A multi-source data-driven tunnel fire source localization methodology and system integrating Bayesian estimation and multi-golden eagle optimization

Yan Li

doi:10.1016/j.firesaf.2025.104469

A multi-source data-driven tunnel fire source localization methodology and system integrating Bayesian estimation and multi-golden eagle optimization

2025-06-30

0

PRE

AI

DOI：10.1016/j.firesaf.2025.104469

Original

Original request for help

Save

Abstract

En 中文

With the increase in urban underground tunnels, fire safety has become a crucial issue, and accurate fire source localization is essential. Previous methods based on sensor arrays or video images have notable limitations. This study proposes a tunnel fire source localization methodology based on multi-source data. A multi-source data fusion model is constructed using Bayesian estimation following data preprocessing. The fitness function is optimized to determine the fire source location by integrating the improved multi-golden eagle optimization algorithm (MGEO). Finally, an intelligent localization system is developed by Unity 3D engine to visualize the fire localization result based on the developed methodology. The effectiveness of the methodology is verified by full-scale experiments and FDS numerical simulations, demonstrating that the MGEO algorithm offers greater accuracy and robustness compared to other algorithms. The results support that the developed methodology and system can provide robust support for tunnel fire safety management and rescue operations.

A multi-source data-driven tunnel fire source localization methodology and system integrating Bayesian estimation and multi-golden eagle optimization

Abstract

Journal

Organization