A Melan bridge is a reinforced concrete arch in which rolled steel I-beam ribs, curved to the arch profile, are embedded in the concrete and act as both primary reinforcement and self-supporting formwork during the pour. It solves the falsework problem on short-to-medium span arches — the steel ribs carry the wet concrete and traffic of the pour, so crews skip the timber centring across the gap below. Joseph Melan patented the system in 1892, and it produced thousands of road and rail arches across Europe and North America in spans up to roughly 60 m.
Melan Bridge Interactive Calculator
Vary span, rise, load, and composite thrust share to see arch thrust, shoe force, and how load shifts from embedded steel ribs into cured concrete.
Equation Used
This calculator uses the standard parabolic-arch thrust estimate for a uniformly loaded Melan arch. The horizontal thrust H rises with load and span squared, and falls as arch rise increases. The selected composite share eta splits the cured thrust between concrete compression and embedded steel ribs.
- Parabolic arch with uniform vertical load per metre of bridge width.
- Horizontal thrust is estimated at the crown/springing line.
- After cure, concrete carries the selected share of compressive thrust.
- Construction staging effects, live-load asymmetry, and rib buckling are not included.
Operating Principle of the Melan Bridge
The Melan system is a composite arch. You roll or fabricate steel I-beam ribs to match the arch curve, set them on the abutments at design spacing — typically 0.6 to 1.2 m on centre across the deck width — then pour reinforced concrete around them. While the concrete is green, the steel ribs alone carry the dead load of the wet pour plus formwork plus workers. Once the concrete cures and bonds to the steel, the section behaves as a composite arch, with the concrete carrying the bulk of compressive thrust and the steel handling tension and bending from live loads, thermal effects, and unsymmetric loading.
The design is centring-free for a reason. On a 30 m masonry or pure-concrete arch you would normally build heavy timber falsework spanning the gap, support the wet concrete on it, then strip the timber after cure. Over a flooded river valley, a rail cut, or a deep ravine, that falsework costs more than the bridge. Embedded steel arch ribs let you skip it entirely — the ribs span the gap on day one and the concrete is poured against them.
Tolerances matter at the rib geometry stage. The rib profile must match the design arch axis within roughly ±10 mm over a 30 m span, otherwise the dead-load deflection of the bare steel under wet concrete will not match the predicted shape and you build in a permanent kink. Rib spacing variation above 5% of nominal causes uneven concrete cover and creates the classic Melan failure mode — corrosion of the embedded flanges where cover drops below 40 mm, followed by spalling of the soffit concrete decades later. Most of the surviving Melan bridges showing distress today fail through this rust-jacking mechanism, not through structural overload.
Key Components
- Steel arch ribs: Rolled I-beams or built-up plate girders curved to the arch profile, typically spaced 0.6 to 1.2 m on centre across the bridge width. Section depth runs roughly 1/60 to 1/80 of the span. The ribs alone must carry wet-concrete dead load plus a 2.4 kPa construction live load without yielding.
- Concrete arch barrel: The cast-in-place concrete that surrounds the ribs and forms the structural arch. Specified at 25 to 35 MPa for most historical Melan work, with minimum 50 mm cover over the embedded steel. The concrete carries roughly 80 to 90% of the compressive thrust at the crown.
- Abutment shoes: Cast or fabricated steel bearings at the springing points that anchor the rib feet into the masonry or concrete abutment. They must transfer the full arch thrust — at the crown of a 30 m arch carrying H20 loading, that thrust runs around 2 to 4 MN per metre of width.
- Spandrel fill or columns: The structure between the extrados of the arch and the underside of the deck. Early Melan bridges used earth or rubble spandrel fill; later ones used open spandrel columns to reduce dead load on longer spans.
- Deck slab: The riding surface and load distribution plate above the spandrels. Cast monolithically with the spandrel walls or columns and tied to the arch barrel via the spandrel reinforcement.
Who Uses the Melan Bridge
Melan bridges show up wherever a short-to-medium-span arch needed to be built across a gap that made conventional centring expensive or dangerous. The system dominated North American and European bridge construction from about 1894 through the 1930s, then faded as prestressed concrete and welded steel girders took over the same span range. You still see thousands of them in service today on secondary roads, rail lines, and park infrastructure.
- Highway bridges: The Walnut Lane Bridge in Philadelphia (1908) used the Melan system for its approach arches before the main prestressed span was added decades later.
- Railway bridges: Austrian and Czech rail authorities built hundreds of Melan-system arch overpasses in spans of 12 to 25 m between 1895 and 1920, many still carrying full ÖBB freight loading today.
- Park and pedestrian crossings: The Topeka Avenue Bridge in Topeka, Kansas (1897) — one of the first Melan bridges in the United States, designed by Edwin Thacher under license from Joseph Melan.
- Aqueducts and canal crossings: Several Melan-system aqueduct arches on the Trent–Severn Waterway and similar early-20th-century water-management infrastructure across Ontario.
- Municipal road network: The Eden Park Melan Arch Bridge in Cincinnati, Ohio (1894), built by the Melan Arch Construction Company and considered one of the earliest American examples — a 21 m clear span still standing.
- Rural farm and forestry crossings: Numerous 8 to 15 m Melan arches built across creeks in the upper Midwest and Pennsylvania in the 1900s and 1910s, where mobile cranes and falsework couldn't reach the gap.
The Formula Behind the Melan Bridge
The defining check on a Melan arch is whether the bare steel ribs alone can carry the wet-concrete pour without yielding or buckling — because once you commit to skipping centring, those ribs are the only thing holding up the structure for the 7 to 14 days before the concrete reaches design strength. The required section modulus depends on span, rib spacing, concrete thickness, and pour-stage live load. At the low end of the typical span range (around 12 m) the ribs are loafing — you can use light rolled sections like W8 or W10 shapes. At the high end (around 50 to 60 m) the ribs become the dominant cost driver and you start needing built-up plate girders. The sweet spot for the system historically sat around 20 to 30 m, where rolled mill sections still worked but the centring-savings were already large.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Sreq | Required section modulus of one steel rib during the pour stage | mm³ | in³ |
| wwet | Wet-concrete plus formwork plus construction live load per rib, per unit length along the arch | N/m | lb/ft |
| L | Effective bending span of the rib between intermediate supports or springing points during the pour | m | ft |
| σallow | Allowable bending stress in the bare steel rib during construction (typically 0.6 × F<sub>y</sub>) | MPa | ksi |
Worked Example: Melan Bridge in a small-municipality Melan road arch
A rural municipality in southwestern Pennsylvania is rehabilitating a creek crossing on a township road and the engineer of record specs a 24 m clear-span Melan arch with steel ribs at 0.9 m on centre, a 350 mm thick concrete arch barrel at the crown thickening to 550 mm at the springing, and 30 MPa concrete. You need to size the rib section modulus so the bare steel can carry the wet-concrete pour across the full 24 m before any concrete strength develops.
Given
- Span L = 24 m
- Rib spacing = 0.9 m
- Concrete thickness (avg) = 0.45 m
- Concrete unit weight = 24 kN/m³
- Construction live load = 2.4 kPa
- Steel Fy = 250 MPa
- σallow = 0.6 × Fy = 150 MPa
Solution
Step 1 — compute the wet-concrete dead load per rib per unit length. Tributary width per rib is 0.9 m, average concrete thickness is 0.45 m, unit weight is 24 kN/m³:
Step 2 — add construction live load (workers, vibrators, hoses) at 2.4 kPa over the 0.9 m tributary width:
Step 3 — total wet-stage load per rib, then required section modulus at the nominal 24 m span. Note this is a simplified bending check; in real design you also account for arch action of the bare ribs, but at this stage we assume the worst case where the rib spans as a beam between springings:
That is a heavy section — roughly a W36×210 or a built-up plate girder. Now check the operating range. At the low end of the typical Melan span range, 12 m, the same load gives Sreq = (11,880 × 144) / (8 × 150 × 106) ≈ 1,425 cm³, which a W18×50 covers comfortably — the system feels effortless and the centring savings dominate. At the high end, 48 m, Sreq jumps to roughly 22,800 cm³ which no rolled section reaches; you must move to a deep built-up plate girder, the rib steel weight per metre of bridge climbs above 1 tonne, and the economic argument against a steel-only bridge starts to disappear.
Result
At the nominal 24 m span the required rib section modulus is roughly 5,700 cm³ — a heavy rolled W36 or a modest built-up girder, achievable but not trivial. At 12 m the same loading needs only 1,425 cm³ and the system is in its sweet spot; at 48 m the demand quadruples to 22,800 cm³ and rolled sections run out, which is exactly why historical Melan bridges above 50 m are rare. If your physical pour shows a measured midspan deflection more than 15% above the predicted bare-rib deflection, the most common causes are: (1) rib bracing not installed between adjacent ribs so they buckle laterally under the wet load, (2) concrete placed faster than the assumed rate so wet-state load concentrates near the crown, or (3) rib camber fabricated short of the design value, which compounds with elastic deflection and produces a permanent sag in the cured arch.
Choosing the Melan Bridge: Pros and Cons
The Melan system competes with a few alternatives in the short-to-medium-span arch bridge market. The decision usually comes down to span, site access for falsework, and whether you need centring-free construction. Here is how it stacks up against a conventional reinforced concrete arch built on falsework and a modern precast concrete arch system.
| Property | Melan bridge | Falsework-built RC arch | Precast concrete arch (e.g. BEBO, CON/SPAN) |
|---|---|---|---|
| Practical span range | 12–60 m | 8–90 m | 3–25 m |
| Centring/falsework requirement | None — ribs self-support | Full timber or steel falsework across span | None — segments crane-lifted into place |
| Construction time for a 24 m span | 6–10 weeks | 10–16 weeks | 2–4 weeks |
| Steel content (kg per m² of deck) | 180–280 kg/m² | 60–120 kg/m² | 40–80 kg/m² (precast cage) |
| Service life (well-detailed example) | 100+ years if cover maintained | 100+ years | 75–100 years |
| Dominant failure mode | Rust-jacking of embedded ribs where cover < 40 mm | Reinforcement corrosion at deck joints | Joint sealant failure between segments |
| Suitability today | Niche — heritage and replication work | Common for long custom arches | Standard for short culvert-class arches |
Frequently Asked Questions About Melan Bridge
Almost always lateral-torsional buckling of the ribs, not flexural under-design. The bare I-beam ribs are slender about their weak axis until concrete sets, and if you didn't install lateral bracing between adjacent ribs at roughly L/8 spacing, the compression flange rolls sideways under load and the rib deflects vertically more than pure beam theory predicts.
Quick diagnostic — sight along the top flanges during the pour. If you see a noticeable S-curve in plan, you've got LTB and need to add cross-bracing immediately. The historical Melan specifications called for transverse tie rods or angle bracing every 2 to 3 m along the arch precisely for this reason.
For a one-off 15 m crossing with truck access, precast wins almost every time. A BEBO or CON/SPAN unit ships in 2 to 6 segments, lifts in with a 50-tonne crane, and you're backfilling within a week. A Melan arch at the same span needs custom rib fabrication, rib erection over the gap, formwork for the spandrels, a full concrete pour, and a 28-day cure before backfill.
Melan only makes sense at 15 m if (a) you cannot get a crane to the site, (b) you're matching an existing heritage arch on the same road, or (c) the span is one of several similar arches and rib fabrication amortises across the project.
That's the classic Melan failure signature — rust-jacking. The bottom flange of each embedded rib has corroded, and as the rust expanded to roughly 6 to 10 times the volume of the parent steel, it cracked the concrete cover off in a strip directly below the flange. You see horizontal spalled stripes at the rib spacing, often 0.6 to 1.2 m apart, with rust staining and exposed steel.
Root cause is almost always insufficient cover at the soffit — the original detailers gave 25 to 30 mm where modern code wants 50 to 60 mm in a wet exposure. Repair means hydrodemolition of the loose concrete, mechanical cleaning of the rib steel back to bright metal, a corrosion-inhibiting primer, and sprayed concrete to restore cover. Patching without addressing the cover deficiency just buys you 10 years.
You should do composite section analysis, but the answer depends on what you're checking. For ultimate compressive thrust at the crown, modelling as plain concrete with the rib area as additional reinforcement gets you within 5 to 8% of the composite answer because the concrete carries 85 to 90% of the thrust anyway.
For live-load bending — particularly unsymmetric loading like a single truck on one half of the arch — you must use the composite section. The steel ribs carry a disproportionate share of the bending moment because they sit at the extreme fibres and have an E roughly 7 times that of the concrete. Treating the section as plain concrete here under-predicts steel stress by 40 to 60% and you will miss a fatigue check on a rail-loaded arch.
Three reasons combined to kill the system around 1940. First, prestressed concrete from Freyssinet's work in the 1930s let you build the same 20 to 40 m span with 30 to 40% less material and no embedded steel corrosion problem. Second, welded plate girders and rolled wide-flange sections got cheaper and longer, so a steel girder bridge beat a Melan arch on first cost above about 25 m. Third, AASHTO and Eurocode load levels climbed — a 1905 Melan arch was sized for an H10 truck of about 9 tonnes, and updating the system to modern HL-93 loading erases the steel-savings that made it attractive.
The system still gets specified occasionally for heritage replications and park bridges where the look matters, but as a routine engineering choice it lost on cost and durability decades ago.
Hold the rib profile to ±10 mm over the full span and ±3 mm in any 3 m chord. The arch is shape-sensitive — a 25 mm error at the crown of a 30 m arch shifts the line of thrust enough to introduce secondary bending moments of 15 to 25% above design at the springings, which the rib alone has to carry until the concrete cures.
Practical rule of thumb: fabricate ribs with a deliberate upward camber of about L/500 to compensate for elastic deflection under wet-concrete load. If you fabricate to the theoretical arch line with no camber, the bare ribs sag under the pour and you cure in a permanent dip at the crown.
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
