Detection of a simulated gas leak in a wind tunnel

Abstract

This paper brings together considerations of gas leak behaviour and leak detector design and use, with a view to improving the detection of low-pressure natural gas leaks. An atmospheric boundary layer wind tunnel has been used to study ground-based releases of methane at full scale over distances of up to 3 m, under controlled conditions. These scales are relevant to the detection of natural gas leaks from mains and services using hand-portable gas detectors. The mean spatial distribution of the leaking gas plume was determined and used to test and fit a Gaussian dispersion model. This was used for subsequent analysis with respect to the ability of gas leak detectors to confirm and locate a leak. For ground-based leaks, gas concentrations drop rapidly with height such that instruments should ideally sample the air from within 100 mm of ground level. The rapid dilution of gas with distance from the source means that instruments with lower limits of detection, ideally of a few parts per million, have much improved ability to detect a leak from greater distances downwind. Finally, observations showed the variable temporal nature of the gas and the potential for confusion when sampling gas at a single point in time and space.