AI-Based Fatigue Prediction and Integrity Optimization for Gas Pipelines under Transient Conditions

Problem addressed:
Gas pipelines operating under transient conditions (e.g., shutdown/restart, hydrate formation, pressure fluctuations) experience variable-amplitude loading that accelerates fatigue crack growth and reduces service life. Current approaches rely on conservative assumptions or costly simulations, leading to overdesign or unexpected failures.

Proposed solution:
Development of a hybrid computational framework combining:

• Fracture mechanics-based crack growth modeling

• Physics-informed machine learning (PINNs)

• Surrogate models for real-time prediction

This approach enables fast and accurate prediction of fatigue life under realistic operating scenarios.

Key applications:

• Pipeline integrity management (oil & gas)

• Risk-based inspection planning

• Digital twins for pipeline systems

Expected impact (economic & industrial):

• Reduction in maintenance and inspection costs

• Improved prediction of remaining useful life (RUL)

• Enhanced safety and reliability of pipeline operations

• Support for decision-making in energy infrastructure

Collaboration sought:

• Industrial partners (oil & gas operators, pipeline companies)

• European research groups in computational mechanics and AI

• MSCA Postdoctoral candidates working on applied engineering problems

What I offer:

• Expertise in computational mechanics, FG materials, and structural optimization

• Experience in modeling transient loading and fatigue behavior

• Integration of AI with physics-based models for industrial applications