DESC0023746

Statement of the problem being addressed: The notable increase in the global average surface temperature relative to pre-industrial levels has motivated enhanced efforts to reduce greenhouse gas emissions. Although emissions of both carbon dioxide and methane have contributed to the problem, emissions of methane are especially concerning due to their much larger global warming potential. Therefore, steps are generally taken by oil and natural gas companies to avoid the unrestrained release of methane into the atmosphere. Natural gas flaring is one of these mitigating practices and involves burning excess natural gas that must be disposed of for economic, safety, or operational reasons. However, gas flares often operate with low combustion efficiency and unburned methane is nonetheless vented to the atmosphere as a result. Furthermore, in the event that flameout, or lean blowout (LBO) occurs, exorbitant methane emissions will ensue and defeat the purpose of the gas flare altogether. General statement of how this problem is being addressed: This proposal seeks to develop a digital engineering software package that can be used to maintain a high combustion efficiency in gas flares. Specifically, the software will detect instances of incomplete combustion using data that is acquired in real time from small, low-cost sensors. This feedback can then be used to modify the gas composition and/or flow rates in the gas flare such that the combustion efficiency is increased and unburned methane emissions are avoided. This work will be performed by Spectral Energies in collaboration with the Georgia Institute of Technology. What is to be done in Phase I: Spectral Energies’ software package will leverage machine learning algorithms that were previously developed by this research team to identify the local flame extinction processes that precede LBO. These local extinction processes are indicators that LBO is imminent, and they are the underlying physical phenomena behind incomplete combustion. The digital engineering software package will be developed and validated using experimental data from combustor test cells at Georgia Tech that appropriately replicate an industrial gas flare. In these experiments, flame imaging data will be acquired simultaneously with combustion efficiency and exhaust gas composition measurements. Therefore, sufficient validation data will be available to confirm that unburned methane does not pass through the flame without activating an alarm from the software. Commercial applications and other benefits: SE’s digital engineering software will allow for the reduction of preventable methane emissions from natural gas flares by ensuring that they operate with a high combustion efficiency. Oil and natural gas companies are anticipated to comprise the largest market for this product because it would allow them to comply with government gas flaring regulations without requiring costly technological upgrades. Although technologies exist that can capture the natural gas that would otherwise be flared, oil and natural gas companies have historically been hesitant to adopt these technologies due to their cost, risk level, and complexity. In contrast, Spectral Energies’ software package only requires input from a few low-cost sensors that could be easily integrated with existing gas flare technology.