Integrating FEM, AI, monitoring and Digital Twins for Reliable Design of Steel Structures under Fire

This proposed doctoral project aims to develop a reliability-based framework for the numerical design and assessment of thin-walled steel structures, integrating advanced finite element modeling (FEM), digital twins, machine learning, and fire resistance analysis. While geometrically and materially nonlinear analysis with imperfections (GMNIA) enables realistic simulations, its sensitivity to modeling assumptions limits its direct use in engineering decision-making, particularly under elevated temperatures.

The research will establish a design-consistent digital twin framework, where numerical models are continuously updated using monitoring and experimental data in compliance with Eurocode requirements. The study focuses on welded and bolted connections of thin-walled steel members, incorporating thermo-mechanical behavior under fire conditions.

A combined numerical–experimental approach will be adopted. Experiments will be designed and conducted to validate FEM simulations and support the development of AI-based surrogate models for efficient and reliable prediction. Bayesian updating will enable continuous model refinement.

The project will deliver validated models, datasets, and a digital twin workflow supporting safe, efficient, and fire-resilient structural design.