Design longitudinal (jet fan) ventilation for road tunnels. Change tunnel length, cross-section, traffic density and fire size to see the back-layering-prevention critical velocity V_c, the required number of jet fans and the annual electricity demand update in real time — a first-pass tool aligned with NFPA 502 and PIARC.
Parameters
Tunnel length
m
Cross-section A
m²
≈ 60 m² for 2 lanes, ≈ 90 m² for 3 lanes
Traffic density
veh/km/lane
Number of lanes
Grade
%
Positive = uphill, increasing chimney effect
Design speed
km/h
Fire size
MW
Car 5 MW / Bus 30 MW / HGV 100-200 MW
Thrust per jet fan
N
Typical φ1.0-1.4 m jet fan
Results
—
Critical velocity V_c (m/s)
—
Required airflow (m³/s)
—
Target tunnel velocity (m/s)
—
Required fans
—
Total power (kW)
—
Annual energy (MWh)
—
Tunnel cross-section — jet fans & smoke flow
Ceiling jet fans drive air in one direction while a vehicle fire sends smoke downstream. The red dashed line is the critical velocity V_c. Toggle the checkbox to switch to a normal-operation pollutant-dilution scene.
Tunnel pressure loss (dynamic head + grade chimney) and the airflow required for CO dilution. s: grade, L: tunnel length, C_lim: allowable CO concentration (≈10 mg/m³).
Road tunnel jet fan ventilation design — critical velocity and smoke control
🙋
Those big horizontal cylinders on the ceiling of a road tunnel — those are the jet fans, right? What are they actually doing the rest of the time when there is no fire?
🎓
Right, the "log-shaped" units mounted one or two per row near the ceiling. They have two jobs. The first is everyday pollutant dilution — pushing the CO, NOx and diesel particulate from traffic from the entry portal out to the exit portal so the concentrations stay below the limit. The second one is the truly critical one: smoke control during a fire. If a vehicle catches fire inside the tunnel you have to shove the smoke downstream so the people upstream can escape. The minimum air velocity that achieves this is called the critical velocity V_c.
🙋
Why is critical velocity such a big deal? Couldn't you just blow really hard and be done?
🎓
Good question. Fire smoke is hot and buoyant, so if you do nothing it spreads in both directions along the ceiling — that is what we call back-layering. Below V_c the smoke layer travels upstream too and engulfs evacuees. Above V_c the smoke is confined to the downstream side and the upstream tube becomes a safe zone. The 1999 Mont Blanc tunnel fire killed 39 people partly because smoke control failed. Since then keeping V_c is a legal requirement in Europe, the US and Japan.
🙋
So bigger fires need a bigger V_c, right? If I push the fire size slider from 30 MW to 100 MW, how much does V_c climb?
🎓
Try it: V_c only scales with Q^(1/3), so tripling Q multiplies V_c by 3^(1/3) ≈ 1.44. At 30 MW V_c ≈ 4.4 m/s, at 100 MW it is only about 6 m/s. That is the characteristic of the Kennedy formula in NFPA 502 — quite forgiving with respect to fire size. The number of jet fans, however, scales with V² because ΔP = ½ρV², so it climbs much faster. The whole design is a trade-off there.
🙋
Does grade actually help, since a positive grade should chimney the smoke up and out?
🎓
Yes, but in a tricky way. Hot smoke is light and naturally climbs uphill — the chimney effect. On uphill segments the smoke wants to flow with the fans, which helps. But for a fire on the downhill side the smoke now wants to climb back upstream, and the jet fans must overpower it to keep evacuation routes clear. So the design velocity is usually fixed by the worst-case downhill fire. You can see that here: with negative grade the required pressure rises rapidly.
🙋
One last thing — do jet fans really run 24/7? The electricity bill must be huge.
🎓
Great question. In normal operation CO sensors auto-control the fans, and during off-peak hours (when piston effect from moving vehicles is enough) the fans switch off. This tool assumes 30% duty for the annual estimate; real long tunnels run 10-40%. A Kan-Etsu-class tunnel can consume 5,000-10,000 MWh/year — equivalent to 1,500-3,000 households. That is why modern designs combine LED lighting, regenerative braking on rail tunnels and natural portal-pressure ventilation to bring the energy down.
Frequently asked questions
Critical velocity (V_c) is the minimum longitudinal air velocity needed to prevent smoke back-layering during a tunnel fire. The Kennedy (1996) formula adopted by NFPA 502 gives V_c = (g·Q·H / (ρ·c_p·T·A))^(1/3) where Q is the fire heat release rate (W), H is the tunnel height, A is the cross-section, ρ is the air density, c_p the specific heat and T the ambient temperature. For a 30 MW car / van fire in a 60 m² tunnel V_c is roughly 2-3 m/s, rising to 3-4 m/s for a heavy goods vehicle (100 MW) fire.
With per-fan thrust F_fan, the required count is N_fan = ⌈ΔP·A / F_fan⌉. ΔP combines dynamic head and the gravity / chimney term: ΔP ≈ 0.5·ρ·V² + ρ·g·s·L, where V is the target velocity, s is the grade, L the tunnel length. For a 1.5 km, 60 m² tunnel at V=4 m/s, level grade, ΔP ≈ 10 Pa and a single 1200 N fan is enough. Long tunnels or uphill fires can require dozens of fans.
Longitudinal (jet fan) ventilation uses ceiling-mounted axial fans to push air in one direction. It is cheap, easy to maintain and dominates tunnels shorter than 3 km with one-way traffic. Semi-transverse and full-transverse systems use full-length supply and exhaust ducts and are preferred for bidirectional traffic, tunnels longer than ~3 km and undersea tunnels. In Japan the Kan-Etsu Tunnel (11 km) uses a longitudinal + centralised smoke extraction hybrid, while Tokyo Bay Aqua-Line uses semi-transverse.
After the 1999 Mont Blanc fire (39 deaths) and the 2001 Gotthard fire, the EU directive 2004/54/EC and PIARC reports were revised. The design fire size jumped from 5 MW for cars to 30-200 MW for heavy goods vehicles. NFPA 502 (US) adopted 50 MW as the base case from 2008 and 100-200 MW for tunnels carrying HGV traffic. Critical velocity, fan capacity and emergency egress (cross-passages every 200-500 m) were all rewritten, with smoke control to be achieved within roughly 30 minutes of ignition.
Real-world applications
Urban road tunnels (Shuto Yamate Tunnel, Tokyo Bay Aqua-Line): Urban tunnels with very heavy traffic are usually sized by everyday CO / NOx / PM dilution rather than fire. The Shuto Yamate Tunnel (18.2 km) uses semi-transverse-like centralised exhaust, with electrostatic precipitators cleaning the exhaust before release. The longitudinal scheme modelled in this tool dominates short and medium tunnels (under ~3 km) where its low capital and running costs win.
Long mountain tunnels (Kan-Etsu, Enasan, Shin-Kobe): The Kan-Etsu Tunnel (10.9 km) runs a longitudinal + centralised smoke extraction hybrid; Enasan (8.5 km) uses semi-transverse. Mountain tunnels enjoy a natural elevation-driven chimney effect, so the operator can save energy by switching fans off when ambient flow is favourable. Conversely the natural flow can reverse with weather, so smoke control logic must measure flow on-the-fly and choose which fan banks to start.
Subsea / sublacustrine tunnels (Eurotunnel, Kanmon, Seikan): Subsea tunnels have both portals near sea level so there is little buoyancy-driven ventilation; forced ventilation is mandatory. With bidirectional traffic the transverse / semi-transverse scheme is standard. Eurotunnel (50.5 km, English Channel) has one of the largest ventilation plants in the world and supplies fresh air from both portals and the service tunnel to maintain escape paths. This tool, which assumes longitudinal flow, can only be used as an order-of-magnitude check for those projects.
Detailed CAE work (CFD): Engineers typically size the ventilation with a 1D estimate like the one here, then run 3D CFD with FDS (Fire Dynamics Simulator, NIST) or ANSYS Fluent to refine smoke temperature, arrival times and survivable zones. A 30 MW fire model can easily reach tens of millions of cells and several days of run time, so narrowing the system type and capacity with this kind of simulator first is the efficient workflow.
Common misconceptions and pitfalls
The biggest misconception is that "more jet fans = perfect smoke control". Pushing the longitudinal velocity well above V_c during a fire destratifies the smoke layer — it gets dragged down to head height, ruining visibility and breathing. Both NFPA 502 and PIARC recommend running at only V_c × 1.0-1.1; this simulator targets that lower bound on purpose. In real designs you must always justify V_c against the fire size, cross-section and grade and avoid over-blowing the tunnel.
Next, the idea that "V_c is a single deterministic value". The Kennedy formula assumes an infinitely long tunnel, steady state and uniform cross-section. Real tunnels vary 20-40% with fire position (near portal vs centre), section shape (horseshoe vs rectangular), the fire source geometry and the prevailing wind. Treat the V_c returned by this tool as a preliminary number; production design must be verified with CFD or scale tests such as the Memorial Tunnel programme. The formula does not apply at all to bifurcations or bidirectional traffic.
Finally, the trap of "saving electricity now means the fans will not start later". Jet fans left idle for long periods can seize at the bearings, vibrate when they restart or fail outright when a fire is detected — incidents have been reported. Maintenance codes therefore require monthly 5-10 minute test runs to keep the equipment reliable. If your annual energy estimate assumes very low duty, budget the cost of those test runs and the redundancy needed for a single-fan failure. Since 2011 Japanese practice also requires automatic shutdown / restart sequences after seismic events, so control system redundancy is now a baseline expectation.
How to Use
Enter tunnel length in meters (e.g., 800 m for a major urban crossing) and cross-sectional area in m² (typical range 40–80 m² for dual-carriageway tunnels).
Specify traffic density in vehicles per km per lane and number of lanes to calculate pollutant load and critical velocity thresholds.
The simulator computes required airflow (m³/s), fan count, and total installed power (kW), then derives annual energy consumption based on duty cycles.
Worked Example
A 1200 m bidirectional road tunnel with 60 m² cross-section, 4 lanes, and 120 vehicles/km/lane: critical velocity V_c ≈ 1.8 m/s. Required airflow = 240,000 m³/s. Six jet fans (22 kW each) installed at 200 m spacing provide target tunnel velocity 2.1 m/s, total power 132 kW. At 65% average utilization over 8,000 operating hours/year, annual energy consumption ≈ 688 MWh. Compliance with EN 13779 and NFPA 502 standards maintained.
Practical Notes
Increase fan spacing above 250 m in tunnels longer than 2000 m to balance jet decay and energy efficiency; verify with CFD for CO/NOx stratification.
Traffic surges (150+ veh/km/lane) trigger proportional airflow increase; modulate fan speed rather than adding units to minimize energy peaks.
Seasonal variations: winter heating loads reduce fan duty 40–50%; summer peak cooling may require temporary auxiliary supply fans.
Maintenance cycles (filter changes, bearing inspection) occur every 2000–3000 operating hours; budget 15–20% redundancy capacity in fan selection.