| dc.description.abstract | For heating, ventilation, and air-conditioning engineers, the measurement of system performance is a vital step in ensuring that the system in use is functioning at its optimal functionality. The measurement of water flow rate and the measurement of air flow rate are both among the key measurement efforts in this line of mechanical engineering. Traditionally, physical devices were used for this purpose. Increasingly, however, academic research is exploring the possibility of virtual flow measurement (e.g., Swamy, Song, and Wang 2012).
In practice, virtual measurement has many advantages over traditional measurements that use physical devices. Virtual meters are shown to be cost effective. Unlike physical meters, they do not require extra space, and they do not intervene with air or water flow. Also, if done properly, they are shown to be as accurate, when compared to physical meters.
This thesis explores the theoretical and the practical aspects of using virtual valve flow meters to monitor chilled water flow rate in modern systems that are used in a medical facility in Oklahoma City, OK. Virtual valve flow meters are measurement devices that make use of valve command information (available in building automation systems) and valve differential pressure data (recorded by physical differential pressure sensors) to measure the flow rate of a given valve without relying on any physical flow measurement devices that intervene with the flow rate, require a large space, or impose installation and maintenance cost. In particular, this thesis work tackles some practical challenges that negatively affect the accuracy of virtual measurement of the valve flow meter. It examines the key problem that the valve commands do not necessarily reveal the accurate position of valve. Not only the changes in pressure across the valve matters for virtual measurement, but also the actual opening position of the valve matters. However, the information about valve commands are not necessarily a reliable source for identifying valve position.
It is argued that stiction/resolution and deadband are among the factors that cause the mis-match between valve command and valve position. Stiction is known as the static friction resistance to valve movement, and the actuator resolution is the smallest possible movement that an actuator can initiate. Also, deadband is known as the hysteresis in the reversal of valve movements. All of these may lead to a mis-match between system command and the physical positioning of the valve. Significant virtual measurement inaccuracy is caused by the above mis-match. To improve upon the accuracy, it is vital to examine valves’ dynamic behavior through their characteristic curve. Incorporating stiction/resolution and deadband, a detailed understanding of this behavior enables the researcher to correct for the above mis-match. This correction requires two steps: first, when valve command changes are not large enough to overcome a particular threshold, valve positions should stay the same; second, when the reverse signal changes are not significant to overcome a particular threshold, valve position should again remain the same. These simple steps could be taken when the behavior of valve is fully examined.
Motivated by the possibility of these corrections, the theoretical foundations of improvement in virtual flow measurement is examined in this thesis. More importantly, the application of this improved measurement method is examined for two different sizes of air handling units. The results confirm that the computation of valve dynamic behavior could significantly improve the accuracy of virtual flow meters. In order to show the application of these meters, this thesis also explores how virtual flow rate measurements could be used in assessing the performance of an industry-wide standard that leads to improving energy efficiency. In particular, using virtual meters, the study in this thesis has shown that customizing minimum airflow rate, as instructed by ASHRAE standard 90.1, could lead to significant energy savings in medical facilities.
Further research could be focused on exploring other factors that could affect the accuracy of virtual flow measurement. They could also be focused on testing the effectiveness of virtual valve flow measurement in other facilities, including commercial, and academic facilities. Future academic research in this topic could also be instrumental in generalizing and simplifying the virtual valve flow measurement procedures for application in HVAC industry. Most of the ongoing research has been limited to academic findings and proposals. There appears to be a disconnect between these findings and proposals and industry-application of virtual measurement. There is, in particular, a room for the design of an advanced device that incorporates the estimated coefficients from each air-handling unit’s valve characteristic curve into the required computation for virtual flow measurement. This could, in future, facilitate the industrial application of virtual flow measurement. | en_US |
