› Bridge Deck Flutter Back
Bridge Wind Engineering

Bridge Deck Aerodynamic Flutter Simulator

A real-time tool to estimate the aerodynamic flutter critical wind speed and vortex-induced vibration lock-in speed of long-span bridge decks. Adjust bridge type, span, deck width and depth, mass and natural frequencies to evaluate the margin against the design wind, the reduced velocity, and the Strouhal frequency — and to build intuition for the wind-resistant design lessons that came out of the 1940 Tacoma Narrows collapse.

Parameters
Bridge type
Sets section-shape factor s_F (streamlined = more stable)
Main span L
m
Deck width B
m
Dominant parameter in the flutter critical speed
Deck depth D
m
Drives the vortex shedding frequency f_v = St U/D
Mass per unit length m
kg/m
Vertical bending freq f_b
Hz
Torsional freq f_t
Hz
Higher f_t / f_b ratio gives more flutter resistance
Design wind V
m/s
10-min mean speed at deck level, 100-year return
Results
—
Mass ratio mu
—
Flutter critical U_cr (m/s)
—
Margin U_cr / V
—
Reduced velocity U_red
—
Vortex shedding f_v (Hz)
—
Lock-in wind U_lock (m/s)
—
Deck cross-section, wind, torsion and vortex shedding

Blue arrows are the oncoming wind, the yellow S-shapes are the alternating Karman vortices, and the deck tilts in torsion. Once the margin drops below 1.2, the amplitude grows and the deck enters the flutter divergence regime.

Flutter critical wind speed vs frequency ratio f_t / f_b
U_cr by bridge type (same conditions)
Theory & Key Formulas

$$\mu = \frac{m}{\pi\,\rho\,(B/2)^{2}}, \qquad U_{cr} = 2.5\,B\,f_{t}\,\frac{\sqrt{\mu}}{s_{F}}$$

Mass ratio mu and Selberg flutter critical wind speed U_cr. m = mass per unit length, rho = air density (1.225 kg/m^3), B = deck width, f_t = torsional natural frequency, s_F = section-shape factor (smaller for more streamlined sections).

$$U_{red} = \frac{V}{f_{t}\,B}, \qquad f_{v} = \frac{St\cdot V}{D}, \qquad U_{lock} = \frac{f_{b}\,D}{St}$$

Reduced velocity U_red, vortex shedding frequency f_v, and VIV lock-in wind speed U_lock. St = 0.12 is a representative Strouhal number for bluff bridge decks; V is the design wind, D the deck depth, f_b the vertical bending natural frequency. Lock-in occurs when f_v approaches f_b.

Bridge deck flutter and aerodynamic instability — the Tacoma Narrows lesson

🙋
The Tacoma Narrows Bridge is that famous video where the deck twists wildly and falls into the river, right? Was the wind really that strong that day?
🎓
Exactly — it collapsed on November 7, 1940. The wind was only about 19 m/s (38 mph), nothing like a typhoon, just a moderate gale. The original design only considered static wind pressure pushing the deck sideways and completely ignored the possibility of the deck twisting itself. Because the deck was a slim plate-girder cross-section, 19 m/s was enough for the aerodynamic forces to couple with the torsional vibration and cross the flutter critical speed. A textbook wind-pressure calculation could never have predicted that failure mode.
🙋
If I switch the Bridge type on the left to "Truss", the margin suddenly drops below 1 and turns red. But a truss looks stronger than a suspension deck — why is it worse?
🎓
Good catch. "Strong" can mean two things: a truss has high static stiffness and bending strength, but aerodynamically it is bad news. The open frame and sharp angles shed lots of vortices, and that turbulent wake couples strongly with the deck's torsional motion. In this tool the section-shape factor s_F captures that effect — 1.3 for trusses, 0.7 for streamlined box girders. Since s_F sits in the denominator of U_cr = 2.5 B f_t sqrt(mu) / s_F, a bigger s_F means a lower flutter speed. That is precisely why the Akashi Kaikyo Bridge chose a streamlined stiffening deck instead of a truss.
🙋
So to raise U_cr we want a heavier, wider, torsionally stiff deck. But what is this "reduced velocity" U_red?
🎓
U_red = V / (f_t * B) is a non-dimensional wind speed: it tells you how many deck widths of wind pass by during one torsional cycle. It is the workhorse of wind-tunnel scaling — match U_red between the model and the prototype and the flow fields are similar. Empirically, flat-deck flutter tends to occur around U_red of 8 to 12, so the default 10 here sits right in the danger zone. Slide the wind up and you also push U_red up, eventually crossing the lock-in band and triggering VIV on top. Real designers always cross-check the U_red at design wind against their wind-tunnel data.
🙋
Vortex-induced vibration and flutter both shake the bridge with wind — what is the practical difference?
🎓
The decisive question is "does the amplitude grow by itself?" VIV is a forced response: the Karman shedding frequency f_v = St U / D happens to hit a natural frequency, the deck vibrates while you are in the lock-in band, and as soon as the wind moves outside that band the motion subsides. So VIV causes fatigue and discomfort but rarely sudden collapse. Flutter is self-excited: once U exceeds U_cr the amplitude grows exponentially and nothing stops it. Tacoma was torsional flutter — the twist amplitude grew to roughly +/- 35 degrees before the deck tore. Remember: VIV brings complaints, flutter brings the bridge down.
🙋
Last one — how do modern long-span bridges actually defend against flutter? Just making them heavier is not enough, right?
🎓
Four ingredients usually combine. (1) Streamlined cross-section — Akashi Kaikyo (1991 m main span) and the Honshu-Shikoku bridges all use smooth flat box decks to suppress vortex shedding in the first place. (2) Central slots — a longitudinal slot through the deck centre lets the upper and lower flows communicate and kills the asymmetric pressure. (3) Edge fairings — small canard surfaces along the deck edges to clean up the flow. (4) Dampers — tuned mass dampers and viscous dampers absorb vibration energy. The design workflow is: estimate U_cr with a quick formula like this tool, measure it in a section-model wind tunnel, and confirm it in a full-aeroelastic model tunnel using a 10 to 30 m scale model. A flutter margin of at least 1.2 to 1.5 over the design wind is the industry baseline.

Frequently asked questions

Flutter is a self-excited oscillation in which aerodynamic forces couple with the deck's vertical bending and torsional motions, so that the amplitude grows by itself. Above a critical wind speed U_cr, the air feeds net energy into the structure each cycle and damping can no longer stop it. The 1940 Tacoma Narrows Bridge suffered torsional flutter at roughly 19 m/s and the deck broke apart within about an hour. Unlike resonance, flutter does not need a matching forcing frequency — once the wind exceeds U_cr the bridge picks its own frequency and runs away.
Selberg's classical estimate assumes a thin-plate-like deck and gives the flutter critical wind speed in terms of deck width B, torsional natural frequency f_t and mass ratio mu = m / (pi * rho * (B/2)^2). A higher mu (heavier or narrower deck) means the wind has less leverage to drive the motion, so U_cr rises. Real deck sections require empirical correction — truss decks shed many vortices and lower U_cr, while streamlined box girders raise it. This tool applies a section-shape factor s_F to capture that trend.
Vortex-induced vibration (VIV) is a forced response that occurs when the Karman vortex shedding frequency f_v = St * U / D matches the deck's vertical bending natural frequency f_b. Amplitudes grow only inside the lock-in wind range and subside once the wind moves past it. Flutter, by contrast, is self-excited: once U exceeds U_cr the amplitude grows exponentially and does not stop. VIV is a fatigue and serviceability problem; flutter is a catastrophic collapse problem.
Modern long-span bridge design uses four layers: (1) section-model and full-aeroelastic wind-tunnel tests to measure U_cr; (2) streamlined box decks (Akashi Kaikyo and Honshu-Shikoku bridges) to suppress vortex shedding; (3) aerodynamic devices like central slots and fairings; (4) tuned mass dampers (TMDs) and viscous dampers to add damping. A flutter margin of 1.2 to 1.5 times the design wind speed is the usual target, matching this tool's margin thresholds. The shift from stiffening trusses to streamlined boxes after Tacoma was driven precisely by this aerodynamic stability requirement.

Real-world applications

Conceptual design of long-span suspension and cable-stayed bridges: Once the main span exceeds about 500 m, securing an adequate flutter critical speed becomes the primary constraint on deck cross-section. The Akashi Kaikyo Bridge (1991 m main span, streamlined box with a central slot), the Honshu-Shikoku bridges, and the Humber Bridge in the UK (1410 m) all start from Selberg-style hand calculations like this tool, then verify the result by section-model and full-aeroelastic wind-tunnel testing.

Retrofitting of older bridges: Many suspension and truss bridges built before the 1980s lack today's required flutter margin. The classic mitigations — TMDs added after construction, edge fairings retrofitted onto the deck, or a longitudinal slot cut into the centre of the deck — have all been applied to older crossings around the world. Switching this tool to the truss bridge type and entering the original deck dimensions makes it intuitive why such retrofits were needed.

Vortex-induced vibration on footbridges: Pedestrian and medium-span bridges with a low deck depth D and low f_b can have their lock-in wind speed U_lock land squarely in the everyday wind range of 5 to 15 m/s, so they rattle on a daily basis. The London Millennium Bridge (lateral synchronous pedestrian excitation) and several auxiliary bridges in Tokyo Bay had to be retrofitted with TMDs after opening. If U_lock falls below the design wind in this tool, that is a clear sign you need a VIV countermeasure during design.

Pre-screening for CFD and wind-tunnel campaigns: A detailed CFD or wind-tunnel campaign for a long-span bridge can cost millions of yen per case. Running candidate deck cross-sections through a quick estimate like this tool — checking whether the flutter margin falls below 1.5 — lets you prune the candidates you take to the wind tunnel. Conversely, if a CFD result deviates by an order of magnitude from this estimate, it is a sanity-check signal that something is off with the modal modelling or the aerodynamic derivatives.

Common misconceptions and pitfalls

The biggest trap is to confuse flutter with resonance. Resonance is a forced response that peaks when the forcing frequency matches a natural frequency, and it dies down as soon as the forcing stops. Flutter is self-excited: above the critical wind speed U_cr the amplitude grows by itself, with no need for the wind to contain a specific frequency. Watching the Tacoma footage and concluding "the wind happened to match the bridge's natural frequency that day" is only half right — the correct statement is "the wind exceeded the critical speed, so the bridge selected its own frequency and diverged". Adding damping is a good remedy for VIV but only slightly raises U_cr; flutter is fundamentally cured by improving the deck cross-section.

The second pitfall is to treat Selberg's number as the answer. The formula U_cr = 2.5 B f_t sqrt(mu) / s_F implemented here is a classical thin-plate approximation multiplied by an empirical section factor. For a real bridge you measure the flutter derivatives (A_i*, H_i*) in the wind tunnel and solve Scanlan's coupled equations to obtain U_cr. This tool is for order-of-magnitude intuition, not for final design numbers. In particular, the gap between s_F = 1.3 (truss) and s_F = 0.7 (box) on real bridges can be a factor of 2 to 3, even wider than the spread shown here.

Finally, do not assume "margin of 1.2 means safe". The design wind V is typically a 10-minute mean speed at a 100-year return period, but the instantaneous peak can be 1.3 to 1.6 times higher, and gust factors push it further. The design wind also has to be adjusted for terrain roughness (urban vs offshore) and for height (a deck 100 m above sea level sees much more than 10 m). The V in this tool is the deck-level representative wind; on a real project you must combine site wind data and topographic effects (canyon winds, slope acceleration). The industry-standard margin is 1.5 or more, and bridges like the Honshu-Shikoku set use margins of 2.0 and above.

How to Use

  1. Enter deck geometry: span length (m), deck width (m), and deck height (m) for your bridge cross-section
  2. Input deck mass per unit length (kg/m) to establish the mass ratio μ
  3. Run simulation to obtain flutter critical wind speed U_cr (m/s), vortex shedding frequency f_v (Hz), and lock-in wind speed U_lock where resonance occurs
  4. Compare ambient wind speed against U_cr margin to verify flutter stability

Worked Example

A cable-stayed bridge deck: span 400 m, width 35 m, height 3.2 m, deck mass 12,500 kg/m. Simulator calculates mass ratio μ ≈ 8.2, flutter critical speed U_cr = 58 m/s, vortex shedding frequency f_v = 0.34 Hz at design wind 15 m/s, and lock-in speed U_lock = 22 m/s. Reduced velocity U_red = 8.1 indicates stable operation below flutter threshold with 36 m/s safety margin.

Practical Notes

  1. Mass ratio μ below 5 requires active damping; most modern decks target μ ≥ 6 for passive stability
  2. Lock-in occurs when vortex shedding frequency synchronizes with deck natural frequency—verify structural damping ratio exceeds 1.5% to prevent lock-in amplification
  3. Reduce deck height and increase mass for flutter-critical long spans (exceed 500 m); Akashi Kaikyo bridge employs aerodynamic fairings to raise U_cr above 100 m/s
  4. Validate inputs against bridge classification wind speed per Eurocode 1 or local codes before design approval