技術論文とプレゼンテーション

Guided ultrasonic waves enable the inspection of large pipeline sections from only a few sensor locations, making them a promising technology for structural health monitoring of critical infrastructure. However, quantitative characterization of corrosion remains challenging because measured signals are strongly influenced by wave propagation, pipe geometry, and defect morphology.

This work presents the simulation framework developed within the EMUS-TWIN project to establish a physics-based digital twin for guided wave inspection. A three-dimensional transient finite element model is implemented in COMSOL Multiphysics® to simulate the propagation of shear horizontal (SH) guided waves in steel pipelines and their interaction with localized corrosion defects. The model incorporates realistic pipe geometries, parameterized defect configurations, and absorbing boundary conditions to accurately capture wave propagation and scattering phenomena.

The numerical model serves as the foundation for a simulation-driven digital twin. Large simulation datasets are generated to investigate the influence of corrosion geometry on the received guided wave signals. These data are subsequently used to train surrogate models, enabling rapid inverse characterization of corrosion through physics-based optimization while avoiding computationally expensive finite element simulations during online operation.

The proposed workflow combines high-fidelity multiphysics simulation, surrogate modeling, and inverse analysis within a unified digital twin framework. It provides an efficient and physically interpretable approach for quantitative corrosion assessment and demonstrates how COMSOL Multiphysics® can be integrated into advanced condition monitoring strategies for industrial pipeline networks.