Potential of Surrogate Modelling for Probabilistic Fire Analysis of Structures

The interest in probabilistic methodologies to demonstrate structural fire safety has increased significantly recent times. However, the evaluation of behavior under loading is computationally expensive even for simple models. In this regard, machine learning-based surrogate modeling provides an appealing way forward. Surrogate models trained simulate engineering (SFE) predict response at negligible computational expense, thereby allowing rapid analyses and design iterations. Herein, a framework proposed analysis exposed structures leveraging modeling. As proof-of-concept (analytical) non-linear model capacity concrete slab advanced (numerical) column are considered. First, procedure training elaborated. Subsequently, developed, followed by evaluate probability density functions capacity. results show that fragility curves developed based on agree with those obtained through direct sampling model, 10–2 quantile predicted error less than 5%. Moreover, cost studies reduced adoption