LEAP-HI: Towards Safe and Efficient Gas Transport Pipeline Network with Learning-enabled Autonomous Risk Assessment Systems (LARAS)

Pipelines play a vital role in transporting energy resources that are essential for America’s economy and the well-being of the American people. A significant challenge facing the domestic pipeline infrastructure is the gradual deterioration of materials and structural components, which can result in serious safety risks. While regular in-line inspection (ILI) of pipelines is crucial for risk management, existing tools are often labor-intensive, expensive, and hard to adapt to different pipeline geometries. This Leading Engineering for America's Prosperity, Health, and Infrastructure (LEAP-HI) award advances fundamental and interdisciplinary research to enable safe, efficient, and automated inspection of gas transport pipelines for accurate risk assessment. The research integrates new knowledge in robotics, non-destructive evaluation (NDE), risk engineering, artificial intelligence (AI), and policy analysis to iteratively develop Learning-based Autonomous Risk Assessment Systems (LARAS). The project will include lab and field testing at different scales, create partnerships with industry and collect stakeholder feedback throughout the planned activities. The broader impacts of the project include the development of new curriculum and undergraduate research opportunities, outreach to local communities and professional societies, and online modules for workforce development. This interdisciplinary project will advance the science and engineering of automated pipeline inspection and risk assessment through the development of novel robotics, sensors, physics models, AI algorithms, and decision-support policies. The project has four core objectives, specifically to: (1) design low-cost and flexible in-pipe inspection robots to autonomously navigate in pipelines with complex configurations; (2) exploit and integrate novel NDE sensing solutions with the pipe inspection robots to collect multimodal data for detecting cracks and leaks; (3) predict failure risks with uncertainty quantification by combining physics-based simulations and machine learning using sensor data and historical accident reports, and; (4) engage stakeholders from industry, government agencies, researchers, and public interest groups to understand the emerging landscape of safety regulations and guide the design and deployment of the risk assessment systems. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.