Wind-Aware Methane Sensor Architecture at a Gas Plant: CFD Analysis of Tree- and Infrastructure-Controlled Plume...

Methane emissions from oil and gas facilities are increasingly monitored using fixed-point sensor networks and continuous monitoring systems, but the effectiveness of these systems depends strongly on sensor location. At the facility scale, plume transport is distorted by buildings, tanks, pipe racks, and surrounding vegetation, so locations that are informative under one wind regime can become ineffective or biased under another. In this study, steady Reynolds-averaged computational fluid dynamics was used to simulate methane dispersion from representative tank and flare-stack release scenarios at a gas plant in west-central Alberta while resolving the combined effects of plant infrastructure and tree canopies. The model domain was 320 m × 320 m × 120 m, and 27 mean-wind cases were examined across three wind directions, three wind speeds (3.6, 9, and 18 km/h), and three ambient temperatures (-15, 5, and 25 °C). The simulations show that wind direction and wind speed are the dominant controls on monitoring design. Temperature enters the present model primarily through density, thereby providing a density-sensitivity check rather than a full atmospheric-stability analysis. Low winds produced broad, wake-sensitive, near-ground plumes, whereas high winds produced 50-70% narrower advective corridors and, in some cases, lifted plume segments above portions of the canopy. Rotation of the wind from south-southwest to south or northeast shifted the informative monitoring sector from one side of the plant to another. The results identify five monitoring rolessource-proximal, dominant-sector downwind, sheltered-wake/canopy-edge, elevated, and background sensorsrather than a single universal layout. The simulations further show that sheltered zones within the plant can locally elevate methane concentrations in the range of 0.5-2 ppm by reducing dilution, potentially biasing release-rate inference if these locations are used without an advective plume and background reference. Overall, the study supports a wind-aware, role-based sensor architecture for gas plants surrounded by vegetation-dense areas.