A Bridge Truss is a rigid framework of straight members arranged in triangles that carries vertical bridge loads through pure axial tension and compression rather than bending. Civil and structural engineers rely on Bridge Trusses for medium-span highway, railway, and pedestrian crossings where a solid girder would be heavy and expensive. Top and bottom chords resist the global bending moment as compression and tension, while diagonal and vertical web members shuttle shear between them. The result — long clear spans, typically 30 to 200 m, built from comparatively light steel or timber sections.
How the Bridge Truss Actually Works
A Bridge Truss converts a beam-bending problem into an axial-force problem. You have a deck spanning from one abutment to the other, and instead of fighting that span with one deep beam, you build a tall framework of triangles. The top chord squeezes (compression), the bottom chord stretches (tension), and the diagonals and verticals — the web members — pass shear back and forth between them. Because every member carries load along its length, you can use slender sections that would buckle or yield instantly if loaded sideways. That is why a 60 m Pratt truss can be built from rolled W-shapes and angles rather than a 3 m deep plate girder.
The geometry has to be right or the whole logic falls apart. Members must meet at clean panel points — the work points where the centroidal axes intersect — within roughly 3 mm on a fabricated steel truss. Eccentricity at a panel point introduces a bending moment the member was never sized for, and you will see it as cracked paint at the gusset plate edge or, worse, as a bowed compression diagonal. The gusset plates themselves are sized for the resultant of all members framing in, and the bolt pattern must develop the full member force without block-shear tearout. If you notice a diagonal humming or vibrating under traffic, your panel-point connection is slipping or the member is fluttering near its Euler buckling load.
Failure modes are well-known and worth naming. Compression chords buckle — historically the failure mode behind the 1907 Quebec Bridge collapse, where the lower chord lacing was undersized. Tension members fail by net-section fracture at the bolt holes if the holes were not deducted in design. Fatigue cracks initiate at gusset-plate weld toes under cyclic truck loading, which is why modern AASHTO detailing demands ground-smooth transitions on Category C details. And corrosion in pack-rust pockets between gusset plates and chord flanges is what brought down the I-35W bridge in Minneapolis in 2007. None of these are exotic — they are predictable consequences of skipping inspection or under-detailing a connection.
Key Components
- Top Chord: The continuous compression member running along the top of the truss. Sized for axial compression plus a column-buckling check using the unbraced length between panel points — typically L/r ratios kept below 120 for main members. On a 60 m highway truss, top chord forces commonly run 2,000 to 6,000 kN.
- Bottom Chord: The tension member along the bottom. Easier to size than the top chord because tension members do not buckle, but the net section through bolt holes must be checked. A typical bottom chord uses a built-up box or W-section with effective net area at least 85% of gross.
- Diagonal Web Members: Carry shear between the chords. In a Pratt truss the diagonals slope toward the centre and act in tension under gravity load — efficient because tension members are lighter. In a Howe truss the diagonals slope outward and act in compression, which suits timber better than steel.
- Vertical Web Members: Subdivide the panels and carry localised floor-beam reactions up into the chord system. Panel spacing typically 4 to 8 m — tighter panels mean lighter floor beams but more connections to fabricate.
- Gusset Plate: The flat steel plate that connects multiple members at a panel point. Thickness usually 12 to 25 mm depending on member forces. The bolt group and plate dimensions must develop the full factored member force with at least 1.5× safety against block shear.
- Floor Beam and Stringers: Transverse floor beams hang from the bottom-chord panel points and support longitudinal stringers that carry the deck. This system delivers wheel loads to the truss only at panel points — never mid-chord — preserving the pure axial-force assumption.
- Lateral Bracing: Horizontal X-bracing in the plane of the top and bottom chords resists wind load and prevents lateral-torsional buckling of the compression chord. Sized for at least 2% of the chord force as a stability bracing demand.
Where the Bridge Truss Is Used
Bridge Trusses appear wherever you need to span 30 m or more economically, in steel, timber, or aluminium. The form dominates short-line railway crossings, pedestrian and cycle bridges, temporary military and disaster-relief spans, and a huge fraction of pre-1970 highway infrastructure still in service. Different industries name the same configuration differently — a railroad calls it a through truss, a structural fabricator calls it a Pratt or Warren depending on diagonal orientation, and military engineers call panelised modular trusses Bailey bridges. Bridge Trusses also show up scaled-down as roof trusses in industrial buildings, and scaled-up as the stiffening trusses inside suspension bridges like the Forth Road Bridge.
- Highway Infrastructure: The Ironton-Russell Bridge replacement in Ohio used a 274 m through-arch truss configuration; thousands of shorter Pratt and Warren highway trusses span 30-100 m across rivers in the US Midwest.
- Railway Engineering: BNSF and CN routinely replace 1920s-era Pratt deck trusses on branch lines with bolted Warren trusses fabricated by companies like Hirschfeld Industries — typical span 40 to 80 m carrying Cooper E80 loading.
- Military and Emergency Bridging: The Bailey bridge developed by Donald Bailey in 1940-41 is a panelised Bridge Truss assembled by hand; modern equivalents include the Mabey Logistic Support Bridge spanning up to 80 m without cranes.
- Pedestrian and Cycling: Continental Bridge and US Bridge supply prefabricated Warren-truss pedestrian spans 20 to 60 m long, weathering-steel weldments delivered on a single flatbed.
- Heritage Restoration: Covered timber Howe trusses across New England — like the 137 m Cornish-Windsor Bridge between Vermont and New Hampshire — use timber compression diagonals with iron tension verticals, a 19th-century Bridge Trusses pattern still serviceable today.
- Industrial and Mining: Conveyor gallery trusses spanning 50 to 150 m at iron-ore terminals like Port Hedland use Warren or modified Pratt configurations to carry belt conveyors plus walkway loads.
The Formula Behind the Bridge Truss
The single most useful calculation on a Bridge Truss is the force in the chords at midspan, because that force drives section selection for the heaviest members. You treat the truss as a simply supported beam, compute the global bending moment M at midspan, and divide by the truss depth d to get the chord force. At the low end of the typical range — short pedestrian spans around 20 m with light pedestrian loading — chord forces stay well under 500 kN and you can use a single rolled HSS or W-section. At the high end — a 100 m highway through-truss carrying HL-93 loading — chord forces push past 5,000 kN and you are into built-up box sections with internal stiffeners. The sweet spot for economical fabrication sits around 40-60 m span with chord forces in the 1,000 to 2,500 kN band, which is exactly why so many North American highway trusses cluster in that range.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Fchord | Axial force in the top or bottom chord at midspan | kN | kip |
| M | Maximum bending moment at midspan from factored uniform load | kN·m | kip·ft |
| w | Uniformly distributed factored load per unit length on the truss | kN/m | kip/ft |
| L | Span between bearing centres | m | ft |
| d | Depth of the truss between top and bottom chord centroids | m | ft |
Worked Example: Bridge Truss in a rural single-lane highway Pratt truss
A county engineering office in Otsego County, New York is sizing a new 48 m single-lane Pratt through-truss to replace a deteriorated 1948 stringer bridge over the Susquehanna headwaters. The truss carries an HL-93 design load that, combined with deck self-weight and an asphalt overlay, produces a factored uniform load of 42 kN/m on the truss. The proposed truss depth is 6.0 m centroid-to-centroid, with 8 panels at 6 m each. The team needs the midspan bottom-chord tension force to select the W-section and verify the bolted splice.
Given
- L = 48 m
- w = 42 kN/m
- d = 6.0 m
Solution
Step 1 — compute the global midspan moment from the factored uniform load:
Step 2 — divide by truss depth to get the nominal chord force at the 6.0 m proposed depth:
That 2,016 kN is comfortably handled by a W14×233 grade 50 section — a workhorse member at any North American steel fabricator.
Step 3 — check the low end of the practical depth range. Architects and clearance constraints sometimes push designers toward shallower trusses. At d = 4.5 m (span-to-depth ratio L/d = 10.7):
That is a 33% jump in chord force for a 25% reduction in depth — you would need a W14×283 or built-up section, and the connection plates grow accordingly. Step 4 — check the high end at d = 8.0 m (L/d = 6, a deeper-than-typical truss):
Chord force drops 25% versus nominal, letting you drop to a W14×176 — but the web members get longer, the verticals carry more, and total steel tonnage often goes up rather than down. The sweet spot for this span sits right around L/d = 8, which is exactly where the 6.0 m proposal lands.
Result
The nominal midspan bottom-chord tension force is 2,016 kN, which sizes cleanly to a W14×233 grade 50 section with a bolted field splice using 22 mm A325 bolts. In practice that means you are designing a member roughly 360 mm deep weighing about 347 kg/m — a section a single rigger can spot from a crawler crane without exotic lifting fixtures. The low-depth case at 4.5 m drives chord force up to 2,688 kN and forces a heavier built-up section, while the high-depth case at 8.0 m relaxes chord force to 1,512 kN but penalises the web members and lateral bracing. If the as-built member shows higher measured strain than predicted, the usual culprits are: (1) deck self-weight underestimated because the asphalt overlay accumulated more lifts than the original drawings show, (2) panel-point eccentricity introducing parasitic bending into the chord, or (3) a slipped bolted splice transferring force through bearing rather than friction, which raises peak strain at the splice plates by 15-25%.
Choosing the Bridge Truss: Pros and Cons
Bridge Trusses compete with plate girders, cable-stayed bridges, and concrete arches across the medium-span range. Choosing between Pratt, Warren, Howe, and the broader family of Bridge Trusses against these alternatives comes down to span, fabrication economy, depth available, and how aggressive your fatigue exposure is. The dimensions below are what real designers compare on.
| Property | Bridge Truss (Pratt/Warren) | Plate Girder | Cable-Stayed Bridge |
|---|---|---|---|
| Economical span range | 30-200 m | 20-90 m | 150-1100 m |
| Span-to-depth ratio | 6 to 12 (deep) | 20 to 28 (shallow) | varies — pylon-driven |
| Steel tonnage per m² of deck (typical) | 180-280 kg/m² | 220-350 kg/m² | 120-220 kg/m² + cable steel |
| Fabrication complexity | High — many connections, gusset plates | Moderate — long welded girders | Very high — pylon, cable anchorages |
| Field erection time (50 m span) | 3-6 weeks | 1-2 weeks | Not applicable at 50 m |
| Fatigue category at typical details | AASHTO C or C′ at gussets | AASHTO B at web-flange welds | AASHTO B/C at cable anchorages |
| Service life with normal inspection | 75-120 years | 75-100 years | 75-120 years |
| Inspection burden | High — every gusset, every pin | Low to moderate | High — cables and anchorages |
Frequently Asked Questions About Bridge Truss
Because uniform-load hand calculations underestimate live-load shear at the supports. End diagonals carry the full reaction shear divided by sin θ of the diagonal, and a concentrated truck axle near the bearing produces shear far higher than (w × L)/2. AASHTO HL-93 design uses an influence-line approach precisely to capture this.
If the discrepancy is more than about 15%, check whether you analysed the truss with the design truck positioned for maximum end-panel shear rather than maximum midspan moment — they are different load positions and they govern different members.
Warren trusses use fewer members — diagonals only, no verticals in the basic form — so they fabricate faster and look cleaner. Pratt trusses add verticals, which gives you intermediate panel points to hang floor beams from and shortens the unbraced length of the compression chord. For a 50 m span carrying a deck on floor beams, Pratt is usually the lower-tonnage answer because the shorter compression chord panels let you drop a section size on the top chord.
Warren wins when the deck is direct-fixed to the top chord (deck truss) and you don't need intermediate panel points, or when architectural cleanliness matters — pedestrian bridges in particular.
Yes, on a fatigue-loaded bridge it is. Slip-critical connections are designed to transfer load through friction at the faying surfaces; once you slip into bearing, the bolt shanks see direct shear and the holes start to ovalise under cyclic traffic. AASHTO requires slip-critical detailing on most truss connections precisely to avoid this.
The usual cause is contaminated faying surfaces — paint, grease, or mill scale where Class B blast-cleaned surface was specified — or under-tensioned bolts. Pull a sample bolt and check tension with a Skidmore-Wilhelm; if it is below the AASHTO minimum pretension, retorque the whole connection.
Pack rust expands to roughly 6-10× the volume of the parent steel, prying connection plates apart and putting the bolts into prying tension they were never designed for. By the time visible plate separation reaches 3 mm, the bolt tension demand has typically doubled, and the gusset-to-chord interface is no longer behaving as a rigid panel point.
The I-35W collapse in 2007 traced back to undersized gusset plates compounded by decades of pack rust at the L11 node. If your inspection report flags any measurable plate separation, get a structural engineer onto it before you defer to the next cycle — the failure mode is non-linear and accelerates rapidly past 3 mm.
Hand and even FEA calculations assume rigid panel-point connections and zero slip in bolted joints. Real trusses have small connection rotations and a measurable contribution from gusset-plate flexibility — together these typically add 10-20% to predicted midspan deflection on a bolted truss.
If you are seeing more than 25% extra deflection, check three things: bolt slip at splices (run a load-unload cycle and look for hysteresis), bearing seizure or rocker-bearing freeze-up at the abutments which transforms a simple span into a partial fixed end, and whether the lateral bracing is engaging — a slack X-brace lets the compression chord sway and adds vertical deflection.
Rarely without strengthening the truss itself. Widening the deck increases dead load and live load proportionally, and the chord forces scale linearly with that load. A 25% deck widening typically pushes chord and end-diagonal forces 25-30% higher, which usually exceeds the original design margin once you also account for HL-93 being heavier than the original H15 or H20 design loading the truss was sized for.
The realistic options are: post the bridge for reduced loading, add cover plates or supplementary members to the governing chord and diagonals (common on 1930s-50s trusses), or replace the superstructure entirely. A load rating per AASHTO MBE is the first step — never widen without one.
For Pratt and Warren highway trusses in the 40-80 m range, total steel tonnage minimises around L/d = 8 to 10. Below L/d = 6 the truss is too deep — chord forces drop but web members and bracing grow faster than the chord savings. Above L/d = 12 the chord forces climb steeply (they scale as 1/d) and you're forced into expensive built-up sections.
The 48 m / 6.0 m example in the worked problem above sits at L/d = 8, which is why it lands cleanly on a rolled W14 section without built-up plates. If you have clearance constraints forcing you above L/d = 12, run the numbers carefully — you may find a plate girder is actually lighter at that depth.
References & Further Reading
- Wikipedia contributors. Truss bridge. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
