ARTIFICIAL INTELLIGENCE IN STEEL STRUCTURAL ENGINEERING: FROM DESIGN OPTIMIZATION TO HEALTH MONITORING

Abstract

The integration of artificial intelligence (AI) into steel structural engineering is transforming traditional workflows by enabling faster, more efficient, and innovative approaches to design, analysis, and health monitoring. As steel structures become increasingly complex, featuring nonlinear material behavior, irregular geometries, and dynamic load responses, conventional methods such as finite element analysis become computationally prohibitive. AI, particularly machine learning (ML), offers data-driven alternatives that accelerate prediction, optimization, and real-time assessment while complementing, not replacing, fundamental mechanics. This review synthesizes advances from 1994 to 2025, focusing on three paradigms: supervised ML for behavior prediction, Inverse Machine Learning (IML) for goal-driven generative design, and Explainable Machine Learning (XML) for trustworthy, interpretable outcomes. Applications span connection behavior modeling, seismic performance forecasting, structural health monitoring via digital twins, and the reuse of sustainable materials. Despite promising results, challenges remain, including data scarcity, lack of code compliance frameworks, and the “black-box” nature of deep learning models. The paper advocates for hybrid physics-AI systems, open data repositories, and regulatory pathways to ensure AI tools are reliable, transparent, and aligned with engineering standards. By bridging empirical intelligence with physical principles, AI holds the potential to enhance safety, reduce costs, and unlock novel steel structural forms for 21st-century infrastructure.

Keywords

Artificial Intelligence, Steel Structures, Machine Learning, Inverse Design.