making could be an effective tool to achieve superior ventilation systems with optimal design and performance. Nevertheless, to develop such a framework, the integration of solid component models that provide insight into the system's aeraulic behavior is vital. In previous research, a simulation framework known as Air Distribution Network Design (ADND) optimization algorithm was developed. The ADND algorithm provides a basic strategy to design centralized air distribution networks. However, the method is missing some features before it can be used in practice. Currently, the method is limited to generating layout by accounting for the ductwork only. Some ventilation system components (e.g., CAV control box) are not yet integrated. This paper presents the development of a new CAV control box model that is typically used in nonresidential buildings, viz., a mechanically controlled damper that maintains airflow to a predefined fixed airflow level. The model aims to predict the aeraulic performance of the control box at any given inlet volumetric flow rate and set airflow rate (i.e., the airflow index at which the CAV box is commissioned to maintain the flow) for diameters between 125 and 250 mm. First, lab setups were built to measure pressure drops for different CAV diameters by varying the inlet airflow rates and set airflow rates. Next, the measurement data was used to develop a model of the CAV control box by training a regression model. Finally, the model was tested and validated on experimental data that was not used in the training set of the regression model. The accuracy of the CAV box model justifies its integration into the ADND algorithm and also its potential to be integrated into common building simulation frameworks. Once integrated, it can be exploited in many applications, including evaluating the performance of designs, automating the iterative balancing process, and optimizing the control strategy of ventilation systems.