Guo et al. (2021) used Fire Dynamics Simulator (FDS) computations of a single 7.5 MW ultra-fast design fire to compare ventilation and smoke extraction in four practical super-long river-crossing subway tunnel configurations — double-line (DLT), unilateral-line (ULT), single-tube (STT), and double-circular (DCT) — and to compare grouped ceiling exhaust layouts (exhaust vent groups, EVGs). They concluded that the ULT showed the lowest ceiling smoke temperature and CO2_2 concentration and that two EVGs at 450 m spacing (2-EVGs) was the least hazardous extraction scheme, with risk ranking 1-EVG > 3-EVGs > 2-EVGs. These conclusions rest on one numerically simulated scenario, with no dedicated physical experiments, one fire size and location, three discrete EVG spacings, and an evacuation analysis decoupled from the smoke field. The open problem is to establish a validated, scenario-robust design basis for such tunnels: confirm or refute the configuration and EVG rankings from any sound evidence (physical experiment, independent or higher-fidelity simulation, or analysis) and across fire size, growth rate, and source location; determine the optimal EVG spacing and exhaust-flow-rate/ventilation-velocity combination; and quantify evacuation consequences with pedestrian behavior coupled to the evolving smoke environment, including fires in the non-river-crossing approach sections.

Background

Super-long river-crossing subway tunnels — metro running tunnels several kilometres long that pass beneath major rivers, such as the roughly 3.85 km river-crossing section examined by Guo et al. (2021) — pose a distinctive fire-safety problem: over most of the river span no intermediate ventilation shafts can be built, escape routes are long, and high riverbed water pressure (reported as up to 0.6 MPa) constrains construction. In tunnel-fire engineering the standard design quantities are the heat release rate (HRR) of the design fire — here 7.5 MW with an 'ultra-fast' t2t^2 growth curve, meaning the fire power grows quadratically in time with the fastest standard growth coefficient — and the tenability metrics used to judge occupant safety: maximum ceiling smoke temperature, carbon dioxide concentration, visibility at 2 m above the floor, and smoke-layer descent height. Smoke is controlled by longitudinal ventilation (a forced airflow along the tunnel axis, 2.3 m/s in the source study) combined with point extraction, in which ceiling vents are opened in groups near the fire; Guo et al. call one such opened cluster an exhaust vent group (EVG), with the flow through each vent fixed at 6.0 m/s, and they compared one, two, and three groups at spacings of 900 m, 450 m, and 300 m. The study was entirely computational, using the Fire Dynamics Simulator (FDS), a large-eddy-simulation computational fluid dynamics code that is the de facto standard for tunnel smoke calculations, validated against published full-scale corridor experiments rather than against experiments reproducing the river-crossing geometries; evacuation was analyzed separately in the agent-based simulator Pathfinder with a uniform walking speed and fixed evacuation direction, then compared with the smoke results through the standard ASET/RSET framework (available versus required safe egress time). This comparative-simulation methodology grew out of the same group's earlier study of the Nanjing Metro Line 10 cross-river tunnel (Guo et al., 2018), which compared single-hole single-track and single-hole double-track structures at different slopes. Guo et al. (2021) concluded that the unilateral-line tunnel (ULT) produced the lowest ceiling smoke temperature and CO2_2 concentration of the four configurations, that the most hazardous extraction scheme was a single EVG and the least hazardous was two EVGs (risk ranking 1-EVG > 3-EVGs > 2-EVGs), and that negatively sloped segments (−0.43°) were the most dangerous locations. Every one of these conclusions is a single-scenario numerical prediction for one design fire placed at the assumed most-unfavorable location (midway between exhaust vents); because all reported parameters come from that single simulated scenario, physical validation, fire-scenario sensitivity, vent-system optimization, evacuation coupling, and fires in the approach sections remain unestablished for this tunnel class.

Problem Statement

For super-long river-crossing subway tunnels of the class studied by Guo et al. (2021) — a roughly 3850 m river-crossing section with a 7.5 MW design fire of ultra-fast quadratic (t2t^2) growth, and grouped ceiling extraction by exhaust vent groups (EVGs, clusters of ceiling exhaust vents opened together near the fire) each operating at 6.0 m/s vent flow under 2.3 m/s longitudinal ventilation (forced airflow along the tunnel axis) — establish a validated and scenario-robust design basis for smoke extraction, jointly answering: (1) Robustness and validation of the rankings: do the numerical conclusions — the advantage of the unilateral-line tunnel in ceiling smoke temperature and CO2_2 concentration among the four cross-sections compared (double-line tunnel, DLT; unilateral-line tunnel, ULT; single-tube tunnel, STT; double-circular tunnel, DCT), and the extraction risk ranking 1-EVG > 3-EVGs > 2-EVGs — survive (a) independent evidence — physical experiments, at full scale or reduced scale reproducing the river-crossing geometries and grouped extraction, or independent or higher-fidelity simulation or re-analysis — and (b) changes in fire heat release rate, fire growth rate, and fire-source location, rather than the single simulated ultra-fast fire at the assumed most-unfavorable midpoint position? (2) Vent-system optimization: how does smoke-extraction performance depend jointly on EVG spacing (beyond the discrete 900/450/300 m layouts), the number of vent groups, exhaust vent flow rate, and longitudinal ventilation velocity, and which configuration minimizes fire risk over the tunnel — including whether better arrangements exist than the 2-EVGs scheme? (3) Evacuation consequence: how does the resulting smoke environment constrain egress when pedestrian wayfinding and walking speed are coupled to the evolving visibility and smoke field, rather than prescribed as uniform and one-directional as in the source study's decoupled evacuation analysis in the Pathfinder agent-based evacuation simulator? (4) Scope extension: how do conclusions (1)-(3) change for fires in the non-river-crossing approach sections of the same super-long tunnel, where different boundary conditions such as intermediate ventilation shafts may be available? These components jointly serve one objective — a validated design basis for fire-safe smoke control and evacuation in super-long river-crossing subway tunnels — with the ranking claims of the source study as the concrete starting point. The coupling of riverbed water pressure with fire-induced thermal loading on the tunnel lining is a related but separate structural fire-resistance question and is not part of this problem's completion.

The four components enumerated above are independently standing sub-questions, not one undifferentiated deliverable: (1) is resolved by an explicit confirm-or-refute verdict on the configuration advantage and on the 1-EVG > 3-EVGs > 2-EVGs ranking, on the basis of any sound evidence and across the fire-size, growth-rate, and source-location variations named above; (2) by a reported optimal combination of EVG spacing, vent-group count, exhaust vent flow rate, and longitudinal ventilation velocity, with the admissible design range and the fire-risk measure it minimizes stated explicitly — fire risk being judged by the tenability metrics of the source study (maximum ceiling smoke temperature, carbon dioxide concentration, visibility 2 m above the floor, and smoke-layer descent height); (3) by a quantified smoke-coupled evacuation consequence; and (4) by the corresponding answers for approach-section fires. Resolving any single component fully answers that sub-question while leaving the remaining components open; the joint objective is met only when all four are addressed.

References

  1. Xinxin Guo, Xuhai Pan, Lijing Zhang, Zhilei Wang, Min Hua, Juncheng Jiang (2021). Comparative study on ventilation and smoke extraction systems of different super-long river-crossing subway tunnels under fire scenarios. Tunnelling and Underground Space Technology 113:103849. DOI 10.1016/j.tust.2021.103849. https://doi.org/10.1016/j.tust.2021.103849
  2. Xinxin Guo, Xuhai Pan, Zhilei Wang, Juan Yang, Min Hua, Juncheng Jiang (2018). Numerical simulation of fire smoke in extra-long river-crossing subway tunnels. Tunnelling and Underground Space Technology 82:82-98. DOI 10.1016/j.tust.2018.08.002. https://doi.org/10.1016/j.tust.2018.08.002
  3. Liangliang Tao, Yanhua Zeng (2022). Effect of Different Smoke Vent Layouts on Smoke and Temperature Distribution in Single-Side Multi-Point Exhaust Tunnel Fires: A Case Study. Fire 5(1):28. DOI 10.3390/fire5010028. https://doi.org/10.3390/fire5010028
  4. Wei Na, Chen Chen (2022). A Study on the Evacuation Spacing of Undersea Tunnels in Different Ventilation Velocity Conditions. Fire 5(2):48. DOI 10.3390/fire5020048. https://doi.org/10.3390/fire5020048
  5. Yaqiang Jiang, Tianhang Zhang, Shuai Liu, Qinli He, Le Li, Xinyan Huang (2022). Full-scale fire tests in the underwater tunnel section model with sidewall smoke extraction. Tunnelling and Underground Space Technology 122:104374. DOI 10.1016/j.tust.2022.104374. https://doi.org/10.1016/j.tust.2022.104374
  6. Yao Hong, Ceji Fu, Bart Merci (2022). Numerical analysis of the performance of a PID control based real-time mechanical ventilation system to prevent smoke back-layering in tunnel fires. Tunnelling and Underground Space Technology 128:104639. DOI 10.1016/j.tust.2022.104639. https://doi.org/10.1016/j.tust.2022.104639
  7. Zhan Wang, Zhi Tang, Zheng Fang, Zhiming He, Tao Yang, Min Zhou, Enshi Wang, Yinhang Xu (2024). Experimental study of smoke characteristics in a double-deck tunnel with multiple point extraction using a branch pipe. Fire Safety Journal 146:104144. DOI 10.1016/j.firesaf.2024.104144. https://doi.org/10.1016/j.firesaf.2024.104144
  8. Yubo Luo, Haoxiang Li, Tongqing Wei, Jialin Ma (2023). Study on the Coupled Smoke Characteristics and Evacuation of Underwater Immersed Tube in Tunnel Fire. International Journal of Frontiers in Engineering Technology 5(6):30-36. DOI 10.25236/ijfet.2023.050605. https://doi.org/10.25236/ijfet.2023.050605
  9. Tian Kun, Jiang Shuping, Xu Pai (2026). Research on Mechanical Response and Damage Thickness of Immersed Tube Tunnel Structure Exposed to High Fire Temperature. Fire and Materials 50(1):11-22. DOI 10.1002/fam.70017. https://doi.org/10.1002/fam.70017
  10. Guanghui Yao, Zhongyuan Yuan, Yuanyi Xie, Nanyang Yu (2025). Smoke control during a fire in tunnels equipped with point extraction ventilation considering different vent settings. Applied Thermal Engineering 279:127673. DOI 10.1016/j.applthermaleng.2025.127673. https://doi.org/10.1016/j.applthermaleng.2025.127673
  11. Zhuyu Shao, Mingshu Bi, Haiyong Cong, Luyang Liu, Fengnan Sun, Yubo Bi, Tianjiao Zhang (2026). Experimental investigation on smoke control and smoke extraction efficiency of two-point extraction system under longitudinal sloped tunnel fires. Tunnelling and Underground Space Technology 174:107656. DOI 10.1016/j.tust.2026.107656. https://doi.org/10.1016/j.tust.2026.107656
  12. Xinyue Zhang, Mingrui Luo, Liangliang Tao, Yanhua Zeng, Yanping Yuan, Zhen Zhang (2025). Analysis of the influence of exhaust vent leakage on the efficiency of tunnel smoke evacuation. Fire Safety Journal 154:104399. DOI 10.1016/j.firesaf.2025.104399. https://doi.org/10.1016/j.firesaf.2025.104399
  13. Xianwang Fan, Huan Zhang, Zhihao Wan, Zhikai Liu, Jiali Liu, Junbin Yang, Sujie Liu, Jiaxuan Pu, Zhaoying Wang, Yan Jiang, Zhangxiang Wu, Shijun You, Wandong Zheng (2024). A simplified model for unsteady airflow analysis in ultra-long tunnels based on the resistance compensation method. Building Simulation 17:1823-1840. DOI 10.1007/s12273-024-1164-z. https://doi.org/10.1007/s12273-024-1164-z