Experimental characterization of the impact of unsteady airflows on tracer gas measurement

Abstract

Natural ventilation induces several issues when trying to measure airflows. Low pressure differences that drive natural airflows are very sensitive. A small pressure drop introduced by an airflow-meter would significantly affect the flow pattern. Furthermore, for window aired buildings or rooms, the implementation of a protocol to characterize the flow through a window is practically hard to implement and often poorly reliable. For these reasons, tracer gas methods are widely used to measure natural airflows. They do not interfere with the flow path, as they do not require any other instruments than concentration sensors. The principle is to inject a tracer gas inside a chamber and to observe the evolution of the concentration of the tracer gas. The knowledge of the emission rate of the gas, as well as its concentration evolution allows describing the airflow. Unfortunately, tracer gas methods are highly sensitive to the measurement noise, which is likely to be significant for concentration sensors. Most used methods, in the literature, use regression techniques to reduce the measurement noise. However, the regression implies steady airflows during the measurement, which should theoretically discard natural airflows. For lack of better methods, these techniques are often used in natural conditions, regardless of the variation of airflows. The present paper experimentally assess the impact of variable airflows on the accuracy of most common tracer gas methods, namely the constant injection and the concentration decay methods. The experiment was implemented in a laboratory cell, taking advantage of a controllable extract fan to simulate natural airflows, while allowing the direct measurement of the airflow in the extract duct. This value provides the reference value for the comparison of tracer gas methods results. Insights are also given, using signal-processing techniques to improve the reliability of tracer gas methods under variable airflows.