APPLICATIONS OF ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING IN STRUCTURAL FIRE RESISTANCE PREDICTION: A SYSTEMATIC LITERATURE REVIEW

Abstract

Structural fire resistance prediction has traditionally relied on empirical formulas and costly experimental testing, which often fail to capture the complex thermal and mechanical behaviors of structures under fire exposure. Recent advances in artificial intelligence (AI) and machine learning (ML) offer promising alternatives to these conventional approaches. In this systematic literature review, we aim to comprehensively map and synthesize the existing research on AI and ML applications in structural fire resistance prediction. Our methodology involved a structured search and rigorous screening of peer-reviewed studies published over the past two decades, followed by a thematic analysis across eight identified dimensions, including fire resistance prediction of structural members, post-fire damage assessment and residual properties, material property prediction under elevated temperatures, and explainable AI integration. The review reveals that neural networks, decision trees, and support vector machines are frequently employed to predict fire-induced responses such as temperature distributions, deflection histories, and load-bearing capacities. We further observe a growing emphasis on hybrid models that combine physics-based principles with data-driven techniques, thereby improving generalizability and trustworthiness. Additionally, the literature highlights applications in advanced composite and strengthened structures, real-time response forecasting, and broader fire safety systems. However, significant challenges remain, including limited high-quality experimental datasets and a lack of standardized validation benchmarks. We conclude that AI and ML hold substantial potential to transform structural fire engineering, though future work must prioritize mechanistic interpretability and robust data curation before these methods can be reliably adopted in practice.