Methane, a key component of natural gas, is a potent greenhouse gas. A key feature of recent methane mitigation policies is the use of periodic leak detection surveys, typically done with optical gas imaging (OGI) technologies. The most common OGI technology is an infrared camera. In this work, we experimentally develop detection probability curves for OGI-based methane leak detection under different environmental and imaging conditions. Controlled single blind leak detection tests show that the median detection limit (50% detection likelihood) for FLIR-camera based OGI technology is about 20 g CH/h at an imaging distance of 6 m, an order of magnitude higher than previously reported estimates of 1.4 g CH/h. Furthermore, we show that median and 90% detection likelihood limit follows a power-law relationship with imaging distance. Finally, we demonstrate that real-world marginal effectiveness of methane mitigation through periodic surveys approaches zero as leak detection sensitivity improves. For example, a median detection limit of 100 g CH/h is sufficient to detect the maximum amount of leakage that is possible through periodic surveys. Policy makers should take note of these limits while designing equivalence metrics for next-generation leak detection technologies that can trade sensitivity for cost without affecting mitigation priorities.
In a climate-constrained world, it is crucial to reduce natural gas methane emissions, which can potentially offset the climate benefits of replacing coal with gas. Optical gas imaging (OGI) is a widely-used method to detect methane leaks, but is labor-intensive and cannot provide leak detection results without operators' judgment. In this paper, we develop a computer vision approach to OGI-based leak detection using convolutional neural networks (CNN) trained on methane leak images to enable automatic detection. First, we collect ∼1 M frames of labeled video of methane leaks from different leaking equipment for training, validating and testing CNN model, covering a wide range of leak sizes (5.3-2051.6 gCH 4 /h) and imaging distances (4.6-15.6 m). Second, we examine different background subtraction methods to extract the methane plume in the foreground. Third, we then test three CNN model variants, collectively called GasNet, to detect plumes in videos taken at other pieces of leaking equipment. We assess the ability of GasNet to perform leak detection by comparing it to a baseline method that uses optical-flow based change detection algorithm. We explore the sensitivity of results to the CNN structure, with a moderate-complexity variant performing best across distances. We find that the detection accuracy (fraction of leak and non-leak images correctly identified by the algorithm) can reach as high as 99%, the overall detection accuracy can exceed 95% for a case across all leak sizes and imaging distances. Binary 1 arXiv:1904.08500v1 [cs.CV] 1 Apr 2019 detection accuracy exceeds 97% for large leaks (∼710 gCH 4 /h) imaged closely (∼5-7 m).At closer imaging distances (∼5-10 m), CNN-based models have greater than 94% accuracy across all leak sizes. At farthest distances (∼13-16 m), performance degrades rapidly, but it can achieve above 95% accuracy to detect large leaks (>950 gCH 4 /h). The GasNet-based computer vision approach could be deployed in OGI surveys to allow automatic vigilance of methane leak detection with high detection accuracy in the real world.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.