The Physics of Bridges: Why They Stand Up (And Sometimes Don't)
Bridges are arguments against gravity, made from steel and stone. Tension, compression, resonance, and the occasional catastrophic failure — structural physics at its most dramatic.
Table of Contents
Standing Up Is the Hard Part
Every bridge is in a fight with gravity. That sounds dramatic, but it’s literally true. The entire weight of the structure — plus everything on it — is constantly being pulled straight down, and the bridge has to redirect those forces sideways into its supports without breaking, buckling, or falling over. That’s the whole job. Everything else is details.
The physics isn’t complicated in principle. Newton’s laws say that if the bridge isn’t accelerating (and we’d prefer it wasn’t), the net force on every part of it must be zero. Every downward force needs an equal upward reaction. Every sideways push needs a sideways push back. The trick is making sure those internal forces don’t exceed what the materials can handle.
And that’s where it gets interesting, because materials are surprisingly picky about how you load them.
Tension and Compression: The Two Ways to Stress a Material
There are really only two things you can do to a piece of material: pull it apart or squeeze it together. Pulling is tension. Squeezing is compression. Every force inside every bridge, no matter how complex the design, resolves into some combination of these two.
Here’s the catch: most materials handle one much better than the other.
Stone and concrete are fantastic in compression. You can stack stone blocks to enormous heights — cathedrals, pyramids, Roman aqueducts — because the compressive strength of stone is impressive, around 100–200 MPa for granite. But try to pull stone apart and it fails at maybe 5–10 MPa. That’s a 20:1 ratio. Stone is basically useless in tension.
Steel is the opposite. Well, not exactly the opposite — steel is strong in both tension and compression, roughly 250–500 MPa either way. But steel’s real superpower is tension. You can draw it into thin cables that support enormous loads. A single steel wire 5 mm in diameter can support about 1,000 kg before snapping. Try building that out of stone.
Wood falls somewhere in between, stronger along the grain than across it, decent in both tension and compression but not spectacular at either.
The history of bridge design is basically the story of engineers figuring out how to route forces through materials in ways that play to their strengths. Stone bridges use arches, which convert loads into compression — perfect for stone. Steel bridges use cables and trusses, which exploit steel’s tensile strength. Get the force routing wrong, and you get cracks. Or worse.
The Arch: Ancient Engineering, Perfect Physics
The arch is one of the oldest structural forms, and the physics behind it is elegant.
A flat beam spanning a gap bends under load. The top surface gets compressed, the bottom surface gets stretched (tension), and the middle — the neutral axis — does neither. The bending moment increases with span length, which is why long flat beams need to be very deep (tall in cross-section) or they’ll fail at the bottom surface where tension is highest.
An arch solves this by curving the load path. Instead of bending, the arch directs forces along its curved shape in pure compression — no tension anywhere, if the arch shape is right. The ideal shape for a uniformly loaded arch is a catenary curve (the shape a hanging chain makes, flipped upside down). The ideal shape for a point load is a triangle.
The Romans didn’t know the word “catenary” — they used semicircles, which are close enough for most spans. But they understood the principle intuitively. Roman arched bridges and aqueducts have survived for 2,000 years, carrying loads far beyond what their builders imagined, because stone in compression is nearly indestructible. The Pont du Gard in southern France still stands after 2,000 years. The mortar has mostly crumbled away. The stones hold themselves up through pure compressive contact — geometry doing the work of glue.
The limitation of arches is that they push outward at their base (horizontal thrust). You need something to resist that push — abutments, buttresses, or the ground itself. This is why arch bridges work well in narrow valleys with solid rock walls and less well spanning wide, flat rivers.
Suspension Bridges: Hanging from Cables
For the longest spans, you need suspension bridges, and the physics shifts completely from compression to tension.
A suspension bridge hangs a road deck from vertical cables (suspenders) connected to two main cables that drape between tall towers. The main cables take the entire load in tension and transfer it to the towers, which take it in compression down to the foundations.
The main cable naturally assumes a parabolic shape under the uniform load of the road deck — similar to a catenary but not quite identical (the difference matters to engineers, less so to physics students). The tension in the cable is highest at the towers, where it must support the full horizontal component of the cable force. This tension is enormous: the main cables of the Golden Gate Bridge each contain 27,572 individual steel wires bundled together, with a total diameter of about 0.92 m, carrying roughly 60,000 tonnes of tension.
Why cables? Because steel wire in tension is the most energy-efficient structural material we have. By drawing steel into thin wires, defects that would initiate cracks under bending or compression are aligned along the wire axis, where they don’t cause failure. The tensile strength of bridge cable wire (about 1,500–1,700 MPa) is much higher than the yield strength of structural steel plates (250–350 MPa). You get more strength per kilogram.
This is also why suspension bridges can span distances that no other design can touch. The Akashi Kaikyo Bridge in Japan has a main span of 1,991 metres. The Çanakkale 1915 Bridge in Turkey spans 2,023 metres. Beam bridges top out at maybe 60 metres. Arches at perhaps 500 metres. Beyond that, cables are the only option.
Resonance: When Vibration Kills
Now for the part where bridges stop standing up.
Every structure has natural frequencies — frequencies at which it vibrates most easily. Push a swing at the right rhythm and it goes higher and higher. That’s resonance, and it’s the same physics whether you’re talking about a playground swing or a 2 km bridge.
The problem arises when an external force happens to match one of the bridge’s natural frequencies. Wind, traffic, pedestrian footsteps — all of these deliver periodic forces to the bridge. If the frequency matches, energy piles up. The oscillation amplitude grows with each cycle. If there’s not enough damping (friction, material deformation, viscous effects) to absorb the energy as fast as it arrives, the amplitude grows until something breaks.
The most famous case is the Tacoma Narrows Bridge collapse in 1940, though the mechanism was more complex than simple resonance — it was aeroelastic flutter, a feedback loop between the wind and the bridge’s twisting motion. The textbook “resonance” explanation is an oversimplification, but the underlying point stands: when periodic forces align with structural natural frequencies, bad things happen.
A more recent and less catastrophic example is the London Millennium Bridge. On its opening day in June 2000, thousands of pedestrians crossed the bridge simultaneously. The slight lateral sway of the bridge caused people to unconsciously adjust their walking rhythm to compensate — which synchronised their footsteps — which amplified the sway — which caused more synchronisation. A classic positive feedback loop. The bridge oscillated so violently that it was closed after two days and retrofitted with 37 viscous dampers and 52 tuned mass dampers at a cost of £5 million.
Nobody was hurt in either case, incidentally. The Tacoma Narrows failure was slow enough that everyone got off (except a dog in an abandoned car). The Millennium Bridge was alarming but not dangerous. But both are textbook demonstrations of why dynamic analysis — understanding how structures respond to time-varying forces, not just static loads — is critical.
Fatigue: The Slow Killer
Most bridge failures don’t happen dramatically. They happen slowly, over decades, through fatigue.
Every time a lorry crosses a bridge, the structural members flex slightly — a tiny stress cycle. One cycle does nothing. Ten thousand cycles do nothing. A hundred million cycles, and microscopic cracks begin to form at stress concentrations — bolt holes, weld toes, sharp corners. The cracks grow a tiny amount with each stress cycle. Eventually, a crack reaches a critical length, and the member fails suddenly — even though the load at the moment of failure is nothing special.
Fatigue was not well understood before the mid-20th century. The I-35W Mississippi River bridge collapse in 2007, which killed 13 people, was caused by gusset plates that had been undersized from original construction in 1967 and gradually fatigued over 40 years of traffic loading. The plates were at about half the thickness they should have been, and nobody caught it during inspections.
The physics of fatigue is materials science at its most sobering. Metals that seem perfectly strong under static loading will eventually fail under repeated cyclic loading at stresses well below their yield strength. The endurance limit — the stress level below which a material can survive essentially infinite cycles — is around 40–50% of the ultimate tensile strength for most steels. Design below this limit, and the bridge lasts indefinitely. Exceed it, even by a modest amount, and the clock starts ticking.
Why Old Bridges Look Overbuilt
You’ve probably noticed that Victorian-era iron and steel bridges look massive. Heavy riveted plates, chunky trusses, enormous members that seem like overkill for the loads they carry.
There’s a reason for this, and it’s not that Victorian engineers were wasteful. They didn’t have finite element analysis software. They couldn’t model every stress concentration, every dynamic load case, every fatigue scenario. So they used large safety factors — typically 4 to 6 times the expected load. If you can’t calculate the exact answer, make everything four times stronger than you think it needs to be and sleep well at night.
Modern bridges look slimmer because computer modelling lets engineers calculate stresses precisely and reduce safety factors (now typically 1.5–2.0) without increasing risk. The physics is the same. The confidence in the analysis is higher.
But there’s an argument — and some structural engineers make it — that the Victorian approach was more robust. A bridge with a safety factor of 5 can tolerate a lot of unexpected loading, poor maintenance, and material degradation before it gets into trouble. A bridge optimised to a safety factor of 1.5 has much less margin. The slender, efficient, beautiful modern bridge is perfectly safe under design conditions. The chunky old railway bridge is safe under conditions nobody thought to check.
A Static Problem That’s Never Really Static
Bridges look like they’re standing still. They aren’t. They flex under traffic, sway in wind, expand and contract with temperature (the Golden Gate Bridge is about 0.5 metres longer on a hot day than a cold one), and vibrate at frequencies you can’t see but instruments can measure. A bridge is a dynamic system pretending to be static, and keeping up the pretence is the entire job of structural engineering.
The physics is forces and materials — Newton’s laws, elasticity, thermodynamics, and wave mechanics. Nothing exotic. Nothing quantum. Just classical mechanics applied with enough care and enough margin to make sure that the argument against gravity keeps winning, every day, for a hundred years.
Frequently Asked Questions
Why did the Tacoma Narrows Bridge collapse?
The Tacoma Narrows Bridge collapsed on November 7, 1940, due to aeroelastic flutter — a self-reinforcing oscillation driven by wind. Contrary to popular belief, it was not simple resonance (the wind was steady, not periodic). What happened was subtler: the wind interacted with the bridge's H-shaped deck cross-section in a way that created alternating vortices. These vortices twisted the deck, which changed the angle at which wind hit it, which amplified the twisting. The feedback loop drove oscillations to catastrophic amplitude in about 45 minutes. The bridge had been visibly oscillating for months before the collapse — earning the nickname 'Galloping Gertie' — but nobody understood the underlying physics well enough to intervene.
What is the strongest type of bridge?
There's no single 'strongest' type — the best design depends on the span, the load, the available materials, and the terrain. For short spans (under 60 m), simple beam bridges work fine. For medium spans (60–300 m), arch bridges and truss bridges are efficient because they convert bending loads into compression and tension, which materials handle better. For long spans (300–2,000 m), suspension and cable-stayed bridges dominate because steel cables are extraordinarily strong in tension. The longest bridge spans in the world are all suspension bridges — the Akashi Kaikyo Bridge in Japan holds the record at 1,991 metres. For very long crossings, you need cables.
Why do soldiers break step when crossing bridges?
When a group of people walks in step, their footfalls deliver periodic forces at the walking frequency — about 2 Hz. If this frequency matches a natural resonance frequency of the bridge, the periodic forcing can drive large oscillations. A famous example is the Millennium Bridge in London, which swayed dramatically on its opening day in 2000 when pedestrians unconsciously synchronised their steps. The fix was adding dampers. Soldiers break step to ensure their footfalls are random rather than periodic, avoiding resonant forcing. For large modern bridges this is rarely a concern because of built-in damping, but the tradition persists — and the physics behind it is sound.