Dynamic modeling and real-time risk assessment of tunnel fires using pressure-state-response (PSR)-enhanced bayesian network
Year: 2025
Type: article
Abstract: Tunnel fires present complex and evolving scenarios owing to confined geometries, limited ventilation, and high traffic density; this often results in severe casualties and significant economic losses. The interplay of multiple fire-related factors introduces high uncertainty in fire evolution, posing challenges to accurate and timely risk assessment. Traditional models fail to capture dynamic progression and are insufficient for evaluating diverse scenarios. To address this, a real-time, quantitative risk-assessment framework is proposed, integrating the pressure-state-response (PSR) model with an enhanced Bayesian network (BN). The PSR structure models fire evolution and emergency response, while the BN quantifies risk using parameters, such as Tunnel temperature, Accident vehicle type, Smoke backflow length , and Smoke layer thickness . Simulations showed that rescue response, Accident vehicle type , and Initial fire scale significantly influenced fire outcomes. Sensitivity analysis identified Heat release rate ( HRR ) and Subsequent fire scale as dominant factors, with “mutual information” of 1.829 and 1.306, respectively. The model was validated using 25 historical tunnel fire cases with a “discrepancy ratio” (DR) of 12 %, thus confirming its predictive capability. Overall, the PSR-enhanced BN was proven to provide a robust and adaptive tool for real-time tunnel fire risk assessment, supporting proactive decision-making and improved safety management.
Source: Results in Engineering
Authors: Yingzhi Zhang, Long Chen, Yuke Shi, Haiying Wang
Institution: Chang'an University
Language: English
Subfield: Safety, Risk, Reliability and Quality
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,940
OpenAlex ID: W4415219761