Abstract
Permanent flow measurements are critical for variable air volume air handling unit controls and operations. In practice, however, the accuracy of airflow measurements is usually compromised by less than ideal measurement locations and insufficient measurement points. Because of extra non-symmetric turbulence at less than ideal measurement locations, a limited number of measurement points usually does not provide full coverage of the velocity distribution in ducts and may lead to significant bias in flow rate measurements. A standard traverse like Equal Area or Log-Tchevycheff could be used to identify representative measurement points by providing a reference mean velocity using one-time traverse measurement under the same flow conditions. However, these selected measurement points might not always represent the mean velocity in ducts under different flow rates due to the non-homothetic velocity distribution changes on a cross-sectional plane under different flow rates. This paper presents a theoretical approach to quantifying overall airflow measurement uncertainty that considers errors caused by non-homothetic velocity distribution changes when fewer measurement points are used. The theoretical analysis for studied four AHUs shows that no prohibitive error can be observed by performing 3-point measurements compared with the pure flow meter errors provided by flow meter manufacturers when flow meters are installed at the inlet of a fan. However, error analysis shows that 9-point measurements are needed for obtaining a 5% margin of error suggested by ASHRAE under low velocities (700 fpm) when the flow meters are installed at the outlet of a fan and measurement locations are at 50% shorter than the required straight duct length as suggested by the ASHRAE Standard.
