NETL is expanding its methane leak detection technology work with the Southwest Research Institute (SwRI) of San Antonio, Texas, to use machine learning-based algorithms for the quantification of fugitive methane emissions – small leaks that typically go undetected along natural gas pipelines and at storage facilities.
Locating and measuring the quantity of fugitive methane emissions will facilitate the natural gas industry to operate most responsibly. Methane quantification is also an important part of Leak Detection and Repair programs for the U.S. Environmental Protection Agency’s inventory reduction and mitigation prioritization foci.
The new phase of research builds upon previous accomplishments in which the Smart Methane Emission Detection System (SLED/M) was developed. SLED/M is an autonomous, real-time methane leak detection technology capable of aerial operation that uses machine learning techniques and an optical sensor.
Conventional methane detection systems, designed to locate relatively larger leaks, suffer from false positives and missed detections, which hamper effectiveness and industry adoption. However, SLED/M provides a substantial reduction in false positives by using a unique machine learning-based algorithm optimized under a variety of environmental conditions. The algorithm enables methane detection without the need for constant operational monitoring, further reducing the potential for human error.
“This newly added task will continue the important research being done at SwRI,” said Joe Renk, the senior federal project manager at NETL overseeing the work. “The team is aiming to quantify the methane emissions using an optical gas imager. They’re also planning to achieve a much-improved representation of the flow rate by analyzing the entire detected methane region rather than a single-point measurement like current laser systems use. Moreover, the overall cost of the system should also be significantly lower because use of an expensive tunable diode laser absorption spectroscopy laser system isn’t required."
This year, SwRI plans to perform testing, collect data and develop the machine learning-based methane detection algorithm to quantify leaks using the existing SLED/M components with the addition of a low-cost longwave infrared thermal imager and lightweight lidar system. Additional tests, conducted in a variety of environmental conditions to increase the algorithm’s robustness, will provide baseline and refinement. The quantification algorithm will then be integrated into the SLED/M system.
SLED/M has already gained industry interest from its proven ability to detect methane, and the integration of methane quantification will provide companies an all-in-one solution that is accurate, reliable, autonomous, affordable and easily integrated with existing operator-owned equipment. As a recent Stanford study found that only one in nine tested methane detection technologies reliably measured the gas, there is a clear need for improved technologies.
This project is another example of how NETL is fulfilling its mission is to discover, integrate and mature technology solutions to enhance the nation’s energy foundation in an environmentally responsible manner