Advanced Offshore Hazard Forecasting to Enable Resilient Offshore Operations

Abstract Hazards in the offshore environment can imperil successful energy operations, whether those operations are conventional, renewable, or for decarbonization. The expanding accessibility of data science and the advanced applications of machine learning (ML) models creates an opportunity to assess potential hazards and the infrastructure they impact. We present a use case demonstrating the combined application of published ML tools to U.S. federal waters of the Gulf of Mexico, an actively explored region for offshore energy that is affected by variable metocean conditions and geologic processes contributing to potential hazards. This paper demonstrates the streamlined use of ML tools to illuminate potential cascading risk events posed by seafloor hazards to pipeline infrastructure, as well as the hypothesized fate and transport of released material in the event of a pipeline rupture. We apply an ML-informed submarine landslide susceptibility tool (ML-LSM) which analyzes spatial features including slope and sediment type using a gradient-boosted decision tree (GBDT) algorithm to produce a susceptibility map in addition to the Climatological and Instantaneous Isolation and Attraction Model (CIIAM), an advanced Lagrangian model to predict material transport based on typical metocean conditions. These models were applied to the Gulf of Mexico to produce spatial results based on the features respective to the seafloor and metocean environments. The Advanced Infrastructure Integrity Model (AIIM) tool, which uses big data and ML models including GBDT and artificial neural networks, was applied to provide a preliminary forecast for pipeline integrity. Intersections of submarine landslide susceptibility, potential for infrastructure integrity loss, and material transport are highlighted amongst the results, which are presented spatially. The combination of these models utilized ML analytics to provide multiscale insights into potential future risks to energy infrastructure. Results of AIIM, in corroboration with various reports, show most existing pipelines are forecasted to be operating past their original design life. The ML-LSM map indicates locations with high landslide susceptibility. Comparing the ML-LSM and AIIM results, we are able identify potentially vulnerable structures that are at greater risk, or those in areas with lower landslide likelihood that may be potential candidates for lifespan extension or retrofitting to support alternative energy strategies. The results of the CIIAM model are integrated to indicate material fate-and-transport complications in the event of infrastructure rupture. Areas of potential bias and recommendations for model tuning are also discussed.