Techno Press
Techno Press

Steel and Composite Structures
  Volume 58, Number 5, March 10 2026 , pages 647-676
DOI: https://doi.org/10.12989/scs.2026.58.5.647
 


Neuro-swarm fire resistance model of concrete-filled steel tube
Andrei Art Geronimo, Dann Carlo Reformado, Earl Jayson Sarmiento, Crispin Lictaoa, Nolan C. Concha

 
Abstract
    This study investigated the behavior of concrete-filled steel tubes (CFST) under extreme fire conditions, focusing on two key fire performance indicators: the fire resistance rating (FRR) and residual strength index (RSI). Advanced prediction models were developed using neural networks optimized with a particle swarm optimization algorithm. A comprehensive experimental database and a diverse range of neural network architectures were utilized. The models demonstrated superior predictive accuracy, as validated through multiple performance metrics and comparisons with existing prediction equations. Furthermore, causal inference techniques were applied to identify the influence and relative importance of each variable. Visualization tools were instrumental in uncovering patterns and correlations that would be difficult to detect through numerical data alone. The proposed FRR and RSI models offer a cost-effective, non-destructive method for assessing and designing CFST elements in concrete structures.
 
Key Words
    concrete-filled steel tubes; fire resistance rating; machine learning; neuro-swarm; residual strength index
 
Address
Andrei Art Geronimo:Department of Civil Engineering, Adamson University, Ermita, Manila, Philippines

Dann Carlo Reformado:Department of Civil Engineering, Adamson University, Ermita, Manila, Philippines

Earl Jayson Sarmiento:Department of Civil Engineering, Adamson University, Ermita, Manila, Philippines

Crispin Lictaoa:Department of Civil Engineering, Adamson University, Ermita, Manila, Philippines

Nolan C. Concha:Department of Civil Engineering, National University, Sampaloc, Manila, Philippines
 

Techno-Press: Publishers of international journals and conference proceedings.       Copyright © 2026 Techno Press
P.O. Box 33, Yuseong, Daejeon 305-600 Korea, Tel: +82-42-828-7996, Fax : +82-42-828-7997, Email: admin@techno-press.com