Effects of crosswind and shroud geometry on performance of low-flow, nonassisted flares

Flaring serves as an important safety and emissions compliance tool in industries such as oil and gas production, refineries, and landfills. Nonassisted, low-flow (≤100 thousand cubic feet per day (MSCFD)), utility (pipe) flares are widely used in practice, yet there are limited studies of real-world conditions. Additionally, while shrouds (windshields) are commonly used to mitigate wind effects, their impact on flare performance is previously undocumented. This study introduces a novel outdoor testing facility designed to evaluate low-flow flares and quantitatively assess their performance with and without shrouds. Experiments were conducted at flare-gas flow rates of 5 to 75 MSCFD using natural gas and an 80% natural gas/20% propane blend (by volume) under crosswind speeds from 0 to over 35 miles per hour (MPH). Combustion efficiency (CE) and methane destruction removal efficiency (DRE_CH4) were determined for all operating conditions. While CE for a baseline utility flare (3-inch diameter pipe equipped with a pilot ignition system) was over 96.5% for crosswinds below 10 MPH, the CE decreased rapidly for crosswinds above 10 MPH, with CE <70% for crosswinds above 30 MPH. The utility flare results were compared with results of prior wind-tunnel studies and prior proposed scaling relationships and incorporated into machine learning (ML) models. The scaling relationships show poor correlation with the body of data, but the ML models yielded good agreement (R² = 0.84) when crosswind turbulence intensity was incorporated as an input parameter. The current work investigated retrofitting a utility flare with different shroud designs, which increased CE ≥96.5% for all conditions, demonstrating the effectiveness of shrouds as practical and cost-effective strategies to improve utility flare performance. The results showed low sensitivity to different shroud designs.<i>Implications</i>: The U.S. Environmental Protection Agency (EPA), industry and other monitoring organizations commonly assume flares operate at 98% destruction efficiency; however, recent aerial surveys have revealed efficiencies as low as 91.1%, resulting in up to five times more methane emissions than expected. Low-flow (≤100 MSCFD) utility flares, widely deployed at oil and gas production sites, have limited performance data under real world conditions. This study addresses that gap by providing new experimental data on low-flow utility flares, identifying a new parameter important for predicting flare efficiency and demonstrating a practical solution for significantly reducing emissions.