Implementation of Artificial Intelligence for Continuous Greenhouse Gas Monitoring and Leak Detection

In the Oil & Gas industry, fugitive methane emissions account for an estimated 57% of total industry emissions, which combined with upstream flaring comprise nearly 2/3rd of the CO2e across the Oil & Gas upstream/midstream/downstream value chains. While hydrocarbons will likely remain an import (dominant) share of the energy & products mix well into the future, decarbonization efforts have been mobilized to ensure that oil & gas is produced and consumed with a lower carbon intensity. Examples of these decarbonization efforts include the Zero Flaring Initiative that many producers around the world have subscribed to and are working toward. However, fugitive emissions and routine venting/flaring remain a tough-to-abate portion of the industry's emissions, and new technologies around emissions detection are being developed to more accurately (and quickly) measure these "invisible" sources greenhouse gas (GHG) emissions. The combination of governmental mandates (e.g. US EPA Super Emitters Program) and carbon pricing mechanisms are increasing the pressure, and also the incentive, for producers to deploy next-generation emissions detection solutions, rather than relying on conventional solutions and theoretical calculations from emissions factors. Within the bucket of continuous emissions monitoring solutions, certain Artificial Intelligence + Metal Oxide Sensor (AI/MoS-based) technologies show promise in terms of the relative cost effectiveness, and also quality & completeness of site-level emissions data capture. This paper evaluates several new emissions monitoring technology types for the purpose of selection for a specific deployment in Oman. The subsequent six month deployment at multiple sites in Oman, characterized by extreme ambient temperature fluctuations and remoteness of the location, resulted in valuable data supporting the efficacy of AI/MoS-based, continuous emissions monitoring technologies in such applications. As such, this study supports the view that AI/MoS should be considered as part of a wider portfolio approach to emissions monitoring within the Oil & Gas industry, across all segments of the value chain and in all areas & environments.